article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the phase diagram of the high temperature superconducting cuprates has been under extensive investigation since their discovery , almost three decades ago.@xcite in addition to high temperature superconductivity , a variety of ground states have been proposed and partly realized upon charge carrier doping . for example , the parent compound la@xmath3cuo@xmath2 , which is an antiferromagnetic ( af ) mott insulator with nel temperature t@xmath7=325 k , @xcite is known to exhibit a short - range glassy phase @xcite and subsequently diagonal stripe order upon doping with e.g. sr ( la@xmath0sr@xmath1cuo@xmath2 ) .
@xcite recently , research has refocused on the highly underdoped cuprates.@xcite however the evolution of the electronic ground state with the very first added charge carriers , in particular in the search for a possible broken symmetry associated with an exotic ground state and subsequent effects on the anomalous normal state properties , remains an enigma . although earlier efforts had in fact suggested that ferroelectricity could be present in la@xmath3cuo@xmath5 and yba@xmath3cu@xmath8o@xmath9,@xcite a low temperature ferroelectric ( fe ) phase with an associated magnetoelectric ( me ) coupling
was recently observed in highly underdoped la@xmath3cuo@xmath5 ( t@xmath10 k).@xcite
notably , ferroelectricity is characterized by broken spatial inversion symmetry and typically emerges in the absence of mobile charge carriers . in the case of lightly oxygen doped la@xmath3cuo@xmath5 ( t@xmath7 = 320
k ) it was proposed that the non - stoichiometric oxygen ions take interstitial positions in the la@xmath3cuo@xmath2 unit cell , causing a displacement in the apical oxygen ions of the cuo@xmath6 octahedra , which are the building blocks of the la@xmath3cuo@xmath2 unit cell @xcite . therefore , local - scale structural cuo@xmath6 distortions take place , breaking the spatial inversion symmetry , resulting in the formation of local electric dipoles.@xcite these dipoles are localized around the oxygen interstitials forming charge clusters , which couple and freeze at low temperature , giving rise to an observable polar state .
other microscopic mechanisms have also been proposed .
these include a model based on polaron formation around the dopant atoms , @xcite a purely magnetic origin associated with dzyaloshinskii - moriya ( dm ) interaction leading to local structural distortions and concomitant local broken inversion symmetry , @xcite and a proposal for the emergence of electric polarization due to the formation of magnetic vortex / anti - vortex pairs at the edges of oriented stripe segments .
@xcite the presence of the me effect in la@xmath3cuo@xmath5 @xcite has also been studied by landau theory , which includes a bi - quadratic coupling between the electric polarization and the magnetic order.@xcite at present , there is no consensus on the origin of the fe phase in these materials .
the question arising therefore is whether ferroelectricity is present in la@xmath3cuo@xmath2 when different dopant ions are introduced into the lattice . to shed further light on the nature of ferroelectricity and magnetoelectricity in la@xmath3cuo@xmath2 system ,
it is important to study other members of the la-214 cuprate family , especially the case where charge carriers originate from sr and li doping .
both of these ions occupy stoichiometric positions in the la@xmath3cuo@xmath2 unit cell and therefore , no dipole moments are directly associated with the dopant sites , in contrast to the case of interstitial excess oxygen ions in la@xmath3cuo@xmath5 .
@xcite here , we report measurements of the electric polarization on sr and li doped la@xmath3cuo@xmath2 single crystals . we show that la@xmath11sr@xmath12cuo@xmath2 exhibits distinct fe behavior along different crystallographic directions .
similar behavior is also shown in lightly oxygen doped la@xmath3cuo@xmath5 samples ( different samples than those investigated in ref . )
studied here for direct comparison to la@xmath11sr@xmath12cuo@xmath2 .
the magnetic field dependence of the electric polarization is anisotropic and it can be tuned through the dm interaction . in la@xmath3li@xmath1cu@xmath4o@xmath2 ( @xmath13 and @xmath14 ) ,
the measured electric polarization is in the order of @xmath15c @xmath16 i.e. , several times higher than for la@xmath11sr@xmath12cuo@xmath2 ( as shown below ) and la@xmath3cuo@xmath5.@xcite the electric polarization in la@xmath3li@xmath1cu@xmath4o@xmath2 is also tunable by an applied magnetic field similar to la@xmath11sr@xmath12cuo@xmath2 and la@xmath3cuo@xmath5 .
these results taken together demonstrate that the fe phase is present in all underdoped la-214 cuprates investigated so far , and is prompted by charge carrier doping .
we propose ferroelectricity may originate from a mechanism that breaks inversion symmetry by local structural distortions of the cuo@xmath6 octahedra due to the presence of the dopant ions and/or clustering of the added holes .
such local symmetry breaking induces electric dipoles which couple and order at low temperature .
the experiments were performed on twinned single crystals .
la@xmath11sr@xmath12cuo@xmath2 crystals were grown using the laser - diode heated floating zone method @xcite .
high purity la@xmath3o@xmath8 ( 99.99 % ) , srco@xmath8 ( 99.9 % ) and cuo ( 99.9 % ) were used as starting materials .
sr ions take la stoichiometric positions and thus the sr concentration ( x = 0.001 ) is higher than the purity of la@xmath3co@xmath8 . on the other hand ,
the sr concentration is comparable to the purity of cuo .
furthermore , the crystals were carefully annealed at 1000 c for 20 h under 1ppm o@xmath3 - ar flowing atmosphere .
the crystal axes were determined by x - ray laue backscattering technique .
the sharp antiferromagnetic transition at 312 k ( see supplementary s1 ) is suggestive of homogeneous distribution of the dopant ions ( either sr or excess oxygen ions that might be included in the unit cell ) into the la@xmath11sr@xmath12cuo@xmath2 samples as any inhomogeneous distribution of the ion dopants would cause either the suppression - broadening of the af peak at 312 k , or the appearance of secondary magnetic peaks at temperatures lower than 312 k. to overcome the difficulty in quantifying experimentally the very low sr concentration , we synthesized ( crystal growth and annealing ) oxygen doped la@xmath3cuo@xmath5 using the same method as la@xmath11sr@xmath12cuo@xmath2 .
@xcite the excellent agreement in the magnetic transition ( supplementary s1 ) , indicate similar excess oxygen doping in both materials , as the very small sr doping will not affect the magnetic transition temperature of la@xmath11sr@xmath12cuo@xmath2 significantly @xcite ( therefore la@xmath11sr@xmath12cuo@xmath17 is the appropriate chemical formula for the sr - doped la@xmath3cuo@xmath2 samples ) .
any possible difference in their transport properties therefore would be attributable to sr doping .
la@xmath3cu@xmath4li@xmath1o@xmath2 single crystals were grown using the conventional lamp heated floating zone technique .
@xcite high purity la@xmath3o@xmath8 ( 99.999 % ) , li@xmath3co@xmath8 ( 99.99 % ) and cuo ( 99.99 % ) were used as starting materials in order to achieve the lowest possible impurity level .
the crystal axes were again determined by x - ray laue backscattering technique .
the samples were annealed at 900 c for 48h , in flowing ar atmosphere .
the li concentration was estimated from x - ray diffraction measurements of the lattice constants .
@xcite all crystals were cut appropriately and pairs of plate - like samples were extracted .
each pair consists of a sample with the thinnest direction along the c - axis and another with the thinnest direction along the ab - plane .
the magnetization of the samples was measured using a commercial mpms quantum design squid magnetometer , in heraklion .
the la@xmath11sr@xmath12cuo@xmath17 crystals exhibit t@xmath7 = 312 k , whereas for the la@xmath3cuo@xmath5 t@xmath7 = 313 k. for la@xmath3li@xmath1cu@xmath4o@xmath2 crystals , t@xmath7 = 255 k ( @xmath18 = 0.01 ) and t@xmath7 = 150 k ( @xmath18 = 0.04 ) , respectively .
the impedance and loss of the samples were measured using an lcr meter , over a frequency range 21 hz - 2 mhz , and the dielectric permittivity was extracted , as described elsewhere.@xcite electric polarization measurements were performed using two different home - built experimental stations in heraklion and singapore employing the pyrocurrent technique .
details of the technique , the measurement protocol and the pyrocurrent data processing can be found in the supplementary ( sections s2 - s4 ) .
( red line ) and out - of - plane p@xmath19 ( black line ) electric polarization as a function of temperature for la@xmath11sr@xmath12cuo@xmath17 .
the corresponding pyroelectric current curves are shown .
b. polarization as a function of applied electric field ( p - e ) hysteresis loops for both p@xmath20 ( red solid circles ) and p@xmath19 ( black solid circles ) at 2 k. broken lines are a guide to the eye . ]
figure [ fig:1]a shows the real part of the in - plane dielectric permittivity @xmath21 for la@xmath22sr@xmath12cuo@xmath17 . at high temperatures , a step - like decrement is observed for all measured frequencies .
this decrement shifts to higher temperature with increasing frequency f , indicative of a common dielectric relaxation process . at low frequencies , an additional dielectric
peak develops which shifts to higher temperature and is suppressed with increasing f. such a peak in the dielectric permittivity can be attributed either to an intrinsic charge relaxor characterized by a diffused phase transition or to spurious effects arising from the electrical contacts such as extrinsic maxwell - wagner phenomena .
@xcite to reveal the intrinsic character of the dielectric peaks , measurements of the dielectric permittivity were repeated several times with renewed contacts , giving consistent experimental results in the temperature regime around the peaks .
thus the electrical contacts do not affect the measured dielectric permittivity .
also , both the temperature t@xmath23 where the dielectric peak occurs and the magnitude of the permittivity at the peak position @xmath24 can not be described by the empirical relations reported for pseudo - fe relaxor behavior .
@xcite nevertheless , such a relaxation process due to the slowing of polar clusters , can be described by the vogel - fulcher ( vf ) relation @xmath25 $ ] , where the characteristic freezing temperature t@xmath26 corresponds to the estimated temperature below which the polar clusters freeze .
@xcite notably , electric polarization may emerge below t@xmath26 in relaxor ferroelectrics .
@xcite figure 1b shows f as a function of t@xmath23 for la@xmath11sr@xmath12cuo@xmath17 .
we estimated the t@xmath27 = ( 7.3 @xmath28 0.3 ) k by fitting the experimental data against the vf relation .
similar analysis of the imaginary part @xcite of the out - of - plane dielectric permittivity @xmath29 gives a freezing temperature t@xmath30 = ( 8.6 @xmath28 0.5 ) k , as ( fig [ fig:1]c and fig [ fig:1]d ) .
therefore , the observed behavior could be due to a fe relaxor behavior , characterized by a diffused phase transition and the freezing of short - range cluster - like order .
the above data analysis also reveals an anisotropic behavior in the charge dynamics in agreement with an earlier report.@xcite figure [ fig:2 ] shows the in - plane ( p@xmath20 ) and the out - of - plane ( p@xmath19 ) electric polarization as a function of temperature for la@xmath11sr@xmath12cuo@xmath17 . for both directions ,
the polarization increases with decreasing temperature , below 10 k. p@xmath19 and p@xmath20 exhibit distinct temperature dependencies .
furthermore , corresponding pyroelectric curves ( fig [ fig:2]a inset ) show that the pyroelectric current local minima occur at different temperatures , namely 4 k and 6.5 k , for the in - plane and the out - of - plane directions , respectively .
these temperatures are also lower than the corresponding freezing temperatures extracted from vf analysis.@xcite notably , a similar anisotropy has been observed in the spin - glass temperature .
@xcite furthermore , the electric polarization curves reverse with reversing the polarity of the applied electric field . in proper fe s
the onset of ferroelectricity is defined as the temperature above which there is no measurable polarization and the corresponding pyroelectric current exhibits a sharp minimum . on the other hand , in relaxor fes
a transition occurs over a large time scale .
similar to our data , the pyroelectric minimum is not sharp , resulting in a broad transition in the polarization with respect to temperature , therefore preventing a precise assignment of a transition temperature .
we assign the transition towards a fe state as the temperature where a minimum ( albeit broad ) occurs in the pyroelectric current .
this assignment is corroborated by the fact that the temperature where the local minimum occurs is robust against both the electric field applied during cooling and the temperature sweep rate .
the p - e hysteresis loops for both p@xmath19 and p@xmath20 , are shown in fig [ fig:2]b ( t = 2 k ) .
the raw data yield a sizable anisotropic electric polarization , with the in - plane and out - of - plane values reaching p@xmath20 = 18 nc@xmath16 and p@xmath19 = 36 nc@xmath16 respectively . the dielectric anomaly and the relaxor behavior in the charge dynamics
indicate a dominant electronic contribution towards the polarization .
in addition , the observation of non - zero polarizations along both the c - axis and ab - plane implies that the electric polarization may arise from an ensemble of charge clusters with varying polarization vectors .
although the intrinsic nature of the different trends in our observations for la@xmath11sr@xmath12cuo@xmath17 ( t@xmath7 = 312 k , n @xmath31 10@xmath32 @xmath33),@xcite it is not clear whether they come from sr doping or from the excess oxygen ions included in the la@xmath3cuo@xmath2 unit cell ( as discussed before ) . to this point , we performed electric polarization measurements on lightly oxygen - doped la@xmath3cuo@xmath5 with t@xmath7 = 313 k ( n @xmath31 10@xmath32 @xmath33).@xcite both p@xmath20 and p@xmath19 increase ( fig [ fig:3 ] ) with decreasing temperature , however p@xmath19(t ) exhibits a different temperature dependence compared to p@xmath20(t ) .
namely , a change in slope is observed at 5 k and 3 k , for out - of - plane and in - plane directions respectively , coinciding with the corresponding pyrocurrent local minima .
these values are also lower than the charge freezing temperatures obtained from the vf analysis ( supplementary s5 ) . at t
= 2 k , p@xmath19 is almost equal to p@xmath20 .
though , both the remnant electric polarization ( p ( e = 0 ) ) and the coercive electric field ( e ( p=0 ) ) differ for the two measured orientations ( inset fig [ fig:3 ] ) . electric polarization appears stronger in the out - of - plane direction ( i.e. larger remnant electric polarization and coercive electric field compared to the ab - plane ) .
we note that a finite conductivity could lead to an expanded `` banana - like '' p - e loop.@xcite however , this is not the case in the p - e loops depicted in the inset of fig .
3 . electrical transport experiments in la@xmath3cuo@xmath2 single crystals @xcite
reveal that the in - plane conductivity is considerably larger than the out - of - plane counterpart .
such high conductivity could result in a relatively lossy dielectric , which can manifest itself as a `` chubby '' banana - like p - e curve , in the ab - direction .
however , in our case the in - plane p - e loop is slimmer than the out - of plane loop , indicating the intrinsic nature of both the enhanced out - of - plane remnant electric polarization and the coercive electric field .
( red line ) and out - of - plane p@xmath34 ( black line ) electric polarization as a function of temperature for lightly oxygen - doped la@xmath3cuo@xmath5 ( t@xmath7 = 313k ) .
polarization as a function of applied electric field ( p - e ) hysteresis loops for p@xmath20 ( red symbols ) and p@xmath19 ( black symbols ) at 2 k. the dash lines are a guide to the eye . ] the above experimental evidence indicates that sr doping contributes to the measured electric polarization .
the distinct dielectric permittivity , anisotropy , remnant polarization and coercive field between la@xmath3cuo@xmath5 and la@xmath11sr@xmath12cuo@xmath17 indicate distinguishing effects between oxygen and sr doping .
we also measured la@xmath35sr@xmath36cuo@xmath17 however the material is not sufficiently insulating to permit transport measurements of the electric polarization . hence , although possible excess oxygen may still affect the observations in la@xmath11sr@xmath12cuo@xmath17 , the above mentioned observations indicate a characteristic effect due to sr doping ( supplementary table s1 ) .
we now turn to the effect of the applied magnetic field .
the spins in the af phase of la-214 are weakly canted out - of - plane due to the presence of a finite dm interaction.@xcite spin canting causes a weak ferromagnetic ( wf ) moment along the c - axis in each cuo@xmath3 plane , although the opposite spin canting in alternate planes leads to a net cancelation of the wf moments . by applying an external magnetic field along the c - axis ,
the wf moments can be aligned in the same direction above a critical magnetic field h@xmath37 , which for our la@xmath11sr@xmath12cuo@xmath17 samples is 6 t ( supplementary section s6 ) .
figure [ fig:4]a shows p@xmath19 as a function of applied magnetic field ( here e@xmath38c and h@xmath38c ) . p@xmath19 decreases abruptly above h@xmath37 = 6 t and the pyroelectric current minimum ( inset in fig [ fig:4]a ) shifts to lower temperatures , indicating a suppression of the fe state as the material enters the wf state ( such abrupt changes are also observed for h@xmath38c and e@xmath39c ) . on the other hand , both p@xmath20 and
the pyrocurrent minimum ( e@xmath38ab and h@xmath38ab ) decrease smoothly with increasing h , as seen in fig [ fig:4]b ( similar changes are observed for e@xmath39ab and h@xmath38ab ) .
hence , the fe order in la@xmath11sr@xmath12cuo@xmath17 is coupled to the underlying af structure and is influenced by the dm interaction in a manner similar to the recent observations for la@xmath3cuo@xmath5 . @xcite the similarity in the fe state among la@xmath3cuo@xmath5 and la@xmath11sr@xmath12cuo@xmath17 cuprates is notable even though oxygen and sr doping do not affect the cuo@xmath6 structure in the same way ; while the excess oxygen take non - stoichiometric positions , sr ions substitute for la . because sr and oxygen dopants take positions outside the oxygen octahedra it is important to study the evolution of ferroelectricity in the la-214 family of cuprates using also li doping , which directly replaces cu ions in the cuo@xmath6 octahedra . among other dopants , such as mg and
zn which can be used to replace cu , li ions exhibit the largest ionic radius ( 0.76 ) compared to 0.74 for zn , 0.72 for mg and 0.73 for cu . furthermore , la@xmath3li@xmath1cu@xmath4o@xmath2 remains insulating up to 4 % , @xcite allowing pyrocurrent measurements even at relatively high dopings .
figure [ fig:5 ] shows the temperature dependence of p@xmath19 for la@xmath3li@xmath1cu@xmath4o@xmath2 ( @xmath18 = 0.01 and @xmath18 = 0.04 ) . in both cases the electric polarization increases with decreasing temperature , below @xmath319 k.
a slope change in the electric polarization occurs at 5 k for @xmath18 = 0.01 and at 3.5 k for @xmath18 = 0.04 , respectively . applying the vf analysis on the corresponding dielectric permittivity data ( supplementary section s7 )
, we find a freezing temperature t@xmath30 = ( 5 @xmath28 0.3 ) k for @xmath18 = 0.01 .
furthermore , t@xmath26 is comparable to the values reported by park _
@xcite suggesting charge glassiness may be associated with the onset of ferroelectricity in la-214 cuprates .
it was not possible to calculate the corresponding t@xmath26 for @xmath18 = 0.04 , since the dielectric permittivity peaks are too broad to define a characteristic peak position ( supplementary section s7 ) .
moreover , the polarization is remarkably large namely , 900 nc@xmath16 for @xmath18 = 0.01 and 800 nc@xmath16 for @xmath18 = 0.04 .
notably , there is no pyroelectric signal along the in - plane direction , in both li - doped samples .
this is probably due to the relatively high in - plane electric conductivity of the sample .
@xcite figure [ fig:6 ] shows the magnetic field dependence of p@xmath19 for both la@xmath3li@xmath1cu@xmath4o@xmath2 samples at t = 2 k ( e @xmath38 c and h @xmath38 c ) . for @xmath18 = 0.01
, p@xmath19 is enhanced below h@xmath37 = 6.5 t , increasing by almost 1.7 times its zero - field value ( fig [ fig:6]a ) before getting suppressed .
theoretical studies suggest this effect may arise due to a dm induced magnetoelectric coupling .
@xcite for @xmath18 = 0.04 , ( h@xmath37 = 4 t ) p@xmath19 is roughly constant up to h@xmath37 , while it decreases when the wf state sets in .
notably , the fe transition temperature is also suppressed above the wf transition for both doping levels ( inset fig [ fig:6]a and fig [ fig:6]b ) , similar to la@xmath11sr@xmath12cuo@xmath17 .
therefore , la@xmath3li@xmath1cu@xmath4o@xmath2 behaves qualitatively similar to la@xmath11sr@xmath12cuo@xmath17 and la@xmath3cuo@xmath5 indicating ferroelectricity is a common ground state in the la-214 cuprates with an associated magnetoelectricity tunable by dm interaction . as a function of applied magnetic field h at t=2 k for la@xmath11sr@xmath12cuo@xmath17 ( e @xmath38c and h @xmath38 c ) .
pyroelectric current as a function of temperature for various h. b. p@xmath20 as a function of h ( e @xmath38 ab , h @xmath38 ab ) .
pyroelectric current as a function of temperature for various h. ] according to the scenario proposed earlier , in la@xmath3cuo@xmath5 charge clusters give rise to ferroelectricity .
@xcite although we can not rule out alternative mechanisms for the emergence of ferroelectricity , our proposal is likely to apply also to la@xmath11sr@xmath12cuo@xmath17 and la@xmath3li@xmath1cu@xmath4o@xmath2 , since local - scale cuo@xmath6 octahedral distortions may cause the spatial inversion symmetry breaking resulting in the formation of electric dipoles .
these dipoles are in turn localized around the dopant ions creating polar clusters , which fluctuate and with decreasing temperature freeze resulting in a measurable ferroelectricity .
theoretical and experimental evidence for both sr and oxygen - doped la@xmath3cuo@xmath2 also indicate the presence of local octahedral distortions correlating to charge inhomogeneities .
@xcite cordero _ et al .
_ reported the development of a significant tilt mode associated with oxygen octahedral distortion below t = 10 k from anelastic spectroscopy measurements on nearly undoped la@xmath0sr@xmath1cuo@xmath2 crystals .
@xcite it was argued that these tilt modes are due to fluctuating low temperature tetragonal structures at domain walls of the low temperature orthorhombic lattice , and are coupled to hole clusters .
additionally , a non - centrosymmetric monoclinic distortion in the cuo@xmath6 octahedra in orthorhombic la@xmath0sr@xmath1cuo@xmath2 and la@xmath3cuo@xmath5 has also been observed .
@xcite it is therefore plausible that the observed ferroelectricity may be linked to local - scale distortions from the ideal octahedral structure .
we note however , further investigation is necessary to narrow down the possibilities for the physical mechanism driving our observations .
@xcite as a function of temperature for la@xmath3li@xmath1cu@xmath4o@xmath2 with @xmath18 = 0.01 ( black solid line ) and @xmath18 = 0.04 ( blue solid line ) .
polarization as a function of applied electric field ( p - e ) slim loops for @xmath18 = 0.01 ( upper inset ) and @xmath18 = 0.04 ( lower inset ) at 2 k are also shown .
broken lines are a guide to the eye . ] as a function of applied magnetic field h for a. la@xmath3li@xmath40cu@xmath41o@xmath2 .
the corresponding pyroelectric current is shown in the inset of the graph , for selected magnetic fields .
b. p@xmath34 vs. magnetic field for la@xmath3li@xmath42cu@xmath43o@xmath2 .
the corresponding pyroelectric signal is shown in the inset . in both cases
e @xmath38 c and h @xmath38 c. ]
we report evidence for the presence of fe and me coupling in underdoped la-214 cuprates , namely la@xmath3cuo@xmath5 , la@xmath11sr@xmath12cuo@xmath17 and la@xmath3li@xmath1cu@xmath4o@xmath2 ( @xmath18 = 0.01 and @xmath18 = 0.04 ) . for both sr and oxygen doped samples , we observed distinct electric polarization behavior along in - plane and out - of - plane crystallographic directions . in all cases the electric polarization
is strongly influenced by dm interaction . considering the above experimental evidence
, the fe order and its related magnetoelectricity appear to be a generic property of la-214 cuprates .
we propose that the observed ferroelectricity originates from the cuo@xmath6 octahedral distortions , occurring around the dopant ions and/or due to the formation of charge clustering , resulting to a local - scale spatial symmetry breaking symmetry .
the work in singapore was supported by the national research foundation , singapore , through a fellowship and grant nrf - crp4 - 2008 - 04 .
the work in greece was partially supported by the european union s seventh framework program ( fp7-regpot-2012 - 2013 - 1 ) under grant agreement n316165 .
b.m.a . acknowledges support from the lundbeckfond fellowship ( grant a9318 ) .
99 j. g. bednorz and k. a. m@xmath44ller z. phys .
b - condensed matter * 64 * , 189 ( 1986 ) ; m. k. wu , j. r. ashburn , c. j. torng , p. h. hor , r. l. meng , l. gao , z. j. huang , y. q. wang , and c. w. chu , phys .
* 58 * , 908 ( 1987 ) .
r. j. birgeneau , m. greven , m. a. kastner , y. s. lee , b. o. wells , y. endoh , k. yamada , and g. shirane , phys .
b * 59 * , 13788 ( 1999 ) . f. c. chou , n. r. belk , m. a. kastner , r. j. birgeneau , and a. aharony , phys .
lett . * 75 * , 2204 ( 1995 ) .
i. raicevi@xmath45 , j. jaroszynski , d. popovi@xmath45 , c. panagopoulos , and t. sasagawa , phys .
* 101 * , 177004 ( 2008 ) .
p. a. lee , n. nagaosa , and x .-
wen , rev .
mod . phys . * 78 * , 17 , ( 2006 ) .
g. r. jelbert , t. sasagawa , j. d. fletcher , t. park , j. d. thompson , and c. panagopoulos , phys .
b * 78 * , 132513 ( 2008 ) . c. panagopoulos , j. l. tallon , b. d. rainford , t. xiang , j. r. cooper , and c. a. scott , phys .
b * 66 * , 064501 ( 2002 ) .
m. matsuda , m. fujita , k. yamada , r. j. birgeneau , y. endoh , and g. shirane , phys .
b * 65 * , 134515 ( 2002 ) .
m. matsuda , y. s. lee , m. greven , m. a. kastner , r. j. birgeneau , k. yamada , y. endoh , p. boni , s .- h .
lee , s. wakimoto , and g. shirane , phys .
b. * 61 * , 4326 , ( 2000 ) .
b. keimer , n. belk , r. j. birgeneau , a. cassanho , c. y. chen , m. greven , m. a. kastner , a. aharony , y. endoh , r. w. erwin and g. shirane , phys . rev .
b * 46 * , 14034 , ( 1992 ) .
e. razzoli , g. drachuck , a. keren , m. radovic , n. c. plumb , j. chang , y .- b .
huang , h. ding , j. mesot and m. shi , phys .
lett . * 110 * , 047004 ( 2013 ) . y. peng , j. meng , d. mou , j. he , l. zhao , y. wu , g. liu , x.dong , s. he , j. zhang , x.wang , q. zhimin wang , s. zhang , f. yang , c. chen , z. xu , t.k .
lee and x.j .
zhou , nature comm .
* 4 * , 3459 ( 2013 ) .
d. mihailovi@xmath45 and i. poberaj , physica c * 185 * , 781 ( 1991 ) .
d. mihailovi@xmath45 , i poberaj and a. mertelj , phys . rev .
b. * 48 * , 16634 ( 1993 ) .
z. viskadourakis , i. radulov , a. p. petrovi@xmath45 , s. mukherjee , b. m. andersen , g. jelbert , n. s. headings , s. m. hayden , k. kiefer , s. landsgesell , d. n. argyriou and c. panagopoulos , phys rev .
b * 85 * , 214502 ( 2012 ) .
m. reehuis , c. ulrich , k. proke@xmath46 , a. gozar , g. blumberg , s. komiya , y. ando , p. pattison , and b. keimer , phys .
b. * 73 * , 144513 , ( 2006 ) .
f. cordero , c. r. grandini , g. cannelli , r. cantelli , f. trequattrini , and m. ferretti , phys .
b * 57 * , 8580 ( 1998 ) .
a. bussmann - holder , a. simon , h. keller , and a. r. bishop , euro .
* 101 * , 47004 ( 2013 ) .
g. seibold , r. s. markiewicz , and j. lorenzana , phys .
b * 83 * , 205108 ( 2011 ) .
g. seibold , m. capati , c. di castro , m. grilli , and j. lorenzana , phys .
b * 87 * , 035138 ( 2013 ) . s. mukherjee , b. m. andersen , z. viskadourakis , i. radulov , and c. panagopoulos , phys .
rev b * 85 * , 140405 ( 2012 ) .
s. mukherjee , b. m. andersen , z. viskadourakis , i. radulov , and c. panagopoulos , j. supercond .
novel magn .
* 26 * , 1649 ( 2013 ) .
t. ito , t. ushiyama , y. yanagisawa , y .
tomioka , i. shindo , and a. yanase , j. crystal growth * 363 * , 264 ( 2013 ) . c. niedermeyer , c. bernhard , t. blasius , a. golnik , a. moodenbaugh , and j. i. budnick , phys .
rev . lett . * 80 * , 3843 ( 1998 ) .
e. stilp , a. suter , t. prokscha , e. morenzoni , h. keller , b. m. wojek , h. luetkens , a. gozar , g. logvenov , and i. bo@xmath47ovi@xmath45 , phys .
b * 88 * , 064419 ( 2013 ) .
t. sasagawa , p. k. mang , o. p. vajk , a. kapitulnik , and m. greven , phys . rev .
b * 66 * , 184512 ( 2002 ) . c. c. wang , s. x. dou , solid state comm .
* 149 * , 2017 ( 2009 ) ; p. lunkenheimer , v. bobnar , a. v. pronin , a. i. ritus , a. a. volkov , and a. loidl , phys .
rev b * 66 * , 052105 ( 2002 ) .
r. pirc and r. blinc , phys .
b * 76 * , 020101 ( 2007 )
. l. e. cross , ferroelectrics * 76 * , 241 ( 1987 ) .
vogel - fulcher analysis can be applied on either the real or the imaginary part of the dielectric permittivity giving similar charge freezing temperatures . since the dielectric peaks of the real part of the out - of - plane dielectric permittivity are not observable , we analyzed the imaginary part of the dielectric data where the corresponding peaks can be identified .
a. n. lavrov , y. ando , s. komiya , and i. tsukada , phys .
lett . * 87 * , 017007 ( 2001 ) .
n. w. preyer , r. j. birgeneau , c. y. chen , d. r. gabbe , h. p. jenssen , m. a. kastner , p. j. picone , and t. thio , phys .
b * 39 * , 11563 ( 1989 ) .
j. f. scott , j. phys .
: cond matter * 20 * , 021001(2008 ) .
cheong , z. fisk , r. s. kwok , j. p. remeika , j. d. thomson , and g. gruner , phys .
b * 37 * , 5916 , ( 1988 ) .
t. thio , t. r. thurston , n. w. preyer , p. j. picone , m. a. kastner , h. p. jenssen , d. r. gabbe , c. y. chen , r. j. birgeneau and a. aharony , phys . rev .
b * 38 * , 905 ( 1988 ) ; t. thio , and a. aharony , phys . rev
. lett . * 73 * , 894 ( 1994 ) .
j. l. sarrao , d. p. young , z. fisk , e. g. moshopoulou , j. d. thompson , b. c. chakoumakos , and s. e. nagler , phys . rev .
b * 54 * , 12014 ( 1996 )
. t. t. park , z. nussinov , k. r. a hazzard , v. a. sidorov , a. v. balatsky , j. l. sarrao , s. w. cheong , m. f. hundley , j. s. lee , q. x. jia , and j. d. thompson , phys rev lett . * 94 * , 017002 ( 2005 ) . s. j. l. billinge , e. s. bo@xmath47 in , m. gutmann , and h. takagi , j. supercond .
novel magn .
* 13 * , 713 ( 2000 ) .
m. hcker , g. d. gu , j. m. tranquada , m. von zimmermann , h. h. klauss , n. j. curro , m. braden , and b. buchner , physica c * 460 * , 170 ( 2007 ) .
d. haskel , e. a. stern , f. dogan , and a. r. moodenbaugh , j. synchrotron rad .
* 8 * , 186 ( 2001 ) and references therein .
e. s. bo@xmath47 in , s. j. l. billinge , g. h. kwei and h. takagi , phys .
b * 59 * , 4445 ( 1999 ) .
e. s. bo@xmath47 in , g. h. kwei , h. takagi , and s. j. l. billinge , phys .
lett . * 84 * , 5856 ( 2000 ) . *
supplementary information * + z. viskadourakis , s. s. sunku , s. mukherjee , b. m. andersen , t. ito , t. sasagawa and c. panagopoulos + and r@xmath48 are the parasitic resistance components as described in the text .
b. pyrocurrent vs. time ( blue solid line ) .
a sharp negative peak is observed when the polarization of the sample vanishes .
the temperature change is represented by the solid black line . c. the integration of the pyrocurrent
peak results in the induced charge curve . dividing by the area of the capacitor we obtain the electric polarization of the sample . in the integration process
we consider the electric polarization above the ferroelectric transition temperature to be zero at the initial state . ]
figure [ fig : s1]a shows the measured out - of - plane magnetization for la@xmath11sr@xmath12cuo@xmath17 .
the peak observed at t = 312 k corresponds to the nel temperature .
a similar peak is observed for the in - plane component of the magnetization again at t = 312 k. figure [ fig : s1]b shows the out - of plane magnetization for la@xmath3cuo@xmath5 ( t@xmath7 = 313 k ) .
figure [ fig : s1]c and fig [ fig : s1]d show the out - of plane magnetization for the li - doped samples , @xmath18 = 0.01 and @xmath18 = 0.04 , respectively .
the corresponding magnetization is comparable to earlier studies .
to measure the electric polarization of our samples we employed the pyrocurrent method .
the sample is placed between two metallic plates to form a capacitor .
an electrometer and a voltage source are connected to form the electric circuit shown in fig [ fig : s2]a .
the capacitor is placed into a cryostat , where we can accurately control the temperature .
when the sample exceeds its ferroelectric transition temperature its electric polarization decreases abruptly to zero .
this induces a current , which is called pyroelectric current or pyrocurrent .
further to the pyrocurrent , resistive and capacitive parasitic electric signals are also included in the total measured current , which can be expressed by @xcite @xmath49 here v is the applied voltage , r and c are the resistance and the capacitance of the sample respectively , p is the electric polarization and a is the area of the capacitor .
the first term corresponds to the resistance of the sample , and can be neglected as long as the sample is highly resistive or when the measurement is performed in v = 0 bias voltage .
the second term relates to the charging current of the capacitor and is negligible for a zero voltage across the circuit .
the third term is related to the current that could be caused when the dielectric permittivity of the sample changes considerably with temperature and can be ignored when a zero bias voltage is applied .
the fourth term corresponds to the pyrocurrent component of the measured current .
it follows that pyroelectric current measurements should be performed under zero bias voltage v = 0 to ensure that the first three parasitic terms do not contribute to the measured current , and thus the measured current will be purely due to the change of the electric polarization with respect to temperature . in this case
we can write eq ( 1 ) as follows : @xmath50 hence , the pyrocurrent is proportional to the sweep rate of the sample s temperature .
thus we can measure pyrocurrent by keeping a constant sweep rate of the temperature ( fig [ fig : s2]b ) .
integrating the pyrocurrent curves , with respect to time gives a quantitative measure of the charge induced due to the depolarization of the sample ( fig [ fig : s2]c ) .
dividing by the area of the capacitor we obtain the electric polarization of the sample .
to measure accurately the electric polarization of our samples we employ the following protocol .
the measurement process begins with the sample well above its ferroelectric transition temperature ( paraelectric state ) .
an electric field is applied and the sample is cooled to 2 k. the electric field is removed to fulfill the conditions described in the former paragraph . the sample is then heated at 3 k / min while measuring time , temperature and current at a constant interval of 0.5 seconds .
we continue to heat the sample well above its ferroelectric transition temperature . in the case of measuring the pyrocurrent in the presence of an applied magnetic field , we apply the magnetic field as soon as the sample reaches at 2 k and before removing the applied electric field ( zero magnetic field cooling ) .
a typical measured current - vs .- time data set is shown in fig [ fig : s3]a . for t
@xmath51 7 k the current increases steeply with increasing temperature , indicating the conductive character of the sample .
although v = 0 during the measurement process , a small but unavoidable voltage arising from the voltage source is still applied to the circuit causing the observed resistive behavior - the level of this voltage depends on the current measurement range used in the current meter , and can be between @xmath15v and 200 mv .
however , in the temperature range 2 k - 6 k there is a local current minimum ( inset fig [ fig : s3]a ) and its sign is reversed ( fig [ fig : s3]b ) upon reversing the electric field during the cooling process .
furthermore , it becomes sharper with increasing the temperature sweep rate suggesting its pyroelectric nature .
thus it is important to distinguish between the pyrocurrent and the resistive components of the current in order to accurately determine the electric polarization of the sample . to this point
, we fit the experimental data above 8 k against a resistivity model.@xcite the best fit is subtracted from the experimental data and the resulting curve is the pure pyroelectric component ( fig [ fig : s3]c ) . beyond 250 seconds there is considerable noise due to the data processing .
thus , the pyrocurrent curve is integrated with respect to time in the range 0 - 250 seconds , and the electric polarization with respect to time is acquired .
we obtain p(t ) by converting time to temperature .
we analyzed the dielectric permittivity data for the lightly oxygen - doped la@xmath3cuo@xmath5 ( t@xmath7 = 313 k ) ( fig [ fig : s4 ] ) .
the out - of - plane component of the permittivity @xmath54c ( fig [ fig : s4]a ) exhibits a step - like decrement ( at t @xmath51 60 k ) with decreasing temperature and shifts to higher temperature with increasing frequency indicating a common dipolar relaxation process .
furthermore , a dielectric peak is observed at t @xmath31 20 k. this peak becomes smoother and shifts to higher temperature with increasing frequency , suggesting a charge relaxor - like similar to la@xmath11sr@xmath12cuo@xmath17 .
figure [ fig : s4]b shows the frequency f as a function of peak temperature t@xmath23 extracted from fig [ fig : s4]a .
application of the vf model gives a freezing temperature t@xmath30 = ( 8 @xmath28 0.5 ) k. similar analysis was performed for the in - plane component of the dielectric permittivity ( fig [ fig : s4]c ) and the calculated freezing temperature was found to be t@xmath27 = ( 5.2 @xmath28 0.4 ) k ( fig [ fig : s4]d ) .
figure [ fig : s5 ] shows the magnetic field dependence of the normalized in - plane ( fig [ fig : s5]b ) and out - of - plane ( fig [ fig : s5]a ) resistivity of la@xmath22sr@xmath12cuo@xmath17 determined from measurements of electric impedance at t = 20 k. for h @xmath38 c ( fig [ fig : s5]a ) we observe a first order phase transition and a corresponding hysteresis at h @xmath31 6 t due to the metamagnetic transition associated with dm interactions . in la@xmath3cuo@xmath5 , the crystal anisotropy and the dm interaction fix the easy axis for the spins to the longer of the two in - plane orthorhombic directions ( the b - axis ) .
the direction of the ( wf ) moments @xmath55 induced by the dm interaction is fixed by the cross product @xmath56 between the dm vector @xmath57 ( oriented along the shorter of the two in - plane orthorhombic directions - the a - axis ) and the af order parameter @xmath58 ( pointing along the b - axis ) so that @xmath55 is oriented along the c - axis , perpendicular to the cuo@xmath3 planes of the crystal structure .
a sufficiently large magnetic field applied along the c - axis can overcome the inter - plane af coupling and induce a discontinuous spin - flop reorientation , causing the so - called wf ( first order ) phase transition .
the critical field is reduced at high temperatures following the decrease in @xmath55 , due to thermal fluctuations in @xmath58(t ) [ 45 ] . for h@xmath39ab
( fig [ fig : s5]b ) , the magnetoresistance varies smoothly because the weak ferromagnetic moments induce a continuous rotation of @xmath58 in the bc - plane .
similar results are obtained for the li doped samples , as shown in fig [ fig : s5]c and fig [ fig : s5]d .
( @xmath33 ) & dielectric permittivity ( @ 1khz ) & t@xmath59 ( k ) & p@xmath19 ( @ 2 k ) + la@xmath3cuo@xmath5 ( t@xmath7=320 k ) & 10@xmath60 & 2000 ( 50 k - out of plane ) & 4.5 & 33 + prb ( 2012 ) & & 3000 ( 50 k - in plane ) & 4.5 & 26 + la@xmath3cuo@xmath5 ( t@xmath7=313k ) & 10@xmath32 & 1200 ( 100 k - out of plane ) & 5 & 30 + present study & & 1300 ( 50 k - in plane ) & 3.5 & 27 + la@xmath11sr@xmath12cuo@xmath2 ( t@xmath7= 312 k ) & 10@xmath32 & 70 k ( 100 k - out of plane ) & 6.5 & 36 + present study & & 290 ( 50 k - in plane ) & 4 & 18 +
similar to la@xmath22sr@xmath12cuo@xmath17 , we also analyzed the dielectric permittivity for la@xmath3li@xmath40cu@xmath41o@xmath2 ( fig [ fig : s6]a ) .
dielectric peaks are observed in the low temperature regime , which are suppressed and move to higher temperature with increasing frequency .
the intrinsic origin of the peaks was confirmed as in the case of la@xmath11sr@xmath12cuo@xmath61 .
vf fits give t@xmath30 = ( 5.0 @xmath28 0.3 ) k ( fig [ fig : s6]b ) .
correspondingly , measurements of the electric polarization discussed in the main text give a fe transition temperature t@xmath59 = 5 k. for la@xmath3li@xmath42cu@xmath43o@xmath2 , we observe step - like changes in the out - of - plane dielectric permittivity , which are suppressed and move to higher temperature with increasing frequency ( fig [ fig : s6]c ) .
such features correspond to a normal dielectric relaxation process . on the other hand , there are no observable peaks corresponding to a relaxor - fe behavior . for t @xmath51 316
hz we observe an additional small change in slope below the step - like transition ( fig [ fig : s6]d ) .
this feature shifts to higher temperature with increasing frequency and vanishes for f @xmath51 10 khz . repeating the experiments several times with renewed contacts resulted to curves with similar features , indicating their intrinsic character .
we can not extract a corresponding f vs.t@xmath23 diagram since it is not possible to define the t@xmath23 hence , we can not apply the vf model to determine a freezing temperature .
+ 1 t. sasagawa , p. k. mang , o. p. vajk , a. kapitulnik , and m. greven , phys .
b * 66 * , 184512 ( 2002 ) .
a. n. lavrov , y. ando , s. komiya , and i. tsukada , phys .
lett . * 87 * , 017007 ( 2001 ) .
r. p. chaudhury , f. yen , b. lorenz , y. y. sun , l. n. bezmaternykh , v. l. temerov , and c. w. chu , phys .
b * 80 * , 104424 ( 2009 ) .
the variable range hopping mechanism of conduction @xmath62 is the most suitable mechanism to explain the hopping conduction in insulating antiferromagnets .
l. benfatto and m. b. silva neto , phys .
b * 74 * , 024415 ( 2006 ) .
z. viskadourakis , i. radulov , a. p. petrovi@xmath45 , s. mukherjee , b. m. andersen , g. jelbert , n. s. headings , s. m. hayden , k. kiefer , s. landsgesell , d. n. argyriou and c. panagopoulos , phys .
b * 85 * , 214502 ( 2012 ) . | we report on the observation of ferroelectricity and magnetoelectricity in la@xmath0sr@xmath1cuo@xmath2 and la@xmath3li@xmath1cu@xmath4o@xmath2 .
combined with earlier works on la@xmath3cuo@xmath5 ( z. viskadourakis _ et al .
_ , phys rev .
b * 85 * , 214502 ( 2012 ) ) these findings establish ferroelectricity as a new generic property in underdoped la-214 cuprates .
furthermore , the measured electric polarization can be tuned by dzyaloshiskii - moriya interaction resulting in a magnetoelectric effect .
it is proposed that ferroelectricity results from local cuo@xmath6 octahedral distortions , associated with the dopant atoms and/or clustering of the doped charge carriers , which break spatial inversion symmetry at the local scale . |
the discovery of the acceleration of the universe through the sneia hubble diagram @xcite , recently awarded by the nobel prize , has been latter confirmed by wide range of data , from more recent sneia data to baos and cmbr anisotropies @xcite . on the other hand
, such overwhelming abundance of observational evidences in favour of the cosmic speed up does not fit in the framework of general relativity ( gr ) making clear that out theoretical background is seriously flawed . the naive interpretation that an unexpected new ingredient in the form of a negative pressured component is driving the acceleration poses difficult questions on its nature and
nurture and introduces further problems hard to be solved .
it has therefore gained more and more attraction the hypothesis that the cosmic speed up is rather the first signal of a breakdown of our understanding of the laws of gravity on cosmological scales .
motivated by this consideration , much attention has been dedicated to @xmath2 theories @xcite where the scalar curvature @xmath3 is replaced by a suitably chosen function @xmath2 in the gravity lagrangian .
looking for a correction to gr , it is instructive to remember that an equivalent formulation is represented by teleparallelism . in this theory ,
torsion , instead of curvature , is responsible of the gravitational interaction @xcite and the weitzenbock connection replaces the levi - civita one on the underlying riemann - cartan spacetime . in this scenario , gravitational interaction is not replaced by geometry and the torsion acts as a force , allowing the interpretation of gravity as a gauge theory of the translation group @xcite . despite conceptual differences , teleparallel gravity and gr yield equivalent dynamics so that the interpretation of the gravitational interaction in terms of a curved or torsioned spacetime is only a matter of convenience , at least at the classical level . in as much the same way as for @xmath2 theories , one can obtain a generalised teleparallel gravity replacing @xmath4 with a generic function @xmath0 thus opening the way to a rich phenomenology .
a particular important consequence is the breakdown of the equivalence with the classical gr with the two theories now predicting a radically different dynamics @xcite .
modified teleparallel gravity preserves the advantage of giving equations that are still second order in field derivatives opposite to the fourth order equations deduced in @xmath2 gravity thus avoiding unpleasant pathologies . on the other hand ,
it suffers from the lack of local lorentz invariance ( lli ) so that all the 16 components of the vierbien are independent and one can not simply fix 6 of them by a gauge choice @xcite . a critical role in generalised teleparallel theories is played by the choice of the functional expression for @xmath0 .
the lack of firmly established theoretical constraints leaves open the way to a wide range of possibilities which can only be validated a posteriori , i.e. by contrasting their predictions with the observational data .
this is the aim of the present work where we focus our attention on two particular classes able to give rise to a phantom - like behaviour of the effective torsion fluid .
we then test these two models against sneia+grb hubble diagram , @xmath1 measurements from cosmic chronometers , baos data and the cmbr distance priors .
although wide , the present dataset only traces the background expansion so that we will also investigate whether further insight into the properties of these models can be obtained by the analysis of the growth factor being this latter a quick way to look at how perturbations evolve in the proposed modified teleparallel scenarios .
the layout of the paper is as follows . in sect.[f(t ) ] , we briefly review the cosmology of @xmath0 gravity and present the two models we are going to analyse . sect.[data ] is devoted to the likelihood analysis and the results , while sect.[growth ] enlarges the study to the growth factor .
finally , conclusions and perspectives are drawn in sect.[conclusions ] .
teleparallelism is a dynamical theory for the vierbein @xmath5 , whose components in a given coordinate basis @xmath6 help in introducing the metric as a subsidiary field : @xmath7 where @xmath8 .
the dynamics is then described by the action @xmath9 e d^4x } + { \cal{s}}_m \ , \label{eq : action}\ ] ] where @xmath10 and @xmath11 is the action for the matter fields . in eq.([eq : action ] ) , a key role is played by @xmath0 which is a differentiable function of the torsion tensor @xmath4 defined as @xmath12 with @xmath13 @xmath14 varying the action with respect to the vierbein components @xmath15 , one gets the field equations @xmath16 where @xmath17 is the matter energy - momentum tensor and subscripts @xmath4 denote differentation with respect to @xmath4 .
one can naively expect that , in order to get the cosmological counterpart of the field equations , one has simply to insert the robertson - walker ( rw ) metric into eqs.([eq : fieldeqs ] ) .
actually , some care is needed since , due to the lack of lli , two pairs of vierbein that lead to the same metric tensor are not equivalent from the point of view of the theory .
nevertheless , in case of spatially flat rw metric , a convenient choice is represented by the diagonal vierbein , i.e. @xcite @xmath18 where @xmath19 is the scale factor as function of cosmic time @xmath20 .
with such a choice , the dynamical equations become : @xmath21 while the torsion scalar reduces to @xmath22 with @xmath23 the usual hubble parameter and @xmath24 the energy density and pressure of the matter component .
note that hereafter we will denote with a prime and with a dot differentiation with respect to @xmath25 and @xmath20 , respectively .
the modified friedmann equations ( [ eq : ftfried ] ) can be rewritten in the usual form by introducing an effective dark torsion fluid with energy density reading @xmath26 since matter still minimally couples to gravity only , its conservation equation will be unaffected so that we still have @xmath27 and @xmath28 for the scaling laws of dust matter and radiation . imposing the bianchi identities and the diffeomorphism invariance of the theory , the conservation equation for the effective torsion fluid reads @xmath29 having set @xmath30 for the equation of state ( eos ) of the dark torsion fluid .
note the coupling to the radiation energy density through the term @xmath31 . for @xmath32 ,
one has @xmath33 and teleparallel gravity goes back to the standard gr , while the choice @xmath34 gives @xmath35 and the @xmath36cdm model is recovered .
moving away from the naive choice @xmath34 gives rise to rich phenomenology in the dynamical evolution of generalised teleparallel theories .
there are actually almost no theoretical hints on the functional form of @xmath0 with , on the contrary , many possible expressions leading to an accelerated expansion .
it is particularly interesting to look at models which are able to give rise to an effective eos ( defined later ) crossing the phantom divide , i.e. , @xmath37 for @xmath38 with @xmath39 the phantom divide redshift .
two recently proposed model of this kind can be obtained setting @xcite ( dot dashed blue ) , @xmath40 ( solid black ) , @xmath41 ( dashed red ) and the fiducial @xmath42 used in the text.,width=283 ] @xmath43 } \\
\end{array } \right .
\ , \label{eq : ftmods}\ ] ] where the subscript @xmath44 denotes present day quantities . in the first case , which we will refer to as the tanh model
, one should set @xmath45 in order to have a dark torsion fluid with a positive energy density , while the same condition leads to @xmath46 for the second choice , referred to hereafter as the exp model .
the constant @xmath47 can be expressed as a function of the other parameters by inserting eq.([eq : ftmods ] ) into the friedmann equations ( [ eq : ftfried ] ) and imposing @xmath48 .
this gives : @xmath49 for the tanh and exp model .
note that we have dropped the @xmath44 subscripts for the dust and radiation present day density parameters @xmath50 to simplify the notation .
plane imposing @xmath51 for the exp model with @xmath52 ( dot dashed blue ) , @xmath40 ( solid black ) , @xmath41 ( dashed red ) and the fiducial @xmath42 used in the text .
note that the lower set of contours are almost perfectly superimposed.,width=283 ] it is instructive to look at the effective torsion fluid eos @xmath53 for the two @xmath0 models we are considering .
we do not give here the lengthy expressions which can be obtained by inserting eqs.([eq : ftmods ] ) into the definition ( [ eq : fteos ] ) , but show how its present day value @xmath54 depend on the @xmath0 parameters in figs.[fig : wzefftanh ] and [ fig : wzeffexp ] for the tanh and exp model , respectively .
since the @xmath36cdm model is known to very well fit the data , it is reasonable to expect that , in order to fit the same data , the parameter space of both models will collapse in a region giving @xmath55 .
fig.[fig : wzefftanh ] shows that , for the tanh model , this will favour values of @xmath56 and that the larger is @xmath57 , the smaller @xmath58 must be in order to have a today effective eos close to the @xmath36cdm one .
this same requirement works in a different way for the exp model .
fig.[fig : wzeffexp ] , indeed , shows that , for given @xmath59 values , there can be more models with @xmath55 but different @xmath60 .
put in other words , the @xmath61 parameters are intrinsically degenerate so that the condition @xmath55 is unable to discriminate among different parameter sets .
actually , when fitting the data , we do not only look at the present day @xmath62 value , but also to its variation with @xmath63 . as we will see
, this helps breaking ( at least partially ) the degeneracy among @xmath64 for the tanh model and @xmath61 for the exp one .
figs.[fig : wzefftanh ] and [ fig : wzeffexp ] are nevertheless useful to quickly grasp what is the region of the parameter space of greater interest for both the tanh and exp @xmath0 models .
the above considerations on the effective eos suggest that the proposed generalised teleparallel scenarios can represent viable alternatives to the @xmath36cdm model . to check whether this is indeed the case , we contrast them with a wide set of cosmological data as described below . in order to answer the question whether @xmath0 gravity can reproduce the observed universe
, we will explore the model parameter spaces by investigating the following likelihood function @xmath65 where the set of model parameters is @xmath66 for the tanh and exp models , respectively , @xmath67 being the hubble constant @xmath68 in units of @xmath69 , and we have set the present day radiation density parameter as @xmath70 with @xmath71 in agreement with wmap7 @xcite constraints .
all the terms in the full likelihood can be written as : @xmath72}}{(2 \pi)^{{\cal{n}}_i/2 } \gamma_i^{1/2}}\ ] ] where @xmath73 depend on the dataset used .
the first term refers to the hubble diagram , i.e. the distance modulus @xmath74 as function of the redshift @xmath63 .
this is related to the underlying cosmological model as @xmath75 } \label{eq : defmu}\ ] ] with @xmath76 the luminosity distance and @xmath77 the comoving distance . as a tracer ,
we use both the union2 sneia dataset @xcite , comprising @xmath78 objects wiht @xmath79 , and @xmath80 grbs probing the redshift range @xmath81 with the distance modulus estimated in @xcite starting from the data in @xcite . for both datasets
, it is @xmath82 @xmath83 ^ 2 } \ , \ ] ] where @xmath84 is the measurement error for the @xmath85-th object , while @xmath86 takes care of the intrinsic scatter of the tracer around the relations used to estimate its @xmath74 value . for sneia , we set @xmath87 , while , for grbs , this is a nuisance parameter we marginalize over . while the hubble diagram probes the integrated expansion rate , the second dataset we use explicitly refers to the hubble parameter @xmath1 as determined through the differential age method @xcite . by using red envelope galaxies as cosmic chronometers @xcite , stern et al .
have estimated @xmath1 over the redshift range @xmath88 @xcite . to this sample ,
we add the @xmath68 determination from local distance ladders derived by the shoes collaboration @xcite and define a total @xmath89 as for the sneia one . while sneia and grbs probe the distance - redshift relation as standardizeable candles , baryon acoustic oscillations ( baos ) work as standard rulers .
we therefore add the term @xmath90 to the full likelihood following the method detailed in @xcite . to this end , three different sets of data are used .
first , one use the 6dfgrs @xcite and the sdss @xcite surveys to determine the scaled volume distance parameter @xmath91^{-\frac{1}{3 } } \ , \label{eq : defdz}\ ] ] with the sound horizon to distance @xmath63 given by @xmath92 and @xmath93 is the drag redshift .
the @xmath94 value at @xmath95 and its error is taken from @xcite , while @xcite gives @xmath94 for @xmath96 and @xmath97 with the corresponding covariance matrix .
further baos constraints come from the wigglez survey @xcite which recommend to use as observable quantity the acoustic parameter @xcite @xmath98 with the volume distance @xmath99 given in eq.([eq : defdz ] ) .
we use the observed values and their covariance matrix for @xmath100 determinations at @xmath101 reported in @xcite .
we finally define the bao likelihood as @xmath102 with @xmath103 ^ 2 \right \ } } \ , \ ] ] @xmath104 } \nonumber \ , \end{aligned}\ ] ] @xmath105 } \nonumber \ , \end{aligned}\ ] ] where @xmath106 is a @xmath107 dimensional vector with the difference between observed and predicted values and @xmath108 the corresponding covariance matrix .
.constraints on the parameters for the tanh model .
columns are as follows : 1 . )
i d parameter , 2 . )
best fit , 3 . ) mean , 4 . )
median , 5 . ) , 6 . )
@xmath109 and @xmath110 confidence levels . [ cols="^,^,^,^,^,^",options="header " , ] the best fit parameters and marginalized constraints for the tanh and exp model are given in tables [ tab : tanhfit ] and [ tab : expfit ] , while fig.[fig : bfplot ] shows the remarkable agreement among the best fit models predictions and the sneia+grb hubble diagram and @xmath1 data .
the overall good quality of the fit may be further appreciated by comparing the model predictions for the bao and cmb quantities with the observed values . for the tanh model , we get @xmath111 @xmath112 @xmath113 so that the best fit tanh model predictions are well within @xmath114 from the observed values . for the exp model , we get @xmath115 @xmath116 @xmath117 still in good agreement with the bao and cmbr data . the left panel in fig.[fig : bfplot ] is quite instructive showing us how the two best fit models can not be discriminated at all based on the hubble diagram data .
this can be easily understood looking at the right panel where one can see that the rate expansion @xmath1 is quite similar among the two models with the largest deviations occurring in the redshift range @xmath118 .
being of order only few @xmath119 , it is easy to understand that the integration needed to compute the distance modulus easily washes out the differences between the two models thus leading to the almost perfectly superimposed curves shown in fig.[fig : bfplot ] .
another way to explain why the two models make so similar predictions is to look at the reconstructed effective torsion fluid eos .
fig.[fig : eosplot ] shows @xmath53 as constrained by the likelihood analysis . in each panel
, the red solid curve refers to the best fit model , while the blue dot - dashed ( dashed ) one gives , for each @xmath63 , the median value ( the @xmath109 confidence range ) as estimated evaluating @xmath120 for the model parameters in the mcmc chain . for both best fit models
( i.e. , look at the red curves ) , the eos stays close to the cosmological constant value @xmath121 up to @xmath122 so that the dynamics is approximately the same as the @xmath36cdm one . on the other hand
, @xmath53 has a radical different behaviour for the two models in the higher redshift regime .
indeed , while the tanh effective torsion fluid has an increasing eos , the exp model eos stays roughly constant to a value slightly smaller than the present day one thus better mimicking a @xmath36cdm scenario .
nevertheless , both best fit models make similar predictions for the cmbr distance priors as a result of the very small contribution of the torsion fluid to the energy budget at the last scattering surface . as a final remark
, it is worth spending some more words on an unusual feature of the exp model .
as one can see from table[tab : expfit ] , the best fit @xmath61 parameters are strongly different from their median values with @xmath60 being outside the @xmath110 confidence range . as a consequence ,
the reconstructed torsion fluid eos for the best fit model is outside the @xmath109 confidence range delimited by the blue dashed lines in the right panel of fig.[fig : eosplot ] .
such an unexpected result is actually a consequence of the strong degeneracy among the model parameters . as already highlighted when commenting fig.[fig : wzeffexp ] , for a given @xmath57 value , there are more than one @xmath60 value giving the same @xmath123 .
this intrinsic degeneracy can only partially be broken by fitting data at different redshifts . however , since most part of the data probes the range @xmath124 , all the couples @xmath61 giving @xmath125 over this range are allowed by the fit .
this is indeed the case for the exp models with @xmath61 set to their best fit values ( leading to @xmath126 ) or to the median ones ( giving @xmath127 ) . for larger @xmath63 ,
the median and best fit models eos are radically different , but the predicted distance priors values are nevertheless almost the same since the torsion fluid becomes subdominant so that the details of its eos turn out to be meaningless . in the following
, we will however set the @xmath61 parameters to their median values since such a choice more reliably takes into account how the data constrain the exp model parameters space .
for the tanh ( red solid ) and exp ( blue dashed ) @xmath0 models superimposed to the @xcite data.,width=283 ] the above analysis , while confirming that both @xmath0 models nicely work at reproducing the available data , has also highlighted that probing only the background expansion does not allow us to conclusively discriminate between them and the @xmath36cdm scenario . in order to deepen the investigation and look for a way to put stronger constraints
, one has to resort to the analysis of the evolution of the perturbations . as a first step ,
we consider here the growth factor @xmath128 with @xmath129 the matter density contrast .
the @xmath130 evolution can easily be determined solving @xcite @xmath131 which is formally identical to the @xmath36cdm one provided the newton constant @xmath132 is replaced by the effective one @xmath133 .
it is worth stressing that eq.([eq : growthfac ] ) actually holds only in the sub - horizon limit , i.e. , under the approximation @xmath134 with @xmath135 the comoving expansion rate .
a much more complicated expression @xcite holds in the general case and must be used in evaluating large scale probes as the matter power spectrum and , for instance , weak lensing observables .
however , for this preliminary investigation , we will use the simpler eq.([eq : growthfac ] ) which is well suited for galactic and galaxy cluster scales .
fig.[fig : gzplot ] shows the growth factor for the best fit tanh and the median exp models together with a compilation of @xmath136 measurement taken from @xcite .
as it is evident , the two models can be easily differentiated from the different @xmath136 behaviour , while their background expansion is almost the same as shown by the very close agreement between the corresponding @xmath137 and @xmath1 curves plotted in fig.[fig : bfplot ] . according to the @xmath136 values ,
structures growth is faster in the tanh than in the exp model so that an observational determination of @xmath136 can easily discriminate among the two @xmath0 scenarios . indeed , using the data plotted , we get @xmath138 for the tanh model indicating a strong disagreement , while better results are obtained for the exp model leading to @xmath139 ( still large , but mainly due to the highest @xmath63 point which is not even fitted by the @xmath36cdm growth factor ) .
these values have , however , to be taken with great caution and should not be considered as a definitive evidence of the exp model working better than the tanh one .
first , most of the measurements of @xmath136 reported in @xcite actually refers to @xmath140 with @xmath141 the bias of the galaxy population used to trace the growth factor .
since the bias is related to the underlying gravity theory , a modified lagrangian can in principle introduce deviations from the standard collapse scenario thus leading to a different and possibly scale dependent bias .
second , @xmath136 is sometimes obtained from a preliminary modelling of the matter power spectrum as in @xcite which partially relies on the assumption of a fiducial @xmath36cdm model to convert from the redshift to the real space clustering . while the background expansion is likely the same as the @xmath36cdm one up to intermediate @xmath63 , the modeling of the power spectrum and its distortions can be different so that the inferred @xmath136 estimate should be taken _
cum grano salis_. motivated by these considerations , we prefer to be conservative and consider fig.[fig : gzplot ] as a strong evidence that accelerating @xmath0 models with a quite similar background expansion can be easily discriminated by their growth factor evolution .
interpreting gravitational interactions in terms of the torsion rather than the scalar curvature leads to the equivalent teleparallel formulation of gr . adding a further @xmath0 term into the gravity lagrangian then modifies general relativity allowing for a wider range of possible dynamical behaviour .
in particular , for some particular choices of the @xmath0 functional expression , it is possible to get an accelerating expansion in a matter only universe with the effective torsion fluid playing the role of dark energy .
motivated by these considerations , we have here considered two generalisations of teleparallel models assigned by the expression ( [ eq : ftmods ] ) for the @xmath0 function .
we have convincingly shown that both models are in excellent agreement with a wide set of cosmological data , from the sneia+grb hubble diagram to the @xmath1 measurements , baos and cmbr distance priors .
having been designed to give an accelerated expansion , the parameter space of both models collapses into a region giving rise to a background dynamics similar to the @xmath36cdm one . as a consequence ,
the two models can be hardly discriminated based on the dataset we have used .
it is worth stressing that this is not a limitation of the data , but rather an intrinsic feature of how the models have been worked out . as such , improving the precision of the measurements or increasing the statistics does not help in discriminating among the @xmath0 models and the @xmath36cdm one . on the contrary , one has to resort to different tracers which are related to the evolution of the perturbations , the simplest one being the growth factor @xmath136 . on small scales
, the impact of torsion only introduces a redshift dependent rescaling of the gravitational constant which now becomes @xmath133 .
this only modification considerably changes the growth factor leading to a scaling of @xmath142 with @xmath63 which strongly depend on the @xmath0 model considered .
as fig.[fig : gzplot ] clearly shows , it is easy to discriminate between the two models provided one has a model independent estimate of the growth factor as function of @xmath63 .
however , to this end , one should carefully take into account how the collapse of structures take place in the modified regime introduced by the effective torsion field . to this end ,
the weak field limit of @xmath0 models must be investigated thus also leading to the determination of the gravitational potential on galactic scales and its impact on the galaxy dynamics . on the largest scales ,
the torsion field modifies the growth of perturbations by altering both the gravitational constant and the friction term @xcite . as a consequence
a richer phenomenology is achieved possibly leading to other ways to discriminate among @xmath0 models and dark energy ones .
two candidate probes are the matter power spectrum @xmath143 and cosmic shear . in the first case ,
the derivation of @xmath143 is likely quite easy to achieve the main deviations from the standard gr one being encoded by the scale dependent growth factor .
however , the comparison with the data asks for two further effects to be taken into account .
first , what we actually observe is the galaxy power spectrum which is related to the matter one through the bias @xmath144 ( if we neglect redshift distortions ) . while for gr models one can assume @xmath144 to be scale independent , this is not guaranteed a priori for @xmath0 models . as a possible way out
, one could rely on a parameterized phenomenological ansatz although this strategy could introduce a degeneracy with the @xmath0 parameters . on smaller scales ,
where @xmath136 is scale independent , nonlinear effects start becoming important so that one should find a way to take them into account .
as far as we know it , a mapping from the linear @xmath143 to the nonlinear one is still unavailable for @xmath0 theories thus preventing a straightforward comparison between predicted and observed @xmath143 .
a somewhat better situation holds for the computation of the cosmic shear power spectrum .
indeed , it probes the full matter distribution along the line of sight and not the clustered component . as such
, the bias plays no role at all so that there is no need to preliminary investigate how galaxy formation takes place in @xmath0 models .
moreover , the cosmic shear power spectrum allows us to probe the largest scales where nonlinear effects may be neglected thus putting more emphasis on the modifications due to the deviations of @xmath0 theory from gr .
we therefore consider cosmic shear as a more promising tool to discriminate among rival teleparallel scenarios and between @xmath0 and gr and deserve a detailed analysis to a forthcoming publication @xcite .
vfc is funded by asi ( agenzia spaziale italiana ) .
nr has been partially supported by a infn / micinn collaboration .
sc acknowledges support from fct - portugal under grant ptdc / fis/100170/2008 . | generalised teleparallel gravity , also referred to as @xmath0 gravity , has been recently proposed as an extended theory of gravitation able to give rise to an accelerated expansion in a matter only universe .
the cosmic speed up is driven by an effective torsion fluid whose equation of state depend on the @xmath0 function entering the modified gravity lagrangian .
we focus on two particular choices for @xmath0 which share the nice property to emulate a phantom divide crossing as suggested by some recent data .
we check their viability contrasting the predicted background dynamics to the hubble diagram as traced by both type ia supernovae ( sneia ) and gamma ray bursts ( grbs ) , the measurement of the rate expansion @xmath1 , the baryon acoustic oscillations ( baos ) at different redshifts , and the cosmic microwave background radiation ( cmbr ) distance priors .
both @xmath0 models turn out to be in very good agreement with this large dataset so that we also investigate whether it is possible to discriminate among them relying on the different growth factors . |
in this paper we show the existence of coverings @xmath2 of @xmath3 by real curves satisfying some specific properties related to the real locus @xmath4 .
part of those constructions is related to recent results on real pencils ( one - dimensional linear systems ) on real curves .
we write @xmath5 to denote the associated riemann surface , @xmath6 to denote its genus and @xmath7 such that @xmath8 is the number of connected components of @xmath9 .
consider a covering @xmath10 and choose an orientation on @xmath11 .
let @xmath12 be a connected component of @xmath4 and consider the restriction @xmath13 .
fixing an orientation on @xmath12 this restriction has a topological degree @xmath14 and we assume the orientations are chosen such that @xmath15 .
we consider the existence of coverings having prescribed values for those topological degrees .
there are trivial natural restrictions between @xmath16 , @xmath17 , @xmath18 and those topological degrees and we prove that there exists such a covering @xmath19 for all values @xmath20 , @xmath21 , @xmath22 and @xmath23 satisfying those restrictions . in particular , for each @xmath24 , @xmath25 with @xmath26 there exists a real curve @xmath27 of genus @xmath21 such that @xmath28 and @xmath4 has exactly @xmath29 connected components and a morphism @xmath30 of degree @xmath29 such that @xmath31 for each component @xmath12 of @xmath4 .
in particular @xmath32 is the union of @xmath29 homeomorphisms . in
@xcite one introduces very special linear systems on real curves satisfying strong clifford - type properties for real curves and in proposition 2.1 of @xcite it is proved that the only very special pencils are exactly those defined by such coverings .
so we obtain a proof for the existence of very special pencils of all possible types . in @xcite one studies some brill - noether properties for pencils on real curves .
we give another proof of part of the results in @xcite making intensive use of results coming from the theory of complex curves . in this way
it is enough to prove the existence of one real curve having a pencil of suited degree . because of our existence results we obtain brill - noether properties for pencils having prescribed topological degree on the connected components of the real locus .
while the arguments in @xcite are restricted to real curves having real points , our arguments also give brill - noether properties for pencils on real curves without real points .
the existence and brill - noether results are also considered in @xcite , @xcite and @xcite without considering the topological degrees . in order to make the arguments more careful with respect to dimension arguments on the real locus of a variety defined over @xmath33 we use universal spaces representing morphisms that are known to be globally smooth because of horikawa deformation theory . a real divisor @xmath34 on a real curve
is called totally non - real in case the support of @xmath34 contains no real point of @xmath27 .
it is known that for a real linear system on a real curve the parity of the restriction of a divisor to a connected component of @xmath4 is constant .
hence if this parity is odd for some component then the linear system does not contain a totally non - real divisor . restricting to linear systems such that the parity is even for all components of @xmath4 ,
there is a sharp result in @xcite concerning the existence of totally non - real divisors in such a linear system . for linear systems of large dimension , it is not easy to find non - trivial examples of such linear systems having no totally non - real divisors . in the case of pencils
our constructions give a lot of such examples coming from coverings having only even topological degrees and at least one of them being non - zero .
one can ask for the existence of such examples coming from coverings with all topological degrees equal to zero . we prove the existence of all types of such coverings of the smallest possible degree 4 . in section 1
we give the constructions that are the base for the existence of coverings with prescribed topological degrees . in section 2
we prove the existence of coverings with prescribed topological degrees . in section 3
we discuss brill - noether problems for pencils on real curves . finally in section 4
we prove the existence of coverings of degee 4 with all topological degrees equal to 0 and having no non - real divisor .
a smooth real curve @xmath27 of genus @xmath17 is a scheme defined over @xmath33 such that the base change @xmath35 , denoted by @xmath36 , is a complete connected smooth complex curve of genus @xmath21 .
we also write @xmath5 to denote the set of closed points of @xmath36 and we consider it as a riemann surface of genus @xmath21 . as usual the non - trivial element of @xmath37 is denoted by @xmath38 and it is called ( complex ) conjugation . for a real curve @xmath27 it induces an @xmath33-involution @xmath39 . for @xmath40 we write @xmath41 instead of @xmath42 and we call it the _ conjugated point _ of @xmath43 . in case @xmath44 then @xmath43 is called a _
real point _ of @xmath27 .
the set of real points is called the real locus @xmath45 .
this locus is a union of @xmath46 connected components , each one of them homeomorphic to a circle .
the invariant @xmath47 is defined by @xmath48 in case @xmath49 is connected and otherwise @xmath28 ( in that case @xmath49 has two connected components ) .
we say a smooth real curve @xmath27 has _ topological type _
@xmath50 if @xmath51 , @xmath52 and @xmath53 . by a theorem of weichold
such real curve exists if and only if either @xmath54 and @xmath55 else @xmath56 , @xmath57 and @xmath58 ( see @xcite , see also @xcite*theorem 5.3 ) .
a triple @xmath50 satisfying those restrictions is called an _ admissable topological type _ for real curves . in case @xmath59
then we call @xmath60 a _ non - real point _ of @xmath27 . for a divisor @xmath61 on @xmath36
we write @xmath62 and we call it the conjugated divisor of @xmath34 . as usual @xmath63 is called the degree of @xmath34 , denoted by @xmath64 .
we say @xmath34 is a real divisor if @xmath65 .
if we say @xmath34 is a divisor on @xmath27 then we mean it is a _ real divisor _ on @xmath36 .
let @xmath66 be an invertible sheaf on @xmath36 .
it can be described by means of trivializations on an open covering @xmath67 of @xmath5 and transition functions @xmath68 ( those are regular functions @xmath69 ) . the conjugated invertible sheaf @xmath70 is defined by means of trivializations on the open covering @xmath71 and transition functions @xmath72 ( for @xmath73 one has @xmath74 ) .
each invertible sheaf @xmath66 on @xmath36 is isomorphic to @xmath75 for some divisor @xmath34 on @xmath36 . in case @xmath76
then we say @xmath66 is an invertible sheaf of degree @xmath1 and we write @xmath77 . in case @xmath78
then @xmath79 .
we say @xmath66 is a _ real invertible sheaf _ if @xmath80 for some real divisor @xmath34 .
we say @xmath66 is _ invariant under conjugation _ if @xmath81 .
in particular a real invertible sheaf is invariant under conjugation .
let @xmath82 be the picard scheme of @xmath27 defined over @xmath33 ( see @xcite ) .
then @xmath83 represents the picard functor on @xmath36 , in particular @xmath84 parameterizes invertible sheaves on @xmath36 . for @xmath85 one
has natural subschemes @xmath86 such that @xmath87 parameterizes invertible sheaves of degree @xmath1 on @xmath36 . the real locus @xmath88 parameterizes invertible sheaves of degree @xmath1 on @xmath36 invariant under conjugation .
let @xmath89 be the sublocus parameterizing real invertible sheaves . in case @xmath90
then @xmath82 also represents the picard functor on @xmath27 and therefore @xmath91 in that case . in case @xmath92
then @xmath89 is a subgroup of @xmath88 with quotient @xmath93 . for @xmath94 with @xmath95 one
has @xmath96 ( see @xcite*proposition 2.2 ) .
we write @xmath97 to denote the complete linear system of an invertible sheaf @xmath66 on @xmath36 . in case @xmath66
is a real invertible sheaf then @xmath98 is the space of real divisors contained in @xmath97 .
one has @xmath99 in a natural way ( meaning @xmath100 can be considered as a projective space @xmath101 and then @xmath102 ) .
as usual we write @xmath103 to denote a linear system of dimension @xmath104 and degree @xmath1 on @xmath36 . in case @xmath66
is a real invertible sheaf of degree @xmath1 then we say a linear subsystem @xmath105 of @xmath97 is defined over @xmath33 if there exists a @xmath1-dimensional linear subsystem @xmath106 such that @xmath107 . in such case
we say @xmath105 is a linear system on @xmath27 .
in particular if @xmath66 is a real invertible sheaf then we say @xmath97 is a linear system on @xmath27 .
we write @xmath3 to denote the projective line defined over @xmath33 and we write @xmath108 to denote the smooth real curve of genus 0 without real points ( defined by the equation @xmath109 in @xmath110 ) .
let @xmath27 be a smooth real curve .
a covering @xmath111 or @xmath112 is a finite morphism defined over @xmath33 .
we say @xmath113 has degree @xmath20 in case the associated covering @xmath114 has degree @xmath20 .
such covering corresponds to a base point free pencil @xmath115 on @xmath36 . in case of @xmath116
this is called a _
real pencil _ on @xmath27 . in case of @xmath112
this is called a _ non - real invariant pencil _ on @xmath27 .
this case only can occur if @xmath92 ( @xmath27 has no real points ) and moreover , in this case the pencil has no real divisor but it is invariant under complex conjugation .
we write @xmath117 to denote the covering of riemann surfaces . in case
@xmath4 is not empty we write @xmath118 to denote the restriction of @xmath119 to the real locus @xmath4 . for a connected component @xmath12 of @xmath4 we write @xmath120 to denote the restriction of @xmath121 to @xmath12 . on @xmath11 , homeomorphic to a circle @xmath122 , we choose an orientation and then for @xmath12 , also homeomorphic to @xmath122 , we choose the orientation such that the degree of @xmath123 , denoted by @xmath124 , is nonnegative .
if @xmath125 and @xmath123 is not ramified at @xmath43 , then the local degree @xmath126 is equal to 1 ( resp .
-1 ) if @xmath123 preserves ( resp . reverses ) the orientation locally at @xmath43 .
let @xmath127 be the connected components of @xmath4 and let @xmath128 for @xmath129 .
we can always assume @xmath130 and then we say @xmath113 has _ topological degree _
@xmath131 ( shortly denoted by @xmath132 ) .
the proof of the existence of coverings with prescribed topological degrees uses an induction argument . to make that argument we start with a smooth real curve @xmath133 and a suited covering @xmath134 . using this covering
we construct a covering @xmath135 with @xmath136 being a real singular nodal curve and we use a real smoothing @xmath137 of @xmath138 .
in this part we describe those constructions giving relations between topological degrees of @xmath139 and of @xmath140
. this will be the base for the induction argument in the next section . in those constructions we start by taking local smoothings of the nodes of @xmath141 having a natural antiholomorphic involution .
those local smoothings glue with the complement @xmath142 of a neighborhood of the nodes of @xmath141 giving rise to a deformation of compact riemann surfaces @xmath143 and holomorphic coverings @xmath144
. this complement @xmath142 can be taken to be invariant under complex conjugation on @xmath141 and this fits with the antiholomorphic involution of the local smoothings under the gluing .
hence we obtain an antiholomorphic involution @xmath145 on @xmath143 .
it is well - known that this defines a smooth real curve @xmath146 inducing the riemann surface @xmath143 such that @xmath145 corresponds to complex conjugation ( see e.g. @xcite*section 4 ) .
moreover the morphism @xmath147 is invariant under complex conjugation hence it comes from a morphism @xmath148 ( indeed , the graph of @xmath149 is a closed subspace of @xmath150 invariant under complex conjugation on @xmath151 ) .
let @xmath133 be a real curve of genus @xmath21 and let @xmath152 be a morphism of degree @xmath1 defined over @xmath33 .
assume @xmath12 is a connected component of @xmath153 and let @xmath43 be a point of @xmath12 such that @xmath154 ( such point @xmath43 does exist because @xmath155 ) . take a copy of @xmath3 and consider the singular curve @xmath156 obtained by identifying @xmath43 on @xmath27 with @xmath157 on @xmath3 .
this singular curve has a natural morphism @xmath138 defined over @xmath33 of degree @xmath158 to @xmath3 having restriction @xmath113 to @xmath133 and the identity to @xmath3 ( see figure [ figure 1 ] ) .
, height=3 ] locally at @xmath43 this situation can be described inside @xmath159 such that the curve @xmath136 has equation @xmath160 and the morphism is locally defined by @xmath161 and we can assume the coordinates are compatible with the real structure on @xmath136 ( this means complex conjugation on @xmath136 corresponds to complex conjugation in @xmath159 ) . let @xmath162 be a small neighborhood of @xmath163 in @xmath159 and @xmath164 a much smaller one and let @xmath165 , @xmath166 . for @xmath167 in @xmath168 we can use @xmath169 to define a holomorphic coordinate at @xmath170 .
consider a local deformation of @xmath136 at @xmath43 defined by @xmath171 with @xmath172 ( @xmath173 very small ) and let @xmath174 ( resp .
@xmath175 ) be the intersection with @xmath162 ( resp .
@xmath142 ) .
we use a gluing of @xmath174 and @xmath176 as follows . for @xmath177
let @xmath178 be the locally defined holomorphic square root function such that @xmath179 .
a point @xmath180 has coordinates @xmath181 or @xmath182 .
we identify @xmath181 with @xmath183 and @xmath182 with @xmath184 .
( one should adapt the description of @xmath175 and @xmath174 to this identification . )
this defines a riemann surface @xmath143 . on @xmath174
we define @xmath185 .
we need to show this behaves well under the previous identification .
since @xmath186 is close to @xmath169 one has @xmath187 is close @xmath188 .
moreover @xmath189 hence @xmath190 . hence under the identification @xmath191 is identified with @xmath192 and similarly @xmath193 is identified with @xmath194 . as mentioned at the beginning of this section we obtain a real smooth curve @xmath146 and a covering @xmath195 of degree @xmath158 . under this deformation the union @xmath196 deforms to a connected component @xmath197 of @xmath198 in case @xmath199 the morphism @xmath140 has ramification on @xmath197 above @xmath200 .
we call this _ the deformation with real ramification_. in order to have an orientation on @xmath197 we have to change the orientation on the attached @xmath11 .
this implies that @xmath201 . of course
, if @xmath202 we also change the orientation of @xmath197 in order to obtain @xmath203 ( see figure [ figure 2 ] ) . in case
@xmath204 the morphism @xmath140 has no ramification on @xmath197 close to @xmath43 .
we call this _ the deformation without real ramification_. we use the orientation on @xmath197 obtained from both the orientation on @xmath12 and the attached @xmath11
. this implies @xmath205 ( see figure [ figure 3 ] ) . in both cases @xmath206
is connected if and only if @xmath207 is connected .
let @xmath133 be a real curve of genus @xmath21 and let @xmath208 be a morphism of degree @xmath1 defined over @xmath33 .
let @xmath209 not a branch point of @xmath113 and assume @xmath210 , a non - real point on @xmath133 ( in particular we assume such non - real point exists ) .
let @xmath136 be the singular curve obtained from @xmath27 by identifying @xmath43 with @xmath41 .
this singular curve has a natural morphism of degree @xmath1 to @xmath3 defined over @xmath33 obtained from @xmath113 ( see figure [ figure 4 ] ) . , height=3 ]
on @xmath136 the singularity ( again denoted by @xmath43 ) is an isolated real point .
locally at @xmath43 this situation can be described inside @xmath159 such that the curve has equation @xmath211 and the morphism is locally defined by @xmath161 with coordinates compatible with the real structure on @xmath136 .
we use @xmath162 and @xmath142 as in construction i. consider the local deformation of @xmath136 at @xmath43 defined by @xmath212 with @xmath213 ( @xmath173 very small ) . a point @xmath214 has coordinates @xmath215 or @xmath216 and @xmath217 corresponds to resp .
@xmath218 or @xmath219 .
we identify @xmath215 ( resp .
@xmath220 ) on @xmath221 with @xmath222 ( resp .
@xmath223 ) on @xmath174 .
this defines a riemann surface @xmath143 . on @xmath174
we define @xmath224 and we check that it behaves well under the identification .
consider @xmath225 on @xmath226 .
since @xmath227 is close to @xmath228 one has @xmath229 is close to @xmath230 , hence @xmath231 .
hence @xmath232 is identified with @xmath233 .
we obtain a smooth real curve @xmath146 of genus @xmath234 together with a morphism @xmath235 of degree @xmath1 defined over @xmath33 . in case
@xmath199 there is a new component @xmath197 of @xmath198 close to @xmath43 and under @xmath140 this maps to the interval @xmath236 $ ] .
in particular @xmath237 .
we call this _ the deformation with real ramification_. in this case @xmath206 is connected if and only if @xmath207 is connected ( see figure [ figure 5 ] ) . in case @xmath204
then @xmath198 contains no real point close to @xmath43 .
we call this the deformation without real ramification . in this case
@xmath206 is always connected ( see figure [ figure 6 ] ) .
let @xmath133 be a real curve of genus @xmath21 and let @xmath208 be a morphism of degree @xmath1 defined over @xmath33 .
assume @xmath60 is a non - real point of @xmath133 such that @xmath238 and @xmath43 , @xmath41 are not ramification points of @xmath239 .
now we let @xmath240 be the union of @xmath133 and @xmath3 identifying @xmath43 with @xmath157 ( still denoted by @xmath43 ) and @xmath41 with @xmath241 ( still denoted by @xmath41 ) .
this singular curve has a natural morphism @xmath138 defined over @xmath33 of degree @xmath158 to @xmath3 having restriction @xmath113 to @xmath133 and the identity to @xmath3 ( see figure [ figure 7 ] ) .
, height=4 ] choose a local deformation over @xmath242 making it smooth at @xmath43 and take the complex conjugated local deformation at @xmath41 and glue it with the remaining part of @xmath136 .
we obtain a smooth real curve @xmath146 of genus @xmath234 together with a morphism @xmath235 of degree @xmath158 defined over @xmath33 .
one has @xmath243 and @xmath11 on @xmath244 deforms to a connected component @xmath197 of @xmath198 such that @xmath245 .
for this construction @xmath206 is connected if and only if @xmath207 is connected ( see figure [ figure 8 ] ) .
consider a double covering @xmath246 defined over @xmath33 .
assume @xmath133 is a real curve without real points and @xmath247 is a morphism of degree @xmath1 defined over @xmath33 ( in particular @xmath1 is even ) .
choose @xmath60 on @xmath3 ( a non - real point ) such that @xmath43 , @xmath41 are not branch points of @xmath239 and choose non - real points @xmath248 on @xmath108 and @xmath249 on @xmath133 such that @xmath250 .
consider the nodal curve @xmath251 obtained by identifying @xmath252 to @xmath170 ( still denoted by @xmath170 ) and @xmath253 to @xmath254 ( still denoted by @xmath254 ) .
this is a singular curve defined over @xmath33 without real points and @xmath113 together with @xmath255 induce a morphism of that singular curve to @xmath3 defined over @xmath33 .
take a local deformation at @xmath170 over @xmath242 smoothing the singularity , take the complex conjugated deformation at @xmath254 and glue it with the remaining part of @xmath136 to obtain a smooth real curve @xmath146 of genus @xmath234 together with a morphism @xmath235 of degree @xmath256 .
this smooth real curve does not have any real point .
assume @xmath133 is a real curve of genus @xmath21 without real points and assume @xmath257 is a morphism of degree @xmath1 defined over @xmath33 .
choose a non - real point @xmath248 on @xmath108 such that @xmath252 , @xmath253 are not branch points of @xmath239 and a non - real point @xmath60 on @xmath133 with @xmath258 .
consider the nodal curve @xmath259 obtained by identifying @xmath252 to @xmath43 ( still denoted by @xmath43 ) and @xmath253 to @xmath41 ( still denoted by @xmath41 ) .
this is a singular curve defined over @xmath33 without real points and @xmath113 together with the identity @xmath260 induces a morphism of degree @xmath158 of that singular curve to @xmath108 defined over @xmath33 . take a local deformation at @xmath43 over @xmath242 smoothing the singularity , take the complex conjugated deformation at @xmath41 and glue it with the remaining part of @xmath136 to obtain a smooth real curve @xmath146 of genus @xmath234 together with a morphism @xmath261 of degree @xmath158 .
let @xmath27 be a real curve of topological type @xmath50 and let @xmath262 be a morphism of degree @xmath1 defined over @xmath33 .
assume @xmath263 and let @xmath264 be the topological degree of @xmath113 .
of course we need @xmath265 . for each @xmath266
the number of real points in @xmath267 counted with multiplicity has the same parity as @xmath268 . since the non - real points of @xmath27
are pairs of conjugated points on @xmath5 it follows @xmath269 . in case @xmath270
then for each @xmath266 the fiber @xmath271 has exactly @xmath272 different points on @xmath273 .
since some fiber needs to contain points of @xmath274 one has @xmath275 in that case , hence @xmath276 in case @xmath277 ( see also @xcite ) .
[ definition1 ] we say @xmath278 is an _ admissible topological degree for morphisms of degree @xmath1 _ if * @xmath279 , * @xmath280 , * @xmath281 in case @xmath277 . for the next theorem
we need some more restriction . in case
there is a morphism @xmath30 of degree @xmath1 with @xmath282 then by definition @xmath283 , hence @xmath284 . since @xmath285 is disconnected it follows @xmath286
is disconnected too , hence @xmath28 .
therefore , in case @xmath48 one always has @xmath287 .
[ theorem2 ] let @xmath288 , let @xmath50 be an admissable topological type of real curves and let @xmath132 be an admissible topological degree for morphisms of degree @xmath1 .
assume @xmath289 in case @xmath54 .
there exists a real curve @xmath27 of topological type @xmath50 having a morphism @xmath262 of degree @xmath1 of topological degree @xmath132 .
[ remarktheorem2 ] the case @xmath290 is contained in e.g. @xcite*section 6 . in that case
there are more restrictions in the case @xmath56 . nevertheless , the existence of real hyperelliptic cruves is the starting point in a lot of cases of the proof of the theorem , even in case 5 which is excluded for real hyperelliptic curves unless @xmath291 .
section [ subsection2.1 ] is the proof of theorem [ theorem2 ] in case @xmath54 and section [ subsection2.2 ] is the proof of theorem [ theorem2 ] in case @xmath56 . in section [ subsection2.3 ]
we prove the corresponding theorem for morphisms to @xmath108 .
first we consider the case @xmath292 .
let @xmath293 be the number of @xmath294 and let @xmath295 be a real hyperelliptic curve of topological type @xmath296 ( such curves do exist , see @xcite*section 6 ) .
let @xmath297 be the associated double covering of @xmath3 then it follows that @xmath298 for @xmath299 . applying construction iii @xmath293 times we obtain @xmath300 with @xmath301 ; @xmath302 and @xmath303 has @xmath29 components @xmath304 such that @xmath305 for @xmath306 and @xmath307 for @xmath308 , in particular @xmath309 has topological type @xmath310 .
for each @xmath311 we use construction i @xmath312 times on @xmath313 and using the deformation without real ramification .
we obtain @xmath314 with @xmath315 , @xmath316 and @xmath317 has @xmath29 components @xmath318 such that @xmath319 for @xmath320 and @xmath321 for @xmath308 hence @xmath322 has topological degrees @xmath132 and @xmath323 has topological type @xmath310 . in case @xmath324 we apply construction i @xmath325 times on @xmath326 using the deformation with real ramification . in case @xmath327 and @xmath328
we apply construction i @xmath325 times on @xmath329 alternating the deformation with and without real ramification .
we obtain a real curve @xmath27 of topological type @xmath310 and a real morphism @xmath30 having degree @xmath1 and topological degree @xmath132 . indeed , in this final step the degrees of the restrictions of @xmath113 do not change since @xmath325 is even .
now assume @xmath330 ( the case of no real points ) .
in this case @xmath1 is even . in case
@xmath331 a general divisor @xmath34 on a real curve @xmath27 of degree @xmath1 gives rise to a base point free linear system @xmath332 .
a general pencil in it defined over @xmath33 gives rise to a real base point free @xmath333 on @xmath27 , hence to a real morphism @xmath30 of degree @xmath1 .
so we can assume @xmath334 .
start with a hyperelliptic curve @xmath295 of genus @xmath335 with @xmath336 .
we apply construction iv @xmath337 times .
again we obtain a real curve @xmath27 of genus @xmath21 with @xmath92 and having a real morphism @xmath30 of degree @xmath1 .
again , let @xmath293 be the number of @xmath294
. assume @xmath338 and @xmath339 .
in particular @xmath340 ( mod 2 ) .
let @xmath136 be a real hyperelliptic curve of genus @xmath21 having @xmath341 , @xmath342 , hence for the hyperelliptic real morphism @xmath343 the unique component @xmath344 of @xmath345 has topological degree 2 ( see @xcite*section 6 ) .
apply @xmath346 times construction i using the deformation without real ramification .
we obtain a real curve @xmath27 of topological type @xmath347 such that @xmath27 has a covering @xmath111 defined over @xmath33 of degree @xmath1 such that @xmath348 .
assume @xmath349 and @xmath350 in which case @xmath351 . in case
all @xmath352 then @xmath353 and therefore @xmath354 ( mod 2 ) . in particular
@xmath355 is odd . since @xmath356 and @xmath357 also @xmath358 .
let @xmath136 be a real hyperelliptic curve of genus @xmath359 having @xmath341 , @xmath360 ( see @xcite*section 6 ) .
hence , for the real hyperelliptic morphism @xmath361 the 2 components @xmath12 of @xmath345 satisfy @xmath362 .
apply @xmath346 times construction iii .
we obtain a real curve @xmath27 of genus @xmath21 having @xmath28 , @xmath363 such that @xmath27 has a covering @xmath364 defined over @xmath33 of degree @xmath1 such that each component @xmath12 of @xmath4 satisfies @xmath365 .
assume not all @xmath352 , hence @xmath366 .
let @xmath136 be a real hyperelliptic curve of genus @xmath367 ( in particular @xmath368 is even ) as in case 1 .
apply @xmath369 times construction iii .
we obtain a real curve @xmath309 of genus @xmath21 having @xmath28 , @xmath52 such that @xmath309 has a covering @xmath370 defined over @xmath33 of degree @xmath371 such that exactly one component @xmath344 of @xmath303 satisfies @xmath372 while the other components @xmath373 satisfy @xmath374 .
we apply @xmath375 times construction i without real ramification starting with @xmath344 and for @xmath376 we apply @xmath312 times construction i without real ramification starting with @xmath273 .
we obtain a real curve @xmath27 of genus @xmath21 satisfying @xmath28 , @xmath52 such that @xmath27 has a covering @xmath377 defined over @xmath33 of degree @xmath378 having topological degree @xmath132 .
assume @xmath379 and @xmath351 .
it follows that @xmath380 and @xmath381 ( mod 2 ) .
in particular @xmath382 is even .
let @xmath136 be a real hyperelliptic curve of genus @xmath382 such that @xmath341 , @xmath342 and for the real hyperelliptic morphism @xmath383 one has that the unique component @xmath344 of @xmath345 satisfies @xmath384 ( see @xcite*section 6 ) .
apply construction i using the deformation with real ramification to @xmath344 and apply @xmath385 times construction iii .
we obtain a real curve @xmath309 of genus @xmath21 having @xmath386 , @xmath387 such that there is a morphism @xmath388 of degree @xmath389 defined over @xmath33 such that each component @xmath273 of @xmath303 satisfies @xmath374 .
for @xmath320 apply @xmath312 times construction i using the deformation without real ramification starting with @xmath273 .
we obtain a real curve @xmath323 of genus @xmath21 having @xmath390 , @xmath391 and such that there is a morphism @xmath392 of degree @xmath393 of topological degree @xmath132 . by assumption @xmath394 is an even non - negative integer . apply @xmath394 times construction i starting with @xmath344 alternating deformation with and without real ramification ( starting with real ramification ) we end up with a real curve @xmath27 of genus @xmath21 satisfying @xmath28 , @xmath395 and such that there is a morphism @xmath30 of degree @xmath1 of topological degree @xmath132 .
assume @xmath379 , @xmath396 and @xmath380 .
we start using the construction of case 3 making a real curve @xmath397 of genus @xmath398 satisfying @xmath399 , @xmath400 and such that there is a real morphism @xmath401 of degree @xmath1 such that , for the components @xmath402 one has @xmath403 . notice that , since @xmath404 ( mod 2 ) one has @xmath405 ( mod 2 ) .
since @xmath379 the set @xmath406 is not empty
. therefore we can apply construction ii @xmath407 times using a deformation with real ramification .
we obtain a real curve @xmath27 of genus @xmath21 such that @xmath28 , @xmath52 and @xmath27 has a real covering @xmath30 of degree @xmath1 of topological degree @xmath132 .
assume @xmath408 .
in this case @xmath1 is even .
let @xmath136 be a hyperelliptic curve of genus @xmath409 ( hence of even genus ) such that @xmath341 , @xmath342 and if @xmath383 is the hyperelliptic morphism then the unique component @xmath344 of @xmath345 satisfies @xmath384 ( see @xcite*section 6 ) .
apply @xmath346 times construction i using the deformation with real ramification .
one obtains a real curve @xmath309 of genus @xmath409 such that @xmath386 and @xmath410 and a morphism @xmath388 of degree @xmath1 such that for the unique component @xmath344 of @xmath303 one has @xmath411 .
it follows that @xmath412 and so we apply @xmath369 times construction ii with real ramification .
we obtain a real curve @xmath27 of genus @xmath21 such that @xmath28 , @xmath52 and there exists a real morphism @xmath262 of degree @xmath1 such that for each component @xmath12 of @xmath4 one has @xmath202 .
assume @xmath27 is a real curve of genus @xmath21 with @xmath92 and assume @xmath413 is a morphism of degree @xmath1 defined over @xmath33 .
consider the associated morphism of riemann surfaces @xmath414 and let @xmath415 be the associated complete linear system on @xmath5 .
we are going to show that @xmath415 is not induced by a linear system on @xmath27 ( i.e. @xmath415 does not contain a real divisor ) .
notice that complex conjugation of divisors on @xmath5 acts on @xmath415 .
indeed , choose a non - real point @xmath249 on @xmath108 and let @xmath416 . since @xmath417 one has @xmath418 .
for any @xmath419 one has @xmath420 .
this implies @xmath421 hence also @xmath422 .
now @xmath119 corresponds to a line @xmath66 in @xmath415 that is invariant under this complex conjugation .
if @xmath415 would be real it would imply @xmath66 is a real line , hence containing a real point .
such real point corresponds to a real divisor and this would be a fiber of @xmath423 invariant under complex conjugation but such fibers do not exist .
this implies @xmath424 corresponds to a point of @xmath425 , hence @xmath426 ( see @xcite ) . in this section
we consider pencils corresponding to points in @xmath427 ( in particular @xmath428 ) .
[ theorem3 ] let @xmath21 and @xmath428 be nonnegative integers . in case @xmath429 and
@xmath27 is a real curve of genus @xmath21 with @xmath92 then there exists a morphism @xmath430 of degree @xmath1 . in case
@xmath431 then there exists a real curve of genus @xmath21 with @xmath92 having a morphism @xmath112 of degree @xmath1 .
first assume @xmath432 .
choose an effective divisor @xmath34 such that @xmath424 corresponds to a point on @xmath433 .
then @xmath415 is not defined over @xmath33 but complex conjugation of divisors is defined on @xmath415 .
choose @xmath434 and let @xmath66 be the line in @xmath415 connecting @xmath435 to @xmath436 .
this corresponds to a morphism @xmath437 . since @xmath438 it follows @xmath119 is equivariant using the complex conjugation on @xmath5 and some antiholomorphic involution on @xmath439 without fixed points
this implies @xmath119 is induced by a morphism @xmath440 defined over @xmath33 .
next assume @xmath441 and @xmath428 .
let @xmath309 be a real curve of genus @xmath355 having a morphism @xmath442 of degree 2 ( such curves exist since @xmath355 is odd , see @xcite*section 6 )
. then use @xmath346 times construction v. [ remark4 ] in his paper @xcite the author obtains theorem [ theorem3 ] by fixing a real divisor @xmath34 on @xmath108 and using the riemann existence theorem using coverings on @xmath443 .
although the author gives no argument for the existence of an antiholomorphic involution on a suited choice of such a covering this reasoning can be applied using klein surfaces ( see @xcite for the foundations of that theory ) .
also in that paper he obtains theorem [ theorem2 ] for the case @xmath330 but in that argument details are omitted . in his paper
@xcite the author also gives an argument for the existence of real curves @xmath27 of given topological type having a morphism @xmath116 of degree @xmath1 . for the case
@xmath28 he uses deformations of hyperelliptic and trigonal curves . the case @xmath48 is contained in @xcite . in that paper
the author uses deformations of singular real curves on suited real rational surfaces using brusotti - type arguments .
we use arguments directly applied to coverings which is more natural and easier . in both cases
it is not clear to us how to obtain the existence of all admissible topological degrees for @xmath113 using those deformations .
we also give some remarks on some of the arguments used in those papers ( especially @xcite ) in remark [ remark9 ]
. in @xcite*section 8 one considers real trigonal curves and one asks about determining the topological types of such curves .
in theorem [ theorem2 ] we obtain that all admissable topological types can be realized as real trigonal curves . a ( in principle similar ) proof is described in @xcite*proposition 2.1 . a different proof using fuchsian and nec ( non - euclidean crystallographic ) groups ( and obtaining much more detailled information concerning the case of cyclic trigonal coverings ) is given in @xcite .
on @xmath87 we consider the closed subspace @xmath444 representing invertible sheafs on @xmath36 satisfying @xmath445 .
this subspace @xmath444 is invariant under conjugation , hence it is defined over @xmath33 .
we write @xmath446 to denote its set of real points and @xmath447 . in case @xmath432 one
has @xmath448 , so we assume from now on that @xmath449 .
the study of those spaces @xmath444 for general complex curves @xmath27 of genus @xmath21 is the brill - noether theory for pencils .
the main results are the following .
first consider the brill - noether number @xmath450 .
for a general complex curve @xmath27 one has @xmath451 in case @xmath452 and @xmath453 in case @xmath454 . also in case @xmath454 then @xmath455 for all complex curves @xmath27 . here general means : the statement holds for all curves @xmath27 on a dense open subset of the moduli space @xmath456 of curves of genus @xmath21 . in @xcite one
considers brill - noether theory of @xmath446 for real curves @xmath27 .
there are important differences with the complex case .
we consider the moduli space @xmath457 of real curves of genus @xmath21 representing the isomorphism classes over @xmath33 of real curves of genus @xmath21 .
we recall some facts on it ( for details see @xcite ) .
it is a semi - analytic real variety of dimension @xmath458 ( as a matter of fact it has a non - empty boundary , see @xcite for its description ) . for each topological type @xmath50
there is a unique connected component @xmath459 of @xmath457 .
this connected component is the quotient of a teichmuller space ( which is a connected real analytic manifold ) by means of a discontinuous action of a modular group .
inside such component @xmath459 , curves having a certain type of real linear system give rise to semi - analytic subvarieties of @xmath459 .
therefore there is in general no description of behavior on a dense open subset of @xmath459 and we need to modify the meaning of general .
[ definition5 ] let @xmath43 be some property concerning real linear systems on real curves .
we say there is _ a general real curve _ of topological type @xmath50 satisfying property p if there exists a non - empty open subset @xmath162 of @xmath459 such that p holds for all curves corresponding to a point of @xmath162 .
[ examplen ] to illustrate the difference with the complex situation we consider @xmath460 for curves of genus 4 ( @xmath461 ) ( see @xcite*section 8) .
a real curve of type ( 4,0,1 ) has no real @xmath462 since all its real divisors have even degree . in case @xmath463 there exists a general real curve of type ( 4,s , a ) having a real @xmath462 . in case @xmath464 there also exists a general real curve of type @xmath465 having no real @xmath462 . in proposition
[ proposition10 ] , as an application of theorem [ theorem2 ] , we give a description of the situation in case @xmath466 .
in @xcite it is proved that in case @xmath454 and @xmath292 then there is a general real curve of type @xmath50 such that @xmath467 . in order to prove this result
the author constructs a real compactification of the real hurwitz space of coverings of degree @xmath1 .
the restriction @xmath292 has a technical cause . working with families of real curves
the author needs the existence of a section defined over @xmath33 .
in particular such a section guaranties that the relative picard scheme of such families represents the picard functor ( see @xcite ) .
now we are going to use the existence results from section [ section2 ] to give another proof . in that way
we obtain a finer statement involving the covering degree . in the argument
we make intensive use of brill - noether theory for complex curves and we work with the space of morphisms to @xmath3 instead of the hurwitz space .
this space of morphisms is defined using hilbert schemes in @xcite*section 4.c .
in particular we do not need relative picard schemes to finish our arguments , hence we do not need the restriction @xmath292 . in case @xmath330 we consider both @xmath468 and @xmath469 . in case @xmath470
is a morphism of degree @xmath1 then @xmath471 corresponds to a linear system @xmath472 for some @xmath473 . in order to study @xmath474 using morphims
, we also need the converse statement .
[ lemma6 ] let @xmath475 and assume @xmath97 is base point free of dimension 1 .
then there is a morphism @xmath470 of degree @xmath1 such that @xmath476 corresponds to @xmath97 .
see @xcite*example 1 .
let @xmath27 be a real curve of topological type @xmath50 and let @xmath333 be a complete base point free linear system on @xmath27 .
we say @xmath333 has topological degree @xmath132 if the morphisms @xmath477 associated to @xmath333 have topological degree @xmath132 .
let @xmath478 be the subspace of @xmath444 parameterizing complete base point free linear systems @xmath333 on @xmath36 .
it is the complement of @xmath479 in @xmath444 , hence it is zariski - open in @xmath444 . in particular
@xmath480 is zariski - open in @xmath446 . because the topological degree is a discrete invariant it is constant on connected components of @xmath480 .
let @xmath481 be the sublocus of @xmath480 such that it parameterizes linear systems @xmath333 of topological degree @xmath132 and let @xmath482 be its closure in @xmath446 ( it is a union of irreducible components of @xmath446 ) .
[ theorem7 ] let @xmath132 be an admissible topological degree of base point free linear systems @xmath333 on real curves of topological type @xmath50 such that @xmath276 in case @xmath54 . 1 . if @xmath452 then there is no general real curve @xmath27 of type @xmath50 such that @xmath482 is not empty .
2 . if @xmath454 then there is a general real curve @xmath27 of topological type @xmath50 such that @xmath482 is non - empty and it is a real algebraic subset of @xmath88 of dimension @xmath483 .
3 . in case @xmath330 and @xmath484 then in case @xmath485 there is no general real curve @xmath27 of topological type @xmath486 such that @xmath469 is not empty . in case @xmath454 then there is a general real curve @xmath27 of topological type @xmath486 such that @xmath487 is not empty and has dimension @xmath488 . in order to prove this theorem we are going to use spaces parameterizing morphisms to @xmath3 .
first we introduce some definition that will be useful for the proof .
[ definition8 ] a _ suited family of curves of genus @xmath21 _ is a projective mophism @xmath489 such that * @xmath490 is smooth , irreducible and quasi - projective , * each fiber of @xmath113 is a smooth connected curve of genus @xmath21 , * for each @xmath491 the kodaira - spencer map @xmath492 is bijective . in case @xmath27
is a given smooth curve of genus @xmath21 then we say @xmath493 is a suited family for @xmath27 if moreover some fiber of @xmath113 is isomorphic to @xmath27 .
if moreover @xmath27 is defined over @xmath33 then we also assume @xmath494 is defined over @xmath33 and @xmath27 is isomorphic over @xmath33 to some fiber of @xmath113 over a real point of @xmath490
. such suited families do exist ( see @xcite*section 3 , in that paper one considers real curves with real points and one needs the family to have a real section , in our situation one does not need to use @xmath495 from loc .
cit . ) . from a suited family
we obtain a morphism @xmath496 and in case @xmath497 we also obtain a morphism @xmath498 . from the bijectivity of the kodaira - spencer map it follows those morpisms to the moduli space
have finite fibers .
therefore it is enough to prove the statements of theorem [ theorem7 ] for real curves corresponding to fibers of @xmath113 at general points of @xmath499 .
let @xmath500 be a suited family of curves of genus @xmath21 defined over @xmath501 ( being @xmath242 or @xmath33 ) .
for a noetherian @xmath490-scheme @xmath502 we consider the set @xmath503 of finite @xmath502-morphisms @xmath504 of degree @xmath1 .
this functor is representable by an @xmath490-scheme @xmath505 ( see @xcite*section 4.c ) .
let @xmath506 be a morphism corresponding to a point @xmath507 $ ] on @xmath508 and consider the associated exact sequence @xmath509 from horikawa s deformation theory of holomorphic maps ( see @xcite , see also @xcite*3.4.2 ) it follows @xmath510}(h_k(\pi))$ ] is canonically identified with @xmath511 and since @xmath512 it follows @xmath513 is smooth of dimension @xmath514 . in case
@xmath497 it follows @xmath515 is a smooth real manifold of dimension @xmath514 ( of course @xmath516 need not be connected ) . in case @xmath517
then for a noetherian @xmath490-scheme @xmath502 we consider the set @xmath518 of finite @xmath502-morphisms @xmath519 of degree @xmath1 .
this functor is represented by an @xmath490-scheme @xmath520 and using the same arguments we obtain @xmath521 is smooth of dimension @xmath514 and @xmath522 is a smooth real manifold of dimension @xmath514 ( again , it need not be connected ) . because @xmath523 the non - empty fibers of @xmath524 ( and @xmath525 ) do have dimension at least 3 . hence the image of @xmath526 ( and of @xmath525 )
has dimension at most @xmath527 .
hence in case @xmath452 then a general point of @xmath528 does not belong to the image of @xmath526 ( or @xmath529 ) .
now assume @xmath454 .
from the arguments in @xcite*appendix it follows that each component of @xmath508 ( or @xmath530 ) dominates @xmath528 .
let @xmath531 ( resp .
@xmath532 ) be the closed subspace of @xmath508 ( resp .
@xmath530 ) consisting of the fibers of @xmath524 ( resp .
@xmath533 ) having dimension greater than @xmath534 .
assume it contains an irreducible component of @xmath535 .
since such component dominates @xmath528 it would follow that it has dimension more than @xmath536 , a contradiction .
hence it follows that @xmath537 ( resp .
@xmath538 ) does not contain irreducible components of @xmath508 ( resp .
@xmath530 ) , hence its dimension is less than @xmath539 .
moreover , those subspaces are invariant under complex conjugation , hence they are defined over @xmath33 .
so we obtain closed subschemes @xmath540 and @xmath541 of @xmath542 defined over @xmath33 and we obtain @xmath543 ( resp .
@xmath544 ) . for a complex curve
@xmath545 corresponding to a general point @xmath29 on @xmath528 it follows from brill - noether theory that @xmath546 has dimension @xmath547 ( if @xmath548 , otherwise it is empty ) . in case @xmath549
this gives rise to a subspace of @xmath550 ( resp .
@xmath551 ) of dimension @xmath552 .
let @xmath553 ( resp .
@xmath554 ) be the closed subspace of @xmath508 ( resp .
@xmath530 ) corresponding to linear systems @xmath333 that are not complete .
then we obtain that those are not components of @xmath508 ( resp .
@xmath530 ) because they have dimension less than @xmath539 . again
@xmath555 and @xmath556 are defined over @xmath33 and we find @xmath557 ( resp .
@xmath558 ) . in case
@xmath452 then from @xmath559 part 1 follows . also in case @xmath330 one
has @xmath560 , hence using lemma [ lemma6 ] the first statement of part 3 follows .
so assume @xmath561 . by theorem [ theorem2 ]
there is a real curve @xmath27 of topological type @xmath50 having a morphism @xmath477 of degree @xmath1 having topological degree @xmath132 .
let @xmath562 be a suited family for @xmath27 then @xmath563 defines @xmath507\in h_k(\pi)(\mathbb{r})$ ] . from the previous consideration it follows we can assume @xmath507\notin h'_k(\pi)(\mathbb{r})\cup h^2_k(\pi)(\mathbb{r})$ ]
. from @xmath507\notin h'_k(\pi)(\mathbb{r})$ ] it follows @xmath564 and from @xmath507\notin h^2_k(\pi)(\mathbb{r})$ ] it follows @xmath563 corresponds to a complete @xmath333 .
hence a general element of @xmath565 corresponds to a real curve @xmath397 of type @xmath50 with @xmath566 .
also @xmath567 implies we obtain @xmath568 , proving part 2 . in case @xmath330 and @xmath484 , by theorem [ theorem3 ] there exists a real curve @xmath27 of topological type @xmath486 having a real morphism @xmath569 of degree @xmath1 .
now we find @xmath507\in h^r_k(\pi ) ( \mathbb{r})$ ] and we can assume @xmath507\notin h'^r_k(\pi ) ( \mathbb{r})\cup h^{2,r}_k(\pi ) ( \mathbb{r})$ ] .
as in the previous case we obtain the second statement of part 3 ( again we also use lemma [ lemma6 ] ) .
[ remark9 ] the proof of theorem [ theorem7 ] is intensively based on known dimension statements in the complex case . in order to be able to use those statements
it is very important to know that @xmath570 implies @xmath571 . to make this conclusion it is very important to know that @xmath535 is smooth . in his paper @xcite mentioned in remark [ remark4 ]
the author uses similar arguments to obtain statements as those in theorem [ theorem7 ] but without considering the topological degree .
however the arguments are applied on the moduli space itself .
although the author claims the locus of @xmath1-gonal curves on the module space is smooth this is not clear ( and in general it is not true ) .
moreover for his arguments he needs that the closure of that locus is smooth at the hyperelliptic or trigonal locus . to solve this technical problem one should use a suited family containing a hyperelliptic or trigonal curve and use the space parameterizing linear systems @xmath333 on fibers of that family . indeed
, according to @xcite that space is smooth .
the construction of that space uses the relative picard scheme and we did not use that space in our arguments in order to be able to use arguments also applicable in case @xmath330 . also in his paper @xcite*end of proof of theorem 0.1 the author uses similar arguments but again not all details are described completely .
in particular the author uses an argument on hurwitz schemes but it is not clear from it that one obtains real coverings with only simple ramification ( in particular , that one obtains a point in the smooth locus of the compactified hurwitz scheme ) . those arguments in @xcite can be replaced by using the hilbert scheme parameterizing morphisms and the details described in this section . in his paper
@xcite the author also shows that ( with the same remarks on the proof as above ) , in case @xmath452 , a general @xmath1-gonal real curve of topological type @xmath50 has a unique linear system @xmath333 .
this fact can be obtained from our arguments ( and fixing the topological degree of the linear system ) . as in @xcite
one uses that a general complex @xmath1-gonal curve has a unique @xmath333 ( see @xcite ) .
since a general complex @xmath1-gonal curve has no multiple @xmath333 ( see @xcite ) one also obtains a similar result for general real @xmath1-gonal curves fixing the topological degrees .
we return to example [ examplen ] , finishing the case of curves of topological type ( 4,1,0 ) .
a non - hyperelliptic complex curve @xmath27 of genus 4 is trigonal . by riemann - roch , in case @xmath21 is a @xmath462 on @xmath27 then @xmath572 is also a @xmath462 on @xmath27 and it is well known that @xmath27 has no other linear systems @xmath462 .
hence @xmath27 has either two ( in case @xmath573 ) or one ( in case @xmath574 ) linear systems @xmath462 .
[ proposition10 ] let @xmath27 be a non - hyperelliptic real curve of topological type @xmath575 .
then @xmath27 has two real linear systems @xmath462 .
one of them has topological degree @xmath576 , the other one has topological degree @xmath577 .
let @xmath502 be a real teichmller space of real curves of topological type ( 4,1,0 ) and let @xmath578 be the associated universal family .
this space @xmath502 is a connected real manifold of dimension 9 and all real curves of topological type ( 4,1,0 ) do occur as fibers of that family
. inside @xmath502 the closed subspace @xmath579 of points having hyperelliptic fibers for @xmath580 has codimension 2 , hence @xmath581 is still connected . inside @xmath582
we consider @xmath583 ( resp .
@xmath584 ) being the subspace of points such that the fiber @xmath27 for @xmath580 has a morphism @xmath585 of degree 3 of topological degree ( 1 ) ( resp .
from @xcite we know @xmath586 and we need to prove @xmath587 too . because of theorem [ theorem2 ] we know @xmath588 , we assume that @xmath589 .
the fiber of a point belonging to the closure of @xmath583 in @xmath590 is a real curve @xmath27 that is the limit of real curves @xmath146 having a morphism @xmath591 of topological degree ( 1 ) .
this corresponds to a linear system @xmath592 on @xmath146 and the limit is a linear system @xmath462 on @xmath27 .
since @xmath27 is not hyperelliptic this linear system has no base points and it is complete , hence it corresponds to a morphism @xmath585 being a limit of those morphisms @xmath593 , hence @xmath563 has topological degree ( 1 ) too .
this implies @xmath583 is closed in @xmath581 hence it is not open in @xmath581 since @xmath581 is connected .
therefore @xmath583 has a boundary point @xmath594 in @xmath581 , let @xmath136 be the fiber of @xmath595 above @xmath594 .
now let @xmath596 be a suited morphism of curves for @xmath136 ( let @xmath597 such that @xmath598 ) .
let @xmath599 be the parameterspace for all coverings of degree 3 to @xmath3 of fibers of @xmath113 above points on @xmath490 .
let @xmath600 ( resp .
@xmath601 ) be the closed open subsets of coverings of topological degree @xmath577 ( resp .
@xmath576 ) . for a covering @xmath602
we write @xmath603 to denote the corresponding invertible sheaf . for @xmath507\in
$ ] it follows from the deformation theory of horikawa that @xmath604}(\pi _ 3)$ ] is surjective if and only if @xmath605 ( indeed @xmath606 is the tangent map of @xmath607 at @xmath507 $ ] ) . in that case for @xmath608)$ ] there is a neighborhood @xmath162 of @xmath169 in @xmath490 such that @xmath609 . in case @xmath610 then @xmath611 , hence @xmath603 is half - canonical .
let @xmath612 be the morphism of degree 3 on @xmath136 of topological degree ( 1 ) .
for each classical neighborhood @xmath162 of @xmath613 in @xmath490 there exists @xmath614 such that @xmath615 for some @xmath616 , hence @xmath617 . from the previous description of the tangent map of @xmath607 it follows @xmath618 is half - canonical on @xmath136 .
in particular @xmath136 has no @xmath462 associated to an invertible sheaf different from @xmath618 and this would imply there is no morphism @xmath619 of topological degree ( 3 ) , hence @xmath620 contradicting @xmath621 .
[ remark11 ] in contrast with our theorem [ theorem7 ] it is already mentioned that for each topological type @xmath622 there is a general real curve @xmath27 having no real @xmath462 . in @xcite
it is also proved that for each topological type @xmath623 there is a general real curve @xmath27 having no real @xmath624 ( while theorem [ theorem7 ] implies there is a general real curve @xmath27 having a real @xmath624 ) .
on the other hand , the main result of @xcite implies that there is no general real curve @xmath27 of topological type @xmath625 having no base point free @xmath624 of topological degree @xmath626 . for real curves without real points
it is proved in @xcite that for each genus @xmath21 there exist general real curves @xmath27 of topological type @xmath486 such that @xmath627 is empty for each @xmath449 .
our theorem [ theorem7 ] implies that there exist general real curves of topological type @xmath628 having a base point free @xmath629 of topological degree @xmath630 . from forthcoming work of the first author it follows that there also exist general real curves @xmath27 of topological type @xmath628 having no such linear system @xmath629 and many similar statements
[ defintion12 ] let @xmath27 be a real curve and let @xmath602 be a morphism defined over @xmath33 .
we define the _ covering number _ @xmath632 as follows .
consider the associated map @xmath633 .
if @xmath634 is not surjective then @xmath635 .
otherwise @xmath632 is the minimal number @xmath1 such that there exist connected components @xmath636 of @xmath4 such that @xmath637 . of course
, in case @xmath92 then @xmath635 by definition .
also @xmath638 if and only if there is a connected component @xmath12 of @xmath4 such that @xmath639 .
in particular @xmath638 in case @xmath640 is odd .
also in case @xmath641 then @xmath632 is equal to 0 or to 1 .
so the case with @xmath642 is the first interesting case to study the possible values for @xmath632 . in the next theorem
we prove there are no further restrictions on @xmath632 for gonality 4 .
let @xmath50 be an admissable topological type for real curves with @xmath263 .
let @xmath643 .
there exists a real curve @xmath27 of topological type @xmath50 such that there is a covering @xmath602 of degree 4 defined over @xmath33 such that @xmath644 and for each connected component @xmath12 of @xmath4 one has @xmath645 . in
case @xmath21 is even take two hyperelliptic @xmath646-curves @xmath648 and @xmath649 of genus @xmath650 having double coverings @xmath651 defined over @xmath33 with the following properties .
let @xmath652 be the connected components of @xmath653 and let @xmath654 then @xmath655 intersects @xmath656 if and only if one of the following holds : let @xmath659 and take @xmath660
. let @xmath661 .
it is a stable real curve of genus @xmath21 having a node @xmath662 and a morphism @xmath663 of degree 4 defined over @xmath33 . locally at @xmath664 the curve @xmath136 is defined over @xmath33 by the equation @xmath160 and the morphism is given by @xmath665 . using a local deformation over @xmath33 given by @xmath171 and gluing with other local coordinates
one obtains a curve @xmath146 defined over @xmath33 having a covering @xmath666 of degree 4 defined over @xmath33 such that @xmath667 .
( the details are as in construction 1 . )
the components of @xmath198 are deformations @xmath668 of @xmath669 for @xmath670 and @xmath671 and of @xmath672 which is one component @xmath673 .
clearly @xmath674 is a deformation of @xmath675 and @xmath676 is a deformation of @xmath677 for @xmath671 , @xmath670 .
it follows that their union is equal to @xmath11 but omitting one of them does not cover @xmath11 .
it follows @xmath678 ( hence @xmath146 is an @xmath646-curve ) and @xmath647 . in case @xmath21
is odd take two hyperelliptic @xmath646-curves @xmath648 of genus @xmath679 and @xmath649 of genus @xmath680 . use the notation @xmath669 for the components of @xmath653 with @xmath681 in case @xmath682 and @xmath683 in case @xmath684 .
let @xmath685 and assume @xmath655 intersects @xmath656 for @xmath686 as in the previous case and @xmath687 ( see figure [ figure 10 ] with @xmath658 ) .
let @xmath688 and take @xmath660 .
let @xmath136 be as before then arguing as before one obtains an @xmath646-curve @xmath146 of genus @xmath21 with a covering @xmath666 of degree 4 defined over @xmath33 such that the components of @xmath198 are deformations @xmath668 for @xmath689 and @xmath671 ( with @xmath690 for @xmath682 ) and @xmath691 of @xmath692 .
the images @xmath693 are deformations of @xmath677 and the image @xmath694 is a deformation of @xmath695 .
it follows @xmath647 .
now we are going to prove the theorem for @xmath646-curves in case @xmath431 . from the previous part of the proof we obtain the existence of an @xmath646-curve @xmath133 of genus @xmath696 having a covering @xmath697 of degree 4 defined over @xmath33 satisfying the following property .
let @xmath698 be the components of @xmath153 , let @xmath699 then @xmath700 intersects @xmath701 for @xmath702 if and only if and the non - empty intersections are connected . in case @xmath704
then @xmath153 has a unique component @xmath344 dominating @xmath11 such that @xmath705 and there is a connected closed subset @xmath706 such that @xmath707 if and only if @xmath708 .
let @xmath709 be different points on @xmath710 in this order ( with @xmath711 most close to @xmath712 ) and let @xmath713 ( see figure [ figure 11 ] with @xmath714 and @xmath715 ) . in case @xmath704 those are points in the inner part of @xmath716 .
let @xmath717 .
then @xmath136 is defined over @xmath33 and it has a covering @xmath718 defined over @xmath33 of degree 4 .
locally at the node @xmath719 the curve @xmath136 is defined over @xmath33 by @xmath160 and the morphism by @xmath720 .
using local deformations over @xmath33 by the equation @xmath171 with @xmath721 one obtains two new real ramification points close to @xmath722 .
gluing one obtains a curve @xmath146 defined over @xmath33 having a morphism @xmath666 of degree 4 defined over @xmath33 with @xmath723 .
( this is similar to applying @xmath724 times construction i with real ramification . ) for each @xmath725 there is a component @xmath726 of @xmath198 that is a deformation of @xmath273 . because of the chosen local deformations , the deformation of @xmath344 is a union of @xmath355 components @xmath727 .
it follows @xmath728 hence @xmath146 is an @xmath646-curve .
the images @xmath729 are deformations of @xmath700 for @xmath730 , the image @xmath731 is a deformation of the connected component of @xmath732 containing @xmath712 , the images of @xmath733 are deformations of the interval between @xmath722 and @xmath734 on @xmath735 for @xmath736 and a deformation of the connected component of @xmath737 not containing @xmath738 for @xmath739 .
now we finish the proof for the orientable case @xmath56 in case @xmath27 is not an @xmath646-curve ( in particular @xmath743 ) .
let @xmath744 , which is an integer . from the prevous part of the proof we obtain the existence of an @xmath646-curve @xmath133 of genus @xmath745 ( hence @xmath153 has @xmath746 connected components ) having a covering @xmath697 of degree 4 defined over @xmath33 satisfying the following property .
let @xmath747 be the components of @xmath153 and let @xmath699 .
then @xmath701 and @xmath748 with @xmath749 do have a point in common if and only if for @xmath756 let @xmath757 and @xmath758 with @xmath759 . let @xmath760 .
using locally real deformations at the nodes @xmath761 one obtains a real curve @xmath146 of genus @xmath762 and a covering @xmath763 of degree 4 defined over @xmath33 .
the components of @xmath198 are a deformation @xmath691 of @xmath764 and deformations @xmath726 of @xmath273 for @xmath765 .
in particular @xmath766 .
the image @xmath767 is a deformation of @xmath768 and the image @xmath769 for @xmath770 are deformations of @xmath700 .
it follows that @xmath742 .
finally we consider the non - orientable case ( hence @xmath54 ) .
first assume @xmath771 .
let @xmath133 be an orientable curve of genus @xmath772 with @xmath773 such that @xmath133 has a covering @xmath697 of degree 4 defined over @xmath33 with @xmath774 .
let @xmath775 such that @xmath776 is not a totally real divisor and let @xmath777 for a non - real pair @xmath60 .
let @xmath778 , it is a real curve of genus @xmath234 having a covering of degree 4 defined over @xmath33 . locally at the node @xmath44 the curve is defined over @xmath33 by @xmath211 and the morphism by @xmath720 . take a local deformation of the the type @xmath779 with @xmath204 and glue it to obtain a real curve @xmath146 ( this corresponds to construction ii without real ramification ) .
clearly @xmath146 is non - orientable , @xmath723 , @xmath766 and it has a covering @xmath666 defined over @xmath33 of degree 4 with @xmath774 .
finally assume @xmath780 .
since @xmath743 we can take a non - orientable real curve @xmath133 of genus @xmath772 such that @xmath773 and there is a covering @xmath697 of degree 4 defined over @xmath33 with @xmath774 .
repeating the previous construction one obtains the curve @xmath146 one is looking for .
marc coppens + katholieke hogeschool kempen + departement industrieel ingenieur en biotechniek + kuleuven dept .
wiskunde groep algebra + kleinhoefstraat 4 + b 2440 geel belgium + e - mail : [email protected] + partially supported by the fund of scientific research - flanders ( @xmath781 ) johannes huisman + universit europenne de bretagne + france + and + universit de brest ; cnrs , umr 6205 + laboratoire de mathmatiques de brest + isstb + 6 , avenue victor le gorgeu + cs 93837 + 29238 brest cedex 3 france + e - mail : [email protected] + home page : http://pageperso.univ-brest.fr/@xmath782huisman + | we consider coverings of real algebraic curves to real rational algebraic curves .
we show the existence of such coverings having prescribed topological degree on the real locus . from those existence results we prove some results on brill - noether theory for pencils on real curves . for coverings having topological degree @xmath0
we introduce the covering number @xmath1 and we prove the existence of coverings of degree 4 with prescribed covering number . 14h51 ;
14p99 real curve , linear pencil , topological degree , morphism , brill - noether theory , divisor |
evidence for the existence of the exotic baryon @xmath0 @xcite has been stimulating theoretical and experimental studies in hadron spectroscopy . the study of multiquark states in qcd will eventually lead to the understanding of the rich structure of hadronic matter .
recently , however , the existence of the @xmath0 has become controversial @xcite .
there are more than 10 experiments that indicate evidence of the @xmath0 , whereas a similar number of experiments claim null results .
in such a situation , confirmation of existence ( or nonexistence ) of the @xmath0 is urgent and crucially important . for this
, it is strongly desired to clarify the reaction mechanism for the production of @xmath0 . among various possibilities ,
a particularly interesting property that is expected to be characteristic to exotic baryons is their strong coupling to two - meson states in transitions to an ordinary baryon .
this is the subject we study in the present paper .
studying two - meson coupling is important for several reasons .
first , a heptaquark model has been proposed in the early stage of development to explain a light mass and a narrow decay width @xcite .
although a quantitative study in particular with a model of hadrons where @xmath0 is regarded as a bound state of @xmath6 system does not work with the present knowledge of hadron interactions , a two - meson contribution to the self - energy of @xmath0 has been shown to be consistent with the expected pattern of the masses of the antidecuplet members @xcite .
second , the importance of two - meson coupling has been implied from an empirical observation of the extended ozi rule @xcite .
the dominance of connected quark lines favors creation of a @xmath7 pair in the transition of @xmath8 , which is naturally associated with the coupling to two mesons , whereas a coupling to a single meson is suppressed .
finally , two - meson coupling plays an important role in reaction studies . without two - meson coupling ,
all the amplitudes for @xmath0 production are proportional to the @xmath9 coupling , which is fixed by the very small decay width of the @xmath0 .
however , two - meson coupling is determined from other source as we will see in the following , independently of the @xmath9 coupling .
therefore , even with the extremely narrow width of @xmath0 , a sizable cross section can be obtained with two - meson coupling . in ref .
@xcite , an analysis of the two - meson coupling is performed in the study of the self - energy of the @xmath0 , assuming that @xmath2 with @xmath10 in the same antidecuplet ( @xmath11 ) .
since the @xmath0 can not decay into @xmath12 channel , the coupling constants are determined from the @xmath13 decay into the @xmath1 channel and flavor su(3 ) symmetry .
two types of lagrangians are found to be important for the self - energy of the baryon antidecuplet .
it is also shown that the two - meson contribution is indeed dominant over a single - meson contribution
. however , the assumption of pure @xmath11 may not be the case in reality .
this point is clarified in ref .
@xcite , where we study the phenomenology of flavor partners for the @xmath0 .
we assign the masses of experimentally known particles in an octet - antidecuplet mixing scheme , finding good fits for @xmath2 and @xmath3 .
the decay width of the @xmath0 is also evaluated in the same scheme , and the @xmath14 case naturally explains the narrow width , in accordance with the quark model estimation @xcite . in both @xmath15 cases ,
we obtain relatively large mixing angles , which implies the importance of the representation mixing . hence ,
combining these two findings , we calculate the two - meson couplings including the representation mixing .
first we determine the coupling constants of @xmath16 from the experimental widths and separate the @xmath11 component from the @xmath17 component .
then , by using su(3 ) symmetry , the coupling constants of @xmath18 can be determined for @xmath2 and @xmath3 , including representation mixing of @xmath17 and @xmath11 .
we focus on the decay channels in which the two pions are correlated in scalar - isoscalar and vector - isovector channels , which are the main decay modes of the resonances and play a dominant role in the @xmath0 self - energy @xcite . as an application of the effective lagrangians , we perform the analysis of the @xmath19 and @xmath20 reactions .
these reactions were studied using effective lagrangian approaches @xcite .
experiments for @xmath19 have been performed at kek @xcite , and a high - resolution experiment for the @xmath20 reaction is ongoing .
we can compare the results with these experiments .
this paper is organized as follows . in the next section ,
we show the framework of representation mixing and relevant experimental information of nucleon decay . in sec .
[ sec : lag ] , the effective interaction lagrangians for nucleons and for the antidecuplet are introduced for both @xmath2 and @xmath3 cases .
the coupling constants are determined in sec .
[ sec : coupling ] by considering the decay widths of @xmath13 resonances and the self - energy of the @xmath0 . with the effective lagrangians ,
the reaction processes @xmath19 and @xmath20 are analyzed in sec .
[ sec : reaction ] .
the final section is devoted to a summary .
let us briefly review the previous study of representation mixing @xcite and summarize the experimental decays of nucleon resonances .
we have performed a phenomenological analysis on the exotic particles using flavor su(3 ) symmetry .
it is found that the masses of @xmath21 and @xmath22 are well fitted in an antidecuplet ( @xmath11 ) representation which mixes with an octet ( @xmath17 ) , with known baryon resonances of @xmath2 or @xmath3 .
the @xmath23 case gives too large a decay width for the @xmath0 , and not enough resonances are well established for @xmath24 to complete the analysis . under the representation mixing , the physical nucleon states are defined as @xmath25 two states @xmath26 and @xmath27 represents @xmath28 and @xmath10 for the @xmath29 case and @xmath30 and @xmath31 for the @xmath3 case .
the mixing angles @xmath32 can be determined by experimental spectra of known resonances as @xmath33 both angles are close to the ideal mixing @xmath34 , in which the nucleon states are classified by the number of strange quarks ( antiquarks ) . in other words , states are well mixed and the effect of mixing of states is important . using these mixing angles and decay widths of nucleon resonances ( @xmath35 ) , we can calculate the decay width of @xmath36 ( @xmath37 ) through the su(3 ) relation between the coupling constants @xmath38 where @xmath39 , @xmath40 , and @xmath41 are the coupling constants of @xmath36 and nucleon resonances . with the known coupling constants @xmath40 and @xmath41
, we obtained @xmath42 mev for @xmath2 and @xmath43 mev for @xmath14 . here
we extend this approach to the three - body decay , as shown in fig .
[ fig : sdecay ] . feynman diagram for the three - body decay of the @xmath13 resonance.,width=302 ] in table [ tbl : exp ] , we show the experimental information of the decay pattern of the nucleon resonances @xmath16 taken from the particle data group ( pdg ) @xcite . for convenience
, we refer to @xmath44 and @xmath45 modes as `` scalar '' ( @xmath46 ) and `` vector '' ( @xmath47 ) , respectively .
there is no information for the scalar decay of @xmath31 .
pdg only shows the fraction decaying via the @xmath1 mode ( 85 - 95% ) and an upper bound for @xmath48 mode ( @xmath4935% ) , although several intermediate states including @xmath44 are shown in the table . for the estimation of the coupling constants ,
we adopt the total branching ratio to @xmath1 channel as the upper limit of the branch for @xmath50 state , @xmath51 .
.experimental information of two - pion decay of nucleon resonances .
`` scalar '' represents the mode @xmath44 and `` vector '' means @xmath45 mode .
values in parenthesis are averaged over the interval quoted in pdg @xcite . [ cols="^,^,^,^,^",options="header " , ] [ tbl : summary ] for the @xmath14 case , with @xmath52 and @xmath53 , the self - energy of @xmath0 becomes @xmath54 in fig .
[ fig : selfcross ] , we plot the cross section of the @xmath19 reaction and the self - energy of @xmath0 by fixing the ratio of coupling constants .
the horizontal line denotes the factor @xmath55 , which is defined by @xmath56 using the maximum value of cross section @xmath57b estimated by the kek experiment , we have @xmath58 in the region plotted in the figure , the coupling constants do not exceed the upper bounds , but the lower limit of @xmath59 appears : @xmath60 to make @xmath61 mev , we have @xmath62 finally , we show the angular dependence of the cross sections . in figs . [
fig : anglecross1o2 ] and [ fig : anglecross3o2 ] , we plot the angler dependence of the differential cross sections at the energy of kek experiment : @xmath63 mev for @xmath19 and @xmath64 mev for @xmath20 . for the @xmath2 case , the contribution from the @xmath46 term has no angular dependence , whereas the @xmath47 term shows a forward peak , owing to the @xmath65-channel exchange of the vector meson propagator .
because of the interference of the two amplitudes , the total result becomes zero at @xmath66 for the @xmath19 reaction . for the @xmath14 case ,
the @xmath46 term varies linearly in @xmath67 , leading to a backward peak .
the @xmath47 term shows a forward peak , which is enhanced by the vector meson propagator .
interference of the two amplitude leads to a clear forward peak for both @xmath19 and @xmath20 reactions .
( color online ) angular dependence of the differential cross section for @xmath19 at @xmath63 mev ( left ) and for @xmath20 at @xmath64 mev ( right ) for the @xmath2 case with @xmath68 and @xmath69 , including a hadronic form factor .
the thick line shows the result with full amplitude .
dash - dotted and dashed lines are the contributions from @xmath46 and @xmath47 terms , respectively.,width=302 ] ( color online ) angular dependence of the differential cross section for @xmath19 at @xmath63 mev ( left ) and for @xmath20 at @xmath64 mev ( right ) for the @xmath14 case with @xmath70 and @xmath71 , including a hadronic form factor .
the thick line shows the result with full amplitude .
dash - dotted and dashed lines are the contributions from @xmath46 and @xmath47 terms , respectively.,width=302 ] in this subsection , we briefly discuss the possible effect from the born terms , as shown in fig .
[ fig : born ] , which have not been taken into account in the present studies . however , there are reasons that the born terms are not important in the present reactions . first , the born terms are proportional to the decay width of @xmath0 and therefore suppressed if the decay width of the @xmath0 is narrow .
second , in the energy region of @xmath0 production , the energy denominator of the exchanged nucleon suppresses the contribution , especially for the @xmath46-channel term in the @xmath19 reaction .
here we would like to confirm this explicitly .
born terms for the reaction : @xmath72-channel diagram for @xmath20 ( left ) ; @xmath46-channel diagram for @xmath19 ( right).,title="fig:",width=151 ] born terms for the reaction : @xmath72-channel diagram for @xmath20 ( left ) ; @xmath46-channel diagram for @xmath19 ( right).,title="fig:",width=151 ] at the tree level , there are @xmath46- , @xmath65- , and @xmath72-channel diagrams . however , assuming @xmath73 for @xmath0 , there is only a @xmath72 channel in @xmath20 ( fig .
[ fig : born ] , left ) , whereas there is only an @xmath46 channel in @xmath19 ( fig .
[ fig : born ] , right ) .
for these terms , we need the yukawa couplings such as @xmath74 and @xmath75 couplings .
there are two schemes to introduce the yukawa couplings , namely , pseudoscalar ( ps ) and pseudovector ( pv ) schemes .
for the construction of the born amplitude , it is reasonable to rely on chiral symmetry , where the two schemes should provide the same results . in this case
the meson - baryon scattering amplitude should be a quantity of @xmath76 or higher , where @xmath77 is a momentum of mesons . in the pv scheme , since each @xmath78 coupling is of @xmath76 , the born amplitude behaves as @xmath79 , consistent with this observation .
in contrast , a naive construction of the born term in the ps scheme leads to an amplitude of @xmath80 .
it is well - known that a scalar exchange term cancel the term of @xmath80 .
however , the interaction of the scalar channel is not well understood .
therefore , we adopt the pv scheme in the following study .
another advantage of the pv scheme is that it can be extended easily to the @xmath14 case , while it is not so simple in the ps scheme @xcite . in this respect ,
our method differs from the previous study of similar reactions @xcite , in which the ps scheme was used .
the interaction lagrangians for @xmath29 are @xmath81 the fields @xmath82 and @xmath83 are defined in eq . , and the kaon field is defined as @xmath84 and the coupling constants are determined as @xmath85 which is determined by @xmath86 mev , and we use @xmath87 the amplitude for @xmath88 is given by @xmath89 and for @xmath90 , @xmath91 in fig . [ fig:1o2_b2 ]
we show the results including born terms .
we can observe that the effect of born terms is indeed small in both reactions .
( color online ) total cross section for the @xmath2 case with @xmath68 and @xmath69 , including hadronic form factor and born terms .
the thick line shows the result with full amplitude .
dash - dotted , dashed , and thin lines are the contributions from @xmath46 , @xmath47 , and born terms , respectively .
the result without born terms are shown by the long - dashed lines.,width=302 ] for the @xmath14 case , the interaction lagrangian can be written as @xmath92 with the same @xmath75 vertex in eq . .
in the nonrelativistic expansion this term yields a @xmath93-wave coupling so that the square of momenta appears in the vertex .
it reduces the contribution further than the @xmath29 case , and therefore , the effect of born terms for @xmath14 is also small .
in this paper , we studied the two - meson couplings of @xmath0 for @xmath2 and @xmath3 .
the effective interaction lagrangians for the two - meson coupling were given , and these coupling constants were determined based on the @xmath17-@xmath94 representation mixing scheme , by using information of the @xmath16 decays .
these values were further constrained in order to provide appropriate size of the self - energy of the @xmath0 . finally , we applied the effective lagrangian to the meson induced reactions @xmath19 and @xmath20 .
we found that there is an interference effect between the two amplitudes of the scalar and vector types , which can help to explain the very small cross section for the @xmath19 reaction observed at kek @xcite . in this case , reflecting the symmetry under exchange of two amplitudes , the large cross section for @xmath20 reaction was obtained as a consequence of the interference .
the interference occurs in both @xmath29 and @xmath3 cases . in table
[ tbl : summary ] , we summarize the results obtained in the present analysis . for a given set of coupling constants , the upper bound of the cross section of the @xmath20 reaction is estimated by maximizing the interference effect .
we observe that the large cross section of the order of millibarns for @xmath20 is obtained for the @xmath29 case , whereas the upper limit of the cross section is not very large for @xmath3 case .
therefore , if large cross sections are observed in the @xmath20 reaction , it would indicate @xmath2 for the @xmath0 . for completeness
, we would like to mention the case where the cross sections for both @xmath19 and @xmath20 reactions are small .
if the cross section of @xmath20 reaction is also small , it could be explained by the small coupling constants and is not an interference effect . for the @xmath2 case
, both coupling constants can be zero within the experimental uncertainties .
however , for the @xmath3 case , there is a lower limit for the @xmath95 , which means that the lower limit is also imposed for the cross sections .
we search for the set of coupling constants that provide the minimum value for the @xmath20 cross section , keeping a @xmath19 cross section to be less than @xmath96b .
we obtain @xmath97b with @xmath98 and @xmath99 .
however , one should notice that the small coupling constants do not guarantee the dominance of two - meson coupling , and the born terms and interference effect may play a role , which is beyond our present scope .
the present analysis provides an extension of effective interactions obtained in ref .
@xcite with representation mixing and @xmath14 .
it is also interesting to apply the present extension to the study of the medium effect of @xmath0 @xcite and the production of @xmath0 hypernuclei @xcite . from the experimental point of view
, the cross section of @xmath20 reaction is of particular importance to the present results . to perform a better analysis for the two - meson coupling ,
more experimental data for three - body decays of nucleon resonances are strongly desired .
we acknowledge manuel j. vicente vacas , eulogio oset , seung - il nam , and koji miwa for useful discussion and comments .
one of the authors ( t. h. ) thanks the japan society for the promotion of science ( jsps ) for support .
this work is supported in part by the grant for scientific research [ ( c ) no.17959600 , t. h. ] and [ ( c ) no.16540252 , a. h. ] from the ministry of education , culture , science and technology , japan .
leps , t. nakano _ et al .
lett . * 91 * , 012002 ( 2003 ) .
k. h. hicks , prog . part .
* 55 * , 647 ( 2005 ) .
p. bicudo and g. m. marques , phys .
d * 69 * , 011503(r ) ( 2004 ) .
t. kishimoto and t. sato , hep - ex/0312003 .
f. j. llanes - estrada , e. oset , and v. mateu , phys .
c * 69 * , 055203 ( 2004 ) .
p. bicudo , phys .
d * 70 * , 111504(r ) ( 2004 ) .
f. huang , z. y. zhang , and y. w. yu , hep - ph/0411222 .
a. hosaka , t. hyodo , f.j .
llanes - estrada , e. oset , j. r. pelez , and m. j. vicente vacas , phys .
c * 71 * , 045205 ( 2005 ) .
d. p. roy , j. phys .
g * 30 * , r113 ( 2004 ) .
t. hyodo and a. hosaka , phys .
d * 71 * , 054017 ( 2005 ) .
a. hosaka , m. oka , and t. shinozaki , phys .
d * 71 * , 074021 ( 2005 ) .
t. hyodo , a. hosaka , and e. oset , phys . lett . *
b579 * , 290 ( 2004 ) .
w. liu and c. m. ko , phys .
c * 68 * , 045203 ( 2003 ) .
y. oh , h. kim , and s. h. lee , phys . rev .
d * 69 * , 074016 ( 2004 ) .
kim , and s. h. lee , phys .
d * 69 * , 014009 ( 2004 ) .
k. imai , talk given at exotic hadrons workshop , may 27 , 2005 ; http://www.hepl.phys.nara-wu.ac.jp/exohad05/files/exohad05imai.pdf , k. miwa ( private communication ) .
particle data group , s. eidelman _
b592 * , 1 ( 2004 ) .
w. rarita and j. s. schwinger , phys . rev . * 60 * , 61 ( 1941 ) .
s. theberge , a. w. thomas , and g. a. miller , phys .
d * 22 * , 2838 ( 1980 ) .
s. i. nam , a. hosaka , and h. c. kim , phys .
b579 * , 43 ( 2004 ) .
nam , a. hosaka , and h .- c .
kim , phys .
d * 71 * , 114012 ( 2005 ) .
d. cabrera , q. b. li , v. k. magas , e. oset , and m. j. vicente vacas , phys .
b608 * , 231 ( 2005 ) .
h. nagahiro , s. hirenzaki , e. oset , and m. j. vicente vacas , phys . lett .
* b620 * , 125 ( 2005 ) . | we evaluate two - meson couplings of @xmath0 , using experimental information of nucleon resonances decaying into @xmath1 channels , in which the two pions are in scalar- and vector - type correlations .
we examine two assignments of spin and parity of @xmath2 and @xmath3 , for which the experimental spectra of known resonances with exotic baryons are properly reproduced by an octet - antidecuplet representation mixing scheme . with the obtained coupling constants , total cross sections of the reactions @xmath4 and @xmath5 are calculated .
substantial interference of two terms may occur in the reaction processes for the @xmath2 case , whereas the interference effect is rather small for the @xmath3 case . |
the authors would like to thank v. barocas , t. stylianopoulos , e. a. sander , e. tuzel , h. zhang , h. othmer , t. jackson , p. smereka , r. krasny , f. mackintosh , a. kabla , r. lee , m. dewberry , l. kaufman and l. jawerth , and for their discussions .
this work is supported by nih bioengineering research partnership grant r01 ca085139 - 01a2 and the institute for mathematics and its applications . | we study the micromechanics of collagen - i gel with the goal of bridging the gap between theory and experiment in the study of biopolymer networks .
three - dimensional images of fluorescently labeled collagen are obtained by confocal microscopy and the network geometry is extracted using a 3d network skeletonization algorithm .
each fiber is modeled as a worm - like - chain that resists stretching and bending , and each cross - link is modeled as torsional spring .
the stress - strain curves of networks at three different densities are compared to rheology measurements .
the model shows good agreement with experiment , confirming that strain stiffening of collagen can be explained entirely by geometric realignment of the network , as opposed to entropic stiffening of individual fibers .
the model also suggests that at small strains , cross - link deformation is the main contributer to network stiffness whereas at large strains , fiber stretching dominates .
since this modeling effort uses networks with realistic geometries , this analysis can ultimately serve as a tool for understanding how the mechanics of fibers and cross - links at the microscopic level produce the macroscopic properties of the network . while the focus of this paper is on the mechanics of collagen
, we demonstrate a framework that can be applied to many biopolymer networks .
collagen is the most abundant animal protein @xcite and its mechanics have been studied in great detail @xcite .
it takes on many morphologies , including skin , tendons , ligaments , individual fibers , and gels .
of particular interest is the mechanics of collagen - i gels , shown in figure [ fig : summary]a .
these gels provide a relatively simple structure that can be noninvasively observed by confocal microscopy @xcite and used as a scaffold for growing artificial tissues @xcite , and as a 3d environment for studying cell motility @xcite and tumor invasion @xcite .
a critical first step in understanding these systems is to develop a model for the collagen gel alone . in this paper
we give a successful theoretical model of the micromechanics of realistic networks .
collagen - i gels belong to a class of materials known as biopolymers .
other examples include actin , found in the cytoskeleton and fibrin , a component of blood clots .
a common feature of biopolymer networks is their ability to strain stiffen by 2 - 3 orders of magnitude at large strains ( figure [ fig : summary]f ) .
the cause of this strain stiffening is not well understood .
storm et al .
@xcite attributed strain stiffening in all biopolymer networks to the pulling out of entropic modes of individual filaments .
their calculation required the assumption that deformations are affine .
later , heussinger et al .
@xcite showed how one could deconstruct the network deformation into a set of floppy modes .
they concluded that accounting for the non - affinity was necessary in describing the elastic properties of the network .
onck and colleagues @xcite have proposed the alternative hypothesis that strain stiffening is due to the rearrangement of the fibers as the network is strained ( figures [ fig : summary]d and [ fig : summary]e ) . resolving this debate has been difficult since almost all theoretical analysis has been on artificially generated networks .
the few examples of quantitative comparisons to experiment in the literature @xcite are not able to quantitatively fit the full stress - strain response of the gel at varying densities using a single set of parameters .
m @xmath0 25.6 @xmath1 m @xmath0 25.6 @xmath1 m ) .
a ) maximal intensity projection along the axis that is perpendicular to the focal plane of the microscope .
b ) projection of 3d network extracted by fire .
c ) reduced network , where elements that do nt contribute to the network stiffness have been removed for improved computational efficiency .
d ) deformation after 50% tension .
e ) deformation after 50% shear .
f ) comparison of the stress - strain response between the model and experiment .
[ fig : summary],width=288 ] in this paper , we bridge the gap between model and experiment .
three dimensional images of fluorescently labeled collagen gels at different densities are imaged by confocal microscopy ( figure [ fig : summary]a ) and the network geometry is extracted using a custom fiber extraction ( fire ) algorithm ( figure [ fig : summary]b ) @xcite . the gel is modeled as a random network of cross - linked fibers , as described below , and the stress - strain response is compared to that measured by an ar - g2 rheometer .
good agreement between model and experiment is obtained by fitting a single parameter , the cross - link stiffness .
the experiments are described in detail in @xcite . in the model for the collagen gel ,
each fiber is treated as a discrete worm - like - chain ( wlc ) @xcite which resists stretching and bending , and each cross - link is treated as a torsional spring , thus more stiff than a freely rotating pin joint but less stiff than a welded joint of fixed angle .
the stretching modulus of an individual fiber is given by @xmath2 , where @xmath3 is the young s modulus and @xmath4 is the cross - sectional area .
the young s modulus of a fiber in aqueous conditions has been estimated to be between 30 - 800 mpa @xcite and we use a modulus of 50 mpa , which fits the data well and is also close to the value chosen by stylianopoulous et al .
( 79mpa ) to fit their model @xcite .
it has been shown that a single fiber will stiffen by a factor of 2 - 4 when strained @xcite .
we choose here to use a constant @xmath3 both to reduce the number of parameters in the model and to see if geometric reorientation of the network is enough to explain strain stiffening .
stylianopoulos and barocas @xcite also explored the bilinear and exponential constitutive relations for the individual fibers and observed only minor effects on the macroscopic network behavior .
the radius of each fiber is @xmath5 nm @xcite .
the bending modulus of the fiber is given by @xmath6 pn-@xmath1m@xmath7 , where @xmath8 @xcite .
no cross - linking agent has been added to the gel and very little is known about the nature of the naturally formed collagen cross - links .
we find that we can fit all the data by setting the torsional spring stiffness to @xmath9 pn-@xmath1 m .
to compare @xmath10 to @xmath11 , we consider @xmath12 , where the mean cross - link spacing is given by @xmath13 @xmath1 m @xcite .
thus , we find that @xmath14 .
one possible reason for a larger @xmath11 could be an increase in fiber radius near the cross - links by a factor of 2 , since bending stiffness scales by @xmath15 .
we assume that in the undeformed state of the network , there are no internal stresses .
thus the fibers have an innate curvature and the cross - links have an equilibrium angle equal to that in their initial configuration .
we ignore entropic contributions to the fiber mechanics . while the geometric persistence length of these fibers has been measured to be 20 @xmath1 m @xcite , the thermal persistence length is much longer @xmath16 cm .
furthermore , in the case that the strain stiffening is dominated by thermal compliance , one would expect to see a decrease in the yield strain with increasing concentration @xcite .
collagen gels , however , have been shown to have a constant yield strain of about 60% for a wide range of concentrations @xcite .
thus the total energy in the network for a given configuration is given below .
@xmath17 here @xmath18 is the number of elements of type @xmath19 , which denotes stretching , bending , and cross - link , @xmath20 is the length of stretching element , @xmath21 and @xmath22 are the bending and cross - link angles respectively , and @xmath23 indicates the difference between the deformed and undeformed state . to calculate the stress - strain relationship of our model network
, we perform a series of 18 incremental strain steps by imposing a small deformation on one face , @xmath24 while holding the opposite face , @xmath25 fixed .
we impose two types of deformations : tension ( figure [ fig : summary]d ) and shear ( figure [ fig : summary]e ) . in a tensile deformation
, we allow the vertices on @xmath25 and @xmath24 to move freely in directions perpendicular to the imposed strain to allow for perpendicular contraction that is seen to occur in these experiments @xcite . in experiments of this type
, the distance between @xmath25 and @xmath24 is on the order of centimeters and the simulated network represents a small region near the center of a sample . in shear , we do not allow the boundary nodes on @xmath25 and @xmath24 to move freely .
we compare the shear results to cone - plate rheometer experiments , where the shear faces are bound to the rheometer . here
, the distance between @xmath25 and @xmath24 is 109 @xmath1 m , and the simulated network is one fourth the length of the experimental sample between the boundaries . in both deformations , all other nodes , including those on the four remaining faces of the network , are free to move .
the minimum energy state of the network at an imposed strain @xmath26 is found using a conjugate gradient method developed by hager and zhang @xcite .
the stress required to hold the network in its current configuration is given by @xmath27 , where @xmath4 denotes the area of @xmath25 .
the results shown in figures [ fig : strainsweep ] and [ fig : ge ] are averaged over all four extracted networks and over all 6 principle shear deformations in the sheared network and all 3 principle tensile directions in the stretched network .
the results from our shear cell experiments are given in figures [ fig : strainsweep ] and [ fig : ge]a .
in addition , in figure [ fig : ge]b , we also present the previously reported tensile modulus of large samples that are centimeters in length @xcite .
figure [ fig : strainsweep]a shows a strain sweep from 0.5% to 100% shear strain in both the model and cone - plate rheometer experiments . at small strains ,
the stress - strain response is linear , as expected , and at larger shear strains , the stress - strain response appears cubic . in figure
[ fig : ge]a , we show that the small strain modulus scales by @xmath28 , where @xmath29 is the collagen density . at this time ,
it is not possible to verify the power law scaling in the model since only densities of 0.5 , 1.2 , and 1.4 were observed .
the fluorescent labeling of the network changes the polymerization properties of the network , causing it to clump at higher densities .
we use this scaling relationship to collapse the curves in figure [ fig : strainsweep]b .
the close agreement between model and experiment indicates that strain stiffening due to the geometric rearrangement of the collagen fibers is enough to explain the strain stiffening seen in experiments . ,
and the number denotes the collagen density .
a ) unscaled results .
note the good agreement between model and data at small and large strains .
b ) when the curves are scaled by @xmath30 relatively good data collapse is achieved .
we denote lines of slope @xmath31 and @xmath32 to guide the eye . at large strains ,
scaling breaks down and the low density curves overtake the high density curves because at large strains , stiffness scales linearly with density .
thus this rescaling serves mainly as a visualization tool and does not represent a true data collapse .
[ fig : strainsweep],width=192 ] is observed for the experiment .
b ) the large strain tensile modulus from the model and from the experiments of roeder et al . @xcite .
results differ by a factor of 2.5 , which is reasonable since the two experimental protocols were different @xcite .
[ fig : ge],width=288 ] at large strains ( @xmath33 in tension , @xmath34 in shear ) , the stress - strain curve of the model becomes linear again , though with a much steeper slope . in figure [ fig : ge]b , we compare the large strain tensile behavior of the model to the experiments of roeder et al .
@xcite .
while our model underestimates their experimental measurement by a factor of 2.5 , we find this to be reasonable since the two experiments used different collagen protocols . in particular , different buffers were used . in figure
[ fig : strainsweep ] , we also explore the case where @xmath35 , such that we have only a network of springs connected at freely rotating pin joints . at low strains ,
the network can be deformed without exerting any stress , but at strains higher than 25% , we see that this simplification adequately describes the gel .
a topic of investigation explored by many is the validity of the assumption that these networks deform affinely @xcite . for brevity
, we state only that the deformations are highly nonaffine in these simulations , as evidenced by a visual inspection of figure [ fig : summary]d and [ fig : summary]e where it is obvious that many fibers leave the volume defined by an affine deformation . in summary , we have presented a microstructural model of a 3d biopolymer gel using a network geometry that is based on the true network architecture .
it differs from previous work in that we use realistic network architectures that have been extracted using the fire algorithm .
we specifically focus on the mechanics of collagen - i networks , but emphasize that this modeling approach is generalizable to other biopolymer networks .
the model has three parameters : @xmath36 .
the fiber radius and tensile modulus can be measured experimentally and the model uses realistic parameters .
the cross - link torsional spring constant must be fit to the data and we used @xmath14 . fitting this single parameter
gives the right strain stiffening behavior for networks at three different densities at strains that vary from 0.5% to 50% .
this result lends support to the hypothesis put forward by onck et al .
@xcite that strain stiffening in polymer networks , and particularly collagen - i gels , is governed by rearrangement of the gel .
another finding of the model is that at strains greater than 25% , the stiffness of the gel is governed almost entirely by stretching of the fibers .
this result is relevant for collagen because cells embedded in these gels are seen to produce deformations of this order of magnitude @xcite . in modeling large systems of this type where the strains are large
, it may be sufficient to treat each fiber as a spring rather than a wlc in order to reduce the computation time .
this work also demonstrates that an understanding of the cross - link mechanics in these systems is critical to understanding their mechanical properties , as has been seen previously @xcite . in much of the theoretical work that has been done on random stick networks ,
the cross - links are treated either as freely rotating pin joints , or welded joints of fixed angle @xcite .
while these are sensible simplifying assumptions in developing a theory , they are not adequate for describing actual networks .
we note that this model has been designed to capture the short time scale ( @xmath37 1 hr ) behavior of the network , where it behaves as an elastic solid .
such behavior requires that the cross - links be relatively fixed @xcite .
this simplified model provides a starting point in the development of a more complete model of collagen gel . ultimately , a more sophisticated approach , such as that taken by rodney et al .
@xcite will be necessary to capture the full dynamic behavior of the gel , where cross - links are allowed to slip and break . |
devising methods for analyzing and predicting time series is currently considered one of the most important challenges in chaotic time - series analysis ( eg .
, see refs .
[ 1 - 3 ] ) . in general , chaotic behavior is observed in relation with nonlinear differential equations and maps on manifolds . _
times series may be construed as being the projections of manifolds onto coordinate axes_. much work in nonlinear dynamics has focused on the building of appropriate model(s ) of the underlying physical process from a time series , with the objective of predicting the _ near - future _ behavior of dynamical systems .
the first step in formulating predictive models is that of specifying / estimating a suitably parameterized nonlinear function of the observation .
this is followed by estimating the parameters of this function .
in general , prediction models are formulated on the basis of the systematic and accurate identification of a _ working hypothesis _ [ 4 ] .
this hypothesis is represented by a set of parameters that form an ansatz .
this paper obtains the coefficients of such an ansatz , which possess information about the data set(s ) , via recourse to a fisher information measure ( fim , hereafter ) based inference procedure .
the leitmotif for obtaining the _ working hypothesis _ by employing an inference procedure is to formulate a prediction model , based on the famed embedding theorem of takens [ 5 , 6 ] .
the conceptual sophistication underlying the takens theorem renders the prediction problem to become an instance of extrapolation .
currently , some of the prominent prediction models based on information theory ( it , hereafter ) are : @xmath0 the framework of plastino et .
al . [ 7 - 10 ] using the maximum entropy ( maxent , hereafter ) method of jaynes [ 11 ] , and @xmath1 the nonparametric models by principe et .
al , ( eg . see [ 12 , 13 ] ) . the work presented herein belongs to a class of models known as _ pseudo - inverse _ models , for reasons described in section 2 and 3 of this paper .
such models have been successfully employed to forecasting tasks in a number of disciplines which include nonlinear dynamical systems [ 7 , 8 ] , financial data forecasting [ 8 ] , prediction of tonic - clonic epileptic seizures from real - time electroencephalogram ( eeg ) data [ 9 ] , and even fraud analysis ( the london interbank offered rate ( libor ) manipulations ) [ 10 ] .
generally , predictive models are of two types , viz .
_ global _ and _ local _ ( see for example ref .
global models are based on training data collected from across the phase space . on the other hand , in local models ,
the training is accomplished by measurements providing data lying in the immediate vicinity of a specific / localized region of the phase space .
_ pseudo - inverse predictive models , including the one presented herein , are essentially global models which possess local characteristics _ [ 10 , 15 ] .
time series prediction has its roots in the theory of optimal filtering by wiener [ 16 ] . in recent times ,
forecasting of chaotic time series has hitherto largely utilized artificial neural networks ( ann s , hereafter ) and other learning paradigms .
commencing from the seminal radial basis function model of casdagli [ 17 ] , some of the notable attempts to study chaotic time series comprise ( but are not limited to ) the time delayed neural network architectures [ 18 ] , recurrent ann s [ 19 ] , maximum entropy ann s [ 20 ] , and support vector machines [ 21 ] . within the perspective of physics - based models ,
the works of crutchfield and mcnamara [ 22 ] and farmer and sidorowich [ 23 ] constitute some of the most prominent efforts .
fim - based studies have recently been acquiring prominence across a spectrum of disciplines ranging from physics and biology to economics ( for eg . ,
see [ 24 ] ) . the prediction model presented in this paper
comprises of two phases : @xmath0 the modeling phase and @xmath1 the prediction phase .
the task of the modeling phase is to obtain the coefficients of the ansatz that suitably parameterizes the nonlinear function of the observed time series ( see section 2 of this paper ) .
this phase establishes the _ working hypothesis _ , and is accomplished with the assistance of the training data .
the prediction phase then generates forecasts based on the set of coefficients obtained in the modeling phase .
the leitmotif for the fim - based model employed in this paper is two - fold .
first , it provides the framework to endow the modeling phase with a quantum mechanical ( qm , hereafter ) connotation .
this is in accordance with wheeler s hypothesis of establishing an information - theoretical foundation for the fundamental theories of physics [ 25 ] , and is accomplished by recourse to the minimum fisher information ( mfi , hereafter ) principle of hber [ 26 , 27 ] .
variational extremization of the fim subject to least squares constraints results in a sturm - liouville equation in a vector setting , hereinafter referred to as the time independent schrdinger - like equation .
consequently , i ) the probability density function ( pdf , hereafter ) of the coefficients of the ansatz , and ii ) the constraint driven pseudo - inverse condition ( that yields the inferred estimate coefficients , fundamental for the _ working hypothesis _ ) , can be specified not only via gaussian ( maxwell - boltzmann ) pdf s [ which are _ equilibrium _ distributions ] , but also in terms of _ non - equilibrium _ distributions [ 24 , 28 - 30 ] , comprising of hermite - gauss polynomials .
this greatly widens the scope of the works presented in refs .
[ 7 - 10 ] , and is accomplished in this paper with the aid of the qm virial theorem [ 31 , 32 ] for normal distributions .
note that in inference problems involving the fim , the gaussian pdf s are obtained as solutions to the lowest eigenvalue by solving the time - independent schrdinger - like equation in section 3 of this paper as an eigenvalue problem , and correspond to the _ ground state _ of the physical schrdinger wave equation ( swe , hereafter ) .
further , the non - equilibrium pdf s correspond to the higher - order eigenvalue solutions of such swe , and are linked to _ excited states _ of the physical swe ( see , for eg . [ 33 , 34 ] ) . from a _
practical _ perspective , this enables the performance of the modeling phase and the concomitant prediction phase to be systematically categorized in terms of an established physics - based framework .
next , the reciprocity relations and the legendre transform structure ( lts , hereafter ) , together with the concomitant information theoretic relations for the fim , in a vector setting and for least squares constraints , are derived .
prior studies have derived reciprocity relations and lts for the fim model [ 35 ] and have analyzed such relations [ 36 - 39 ] .
recently , these works have been qualitatively extended to the case of the relative fisher information ( rfi , hereafter ) [ 40 - 42 ] by venkatesan and plastino by deriving the reciprocity relations and lts [ 43 ] . a connection between the celebrated hellmann - feynman theorem , the reciprocity relations , and lts for the rfi has been established in [ 44 ] , in addition with a unique inference procedure to obtain the energy eigenvalue without recourse to solving the time - independent schrdinger - like equation .
these prior works differ from the analysis presented in this paper in two significant aspects - @xmath0 they treat the scalar case and @xmath1 the prior knowledge encoded in the observed data are introduced as constraints into the variational extremization procedure in the form of expectations of the powers of the scalar independent variable .
the reciprocity relations and the lts for the time - independent schrdinger - like equation derived in this paper , despite possessing a vector form and least squares constraints , mathematically resemble those derived in [ 35 ] .
this augurs well with regards to the possibility of translating the entire mathematical structure of thermodynamics into the fisher - based model presented in this paper .
the distinctions in the reciprocity relations and lts derived in this paper vis - - vis earlier referenced works [ 36 - 39 ] result in the information theoretic relations derived from these relations being qualitatively different from those obtained in the scalar case .
this fact evidences the distinction between the results presented in this paper and those demonstrated in refs .
[ 36 - 39 ] , based on physics and on systems theoretic [ 45 ] considerations . _ of interest is an expression that infers the fim of the modeling phase just in terms the observed data , hereafter referred to as the empirical fim_. such relation , which is a solution of a linear pde derived from the reciprocity relations together with the lts that infers the fim without recourse to the time - independent schrdinger - like equation , has no equivalent in the maxent model
. the goals of this paper are * @xmath0 to provide an overview of the solution procedure .
this is done in section 2 , * @xmath1 to : @xmath3 introduce the mfi principle in a vector setting and using least square constraints , @xmath4 derive a systematic procedure for the inference of exponential pdf s of the modeling phase with the aid of the qm virial theorem , and @xmath5 obtain the constraint driven pseudo - inverse condition that yields the estimate of the coefficients comprising the _ working hypothesis _ ( see section 2 of this paper ) by invoking the qm virial theorem .
this three - fold objective is performed in section 3 .
note that for normal pdf s the solutions of the mfi and maxent principles are known to coincide [ 24 , 46 ] .
this paper focuses on the normal distribution to demonstrate that the results of the maxent model can be derived from qm considerations and interpreted within the framework of estimation theory , which is not possible within the ambit of the maxent framework , * @xmath2 to derive the reciprocity relations and the lts for the fim in a vector setting using square constraints , analyzing the concomitant information theoretic relations .
the _ empircal fim _ is derived , and a preliminary analysis of its properties is performed .
this is accomplished in section 4 , * @xmath6 to computationally demonstrate the efficacy of the prediction framework for the mackey - glass ( m - g , hereafter ) delay differential equation ( dde , hereafter ) [ 47 ] , for a 5 minute electrocardiogram ( ecg , hereafter ) segment of record 207 of the mit - beth israel deaconess hospital ( mit - bih , hereafter ) arrythmia database [ 48 ] ( considered to be one of the most challenging records in the mit - bih arrhythmia database ) for the modified lead ii ( mlii , hereafter ) , and for the single ecg signal in record cudb / cu02 of the around 8.5 minute creighton university ventricular tachyarrhythmia ( vta , hereafter ) database [ 49 ] .
the ecg data are obtained from the physionet online repository [ 50 ] .
this is demonstrated in section 5 of this paper .
the leitmotif of this exercise is as follows .
+ + an obvious practical advantage of the pseudo - inverse model presented in this paper over a least squares approach in ordinary euclidian space is that the former requires the moore - penrose pseudo - inverse [ 51 ] of the embedding matrix @xmath7 ( defined in section 2 of this paper ) and therefore , the estimate of the coefficients of the ansatz comprising the @xmath8 @xmath9 ( see sections 2 and 3 of this paper ) derived via inference from the training data .
this can be achieved even when * w * is nearly singular .
the fact that the estimates @xmath9 are defined even when * w * is singular ( or nearly singular ) can in principle result in very volatile forecasts , on account of ill - conditioning .
note that ill - conditioning occur in the presence of a near - singular * w * , which in turn might occur if many lags of the observed data @xmath10 are present .
+ + the leitmotif for the choice of the benchmarks on which to test the prediction model is as follows .
the m - g equation with delay @xmath11 has a high embedding dimension [ 18 ] .
thus * w * displays more lags as compared to most prominent models describing low dimensional chaos [ 3 , 14 ] . as is described in section 2 of this paper , the rationale being that the number of lags in * w * depends upon the embedding dimension .
as evidenced in section 5 of this paper , the forecast of the m - g dde is stable and accurate .
next , ecg s of patients suffering from serious cardiac related ailments possess artifacts which are representative of various conditions of a diseased heart .
these artifacts are noted in the reference annotations as episodes ( transients ) . it is demonstrated that even for the most challenging cases , the model presented in this paper accurately forecasts these episodes without any signs of volatility , thereby demonstrating the accuracy and robustness of the pseudo - inverse model .
this is established for cases where the original signal possesses highly erratic / volatile behavior .
numerical examples for exemplary cases are provided . to the best of the authors knowledge , these objectives have never hitherto been accomplished .
given a signal * x * from an unknown dynamical system @xmath12 , the corresponding time series consists of a sequence of @xmath13 _ stroboscopic _ measurements : @xmath14 made at intervals @xmath15 .
the state space is reconstructed using the time delay embedding [ 1 , 5 , 6 ] , which uses a collection of coordinates with time lag to create a vector in @xmath16-dimensions , on a system considered to be in a state described by @xmath17 at discrete times @xmath18 where @xmath19 is the time lag , and @xmath16 is the embedding dimension of the reconstruction .
it is known from takens theorem ( eg . see refs . [ 5 , 6 ] ) that for flows evolving to compact attracting manifolds of dimension @xmath20 ; if @xmath21 for the forecasting time @xmath22 , @xmath23 ( time samples in this paper ) , there exists a functional form of the type @xmath24 where @xmath25,\ ] ] and @xmath26 .
a _ non - unique _
ansatz for the mapping function of this form ( employing the einstein summation convention ) is specified as [ 9 ] @xmath27 where @xmath28 and @xmath29 is the polynomial degree chosen to expand the mapping @xmath30 .
the number of parameters in ( 4 ) corresponding to @xmath31 terms ( the degree ) , is the combination with repetitions @xmath32 the length of the vector of parameters , @xmath33 is @xmath34 other forms of ansatz are encountered in [ 52 ] .
it is important to note that specifying an ansatz of a form , such as that defined in ( 4 ) , has its roots in signal processing [ 53 ] . as an information recovery criterion
, the vector of coefficients @xmath33 is obtained via inference by invoking the mfi principle .
the objective is to achieve a model possessing high predictive ability .
computations are made on the basis of the information given by @xmath35 points of the time series .
these constitute the _ training data _ obtained from the observed signal , whose utility is to infer the coefficients @xmath33 .
@xmath36 ; n=1, ... ,m.\ ] ] given the data set ( 7 ) , the parametric mapping ( 2 ) can be re - stated as @xmath37 here , ( 7 ) can be expressed in vector - matrix form as @xmath38 where @xmath39 and @xmath7 is a rectangular matrix with dimensions @xmath40 , and whose @xmath41 row is:@xmath42 $ ] .
the _ working hypothesis _ is established in section 3 via inference of the coefficients from the observed data by invoking the mfi principle . here ,
@xmath43 $ ] .
it is assumed that the probability associated with @xmath33 is @xmath44 .
note that @xmath33 is assumed to be a continuous random variable .
alternately , @xmath44 may be defined as the _ empirical distributions _ of the observations @xmath45 [ 54 ] .
the fim is extremized subject to the constraints @xmath46 and the normalization condition @xmath47 note that @xmath48 , where @xmath49 is the number of parameters of the model .
also @xmath50 denotes the expectation evaluated with respect to @xmath51 .
section 3 derives the constraint driven pseudo - inverse condition for normal distributions by invoking the qm virial theorem as @xmath52 where : @xmath53 is the moore - penrose _ pseudo - inverse _ [ 51 ] .
note that as stated in sections 1 and 3 , unlike the maxent model the fim - based framework presented herein also allows for @xmath44 described by hermite - gauss solutions .
such extensions of the present model and the subsequent effects on the pseudo - inverse condition are beyond the scope of this paper , and will be presented elsewhere . the _ prediction phase _ commences once the pertinent parameters @xmath9 are determined from the @xmath35 training data in the _
modeling phase_. these are employed to predict _ new _ series values @xmath54 where @xmath55 is a matrix of dimension @xmath56 .
note that @xmath57 is such that @xmath58 _ new _ time series values may be evaluated after the training data has been reconstructed .
the prediction phase is essentially the implementation of ( 10 ) , for temporal indices @xmath59 , where , @xmath60 is the sum of both the training data and the _ new _ data to be predicted _ after _ completion of the modeling phase . _
it is important to note that the process of inference necessitates the re - definition of the _ working hypothesis _ to account for ( 10 ) now superseding ( 9 ) .
the obvious reason being that the process of inference can only evaluate @xmath9 and not @xmath33_. the value of @xmath57 should be suitably bounded to facilitate the comparison between the predicted signal obtained from the solution of ( 13 ) , with the original signal .
this is done in order to judge the fidelity of the prediction through _ both _ visual inspection and analysis ; viz .
calculation of the mean squared error ( mse , hereafter ) between the original and the predicted signal . in this paper ,
given the original signal represented by the column vector @xmath61 , @xmath62 - \max \left\ { { t , d } \right\}$ ] .
_ note that this non - unique fiduciary bound does not in any way constrain the it - based prediction model , and there is nothing that prevents the value of @xmath57 from exceeding this bound should the situation require it_. to evaluate the mse , defining the exact measurement from the original signal as @xmath10 , and the corresponding results of the predictive model as @xmath63 @xmath64
consider the probability @xmath65 where @xmath66 is a vector parameter . specializing the focus to a class of probability distributions exhibiting translational invariance where @xmath67 , and assuming without any loss of generality , that the elements of the vector * a * are _ a priori iid _ , the fi matrix with vector entries acquires the form of a diagonal matrix .
the fim [ 24 , 55 ] now takes the form @xmath68 = \int { \frac{1}{{f\left ( \textbf{a } \right ) } } } \left ( { \frac{{d f\left ( \textbf{a } \right)}}{{d \textbf{a } } } } \right)^2d\textbf{a}=\sum\limits_i { \int { f\left ( \textbf{a } \right ) } \left ( { \frac{{\partial \ln f\left ( \textbf{a } \right)}}{{\partial a_i } } } \right)^2 d\textbf{a}}= \sum\limits_i { \int { \frac{1}{{f_i \left ( { a_i } \right)}}\left ( { \frac{{\partial f_i } } { { \partial a_i } } } \right)^2 da_i } } = \sum\limits_i { i_i } .
\end{array}\ ] ] note that in ( 16 ) , for iid entries of the vector * a * , the fim is the trace of the fi matrix which is identical to the scalar case [ 24 ] , and @xmath69 is the @xmath70 diagonal element of the diagonal fi matrix . the derivation of ( 16 ) is described in the appendix . with the aid of real valued amplitudes defined by @xmath71 the fim ( 16 ) may be compactly expressed as @xmath72 = 4\int { \left ( { \frac{{d \psi \left ( \textbf{a } \right)}}{{d \textbf{a } } } } \right)^2 d\textbf{a}}\ ] ] which is extremized subject to the constraint defined by eq .
( 10 ) and the normalization condition eq . ( 11 ) , in section 2 .
a lagrangian can be specified of the form @xmath73 = \int { \left\ { { 4\left ( { \frac{{d\psi \left ( \textbf{a } \right)}}{{d\textbf{a } } } } \right)^2 - \mathop \lambda \limits^ \to \textbf{w}\textbf{a}\psi ^2 \left ( \textbf{a } \right ) - \lambda _ 0 \psi ^2 \left ( \textbf{a } \right ) } \right\}d\textbf{a}},\ ] ] where @xmath74 is the vector of lagrange multipliers associated with the constraint ( 10 ) .
( 19 ) is re - expressed with its constraint terms described in component - wise form as @xmath73 = \int { \left\ { { 4\left ( { \frac{{d\psi \left ( \textbf{a } \right)}}{{d\textbf{a } } } } \right)^2 - \sum\limits_{k = 1}^m { \lambda _ k \sum\limits_{i = 1}^{n_c } { w_{ki } } } a_i \psi ^2 \left ( \textbf{a } \right ) - \lambda _ 0 \psi ^2 \left ( \textbf{a } \right ) } \right\}d\textbf{a}}\ ] ] variational extremization of ( 19 ) with respect to @xmath75 , and multiplying the resultant by 2 yields @xmath76 where @xmath77 is the empirical pseudo - potential . here ,
( 21 ) bears a resemblance to the swe in a vector setting with @xmath78 . redefining ( 18 ) in terms of the pdf @xmath79
one finds after invoking the qm virial theorem [ 31 ] @xmath80 eqs .
( 22 ) yields @xmath81d\textbf{a } = 0.\ ] ] substituting the expression for @xmath82 in ( 21 ) into ( 23 ) results in @xmath83d\textbf{a } = 0.\ ] ] solving ( 24 ) yields @xmath84.\ ] ] setting @xmath85 results in the pdf @xmath86 = \frac{{\exp \left [ { - \frac{{\textbf{w}\textbf{a}^t \textbf{a}}}{{2\sigma ^2 } } } \right]}}{{\tilde z } } = \frac{{\exp \left [ { - \frac{{\sum\limits_{k = 1}^m { w_{ki } \sum\limits_{i = 1}^{n_c } { a_i^2 } } } } { { 2\sigma ^2 } } } \right]}}{{\tilde z } } ; \\ \\ \tilde z = \int { \exp \left [ { - \frac{{\textbf{w}\textbf{a}^t \textbf{a}}}{{2\sigma ^2 } } } \right]d\textbf{a } } .
\end{array}\ ] ] note that @xmath87 denotes the _ statistical dispersion _ , and @xmath88 is the canonical partition function . the above analysis is presented in a more familiar form by invoking the translational invariance property of the fim by specifying @xmath89 here , ( 28 ) has the effect of transforming ( 21 ) to @xmath90 where the translated empirical pseudo - potential is defined by @xmath91 and the translated normalization lagrange multiplier is now @xmath92 it is noteworthy to mention that ( 28 ) is identical to the so - called _ zero - mean _ form of the swe employed in many works on fim - based inference [ 56 ] .
( 23 ) is now re - cast as @xmath93d\textbf{r } = 0.\ ] ] note that @xmath94 where , @xmath95 denotes the expectation evaluated with respect to @xmath96 . substituting ( 30 ) into ( 32 ) and solving yields @xmath97.\ ] ] setting @xmath98 results in the pdf @xmath99\\\ ] ] thus @xmath100}}{{\tilde z } } = \frac{{\exp \left [ { - \frac{{\sum\limits_{k = 1}^m { w_{ki } \sum\limits_{i = 1}^{n_c } { \left ( { a_i - \left\langle { a_i } \right\rangle } \right)^t \left ( { a_i - \left\langle { a_i } \right\rangle } \right ) } } } } { { 2\sigma ^2 } } } \right]}}{{\tilde z}};\\ \\ \tilde z = \int { \exp \left [ { - \frac{{\textbf{w}\left ( { \textbf{a } - \left\langle \textbf{a } \right\rangle } \right)^t \left ( { \textbf{a } - \left\langle \textbf{a } \right\rangle } \right)}}{{2\sigma ^2 } } } \right]d\textbf{a } } . \end{array}\ ] ] solving ( 37 ) yields @xmath101.\ ] ] from ( 28 ) @xmath102 with the aid of ( 28 ) , ( 32 ) , and ( 36 ) , ( 37 ) yields the matrix fim in the form [ 45 ] @xmath103 = \frac{{\textbf{w}^t \textbf{w}}}{{\sigma ^2 } } \ ] ] for normal distributions , the cramer - rao bound is always saturated [ 24 , 45 ] .
thus , the diagonal covariance matrix is of the form @xmath104 to formally establish the pseudo - inverse relation , ( 10 ) , ( 17 ) , and ( 35 ) yield @xmath105 thus @xmath106 the pseudo - inverse condition ( 43 ) may be readily shown to be an _ efficient estimator _ of * a*.
the basic mathematical apparatus and theoretical considerations for deriving the reciprocity relations and the lts have been established [ 28 , 43 ] .
thus , only the pertinent fim - relations in a vector setting for least square constraints are to be stated . multiplying ( 21 ) by @xmath107 and integrating yields in vector form after re - arranging the terms
@xmath72=\lambda_0 + \mathop \lambda \limits^ \to \textbf{w}<\textbf{a}>\ ] ] to treat the component - wise case , the following definition is invoked @xmath108 yielding @xmath72 = \lambda _ 0 + \sum\limits_{k = 1}^m { \lambda _ k \sum\limits_{i = 1}^{n_c } { w_{ki } \left\langle { a_i } \right\rangle } } .\ ] ] taking derivatives of ( 46 ) with respect to @xmath109 results in @xmath110}}{{\partial \lambda _ k } } = \frac{{\partial \lambda _ 0 } } { { \partial \lambda _ k } } + \sum\limits_{i = 1}^{n_c } { w_{ki } \left\langle { a_i } \right\rangle } + \sum\limits_{\scriptstyle j = 1 \hfill \atop \scriptstyle j \ne k \hfill}^m { \lambda _ j \frac{{\partial \sum\limits_{i = 1}^{n_c } { w_{ji } \left\langle { a_i } \right\rangle } } } { { \partial \lambda _ k } } } .\ ] ] specifying @xmath111 in ( 47 ) , yields the generalized fisher - euler theorem in a vector setting for least squares constraints @xmath110}}{{\partial \lambda _ k } } = \sum\limits_{\scriptstyle j = 1 \hfill \atop \scriptstyle j \ne k \hfill}^m { \lambda _ j \frac{{\partial \sum\limits_{i = 1}^{n_c } { w_{ji } \left\langle { a_i } \right\rangle } } } { { \partial \lambda _ k } } } .\ ] ] setting @xmath112 with the aid of ( 45 ) , ( 50 ) takes the form @xmath113 with the aid of ( 21 ) , ( 44),(45 ) and ( 50 ) , the following relation is obtained @xmath114 = - \sum\limits_{k = 1}^m { \lambda _ k \left\langle { \theta _ k \left ( { a_i } \right ) } \right\rangle } \\ \end{array}\ ] ] taking derivatives of the third term in ( 52 ) yields @xmath115}}{{\partial \lambda _ j } } = - \left\langle { \theta _ j \left ( { a_i } \right ) } \right\rangle - \sum\limits_{k = 1}^m { \lambda _ k \frac{{\partial \left\langle { \theta _ k \left ( { a_i } \right ) } \right\rangle } } { { \partial \lambda _ j } } } \ ] ] consider the relation that underlies the basis for the lts [ 26 , 43 ] @xmath116 taking derivatives of ( 54 ) with respect to @xmath117 and comparing the ensuing results with ( 53 ) yield the reciprocity relation @xmath118 likewise , taking derivatives of ( 53 ) with respect to @xmath109 yields the reciprocity relation ( 48 ) . substituting ( 55 ) into ( 52 ) yields a linear pde to infer the fim without having to solve the vector independent schrdinger - like equation @xmath119 = - \sum\limits_{k = 1}^m { \left\langle { \theta _ k \left ( { a_i } \right ) } \right\rangle } \frac{{\partial i\left ( { \left\langle { \theta _ 1 \left ( { a_i } \right ) } \right\rangle , ... , \left\langle { \theta _ m \left ( { a_i } \right ) } \right\rangle } \right)}}{{\partial \left\langle { \theta _ j \left ( { a_i } \right ) } \right\rangle } } .\ ] ] specifying @xmath119 = \sum\limits_{k = 1}^m { i_k \left [ \psi \right ] } = \sum\limits_{k = 1}^m { \exp \left [ { g\left ( { \left\langle { \theta _ k \left ( { a_i } \right ) } \right\rangle } \right ) } \right]},\ ] ] and substituting ( 57 ) into ( 56 ) yields @xmath120 where @xmath121 is a constant of integration . invoking ( 10 ) in ( 58 ) yields a candidate _ empirical fim _ for the modeling phase , defined solely in terms of the training data @xmath122 the utility and practical implementation is the task of ongoing work
. some of the potential implications of ( 59 ) are briefly discussed in section 6 . taking the derivative of ( 58 ) with respect to @xmath123 yields @xmath124 here , ( 60 )
describes a monotonically decreasing @xmath125 in the @xmath123-direction . differentiating ( 60 ) yields @xmath126 where @xmath127 is the krnecker delta . here , ( 61 ) establishes the convexity of the fim derived in this section , thereby guaranteeing the existence of its inverse .
this section demonstrates the efficacy of the prediction model with the aid of the m - g dde and two ecg signals .
the embedding dimension for all examples is evaluated using the false nearest neighbor method [ 57 ] all numerical examples are evaluated for values of the forecasting time t= 1 and 5 . the largest positive lyapunov exponent ( lle , hereafter )
is one of the simplest indicators of chaotic behavior . from refs .
[ 17 , 58 , 59 ] , it is evident that m - g dde , with the delay time @xmath128 secs . , possess a positive lle . likewise , from [ 60,61 ] , it has been established that the ecg signals comprising record 207 of the mit - bih arrhythmia database possess a positive lle , while [ 60 ] demonstrates that the signal comprising cudb / cu02 also has a positive lle .
the famed m - g dde is described by @xmath129 where , @xmath130 secs . here , @xmath131 .
integration with a fourth - order runge - kutta routine yields the original solutions .
the value of the embedding dimension is @xmath132 .
the total number of points in the original solution is @xmath133 .
the procedures described in sections 2 and 3 are then employed for @xmath134 $ ] secs . to obtain @xmath9 , for @xmath135 .
these results and the concomitant mse values are depicted in figs . 1 and 2 , respectively . here ,
1 clearly demonstrates that the predicted results faithfully capture the dynamics embedded in the chaotic m - g time series. fig . 2 expectedly demonstrates a slight distortion of the predicted signal vis - - vis the original signal , as a consequence of long - term forecasting .
it may be argued that the number of coefficients @xmath136 is high and can forecast just about any signal .
this argument is not only tenuous at best for the case of chaotic signals , but is also orthogonal to the very reason causing the choosing of such a high value of @xmath137 .
specifically , section 1 explicitly discusses the possible singularity ( or near - singularity ) of the embedding matrix * w*. as stated therein , large number of lags , * w * can result in volatile forecasts owing to ill - conditioning .
the m - g dde in this example has a higher embedding dimension than other prominent models ( such as the lorenz , hnon , etc . )
( see , for eg .
[ 3 ] ) , and thus the resulting * w * would be more prone to result in volatile forecasts . as is evidenced by figs . 1 and 2
, this is not the case and the forecasts are clearly accurate and stable .
it is noteworthy to mention that the coefficients obtained from the training data during the modeling phase , which form the basis on which further prediction is done over a much larger time period and data sample size ( as compared to those in the modeling phase ) , are unique to the specific data set under consideration .
specifically , coefficients obtained from different data sets , for example @xmath0 the m - g dde with a different value of the lag @xmath138 or @xmath1 another nonlinear dynamics model , yield erroneous predictions if applied to a data set which differs from the one(s ) they were obtained from .
this issue is the task on ongoing studies briefly described in section 6 within the context of the results described in sections 3 and 4 , and will be presented elsewhere .
this sub - section employs a 300 secs .
ecg signal to demonstrate that the model described in this paper accurately predicts episodes ( transients ) which are the artifacts of a diseased heart over a reasonable period of time , even for a highly erractic / volatile signal .
the annotations are described in [ 62 ] .
the signal is extracted from data obtained as a .mat data file from [ 63 ] .
the sampling frequency of the data @xmath139 [ 64 ] , the number of samples being @xmath140 for a total duration of @xmath141 .
the rationale for the choice of 300 secs .
sample is to ensure that the portion of the signal , both during the modeling phase and the prediction phase that follows , possess sufficiently identifiable episodes which are documented in the reference annotations [ 65 ]
. it would be desirable to conduct the study over the entire duration of the signal spanning around 30 mins . however , this would yield simulation results that are visually incoherent , and hence the truncation of the signal length / duration .
the number of training samples from which the values of @xmath9 is obtained is @xmath142 @xmath143 $ ] secs . for forecasting times of @xmath144 ( @xmath145 and @xmath146 ( @xmath147 . in all cases , @xmath148 , @xmath149 , and @xmath150 .
all simulation results are presented for the mlii lead .
the simulation results depicting the predicted ecg signal superimposed over the original record 207 signal for the mlii lead for t=1.0 and 5.0 respectively and the concomitant mse s are presented in figs . 3 and 5 , respectively .
both figs .
( 3 ) and ( 5 ) demonstrate that even for a relatively small duration of the modeling phase which comprises 1/6-th of the duration of the entire prediction exercise comprising of both the modeling phase and the prediction of new data , the prediction results are of a high quality .
the case corresponding to @xmath146 shows greater `` undershoots '' and `` overshoots '' of the peaks of the signal as compared to the case of @xmath144 .
this degradation of prediction performance is expected .
_ the results of the mse s for both cases of the forecasting time shows divergent peaks .
these have been analyzed and found to be the result of the highly erractic / volatile nature of the signal .
it is important to note that they do not constitute any volatility in the prediction since the profiles of the predicted signals are demonstrated to be very much in accord with the original signal .
further , in both cases , following every divergent peak which can even be visually related to erratic signal quality in figs .
( 3 ) and ( 5 ) , the prediction returns to `` normalcy '' which is defined by a low mse value .
this would not be the case of a volatile prediction cased by ill - conditioning or any other factor ( see section 1 ) , because the errors cased by volatile predictions tend to cascade_. figs .
( 4 ) and ( 6 ) focus on specific regions of interest .
. 4(a ) and 6(a ) depict the overlaid overlaid and predicted signals of the modeling phase for the cases of the forecasting time @xmath144 and @xmath146 , respectively . to establish the accuracy of the fim based model , the prediction of the episodes corresponding to the various conditions of the diseased heart documented in the reference annotations [ 65 ] , that can be visually determined in the period @xmath151 $ ] secs .
are demonstrated in figs .
4(b)-(d ) and 6(b)-(d ) for @xmath144 and @xmath146 , respectively .
on inspection of figs .
4(b ) and 6(b ) , @xmath0 the instance of ventricular tachycardia identified by `` + '' and defined by `` ( vt '' in the reference annotations at 38.522 secs . ,
immediately followed by three instances of premature ventricular contraction identified by `` v '' and @xmath1 the onset of ventricular flutter / fibrillation identified by `` [ '' at 40.736 secs . in the reference annotation followed by an instance of ventricular flutter identified by `` + '' and defined by `` ( vfl '' in the reference annotations at 40.803 secs . and the subsequent termination of the ventricular flutter / fibrillation identified by `` ] '' at 50.972 secs . in the reference annotation can be easily identified .
this region is of particular importance since it spans _ both _ the modeling phase from which the _ working hypothesis _
is determined from the training data , _ and _ the prediction of new data values . on inspection of figs .
4(c ) and 6(c ) , @xmath0 the onset of ventricular flutter / fibrillation identified by `` [ '' at 54.764 secs . in the reference annotation followed by an instance of ventricular flutter identified by `` + '' and defined by `` ( vfl '' in the reference annotations at 54.869 secs . and the subsequent termination of the ventricular flutter / fibrillation identified by `` ] '' at 50.972 secs . in the reference annotation and
@xmath1 the instance of ventricular tachycardia identified by `` + '' and defined by `` ( vt '' in the reference annotations at 61.839 secs .
( 1:01.839 mins . ) , immediately followed by three instances of premature ventricular contraction identified by `` v '' , can be easily identified . finally , on inspection of figs .
4(d ) and 6(d ) , the onset of ventricular flutter / fibrillation identified by `` [ '' at 269.467 secs .
( 4:29.467 mins . ) in the reference annotation followed by an instance of ventricular flutter identified by `` + '' and defined by `` ( vfl '' in the reference annotations at 129.586 secs .
( 4:29.586 mins . ) and the subsequent termination of the ventricular flutter / fibrillation identified by `` ] '' at 240.906 secs .
( 4:40.906 mins . ) in the reference annotation , can be easily identified . in all cases depicted in figs .
( 4 ) and ( 6 ) , it is observed that the quality of the prediction is high , with the case of the example with @xmath146 being marginally degraded vis - - vis the case with @xmath144 , which is expected .
this sub - section demonstrates the robustness of model described in this paper for an extended ecg signal , even for a highly erractic / volatile signal displaying the symptoms of cardiac vta .
the signal is extracted from data obtained as a .mat data file from [ 66 ] .
the sampling frequency of the data @xmath152 , and the number of samples is @xmath153 , for a total duration of @xmath154 [ 67 ] . in order to study the robustness of the prediction performance of the fim - based model , the number of training samples from which the values of @xmath9 is obtained is chosen to be @xmath155 @xmath156 $ ] secs . for @xmath144 ( @xmath157 and @xmath146 ( @xmath158 . in all cases , @xmath148 , @xmath149 , and @xmath150 .
the high quality of the model is attested by the fidelity of the predicted ecg profiles with the original signals , and , the mse s .
the simulation results depicting the predicted ecg signal superimposed over the original signal for t=1.0 and 5.0 and the concomitant mse s , are presented in figs . 7 and 8 , respectively .
similar to the case described in section 5.2 , the results of the mse s for both forecasting times depict divergent peaks .
these have been analyzed and found to be the result of the nature of the signal .
again , note that these divergent peaks in the mse s do not constitute any volatility in the prediction since the profiles of the predicted signals are demonstrated to be very much in accord with the original signal .
further , in both cases , following every divergent peak which can even be visually related to erratic signal quality in figs .
( 7 ) and ( 8) , the prediction returns to `` normalcy '' ( defined by a low mse value ) .
this would not be the case in a volatile prediction cased by ill - conditioning or any other factor ( see section 1 ) , because the errors cased by volatility of the forecasting tend to cascade .
a convenient framework for the modeling and forecasting of time series has been developed within the ambit of a fim - based inference model . the modeling phase from which the _ working hypothesis _ is derived has been provided with a qm connotation , by the formulation of a time independent schrdinger - like equation in a vector setting , employing least squares constraints .
this has been achieved by invoking the mfi principle .
apart from the obvious theoretical implications , this allows for the systematic derivation and categorization of the _ working hypothesis _ and the subsequential forecasting phase .
the pdf and the pseudo - inverse relations have been self - consistently inferred by invoking the qm virial theorem for the case of normal distributions .
the reciprocity relations and the lts for the modeling phase have been derived for the modeling phase .
this results in an intriguing form of the fim for the modeling phase , which defines the _ working hypothesis _ , described solely in terms of the observed data ( the _ empirical fim _ ) .
the possible utilities of this form of the fim are to derive principled expressions and values for the statistical dispersion .
this fim expression has no equivalent in prior maxent models [ 7 - 10 ] , which treat the statistical dispersion as an ad - hoc scaling [ 15 ] .
the fim - based model has been numerically tested , and its efficacy proven for the mackey - glass dde , the ecg signal for the mlii lead from record 207 of the mit - bih cardiac arrhythmia database , and the ecg from record cudb / cuo2 of the creighton university vta database . the forecasting is consistently demonstrated to be of very high quality , and does not suffer from any signs of volatility caused by ill - conditioning of the embedding matrix * w * or any other factor .
the numerical experiments on the ecg demonstrate that the model presented in this paper is able to forecast salient episodes documented in the reference annotations for the said signals , to a very high degree of accuracy .
ongoing work is focused on a two - pronged approach .
first , the _ empirical fim _ derived in section 4 has been investigated within the context of its relationship to the fim employed in section 3 , and the subsequent effects on the forecasting .
a fiduciary time dependence is induced into the modeling phase via a sliding window analysis .
this allows for the quality of prediction to be related to fundamental results governing the fim , viz .
i - theorem _ ( the fisher - equivalent of the h - theorem ) [ 24 ] , for both normal and non - equilibrium distributions .
next , a principled comparison of the fim - based model with existing nonparametric prediction models [ 12 , 13 ] is in progress .
these works will be presented elsewhere .
for the fisher information matrix @xmath159 $ ] , each element is defined by @xmath160da.\ ] ] for * a * possessing _ a - priori _ iid entries @xmath161 substituting ( a.2 ) into ( a.1 ) yields @xmath162da_k\ ] ] for @xmath163 @xmath164 since all other integrals @xmath165 integrate to unity because of normalization . for @xmath166 , @xmath167 thus , @xmath159 $ ] is a diagonal matrix with each element defined by ( a.5 ) .
note that @xmath168 , as used in eq .
( 16 ) .
eckman , d. ruelle , rev .
phys . * 15 * ( 1985 ) 617 - 656 .
h. d. i. abarbanel , r. brown , j. j. sidorowich , l. sh .
ysimring , rev .
* 65 * ( 1993 )
1331 - 1392 . h. kantz , t. schreiber , _ nonlinear time series analysis _ , cambridge univ .
press , cambridge u.k , 1999 .
j. rissanen , ann . stat . * 14 * ( 1989 ) 1080 - 1100 .
f. takens , `` detecting strange attractors in turbulence '' , in _ dynamical systems and turbulence _ : lecture notes in mathematics , volume 898 , springer , berlin , 1981 , 366 - 381 .
t. sauer , j. a. yorke , m. casdagli , j. stat .
* 65 * ( 2000 ) 579 .
l. diambra , a. plastino , phys .
lett . a * 216 * ( 1996 ) 278 - 282 .
m. t. martn , a. plastino , v. vampa , g. judge , physica a * 405 * ( 2014 ) 63 - 69 .
m. t. martn , a. plastino , v. vampa , entropy * 16 * ( 2014 ) 4603 - 4611 .
a. f. bariviera , m. t. martn , a. plastino , v. vampa , physica a * 449 * ( 2016 ) 401 - 407 .
e. t. jaynes , phys . rev .
* 106 * ( 1957 ) 620 - 630 .
j. c. principe , _ information theoretic learning - renyi s entropy and kernel perspectives _ , springer , new york , 2010 .
w. liu , j. c. principe , s. haykin , _ kernel adaptive filtering : a comprehensive introduction _ , wiley , hoboken nj , 2010 .
h. d. i. abarbanel , _ analysis of observed chaotic data _ , springer , new york , 1996 .
l. diambra , physica a * 278 * 2000 ) 140 - 149 . n. wiener .
_ extrapolation , interpolation , and smoothening of stationary time series with engineering applications _ , wiley , new york , 1949 .
m. casdagli , physica d * 35 * ( 1989 ) 335 - 356 .
j. c. principe , a. rathie , j .-
kuo , intl .
j. of bifurcation and chaos * 2 * 1992 989 - 996 .
d. mandic , j. chambers , _ recurrent neural networks for prediction _ , wiley , chichester , 2001 .
l. diambra , a. plastino , phys .
e * 52 * ( 1995 4557 - 4560
. n. i. sapankevych , r. sankar , ieee comput .
intell . mag .
, * 4 * ( 2009 ) 24 - 38 .
j. p. crutchfield , b. s. mcnamara , complex systems , * 21 * ( 1985 ) 417 .
j. d. farmer , j. j. sidorowich , phys .
* 59 * ( 1987 ) 845 .
b. r. frieden , _ science from fisher information - a unification _ , cambridge university press cambridge , 2004 . j.a .
wheeler , in zurek w. h. ( ed . ) : _ complexity , entropy and the physics of information _ , addison wesley , new york , 3 - 28 , 1991 .
j. hber , _ robust statistics _ , wiley , newy york , 1981 .
b. r.frieden , phys .
a , * 41 * ( 2000 ) 4265 - 4276 ; optics lett , * 14 * ( 1989 ) 199 - 201 . b. r. frieden , a. plastino , a. r. plastino , b. h. soffer , phys . rev .
e * 66 * ( 2002 ) 046128 .
frieden , a. plastino , a. r. plastino , b. h. soffer , phys .
lett . a * 304 * ( 2002 ) 73 - 78 .
j. s. dehesa , .
g. martn , p. snchez - moreno , complex anal . and
oper . th .
* 6 * ( 2012 ) 585 - 601 .
w. greiner , _ quantum mechanics .
an introduction _ , springer , berlin , 2012 . f. n. fernandez and e. a. castro , _ hypervirial theorems _ ,
lecture notes in chemistry , vol .
43 , springer - verlag , berlin , 1987 .
r. c. venkatesan , , , 487 , ieee press , piscataway , nj , 2007 .
r. c. venkatesan , `` encryption of covert information through a fisher game '' , in _ exploratory data analysis using fisher information _ , frieden , b.r . and gatenby , r.a . ,
( eds . ) , springer- verlag , london , 181 - 216 , 2006 .
b. r.frieden , a. plastino , a. r. plastino , b. h. soffer , phys .
e , * 60 * ( 1999 ) 046128 .
s. p. flego , a. plastino , a. r. plastino , ann .
* 326 * ( 2011 ) 2533 - 2543 .
s. p. flego , a. plastino , a. r. plastino , physica a * 390 * ( 2011 ) 2276 - 2282 .
s. p. flego , a. plastino , a. r. plastino , physica a * 390 * ( 2011 ) 4702 - 4712 .
s. p. flego , a. plastino , a. r. plastino , physica scripta , * 85 * ( 2012 ) 055002 - 055008 .
c. villani , _ topics in optimal transportation _ , graduate studies in mathematics vol.*58 * , american mathematical society , 2000
. g. blower , _ random matrices : high dimensional phenomena _ , london mathematical society lecture notes , cambridge university press , cambridge , 2009 .
zegers , entropy * 17 * ( 2015 ) 4918 - 4939 ; b. r. frieden , b. h. soffer , phys . lett .
a * 374 * ( 2010 ) 3895 - 3898 .
r. c. venkatesan , a. plastino , phys .
a , * 378 * ( 2014 ) 1341 - 1345 .
r. c. venkatesan , a. plastino , ann .
* 359 * ( 2015 ) 300 - 316
. s. m. kay , _ fundamentals of statistical signal processing , vol i : estimation theory _
, prentice - hall signal processing series , 1993 . m. casas , f. pennini , a. plastino , phys . lett . a * 235 * ( 1997 ) 457 - 463 .
m.c.mackey , l.glass , science * 197 * ( 1977 ) 287 - 289 . g. b. moody , r. g. mark , ieee eng . in med and biol . *
20 * ( 2001 ) 45 - 50 .
f. m. nolle , f. k. badura , j. m. catlett , r. w. bowser , m. h. sketch , `` crei - gard , a new concept in computerized arrhythmia monitoring systems '' , _ computers in cardiology 1986 _ * 13 * , 515 - 518 , ieee press , piscataway , nj , 1986 .
a. l. goldberger , l. a. n. amaral , l. glass , j. m. hausdorff , p. ch .
ivanov , r. g. mark , j. e. mietus , g. b. moody , c .- k .
peng , h. e. stanley , circulation , * 01*(2000 e215-e220 .
g. h. golub , c. l. van loan , _ matrix computations _
, third ed . , johns hopkins univ . press , baltimore , 1995 .
l. diambra , c. p. malta , phys .
e * 57 * ( 1999 ) 929 - 937 .
b. pompe , `` mutual information and relevant variables for predictions '' , in _ modeling and forecasting financial data techniques for nonlinear dynamics _ , soofi , a. s. and cao , l. , ( eds . ) , springer , new york , pp .
61 - 92 , 2002 .
a. b. owen , _ empirical likelihood _ ,
chapman & hall / crc , boca raton , 2001 .
d. guo , s. shamai ( shitz ) and s. verd , _ the interplay between information and estimation measures _ , foundations and trends in signal processing ser .
, now publishers , boston , 2012 . m. casas , a. plastino , a. puente , phys .
a * 248 * ( 1998 ) 161 - 166 .
m. kennel , r. brown , h. d. i. abarbanel , phys .
rev . a;*45 * ( 1992 ) 3403 - 3411 .
p. i. grassberger , i. procaccia , physica d , * 9 * ( 1983 ) 189 - 208 . j. d. farmer , physica d , * 4 * ( 1982)366 - 393 .
r. b. govindan , k. narayanan , m. s. gopinathan , chaos , * 8 * ( 1998 ) 495 - 502 .
a. casaleggio , s. braiotta,``study of the lyapunov exponents of ecg signals from mit - bih database '' , _ computers in cardiology 1995 _ , 697 - 700 .
ieee press , piscataway , nj , 1995 ; a. casaleggio , s.braiotta , chaos sol .
* 9 * ( 1997 ) 1591 - 1599 . | a robust prediction model invoking the takens embedding theorem , whose _ working hypothesis _ is obtained via an inference procedure based on the minimum fisher information principle , is presented .
the coefficients of the ansatz , central to the _ working hypothesis _ satisfy a time independent schrdinger - like equation in a vector setting .
the inference of i ) the probability density function of the coefficients of the _ working hypothesis _ and ii ) the establishing of constraint driven pseudo - inverse condition for the modeling phase of the prediction scheme , is made , for the case of normal distributions , with the aid of the quantum mechanical virial theorem .
the well - known reciprocity relations and the associated legendre transform structure for the fisher information measure ( fim , hereafter)-based model in a vector setting ( with least square constraints ) are self - consistently derived .
these relations are demonstrated to yield an intriguing form of the fim for the modeling phase , which defines the _ working hypothesis _ , solely in terms of the observed data .
cases for prediction employing time series obtained from the : @xmath0 the mackey - glass delay - differential equation , @xmath1 one ecg sample from the mit - beth israel deaconess hospital ( mit - bih ) cardiac arrhythmia database , and @xmath2 one ecg from the creighton university ventricular tachyarrhythmia database .
the ecg samples were obtained from the physionet online repository .
these examples demonstrate the efficiency of the prediction model .
numerical examples for exemplary cases are provided . fisher information , time series prediction , working hypothesis inference , minimum fisher information , takens theorem , generalized vector fisher - euler theorem , legendre transform structure , mackey - glass equation , ecg s .
pacs : 05.20.-y ; 2.50.tt ; 0.3.65.-w ; 05.45.tp |
let @xmath1 be an @xmath2-dimensional sequence in the unit cube @xmath3 , @xmath4 , @xmath5 where @xmath6 , if @xmath7 and @xmath8 , if @xmath9 .
we define the star _ discrepancy _ of a @xmath10-point set @xmath11 as @xmath12 let @xmath1 be an arbitrary sequence in @xmath3 . according to the well - known conjecture ( see e.g. [ 1 , p.67 ] , [ 3 , p.32 ] ) @xmath13 * definition 1*. _ let @xmath14 , and @xmath15 be integers and let @xmath16 . a @xmath17-net in base @xmath18 is a point set @xmath19 of @xmath20 points in @xmath21 such that every subinterval @xmath22 of volume @xmath23 which has the form @xmath24 with integers @xmath25 and @xmath26 for @xmath27 , contains exactly @xmath28 points of @xmath19 . _ + if @xmath29 , we obtain a classical @xmath30-net . for @xmath31 ,
where @xmath32 and @xmath33 , we define the truncation @xmath34_m = \sum_{1 \leq j \leq m } x_{j}b^{-j}$ ] .
if @xmath35 , then the truncation @xmath36_m$ ] is defined coordinatewise , that is , @xmath36_m = ( [ x^{(1)}]_m , . . .
, [ x^{(s)}]_m)$ ] . + * definition 2*. _ let @xmath14 , and @xmath15 be integers and let @xmath16 . a sequence @xmath37 of points in @xmath21 is a @xmath38-sequence in base @xmath18 if for all integers @xmath39 and @xmath40 the points @xmath41_{m}$ ] with @xmath42 form a @xmath17-net in base @xmath18 . _
+ if @xmath43 , we obtain a classical @xmath44-sequence . for @xmath45 , and @xmath46
where @xmath47 , we define the ( @xmath18-adic ) digitally shifted point @xmath48 by @xmath49 , where @xmath50 and @xmath51 . for higher dimensions
@xmath52 let @xmath53 .
for @xmath54 we define the ( @xmath18-adic ) digital shifted point @xmath55 by @xmath56 . for @xmath57 , we define @xmath58 . + for @xmath45 , where @xmath59 , @xmath60 @xmath61 and @xmath62 we define the absolute valuation @xmath63 of @xmath64 by @xmath65 .
let @xmath66 for @xmath67 . + * definition 3 .
* _ a digital point set @xmath68 in @xmath3 is @xmath69admissible in base @xmath18 if _
@xmath70 _ a sequence @xmath71 in @xmath3 is @xmath69admissible in base @xmath18 if _
+ by [ 3 , p. 60 ] @xmath73 for every @xmath0-sequence @xmath74 . in this paper
we prove that this estimate is exact for digitally shifted @xmath69admissible @xmath0 sequences and in particulary for digitally shifted niederreiter s sequence ( see e.g. [ 1]-[4 ] ) .
this result supports the conjecture ( [ 2 ] ) . in [ 2 ] we prove that @xmath0 sequences from [ 1 , section 8 ] are @xmath69admissibles . + * theorem 1 . *
_ let @xmath75 , @xmath76 , @xmath77_m \;| \;y
\in [ 0,1)\}$ ] , @xmath78 be a @xmath69admissible @xmath79 net in base @xmath18 , @xmath80 . then _
@xmath81 + * theorem 2 . *
_ let @xmath82 , @xmath76 , @xmath83 be an @xmath69admissible @xmath0 sequence in base @xmath18 .
then _ @xmath84 + * theorem 3 . *
_ let @xmath82 , @xmath83 be a generalized niederreiter sequence with generating polynomials @xmath85 , ( see [ 1 , p.266 ] , [ 4 , p. 242 ] ) , @xmath86 @xmath87 , @xmath88 , @xmath89 , @xmath90 .
then ( [ 10 ] ) hold . _
* lemma 1 . * _ let @xmath91 , @xmath76 , @xmath92 be an @xmath69admissible @xmath93 net in base @xmath18 , @xmath94 , @xmath95 , @xmath96 , @xmath97 $ ] , @xmath98 , @xmath99 @xmath100 , @xmath101 , @xmath102 , @xmath103 @xmath104 , @xmath105 and let @xmath106 , @xmath107 and @xmath108 , @xmath109 , @xmath110 .
let there exists @xmath111 such that @xmath112_{\dot{m}_i } = \gamma^{(i)}$ ] , @xmath113 , and @xmath114 . then _ + @xmath115 + * proof .
* let @xmath116 , @xmath117 , @xmath118 , @xmath119:\ ; [ ( r_i-\ddot{m}_i-1)/(d_0\hat{e } ) ] \in b_i \}$ ] , @xmath120 , @xmath121 , @xmath122 and @xmath123 .
we have @xmath124 .
let @xmath125_{r_i-1 } + g_ib^{-r_i } , [ \gamma^{(i)}]_{r_i-1 } + ( g_i+1)b^{-r_i } \big ) . \ ] ] similarly to [ 3 , p. 37,38
] , from ( [ 1 ] ) we have that @xmath126 , where + @xmath127 consider @xmath128 .
bearing in mind that @xmath129 is a @xmath93 net , we obtain @xmath130 .
hence @xmath131 .
+ consider @xmath132 .
it is easy to verify that @xmath133 .
+ consider @xmath134 .
we see that @xmath135 .
hence @xmath136 . taking into account that @xmath137 and @xmath138 \notin b_i $ ]
, we get @xmath139 with some @xmath140 , @xmath141 .
hence @xmath142 thus @xmath143 .
we have a contradiction .
hence @xmath144 and @xmath145 .
+ consider @xmath146 .
suppose that @xmath147 for some @xmath148 and @xmath149 .
then @xmath150_{r_i } = [ \gamma^{(i)}]_{r_i } -b^{-r_i } $ ] , @xmath151 .
hence @xmath152 for @xmath153 , @xmath151
. therefore @xmath154 by ( [ 3 ] ) and conditions of the lemma , we have a contradiction .
thus @xmath155 .
we have @xmath156 .
we derive @xmath157 with @xmath158 + let @xmath159 , @xmath160 $ ] and @xmath161 for @xmath162 $ ] . by ( [ in10 ] ) , we get @xmath163 for @xmath162 $ ] .
hence @xmath164 , where @xmath165 , \;i \in [ 1,\dot{s } ] \}.\ ] ] let @xmath166 , for @xmath162 $ ] .
we have @xmath167\;\for\ ; \hat{j}_{i } \in [ \dot{m}- \dddot{m } , \dot{m}-1 ] , \qquad \qquad\qquad \qquad\qquad \qquad\qquad \qquad\qquad \qquad\qquad \qquad \nonumber \\ \for\ ; i \in[1,{\dot{s}}-1],\ ; \with \;\dddot{m}=[\dot{m}/({\dot{s}}-1)].\ ;
{ \rm thus } \;\#a_6 \geq \#a_7,\ ; \where \qquad \qquad\qquad \qquad\qquad \qquad\qquad \qquad \qquad\qquad \qquad \nonumber\end{gathered}\ ] ] @xmath168 , \quad i \in [ 1,\dot{s}],\ ] ] @xmath169,\quad i \in [ 1,\dot{s}-1 ] \qquad \ad \qquad j_{\dot{s}}=(\dot{s } -1)\hat{e}\dot{m } + 1- j_1- ... -j_{{\dot{s}}-1 } \}.\ ] ] we obtain @xmath170 , where @xmath171 , \ ; i \in [ 1,\dot{s}-1 ] \}.\ ] ] therefore @xmath172 @xmath173 for @xmath114 .
therefore @xmath174 .
+ thus lemma 1 is proved .
+ * proof of theorem 1 . * using lemma 1 with @xmath175 , @xmath176 @xmath177 , @xmath178 , @xmath179 , @xmath180 , @xmath181 , and @xmath182_m$ ] , we obtain the assertion of theorem 1 .
* proof of theorem 2 .
* according to [ 3 , lemma 3.7 ] , we have @xmath183 by ( [ 3 ] ) and [ 1 , lemma 4.38 ] , we have that @xmath184 is a @xmath69admissible @xmath185net in base @xmath18 .
using lemma 1 with @xmath186 , @xmath187 , @xmath176 @xmath188 , @xmath178 , @xmath179 , @xmath189 , @xmath190 , and @xmath191_m,{-q / b^m})$ ] , we get the assertion of theorem 2 .
* lemma 2 . *
_ let @xmath192 be a @xmath193 sequence in base @xmath18 .
then @xmath192 is @xmath194admissible . _ + *
* suppose that @xmath192 is not a @xmath195 admissible .
then there exists @xmath196 with @xmath197 @xmath198 .
let @xmath199 , and let @xmath200 @xmath201 .
hence @xmath202 .
let @xmath203e_i \geq d_i -e_i + 1 $ ] , @xmath204_{\dot{d}_i } b^{\dot{d}_i}$ ] @xmath205 and let @xmath206 .
we have @xmath207 for all @xmath208 $ ] , @xmath209 $ ] .
hence @xmath210 .
we derive @xmath211 let @xmath212 where @xmath213 .
it is easy to see that @xmath214 with some @xmath215 .
hence @xmath216 .
thus @xmath217 .
bearing in mind ( [ 70 ] ) , we obtain that @xmath192 is not @xmath218 sequence .
we have a contradiction . hence lemma 2
is proved .
00 dick , j. , pillichshammer , f. , digital nets and sequences , discrepancy theory and quasi - monte carlo integration , cambridge university press , cambridge , 2010 .
levin , m.b .
, on the lower bound of the discrepancy of @xmath0 sequences : ii , http://arxiv.org/abs/1505.04975 niederreiter , h. , random number generation and quasi - monte carlo methods , in : cbms - nsf regional conference series in applied mathematics , vol .
63 , siam , 1992 .
tezuka , s. , on the discrepancy of generalized niederreiter sequences .
journal of complexity , 29 ( 2013 ) , 240 - 247 . | 0.5we find the exact lower bound of the discrepancy of shifted niedereiter s sequences .
+ _ to cite this article : a. nom1 , c. r. acad .
sci .
paris , ser .
i .
_ * rsum * 0.5*sur la limite infrieure de la discrpance de @xmath0 squences : i * + nous trouvons une limite infrieure de la discrpance de squences dcals de niedereiter .
+ _ pour citer cet article : a. nom1 , c. r. acad .
sci .
paris , ser .
i . _ 0.5 , reu le * * * * * ; accept aprs rvision le + + + + + + prsent par |
in population genetics , one way to explain disparity is to observe how many genes appear only once in the sample .
a gene carried by a single individual is the result of two possible events : either the gene comes from a mutation that appeared in an external branch of the genealogical tree , either this gene is of the ancestral type and mutations occured in the rest of the sample ( see figure [ fig : graph1 ] ) . we suppose that events of the second type occur in a much less frequent way than events of the first type ( it is indeed the case when the size of the sample goes big ) .
the total number of genes carried by a single individual is then closely related to the so - called total external length , which is the sum of all external branch lengths of the tree .
the bolthausen - sznitman coalescent ( see for instance @xcite ) is a well - known example of exchangeable coalescents with multiple collisions ( see @xcite for a proper definition of this type of coalescents ) .
it was first introduced in physics , in order to study spin glasses but it has also been thought as a limiting genealogical model for evolving populations with selective killing at each generation , see for instance @xcite .
recently , berestycki et al . in @xcite noted that this coalescent represents the genealogies of the branching brownian motion with absorption .
the bolthausen - sznitman coalescent @xmath1 , is a continuous time markov chain with values in the set of partitions of @xmath2 , starting with an infinite number of blocks / individuals . in order to give a formal description of the bolthausen - sznitman coalescent , it is sufficient to give its jump rates .
let @xmath3 , then the restriction @xmath4 of @xmath1 to @xmath5=\{1,\dots , n\}$ ] is a markov chain with values in @xmath6 , the set of partitions of @xmath5 $ ] , with the following dynamics : whenever @xmath7 is a partition consisting of @xmath8 blocks , any particular @xmath9 of them merge into one block at rate @xmath10 so the next coalescence event occurs at rate @xmath11 note that mergers of several blocks into a single block is possible , but multiple mergers do not occur simultaneously .
moreover , this coalescent process is exchangeable , i.e. its law does not change under the effect of a random permutation of labels of its blocks .
one of our aims is to study the total external length of the bolthausen - sznitman coalescent .
more precisely , we determine the asymptotic behaviour of the total external length @xmath12 of the bolthausen - sznitman coalescent restricted to @xmath6 , when @xmath13 goes to infinity , and relate it to its total length @xmath14 ( the sum of lengths of all external and internal branches ) . a first orientation can be gained from coalescents without proper frequencies .
for this class mhle @xcite proved that after a suitable scaling the asymptotic distributions of @xmath12 and @xmath14 are equal .
now the bolthausen - sznitman coalescent does not belong to this class , but it is ( loosely speaking ) located at the borderline . also it is known for the bolthausen - sznitman coalescent @xcite that @xmath15{d}z,\ ] ] where @xmath16{d}$ ] denotes convergence in distribution and @xmath17 is a strictly stable r.v . with index 1 , i.e. its characteristic exponent satisfies @xmath18=\frac{\pi}{2}|\theta|-i\theta\log|\theta| , \qquad \theta \in \mathbb{r}.\ ] ] in their recent work , dhersin and mhle @xcite showed that the ratio @xmath19 converges to 1 in probability .
thus one might guess that @xmath12 satisfies the same asymptotic relation with the same scaling .
it is a main result of this paper that such a conjecture is almost , but not completely true .
+ let us consider @xmath20 .
we denote by @xmath21 the size of the @xmath9-th jump , i.e the number of blocks that the markov chain loses in @xmath9-th coalescence event .
we also denote by @xmath22 for the number of blocks after @xmath9 coalescence events .
observe that @xmath23 and @xmath24 .
since the merging blocks coalesce into 1 , there are @xmath25 blocks involved in @xmath9-th coalescence event and , for @xmath26 , @xmath27 let @xmath28 be the number of coalescence events .
more precisely @xmath29 according to iksanov and mhle @xcite ( see also @xcite ) , @xmath28 satisfies the following asymptotic behaviour @xmath30{d } z.\ ] ] the main result of this paper describes the behaviour of the total internal length @xmath31 , when @xmath13 goes to @xmath32 . in order to do so
, we introduce the r.v . @xmath33 that represents the number of internal branches after @xmath9 coalescence events . in other words , it is the number of remaining blocks which have already participated in a coalescence event .
note that at time 0 all branches are external i.e. @xmath34 .
let @xmath35 be a sequence of i.i.d .
standard exponential r.v . , also independent from the @xmath22 and @xmath33 , thus from , we have @xmath36 our main result is the following weak law of large numbers for @xmath31 . here @xmath16{\mathbb{p } }
$ ] denotes convergence in probability .
[ th : cvi ] for the total internal length of the bolthausen - sznitman coalescent , we have @xmath37{\mathbb{p } } 1.\ ] ] now noting that @xmath38 and using and our main result , we deduce the asymptotic distribution of the total external length @xmath39 .
[ cor : cve ] for the total external length of the bolthausen - sznitman coalescent , we have @xmath40{d}z-1.\ ] ] observe that the bolthausen - sznitman coalescent can be seen as a special case ( @xmath41 ) of the so - called @xmath42-coalescent which class is defined for @xmath43 .
work @xcite shows that in the case @xmath44 the variable @xmath45 converges in law to a random variable defined in terms of a driftless subordinator depending on @xmath46 . for @xmath47
, we refer to @xcite where it is proven that @xmath48 converges weakly to a stable r.v . of index @xmath46 , @xmath49 being a constant also depending on @xmath46 ( see also @xcite ) . in kingman
s case ( @xmath50 ) a logarithmic correction appears and the limit law is normal ( see @xcite ) .
the remainder of the paper is structured as follows . in section 2 ,
we prove our main results using a coupling method which was introduced in @xcite that provides more information of the chain @xmath51 .
finally , section [ sec : mut ] is devoted to the asymptotic behaviour of the number of mutations appearing in external and internal branches of the bolthausen - sznitman coalescent .
in this section , we use the coupling method introduced in @xcite in order to study the number of jumps @xmath28 .
let @xmath52 be a sequence of i.i.d .
random variables with distribution @xmath53 note that @xmath54 let @xmath55 .
it is well - known , see for instance @xcite , that @xmath56{d}z,\ ] ] where @xmath17 is the stable random variable that appears in .
we have the following functional limit result , with a limit , which is certainly a lvy process .
[ lem : fltv ] the process @xmath57 defined by @xmath58 converges weakly in the skorohod space @xmath59 $ ] .
we first verify that the convergence of finite - dimensional distributions holds .
let @xmath60 , from ( [ gdh ] ) , we deduce @xmath61{d}tz+t\log t.\ ] ] similarly , if we take @xmath62 , then @xmath63{d}(z_1,z_2),\ ] ] where @xmath64 and @xmath65 are independent random variables distributed as @xmath66 and @xmath67 , respectively .
the mapping theorem implies that @xmath68{d}(z_1,z_1+z_2).\ ] ] a set of three or more time points can be treated in the same way , and hence the finite - dimensional distributions converge properly .
we now check tightness via aldous criterion .
let @xmath69 be a @xmath70-stopping time and @xmath71 a sequence of positive numbers such that @xmath72 , as @xmath13 increases .
then for @xmath73 , we have @xmath74 which converges to 0 .
this completes the proof . in
what follows , we use the following notation . for
a stochastic process @xmath75 and a function @xmath76 write @xmath77 as @xmath78 , if @xmath79 is stochastically bounded as @xmath80 , i.e. if @xmath81 .
we also write @xmath82 as @xmath80 , if @xmath79 goes to @xmath83 in probability .
+ from the above result , we deduce @xmath84 we now define recursively @xmath85 , a sequence of stopping times such that @xmath86 and @xmath87 with the convention @xmath88 . in other words ,
the sequence @xmath89 is the collection of indices of the r.v.s @xmath90 such that their sum does not exceed @xmath91 .
it is proved in @xcite that @xmath28 and @xmath92 are equal in law , and that the terms of the block - counting markov chain of the bolthausen - sznitman coalescent can be represented as @xmath23 , and @xmath93 next , we define @xmath94 the first time that the random walk meets or exceeds @xmath13 , and @xmath95,\ ] ] with the convention @xmath96 .
our first result allows us to consider the random walk instead of the process of disappearing blocks until time @xmath97 .
[ lem : grolem ] let @xmath98 .
then as @xmath13 goes to @xmath32 , we have @xmath99{\mathbb{p}}1.\ ] ] moreover , @xmath100 for @xmath13 sufficiently large . in order to prove this proposition ,
we first show that a similar result holds for the family of stopping times @xmath101 where @xmath49 is a positive constant , and then note that for @xmath102 , @xmath103{}1.\ ] ] hence , the proof of proposition [ lem : grolem ] relies on the following lemma .
[ lem : grolem1 ] let @xmath104 .
then as @xmath13 goes to @xmath32 , we have @xmath105{\mathbb{p}}1.\ ] ] moreover , @xmath106 for @xmath13 sufficiently large .
first we prove the result for @xmath107 .
observe , from , that for any @xmath73 @xmath108 { } 1.\ ] ] now , since @xmath109 , as @xmath13 increases ( which follows from ) , we get @xmath110 for simplicity , we write @xmath111 instead of @xmath112 .
then , it follows @xmath113 thus if we take @xmath114 in ( [ eq : maxv ] ) , we deduce @xmath115 { } 0.\ ] ] on the event @xmath116 , it is clear that @xmath117 hence , from ( [ eq : maxv ] ) and ( [ eq : siginfeta ] ) , we obtain @xmath118 in particular , since @xmath119 , we get @xmath120{{{\mathbb p}}}1.\ ] ] next , we note @xmath121 and from ( [ eq : abton ] ) , it is clear @xmath122 then on the event @xmath123 , it follows from that @xmath124 finally using ( [ eq : asser2step1 ] ) and the strong markov property for @xmath125 , we deduce @xmath126 then , putting all the pieces together , we get @xmath127 and since @xmath107 , @xmath128 next , we will prove ( [ eq : siginfeta ] ) , ( [ eq : ratio succ ] ) , ( [ eq : asser2step1 ] ) , ( [ eq : tau - eta ] ) and ( [ eq : recu1 ] ) for any @xmath129 . we show that this claim holds for @xmath130 for any @xmath131 , using induction on @xmath132 .
the proof for @xmath133 has just been done .
for the induction step suppose that the asymptotics in ( [ eq : siginfeta ] ) to ( [ eq : recu1 ] ) hold for @xmath130 . for simplicity ,
we write @xmath134 .
the idea is to use the strong markov property at the stopping time @xmath135 and apply the above results for @xmath136 to the markov chain @xmath137 started at @xmath138 ( instead of @xmath139 ) .
define the family of stopping times @xmath140 observe that @xmath141 .
hence , using the strong markov property at the stopping time @xmath135 and the behaviour in ( [ eq : asser2step1 ] ) , with @xmath142 , we get @xmath143{{{\mathbb p}}}1.\ ] ] then , from this asymptotic behaviour and the induction hypothesis taken in ( [ eq : asser2step1 ] ) , @xmath144{{{\mathbb p}}}1.\ ] ] from this behaviour , from ( [ eq : asser2step1 ] ) and from @xmath145 we obtain for @xmath146 @xmath147 { } 1.\ ] ] now , on the event @xmath148 , using the strong markov property at @xmath135 and ( [ eq : siginfeta ] ) with the initial state @xmath149 , we get @xmath150 { } 0 .\ ] ] the induction hypothesis gives @xmath151 , as @xmath13 goes to @xmath32 .
these two facts together lead to @xmath152 { } 1,\ ] ] for @xmath153 $ ] . from ( [ eq : ratio succ ] ) and again the strong markov property at @xmath135 , we get @xmath154 from the above behaviour , ( [ eq : hetaetazeta ] ) and the induction hypothesis , we have @xmath155 for @xmath153 $ ] . again the strong markov property
together with the above behaviour give us for all @xmath153 $ ] , @xmath156 from ( [ eq : recu1 ] ) and the strong markov property , we get @xmath157 we know from the induction hypothesis that @xmath158 , as @xmath13 goes to @xmath32 .
we then obtain , from ( [ eq : hetaetazeta ] ) and using again the induction hypothesis , @xmath159 for all @xmath153 $ ] .
hence the induction is complete and the behaviour in ( [ eq : siginfeta ] ) to ( [ eq : recu1 ] ) hold for any @xmath129 .
we first recall the definition of @xmath97 in ( [ eq : defth ] ) and define @xmath160 and @xmath161 . from ( [ eq
: tau - eta ] ) , it is clear @xmath162 { } 1,\ ] ] and from ( [ eq : siginfeta ] ) , we deduce @xmath163 { } 1.\ ] ] thus the first asymptotic behaviour in ( [ eq : aser2 ] ) holds .
also note @xmath164 { } 1,\ ] ] then the second asymptotic behaviour in ( [ eq : aser2 ] ) follows from ( [ eq : asser2step1 ] ) . from ( [ eq : recu1 ] ) ,
we get @xmath165 which gives ( [ eq : aser3 ] ) for @xmath166 .
also @xmath167 this give us ( [ eq : aser3 ] ) .
this completes the proof .
we first define @xmath168 which is obtained by replacing the exponential random variables @xmath169 s by their mean and approximating the denominator .
similarly , we define @xmath170}{{x^{(n)}_{k}}},\ ] ] which is obtained by replacing the random variables @xmath33 by its conditional expectation .
this new formulation is of interest . indeed ,
similar as in @xcite it is possible to determine @xmath171 via a recursive formula .
let @xmath172 be the number of external branches after @xmath9 jumps , @xmath173 , and we take conditional expectation to each @xmath172 with respect to @xmath174 and @xmath175 .
observe that @xmath176 is the number of external branches which participate to @xmath9-th coalescent event .
hence , this random variable is distributed as an hypergeometric r.v . with parameters
@xmath177 , @xmath175 and @xmath178 .
it is then clear @xmath179= { z^{(n)}_{k-1}}-\big(1+{u^{(n)}_{k}}\big)\frac{{z^{(n)}_{k-1}}}{{x^{(n)}_{k-1}}},\ ] ] then @xmath180 = { { \mathbb e}}\big[{z^{(n)}_{k-1}}\big|{x^{(n)}_{}}\big]\frac{{x^{(n)}_{k}}-1}{{x^{(n)}_{k-1}}},\ ] ] and @xmath181}{{x^{(n)}_{k } } } = \prod_{i=1}^k\left(1-\frac{1}{{x^{(n)}_{i}}}\right).\ ] ] finally , since @xmath182 , it follows @xmath183 this last expression is a good way to understand the asymptotic behaviour of the total internal branch .
the following lemma provides the asymptotic behaviour of @xmath184 .
[ lemm : aincv ] as @xmath13 goes to @xmath32 , @xmath185{}1,\ ] ] in probability .
let @xmath73 and take @xmath97 as in ( [ eq : defth ] ) .
we also let @xmath186 and @xmath187 and consider @xmath171 as it is given in .
we now split @xmath171 in two parts , as follows @xmath188 and @xmath189 note that @xmath190 which implies that @xmath191 { } 0,\qquad \textrm { almost surely.}\ ] ] then it is enough to study the behaviour of @xmath192 . in order to do so , we first note @xmath193 ( this can be viewed as two bonferroni inequalities for independent events with entrance probabilities @xmath194 . ) on the one hand , @xmath195 and thus @xmath196 from ( [ eq : aser3 ] ) , we get @xmath197 from the fact that @xmath198 , as @xmath80 , we deduce @xmath199 on the other hand by inverting the sums , we obtain @xmath200 using ( [ eq : aser3 ] ) , we obtain @xmath201 and finally @xmath202 putting all the pieces together give us @xmath203{{{\mathbb p}}}1,\ ] ] which ends the proof . in order to prove theorem [ th : cvi ] , we just need to control our approximation .
this is the aim of the next two lemmas .
[ lem : app1 ] as @xmath13 goes to @xmath32 , @xmath204 recall that @xmath174 denotes the markov chain @xmath205 .
a simple computation gives us @xmath206 conditionally on @xmath207 , the random variables @xmath208 are independent with zero mean .
this implies @xmath209 = \sum_{k=1}^{\tau^{(n)}-1 } \left(\frac{{y^{(n)}_{k } } } { { { x^{(n)}_{k}}}}\right)^2 \leq{\tau^{(n)}}\leq n,\ ] ] where the inequality follows from the fact that @xmath210 a.s .
chebychev s inequality implies @xmath211 again using that @xmath210 a.s . , we get @xmath212 it is a classical result ( used also for the total length of kingman coalescent ) that @xmath213{{{\mathbb p}}}1,\ ] ] which implies that @xmath214 this completes the proof .
[ lem : app2 ] as @xmath13 goes to @xmath32 @xmath215 we proceed similar as in @xcite . recall that @xmath172 is the number of external branches after @xmath9 coalescing events . since @xmath216 , @xmath217}{{x^{(n)}_{k}}}\ ] ] also recall that @xmath218 has a conditional hypergeometric distribution , given @xmath219 .
therefore @xmath220 where @xmath221 denotes a random variable with conditional hypergeometric distribution with parameters @xmath177 , @xmath175 and @xmath178 as above , centered at its ( conditional ) expectation . for @xmath222\ ] ] it follows @xmath223 iterating this linear recursion we obtain because of @xmath224 @xmath225 and consequently @xmath226 now , since the @xmath227 are centered hypergeometric variables , they are uncorrelated , given @xmath174 . also from the formula for the variance of a hypergeometric distribution @xmath228 \le ( { u^{(n)}_{j}}+1)\frac{{z^{(n)}_{j-1}}}{{x^{(n)}_{j-1}}}\ ] ] thus , since @xmath229 a.s .
@xmath230 \le ( { u^{(n)}_{j}}+1 ) .\ ] ] putting everything together we obtain @xmath231 \le \sum_{j=1}^{\tau^{(n)}-1 } \frac{{u^{(n)}_{j}}+1}{({x^{(n)}_{j}})^2}\big(\sum_{k = j}^{\tau^{(n)}-1}\prod_{i = j+1}^k \big ( 1- \frac 1{{x^{(n)}_{i}}}\big)\big)^2.\ ] ] the product can be estimated by 1 , thus @xmath231 \le \sum_{j=1}^{\tau^{(n)}-1 } \frac{{u^{(n)}_{j}}+1}{({x^{(n)}_{j}})^2 } ( \tau^{(n ) } - j)^2.\ ] ] by means of @xmath232 @xmath231 \le \sum_{j=1}^{\tau^{(n)}-1 } ( { u^{(n)}_{j}}+1 ) \le n+
\tau_n \le 2n .\ ] ] now an application of chebychev s inequality gives the claim .
let us now suppose that mutations occur along genealogical trees according to a poisson process of intensity @xmath233 .
we write by @xmath234 for the total number of mutations in the bolthausen - sznitman @xmath13-coalescent .
the poissonian representation implies that , conditionally on @xmath235 , @xmath234 is distributed as a poisson r.v . with parameter @xmath236 .
mutations can be divided as external and internal according to the type of the branches where they appear and we denote them by @xmath237 and @xmath238 , respectively .
let @xmath241 be a poisson process with parameter @xmath233 .
we first note that conditionally on @xmath31 , @xmath242 has the same distribution as @xmath243 .
this implies @xmath244={{\mathbb e}}\big[{{\mathbb e}}\big[{m^{(n)}_{i}}|{i^{(n)}}\big]\big]=\mu{{\mathbb e}}\big[{i^{(n)}}\big]\ .\ ] ] since @xmath245 a.s .
, thanks theorem [ th : cvi ] , we deduce that @xmath246 in probability and @xmath247}\stackrel{d}{= } \frac{n_{{i^{(n)}}}}{\mu{i^{(n)}}}\frac{{i^{(n)}}}{{{\mathbb e}}\big[{i^{(n)}}\big]}\xrightarrow[n\to\infty]{{{\mathbb p}}}1,\ ] ] thanks to theorem [ th : cvi ] .
therefore , the first result follows from @xmath248=\mu{{\mathbb e}}[{i^{(n)}}]\sim \mu n/(\log n)^2 $ ] , as @xmath80 .
, asymptotic results for coalescent processes without proper frequencies and applications to the two - parameter poisson - dirichlet coalescent , _ stochastic process .
_ * 120 * ( 2010 ) , no .
11 , 21592173 . | in this paper , we study a weak law of large numbers for the total internal length of the bolthausen - szmitman coalescent . as a consequence ,
we obtain the weak limit law of the centered and rescaled total external length .
the latter extends results obtained by dhersin @xmath0 mhle @xcite .
an application to population genetics dealing with the total number of mutations in the genealogical tree is also given . |
small scale magnetic elements or magnetic flux concentrations are described by flux tubes , often with kg field strengths , located in intergranular downflow lanes @xcite . studying the intensity contrast of magnetic elements relative to the qs in the continuum , and line core of spectral lines , is of importance , because it provides information about their thermal structure .
due to their enhanced brightness , particularly in the cores of spectral lines @xcite and in the uv @xcite , these elements are believed to contribute to the variation of the solar irradiance , especially on solar cycle time scales @xcite .
the contrast in the visible , and in the uv spectral ranges contributes by 30% and 60% respectively , to the variation of the total solar irradiance ( tsi ) between minimum and maximum activity @xcite and spectral lines contribute to a large part of this variation @xcite . the magnetic flux in magnetic elements is also believed to be responsible for the structuring and heating of the chromosphere and corona .
the relationship with chromospheric heating is indicated by the strong relationship between excess brightening in the core of the caii k - line and the photospheric magnetic flux ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
+ several studies have been carried out to investigate how the brightness at given wavelength bands depends on the photospheric magnetic field at disc centre , and different results have been obtained .
@xcite and @xcite made a pixel - by - pixel comparison of the longitudinal magnetogram signal and the continumm intensity at 676.8 , 525 , 557.6 , and 630.2 nm , using simultaneous continuum filtergrams , in addition to line centre filtergrams at 676.8 nm in @xcite , and magnetograms of magnetic features in active regions at disc centre , acquired by the 50 cm swedish solar vacuum telescope with a 0.3@xmath2 resolution .
they obtained negative contrasts in the continuum ( darker than the average qs ) for all magnetogram signals ( as long as the observations were carried out almost exactly at solar disc centre ) , and an increase in the line core brightness with the magnetogram signal until 600 g followed by a monotonic decline .
@xcite applied the same method at roughly the same spatial resolution as @xcite to quiet - sun network data .
their scatterplot showed positive contrast values ( brighter than the average qs ) for magnetic fields larger than 200 g , until nearly 500 g , and negative contrasts for higher magnetic fields .
@xcite used low resolution ( 4@xmath2 ) , but simultaneous magnetograms and full - disc continuum intensity images of the nii 6768 absorption line , recorded by the michelson doppler imager ( mdi ) on - board the solar and heliospheric observatory ( soho ) .
they found that the contrast close to disc centre ( at @xmath3 ) initially increases slightly with the magnetic field , before decreasing again for larger fields .
@xcite made the same pixel - by - pixel study of the continuum contrast at 630.2 nm with the longitudinal magnetic field in the quiet - sun network and in active region ( ar ) plage near disc centre , using data from the solar optical telecope on - board hinode ( 0.3@xmath2 spatial resolution ) .
they found that for both the qs and ar , the contrast initially decreases and then increases for weak fields , until reaching a peak at @xmath4 700 g , to decrease again for stronger fields , even when the pores in their ar fields of view ( fovs ) were expilicitly removed from their analysis .
+ they explained the initial , rapid decrease to have a convective cause , with the bright granules harbouring weaker fields than the intergranular lanes .
the following increase in contrast is due to the magnetic elements being brighter than the average qs .
to explain the final decrease at high field strengths , they argued that due to their limited spatial resolution ( 0.3@xmath2 for hinode / sp ) , many flux tubes were not resolved , and therefore , the field strength of small bright elements is attenuated , while those of bigger structures ( such as micropores ) , which are darker than the mean qs , were weakly affected by the finite resolution , so that the average contrast showed a peak at intermediate field strengths , and decreased at higher strengths . at a constant spatial resolution of
1@xmath2 achieved by the helioseismic and magnetic imager ( hmi ) on - board the solar dynamics observatory ( sdo ) , @xcite studied the dependence of the continuum and line core contrasts in the fei line at 6173 , of network and faculae regions on disc position and magnetogram signal , using simultaneous full - disc magnetograms and intensity images . for a quiet - sun region at disc centre ( @xmath5 ) , their scatterplots of the continuum intensity against @xmath6 exhibited a peak at @xmath7 200 g .
+ @xcite simulated a plage region using the muram code @xcite .
they studied the relation between the continuum contrast at 630.2 nm and the vertical magnetic field both , at the original muram resolution and at the resolution of telescopes with 1.0 m and 0.5 m apertures . for the original resolution , the contrast monotonically increased with increasing field strength , confirming the expectations of a thin flux - tube model @xcite , which predicts that for higher field strengths , the flux tubes get more evacuated , which leads to lateral inflow of heat from the hot walls of the evacuated flux concentrations , and therefore an increase in brightness as a result of the optical depth surface depression , which allows deeper layers to be seen . according to @xcite , at the resolution of a 1.0 m telescope the average simulated contrast saturates for stronger fields , while it shows a turnover at the resolution corresponding to 0.5 m. this points to a non - trivial effect of finite spatial resolution on the relation between continuum brightness , and photospheric magnetic field .
+ in contrast to continuum radiation , spectral lines display a relatively monotonic increase in rest intensity with magnetic flux . of particular importance
is the caii h line , due to its formation in the chromosphere .
@xcite showed by using simultaneous observations of the calcium network and photospheric magnetic field that the line core of fei at 5250.2 and the caii k emission increases with the magnetic field until his limit of about 500 g .
@xcite carried out a quantitative study of the relationship between the caii k emission and magnetic flux density in an active region ( outside sunspots ) . after subtracting the basal flux ( the non - magnetic contribution of the chromospheric emission )
they found that the relation follows a power law , with an exponent of 0.6 .
@xcite found the same relation with a power - law exponent of 0.66 , for a quiet - sun region at disc centre , while @xcite reported a value of 0.2 ( including the internetwork ) , and 0.40.5 for the network , with a strong dependence of the power exponent on the magnetic field threshold .
@xcite in their study of the correlation between emissions at different chromospheric heights with the photospheric magnetic field in a quiet - sun region close to the disc centre , found an exponent of 0.31 for caii k. + here , we study the contrast at a number of wavelengths in the nuv ( between 214 nm and 397 nm ) and the visible ( around 525 nm ) in the qs close to solar disc centre . we employ high - resolution , seeing - free measurements of both , the intensity and the magnetic field obtained with the @xmath0 balloon - borne observatory .
+ the structure of the rest of the paper is as follows . in sec .
2 , we describe the data used for this analysis , in addition to presenting the detailed data reduction steps . in sec . 3 , we present our results , and compare them to the literature . in sec . 4 , we summarize and discuss the results .
we use a time series recorded during the first @xmath0 flight on 2009 june 9 , between 14:22 and 15:00 ut .
@xmath0 is composed of a telescope with a 1.0 m diameter main mirror mounted on a gondola with two post - focus instruments @xcite .
it carries out its observations hanging from a stratospheric balloon .
the images are stabilized against small - scale motions by a tip - tilt mirror placed in the light distribution unit connected to a correlation tracker and wavefront sensor @xcite .
@xmath0 carried two instruments , a uv imager ( sufi ) and a magnetograph ( imax ) . + the imaging magnetograph experiment ( imax ; * ? ? ?
* ) acquired the anaylsed spectropolarimetric data by scanning the photospheric fei line at 5250.2 ( land factor @xmath8 = 3 ) at five wavelengths positions ( 4 within the line at @xmath9 , @xmath10 , @xmath11 , @xmath12 m and one in the continuum at @xmath13 m from the line centre ) , with a spectral resolution of 85 m , and measuring the full stokes vector @xmath14 at each wavelength position .
the total cadence for the v5 - 6 observing mode ( v for vector mode with 5 scan positions and 6 accumulations of 250 ms each ) was 33s .
we use the phase - diversity reconstructed data with a noise level of 3@xmath1510@xmath16 , and achieving a spatial resolution of @xmath17 ( see @xcite for more details on the instrument , data reduction and data properties ) . for the nuv observations , we use the data acquired by the @xmath0 filter imager ( sufi ;
* ) quasi - simultaneously with imax , in the spectral regions 214 nm , 300 nm , 313 ( oh - band ) nm , 388 nm ( cn - band ) , and 397 nm ( core of caii h ) , at a bandwidth of 10 nm , 5 nm , 1.2 nm , 0.8 nm , and 0.18 nm , respectively .
the cadence of the sufi data for a given wavelength was 39s .
we analyse data that were reconstructed using wave - front errors obtained from the in - flight phase - diversity measurements , via an image doubler in front of the ccd camera ( level 3 data , see * ? ? ?
* ; * ? ? ?
the sufi data were corrected for stray light by deconvolving them with the stray light modulation transfer function ( mtfs ) derived from comparing the limb intensity profiles recorded for the different wavelengths , with those from the literature @xcite .
but for stokes @xmath18 line core normalized to the local continuum . ] in this study we use the line - of - sight ( los ) component of the magnetic field vector @xmath19 , which was retrieved from the reconstructed stokes images by applying the spinor inversion code @xcite .
such an inversion code assumes that the five spectral positions within the 525.02 nm fei line are recorded simultaneously .
this is not the case for the imax instrument that scans through the positions sequentially with a total acquisition time of 33s . to compensate for the solar evolution during this cycle time we interpolated the spectral scans with respect to time .
we also corrected the data for stray light because we expect that the stray - light contamination has a serious effect on the inversion results , in particular on the magnetic field results in the darker regions ( intergranular lanes , micropores ) .
unfortunately , feller et al .
( in preparation ) could only determine the stray - light mtf for the stokes @xmath18 continuum images but not for the other spectral positions .
since a severe wavelength dependence of the stray light can not be ruled out , we decided to apply a simplistic global stray - light correction to the imax data by subtracting 12% ( this value corresponds to the far off - limb offset determined in the continuum by feller et al .
, in preparation ) of the spatial mean stokes profile from the individual profiles .
after applying the time interpolation and the stray - light correction , the cleaned data were inverted with the traditional version of the spinor code . in order to get robust results ,
a simple one - component atmospheric model was applied that consists of three optical depth nodes for the temperature ( at @xmath20 ) and a height - independent magnetic field vector , line - of - sight velocity and micro - turbulence .
the spectral resolution of the instrument was considered by convolving the synthetic spectra with the spectral point - spread function of imax ( see bottom panel of fig . 1 in * ? ? ?
the spinor inversion code was run five times in a row with ten iterations each .
the output of a run was smoothed and given as initial atmosphere to the following run .
the strength of the smoothing was gradually decreased which lowered the spatial discontinuities in the physical quantities caused by local minima in the merit function .
the final los velocity map was then corrected by the etalon blueshift which is an unavoidable instrumental effect of a collimated setup ( see * ? ? ?
* ) and a constant velocity was removed from the map so that the spatially averaged velocity is zero . the inversion strategy used for the 2009 imax data analysed in this paper is identical to the one applied to the 2013 data which is described in more detail by @xcite . when comparing the sufi and imax data , we need to align the two data sets with each other and transform them to the same pixel scale . because some of the sufi wavelengths show granulation ( at 300 nm , 313 nm , 388 nm , and 214 nm ) ,
these were aligned with imax stokes @xmath18 continuum images .
the sufi caii h images at 397 nm were aligned with imax stokes @xmath18 line - core images ( see sect .
[ contrast_def ] for the derivation of the line - core intensity ) , since both wavelength bands sample higher layers in the photosphere and display reversed granulation patterns . in a first step , the plate scale of sufi images of roughly 0@xmath2102 pixel@xmath22 were resampled via bi - linear interpolation to the plate scale of imax ( 0@xmath2105 pixel@xmath22 ) . after setting all the images to the same pixel size , imax images with @xmath23 fov
were first trimmed to exclude the edges lost by apodisation .
the usable imax fovs of @xmath24 were then flipped upside down , and trimmed to roughly match the fov of the corresponding sufi images of @xmath25 .
a cross - correlation technique was used to compute the horizontal and vertical shifts with sub - pixel accuracy .
after shifting , the imax and sufi fovs were trimmed to the common fov ( cfov ) of all data sets of @xmath26 .
this value is smaller than the fov of individual sufi data sets since the images taken in the different sufi filters are slightly shifted with respect to each other due to different widths and tilt angles of the used interference filters .
figure [ images ] shows from right to left , an imax stokes @xmath18 continuum contrast image with an effective fov of @xmath24 , with the cfov overlaid in red , an imax stokes @xmath18 line core contrast image trimmed to the cfov , and the corresponding resampled and aligned sufi contrast images at 397 nm , 388 nm , 313 nm , 300 nm , and 214 nm ( see sect . [ contrast_def ] for the definition of intensity contrast ) .
+ the reversed granulation pattern in the line - core image is more visible if normalized to the local continuum intensity as shown in fig .
[ lc_qs ] . the relative intensity ( hereafter referred to as contrast ) , @xmath27 at each pixel for each wavelength band ,
@xmath28 is computed as follows : @xmath29 where @xmath30 and @xmath31 are the imax stokes @xmath18 continuum and line - core intensity contrasts , respectively .
@xmath32 , @xmath33 , @xmath34 , @xmath35 and @xmath36 are the sufi intensity contrasts at 214 nm , 300 nm , 313 nm , 388 nm and 397 nm , respectively .
+ @xmath37 is the mean quiet - sun intensity averaged over the entire common fov . when comparing sufi contrast with imax - based magnetic field parameters this common fov is @xmath26 , when comparing imax continuum or line - core intensity with imax magnetic field , the full usable imax fov of @xmath24 is employed .
+ the line - core intensity , @xmath38 at each pixel was computed from a gaussian fit to the 4 inner wavelength points of the stokes @xmath18 profile . for comparison later in sect .
[ scatter_ca ] , we also compute the line core by averaging the imax stokes @xmath18 intensity at @xmath10 m@xmath39 and @xmath11 m@xmath39 from the line centre : @xmath40 each scatterplot in sect .
[ plots ] contains data from all the 40 available images in the time series .
pixel - by - pixel scatterplots of the imax continuum and line core contrasts are plotted vs. the longitudinal component of the magnetic field , @xmath1 in figs .
[ cont_con ] and [ lc_con ] .
+ the contrast values are averaged into bins , each containing 500 data points , which are overplotted in red on figs .
[ cont_con ] and [ lc_con ] , as well as on figs .
[ b_c_hinode ] and [ sufi ] that are discussed later . ,
but for the imax line - core contrast derived from gaussian fits to the stokes @xmath18 line profiles .
] for the continuum contrast ( fig . [ cont_con ] )
, the large scatter around @xmath41 is due to the granulation . at very weak fields , the average contrast decreases with increasing field strength , because only the weakest fields are present in granules , while slightly stronger fields , close to the equipartition value , are concentrated by flux expulsion in the dark intergranular lanes @xcite .
these weak fields are typically at or below the equipartition field strength of around 200400 g at the solar surface ( e.g. , * ? ? ?
* ) , which corresponds roughly to 120240 km about 1 scale height above the solar surface , a very rough estimate of the height at which fei 525.02 nm senses the magnetic field .
fields of this strength have little effect on the contrast , so that these pixels are darker than the mean quiet - sun intensity ( shown as the horizontal dashed red line ) .
the contrast reaches a minimum at approximately 80 g , then increases with increasing field strength , as the pressure in the flux tubes decreases , and these brighten , becoming brighter than the mean qs at around 600 g .
together , these various effects give rise to the fishhook " shape of the continuum contrast curve , as described by @xcite .
the contrast then saturates at larger field strengths .
cccc[!htb ] 90 & 0.17 @xmath420.001 & [email protected] & 7.62 + 130 & 0.18 @xmath420.001&[email protected]&4.04 + 170 & [email protected]&[email protected]&3.45 + 210 & 0.19@xmath42 0.002&[email protected]&3.26 + 250 & [email protected]&[email protected]&3.05 + [ cont_log ] cccc[!htb ] 100 & [email protected] & [email protected] & 6.3 + 140 & 0.80 @xmath420.002&[email protected]&3.15 + 180 & [email protected]&[email protected]&2.55 + 220 & 0.81@xmath42 0.003&[email protected]&2.28 + 240 & [email protected]&[email protected]&2.22 + [ lc_log ] the contrast values reached in the line core data ( fig .
[ lc_con ] ) are much higher than those in the continuum , in agreement with @xcite and @xcite , and are on average larger than the mean qs intensity for @xmath43 50 g . the high average contrast values ( both in the continuum and line core ) with respect to the mean qs for strong magnetic fields proves the enhanced brightness property of small scale magnetic elements present in our data .
moreover , the average continuum contrast values reported here , are higher than the ones measured by @xcite , partly due to our higher spatial resolution , but partly also due to the shorter wavelength of 525 nm vs. 630 nm of the hinode data employed by @xcite .
in addition , our plots do not show any peak in the contrast at intermediate field strengths , nor a downturn at higher values as reported by @xcite and @xcite , or a monotonic decrease as obtained by @xcite .
+ to reproduce the scatterplot obtained by @xcite , and to show the effect of spatial resolution on the shape of the @xmath30 vs. @xmath1 scatterplot , we degrade our data ( stokes @xmath18 and @xmath44 images ) to the spatial resolution of hinode , with a gaussian of 0@xmath2132 _ fwhm_. then , the magnetic field is computed at each pixel , using the centre of gravity ( cog ) technique @xcite for the degraded stokes images .
figure [ b_c_hinode ] shows the corresponding scatterplot , based on the degraded contrast and magnetic field images , with the data points binned in the same manner as for the undegraded data .
one can clearly see both , a decrease in the contrast values and a leftward shift of data points towards lower magnetogram signals . upon averaging , the binned contrast peaks at intermediate field stengths , and turns downwards at higher values . + , but after degrading the underlying continuum stokes images with a gaussian of fwhm 0@xmath2132 to mimic the spatial resolution of the hinode spectropolarimter . ] to derive a quantitative relationship between the continuum contrast and the quiet - sun magnetic field , we tried fitting the scatterplot with a power - law function of the form : @xmath45 this function could not represent our scatterplots since the log log plots ( not shown ) did not display a straight line in the range of @xmath1 where we expected the fit to work , i.e. above the minimum point of the fishhook shape described earlier .
consequently , we looked for other fitting functions , and we found that the scatterplots could be succesfully fitted with a logarithmic function , which is the first time that it is used to describe such contrast curves : @xmath46 the fit represents the data quite well ( the lin
log plots show a straight line ) for data points lying above 90 g for the continuum contrast vs @xmath1 , and from 100 g for the line - core contrast vs. @xmath1 . to investigate how the quality of the fit and the best - fit parameters depend on the magnetic field threshold below which all the data points are ignored , we list in tables [ cont_log ] and [ lc_log ] the best - fit parameters for the continuum and line core contrast variation with @xmath1 , respectively , along with the magnetic flux threshold , and the corresponding @xmath47 values . fitting the original data points or the binned values
returns similar results .
we have plotted and tabulated the curves obtained by fitting the binned values .
the @xmath47 values are large for smaller thresholds , and decrease with increasing threshold where less points are fitted .
in contrast to this , the best - fit parameters show only a rather small variation with the threshold used , which is not the case with the power - law fit used later in sec .
[ scatter_ca ] when describing the relationship between the caii h emission and @xmath1 , a relationship that has traditionally been described with a power - law function .
+ in order to test the validity of the binning method used to represent the trend of the scatterplots throughout the paper , and to which the parametric models described above ( logarithmic and power - law models ) are fitted , we also apply non - paramteric regression ( npr ) methods to the data points .
these methods do not require specific assumptions about how the data should behave , and are used to find a non - linear relationship between the contrast and magnetic field by estimating locally the contrast value at each @xmath1 , depending on the neighbouring data points .
we show in fig .
[ bc_smooth ] the scatterplot of the imax continuum contrast vs. @xmath1 discussed earlier in this section and depicted in fig .
[ cont_con ] .
we plot in blue the logarithmic fit extrapolated to small @xmath1 values .
the red curve is the graph joining the binned contrast values and the green curve is the regression curve obtained after applying one of the npr techniques that are described in detail in the appendix .
the npr curve fits the data , including the fishhook shape at small values of @xmath1 , and lies extremely close to the curve produced by binning contrast values .
this agreement gives us considerable confidence in the binned values we have used to compare with the simple analytical model functions .
the most relevant conclusion that can be drawn from fig .
[ bc_smooth ] is that our binning method is appropriate to represent the behaviour of the data points , and that it is valid to fit the logarithmic or power - law models to the binned values of the data points .
this test was repeated also for the scatterplots analysed in the next sections and the results are discussed in appendix . [ npr_tests ] . of the imax continuum vs. @xmath1 .
the red curve is the binned values of the contrast .
the green curve the non - parametric regression curve obtained from applying the kernel smoothing technique , with a local - polynomial of order @xmath48 as a regression method ( check appendix for more details ) .
the blue curve is the logarithmic fit applied to data points starting at 90 g and extrapolated to smaller @xmath1 values . ]
the pixel - by - pixel scatterplots of the contrast at 214 nm , 300 nm , 313 nm , and 388 nm vs. @xmath1 are shown in fig .
the data points are binned following the procedure described in section [ plots_b_c ] .
+ for all wavelengths in this range , the contrasts are much larger than in the visible , especially at 214 nm where the contrast is greatly enhanced ( see * ? ? ?
* ) . the averaged contrast increases with increasing field strength , even for higher @xmath1 . also , the fishhook shape is well visible at all uv wavelengths , and the minimum in the contrast occurs at similar @xmath1 values ( 30 g
50 g ) , while contrast @xmath49 1 is reached at somewhat different @xmath1 values , ranging from 65 g to 220 g for the different uv wavelengths .
the data points are fitted with a logarithmic function ( eq . [ log ] ) , since it describes the fitted relation better than a power - law function . table [ uv_log ] lists the best - fit parameters for the different spectral regions , from a threshold of 90 g , at which the fits start to work , along with the corresponding @xmath47 values .
the logarithmic function represents well the contrast vs. @xmath1 relationship for all wavelengths in the nuv .
a test - wise fit for different magnetic field thresholds showed that the fit results are quite insensitive to the threshold .
+ cccc[!htb ] 214 & [email protected] & [email protected]&2.36 + 300 & [email protected]&[email protected]&3.44 + 313&[email protected]&[email protected]&2.83 + 388&[email protected]&[email protected]&2.54 + [ uv_log ] a scatterplot of the contrast in the sufi 397 nm caii h band vs. @xmath1 is shown in fig .
+ the caii h spectral line gets considerable contribution from the lower chromosphere .
this was shown in @xcite(see their fig .
2c ) , and @xcite ( their fig .
1 ) who determined the average formation heights of this line as seen through the wide and the narrow @xmath0/sufi caii h filter , respectively , by convolving the spectra with the corresponding transmission profile and computing the contribution function for different atmospheric models .
the model corresponding to an averaged quiet - sun area returned an average formation height of 437 km .
as pointed out in sects .
[ plots_b_c ] and [ uv_vs_b ] a logarithmic function fits the contrast vs. @xmath1 relationship better than a power - law function .
nonetheless , we have fitted the caii h contrast vs. @xmath1 relation with both , a power law ( described by eq . [ pl ] ) and a logarithmic function ( described by eq .
[ log ] ) .
we first discuss the power law fits , as these have been widely used in the literature ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as pointed out by @xcite the best - fit parameters of the power law depend significantly on the @xmath1 threshold below which the fit is not applied . to investigate this dependence
, we fit the data points exceeding different threshold values , which are listed in table [ ca_pl ] along with the corresponding best - fit parameters and @xmath47 values .
the data are well - represented by a power - law function , for points lying above 190 g . however , have strongly different best - fit parameters , depending on the threshold in @xmath1 applied prior to the fit .
the data were also succesfully fitted by a logarithmic function ( eq . [ log ] ) .
the fit works from a lower threshold ( 50 g ) than the power - law fit , and the best - fit parameters vary only slightly with the threshold as can be seen in table [ ca_log ] .
+ although the log function produces a reasonable fit starting already from 50 g , the @xmath47 is rather large for this threshold and drops to values close to unity only for a @xmath1 threshold @xmath49 190 g , although the fit parameters remain rather stable .
+ ccccc[!htb ] 190 & 0.61 @xmath420.21&[email protected]&[email protected]&0.91 + 210 & [email protected]&[email protected]&[email protected]&0.86 + 230 & 0.25@xmath42 0.09&[email protected]&[email protected]&0.81 + 250 & [email protected]&[email protected]&[email protected]&0.72 + [ ca_pl ] cccc[!htb ] 50 & [email protected]&[email protected]&3.90 + 90 & [email protected]&[email protected]&1.67 + 170&[email protected]&[email protected]&1.04 + 190&[email protected]&[email protected]&1.00 + 210&[email protected]&[email protected]&0.96 + 230&[email protected]&[email protected]&0.94 + 250&[email protected]&[email protected]&0.90 + [ ca_log ] to better compare the magnitudes of the contrast at the different wavelengths studied here , we plot in fig .
[ bins ] all the scatterplots composed of the binned points obtained in the previous sections for the nuv wavelengths at 214 nm , 300 nm , 313 nm , 388 nm , and 397 nm ( from figs .
[ sufi ] and [ b_ca ] ) , and at the visible wavelengths , i.e. the continuum at 525 nm ( from fig .
[ cont_con ] ) and line core derived from the gaussian fit to the imax stokes @xmath18 profiles ( from fig .
[ lc_con ] ) .
also plotted is the approximate line - core value obtained from the average of the @xmath10 m@xmath39 and @xmath11 m@xmath39 line positions of the stokes @xmath18 line profile using eq .
[ lc_40 ] .
several qualitative observations can be made from this graph .
firstly , the contrast reached at 214 nm is higher than the caii h contrast .
we could not find an instrumental reason for this , e.g. the calcium images were not overexposed and the response of the detector was quite linear .
therefore , we believe that the larger contrast seen in the 214 nm wavelength band is intrinsic . we expect that it is due to the very large density of lines at 214 nm and to the relatively broad caii h filter of 1.8 , which has considerable contributions from the photosphere .
the much higher temperature sensitivity of the planck function at short wavelengths also plays a role .
secondly , the large difference between the line core contrasts derived from the two different methods , with the gaussian fits to the line profile giving almost twice the contrast ( and almost reaching that of the caii h line core at 397 nm for higher fields ) suggests that the sum of intensities at @xmath50 m@xmath39 from the line centre is not a good approximation of the line core itself . at the other nuv wavelengths the contrast depends mainly on the number and the temperature sensitivity of the molecular lines in the passbands
( see * ? ? ?
thus the 388 nm has a larger contrast than both , 300 and 312 nm , due to the large density of cn lines in the former .
the constant continuum contrast reached in our scatterplots for field strengths higher than 1000 g ( see fig . [ cont_con ] ) , confirms the previous results of @xcite . they compared the relation between the bolometric intensity contrast and magnetic field strength for mhd simulations degraded to various spatial resolutions . for field strengths
higher than 300400 g they found a monotonic increase at the original full resolution , a saturation at the spatial resolution of a 1-m telescope , while at the spatial resolution of a 50 cm telescope a turnover at around 1000 g followed by a contrast decrease .
our results also agree with the analysis presented by @xcite , who found that even at hinode / sp resolution , the strong magnetic features are not resolved , leading to a turnover in their scatterplots for higher fields , similar to the behaviour obtained by @xcite , for a quiet - sun region .
the @xmath0/imax data display a saturation of the contrast at its maximum value in the visible continuum , in qualitative agreement with the work of @xcite , although the wavelength range of their simulated contrast is different , which indicates the need to repeat the study of @xcite , but for the actual measured wavelength bands and to compare these results quantitavely with the @xmath0 data .
interestingly , after degrading the @xmath0/imax data to hinode s spatial resolution , a peak and downturn of the contrast were reproduced ( see fig .
[ b_c_hinode ] ) . the higher spatial resolution reached by imax allowed us to constrain the effect of spatial resolution on the relation between continuum brightness at visible wavelengths , and the los component of the photospheric magnetic field . at a resolution of 0.15@xmath2
( twice that of hinode / sp ) , magnetic elements in the quiet - sun internetwork start to be spatially resolved ( see * ? ? ?
* ) , leading to a constant and high contrast becoming visible in strong magnetic features .
the relationship between the intensity and the photopsheric magnetic field for several wavelengths in the nuv provided new insights into the quantitative relation between the two parameters .
the wavelength range between 200 and 400 nm is of particular importance for the variable sun s influence on the earth s lower atmosphere , as the radiation at these wavelengths affects the stratospheric ozone concentration ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
although there is some convergence towards the level of variability of the solar irradiance at these wavelengths ( yeo et al .
2015 ) , there is a great need for independent tests of the employed modelled spectra in the uv .
such uv data are also expected to serve as sensitive tests of mhd simulations .
+ of the uv wavelengths imaged by @xmath0 only the caii h line ( discussed below ) and the cn band head at 388 nm have been observed at high resolution earlier .
e.g. @xcite obtain a contrast of 1.48 in bright points with the sst , which is close to the mean contrast of 1.5 we find for large field strengths .
most of the studies made so far on the relationship between the caii h emission and the photospheric magnetic field were carried out using ground - based data and different results have been obtained concerning the form of this relation . as mentioned in the introduction , several authors fitted their data with a power - law function , obtaining power - law exponents that varied from 0.2 @xcite to 0.66 @xcite .
we were also able to fit our caii h data with a power - law function , obtaining different exponents for different thresholds of the magnetic field strength from which the fit started ( see table [ ca_pl ] ) .
a nearly equally good fit was provided by a logarithmic function , starting from lower field strengths , and showing no strong variations of best - fit parameters with the threshold ( see table [ ca_log ] ) .
other advantages of the logarithmic fit are that it has a free parameter less than the power law fit and that it also works well and equally independently of the threshold for the other observed wavelengths ( see tables [ cont_log][uv_log ] ) , whereas the power - law fit did not lead to reasonable results . +
as mentioned earlier in this paper , irradiance changes from below 400 nm are the main contributors to the tsi variations over the solar cycle .
the magnetic flux from small - scale magnetic elements in the qs is believed to contribute considerably to not just these changes @xcite , but likely lie at the heart of any secular trend in irradiance variations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , which are particularly uncertain , but also particularly important for the solar influence on our climate .
this contribution depends on the size and position on the solar disc of these elements and possibly on their surroundings .
here we studied the intensity contrast of a quiet - sun region near disc centre .
next steps include repeating such a study for mhd simulations , carrying out the same study for different heliocentric angles and extending it to active region plage , so that the results can be used to test and constrain the atmosphere models used to construct spectral solar irradiance models .
the german contribution to sunrise and its reflight was funded by the max planck foundation , the strategic innovations fund of the president of the max planck society ( mpg ) , dlr , and private donations by supporting members of the max planck society , which is gratefully acknowledged .
the spanish contribution was funded by the ministerio de economa y competitividad under projects esp2013 - 47349-c6 and esp2014 - 56169-c6 , partially using european feder funds .
the hao contribution was partly funded through nasa grant number nnx13ae95 g .
this work was partly supported by the bk21 plus program through the national research foundation ( nrf ) funded by the ministry of education of korea .
kernel smoothing is a non - parametric regression technique where a non - linear relationship between two quantities , in our case the los component of the magnetic field @xmath1 , and the corresponding contrast value @xmath51 is approximated locally . in the following we write @xmath52 for @xmath1 for simplicity . at each field value @xmath53
we want to find a real valued function @xmath54 to compute the corresponding contrast value , @xmath55 or the conditional expectation of @xmath51 given @xmath53 , i.e. @xmath56 $ ] , which is the outcome of the smoothing procedure , and is called _ the smooth _ @xcite .
the curve joining the smooth values at each @xmath53 in our data set is called the non - parametric regression or npr curve . to do this , the function @xmath54 is estimated at the @xmath57 neighboring data points , @xmath58 , falling within a bandwidth @xmath59 around @xmath53 , with @xmath60 ranging from @xmath61 to @xmath57 , and weighted by a kernel density function .
the latter is defined for @xmath53 as : @xmath62 here @xmath63 is a kernel centered at @xmath53 , giving the most weight to those @xmath64 nearest @xmath53 and the least weight to points that are furthest away .
the shape of the kernel is determined by the type of the used kernel ( often gaussian ) , and by the magnitude of the bandwidth or _ smoothing parameter _
+ following a derivation that can be found in @xcite and @xcite , the following expression for @xmath55 is found : @xmath65 for local averaging , we compute at each @xmath53 a weighted average of all the @xmath57 data points @xmath58 that fall within the bandwidth @xmath59 , with @xmath60 ranging from @xmath61 to @xmath57 . hence , the process of kernel smoothing defines a set of weights @xmath66 for each @xmath53 and defines the function @xmath54 as : apart from local averaging , the second kernel regression method we tested is local - polynomial smoothing . there
the set of @xmath57 data points @xmath58 around each field strength value @xmath53 are fit with a local polynomial of degree @xmath69 : as mentioned in the main text , these techniques are applied to the various scatterplots shown in the paper to test the validity of our binning method . in fig .
[ bc_smooth ] we showed the scatterplot of the imax continuum contrast vs. @xmath1 with the npr curve plotted in green .
it should be mentioned here that using a local averaging ( eq . [ sa ] ) or a local - polynomial fit ( eq . [ taylor ] ) returns the same result , except at the boundaries , where the curve resulting from the local - polynomial fit is smoother .
the reason is the non - equal number of data points around @xmath52 when the latter is close to the boundary , which leads to a bias at the boundary upon locally averaging the contrast values there , meaning that the computed average will be larger than one would get if data points were symmetrically distributed around @xmath53 . using a high order polynomial reduces this bias at the boundaries , but increases the variance .
we have tested different orders and saw that linear , quadratic and cubic polynomials give very similar results .
for all the scatterplots analysed here , we only show the cubic local - polynomial ( @xmath48 ) regression curve with a bandwidth of 6 g .
+ figure [ sufi_nonparam ] shows the scatterplots analysed in sect .
[ uv_vs_b ] , for the nuv contrast vs. @xmath1 at 214 nm , 300 nm , 313 nm , and 388 nm .
the binned data points are plotted in red , the npr curves in green , and the logarithmic fits ( starting from 90 g ) in dashed blue .
the non - parametric regression curves again agree almost perfectly with the binned data ( although the npr curves are smoother , with less scatter than the binned values at small @xmath1 values ) , and they agree very well with the logarithmic fits for almost all magnetic field values above 90 g . + finally , we show in fig .
[ b_ca_smooth ] the scatterplot of sufi 397 nm caii h contrast vs. @xmath1 .
the npr curve is overplotted , along with both , the logarithmic fit starting at 50 g , and the power - law fit starting at 190 g .
in addition to the similarity between the binned values plotted in red and the npr curve ( in green ) , it can be inferred from this plot that both parametric model fits reliably represent the variation of the caii h contrast vs. @xmath1 , since they also agree with the non - parametric regression curve for almost all values of @xmath1 larger than the threshold above which the fits are applied .
+ of the caii h contrast vs. @xmath1 .
the red curve is the binned contrast values .
the npr curve is plotted in green , with a bandwidth of @xmath59=6 g , the dark blue dashed curve is the logarithmic fit to the binned curve lying above @xmath1 = 50 g , and the light blue dashed curve is the power - law fit starting at 190 g . ] | we investigate the relationship between the photospheric magnetic field of small - scale magnetic elements in the quiet sun ( qs ) at disc centre , and the brightness at 214 nm , 300 nm , 313 nm , 388 nm , 397 nm , and at 525.02 nm . to this end
we analysed spectropolarimetric and imaging time series acquired simultaneously by the imax magnetograph and the sufi filter imager on - board the balloon - borne observatory @xmath0 during its first science flight in 2009 , with high spatial and temporal resolution .
we find a clear dependence of the contrast in the near ultraviolet ( nuv ) and the visible on the line - of - sight component of the magnetic field , @xmath1 , which is best described by a logarithmic model .
this function represents well the relationship between the caii h - line emission and @xmath1 , and works better than a power - law fit adopted by previous studies .
this , along with the high contrast reached at these wavelengths , will help with determining the contribution of small - scale elements in the qs to the irradiance changes for wavelengths below 388 nm . at all wavelengths including the continuum at 525.40 nm the intensity contrast does not decrease with increasing @xmath1 .
this result also strongly supports that @xmath0 has resolved small strong magnetic field elements in the internetwork , resulting in constant contrasts for large magnetic fields in our continuum contrast at 525.40 nm vs. @xmath1 scatterplot , unlike the turnover obtained in previous observational studies .
this turnover is due to the intermixing of the bright magnetic features with the dark intergranular lanes surrounding them . |
ultra - stable lasers @xcite are essential in high - resolution laser spectroscopy @xcite , optical frequency standards @xcite , gravitational wave detection @xcite , and fundamental tests of physics @xcite . to achieve such a frequency stabilized local oscillator , the laser is typically servo locked to a high - finesse reference cavity by using pound - drever - hall ( pdh ) technique @xcite . within the servo locking bandwidth ,
the stability of the locked laser is determined by the stability of the optical path - length of the reference cavity . at low frequencies ( below 100 hz )
the most severe fluctuations of the cavity length are caused by environmental disturbances such as seismic and acoustic noise . at higher frequencies
these disturbances are usually damped by active vibration isolation .
it can be shown that the sensitivity of the reference cavity to the low frequency environmental disturbances can be minimized by optimizing cavity geometry and mounting method @xcite .
while the sensitivity of a reference cavity to mechanical vibrations can be suppressed such that the influence of these perturbations can be ignored , there still exists a fundamental thermal noise limit to the level of stability achievable with the reference cavity @xcite .
this limit is attributed to brownian motion of the reference cavity spacer , mirror substrates , and optical coatings @xcite . as have been widely reported before @xcite , using short length cavities ( @xmath22 cm ) constructed entirely from ule glass , the relative frequency stability characterized by allan deviation
was limited by thermal noise to the level of @xmath23 .
several methods can be used to lower the thermal noise limit , such as increasing the reference cavity spacer length , increasing the laser spot size , and using high mechanical @xmath24 materials . of these methods , using high @xmath24 materials offer the biggest improvements . to lower the thermal noise limit of reference cavities , a commonly used high @xmath24 material is fused silica ( fs ) .
because fs has a much larger coefficient of thermal expansion ( cte ) ( @xmath25 ppb / k ) than that of ule ( @xmath26 ppb / k ) @xcite , it is not appropriate to use fs as cavity spacer material . since the dominant thermal noise contribution comes from the mirror substrates , it would be more suitable to just use fs as mirror substrates , and keep ule as cavity spacer material .
replacing the ule mirrors of a short ule cavity by fs mirrors can lower the thermal noise limit approximately by a factor of two @xcite . combining with a longer length cavity spacer , it is then possible to reach a thermal noise limit of @xmath21 .
however , the cte mismatch between fs and ule can cause a problem .
temperature changes inside the reference cavity can create stress which bends the mirror substrates , and introduces a larger effective cte for the composite cavity .
this problem can be alleviated by setting the operating temperature of the reference cavity at the zero crossing temperature of the effective cte .
if a ule spacer has its zero crossing temperature of cte at room temperature , then the zero crossing temperature of the effective cte will be much lower . however , this is not very convenient , because cooling a reference cavity is technically more challenging than simply heating the reference cavity .
if the temperature is set too low , it can introduce extra problems , like water condensation on cavity housing , _ etc .
_ for practical purpose , it is preferable to stabilize the reference cavity temperature a little bit above room temperature .
presently there are several approaches to achieve this objective .
one is to find a ule piece that has negative cte at room temperature , and in combination with positive cte of fs , it is possible to make the zero crossing temperature of the effective cte to be above room temperature @xcite .
another method is to contact ule rings to the back side of fs mirrors to prevent the mirror bending due to cte mismatch @xcite . as a consequence
, the zero crossing temperature of the effective cte can be above room temperature .
both methods are used by jiang _
et al . _ , and a fractional frequency instability of @xmath27 was achieved @xcite .
the above two approaches have their limitations .
the first approach relies on the availability of ule material with negative cte at room temperature .
this requires the good will of the ule manufacturer , and might not be always available . the second approach in principle
can be used to tune the zero crossing temperature over a range of about 30 @xmath28c , but it only works for a smaller diameter ule spacer @xcite . to lower the thermal noise of the reference cavity , it is common to choose a longer cavity length with bigger diameter .
the tuning range then is limited to only about @xmath29 @xmath28c for this geometry according to our calculations .
normal ule has the zero crossing temperature of cte varying between 5 @xmath30c and 35 @xmath30c @xcite , so with this limited tuning range it is not always possible to raise the zero crossing temperature of the effective cte to above room temperature . another approach is to use a re - entrant mirror design for the reference cavity @xcite . by using a re - entrant mirror design ,
the direction of fs mirror thermal expansion is opposite to that of ule spacer , so in principle they can be compensated better , and introduce less stress to the mirrors . in reality the tuning range of the zero crossing temperature difference is about 20 @xmath30c @xmath31 30 @xmath30c .
consequently , this cavity design is only suitable for cases when ule spacers have a zero crossing temperature lower than 5 @xmath30c . in this paper , we analyze a modified re - entrant mirror design for the reference cavity , where the ule rings for the re - entrant fs mirrors are replaced with fs rings . with this design modification ,
the zero crossing temperature of the effective cte can be tuned to much greater range , and it is very easy to design a composite cavity so that the zero crossing temperature of the effective cte is above room temperature . since a low thermal noise reference cavity design is central to this paper , a detailed discussion of the thermal noise limit of a ule cavity is presented in sec .
@xmath32 . based on our calculation results , a particular geometry of the reference cavity
is chosen , and re - entrant fs mirrors are used to replace the traditional ule mirrors . with this geometry , in sec .
@xmath33 we analyze the zero crossing temperature of the composite cavity with finite elements analysis ( fea ) , and compare the results with that of the other approaches . for completeness , in sec .
@xmath34 we also design the cavity to be vibration - insensitive .
finally we concluded in sec . @xmath35 .
the stability of a reference cavity is limited by an inevitable mechanical thermal fluctuation due to the brownian motions of materials . from the fluctuation - dissipation theorem ,
the most significant parameters in the calculation of the thermal noise spectrum for a given structure are the mechanical quality @xmath24 factors of materials .
materials with higher @xmath24 factors generate less mechanical dissipation , thus contribute smaller thermal noise . to get an order of magnitude estimation of their contributions , using the formulas given in @xcite , the power spectral density of mirror displacement due to thermal noise , @xmath36 , which typically dominates the spacer contribution ,
is given by @xmath37 where @xmath38 is the boltzmann s constant , @xmath39 is the temperature , @xmath40 is the poisson s ratio of mirror substrate , @xmath41 is the young s modulus of mirror substrate , @xmath42 is the beam radius , @xmath43 is the mechanical quality factor of the mirror substrate ( coating ) , and @xmath44 is the coating thickness ( @xmath45 @xmath46 m ) . in addition , the spacer contribution can be written as @xmath47 where @xmath48 and @xmath49 are the spacer length and radius , respectively , @xmath50 and @xmath51 is the young s modulus and mechanical quality factor of the spacer , respectively .
the total reference cavity displacement power spectral density due to thermal noise is @xmath52 from which we can calculate the allan deviation of the relative stability of the cavity as @xmath53 from eq .
( [ gm ] ) and eq .
( [ gsp ] ) , we can see that by choosing materials of higher @xmath24 factors , the thermal noise limit can be efficiently lowered . among the low cte materials that are commonly used in ultra - stable cavities
, ule has a @xmath24 factor of @xmath54 that is more than one order of magnitude larger than that of zerodur ( @xmath55 ) .
furthermore , fs has an even larger @xmath24 factor of about @xmath56 @xcite . in cases where thermal noise limits are the ultimate limiting factor of achievable cavity stability
, it is very effective to use a higher @xmath24 factor material ( like fs ) as mirror substrate while keeping ule as spacer material . according to eq .
( [ allan ] ) , using longer cavity can also have the benefit of suppressing the thermal noise .
in addition , the total thermal noise limit also depends on the laser beam sizes on the reflecting mirror surfaces .
larger beam radius contributes a lower thermal noise .
taking into account of all these effects , we calculated the thermal noise limits of cavities with different dimensions and configurations , expressed in both allan deviation and frequency spectral density , and the data are listed in table [ table ] . by replacing ule mirrors with fs mirrors and increasing the cavity length from 10 cm to 30 cm , the thermal noise limit can be suppressed by nearly one order of magnitude .
consequently in the following we will consider the design of a long , fs mirror substrate cavity .
notice here that for longer cavity , the spacer radius is also increased .
this is due to the fact that longer cavity is more susceptible to vibration influence , and larger cavity can help to suppress this effect .
the result is shown in fig .
[ vibration ] . the detail
is explained in sec . @xmath34 .
.thermal noise limits of optical reference cavities with different configurations and dimensions .
@xmath48 , @xmath49 , and @xmath57 are the spacer length , spacer radius and approximate mode spot sizes ( intensity @xmath58 radius ) on the plano - confocal cavity mirrors , respectively .
@xmath59 is the allan deviation of cavity stability and @xmath60 is the root cavity frequency noise spectral density at 1 hz for the 1064 nm light . [
cols="^,^,^,^,^,^ " , ]
the cavities designed with re - entrant fs mirrors and fs rings have an excellent cte zero crossing temperature tuning property with a very low thermal noise limit . in order to reach this thermal noise limit ,
the cavity should be further designed to be vibration insensitive .
a quantitative analysis of the elastic deformation of reference cavities provides valuable guidance for cavity design .
since the time - dependent vibration perturbation consists of a broad spectrum , in principle a full dynamic analysis of the cavity deformation is required to correctly predict the cavity vibration sensitivity .
this dynamic analysis to our knowledge is not available in the literature .
however , we can greatly simplify the process by performing static analysis .
a static analysis can be used to approximate dynamic analysis because the following two conditions are satisfied .
first , only the vibration modes with frequency lower than 100 hz are of interest , because the modes with higher frequencies will be significantly damped by either passive or active vibration isolations used in all reference cavities systems .
second , the considered cavity dimension is much smaller than the cavity vibration characteristic wavelength , meaning that all particles move in - phase inside each eigenmode .
thus the dynamic response of the cavity can be mimicked by applying a static force with the random acceleration frozen at that moment .
for example , consider ule spacer material ( with @xmath61 of 2.21 g / cm@xmath14 and @xmath63 of 67.6 gpa ) , its dispersion relation is @xcite @xmath64 for oscillation frequency of 100 hz , the characteristic wavelength @xmath65 of 55 m is much greater than the simulated 0.3 m cavity length .
consequently , at these low frequencies considered here the static analysis is a reliable substitute for the dynamic analysis . in static analysis the elastic deformation of cavities caused by a gravity - like acceleration
are studied . to understand the principle , for simplicity
we assume the cavity is held in a plane instead of several mounting points .
if the cavity is horizontally held on a plane from bottom , compression caused by acceleration normal to the supporting plane will cause expansion in length because of the non - zero poisson s ratio ; on the other hand , if the cavity is hung on the plane from top , the deformation is reversed , the cavity elongates in height and shortens in length . when the cavity is held on its symmetry plane ,
the lengthening of the upper part and the shortening of the bottom part compensate each other and the axial length of the cavity remains unchanged under vertical accelerations . in practice , cavities are held on discrete points instead of the complete plane , which will cause additional bending deformation as well . the bending effect along with the simple compression or expansion , contribute to the total cavity length variation . in this case , the optimal supporting plane is no - longer the symmetry plane and the cavity length variation under acceleration is dependent on both the discrete supporting positions and the mounting plane of those discrete points .
the bending of cavity will cause mirror tilting as well .
if the cavity is supported on points locating at the two opposite sides along the horizontal axial length , it will bend downwards under vertical acceleration and the mirrors at both sides tilt correspondingly ; if the cavity is supported on the middle of its length , as an extreme , it will bend upwards , and the mirrors tilt in an opposite way .
there must be a critical position along the length of the cavity ( airy points ) , on which the cavity is supported , so that the mirror tilting is eliminated by the compensation of both bending processes . to eliminate both cavity length variation and mirrors tilting simultaneously in the case of our cavity design
, we apply a notched structure @xcite to the cavities as shown in fig .
[ vibration1 ] . in simulations ,
we implement fea analysis by applying a gravity - like force on the cavity vertically with @xmath66 m@xmath67 , and the support with @xmath68 mm@xmath6 square on the notch surface is constrained to simulate the mounting condition .
optimal dimensions of the notched structure are searched so that the two kinds of critical mounting positions with zero length variation and zero mirrors tilting of the cavity overlap , while maintaining a less rigorous requirement of mounting accuracy and machining tolerance .
the definitions of parameters about the dimensional and mounting configurations are also illustrated in fig .
[ vibration1 ] . and @xmath69 ) of the notched structure , and positions of mounting supports ( @xmath70 and @xmath71 ) on the notch surface .
( b ) fea simulations of the deformation of a quarter section of the cavity with optimal notch sizes and supporting positions under vertical vibration.,width=377 ] during simulations , cavities with different notch sizes ( @xmath72 and @xmath69 ) are examined to search for an optimal dimension with which an overlap of the above two critical mounting positions is available . in most cases ,
the mounting positions satisfying those two conditions do not overlap , which means only one of the effects can be eliminated .
analysis results are shown in fig .
[ vibration2 ] in which the mounting positions with zero length variations and zero mirrors tilting of different notch sizes are compared .
calculations reveal that the mounting positions with zero mirror tilting angles are not sensitive to the changes of the notch sizes , while the zero crossing mounting positions ( where the length variations of cavities are immune to vertical vibrations ) are much more sensitive to the notch sizes , as shown in fig .
[ vibration2](a ) and its inset .
it is also found that at a particular value of @xmath69 ( or @xmath72 ) there is a corresponding @xmath72 ( or @xmath69 ) that can result in the overlap of these two critical positions .
another important parameter is the slope of the cavity length variation curve , which reflects the sensitivity of vibration immunity to mounting positions accuracy and machining tolerance .
different slopes of cavities with different notch sizes are shown in fig .
[ vibration2](b ) .
it indicates that the sensitivity of the frequency variation as mounting positions changes in the transverse ( @xmath73 ) direction is not sensitive to the notch sizes , neither @xmath72 nor @xmath69 , and has a value of approximately 42 khz / g / mm .
in contrast , the sensitivity with mounting positions changes in the axial ( @xmath74 ) direction steadily decreases with decreasing notch sizes of both @xmath72 and @xmath69 . as a result ,
smaller notches are much more preferred when optimizing the cavity dimensions .
we choose dimensions of @xmath72=28 mm and @xmath69=49 mm and the corresponding optimized vibration sensitivity of the cavity are shown in fig .
[ vibration3 ] . , for cavity mounting with different notch sizes .
( b ) the slopes near the zero crossing positions of curves related to the dependence of resonance frequency deviation on mounting positions under vertical vibrations with different notch sizes.,width=453 ] .
( b ) fea simulations of the axial length variations at the mirror center and mirror tilting angles at different mounting positions of @xmath70 .
the cavity has notch sizes of @xmath72=28 mm and @xmath69=49 mm , under a gravitational acceleration of @xmath66 m@xmath67.,width=453 ] with the optimized notch dimensions , both the length variation and the mirror tilting have a value of zero at the same mounting position . as can be seen in fig .
[ vibration3 ] , at the zero crossing point , the frequency variation has a very small slope against mounting position deviation in @xmath74 direction , which is 13.1 khz / g / mm , corresponding to a relative frequency instability of @xmath75 /g / mm at a laser wavelength of 1064 nm . the slope of mirror tilting in the axial direction is about @xmath76 nrad / g / mm , corresponding to about @xmath77 khz / g / mm , thus the relative frequency instability of @xmath78 /g / mm if the optical axis is displaced from the mechanical axis by @xmath79 mm .
comparably , the slope of frequency variation in transverse ( @xmath73 ) direction is four times larger than that in @xmath74 direction , as shown in fig .
[ vibration3](b ) , indicating that the cavity length variation is more sensitive and critical to the accuracy of mounting positions in @xmath73 direction .
the slope of @xmath80 khz / g / mm corresponds to a relative frequency instability of @xmath81/g / mm at @xmath82 nm .
the sensitivity of mirror tilting to mounting accuracy in @xmath73 direction is very small ( @xmath83 /g / mm at @xmath79 mm optical axis displacement ) and can be neglected .
in conclusion , we have proposed a new ultra - stable optical reference cavity design with ule spacer , re - entrant fs mirrors and fs rings based on extensive fea simulations for different cavity structures .
the designed cavity has an ultra - low thermal noise limit at the levels of @xmath0 .
the zero crossing temperature of the effective cte of the designed cavity can be easily tuned to above room temperature with any ule spacer materials , as long as the cte zero crossing temperature of the selected ule spacer is measured beforehand .
currently the cte zero crossing temperature of the ule spacer can be measured with an accuracy of @xmath84 @xmath30c at reasonable cost ( stable laser systems ) .
this accuracy level is more than enough for our design purpose .
no matter what the zero crossing temperature of the selected ule spacer is measured to be , one can always design a cavity with compatible fs rings thickness such that the zero crossing temperature of the effective cte is above room temperature .
the design is applicable for cavities with different lengths .
in addition , we have designed the reference cavity to be vibration insensitive through fea structure optimization .
ultra - low thermal noise limit reference cavities play a key role in realizing ultra - narrow linewidth lasers .
reference cavities based on ule and fs materials are still very popular choices for lab - based or portable ultra - narrow linewidth lasers . to further reduce the thermal noise limit to below @xmath0 , ule based reference cavities are no longer adequate . to this end , there are currently strong interests in the developments of cryogenic cavities based on single crystal silicon or sapphire @xcite .
we hope the technique discussed in this paper can also stimulate ideas in the design of these cryogenic cavities .
we thank yan - yi jiang and long - sheng ma for helpful discussions , and m. notcutt for technical suggestions .
the project is partially supported by the national basic research program of china ( grant no.2012cb821300 ) , the national natural science foundation of china ( grant number 61108025 and 11174095 ) , and program for new century excellent talents by the ministry of education . | an ultrastable optical reference cavity with re - entrant fused silica mirrors and a ule spacer structure is designed through finite element analysis .
the designed cavity has a low thermal noise limit of @xmath0 and a flexible zero crossing temperature of the effective coefficient of thermal expansion ( cte ) .
the cte zero crossing temperature difference between a composite cavity and a pure ule cavity can be tuned from @xmath1c to @xmath2c , which enables operation of the designed reference cavity near room temperature without worrying about the cte zero crossing temperature of the ule spacer .
the design can be applied to cavities with different lengths .
vibration immunity of the cavity is also achieved through structure optimization . 99 b. c. young , f. c. cruz , w. m. itano , and j. c. bergquist , `` visible lasers with subhertz linewidths , '' 37993802 ( 1999 ) .
w. h. oskay , w. m. itano , and j. c. bergquist , `` measurement of the @xmath3hg@xmath4 5d@xmath56s@xmath6 @xmath6d@xmath7 electric quadrupole moment and a constraint on the quadrupole shift , '' 163001 ( 2005 ) .
s. a. diddams , th .
udem , j. c. bergquist , e. a. curtis , r. e. drullinger , l. hollberg , w. m. itano , w. d. lee , c. w. oates , k. r. vogel , and d. j. wineland , `` an optical clock based on a single trapped @xmath3hg@xmath4 ion , '' science * 293 , * 825 ( 2001 ) .
h. s. margolis , g. p. barwood , g. huang , h. a. klein , s. n. lea , k. szymaniec , and p. gill , `` hertz - level measurement of the optical clock frequency in a single @xmath8sr@xmath4 ion , '' science * 306 , * 1355 ( 2004 ) .
t. schneider , e. peik , and c. tamm , `` sub - hertz optical frequency comparisons between two trapped @xmath9yb@xmath4 ions , '' 230801 ( 2005 ) .
r. le targat , x. baillard , m. fouch , a. brusch , o. tcherbakoff , g. d. rovera , and p. lemonde , `` accurate optical lattice clock with @xmath10sr atoms , '' 130801 ( 2006 ) .
t. rosenband , p. o. schmidt , d. b. hume , w. m. itano , t. m. fortier , j. e. stalnaker , k. kim , s. a. diddams , j. c. j. koelemeij , j. c. bergquist , and d. j. wineland , `` observation of the @xmath11s@xmath12 @xmath13 @xmath14p@xmath12 clock transition in @xmath15al@xmath4 , '' 220801 ( 2007 ) .
y. h. wang , r. dumke , t. liu , a. stejskal , y. n. zhao , j. zhang , z. h. lu , l. j. wang , th .
becker , and h. walther , `` absolute frequency measurement and high resolution spectroscopy of @xmath16in@xmath4 @xmath17 @xmath11s@xmath12-@xmath18 @xmath14p@xmath12 narrowline transition , '' opt . commun . * 273 , * 526 ( 2007 ) .
a. d. ludlow , t. zelevinsky , g. k. campbell , s. blatt , m. m. boyd , m. h. g. de miranda , m. j. martin , j. w. thomsen , s. m. foreman , jun ye , t. m. fortier , j. e. stalnaker , s. a. diddams , y. le coq , z. w. barber , n. poli , n. d. lemke , k. m. beck , and c. w. oates , `` sr lattice clock at @xmath0 fractional uncertainty by remote optical evaluation with a ca clock , '' science * 319 , * 1805 ( 2008 ) .
t. akatsuka , m. takamoto , and h. katori , `` optical lattice clocks with non - interacting bosons and fermions , '' nature phys .
* 4 , * 954 ( 2008 ) .
n. d. lemke , a. d. ludlow , z.w .
barber , t. m. fortier , s.a .
diddams , y. jiang , s. r. jefferts , t. p. heavner , t. e. parker , and c.w .
oates , `` spin-1/2 optical lattice clock , '' 063001 ( 2009 ) . b. willke , k. danzmann , m. frede , p. king , d. kracht , p. kwee , o. puncken , r. l. savage , b. schulz , f. seifert , c. veltkamp , s. wagner , p. we@xmath19els , and l. winkelmann , `` stabilized lasers for advanced gravitational wave detectors , '' class .
. grav .
* 25 , * 114040 ( 2008 ) .
t. rosenband , d. b. hume , p. o. schmidt , c. w. chou , a. brusch , l. lorini , w. h. oskay , r. e. drullinger , t. m. fortier , j. e. stalnaker , s. a. diddams , w. c. swann , n. r. newbury , w. m. itano , d. j. wineland , and j. c. bergquist , `` frequency ratio of al@xmath4 and hg@xmath4 single - ion optical clocks ; metrology at the 17th decimal place , '' science * 319 , * 18081812 ( 2008 ) .
s. blatt , a. d. ludlow , g. k. campbell , j. w. thomsen , t. zelevinsky , m. m. boyd , j. ye , x. baillard , m. fouch , r. l. targat , a. brusch , p. lemonde , m. takamoto , f. l. hong , h. katori , and v. v. flambaum , `` new limits on coupling of fundamental constants to gravity using @xmath10sr optical lattice clocks , '' 140801 ( 2008 ) .
r. w. p. drever , j. l. hall , f. v. kowalski , j. hough , g. m. ford , a. j. munley , and h. ward , `` laser phase and frequency stabilization using an optical resonator , '' appl .
phys .
b * 31 , * 97105 ( 1983 ) .
m. notcutt , l. s. ma , j. ye , and j. hall , `` simple and compact 1-hz laser system via an improved mounting configuration of a reference cavity , '' 1815 ( 2005 ) .
l. s. chen , j. l. hall , j. ye , t. yang , e. j. zang , and t. c. li , `` vibration - induced elastic deformation of fabry - perot cavities , '' 053801 ( 2006 ) .
t. nazarova , f. riehle , and u. sterr , `` vibration - insensitive reference cavity for an ultra - narrow - linewidth laser , '' appl .
phys .
b * 83 , * 531 ( 2006 ) . s. a. webster , m. oxborrow , and p. gill , `` vibration insensitive optical cavity , '' 011801(r ) ( 2007 ) . j. alnis , a. matveev , n. kolachevsky , th .
udem , and t. w. hnsch , `` subhertz linewidth diode lasers by stabilization to vibrationally and thermally compensated ultralow - expansion glass fabry - prot cavities , '' 053809 ( 2008 ) .
j. millo , d. v. magalhes , c. mandache , y. le coq , e. m. l. english , p. g. westergaard , j. lodewyck , s. bize , p. lemonde , and g. santarelli , `` ultrastable lasers based on vibration insensitive cavities , '' 053829 ( 2009 ) .
y. n. zhao , j. zhang , a. stejskal , t. liu , v. elman , z. h. lu , and l. j.wang , `` a vibration - insensitive optical cavity and absolute determination of its ultrahigh stability , '' opt .
express * 17 , * 8970 ( 2009 ) .
y. n. zhao , j. zhang , j. stuhler , g. schuricht , f. lison , z. h. lu , and l. j.wang , `` sub - hertz frequency stabilization of a commercial diode laser , '' opt .
commun .
* 283 , * 4696 ( 2010 ) .
k. numata , a. kemery , and j. camp , `` thermal - noise limit in the frequency stabilization of lasers with rigid cavities , '' 250602 ( 2004 ) .
a. yu .
nevsky , m. eichenseer , j. von zanthier , and h. walther , `` a nd : yag laser with short - term frequency stability at the hertz - level , '' 91100 ( 2002 ) .
m. notcutt , l. s. ma , a. d. ludlow , s. m. foreman , j. ye , and j. l. hall , `` contribution of thermal noise to frequency stability of rigid optical cavity via hertz - linewidth lasers , '' 031804(r ) ( 2006 ) .
a. d. ludlow , x. huang , m. notcutt , t. zanon - willette , s. m. foreman , m. m. boyd , s. blatt , and j. ye , `` compact , thermal - noise - limited optical cavity for diode laser stabilization at @xmath20 , '' 641643 ( 2007 ) .
s. a. webster , m. oxborrow , s. pugla , j. millo , and p. gill , `` thermal - noise - limited optical cavity , '' 033847 ( 2008 ) .
d. guyomarch , g. hagel , c. zumsteg , and m. knoop , `` some aspects of simulation and realization of an optical reference cavity , '' 063802 ( 2009 ) .
r. w. fox , `` temperature analysis of low - expansion fabry - perot cavities , '' opt .
express * 17 , * 15023 ( 2009 ) .
t. legero , t. kessler , and u. sterr , `` tuning the thermal expansion properties of optical reference cavities with fused silica mirrors , '' j. opt .
soc . am .
b * 27 , * 914 ( 2010 ) .
y. y. jiang , a. d. ludlow , n. d. lemke , r. w. fox , j. a. sherman , l. -s .
ma , and c. w. oates , `` making optical atomic clocks more stable with @xmath21 level laser stabilization , '' nature photonics * 5 , * 158 ( 2011 ) .
ule product information sheet , corning advanced optics , 334 county rt .
, 16 , canton , new york 13617 ( 2006 ) . http://www.corning.com/docs/specialtymaterials/pisheets/ulebro91106.pdf .
s. webster and p. gill , `` low - thermal - noise optical cavity , '' proc .
2010 ieee int .
freq .
control symp .
, 470 ( 2010 ) .
k. numata , g. b. bianc , n. ohishi , a. sekiya , s. otsuka , k. kawabe , m. ando , and k. tsubono , `` measurement of the intrinsic mechanical loss of low - loss samples using a nodal support , '' phys .
lett .
a * 276 * , 37 ( 2000 ) .
t. kessler , c. hagemann , c. grebing , t. legero , u. sterr , f. riehle , m. j. martin , l. chen , and j. ye , `` a sub-40 mhz linewidth laser based on a silicon single - crystal optical cavity , '' arxiv:1112.3854v1 , ( 2011 ) . m. nagel and a. peters , `` towards an ultra - stable optical sapphire cavity system for testing lorentz invariance , '' arxiv:1112.3623v1 , ( 2011 ) . |
the opera observation @xcite of superluminal neutrinos has been ruled out by the measurement by icarus @xcite which puts an upper bound on the superluminality of neutrinos at @xmath3 .
an phenomenological consequence of neutrinos being superluminal was the observation by cohen and glashow ( gc)@xcite and others @xcite that superluminal neutrinos are kinematically allowed to emit pairs of @xmath4 and would thereby lose most of their energy during the cern - gran sasso flight of 730 km @xcite .
icarus experiment @xcite searched for the cerenkov emission of electrons in the same cern - cngs to gran sasso neutrino beams , where according to the gc calculation , @xmath5 of the neutrinos are expected to decay and no anomalous @xmath4 events were detected . using this a much stronger bound @xmath6 can be put on the superluminal neutrinos .
another problem with superluminal neutrinos which has been discussed @xcite is that the change in neutrino energy - momentum relation restricts the phase - space for the @xmath7 process and the pion lifetime would be larger for the opera neutrinos produced by the pion decay from the cern - cngs beam . in a measurement by minos @xcite it was found that muon neutrinos of average energy 3 gev traversing a distance 730 km exceed @xmath8 by an amount , @xmath9 .
this however is in contrast to the neutrino observations from supernova sn 1987a @xcite where over a flight path of @xmath10 kpc , the neutrinos with energy in the band @xmath11 mev all arrived within a time span of @xmath12 sec and the optical signal arrived after @xmath13 hours of the neutrino signal ( consistent with prediction of supernova models ) from which it is inferred that @xmath14 . in the present paper
we calculate the rates for @xmath4 radiation from neutrinos and pion decay assuming that the neutrinos obey the energy momentum relation of the form @xmath15 .
this is motivated by horava - lifshitz type field theories @xcite where the higher derivative terms break lorentz invariance at high scale but help in removing ultraviolet divergence .
we find that for @xmath1 the bounds on @xmath2 from direct time of flight measurement @xcite are much more stringent than from constraining the cerenkov decay process during the cern to gran - sasso flight of neutrinos @xcite .
this is unlike the case for @xmath16 studied by glashow - cohen @xcite and @xcite where the constraint on lorentz violation from the kinematically forbidden processes is more stringent than from the time of flight measurement .
models which explain the opera result of superluminal neutrinos and which have a bearing on the question of cerenkov emission from neutrinos or the pion decay kinematics fall broadly in the following categories : 1 .
deformed lorentz symmetry models @xcite where the dispersion relations change from the usual @xmath17 form to a different form which is still covariant under the modified lorentz transformations . in this picture the processes which are forbidden in one reference frame ( like the rest frame of the massive neutrinos ) will be forbidden also in the lab frame .
lorentz invariance violation as in lifshitz type field theories @xcite , from a gauge singlet susy sector @xcite or a hidden sector@xcite , environmental couplings @xcite , dynamical symmetry breaking @xcite , fermi - point splitting @xcite , space - time fluctuations @xcite and string theory @xcite . in this paper
we we consider the following general dispersion relation motivated by horava - lifshitz theories @xmath18 where @xmath19 etc .
the difference between the superluminal neutrino velocity and the speed of light ( taken to be 1 ) is then given by @xmath20 the icarus time of flight experiment @xcite has observed neutrinos and the time difference between the neutrino time of flight ( tof ) and the calculated photon tof is @xmath21 for neutrino energy @xmath22 gev and a distance @xmath23 from cern - cngs to the detector in gran sasso .
this corresponds to a neutrino superluminality by the amount @xcite , ( e=12.5gev)= < 2.3 10 ^ -7 [ d17 ] from ( [ delta ] ) and ( [ d17 ] ) we put constraints on @xmath2 from the tof experiment @xcite .
the cerenkov decay constraint comes from the earlier icarus experiment @xcite where the expected number of cc neutrino events is @xmath24 and the observed number is @xmath25 .
this corresponds to a bound on the decay length @xmath26 of neutrinos given by 0.04 < ( - c /731.2 km ) .
[ cerenkov ] we calculate the neutrino cerenkov decay length @xmath27 using the generalised dispersion relations ( [ disp ] ) and constraint @xmath2 from the experimental bound ( [ cerenkov ] ) .
the upper bounds on the lorentz violating parameter @xmath2 for different @xmath28 obtained from the tof experiment @xcite and cerenkov decay experiment @xcite are displayed in table [ xin ] . .
upper bounds on the lorentz violating parameter @xmath2 for different @xmath28 obtained from the tof experiment @xcite and neutrino cerenkov decay experiment @xcite [ cols="^,^,^ " , ] writing @xmath29 , where @xmath30 is the angle between @xmath31 and @xmath32 , and @xmath33 we see from the argument of the @xmath34-function in eq.([pi - rate ] ) @xmath35 while the derivative of the argument of @xmath34-function with respect to @xmath36 yields @xmath37 substituting this in eq.([pi - rate ] ) we get @xmath38\ ] ] the limits of the @xmath39 integral are fixed by taking @xmath40 in eq.([cos ] ) @xmath41 k_{min}=&\frac{m_\pi^2-m_\mu^2-\xi_nk_{min}^{n-1}(e_\pi - k_{min})}{2(e_\pi+q ) } \label{kmax}\end{aligned}\ ] ] we solve these polynomial equations for @xmath42 and @xmath43 numerically to obtain the kinematically allowed limits of neutrino momentum . using these limits to integrate over the neutrino momentum
@xmath39 we get the decay rate for pion .
the ratio of pion decay rate thus calculated to the standard model prediction _ 0()=(1-)^2 for different @xmath28 is shown in the table [ pitab ] .
the @xmath44 case has also been dealt with in @xcite and we are in broad agreement with their result . in fig .
[ pifig ] an approximate numerical calculation of the pion decay width is plotted as a function of pion momentum for different @xmath28 . ) for different @xmath28.,scaledwidth=80.0% ] the decay width for the 100 gev pions decreases by @xmath45 for @xmath46 and is smaller at higher @xmath28 .
we have computed neutrino processes assuming a power law correction to the neutrino energy - momentum relations .
we conclude that for steep power law ( @xmath47 ) dispersion relations the contraint from the time of flight experiments is more stringent than from the measuremnet of the cerenkov @xmath4 emission or from the change in the width of pion decay .
our calculation of neutrino cerenkov emission and pion decay width in lorentz violating theories can be applied for putting bounds on lorentz violating parameters from the analysis of high energy cosmic rays .
also , future experiments for measuring neutrino velocities performed at higher energies will put strong constraints on the higher derivative lorentz violation theories @xcite . d. m. mattingly , l. maccione , m. galaverni , s. liberati , g. sigl , `` possible cosmogenic neutrino constraints on planck - scale lorentz violation , '' jcap * 1002 * , 007 ( 2010 ) [ arxiv : [ hep - ph ] ] . j. m. carmona , j. l. cortes , `` constraints from neutrino decay on superluminal velocities , '' [ arxiv : [ hep - ph ] ]
. m. j. longo , `` tests of relativity from sn1987a , '' phys
. rev .
d * 36 * , 3276 ( 1987 ) .
d. fargion and d. darmiento , `` inconsistence of super - luminal opera neutrino speed with sn1987a neutrinos burst and with flavor neutrino mixing , '' arxiv : [ astro-ph.he ] .
p. horava , `` quantum gravity at a lifshitz point , '' phys .
rev . * d79 * , 084008 ( 2009 ) [ arxiv : [ hep - th ] ] . j. magueijo , `` neutrino oscillations and superluminal propagation , in opera or otherwise , '' arxiv : [ hep - ph ] . z. lingli and b. q. ma , `` neutrino speed anomaly as a signal of lorentz violation , '' arxiv : [ hep - ph ] .
z. chang , x. li and s. wang , `` opera superluminal neutrinos and kinematics in finsler spacetime , '' arxiv : [ hep - ph ] .
g. guo and x. -g .
he , `` dispersion relations explaining opera data from deformed lorentz transformation , '' arxiv : [ hep - ph ] .
j. alexandre , `` lifshitz - type quantum field theories in particle physics , '' int .
j. mod .
phys . * a26 * , 4523 ( 2011 ) [ arxiv : [ hep - ph ] ] .
g. dvali and a. vikman , `` price for environmental neutrino - superluminality , '' arxiv : [ hep - ph ] .
a. kehagias , `` relativistic superluminal neutrinos , '' arxiv : [ hep - ph ] . m. matone , `` neutrino speed and temperature , '' arxiv : [ hep - ph ] .
r. b. mann and u. sarkar , `` superluminal neutrinos at the opera ?
, '' arxiv : [ hep - ph ] . i. oda and h. taira , arxiv : [ hep - ph ] . f. r. klinkhamer , `` superluminal muon - neutrino velocity from a fermi - point - splitting model of lorentz violation , '' arxiv : [ hep - ph ] . f. r. klinkhamer and g. e. volovik , `` superluminal neutrino and spontaneous breaking of lorentz invariance , '' pisma zh . eksp .
fiz . * 94 * , 731 ( 2011 ) [ arxiv : [ hep - ph ] ] . s. i .
nojiri and s. d. odintsov , `` could the dynamical lorentz symmetry breaking induce the superluminal neutrinos ?
, '' eur .
j. c * 71 * , 1801 ( 2011 ) [ arxiv : [ hep - ph ] ] .
f. r. klinkhamer , arxiv : [ hep - ph ] .
g. amelino - camelia , j. r. ellis , n. e. mavromatos and d. v. nanopoulos , int .
j. mod .
a * 12 * , 607 ( 1997 ) [ arxiv : ] . t. li and d. v. nanopoulos , `` background dependent lorentz violation from string theory , '' arxiv : [ hep - ph ] . | we compute various processes involving neutrinos in the initial and/or final state and we assume that neutrinos have energy momentum relation with a general power law @xmath0 correction due to lorentz invariance violation . we find that for @xmath1 the bounds on @xmath2 from direct time of flight measurement are much more stringent than from constraining the neutrino cerenkov decay process .
2.5 cm + 0.5 cm _ physical research laboratory , ahmedabad 380009 , india _ + |
one of the longstanding and intriguing problems of hadron physics is the identification of the states that can be assigned as the first radial excitations of the nucleon and @xmath7 .
it is well recognized that the crucial role in the identification of the roper resonance n(1440)@xmath8 as a predominantly first radial excitation of the three - quark ( @xmath9 ) ground state belongs to the measurements by the clas collaboration @xcite that resulted in the determination of the electrocouplings of this resonance with the proton in a wide range of @xmath10gev@xmath4 .
comparison of the @xmath11n(1440)@xmath8 transition amplitudes extracted from these data @xcite with the predictions of the lf relativistic quark models ( lf rqm ) @xcite provided strong evidence for the n(1440)@xmath8 as a member of the multiplet @xmath12_r$ ] , with additional non-3-quark contributions needed to describe the low @xmath13 behavior of the amplitudes .
our goal in this paper is computation of the @xmath5 transition amplitudes in the lf rqm .
comparison of the results obtained in the quark model with the amplitudes that are expected to be extracted from experimental data will provide important test for the commonly expected asignment of the @xmath6 as the first radial excitation of the @xmath7 .
very recently , the clas data on the differential cross sections of exclusive process @xmath14 were reported in the range of @xmath15gev@xmath4 , and the invariant mass range of the @xmath16 final state @xmath17 gev @xcite .
these data combined with the earlier clas data @xcite on the cross sections and longitudinally polarized beam asymmetries for this reaction in the lower mass range @xmath18gev and at close values of @xmath13 allowed the extraction of the electroexcitation amplitudes of the resonances @xmath19 , @xmath20 , and @xmath21 in the third resonance region .
the isotopic pairs of the resonances from this region : @xmath6 and @xmath22 , @xmath23 and @xmath24 , and @xmath25 and @xmath26 , could not be separated from each other using data from a single isospin channel .
currently new data are in preparation by the clas collaboration for the @xmath27 process in the same kinematics region as the data in the @xmath28 channel @xcite , as well as at lower @xmath13 .
the two - channel analysis will allow the extraction of the electroexcitation amplitudes of all resonances from the third resonance region including the @xmath6 .
the approach we use is based on the lf dynamics and is formulated in refs .
@xcite . in numerous applications ( see refs . @xcite and references therein )
, this approach was utilized for the investigation of nucleon form factors and electroexcitation of nucleon resonances . in this work
we study the electroexcitation of the @xmath6 in parallel with that of the @xmath7 , where we complement the results obtained earlier in ref .
@xcite by computing all three form factors that describe the transition @xmath2 . in refs .
@xcite it was shown that there are difficulties in the utilization of the lf approaches for hadrons with spin @xmath29 . in the approach of ref .
@xcite , these difficulties limit the number of transition amplitudes that can be investigated for the @xmath7 and @xmath6 .
reliable results can be obtained only for two of the three transition form factors .
they are based on the utilization of longitudinal components of the electromagnetic current @xmath30 . for the @xmath7 ,
the results obtained for two transition form factors have been presented in ref . @xcite . in the present work ,
we complement these results by calculating the third transition form factor using @xmath31 . as was shown in ref .
@xcite , these results are less reliable , as the matrix elements of transverse components of the electromagnetic current can contain contributions that violate impulse approximation , i.e. contributions of diagrams containing vertices like @xmath32 .
similar problem exists in the lf rqm of refs .
@xcite , where the requirement of rotational covariance can not be satisfied without introducing two- and three - body current operators . for this reason ,
the results for the electroexcitation amplitudes for the resonances with spins @xmath33 are presented in ref .
@xcite along with curves which show the uncertainty that can be caused by the violation of the rotational covariance .
when presenting our results we also demonstrate the uncertainty that can arise due to the inclusion of the transverse components of the electromagnetic current .
an important aspect in the comparison of the transition amplitudes obtained in theoretical approaches with the amplitudes extracted from experimental data is their sign ( see , for example , ref .
the results on the @xmath34 transition amplitudes extracted from experimental data contain an additional sign related to the vertex of the resonance coupling to the final state hadrons .
in the electroproduction of pions on nucleons this is the relative sign between the @xmath35 and @xmath36 vertices .
for the roper resonance , this sign was found in refs . @xcite and @xcite using , respectively , the @xmath37 model and the approach based on pcac in the way suggested in ref .
the results obtained in both approaches are consistent with each other . in sec .
[ delta_amplitudes ] , we determine the relative signs of the vertices @xmath36 , @xmath38 , and @xmath39 using the approach based on pcac . our goals and the ranges of @xmath13 , where we make predictions , for the resonances @xmath7 and @xmath6 are different . for the @xmath6 , we make predictions that are of interest to reveal the nature of this resonance using the existing and future clas data at @xmath40gev@xmath4 . for the @xmath7 ,
our goal is to make predictions up to @xmath41gev@xmath4 .
these results will be important for the interpretation of future data on @xmath42 that are expected with the jefferson lab @xmath41gev upgrade . in sec .
[ delta_amplitudes ] we present the lf rqm formalism to compute the @xmath43 transition amplitudes .
the results for both resonances are presented and discussed in sec .
[ results ] and summarized in sec . [ summary ] .
the @xmath2 and @xmath5 amplitudes have been evaluated within the approach of ref .
@xcite where the lf rqm is formulated in the infinite momentum frame ( imf ) .
the imf is chosen in such a way , that the initial hadron moves along the @xmath44-axis with the momentum @xmath45 , the virtual photon momentum is @xmath46 , the final hadron momentum is @xmath47 , and @xmath48 ; @xmath49 and @xmath50 are masses of the nucleon and @xmath51 , respectively . in this frame , the matrix elements of the electromagnetic current for the @xmath43 transition have the form : @xmath52 where @xmath53 and @xmath54 are the projections of the hadron spins on the @xmath44-direction . in eq .
( [ eq : sec1 ] ) , it is supposed that the photon interacts with quark @xmath55 ( the quarks in hadrons are denoted by @xmath56 ) , @xmath57 is the charge of this quark in units of @xmath58 ( @xmath59 ) ; @xmath60 and @xmath61 are wave functions in the vertices @xmath62 ; @xmath63 and @xmath64 ( @xmath65 ) are the quark momenta in imf ; @xmath66 is the phase space volume ; @xmath67 corresponds to the vertex of the quark interaction with the photon : @xmath68 where @xmath69 is the fraction of the initial hadron momentum carried by the quark .
let @xmath70 be the three - momenta of initial quarks in their c.m.s .
: @xmath71 .
the sets of the quark three - momenta in the imf and in the c.m.s . of the quarks are related as follows : @xmath72 according to results of ref .
@xcite , the wave function @xmath60 is related to the wave function in the c.m.s . of quarks through melosh matrices
@xcite : @xmath73 here we have separated the flavor - spin - space ( @xmath74 ) and spatial ( @xmath75 ) parts of the c.m.s
. wave function .
the melosh matrices are @xmath76 where @xmath77 is the quark mass and @xmath78 is invariant mass of the system of initial quarks : @xmath79 in the c.m.s . of quarks : @xmath80 for the final state quarks , the quantities defined by eqs .
( [ eq : sec5]-[eq : sec9 ] ) are expressed through @xmath64 , @xmath81 , and @xmath82 . the phase space volume in eq .
( [ eq : sec1 ] ) has the form : @xmath83 to study sensitivity to the form of the quark wave function , we employed two forms of the spatial wave function : @xmath84 and that were used , respectively , in refs . @xcite and @xcite .
the parameters @xmath86 and @xmath87 are determined by the conditions : @xmath88 to distinguish between ground state @xmath89 and the @xmath90 , considered as the member of the multiplet @xmath12_r$ ] , we have used in eqs .
( [ eq : sec11]-[eq : sec15 ] ) notations @xmath51 and @xmath91 .
other parameters of the model , namely , the quark mass @xmath77 and the oscillator parameter @xmath92 , were found in ref .
@xcite from the description of nucleon form factors up to @xmath93gev@xmath4 : @xmath94 the @xmath13-dependence of the constituent quark mass ( [ eq : nuc2],[eq : nuc3 ] ) is in qualitative agreement with the qcd lattice calculations and dyson - schwinger equations @xcite , where the running quark mass is generated dynamically . the parameters ( [ eq : nuc1 ] ) and the parameterizations ( [ eq : nuc2],[eq : nuc3 ] ) have been used in the present calculations . for both resonances , the results for the transition amplitudes obtained with the wave functions ( [ eq : sec11],[eq : sec12 ] ) and ( [ eq : sec13],[eq : sec14 ] ) turned out very close to each other .
electroexcitation of the states with @xmath95 on the nucleon is described by three form factors @xmath96 , @xmath97 , and @xmath98 , which we define according to refs .
@xcite in the following way : @xmath99 where @xmath100 and @xmath101 are , respectively , the dirac and rarita - schwinger spinors .
these form factors have been found through the matrix elements ( [ eq : sec1 ] ) using the relations : @xmath102 , \label{eq : sec21 } \\ & & \frac{1}{2{\rm p}}<\delta,\frac{3}{2}|j_{em}^{0,z}|n,-\frac{1}{2 } > the relations between form factors @xmath96 , @xmath97 , and @xmath98 and the @xmath103 helicity amplitudes and the jones - scadron form factors @xmath104 , @xmath105 , and @xmath106 @xcite are given in the appendix . in the approach based on pcac , the relative signs of the @xmath36 , @xmath38 , and @xmath39 vertices
are determined according to refs .
@xcite by the relative signs of the following expressions : @xmath107 where @xmath108 denotes the states @xmath86 , @xmath89 , and @xmath90
. numerical calculation of @xmath109 , @xmath110 , and @xmath111 with the wave functions ( [ eq : sec11]-[eq : sec14 ] ) gives positive relative signs for the @xmath36 , @xmath38 , and @xmath39 vertices .
we present the results for the @xmath7 in terms of the @xmath112 magnetic - dipole transition form factor in the ash convention @xcite ( fig . [ gm ] ) and the ratios @xmath113 and @xmath114 ( fig .
[ ratios ] ) .
these observables are commonly used to present the results on the @xmath7 extracted from experimental data on the electroproduction of pions on nucleons . the @xmath112 magnetic - dipole form factor in the ash convention
is related to the jones - scadron form factor defined in the appendix as follows : @xmath115 the ratios @xmath0 and @xmath1 are related to the jones - scadron form factors by : @xmath116 where @xmath117 is the virtual photon 3-momentum in the c.m.s . of the reaction @xmath118 : @xmath119 as it was mentioned in the introduction , in the approach that we utilize @xcite , the results are reliable that are obtained through longitudinal components of the electromagnetic current @xmath30 , i.e. the results for the form factors @xmath120 and @xmath121 ( [ eq : sec21],[eq : sec22 ] ) .
these results have been presented and discussed in ref .
@xcite . in this paper , we complement the results for @xmath120 and @xmath121 by calculating the third transition form factor @xmath122 using @xmath31 ( [ eq : sec23 ] ) .
this allows us to present the predictions in a more convenient way in terms of @xmath123 and @xmath0 and @xmath1 . in order to demonstrate the uncertainty that can arise due to inclusion of the transverse components of the electromagnetic current , we also present in figs .
[ gm],[ratios ] results that correspond to the values of @xmath122 taken with @xmath124 deviation from the values obtained using the relation ( [ eq : sec23 ] ) .
it is known that at relatively small @xmath13 , nearly massless pions generate pion - loop contributions that significantly alter three - quark contribution to @xmath112 @xcite .
it is expected that the corresponding hadronic component , including contributions from other mesons , will be rapidly losing strength with increasing @xmath13 . from the description of the data on pion electroproduction on proton within dynamical reaction model @xcite
, it follows that the contribution associated with the meson - baryon contribution to @xmath112 ( dashed - dotted curve in fig .
[ gm ] ) can be neglected above @xmath125gev@xmath4 .
therefore , the weight of the @xmath9 contribution to the @xmath7 : @xmath126 was found in ref .
@xcite from the description of the form factors @xmath120 and @xmath121 at @xmath127gev@xmath4 : @xmath128 the uncertainty of @xmath129 is caused mainly by the systematic uncertainties of the data on @xmath130 at these @xmath13 . we have used the value of @xmath129 from eq .
( [ eq : sec28.1 ] ) to find the three - quark contributions to @xmath130 and @xmath0 and @xmath1 , that are presented in figs .
[ gm],[ratios ] . from the discussion above
, it follows that at @xmath131gev@xmath4 , meson - baryon contributions alter the three - quark contribution to @xmath112 @xcite . with this , for the magnetic - dipole form factor , these contributions definitely result in better agreement with experiment . above @xmath132gev@xmath4
, we expect that the @xmath112 transition will be determined by the three - quark contribution only .
therefore , we consider our results at these @xmath13 as predictions for the @xmath112 transition amplitudes obtained within nonperturbative approach .
for the form factor @xmath130 , the spread of our results caused by uncertainties in the form factor @xmath122 is insignificant , and we have definite predictions up to @xmath41gev@xmath4 . according to these predictions , above @xmath133gev@xmath4 the behaviour of the ratio @xmath134 becomes more flat in comparison with that at lower @xmath13 .
the similar @xmath13-dependence is observed for the proton magnetic form factor @xcite . for the jones - scadron magnetic - dipole form factor @xmath135 and the proton magnetic form factor @xmath136 ,
the @xmath13-dependences at @xmath137gev@xmath4 practically coincide . for the @xmath112 transition relative to @xmath138 : @xmath139 .
the full boxes are the clas data extracted in the analysis of ref .
@xcite , the open boxes correspond to the data from ref .
the bands show the model uncertainties of these data @xcite .
the thin solid curve is the result of the global analysis of the mainz group @xcite .
the results from other experiments are : open triangles @xcite , open cross @xcite , open rhombuses @xcite , and open circle @xcite .
the thick solid curve presents our results ; the dashed curves are our results corresponding to @xmath140 deviation of @xmath122 from the values obtained using the relation ( [ eq : sec23 ] ) ; the dotted curves show the uncertainty of our results ( given by the solid curve ) that is caused by the uncertainty of @xmath129 ( [ eq : sec28.1 ] ) .
the dashed - dotted curve is meson - baryon contribution from refs .
[ gm],width=287 ] for the ratios @xmath0 and @xmath1 , the spread of predictions grows from @xmath141 to @xmath142 , when @xmath13 is increasing from @xmath143 to @xmath41gev@xmath4 . nevertheless , for the ratio @xmath1 one can definitely conclude , that according to our predictions @xmath1 will continue to grow and within the @xmath144 gev@xmath4 limit
will not reach the value predicted in pqcd , i.e. @xmath145 with undefined sign and magnitude . on the other hand , in holographic qcd in the large @xmath146 limit
the @xmath1 ratio is predicted at the specific asymptotic value : @xmath147 @xcite .
the data show the correct trend , but are projected to reach only 40 to 50% of that value at @xmath148 gev@xmath4 . for the @xmath112 transition .
the legend for experimental data and thick solid and dashed curves is as for fig .
[ ratios],width=287 ] the results for the resonance @xmath6 are presented in fig .
[ hel_ampl ] in terms of the @xmath149 helicity amplitudes .
the predictions of the lf rqm approach from ref .
@xcite are also shown .
the common sign of the amplitudes has been found in our approach and in ref .
@xcite due to additional computation of the relative signs of the @xmath36 , @xmath38 , and @xmath39 vertices using different approaches . in section [ intro ]
we have discussed the difficulties in the utilization of the lf approach for hadrons with spin @xmath29 . in the approach of ref .
@xcite the uncertainty that can be caused by these difficulties for the @xmath6 nearly coincides with the longitudinal helicity amplitude @xmath150 . in our approach
these uncertainties are presented by dashed curves that correspond to the results obtained with the values of @xmath122 taken with @xmath124 deviation from the values obtained using the relation ( [ eq : sec23 ] ) . from the presented results we conclude that independently of uncertainties , both approaches give close predictions for the transverse helicity amplitudes : these amplitudes ,
being negative at @xmath151 , change their signs at @xmath152gev@xmath4 and become positive . with this , the values of transverse amplitudes at @xmath151 are in good agreement with the rpp estimates @xcite as in the case of the roper resonance , these features will be crucial for conclusions on the nature of the resonance @xmath6 that will be obtained from the comparison with the future data on the @xmath149 helicity amplitudes .
we have employed the lf rqm to evaluate the quark core contributions to the transitions @xmath2 and @xmath5 .
our previous evaluation of the 3-quark core contribution to the @xmath7 based on the @xmath2 data up to @xmath153 gev@xmath4 allowed us to make projections into unmeasured territory of @xmath148 gev@xmath4 .
this region may be covered in upcoming measurements with clas12 at the jefferson lab 12 gev upgrade .
the projections are made for the magnetic - dipole form factor and electric and scalar quadrupole ratios @xmath154 and @xmath155 .
predictions for the 3 electrocoupling amplitudes are also made for the @xmath6 in the range @xmath156
gev@xmath4 assuming this state is the first radial excitation of the @xmath7 . the predicted very rapid transition from large negative values at the real photon point to large positive values with maxima near @xmath157 gev@xmath4 , and a slow falloff with @xmath13 for the two transverse amplitudes , should be readily accessible to experimental exploration .
this work was supported by the us department of energy under contract de - ac05 - 06or23177 and the national science foundation and state committee of science of republic of armenia , grant-13 - 1c023 . | the magnetic - dipole form factor and the ratios @xmath0 and @xmath1 for the @xmath2 transition are predicted within light - front relativistic quark model up to photon virtuality @xmath3gev@xmath4 .
we also predict the helicity amplitudes of the @xmath5 transition assuming the @xmath6 is the first radial excitation of the ground state @xmath7 . |
chemistry on dust - grain surfaces is crucial to the chemical evolution and enrichment of the interstellar medium .
the simplest and most abundant interstellar molecule , h@xmath0 , has long been known to form exclusively on dust - grain surfaces ( gould & salpeter 1963 ) before rapidly returning to the gas phase , where it participates in ion - molecule reactions that may increase the chemical complexity of the gas phase ( dalgarno & black 1976 ) . however , in both quiescent clouds and in protostellar envelopes , the accretion onto the grains of gas - phase atoms and molecules also leads to the build - up of molecular ice mantles ( e.g. whittet et al . 1989 ; chiar et al .
1995 ; gibb et al .
2000 ; boogert et al .
the observed dust - grain ices are composed primarily of water ice ( h@xmath0o ) , as well as other simple hydrides such as methane ( ch@xmath3 ) and ammonia ( nh@xmath4 ) , and other organic species including carbon monoxide ( co ) and carbon dioxide ( co@xmath0 ) .
these ices are understood to form primarily at low dust temperatures ( @xmath510 20 k ) , with the hydrides forming mainly by the surface diffusion and repetitive addition of atomic hydrogen to other atoms ( e.g. hasegawa et al .
atomic h remains significantly mobile even at low temperatures ( e.g. manic et al .
2001 ) , which allows it to diffuse around the surface of the grain ( or the surface of the ice mantle ) , thermally hopping from one binding site ( i.e. potential well ) to another , until it meets a reactive species and forms a new molecule .
solid - phase water , thus formed , may account for the majority of the oxygen budget in many dense interstellar regions .
the porosity of both the dust - grain ices and of the grains themselves has recently been suggested as an important influence on grain - surface chemistry .
chemical models that treat the formation of h@xmath0 within pores in the grains ( perets & biham 2006 ) and the formation of other molecules in pores within the ices ( taquet et al .
2012 ) have demonstrated significant albeit moderate
effects in each case .
these models combine rate equation - based treatments of the chemical kinetics with generic assumptions regarding the characteristics of the pores .
experiments on ice mixtures ( collings et al .
2004 ; fayolle et al .
2011 ) have also shown that pores in laboratory water ice mixtures may absorb more volatile species , such as co and co@xmath0 , to be released at higher temperatures .
amorphous solid water ( asw ) ices formed in the laboratory by deposition of water molecules directly onto a cold substrate show a large degree of microporosity , defined as pores with diameter @xmath620 ( raut et al . 2007 and references therein ) ; however , it is unclear whether ices formed by such means are representative of the structure of interstellar dust - grain ices formed largely by surface - chemical processes over many thousands of years .
the large - scale morphology of interstellar dust grains is also likely to be important to the ice chemistry and structure .
interplanetary dust grains appear to be irregular in shape , and interstellar dust models imply `` fluffy '' morphologies ( mathis 1996 ) ; interstellar dust coagulation simulations also appear to confirm this view ( ossenkopf 1993 ; ormel et al
the chemical implications of such morphologies can not be investigated using current dust - grain chemical models , which do not specify actual grain shapes .
various models currently exist for the simulation of interstellar dust - grain surface chemistry , which may be divided into broad two groups .
the most common is that which uses generalized reaction rates for surface species , based on representative values for the binding energies and surface - diffusion barriers for each atom or molecule .
this chemical system may be simulated using either numerical solver routines to integrate ordinary differential equations ( e.g. hasegawa et al . 1992 ; garrod & herbst 2006 ; garrod 2013 ) , or using monte carlo ( mc ) techniques ( charnley 2001 ; vasyunin et al .
2009 ; vasyunin & herbst 2013 ) .
however , in either case , little structural information about the ice can be obtained , because the _ positions _ of atoms and molecules are not explicitly modeled , other than ( in some cases ) by depth .
furthermore , the effects of local binding properties of the surface ( caused by its shape or molecular composition ) can not be accounted for by this approach ; the models implicitly assume a perfectly smooth grain with uniform binding . unfortunately
, heterogeneous surface effects appear to be important , even in very simple systems such as that of h@xmath0 formation on the simplest non - smooth surface ( cuppen & garrod 2011 ) .
the other type of model uses a microscopically - exact kinetic monte carlo approach ( chang , cuppen & herbst 2005 ; cuppen & herbst 2007 ; cuppen et al .
2009 ) , which allows the behavior of each individual surface particle to be simulated .
these models begin with a square section of grain surface , with periodic boundary conditions to imitate the closed surface of a dust grain . using a
so - called _ on - lattice _
approach , atoms and molecules on the surface are allowed to diffuse between pre - defined lattice positions , with diffusion barriers and binding energies dependent on the pair - wise interations between species in adjacent binding sites .
every diffusion , desorption , accretion and reaction event is treated explicitly ; consequently , the method is more computationally demanding than those based on simple rate equations , and requires a simpler network of species and reactions . while the latter method produces valuable structural information , it retains several limitations .
the use of an on - lattice approach means that the surface species are not fully free to adapt their positions to local conditions , and the sizes of individual molecules are not taken into account in any way .
furthermore , the method does not treat chemistry on a three - dimensional grain surface , and so the effects of large - scale grain morphologies on the surface chemistry and ice structure can not be investigated . an _ off - lattice _
approach requires information concerning the possible diffusion pathways of a surface particle , as well as its final position , as these considerations are no longer guided by the fixed lattice .
off - lattice techniques have been used in other fields ( e.g. konwar et al .
2011 , and references therein ) , using pre - calculated , generic diffusion pathways .
such information may be obtained through intensive molecular dynamics calculations , and incorporated into kinetic monte carlo models of diffusion on regular surface structures . however , for an amorphous ice surface , composed of multiple different chemical species , the number of calculations and data to be stored using such a method would be extremely large .
the long timescales involved in interstellar chemistry would also likely make such intensive approaches prohibitively slow . here ,
a new off - lattice astrochemical kinetic monte carlo model of interstellar grain - surface chemistry is presented .
a simple method is devised for the _ ad hoc _ calculation of the post - hop positions of diffusing particles , under the assumption that thermal hopping occurs through easily - identified surface - potential saddle points .
particle positions are determined by local minima in the surface potentials felt by individual particles , as defined by the binding partners of each surface species . in this way
, the positions and chemical kinetics of every particle on a dust grain may be traced as atoms accrete from the gas phase and an ice mantle is formed .
the model allows chemistry to be simulated on a three - dimensional dust grain of arbitrary size and shape , whose surface is explicitly defined by the positions of its constituent atoms .
the morphology and porosity of the resultant ice structure is determined purely by the chemical kinetic processes occurring on the grain / ice surface .
results are presented for a preliminary version of the new model , as applied to a range of gas densities , using a simple chemical network and a small , idealized dust grain .
future work will consider the parameter space in greater depth .
section 2 details the new method ; the results are presented in section 3 .
a discussion of the model and its initial results are provided in section 4 .
the conclusions of this work are outlined in section 5 .
the model presented here is used to trace the microscopic chemical kinetics that occur on an interstellar dust - grain / ice surface as the result of the accretion and diffusion of atoms derived from the gas phase .
the model starts with a three - dimensional dust grain ( see fig .
1 ) , whose surface is explicitly defined by the positions of the constituent atoms .
the arrival of a particle from the gas phase results in its binding in a local potential minimum , whose position and strength are determined by the combined pair - wise interaction potentials between the particle and the nearest surface atoms .
the number of surface binding partners is dependent on the morphology of the surface , including the presence of other accreted surface particles .
the strength of the pair - wise interaction potential between specific chemical species is defined explicitly within the model ( see table 1 ) .
binding is treated as a `` physisorption '' or van der waals - type interaction . no _ chemical _ bonding to the grain surface itself
is considered .
in general , a surface particle may undergo thermal desorption into the gas phase , or may diffuse to a binding site ( i.e. surface potential minimum ) adjacent to its current location , via thermal hopping .
the interaction potentials felt by the particle determine the rates at which such processes may occur , and thus determine the competition with similar processes for other surface species or with further accretion from the gas phase .
details of each of these model elements are presented in the subsections below . in this preliminary model ,
all atoms and molecules including atoms that constitute the dust grain itself are assigned a uniform radius , with an assumption of spherical symmetry .
experimentally - determined charateristic radii for water molecules , which comprise the majority of the dust - grain surface ice in the models presented here , are dependent on the overall ice structure .
measured values for the intermolecular distance in amorphous solid water ( asw ) ice can range to as high as 3.3 (angell 2004 ) , while static , non - polarizable water models produce values around 3.16 (vega et al .
2009 ) , with radii centered on the oxygen atom .
an intermolecular spacing of @xmath7 is chosen for the current model , which for practical purposes defines the hard - sphere radius of a water molecule to be 1.6 ; this value is applied to all other chemical species .
canonical van der waals radii for oxygen and hydrogen atoms are 1.52 and 1.20 , respectively ( bondi 1964 ) .
the choice of a water - specific generic value ensures that the sizes of the dust - grain ice mantles produced by the models are appropriate .
future models will incorporate van der waals radii from the literature , allowing the ice structure to reflect more accurately the varying sizes of the constituent atoms and molecules .
the model adopts the numerical monte carlo ( mc ) approach suggested and proved by gillespie ( 1976 ) , and outlined below ; the general method has been successfully used in the past for astrochemical simulations by several authors ( e.g. charnley 2001 ; chang , cuppen & herbst 2005 , cuppen & herbst 2007 , vasyunin et al .
2009 ) . in common with the work of cuppen
, the new model takes a microscopic view of the grain - surface chemistry , but otherwise uses a new methodology for the treatment of the chemistry and the physical positions of atoms and molecules on the grain . gillespie ( 1976 ) treats the time evolution of a chemical kinetic system as a series of consecutive events
. the two key quantities to be determined by the probabilistic approach at any time , @xmath8 , are : ( i ) which event occurs next , and ( ii ) how much time elapses between this and the previous event . to determine this information , each possible chemical process , @xmath9 , is assigned a rate , @xmath10 ( in units of s@xmath11 ) . in the new model ,
such processes include an accretion event , a desorption event , or an individual thermal hop from one binding site to another ; in the case of desorption and diffusion , specific rates are calculated for each individual surface particle .
following gillespie , each process is then assigned a probability equal to its own rate divided by the sum of all rates , @xmath12 .
a computer - generated random number , @xmath13 , then chooses from these probabilties which process will occur next .
another random number , @xmath14 , determines the time elapsed since the last event , equal to @xmath15 .
the populations , positions , rates , or any other relevant information for each affected chemical species is then updated and the above steps are repeated until the user - defined end - time of the simulation is reached . in the model , this method is used to determine the overall sequence of thermal hopping , reaction , desorption or accretion of every surface atom / molecule , according to their individual rates .
the new model takes explicit account of the positions of particles on the grain surface , without using a pre - determined position lattice .
however , it _ does _ require the grain - surface structure to be defined initially .
the model will allow chemical simulations on any arbitrary grain size and structure . for this purpose ,
a simple code has been constructed to produce coordinates for every atom in an approximately spherical dust - grain of radius 5 atoms ; see fig .
1 . the grain may be simplistically considered to be composed of carbon atoms , although the purpose of the generated grain is to provide a simple surface structure upon which the new model may be initially tested .
atoms are positioned in a simple cubic structure , and assigned the representative radius of 1.6 ( which varies only slightly from the canonical van der waals radius for carbon atoms of 1.70 ; bondi 1964 ) .
the grain has a diameter of approximately 32 .
future applications of the generating code will be used to produce alternative / arbitrary grain sizes and morphologies , as well as more specific surface structures , compositions , and atomic sizes .
the total rate of accretion of gas - phase material onto the grain is dependent on the cross - section presented by the grain and its surface ice mantle . in simpler models ,
this rate is typically approximated by assuming a spherical grain , using a mean radius corresponding to either a bare grain ( hasegawa et al . 1992 ) or the combined grain and mantle ( acharyya et al . 2011 ) .
however , unlike most past models , the new method explicitly tracks the position of each surface species , and allows the non - uniform build - up of surface structures .
thus , the treatment of accretion adopted in this model must account for the _ trajectory _ of the incoming particle as well as the rate . in the new model , accretion from the gas phase
is handled in several stages ; firstly , a basic accretion rate is constructed for each gas - phase chemical species , based on the cross - section of the smallest sphere that can contain the entire grain and ice mantle . following hasegawa et al .
( 1992 ) , @xmath16 for the sphere bounding a single dust grain , where @xmath17 and @xmath18 are the thermal velocity and concentration of the accreting gas - phase species , @xmath9 , and @xmath19 , where @xmath20 is the radius of the sphere .
( because the bounding sphere is by definition larger than the combined mantle and dust grain , this basic rate is larger than the effective rate ultimately produced by the model , as explained below ) .
an independent accretion rate for each gas - phase species is constructed in this way . at such time that the monte carlo routine selects , according to the rate given by equation ( 1 ) , the accretion of a particular species to be the next process in the sequence , a randomized entry point into the bounding sphere is then generated in spherical coordinates : @xmath21 where @xmath22 and @xmath23 are random numbers between 0 and 1 .
equations ( 2 ) and ( 3 ) result in a uniform distribution of points over the surface of the bounding sphere .
expressions of the same form as equations ( 2 ) and ( 3 ) , using new random numbers , are then used to determine a trajectory for the incoming particle that is randomized over solid angle . however , only _
inward _ trajectories are considered ( outward trajectories are reversed to point inward ) , thus avoiding the double - counting of cases where a particle enters from an alternative point on the sphere .
the randomly - generated trajectory is traced through the interior of the sphere until a grain / ice particle is met by the incoming particle . at this stage ,
the accretion is judged to be successful , and another routine finds the nearest potential well to the contact point , assuming lennard - jones potentials for all interactions between the accreting particle and the grain / ice particles ( see section 2.3.2 ) .
if the trajectory of the incoming particle does not intersect with any atoms / molecules on the grain or surrounding ice mantle , then the accretion was unsuccessful .
in such a case , the clock is nevertheless advanced , and the monte carlo selection process continues as usual .
the entry of a particle into the bounding sphere is still a valid process , even if no accretion ultimately results .
the choice of the radius of the bounding sphere could be made arbitrarily large ( producing very low accretion efficiency for particles entering the bounding sphere ) , with no loss of accuracy to the model results .
the purpose of the adoption of the smallest bounding sphere possible is simply to avoid unnecessary and time - consuming calculations . in the results presented , only atomic h and o
are allowed to accrete from the gas phase , with an assumed sticking coefficient of unity .
fractional abudances of @xmath24 are adopted for each species .
the overall gas density is varied in the models , but the fractional abundances are assumed to remain constant over time , and no gas - phase chemistry is treated in the current model . each of these restrictions may be lifted in future work .
monte carlo treatments of _ gas - phase _ chemistry need consider only the evolution of , and rates associated with , the _ total _ population of each gas - phase chemical species . however , a microscopically - exact treatment of ice structure requires the positions and behavior of individual grain - surface atoms and molecules to be simulated explicitly . in this model ,
rates for diffusion and desorption are calculated for each individual grain - surface atom or molecule .
diffusion and desorption are treated as thermal processes , with individual rates of the form : @xmath25 where @xmath26 is a specific atom or molecule , @xmath27 is the characteristic vibrational frequency of the species in question , @xmath28 is the dust temperature , and @xmath29 is either the binding energy or diffusion barrier , depending on which process is being considered . in this paper , the terms `` thermal desorption '' and `` evaporation '' are considered synonymous . in practice ,
the monte carlo procedure initially selects a _ particle _ to undergo a thermal process , using the rate of the fastest thermally - activated process allowed for each surface particle in order to make the selection ( although these rates also compete with the accretion rates described above ) . from here , a random selection is made between desorption and each and every diffusion process that the chosen particle may undergo , weighted according to each individual rate .
this two - stage approach is necessary because , for example , the probability that a particle attains sufficient energy to diffuse also contains the small probability that it attains sufficient energy to desorb entirely ; equation ( 4 ) represents the rate at which a particle reaches an energy of @xmath30 _ or higher_. these processes are therefore in competition with one - another and must be treated as such . in the same way , for the case of a particle that has , for example , four available diffusion paths with equal energy barriers , the diffusion rate is not increased four - fold , but split four ways .
the atoms that constitute the dust - grain surface itself are not allowed to diffuse or desorb , but do contribute to the binding energies of surface - bound species .
the treatments applied to desorption and diffusion processes are described in more detail below .
when an atom has just accreted onto the grain surface , it initially resides in a potential well whose position and binding energy are calculated at the time of accretion .
( it should be noted that the position assigned to each particle represents only its mean position ; in reality , all particles would be vibrating in their potential wells ) . in the model ,
a particle is considered to be bound to a surface if it is bonded to 3 or more other atoms / molecules these could be atoms of the dust - grain itself , or other particles that are themselves bound to the grain surface .
two particles , a and b , are considered to be bonded to each other if their center - to - center separation , @xmath31 , is approximately equal to @xmath32 ( specifically , in the range @xmath33 ; closer bonding is explicitly prohibited in the model , in consideration of van der waals repulsion ) .
for each individual atom / molecule , the binding energy ( i.e. the energy required to desorb / evaporate into the gas phase ) , @xmath34 , is equal to the sum of the interactions , @xmath35 , with the other particles to which it is bound ( n.b . equation ( 5 ) , below , is _ not _ used for these calculations ) .
this value is used to evaluate the evaporation rate given by eq .
( 4 ) , and in the calculation of @xmath27 , following hasegawa et al .
any particles that are spatially `` boxed - in '' , under the hard - sphere approximation , are not allowed to evaporate or diffuse .
thus , molecules within bulk ice mantles are fixed in place unless the outer - surface molecules are removed .
interaction potentials are defined for the pairing of every type of chemical species in the model , as shown in table 1 .
potentials between any chemical species and a particle in the dust grain itself are chosen to reproduce approximately the binding energies used in previous , rate - based gas - grain chemical models , under the assumption that binding to the bare grain involves bonding with 4 dust - grain atoms . for example , the binding energy of atomic hydrogen assumed by garrod & herbst ( 2006 ) of 450 k may be compared with a value in this model of 400 k , which would be obtained for bonding to a flat face of the dust grain used in the model ( see figs . 1 & 2 ) .
bonding to only three grain atoms , or to as many as 5 , is possible in some positions on the bare grain surface , which would produce a binding energy of 300 k , or 500 k , respectively . for atomic oxygen ,
the 200 k interaction potential with a dust - grain atom adopted in this model reproduces the commonly - used binding energy of 800 k ( tielens & allamandola , 1987 ) , for most positions on the grain .
interaction potentials between particles that are not a part of the grain itself are nevertheless guided by the grain - binding values .
in general , the interaction potential of a pairing a b is assigned the lower of the values for a gr and b gr ( where `` gr '' represents a dust - grain atom ) .
an exception is permitted for species expected to demonstrate strong hydrogen - bonding properties .
thus , the interaction potentials involving h@xmath0o , h@xmath0o@xmath0 , and oh ( when bound to one of those two molecules ) are augmented by several hundred kelvin . for h@xmath0o h@xmath0o and h@xmath0o@xmath0h@xmath0o@xmath0 bonding , the a
gr values are doubled .
this produces a 4-partner binding energy for water of 4000 k , although in practice the model may yield water - ice that is sufficiently amorphous to produce many surface water molecules that are bound to 5 or 6 other waters .
the generic value adopted by garrod & herbst ( 2006 ) was 5700 k. the binding treatment assumed throughout the model may be characterized as a `` hard - sphere , nearest - neighbor '' approach .
the interaction potentials assumed in the model are isotropic ; polar molecules may in reality exhibit more directionality , but such considerations are currently outside the scope of this model . because the model uses an `` off - lattice '' approach , the positions of the adjacent potential wells into which a particle could diffuse are not initially specified , and must be calculated as needed from the current physical state of the system , as defined by the positions of all nearby particles .
the method set out for this model allows the energy barrier for each possible diffusion pathway of a given surface particle to be calculated at the same time as the total binding energy , prior to any calculation of a final _ post - diffusion _ position associated with it .
the putative final position resulting from a thermal hop need therefore only be calculated if that particular hop has actually been selected as the next process .
to determine diffusion barriers , the model requires information about the possible _ pathways _ that a diffusing particle may take from the current binding site to another . for the illustrative case of an h atom bound to a flat face of the grain ( see fig .
2 ) , the atom is bound to four surface - binding partners .
the atom may also be seen to have four possible diffusion pathways , corresponding to diffusion along the surface _ between pairs of binding partners_. while in theory there is an infinite number of paths that a diffusing atom may take , these four constitute those that pass through a saddle - point in the surface potential : they are , by far , the most probable paths . for each saddle - point pair
, a diffusion barrier is defined equal to the total binding energy , _ minus _ the interaction potentials of the diffusing particle with each of the pair of atoms through which the saddle - point would pass . in other words
, diffusion requires the breaking of two of the four bonds , while the other two remain intact , even after the hop is finished ( fig .
2 ) . in a different case where , for example
, the atom were bound to only 3 binding partners , only 1 bond would need to be broken . this approach is consistent with typical values of the ratio of diffusion barriers to binding energies , estimates of which can range from @xmath50.3 0.8 in current models ( see e.g. garrod & pauly 2011 ) .
such values may be achieved through surface binding to different numbers of binding partners ; the minimum of 3 partners ( of equal interaction potential ) would produce a value @xmath36 , while values closer to unity may be attained for binding to highly irregular ( rough ) surfaces .
the calculated diffusion barriers are stored and used to determine the selection of a particle and a thermal process as described in sections 2 and 2.3 .
if diffusion is selected , the next step is to determine the final position of the particle after diffusion through the selected saddle point . in order to determine this position ,
the diffusion process is treated as a _ rotation _ around the saddle - point pair ( with which the bonds remain intact following diffusion ) .
the particle is rotated at a fixed radius around the axis joining the saddle - point pair , until it is close enough to be considered `` bonded '' to the atoms on the other side of the saddle - point . at this stage ,
the position of the particle is optimized iteratively , to find the deepest part of the combined interaction potentials of the particles to which it is now bound .
each of these interaction potentials is treated as a lennard - jones 612 potential : @xmath37\ ] ] where @xmath38 is the current center - to - center separation , @xmath35 is the depth of the potential , and @xmath39 is the separation at which the potential minimum is reached , which is set to 3.2 for the present model .
( this expression is used wherever distance - dependent pair - wise potentials are required in the model ; for example , the binding positions of newly - accreted particles are also calculated iteratively using this expression . )
the derivatives of the potentials are used to calculate the net force vector resulting from the combined field at the current position .
the position of the particle is advanced incrementally according to the force vector , to yield the position of the potential minimum to a tolerance of 0.005 .
after the position is optimized in this way , the new binding energy is , again , taken simply as the sum of the strengths of the interaction potentials , @xmath35 , with all the species to which the particle is bound ( i.e. @xmath40 ) ; new diffusion barriers for all _ viable _ saddle - point pairs are also calculated at this time . the viability of saddle - point pairs is defined in the model as those pairs for which a rotation would be unhindered by other particles to which the diffusing species is _ already _ bound .
for example , in the case shown in fig .
2 , while there are four pairs that would allow rotation ( and thus diffusion ) , a pairing of _ diagonally - arranged _ particles from this group ( fig .
2c ) is disallowed , because rotation / diffusion around the axis between the diagonal particles would be obstructed in either direction by another binding partner .
all rotation - pairings of particles to which the species in question is bound ( i.e. @xmath41 pairings , for @xmath42 bonds ) are initially allowed .
each pairing is then assessed according to the following test : a plane is defined by the initial positions of the diffusing particle and each of the particles of the rotation pairing . if the _ other _ particles to which the surface species is bound are either _ all above _ or _ all below _ the plane , then diffusion is allowed ( in a direction _ away _ from those particles ) .
otherwise , diffusion is disallowed , as a collision would result ( as for the diagonal pairing shown in fig .
2c ) . typically , only pairings between adjacent ( although not necessarily _ bonded _ ) particles are allowed by this test , although irregular surface morphologies may allow other cases to be permitted . in this way , the possible diffusion paths on an arbitrary surface may be determined .
species for which all rotation pairs are forbidden are considered `` boxed in '' , and are thence also prohibited from desorption , until such time as this condition changes due to diffusion or evaporation of other particles .
it should be noted that , following accretion , desorption , or diffusion , the diffusion and desorption properties of all affected species such as binding partners both prior to and following a hop are re - calculated , and the rates and diffusion viabilities are re - assessed .
the rotation itself should not be considered an accurate description of the _ motion _ of the particle , but rather a convenient method for finding the potential minimum to which the particle hops , via a potential saddle - point .
actual diffusion paths for arbitrary binding structures would require intensive molecular dynamics or quantum chemical calculations . in the case given in the example ( fig .
2 ) , the diffusion barrier associated with each of the four possible hops is the same . however
, if the particle were bound to different chemical species , differences in the magnitudes of the interaction potentials would lead to differences in the probabilities of diffusion along each path . on an irregular surface
, the particle could be bound to 3 or more other particles ; the _ number _ of binding partners as well as their chemical species is therefore important to the energy barrier to diffusion ( i.e. the energy required to break all but two of the bonds ) .
the model is also optimized such that previous hops of the currently - active particle are stored in memory , so long as no other particles diffuse , desorb or accrete in the meantime .
this decreases run times by several orders of magnitude ; the calculations are often dominated by the diffusion of a single weakly - bound atom ( until it finds a stronger binding site ) . in the case where either a diffusion or accretion event leads to a particle being bonded to a species with which it is allowed to react
, reaction is assumed to occur immediately .
the reaction product inherits the position of the _ stationary _ reactant , but the other physical characteristics of the new particle are re - calculated .
both the so - called `` langmuir - hinshelwood '' ( diffusive ) and `` eley - rideal '' ( prompt ) surface - reaction mechanisms are treated automatically in this model .
the current network includes the species h , o ,
h@xmath0 , o@xmath0 , oh , h@xmath0o and h@xmath0o@xmath0 , commonly referred to as the `` water system '' .
no activation energy barrier - mediated reactions are currently considered in the network ; see table 2 .
many models include a number of other grain - surface processes not included here , such as photo - desorption and photo - dissociation .
the latter is often suggested as an important mechanism for the formation of complex organic molecules in the interstellar medium ( e.g. garrod & herbst 2006 ) .
all such processes are expected to be included in future versions of the model , but are omitted here so that the basic approach may be tested .
the inclusion in the model of photo - induced desorption based on estimated or measured rates ( berg et al .
2009 ) would be technically trivial .
photo - dissociation would be somewhat more complex , as it would require the positioning of two surface products that originated from a single molecule .
consideration of the formation of complex organic molecules using the new model is anticipated , but will depend on the adoption of a significantly larger reaction network that includes co and its related products .
furthermore , only _ surface _
processes are considered in the current model ; photo - processing of the sub - surface ice - mantle material would involve a considerable increase in technical complexity .
figure 1 shows a simulation of grain - surface chemistry using the method described above .
the simulation uses a gas density of @xmath43 @xmath2 , and dust and gas temperatures of 10 k. the images show accurately the positions of each simulated particle ; however , the size and appearance of each particle in the image is chosen purely for ease of visual identification . as described above , every atom and molecule in the present model
is treated as a uniform sphere of radius 1.6 ; no inference as to the polar orientation of molecules should be drawn .
the images are constructed using the free - ware ray - tracing software _ pov - ray_. panel ( b ) shows an instance wherein 2 oxygen atoms and 1 hydrogen atom have been accreted from the gas phase .
all are free to diffuse , but the hydrogen atom is significantly more mobile than the oxygen atoms . following several thermal hops , the hydrogen atom meets one of the oxygen atoms , with which it reacts to form oh . the further accretion of an oxygen and a hydrogen atom from the gas phase results in the formation of o@xmath0 and h@xmath0o .
a video of this sequence may be found in the online version of the journal .
figure 3 shows another simulation with a gas density of @xmath43 @xmath2 , extended for a much longer period ; 1000 water molecules are formed over 350 yr .
( an identical simulation is shown in figs.[f_surface]d and [ f_cross]d , extended to 200,000 water molecules , or 8,471 yr ) .
the resultant ice mantle is composed primarily of water molecules , as expected . some molecular oxygen ( o@xmath0 ) and molecular hydrogen ( h@xmath0 )
are also present ; panel ( b ) shows an h@xmath0 molecule a little below the center of the image . to the left of this
may be seen both an o@xmath0 molecule and a hydrogen peroxide molecule ( h@xmath0o@xmath0 ) . h@xmath0 and h@xmath0o are in fact formed in approximately equal quantities , but the molecular hydrogen evaporates rapidly unless trapped in a strong binding site , typically one with @xmath444 binding partners .
panel ( c ) shows that the h@xmath0 indicated in panel ( b ) has left that binding site . in rate equation - based models , the formation of any significant quantity of h@xmath0o@xmath0 by the direct addition of two oh
radicals at low temperature would be unlikely , due to the low diffusion rate of oh . such methods treat only the direct diffusion of each stated reactant .
however , because the model presented here considers the explicit positions of all particles at all times , h@xmath0o@xmath0 may instead be formed by the diffusion of an o atom to a position that is in contact with an oh radical , followed by the surface diffusion of an h atom onto the oxygen atom .
this results in an immediate two - stage reaction : h + o + oh @xmath45 oh + oh @xmath45 h@xmath0o@xmath0 .
the occurrence of this process is dependent only on the relatively fast diffusion of hydrogen and oxygen , and not oh . such effects , which are essentially three - body surface reactions , are not generally taken into account in rate - based models .
however , a similar process for the surface reaction of h + o + co was considered by garrod & pauly ( 2011 ) , who found this mechanism to be the primary formation route for co@xmath0 ice at low temperatures , due to the immobility of co and oh . because the new model takes explicit account of physical positions and structure , all such multi - stage processes are included by default . the formation of the ice mantle shown in fig .
3 is not uniform ; panels ( b ) and ( c ) in particular show the formation of a `` hole '' in the ice , corresponding to a region on the grain whose geometry results in weaker surface potential minima . in the center of the hole shown in panel ( c )
is a single dust - grain atom , around which six other dust - grain atoms reside .
this arrangement produces six contiguous potential minima that allow bonding to only 3 dust - grain atoms , rather than the typical 4 .
moderately mobile species that are momentarily bound in these weak sites rapidly diffuse out again .
thus , the geometry of the dust grain alone may be seen to influence the consequent ice - mantle structure .
there exist 8 such regions on the grain used in these models , where similar behavior is found . in this particular model , these holes eventually close over , as molecules build up at the edges of the holes , presenting positions of stronger binding within . under other conditions
, such holes may progress to form pores within the ice structure ( see section 3.1.1 ) . in order to investigate the formation of structure in the ices
, simulations have been run to produce ice mantles several times thicker than the diameter of the underlying grain .
four sets of models have been produced , corresponding to gas densities in the range @xmath46
@xmath47 @xmath2 . for each density
, the model has been run three times , adopting a different initial random number seed .
this is done to demonstrate that the behavior of the models is a general feature and not dependent on the specific random numbers used in each run .
all the simulations are run until 200,000 water molecules have been formed on the grain .
as before , dust and gas temperatures are held at 10 k and the gas - phase fractional abundances of h and o atoms are fixed at @xmath24 .
in addition , a single model has been run in which water molecules are accreted directly from the gas phase , with no accretion of h and o atoms .
an arbitrary gas density of @xmath47 @xmath2 is used ; however , for the consideration of ice structure , the precise rate of accretion is unimportant , due to the lack of surface chemistry and the neglible rate of water diffusion caused by the low temperature .
figure 4 shows the final state of each density model ( run 1 is shown in each case ) , including pure accretion ( panel a ) .
hydrogen molecules ( h@xmath0 ) are highlighted in blue to aid the eye .
a number of features are immediately apparent ; firstly , the large- and small - scale structure of the ice is strongly dependent on the gas density .
the most irregular structure results from the higher - density simulations , culminating in the accretion - only model in this case , water molecules remain wherever they land as the result of accretion . in the simulations with active chemistry , higher densities result in more rapid hydrogenation of surface oxygen atoms , before they have the opportunity to find alternative , stronger binding sites .
the higher - density ice mantles exhibit a highly porous , or `` creviced '' structure .
panels ( b ) and ( c ) in particular show a `` cauliflower - like '' structure , wherein the crevices act to isolate large - scale nodules of ice . in the lower density simulations ,
the nodules become broader and the crevices less deep .
the lowest - density model indeed shows a fairly smooth ice structure with no deep crevices or voids . in the accretion - only simulation , the ice appears irregular on all scales .
the small - scale porous structure is largely the result of the lack of surface diffusion , such that new species are unable to move even a short distance into the stronger binding sites that are provided by areas of extreme small - scale curvature of the ice surface .
however , the large - scale irregularity of the ice is caused by the combination of fully - randomized trajectories with a three - dimensional grain ; the growth of randomly - produced surface irregularities , or `` bulges '' , is amplified by their protrusion into the accretion field , picking up more material from a greater solid angle , while blocking out those trajectories for other regions of the grain / ice surface .
such effects are still prevalent in the active - chemistry models panel ( c ) shows an ice mantle whose underlying dust grain is centered in the image ; a large protrusion may be seen on the right .
while a moderate degree of surface diffusion may act to close over the small - scale pores or crevices in the ice , atoms of oxygen are required to diffuse a significant distance over the ice surface to counteract the large - scale irregularities .
consequently , while smooth on the small scale , the lowest - density simulation ( panel e ) nevertheless shows an overall irregular shape .
figure 5 shows cross sections of the same final ice mantles .
each cross section passes through the dust - grain center at an arbitrary angle , showing molecules that sit within a `` slice '' @xmath48 deep this allows space for up to two molecules along a line of sight .
videos of panels ( a ) , ( b ) and ( e ) are available in the online version of the journal , showing cross sections over 360 degrees .
the detail of the porosity within the ice mantles is clearly visible in figure 5 .
panel ( a ) shows significant and fairly uniform porosity , with pores or crevices that pass deep within the mantle almost down to the dust grain itself .
the highest - density simulation with active chemistry is not as extreme , but also shows very deep pores / crevices in places , and exhibits some closed pores .
the degree of porosity falls away at lower gas densities , until , for @xmath49 @xmath2 , there are essentially no open pores and the ice is extremely compact .
one of the most striking features of fig .
5 is the arrangement of hydrogen molecules ( h@xmath0 ) .
shown in blue in figs . 4 and 5 for ease of identification , h@xmath0 is almost completely segregated from the water molecules , especially in the higher - density simulations .
furthermore , the h@xmath0 molecules are arranged in `` veins '' that are similar in appearance to the empty porous structures . as may be seen three - dimensionally in fig .
4 , the h@xmath0 indeed fills in crevices between larger structures .
the veins of h@xmath0 are typically no more than around 5 molecules across ( commensurate micro - porous structure ) , while many of the unfilled pores are significantly larger .
following its formation on the grains , an h@xmath0 molecule may diffuse around the surface .
if it finds a weak binding site , it may evaporate entirely , but if it finds a strong binding site such as may be found in a pore , where multiple binding potentials converge then it may remain bound in place .
the addition of further hydrogen molecules could then contain it , and allow a pore to fill up with hydrogen .
the presence of larger , unfilled pores in the higher - density simulations suggests that these pores have an insufficient degree of inward curvature to produce binding potentials great enough to retain h@xmath0 for long enough before the deposition of new material can lock it into place .
figure [ bar_coord ] shows a bar chart of the number of surface bonds of each molecule ( of any species ) within the ice mantle , for each of the simulations shown in figs . 4 and 5 .
the accretion - only model shows a very different distribution from the active - chemistry models .
approximately 35% of all molecules in this simulation experience only 3 5 bonds , making them ( typically ) surface molecules .
conversely , only around 6% of molecules in even the highest - density active - chemistry simulation are so weakly bound , and the number of species with only 3 bonds is negligible .
however , this does not imply that potential minima that would allow only 3 bonds do not exist only that they are unoccupied .
pores , and the associated h@xmath0 veins , may begin to form in such positions , with ice structure building up around them .
the porosity visible in figs . 4 and 5
is reflected in the bonding distributions of fig .
6 , with the most compact ice ( produced by the lowest - density model ) showing the greatest bias toward 10 12 bonds , which are necessarily bulk ice molecules .
such distributions may therefore be useful as a measure of porosity in these simulations .
figure 7 shows the fraction of bonds shared with an h@xmath0 molecule , averaged over all h@xmath0 molecules .
the highest density model shows the greatest degree of clustering of h@xmath0 molecules .
based on a purely statistical comparison of abundances , the expected value would be around 13
14% for this density ( based on values in table 3 ) .
for the lowest density model , the statistical expectation would be @xmath58% . in each case , the actual fractional bonding of h@xmath0 to h@xmath0 is close to 3 times higher than the statistical expectation .
it should also be borne in mind that the binding potentials used for h@xmath0
( table 2 ) are the same for all binding partners .
the h@xmath0 clustering is thus the result of the structural arrangement of the ice mantle to produce pores or concentrations of strong binding sites in which h@xmath0 molecules may become trapped , rather than a preference for h@xmath0 molecules to bind to each other , per se . figure 8 shows time - dependent grain - surface chemical abundances for each of the stable molecules included in the model , for all four gas - density values .
results for all three runs at each density are plotted in the same panel .
the end - time of each model corresponds to the time at which 200,000 water molecules have been formed ; the precise value varies according to the gas density and , to some degree , according to random variation between runs at the same density . for o@xmath0 and h@xmath0o@xmath0 , which have generally low abundances
, there is some random fluctuation apparent at early times , and there are small systematic differences between runs of the same density .
however , as a proportion of the total quantities , these variations fall away as the ice mantles grow .
table 3 shows the final values for each model run , as well as the total amount of h@xmath0 formed , much of which evaporates back into the gas phase .
h@xmath0o@xmath0 shows the largest variation between same - density runs , of around 7 - 8% , but variations for the other species are typically 1 - 2%
. the elevated value for h@xmath0o@xmath0 may be due to its more intricate formation mechanism in this model ( see section 3 ) . for reference ,
table 3 also indicates end - times of the models , @xmath50 , and the greatest radial distance of any ice - mantle particle from the centroid of the grain , @xmath51 , achieved by the end of each run .
abundances of species other than h@xmath0o are seen to be somewhat lower for the lower density models , including h@xmath0 . however , the total amount of h@xmath0 formed ( see table 3 ) is much less affected than the quantity that is stored on the grain at each gas density .
accreted atomic hydrogen is typically able to remain on the grain for long enough for another reactive species to accrete and react with it , for all density models . in the case where the product is h@xmath0 ,
it is more likely to evaporate for the lower density models , which have a more regular surface that provides fewer strong - binding positions or pores . in each of the simulations ,
the amount of h@xmath0 formed is only a small amount larger than the total water formed .
the running of the new model over various density conditions allows an assessment of the computational efficiency to be made .
each simulation was run with an _ intel xeon x5482 _ cpu , using a single thread .
the run - time for all of the three higher - density simulations ( @xmath52 , @xmath53 , and @xmath47 @xmath2 ) is approximately 90 95 hours , depending on which random - number seed run is considered .
only the lowest - density simulations show an apparent density dependence , taking @xmath5180 hours to run .
the majority of cpu - time is used up on the calculation of diffusion pathways and on the assignment of position and pairing information associated with each surface hop . because the model keeps diffusion paths in memory , the most intensive calculations are typically done only once per particle , providing the positions of surface potential minima for an individual , mobile particle for large numbers of subsequent , identical hops , prior to reaction , desorption or trapping in a strong binding site .
it appears that the larger drain on cpu - time in the low - density case is caused by the relative smoothness of the ice surface . in this case
, a diffusing particle ( typically atomic hydrogen ) on the grain has a smaller probability of becoming trapped in a strong binding site ( thus ceasing its diffusion ) , while the probability of another particle being accreted onto the grain , and subsequently reacting with the first particle , is smaller ( due to lower gas density ) .
this results in a longer sequence of diffusion before a reaction occurs .
it is to be expected that the use of larger grains would ameliorate this effect , as the current model dust grain is extremely small , providing low overall rates of accretion .
conversely , higher temperatures would most likely make it worse , due to the resulting faster diffusion , although desorption rates would also increase .
the inclusion of a broader network of surface species is expected to make the surface more heterogeneous , such that the smoothness of the surface potential would be reduced , minimizing the possibility that surface species may spend long periods of cpu - time without being trapped in a strong binding site .
this model allows porous structures within the ice mantle to form naturally , as the result of chemical and physical conditions .
however , the consideration of porosity in simulations of interstellar dust - grain chemistry is not new ; perets & biham ( 2006 ) used a rate - based model to study the effects of porosity within the dust - grain structure itself .
the models indicated that porosity could increase the efficiency of h@xmath0 formation over a wider range of temperatures than would otherwise be possible .
taquet et al .
( 2012 ) extended the treatment of perets et al . to a full gas - grain chemical model that included the formation of a porous ice mantle , using pre - defined values for the number and sizes of pores .
they found only moderate increases in the production of certain key ice - mantle species .
cuppen & herbst ( 2007 ) used a monte carlo technique to treat the formation of ice on a surface with periodic boundary conditions , using a fixed lattice for the positions of particles on the grain or within the ice mantle . depending on the physical conditions ,
ice structures of varying compactness and porosity were produced , including some with tower - like structures of apparently single - molecule thickness .
the models that show the most extreme porosity , however , were obtained assuming temperatures rather higher than the 10 k used here . in the present model , only a limited reaction set
is considered , which does not include any activation barrier - mediated reactions .
nevertheless , the basic reactions that allow the formation of the most important species , such as h@xmath0 and h@xmath0o , are indeed present .
the use of an off - lattice approach in this model requires that the binding of atoms or molecules to the surface involve at least 3 other binding partners .
thus , the `` sky - scraper '' structures produced by the models of cuppen & herbst are not found here , both because binding to a single partner can not occur and because the direction of the binding is not pre - determined , but depends on the arrangement of the binding partners .
perhaps most importantly , the off - lattice approach of the present model allows the effects of the curvature of the surface to be automatically taken into account .
inward curvature , such as may be found within the pores , produces stronger binding conditions resulting from the ability to bind with a greater number of partners ; likewise , the outward curvature found on the bare grain , and on the `` nodules '' or `` bulges '' that form on the ice , tend to minimize the number of binding partners .
these curvature effects drive mobile particles into the pores as the pores are forming , so that they become filled with h@xmath0 , producing veins .
larger pores which are formed in the higher - density models , and which experience less extreme curvature tend not to fill up in this way
. the kind of micro - pores investigated by taquet et al .
( 2012 ) , which appear to be comparable in size to the h@xmath0 veins that form here , may not therefore be long - lived enough to affect the chemistry significantly .
the present model indicates that such pores would be filled up , although others may develop in their place .
in general , however , the low density model ( which perhaps provides the most appropriate comparison ) tends to predict a compact , albeit segregated , ice structure .
the gas - phase chemical abundances used in the model presented here represent dense - cloud conditions , under which most hydrogen takes the form of h@xmath0 , resulting in similar h and o gas - phase abundances
. however , higher values for h are entirely plausible , which would result in greater h fluxes onto the grains , and therefore a shorter time period before surface oxygen atoms could find alternative binding sites .
it is therefore likely that a cloud with large quantities of h not yet converted to h@xmath0 should produce more porous dust - grain ices .
it is also probable that , other model quantities being equal , the smallest of these pores would rapidly fill up .
the consideration here of a model that involves the direct accretion of water onto the grain allows a basic comparison with laboratory amorphous solid water ( asw ) ices , which are also formed by the deposition of water molecules directly onto a cold surface .
it is well known that the angle of deposition has a strong influence on the degree of porosity in the ice ( e.g. kimmel et al .
2001 ; raut et al .
the present model does not investigate the effect of deposition angle directly , but uses a completely randomized field of accreting particles .
the result is an extremely porous ice , which is qualitatively different from the forms produced by active grain - surface chemistry , because of the minimal diffusion of the water molecules following accretion . in the active - chemistry models
, the slow formation of the ice from its constituent atoms allows voids and strong binding sites to be filled more effectively , so that the lowest - density simulations show an absence of pores , and a relatively small quantity of h@xmath0 veins as compared with higher gas - density ( i.e. faster - accretion ) models .
it therefore seems appropriate to suggest that interstellar ice analogs formed in the laboratory may be significantly more porous than actual interstellar ices .
this would imply that effects observed in laboratory ices , relating to the absorption of other molecules into pores and their subsequent release at higher temperatures ( e.g. collings et al .
2004 ) , may be of less importance in interstellar ice mantles . in mitigation
, the present model does not include any mechanism for super - thermal surface diffusion of the accreted particle , or indeed of newly - formed surface molecules .
the acceleration of the accreting particle into a surface potential well , or the release of chemical energy from the surface formation of a molecule , could provide sufficient energy for diffusion .
this could allow an otherwise immobile particle to find a nearby binding site with stronger binding properties .
however , in view of the experimental evidence for significant porosities in laboratory asw ice ( e.g. westley et al .
1998 , and references therein ) , the importance of this effect may be small .
as shown in table 3 , for decreasing density the h@xmath0:h@xmath0o production ratio approaches 1:1 .
this may be explained through a simple analysis : taking the 1:1 gas - phase abundance ratio for h : o used in the present model , the ratio of accretion rates is 4:1 , due to the 16 times higher mass of oxygen , as entered into eq .
( 1 ) . under conditions where atomic - hydrogen mobility dominates all surface processes , and accretion dominates evaporation
, this rate should indeed produce a 1:1 formation rate for h@xmath0:h@xmath0o .
( for every 1 oxygen that accretes , 4 h - atoms will accrete ; 2 are used up , as they rapidly find the o / oh on the surface , to form water .
the other two will on average meet each other and form h@xmath0 . )
this analysis ignores o@xmath0 and h@xmath0o@xmath0 formation , each of which would otherwise remove 4 and 2 more hydrogen atoms , respectively , if their oxygen atoms had instead formed surface water .
this difference accounts for around 40% of the h@xmath0 excess over h@xmath0o in the low - density models .
it may be seen , therefore , that for the lowest density simulations in particular , the production of h@xmath0 is very close to what may be called a `` smooth - grain '' limit , in which hydrogen - atom mobility is not significantly affected by the presence of strong binding sites . with increasing density ,
there are more strong binding sites ( due to the more irregular ice structure ) , while the waiting period for the accretion of a new particle is also shorter .
these effects heighten the probability that an accreted hydrogen atom will wait for another h to react with it , rather than to diffuse to find a free oxygen atom elsewhere on the dust / ice surface , thus producing a greater proportion of molecular hydrogen . for gas densities
less than @xmath5@xmath46 @xmath2 , one should expect that the ice surface would be similarly smooth , and perhaps yet more spherical , while the waiting period for accretion would be greater , all of which should produce h@xmath0:h@xmath0o ratios even closer to unity . under such circumstances ,
the use of rate - based models that employ generalized binding and diffusion characteristics should not produce strongly divergent results for h@xmath0 production from those obtained here , providing appropriate characteristic values were chosen .
such an approach would also require that the stochastic behavior of the chemistry be treated adequately ( e.g. the use of the modified rate equations of garrod 2008 and garrod et al .
2009 ) , as the scenario described clearly falls into the case where standard rate equations should fail .
it is unclear , however , whether such an approach could accurately reproduce the quantity of the resultant h@xmath0 that is retained on the grain surface , which is still determined by heterogeneous structural considerations .
the choice of physical conditions investigated in the present study has been limited to the gas density , which controls the rate of accretion onto the grain .
the gas density demonstrates a clear effect on the resultant ice structure , and thence on the retention of h@xmath0 molecules on the grain .
the relative formation rates of o@xmath0 and h@xmath0o@xmath0 are also seen to be affected by the density variation , due to the competition between chemical reactions whose rates are dependent on surface diffusion .
each of the effects mentioned above concern the interplay between accretion , diffusion and desorption .
the gas density is thus not the only parameter which may affect the chemistry and structure of the ice formed on the grain .
future work will address the importance of dust temperature ; however , one may predict that temperatures higher than the 10 k used here would produce more diffusion of oxygen atoms , meaning more compact ices .
likewise , lower temperatures would result in more porosity , and greater retention of h@xmath0 within those structures .
the present , preliminary model uses a very small grain , of radius 16 , as compared to a canonical value of 0.1 @xmath54 m ( = 1000 ) .
however , the final ice - mantle radii are significantly larger , at around 150 190 .
future models will investigate chemistry on larger grains .
the application of the model to a porous dust grain , of explicitly defined structure , would also be quite possible .
the model presented in this paper is the first off - lattice monte carlo kinetic model of interstellar dust - grain chemistry .
the model allows the full three - dimensional simulation of chemical kinetics and ice structure on a grain surface defined by the positions of its constituent atoms .
the use of an off - lattice technique allows the precise positions , binding strengths and diffusion probabilities of grain - surface particles to be determined according to the interaction potentials with binding partners ; particles are not fixed in a pre - determined lattice structure .
the model also allows arbitrary morphologies and structures to be used for the underlying dust grain .
this includes not only the local surface roughness of a particle , but the degree of grain porosity , as well as the choice of a spherical , spheroidal , irregular , or any other grain shape .
the model opens up a large new parameter space that will be investigated in future , in conjunction with a more extensive chemical network . 1 .
the shape of the ice mantle is irregular , even for low gas - density models .
the porosity of the ice is strongly dependent on the gas density , with higher densities producing greater porosity .
3 . the small - scale curvature of local regions of the ice / grain surface has a strong effect on the chemistry and structure of the ice formed on top .
such considerations can only be treated using an off - lattice approach .
4 . inward
curvature on small scales as found with micropores allows binding with a greater number of binding partners than on a relatively flat surface .
this increases desorption and diffusion barriers , encouraging the trapping of mobile species in micropores and strong surface binding sites .
micropores are formed in the ices , but are rapidly filled with h@xmath0 molecules , forming veins of h@xmath0 .
thus , h@xmath0 shows significant segregation from water molecules in all models .
larger pores remain unfilled , as they do not provide sufficient small - scale curvature of the surface to significantly increase binding potentials .
7 . for gas densities appropriate to dark interstellar clouds ( @xmath46 @xmath2 ) ,
the ice mantle is fairly smooth and non - porous , albeit with veins of h@xmath0 throughout .
. direct deposition of water molecules , such as is used to produce laboratory amorphous water ice , results in far greater porosity than is achieved when the ice is formed by surface reactions between accreted atoms .
, then another hydrogen atom has landed , to react with oh , forming h@xmath0o . _
a video of this sequence is availble in the online version of the journal , showing one thermal hop per frame._,title="fig:",scaledwidth=49.0% ] , then another hydrogen atom has landed , to react with oh , forming h@xmath0o . _
a video of this sequence is availble in the online version of the journal , showing one thermal hop per frame._,title="fig:",scaledwidth=49.0% ] , then another hydrogen atom has landed , to react with oh , forming h@xmath0o . _
a video of this sequence is availble in the online version of the journal , showing one thermal hop per frame._,title="fig:",scaledwidth=49.0% ] , then another hydrogen atom has landed , to react with oh , forming h@xmath0o . _
a video of this sequence is availble in the online version of the journal , showing one thermal hop per frame._,title="fig:",scaledwidth=49.0% ] @xmath2 .
the `` hole '' in the ice structure is caused by the comparatively weak surface potential of the underlying grain surface in that region , resulting from its geometry . _
a video of this simulation is availble in the online version of the journal , showing the formation of 1 water molecule per frame._,title="fig:",scaledwidth=32.0% ] @xmath2 .
the `` hole '' in the ice structure is caused by the comparatively weak surface potential of the underlying grain surface in that region , resulting from its geometry .
_ a video of this simulation is availble in the online version of the journal , showing the formation of 1 water molecule per frame._,title="fig:",scaledwidth=32.0% ] @xmath2 .
the `` hole '' in the ice structure is caused by the comparatively weak surface potential of the underlying grain surface in that region , resulting from its geometry .
_ a video of this simulation is availble in the online version of the journal , showing the formation of 1 water molecule per frame._,title="fig:",scaledwidth=32.0% ] lrrrrrrrr h & 100 & 100 & 50 & 100 & 100 & 100 & 100 & 100 + h@xmath0 & 50 & 50 & 50 & 50 & 50 & 50 & 50 & 50 + o & 200 & 100 & 50 & 200 & 200 & 200 & 200 & 200 + o@xmath0 & 300 & 100 & 50 & 200 & 300 & 300 & 300 & 300 + oh & 400 & 100 & 50 & 200 & 300 & 400 & 500 & 600 + h@xmath0o & 500 & 100 & 50 & 200 & 300 & 500 & 1000 & 1000 + h@xmath0o@xmath0 & 600 & 100 & 50 & 200 & 300 & 600 & 1000 & 1200 + rcccccccccccc @xmath46 & 1 & & @xmath55 & 147.1 & & 3 & 1 & 3 & 1,238 & 240 & 18,284 & 207,313 + & 2 & & @xmath56 & 152.6 & & 6 & 3 & 3 & 1,241 & 261 & 18,399 & 206,096 + & 3 & & @xmath57 & 149.1 & & 1 & 4 & 0 & 1,260 & 257 & 18,407 & 208,248 + + @xmath52 & 1 & & @xmath58 & 153.9 & & 6 & 7 & 8 & 1,599 & 304 & 23,011 & 209,973 + & 2 & & @xmath59 & 148.3 & & 9 & 4 & 5 & 1,547 & 301 & 22,998 & 206,612 + & 3 & & @xmath60 & 160.0 & & 7 & 6 & 8 & 1,523 & 325 & 23,494 & 208,044 + + @xmath53 & 1 & & @xmath61 & 164.9 & & 37 & 30 & 19 & 3,073 & 620 & 28,872 & 217,227 + & 2 & & @xmath62 & 176.0 & & 27 & 23 & 19 & 2,974 & 618 & 29,216 & 217,463 + & 3 & & @xmath63 & 178.3 & & 27 & 20 & 18 & 3,031 & 596 & 29,620 & 216,183 + + @xmath47 & 1 & & @xmath64 & 191.9 & & 72 & 42 & 52 & 5,308 & 1,321 & 32,824 & 225,258 + & 2 & & @xmath65 & 177.3 & & 60 & 53 & 45 & 5,385 & 1,233 & 31,484 & 228,142 + & 3 & & @xmath66 & 183.6 & & 81 & 42 & 42 & 5,381 & 1,281 & 33,319 & 227,244 + gibb , e. l. , whittet , d. c. b. , schutte , w. a. , boogert , a. c. a. , chiar , j. e. , ehrenfreund , p. , gerakines , p. a. , keane , j. v. , tielens , a. g. g. m. , van dishoeck , e. f. , & kerkhof , o. 2000 , , 536 , 347 | the first off - lattice monte carlo kinetics model of interstellar dust - grain surface chemistry is presented .
the positions of all surface particles are determined explicitly , according to the local potential minima resulting from the pair - wise interactions of contiguous atoms and molecules , rather than by a pre - defined lattice structure .
the model is capable of simulating chemical kinetics on any arbitrary dust - grain morphology , as determined by the user - defined positions of each individual dust - grain atom .
a simple method is devised for the determination of the most likely diffusion pathways and their associated energy barriers for surface species .
the model is applied to a small , idealized dust grain , adopting various gas densities and using a small chemical network .
hydrogen and oxygen atoms accrete onto the grain , to produce h@xmath0o , h@xmath0 , o@xmath0 and h@xmath0o@xmath0 .
the off - lattice method allows the ice structure to evolve freely ; ice mantle porosity is found to be dependent on the gas density , which controls the accretion rate .
a gas density of @xmath1 @xmath2 , appropriate to dark interstellar clouds , is found to produce a fairly smooth and non - porous ice mantle . at all densities , h@xmath0 molecules formed on the grains
collect within the crevices that divide nodules of ice , and within micropores ( whose extreme inward curvature produces strong local potential minima ) .
the larger pores produced in the high - density models are not typically filled with h@xmath0 .
direct deposition of water molecules onto the grain indicates that amorphous ices formed in this way may be significantly more porous than interstellar ices that are formed by surface chemistry . |
how much does knowledge of a system s thermal equilibrium fluctuations at one state point tell us about its behavior at other state points ? in principle , complete knowledge of the fluctuations provides enough information to determine the partition function and thus the density of states , from which the free energy at other state points may be calculated . in practice , only second order moments of the fluctuations may be determined reliably .
these generally give little knowledge of the system s properties away from the state point in question .
it was recently shown that van der waals type liquids are `` strongly correlating '' in the sense defined below @xcite ; this implies that such liquids have a `` hidden '' approximate scale invariance @xcite . because of this , important global information about the system may be obtained from knowledge of the virial and potential energy second - order moments at one state point .
this unusual situation in statistical mechanics is the focus of the present note . in order to briefly recapitulate the definition of the class of strongly correlating liquids ,
recall that the pressure @xmath0 is a sum of the ideal gas term @xmath1 and a term reflecting the interactions , @xmath2 , where @xmath3 is the so - called virial : @xmath4 this equation is usually thought of as describing thermodynamic averages , but it also describes the instantaneous fluctuations .
the instantaneous ideal - gas pressure term is a function of the particle momenta .
the instantaneous virial @xmath3 is a function of the particle positions , @xmath5 . in the same way , of course
, the energy is a sum of the kinetic energy and the potential energy @xmath6 . if @xmath7 is the ( instantaneous ) potential energy minus its thermodynamic average and @xmath8 the same for the virial , at any given state point the @xmath9 correlation coefficient @xmath10 is defined by ( where sharp brackets denote equilibrium nvt ensemble averages ) @xmath11 we define `` strongly correlating liquids '' by requiring @xmath12 @xcite .
strongly correlating liquids include @xcite the standard lennard - jones ( lj ) liquid ( and the classical lj crystal ) , the kob - andersen binary lj liquid as well as other binary lj type mixtures , a dumbbell - type liquid of two different lj spheres with fixed bond length , a system with exponential repulsion , a seven - site united - atom toluene model , the lewis - wahnstrm otp model , and an attractive square - well binary model .
liquids that are not strongly correlating include water and methanol @xcite . the physical understanding developed in refs .
@xcite is that strong @xmath9 correlations are a property of van der waals liquids and some or all metallic liquids .
liquids with directional bonding like covalent and hydrogen - bonding liquids do not have strong @xmath9 correlations .
likewise , ionic liquids are not expected to be strongly correlating because of the different distance dependence of their short - range repulsions and the coulomb interactions competing interactions spoil the correlations . strongly correlating liquids appear to have simpler physics than liquids in general , an observation that has particular significance for the highly viscous phase @xcite . it has been shown that supercritical ( experimental ) argon is strongly correlating @xcite , that strongly correlating viscous liquids have all eight frequency - dependent thermoviscoelastic response functions @xcite given in terms of just one @xcite ( are `` single - parameter liquids , '' i.e. , have dynamic prigogine - defay ratio close to unity @xcite ) , that strongly correlating viscous liquids obey density scaling , i.e. , that their relaxation time @xmath13 depends on density @xmath14 and temperature as @xmath15 @xcite , and that even complex systems like a biomembrane may exhibit significant @xmath9 correlations for their slow degrees of freedom @xcite .
the property of strong virial / potential energy correlation is maintained even out of equilibrium and during crystallization @xcite .
whenever equilibrium fluctuations of the virial are plotted versus those of the potential energy for a strongly correlating liquid , an elongated ellipse appears @xcite .
the `` slope '' @xmath16 of this ellipse in the @xmath9 diagram is given by ( nvt averages ) @xmath17 this quantity , which is weakly state - point dependent , is the number entering into the density scaling relation @xmath15 @xcite .
thus for strongly correlating liquids knowledge of equilibrium fluctuations at one state point provides a prediction about how the relaxation time varies with density and temperature . what causes strong @xmath9 correlations ?
a hint comes from the well - known fact that an inverse power - law pair potential , @xmath18 where @xmath19 is the distance between two particles @xcite , implies perfect correlation @xcite . in this case
@xmath20 . in simulations of the standard lj liquid we found @xmath21 which corresponds to @xmath22 @xcite .
although this may seem puzzling at first sight given the expression defining the lj potential , @xmath23 $ ] , if one fits the repulsive part of the lj potential by an inverse power law , an exponent @xmath22 is required @xcite .
this is because the attractive @xmath24 term makes the repulsion steeper than the bare repulsive @xmath25 term would imply .
reference @xcite gave a thorough discussion of the correlations with a focus on the standard single - component lj liquid , including also a treatment of the classical crystal where @xmath26 at low temperature . according to ref .
@xcite the @xmath19-dependent effective exponent @xmath27 which controls the correlation is not simply that coming from fitting the repulsive part of the potential , but rather @xmath28 .
this number is approximately 18 around the lj minimum .
in fact , the lj potential may here be fitted very well with an `` extended '' inverse power - law potential @xcite , @xmath29 with @xmath22 ; for this particular potential of course @xmath30 . at constant volume the linear term contributes little to the virial and potential - energy fluctuations : when one nearest - neighbor interatomic distance increases , another decreases in such a way that the sum is almost constant . thus almost correct
boltzmann probability factors are arrived at by using the inverse power - law approximation , implying that thermal fluctations at one state point are well described by this approximation .
the equation of state , however , is poorly represented by the inverse power - law ( ipl ) approximation @xcite , and for instance the ipl relation @xmath31 does not apply for liquids that are not 100% correlating . inspired by the results summarized above , in this note we demonstrate a number of results based on one single assumption , the existence of curves in the phase diagram of a strongly correlating liquid on which the state points have configurations that are statistically equivalent in the canonical ensemble .
such curves will be referred to as _ isomorphs_.
two state points ( 1 ) and ( 2 ) with temperatures @xmath32 and @xmath33 and densities @xmath34 and @xmath35 , respectively , are said to be _ isomorphic _ if they obey the following : any two configurations of state points ( 1 ) and ( 2 ) , @xmath36 and @xmath37 , which may be scaled into one another , i.e. obey @xmath38\,=\,\rho_2^{1/3}[\mathbf{r}_i^{(2)}-\mathbf{r}_j^{(2)}]\,\,(i , j=1, ... n)\,,\ ] ] have proportional boltzmann statistical weights : @xmath39 here @xmath40 is the potential energy function and @xmath41 is a constant , which in general depends on both state points . an ipl liquid obeys eq .
( [ eq : rel2 ] ) with @xmath42 for states that obey @xmath43 .
no systems except ipl liquids obey eq .
( [ eq : rel2 ] ) rigorously . nevertheless , based on computer simulations we show below that the existence of isomorphs is a good approximation , thus providing new insight into the physics of strongly correlating liquids . before presenting the simulation results we briefly summarize some predicted properties of isomorphs , leaving their more detailed justification to a future publication : 1 . _
two isomorphic states have the same excess entropy : _
the excess entropy is by definition the entropy in excess to that of an ideal gas at same volume and temperature .
this quantity is always negative .
the excess entropy equals @xmath45 where @xmath46 is the normalized canonical configuration space probability .
equation ( [ eq : rel2 ] ) implies that isomorphic points have the same normalized canonical configuration space probability , thus the same excess entropy . in other words
, isomorphs are configurational adiabats . in particular
, the specific heat along an isomorph equals the ideal gas specific heat .
isomorphic states have the same average reduced relaxation time and the same reduced diffusion constant ( i.e. , they lie on the same``isochrone'')_. this prediction applies whether the dynamics is brownian or given by newton s equations .
it follows from the fact that the potential energy landscapes of two isomorphic state points are identical , except for an overall scaling that is compensated by the adjusted temperature and an additive constant .
isomorphic states have the same ( reduced ) dynamics , e.g. , the same relaxation functions .
_ this follows from the fact that their scaled potential - energy landscapes are identical .
property 3 implies property 2 , of course .
no aging is associated with jumps between two isomorphic state points . _
if density and temperature are changed instantaneously from one state point to values characterizing a new state point which is isomorphic with the initial state , the system will instantaneously be in equilibrium .
this is because all normalized boltzmann probability factors are identical for the two systems .
thus isomorphs are predicted to be `` wormholes '' between which one can jump instantaneously from equilibrium to equilibrium , even when the states themselves are characterized by long relaxation times .
_ isomorphic states have the same scaled radial distribution functions . _
this follows from the fact that their normalized boltzmann probability factors are identical .
we emphasize once again that no systems except ipl liquids have state points that are exactly isomorphic .
thus whenever state points are referred to as as isomorphic , more correctly one should say that they are _ approximately _ isomorphic .
note that properties 2 and 3 are consistent with the intriguing discovery by ngai , casalini , capaccioli , paluch , and roland @xcite that for varying pressure and temperature states with the same relaxation time have the same shape of the alpha relaxation dielectric loss peak ( this should only to apply for strongly correlating liquids , for instance not for hydrogen - bonding liquids ) .
the concept of isomorphs may shed light on the excess - entropy scaling proposed long ago by rosenfeld , according to which the dimensionless transport coefficients ( diffusion constant , viscosity , thermal conductivity , etc ) of dense fluids are all functions of the configurational entropy @xcite . for any strongly correlating liquids
this property follows from eq .
( [ eq : rel2 ] ) , but we can not explain rosenfeld scaling of dense fluids generally ( see also ref .
@xcite ) . recall that density scaling is the fact that the reduced relaxation time is a function of @xmath47 @xcite .
the exponent @xmath16 entering into this expression is given by eq .
( [ gamma ] ) @xcite .
this makes it easy to identify approximate isomorphs in computer simulations of strongly correlating liquids : in the simulations presented below two state points are regarded as isomorphic in the simulations below if @xmath48 .
the well - known kob - andersen 80:20 binary lennard - jones liquid ( ka ) @xcite is a nice model system that is easily supercooled .
this system is a strongly correlating liquid @xcite .
the slope @xmath16 varies slightly with state point , but at low and moderate pressures and temperatures this quantity is always close to 6 . at the wavevector corresponding to the first peak of the structure factor as function of the reduced time @xmath49 for five state points of the ka liquid with 8000 particles
the two isomorphic points ( i.e. , with the same @xmath50 ) have identical scaled self part of the intermediate scattering functions.,width=377 ] figure [ figure1 ] reports results from molecular dynamics simulations of the ka mixture with @xmath51 particles .
the figure shows the self part of the intermediate scattering functions @xmath52 at the wavevector corresponding to the first peak of the structure factor for five state points as functions of reduced time ( i.e. , time scaled by @xmath53 ) .
the simulations were performed in the nvt ensemble using the nose - hoover thermostat with characteristic time equal to 0.5 in md units .
two of the five state points in the figure are isomorphic .
these two points not only have the same average relaxation time , but the same relaxation behavior including the short - time `` cage - rattling '' contributions .
this figure confirms isomorph properties 2 and 3 .
figure [ fig : figure2](a ) gives aa radial distribution functions for various isomorphic state points of the ka system , where a is the large particle .
temperature varies by more than a factor of two
. nevertheless , fig .
[ fig : figure2](b ) shows that when plotted as function of the reduced distance ( @xmath54 ) , there is good data collapse .
this confirms isomorph property 5 .
the inset of ( b ) focuses on the radial distribution function peak , adding data for a second collection of isomorphic state points .
finally , fig .
[ fig : figure2](c ) shows two radial distribution functions for water , our favorite non - strongly correlating liquid water s correlation coefficient @xmath10 is close to zero , a fact that reflects the existence of a density maximum .
the two water state points have the same reduced diffusion constant ; thus if water had isomorphs , the two radial distribution functions should be identical except for an overall scaling .
this is not the case .
figure [ fig : figure3](a ) shows the time evolution of the potential energy when a temperature and density jump is made on a ka system in thermal equilibrium .
the jump brings the system to a new state point that is isomorphic with the initial point ( the state points in question are the two isomorphic state points of fig .
the jump was obtained in the following way : we instantaneously increased the box length without changing any particle positions ( the overshoot is due to this ) and simultaneously changed the thermostat temperature to the final temperature .
there is no sign of relaxation towards equilibrium after the jump .
thus the system is instantaneously in equilibrium after the jump , as predicted for jumps between isomorphic state points ( property 4 ) .
figure [ fig : figure3](b ) extends on the point made in ( a ) that a jump between two isomorphic state points conserves thermal equilibrium ( @xmath55 ) .
in contrast , the @xmath56 and @xmath57 jumps both have slow relaxations to equilibrium after the jump was applied at @xmath58 .
the inset is a scatter plot of virial vs potential energy for the three different starting points ( same volume ) and for state point ( 2 ) , which is isomorphic with state point ( 1 ) but has a different volume .
the curves show relaxation towards the equilibrium value represented by the dashed line , the curve that oscillates around the average value from the beginning shows the @xmath55 jump .
again , we see that instantaneous equilibration is a feature of jumps between isomorphic points .
the preservation of equilibrium for jumps between state points on an isomorph has consequences that are relevant for generic aging experiments .
consider the inset of fig .
[ fig : figure3 ] ( c ) .
suppose we start in equilibrium at state ( 1 ) and change temperature and volume to @xmath59 and @xmath60 .
state ( 2 ) has volume @xmath60 and lies on the same isomorph as state ( 1 ) .
now suppose that instead of jumping directly from point ( 1 ) to point ( 3 ) , one first jumps from ( 1 ) to ( 2 ) and then immediately after to point ( 3 ) . in that case , whenever relaxation times are long , the system never `` discovers '' that it spent a short amount of time in state ( 2 ) ( where it immediately equilibrated ) . in other words ,
the relaxation behavior going from ( 2 ) to ( 3 ) must be the same as for the ( 1 ) to ( 3 ) experiment .
figure [ fig : figure3 ] ( c ) confirms this . , @xmath56 , and @xmath57 .
the @xmath55 jump is between isomorphic state points and thus predicted not to involve any change beyond the immediate change ( compare ( a ) ) ; the two other jumps are not between isomorphic state points .
( c ) results of the potential energy relaxation towards equilibrium for the ka system comparing the jumps @xmath61 and @xmath62 .
the two relaxations are predicted to be identical.,title="fig:",width=340 ] , @xmath56 , and @xmath57 .
the @xmath55 jump is between isomorphic state points and thus predicted not to involve any change beyond the immediate change ( compare ( a ) ) ; the two other jumps are not between isomorphic state points .
( c ) results of the potential energy relaxation towards equilibrium for the ka system comparing the jumps @xmath61 and @xmath62 .
the two relaxations are predicted to be identical.,title="fig:",width=340 ] , @xmath56 , and @xmath57 .
the @xmath55 jump is between isomorphic state points and thus predicted not to involve any change beyond the immediate change ( compare ( a ) ) ; the two other jumps are not between isomorphic state points .
( c ) results of the potential energy relaxation towards equilibrium for the ka system comparing the jumps @xmath61 and @xmath62 .
the two relaxations are predicted to be identical.,title="fig:",width=340 ]
( b ) two isomorph predictions according to eqs . ( [ isou ] ) and ( [ isow ] )
compared to two series of state points ( marked by blobs ) that each have the same reduced relaxation time .
( c ) the `` master isomorph '' , collapsing two ka isomorphs and one for the wahnstrm binary lj liquid by scaling with @xmath63 defined as the virial on the same isomorph when @xmath64.,title="fig:",width=340 ] ) .
( b ) two isomorph predictions according to eqs . ( [ isou ] ) and ( [ isow ] )
compared to two series of state points ( marked by blobs ) that each have the same reduced relaxation time .
( c ) the `` master isomorph '' , collapsing two ka isomorphs and one for the wahnstrm binary lj liquid by scaling with @xmath63 defined as the virial on the same isomorph when @xmath64.,title="fig:",width=340 ] ) .
( b ) two isomorph predictions according to eqs . ( [ isou ] ) and ( [ isow ] )
compared to two series of state points ( marked by blobs ) that each have the same reduced relaxation time .
( c ) the `` master isomorph '' , collapsing two ka isomorphs and one for the wahnstrm binary lj liquid by scaling with @xmath63 defined as the virial on the same isomorph when @xmath64.,title="fig:",width=377 ] is it possible to say anything general about the isomorphic lines in the @xmath9 diagram ? consider a system characterized by an arbitrary number of @xmath65-@xmath66 lj interactions : @xmath67 where @xmath68^m $ ] .
for the total potential energy we get @xmath69 the definition of the virial implies that @xmath70 denoting a reference point by `` * '' , if we define the relative density @xmath71 , one has at points that are isomorphic wiht the reference point , because the scaled radial disbributions functions are identical , @xmath72 and @xmath73 these two equations parametrize the isomorph curves in the @xmath74-plot for any @xmath65-@xmath66 lj system . in particular , they imply that along an isomorph @xmath75 , which leads to @xmath76 ( where @xmath77 is the configurational pressure ) .
this expresses the fact that an isomorph is an excess entropy adiabat ( property 1 ) .
the two parameters @xmath78 and @xmath79 can be calculated by solving eqs .
( [ generalu ] ) and ( [ generalw ] ) at the reference point .
note that the predicted isomorphs are independent of the parameters characterizing the potential .
thus all kinds of different 12 - 6 lj - type liquids are predicted to have the same isomorphs ( provided an identical reference point exists ) .
figure [ fig : figure4 ] compares the above to simulations . in fig .
[ fig : figure4 ] ( a ) eq .
( [ isou ] ) is tested ; the red line in this plot has slope @xmath80 indicating that @xmath81 as predicted .
figure [ fig : figure4 ] ( b ) shows two predicted isomorphs according to eqs .
( [ isou ] ) and ( [ isow ] ) in the @xmath9 diagram ( dotted lines ) compared to two collections of isomorphic state points .
the agreement is good , with some deviations only when @xmath82 .
finally , fig .
[ fig : figure4 ] ( c ) shows that a `` master isomorph '' is obtained by scaling both @xmath6 and @xmath3 by @xmath63 defined as the virial for the point on the same isomorph with @xmath64 .
we here also give data from simulations of the wahnstrm binary lj liquid , which has quite different properties from the ka system .
all systems collapse to the master isomorph a parabola which is identified as follows . equations ( [ isou ] ) and ( [ isow ] ) imply @xmath83 and @xmath84 .
the requirement @xmath85 implies @xmath86 ^ 2=(w-2u)/w^*-2u^*)$ ] . since @xmath87 , in terms of the variables @xmath88 and
@xmath89 the master isomorph equation is @xmath90 .
strongly correlating liquids have lines in their state diagram `` isomorphs '' with the property that different points on an isomorph have potential energy landscapes that to a good approximation may be scaled into one another .
a number of static and dynamic properties are invariant along an isomorph .
an isomorph acts as wormhole in the state diagram .
w. kauzmann , chem . rev . * 43 * , 219 ( 1948 ) ; g. harrison , _ the dynamic properties of supercooled liquids _ ( academic , new york , 1976 ) ; s. brawer , _ relaxation in viscous liquids and glasses _ ( american ceramic society , columbus , oh , 1985 ) ; i. gutzow and j. schmelzer , _ the vitreous state : thermodynamics , structure , rheology , and crystallization _ ( springer , berlin , 1995 ) ; m. d. ediger , c. a. angell , and s. r. nagel , j. phys . chem . *
100 * , 13200 ( 1996 ) ; r. v. chamberlin , phase transitions * 65 * , 169 ( 1999 ) ; c. a. angell , k. l. ngai , g. b. mckenna , p. f. mcmillan , and s. w. martin , j. appl .
* 88 * , 3113 ( 2000 ) ; c. alba - simionesco , c. r. acad .
paris ( ser .
iv ) * 2 * , 203 ( 2001 ) ; p g. debenedetti and f. h. stillinger , nature * 410 * , 259 ( 2001 ) ; k. binder and w. kob , _ glassy materials and disordered solids : an introduction to their statistical mechanics _
( world scientific , singapore , 2005 ) ; f. sciortino , j. stat . mech .
, p05015 ( 2005 ) ; j. c. dyre , rev .
mod . phys . *
78 * , 953 ( 2006 ) .
n. l. ellegaard , t. christensen , p. v. christiansen , n. b. olsen , u. r. pedersen , t. b. schrder , and j. c. dyre , j. chem . phys .
* 126 * , 074502 ( 2007 ) ; n. p. bailey , t. christensen , b. jakobsen , k. niss , n. b. olsen , u. r. pedersen , t. b. schrder , and j. c. dyre , j. phys . : condens .
matter * 20 * , 244113 ( 2008 ) .
a. tlle , rep .
. phys . * 64 * , 1473 ( 2001 ) ; c. dreyfus , a. aouadi , j. gapinski , m. matos - lopes , w. steffen , a. patkowski , r. m. pick , phys .
e * 68 * , 011204 ( 2003 ) ; c. alba - simionesco , a. cailliaux , a. alegria , and g. tarjus , europhys . lett . * 68 * , 58 ( 2004 ) ; r. casalini and c. m. roland , phys .
e * 69 * , 062501 ( 2004 ) ; c. m. roland , s. hensel - bielowka , m. paluch , and r. casalini , rep . prog .
phys . * 68 * , 1405 ( 2005 ) .
w. g. hoover and m. ross , contemp . phys .
* 12 * , 339 ( 1971 ) ; w. g. hoover , d. a. young , and e grover , j. chem . phys . *
56 * , 2207 ( 1972 ) ; y. hiwatari , h. matsuda , t. ogawa , n. ogita , and a. ueda , prog .
phys . * 52 * , 1105 ( 1974 ) ; l. v. woodcock , phys .
lett . * 54 * , 1513 ( 1985 ) ; p. g. debenedetti , f. h. stillinger , t. m. truskett , and c. j. roberts , j. phys . chem . * 103 * , 7390 ( 1999 ) ; e. la nave , f. sciortino , p. tartaglia , m. s. shell , and p. g. debenedetti , phys .
e * 68 * , 032103 ( 2003 ) ; g. rickayzen and d. m. heyes , phys . rev .
e * 71 * , 061204 ( 2005 ) ; r. casalini , u. mohanty , and c. m. roland , j. chem . phys . * 125 * , 014505 ( 2006 ) . | a liquid is termed strongly correlating if its virial and potential energy thermal equilibrium fluctuations in the nvt ensemble are more than 90% correlated [ phys .
rev . lett . * 100 * , 015701 ( 2008 ) ] .
the fluctuations of a strongly correlating liquid are well approximated by those of an inverse power - law intermolecular potential .
building on this fact we here define `` isomorphic lines '' in the state diagram of a strongly correlating liquid .
it is shown from computer simulations of the kob - andersen binary lennard - jones liquid that no aging is associated with jumps between two isomorphic points .
isomorphic state points have the same excess entropy , the same reduced average relaxation time , the same dynamics , and the same scaled radial distribution functions .
finally we calculate the equation for isomorphs in the virial / potential energy diagram for lennard - jones type liquids and show that all such 12 - 6 liquids have the same isomorphs ; these may be scaled into a master isomorph . |
the energy distribution of primary @xmath0-rays provides information on both the level density and the @xmath0-ray strength function .
the present aim is to study the level density and @xmath0-strength function of the nuclei @xmath7er and to compare with other nuclei in this mass region .
in addition , thermodynamic and electromagnetic properties of the two nuclei will be investigated .
most of the experimental information on the level density of rare earth nuclei originates from the neutron - resonance spacing at the neutron - separation energy @xcite and direct counting of levels in the vicinity of the ground state @xcite .
experimental knowledge of the @xmath0-strength function is mainly based on the study of photo - absorption cross sections @xcite and radiative neutron capture @xcite .
a new method @xcite derives the level density and @xmath0-strength function simultaneously from a set of primary @xmath0 spectra .
this technique has proven to give a valuable supplement to the more traditional methods for level density extraction , and to our knowledge , it represents the least model - dependent method to obtain the @xmath0 strength function over a wide energy region below the neutron - separation energy .
the nuclear level density is closely related to the thermodynamic properties of nuclei , and can therefore be utilized to deduce e.g. entropy , temperature and heat capacity of nuclei .
the thermodynamic observables may reveal phase transitions .
the @xmath0-strength function is a measure for the average electromagnetic properties of nuclei and has a fundamental importance for the understanding of nuclear structure and reactions involving @xmath0 rays @xcite . at @xmath8 mev of @xmath0 energy ,
a bump is observed in the @xmath0-strength function of rare earth nuclei from ( @xmath3he,@xmath9 ) experiments @xcite , and is probably of the same origin as the pygmy resonance found in @xmath10 reactions @xcite .
section ii describes the experimental methods . in sect .
iii the experimental level density and @xmath0-strength function of @xmath1er and @xmath2er are obtained .
section iv examines thermodynamic properties within the micro - canonical and the canonical ensemble , while electromagnetic properties of the two nuclei are discussed in sect . v. concluding remarks are given in sect .
the experiment was carried out at the oslo cyclotron laboratory .
the reactions employed were the neutron pick - up ( @xmath3he,@xmath4 ) and the inelastic scattering ( @xmath3he,@xmath3he@xmath0 ) with a beam energy of @xmath11 mev , populating @xmath1er and @xmath2er with high excitation energy and low spin .
the @xmath12-value of the ( @xmath3he,@xmath9)-reaction is @xmath13 mev .
the self - supporting @xmath2er target was isotopically enriched to @xmath14% and had a thickness of @xmath15 mg/@xmath16 .
the charged particles and @xmath0 rays were recorded with the detector array cactus @xcite , which contains 8 si particle telescopes and 27 nai @xmath0-ray detectors .
each telescope is placed at an angle of 45@xmath17 relative to the beam axis , and comprises one si front and one si(li ) back detector with thickness @xmath18 and @xmath19 @xmath20 m , respectively .
the nai @xmath0-detector array , having a resolution of @xmath21 % at @xmath0 energy @xmath22 mev and a total efficiency of @xmath23 % , surrounds the target and particle detectors .
in addition , two ge detectors were used to monitor the spin distribution and selectivity of the reactions .
the excitation energy of the resulting @xmath1er and @xmath2er nuclei are determined by means of reaction kinematics of the ejectile . by setting proper gates in the particle spectra ,
each coincident @xmath0 ray can be assigned to a @xmath0 cascade originating from a specific excitation energy .
the data are sorted into a matrix of @xmath24 energy pairs , @xmath25 being the excitation energy .
examples of the recorded @xmath0 spectra , the so called raw spectra , from two excitation energies are shown in the left panel of figs .
[ fig : adata ] and [ fig : tdata ] for @xmath1er and @xmath2er , respectively .
note that the statistics of @xmath2er is about twice as good as for @xmath1er , since the ( @xmath3he,@xmath3he ) reaction has a higher cross section than the ( @xmath3he,@xmath9 ) reaction . in order to determine the true @xmath0-energy distribution , the @xmath0 spectra
are corrected for the response of the nai detectors with the unfolding procedure of ref .
@xcite . unfolded @xmath0 spectra are shown in the central panel of figs .
[ fig : adata ] and [ fig : tdata ] for the two nuclei .
the now corrected @xmath24 matrix comprises the @xmath0-energy distribution of the total @xmath0 cascade , and make it possible to derive the primary @xmath0 matrix according to the subtraction technique of refs . @xcite .
primary @xmath0 spectra can be seen in the right panel of figs .
[ fig : adata ] and [ fig : tdata ] .
the method of extracting the primary @xmath0 spectra is based on the assumption that the decay properties of a bin of excited states are independent of whether the states are directly populated through the nuclear reaction or from de - excitation from higher excited states .
this is believed to be approximately fulfilled because the long life time of excited states give the nucleus time to thermalize prior to the @xmath0 decay .
the primary @xmath0 matrix provides information on both the level density and the @xmath0-ray strength function , enabling a simultaneous determination of the two functions .
the fundamental assumption behind the extraction procedure is the brink - axel hypothesis @xcite , where the probability of @xmath0-decay in the statistical regime , represented by the primary @xmath0 matrix @xmath26 , can be expressed simply as a product of the final - state level density @xmath27 and a @xmath0-energy dependent factor @xmath28 @xmath29 the @xmath0-energy dependent factor is proportional to @xmath30 , where @xmath31 is the @xmath0-ray strength function for the multipolarity @xmath32 . during the last few years
a method for simultaneous deduction of level density and @xmath0-strength function from primary @xmath0 spectra has been developed at the oslo cyclotron laboratory @xcite . in the previously published results on @xmath1er @xcite ,
the level densities and @xmath0-strength functions were still unnormalized .
the procedure is now fully replaced by an iteration technique deriving the level density and the @xmath0-energy dependent factor by a @xmath33-fit to the experimental primary @xmath0 spectra @xcite .
this technique is described in detail in ref .
@xcite , and only an outline of the normalization procedures is given here .
it is important to note that the extracted experimental quantity @xmath34 is in fact the density of levels accessible to the nuclear system in the @xmath0-decay process .
this quantity is interpreted as the level density , but may be influenced by selection rules in the @xmath0 decay . for a thermalized nucleus in the continuum , the interpretation is approximately valid @xcite
. equation ( [ eq : ab ] ) has an infinite number of solutions .
it can be shown @xcite that all equally good solutions of eq .
( [ eq : ab ] ) can be obtained by the transformations @xmath35 \rho(e - e_{\gamma } ) , \label{eq : array } \\
\tilde{f}(e_{\gamma } ) & = & b \exp(\alpha e_{\gamma } ) f(e_{\gamma } ) \label{eq : array2}\end{aligned}\ ] ] of any particular solution ( @xmath36 ) . consequently
, neither the slope nor the absolute value of the two functions can be obtained through the iteration procedure , but the three variables @xmath37 , @xmath38 and @xmath9 of eqs .
( [ eq : array ] ) and ( [ eq : array2 ] ) have to be determined independently to give the best physical solution of the level density and @xmath0-strength function . the parameters @xmath37 and @xmath9 of eq . (
[ eq : array ] ) can be determined by fitting the level density from the iteration procedure to the number of known discrete levels @xcite at low excitation energy and to the level density estimated from neutron - resonance spacing data at high excitation energy @xcite .
this normalization procedure is shown for @xmath2er in the right panel of fig .
[ fig : normrho ] .
the deduced level density is fitted to the discrete levels ( between the arrows in the upper panel of fig .
[ fig : normrho ] ) as far up in energy as we can assume that all levels are known . at high excitation energies the deduced level density
is fitted ( between the arrows in the lower panel of fig .
[ fig : normrho ] ) to a fermi - gas approximation of the level density ( line ) forced to pass through the level - density estimate at the neutron - separation energy ( filled square ) obtained from the neutron - resonance spacing data .
unfortunately , @xmath39er is an unstable nucleus ( @xmath40 h ) , making the neutron capture @xmath39er(n,@xmath0)@xmath1er difficult to investigate .
thus , there is no neutron - resonance data available for @xmath1er . in this case
the level density at the neutron - separation energy is estimated with the fermi - gas expression @xmath41 where the level density parameter @xmath42 , the shifted excitation energy @xmath43 , and the spin - cutoff parameter @xmath44 is defined by @xmath45 , while @xmath37 is the mass number , @xmath46 the pairing - gap parameter and @xmath47 .
all parameters are calculated according to ref .
@xcite , which provides a level density in acceptable agreement with the experimental level density at the neutron - separation energy for @xmath2er .
also the nuclei @xmath48dy and @xmath49yb show conformity between the experimental and theoretical level density at the neutron - separation energy @xcite .
the left panel of fig .
[ fig : normrho ] shows the normalization of @xmath1er , where the open square represents the theoretically calculated level density at the neutron - separation energy with an error of @xmath50% @xcite .
the normalized level density of @xmath1er is compared to the level densities of @xmath51dy and @xmath52yb @xcite in the left panel of fig .
[ fig : rhosmnl ] .
the three level - density functions coincide well at high excitation energies , indicating that the chosen value for the level density at the neutron - separation energy for @xmath1er is reasonable . for excitation energies below @xmath53 mev the experimental level density of the erbium nucleus
is seen to be lower than for the two others . in this energy region
most levels in the three nuclei are believed to be known . from counting of known levels @xcite , it is found that between @xmath54 and @xmath53 mev , the level density of @xmath51dy and @xmath52yb indeed exceed the level density of @xmath1er with the same degree as observed experimentally .
the right panel of fig .
[ fig : rhosmnl ] compares the level density of @xmath2er with the level densities of @xmath55dy and @xmath56yb @xcite .
the two latter nuclei were populated through ( @xmath3he,@xmath9 ) reactions .
the @xmath56yb - level density is seen to be lower than the two others below @xmath54 mev . according to @xcite
, this agrees with reality below @xmath57 kev . above this limit
the level densities from counting of known levels start to decrease , making it evident that not all levels are known . in ref .
@xcite the level density of @xmath51dy extracted from two different reaction mechanisms are compared .
the primary @xmath0 spectra from the ( @xmath3he,@xmath9 ) pick - up reaction and the inelastic ( @xmath3he,@xmath3he ) reaction are found to provide level densities with very good agreement except below @xmath58 mev of excitation energy .
there , the level density obtained from the inelastic @xmath3he scattering overestimates the number of levels by a factor of @xmath8 , because this reaction populates more collective excitations with similar structure as the ground - state band .
this causes a relatively large @xmath0-transition rate from the direct populated states to the ground band , which in turn gives an overestimation of the level density there .
the same is observed in the level density of @xmath52yb deduced from the inelastic reaction @xcite .
this effect of overestimation can not be ruled out in the level density of @xmath2er derived from the ( @xmath3he,@xmath3he ) reaction , however , the experimental data fit the known level density at low excitation energy very well , see fig . [
fig : normrho ] .
a difference between these ( @xmath3he,@xmath3he ) experiments is the spin of the target nuclei .
the @xmath2er target has a spin of @xmath59 making excitations to a large number of states probable , with a resulting fragmentation of collective strength and a lower probability for subsequent @xmath0 transitions directly into the ground - state band than in the even - even nuclei .
thus , the accessible states in the @xmath0 decay seem to approximates the level density quite well .
blatt and weisskopf @xcite suggested the ratio of the partial radiative width @xmath60 and the level spacing of the initial states @xmath61 with equal spin and parity @xmath62 to describe the @xmath0 decay in the continuum .
the corresponding definition of the @xmath0-ray strength function is given by @xmath63 , where @xmath64 denotes the electric or magnetic character , and @xmath65 defines the multipolarity of the @xmath0 transition .
after the normalization of the level density , the parameter @xmath9 of eq .
( [ eq : array2 ] ) is already fixed , and the slope , i.e. the exponential @xmath66 , is included in the @xmath0-energy dependent functions shown for @xmath1er and @xmath2er in fig .
[ fig : f ] . still , the normalization constant @xmath38 of eq .
( [ eq : array2 ] ) remains to be determined .
the @xmath0-energy dependent factor @xmath28 is proportional to the sum of @xmath67 for all possibilities of @xmath64 and @xmath65 .
we assume that the @xmath0 decay in the continuum of nuclei with low spin is mainly governed by electric and magnetic dipole radiation and that the accessible levels have equal numbers of positive and negative parity states .
thus , the observed @xmath68 can be expressed by a sum of the e1 and m1 @xmath0-strength functions only : @xmath69 e_{\gamma}^3 .
\label{eq : prim1}\ ] ] the average total radiative width of neutron resonances @xmath70 @xcite with excitation energy equal to the neutron - separation energy @xmath71 , spin @xmath72 , and parity @xmath73 @xmath74 can be written in terms of @xmath68 by means of eq .
( [ eq : prim1 ] ) . with the experimental level density @xmath34
already normalized , the normalization constant @xmath38 can be deduced .
assuming @xmath75-neutron capture , @xmath76 take the values @xmath77 , where @xmath78 are the spin and parity of the target nucleus in the @xmath10 reaction .
the experimental value of the total radiative width is the average over the possible spins of the compound state ( @xmath79 ) .
since we expect the @xmath0 decay to be governed by dipole transitions , the second sum is restricted to possible final states with spin @xmath80 and parity @xmath81 accessible by dipole radiation .
the methodical difficulties in the primary @xmath0-ray extraction , prevents determination of the functions @xmath28 and @xmath82 in the interval @xmath83 mev and @xmath84 mev , respectively .
in addition , the data at the highest @xmath0-energies , above @xmath85 mev , suffer from poor statistics .
therefore , extrapolations of @xmath68 and @xmath34 were necessary in order to calculate the integral of eq .
( [ eq : norm ] ) . for the level density
a fermi - gas extrapolation is used ( see fig .
[ fig : normrho ] ) , and for the @xmath0-energy dependent factor a pure exponential of the form @xmath86 is utilized ( see fig . [
fig : f ] ) .
the contribution of the extrapolation to the total radiative width in eq .
( [ eq : norm ] ) does not exceed @xmath87% , thus the errors due to a possibly poor extrapolation are expected to be of minor importance @xcite .
values for the neutron - resonance radiative width are given in @xcite , and for @xmath2er we find @xmath88 mev . since the radiative width of @xmath1er is unknown , its value is taken as the average of @xmath70 of the neighbouring isotopes .
this is justified by the general slow variance in the @xmath70 values of neighbouring isotopes for other elements . from interpolation
, we thus adopt the value @xmath89 mev for @xmath1er .
the normalized @xmath0-strength functions of @xmath1er and @xmath2er are compared to the strength functions of @xmath51dy and @xmath52yb , and @xmath55dy and @xmath56yb , respectively , in fig .
[ fig : sfsmnl ] .
all strength functions are increasing smoothly with @xmath0 energy except for a bump around @xmath90 mev .
the @xmath0-strength functions of neighbouring isotopes display a striking qualitative similarity .
this can be expected for nuclei with approximately the same charge distribution .
the location of the bump can however be seen to move towards higher @xmath0 energies with increasing mass number .
for a statistical description of hot nuclei , the micro - canonical thermodynamics is the proper theory . within this frame ,
the system is isolated , giving a well defined energy .
however , the canonical ensemble , permitting heat exchange , and the grand - canonical ensemble , which in addition allows particle exchange , are often used due to mathematical difficulties with detailed calculations within the micro - canonical ensemble . in this work ,
both the micro - canonical and the canonical ensemble will be utilized to discuss thermodynamic properties of the @xmath7er nuclei experimentally . in particular we will focus on different ways , within the two ensembles , to obtain experimental values of the critical temperature for the pair - breaking process and the general quenching of pair correlations .
the micro - canonical ensemble gives a detailed description of the breaking of one , two , three , ... nucleon pairs as a function of excitation energy , while the canonical ensemble reveals the general average properties of this phase - like transition .
the proper definition of thermally driven first- and second - order phase transitions in systems with few particles is a long standing problem , which will not be discussed in the present experimental work .
the micro - canonical partition function is simply the multiplicity of nuclear states , which experimentally corresponds to the level density of accessible states .
thus , the experimental level density @xmath82 is our starting point for the extraction of thermodynamic properties of nuclei .
the entropy is determined by @xmath91 where the boltzmann constant for simplicity is set to unity ( @xmath92 ) , and the normalization constant @xmath93 can be adjusted to fulfill the condition of the third law of thermodynamics ; @xmath94 when @xmath95 , @xmath96 being the nuclear temperature @xmath97 the ground - state band of even - even nuclei is assumed to have @xmath98 . therefore , @xmath93 is determined so that the entropy of the ground - state band in @xmath1er is approximately zero
. figures [ fig : s ] and [ fig : mu ] show the entropy and the temperature , respectively , deduced in the micro - canonical ensemble for @xmath1er and @xmath2er .
the entropy curve plotted in a linear scale is essentially identical to the level - density curve in a logarithmic scale .
the small bumps in the entropy curves of fig .
[ fig : s ] are enhanced through the differentiation performed in eq .
( [ eq : t ] ) to obtain the temperatures of fig .
[ fig : mu ] .
the specific heat can be determined from differentiating the temperature @xmath99 and is displayed in the lower panel of fig .
[ fig : mu ] for @xmath1er and @xmath2er .
the double differentiation of the entropy has introduced strong fluctuations in the specific heat .
still it is distinct that the oscillations in the temperature give rise to successive positive and negative heat capacity .
the spectacular feature of negative heat capacity is a direct consequence of the decrease in the micro - canonical temperature , and has recently been observed experimentally in the critical region of nuclear fragmentation in au quasi - projectile sources formed in au+au collisions @xcite .
an anomalous decrease in the temperature means an unusual increase in the entropy , i.e. a more than normal opening of new domains of the phase space , e.g. opening of new degrees of freedom as more particle pairs are broken . the successive bumps of figs .
[ fig : s ] and [ fig : mu ] indicate that the transition from a paired to an unpaired phase is a gradual process of breaking up more and more nucleon pairs .
when we interpret the area of negative heat capacity , corresponding to the region where the temperature has a negative slope , as the location of the break up of a nucleon pair , a signal of the break up of the first pair can be seen around @xmath100 mev in @xmath2er in fig .
[ fig : mu ] .
this value is close to twice the pairing - gap parameters @xmath101 mev and @xmath102 mev , which are the expected cost of breaking of a neutron and proton pair , respectively .
the pairing - gap parameters are calculated from the empirical masses of a sequence of isotopes or isotones @xcite of @xmath2er . the first negative slope in the temperature as a function of excitation energy in @xmath55dy , @xmath51dy , @xmath56yb , and @xmath52yb @xcite also coincide roughly with @xmath103 .
a similar argument is however less successful for @xmath1er .
deviations from @xmath103 in the localization of the break up of the first pair of particles can be due to the influence of several structural effects in the nuclei , such as e.g. the fermi - level position in the nilsson single - particle scheme , variation in the density of single - particle orbitals , and two quasi - particle couplings to collective degrees of freedom .
it has been shown @xcite that phase transitions of finite - size systems can be observed in the micro - canonical ensemble without invoking the thermodynamic limit . according to @xcite
, a phase transition of first order can be recognized from the caloric curve in the region of co - existing phases , i.e. in the excitation energy region where the phase transition takes place . there , both the phase the system is leaving and the phase the system is attending are existing simultaneously . for this co - existency
the system has to `` pay '' an amount of entropy which is eventually returned as the whole system is converted into the new phase @xcite .
the non - oriented area between the inverse temperature @xmath104 and the line @xmath105 is twice this entropy .
the critical temperature has to satisfy the condition that the oriented area between @xmath104 and @xmath105 equals zero .
the process studied in the present work , namely the breaking of nucleon pairs and the general quenching of pair correlations , deviates from the process of multifragmentation @xcite . in the latter process ,
the geometrical surfaces of the fragments play a crucial role . in the depairing process
, the long - range two - nucleon force has correlation lengths longer than the diameter of the nucleus .
thus , the phase - like situation of having a pair coupled or not , reveals no geometrical surface of interaction . in order to qualitatively analyze our experimental findings along the procedure of ref .
@xcite , the excitation energy region around the outstanding bump in the temperature of @xmath2er is expanded , and the entropy , inverse temperature and heat capacity of @xmath2er are plotted in fig .
[ fig : gross ] . the qualitative similarity with the picture of ref .
@xcite suggests that the general process of breaking a nucleon pair around @xmath106 appears like a phase transition of first order .
accepting this theoretical result , however , the similar experimental signals of breaking of further nucleon pairs indicate that the process of quenching of pairing correlations is a series of first order phase transitions .
this interpretation might seem physically unattractive , and therefore we are anticipating further theoretical clarifications . the convex shape of the entropy curve of fig .
[ fig : gross ] indicates a transition from the one quasi - particle regime to the three quasi - particle regime .
the transition takes place between the two tangent points of the entropy and its enveloping curve ( dashed line in the entropy plot ) .
the transition region is marked with two dotted vertical lines .
the critical temperature for breaking of the first nucleon pair in @xmath2er is derived to be @xmath107 mev ( horizontal dashed line in the caloric curve of fig . [
fig : gross ] ) .
the inter - phase entropy could in principle be expected to be equal to the envelope entropy , that is , an interpolation of the entropies before and after the pair - breaking process . but compared to this expectation , the entropy is seen to be lowered by @xmath108 units , corresponding to the sum of the primary loss and the final gain in entropy through the transition .
it is unclear how this entropy - mixing contribution shall be interpreted in a small nuclear system .
the first law of thermodynamics states that supplied heat equals the change in internal energy plus the work performed by the system : @xmath109 .
since the nucleus is an isolated system the work @xmath110 is zero , and the increase in energy equals the supplied heat through the phase transition . in macroscopic systems , where the temperature is constant through a phase transition of first order , this increase in energy is the latent heat .
we interpret the transition - region energy of approximately @xmath111 mev ( see fig . [
fig : gross ] ) as the latent heat , even though the temperature is not constant .
we note that this value is also very close to the expected cost @xmath103 , of breaking up a particle pair .
the dashed lines in the specific heat are purely meant as guiding lines towards the theoretical prediction of poles between the successive positive and negative branches .
the derivative of the entropy - envelope curve in fig .
[ fig : gross ] equals the critical temperature .
the somewhat stronger concave shape of the micro - canonical entropy of @xmath2er compared to @xmath1er ( see fig .
[ fig : s ] ) suggests that increasing temperature is necessary to break further nucleon pairs of @xmath2er , whereas the critical temperature for breaking of particle pairs is rather constant in @xmath1er . after the break up of the first particle pair in @xmath2er ,
the slope of the entropy becomes however more constant .
a linear fit to the entropy above @xmath112 mev gives a critical temperature of @xmath113 mev for @xmath1er and @xmath114 mev for @xmath2er , indicating that the critical temperature for breaking of nucleon pairs in @xmath2er has a slight increase .
this can be understood from the blocking effect of the unpaired nucleon , increasing the distance to the fermi surface for low - lying orbitals with coupled pairs .
the critical temperature for breaking of the first particle pair in @xmath2er could be nicely determined from equal areas in the caloric plot .
unfortunately , the thermodynamic quantities of small systems are strongly fluctuating .
thus , the critical temperature for each back bend of the temperature , e.i . breaking of a particle pair , can generally be difficult to obtain by this method .
the slope of the entropy - envelope curve can however always serve as a measure .
the transformation from the micro - canonical to the canonical ensemble is performed by the canonical partition function @xmath115 the partition function is a laplace transform of the multiplicity of states @xmath116 , where @xmath117 is the level density of accessible states in the nuclear reaction at the discrete energy @xmath25 given in energy bins @xmath118 .
the thermal average of the excitation energy in the canonical ensemble is @xmath119 by the laplace transform in eq .
( [ eq : z ] ) much of the information contained in the micro - canonical level density becomes smeared out .
this smoothing effect can be quantified by the standard deviation for the thermal average of the energy @xmath120 which e.g. gives @xmath121 mev at @xmath122 mev .
thus , fine structure in the thermodynamic observables in the micro - canonical ensemble will not be visible in the canonical ensemble .
the lines in the upper panel of fig .
[ fig : mu ] display the smooth variance of the canonical temperature as a function of the thermal average of the excitation energy .
the mathematical justification of eqs .
( [ eq : z ] ) and ( [ eq : e ] ) is that the summation is performed to infinity .
however , for the typical temperatures studied here , @xmath123 mev , the level density should be known up to @xmath124 mev , which was found @xcite to be a sufficiently high upper limit for the summation .
the experimental level density is only covering the excitation energy region from zero to about @xmath125 mev . in the region of and above the neutron - separation energy @xmath71
, the fermi - gas model is believed to describe the nuclear properties .
therefore , the experimental level density is extrapolated to higher excitation energies with the fermi - gas approximation of ref .
@xcite , see the solid lines of fig .
[ fig : normrho ] .
the heat capacity in the canonical ensemble is the derivative of the thermal average of the excitation energy @xmath126 and it is shown for @xmath7er as a function of the thermal average of the excitation energy in fig .
[ fig : mu ] and as a function of temperature in fig .
[ fig : crit ] . the averaging done by the partition function can be seen to provide a smooth energy and temperature dependence of the heat capacity .
the heat capacity of both nuclei shows an @xmath127 shape as a function of temperature .
this feature is interpreted @xcite as a fingerprint of a phase transition from a state with strong pairing interaction to a state where the pairing correlations are quenched .
because of the smoothing performed by the canonical - partition function , discrete transitions between the different quasi - particle regimes , as observed within the micro - canonical ensemble , are hidden , and only the phase transition related to the quenching of the pair correlations as a whole can be seen . the shape of the heat - capacity curve is related to the level - density . in general
a constant - temperature level density gives a pronounced @xmath127 shape @xcite , while a pure fermi - gas level density provides a linear heat capacity @xmath128 ; see the dashed lines of fig .
[ fig : crit ] .
the more prominent @xmath127 shape in @xmath1er is due to a quicker rise in the heat capacity .
this is the same phenomenon as observed in the micro - canonical ensemble , where the concave shape of the @xmath2er entropy was assigned to the blocking effect of the odd neutron .
shell model monte carlo calculations @xcite on various fe isotopes have shown that the pairing - phase transition is strongly correlated with the suppression of neutron pairs with increasing temperature .
it is also observed that the reduction of neutron pairs is much stronger in the even- than in the odd - mass isotopes , giving rise to a more pronounced @xmath127 shape in the even nuclei . the same difference in the heat capacity
is also observed experimentally between @xmath55dy and @xmath51dy , and @xmath56yb and @xmath52yb @xcite . in order to extract a critical temperature for the quenching of pair correlations from the canonical data ,
two different approaches has been tested .
first , the local maximum of the heat capacity relative to the fermi - gas approximation of fig .
[ fig : crit ] have been determined to be @xmath129 mev for @xmath1er and @xmath130 mev for @xmath2er , see the arrows of fig .
[ fig : crit ] .
this method relies very much on the fermi - gas extrapolation of the level density .
the second approach depends on the assumption that the nuclear level density can be approximated by a constant - temperature expression at low excitation energies . within the canonical ensemble
this assumption provides the relations @xmath131 @xmath132 where @xmath133 is the constant - temperature parameter . by fitting @xmath134 as a function of @xmath135 to a straight line of slope @xmath136
, @xmath133 can be determined .
to minimize the dependency of the extracted critical temperature on the extrapolated level density , the fit is performed between @xmath137 mev . according to eq .
( [ eq : e ] ) this corresponds to energies in the level density curve up to @xmath21 mev .
then @xmath133 is identified with the critical temperature , since @xmath138 has a pole at @xmath139 .
the critical temperature is determined to @xmath140 mev for @xmath1er and @xmath141 mev for @xmath2er .
the dashed - dotted lines of fig .
[ fig : crit ] describe eq .
( [ eq : cvt ] ) with asymptotes at @xmath142 .
the different values derived for @xmath143 are quoted in table [ tab : crit ] . the micro - canonical caloric - curve method ( fig .
[ fig : gross ] ) is only supposed to provide the critical temperature for breaking of the first particle pair , while the others should give the critical temperature for the total quenching of the pairing correlations .
the canonical constant - temperature approach relies on experimental data , but it can be questionable whether eqs .
( [ eq : invt ] ) and ( [ eq : cvt ] ) describes the nucleus well at low excitation energies .
the method of deducing @xmath143 from the local maximum of the canonical heat capacity relative to the fermi - gas approximation depends highly on the extrapolation of the level density .
the most straight forward method is to obtain the temperature directly from the slope of the micro - canonical entropy .
there seem to be a tendency for the critical temperature in @xmath1er to be slightly lower than in @xmath2er .
all results are however equal within the approximated uncertainties , with a mean value of @xmath144 mev for both nuclei .
the results for @xmath1er agrees with recent finite - temperature hartree fock bogoliubov calculations for @xmath145er @xcite , in which the critical - temperature estimate extracted is @xmath146 mev .
this value is somewhat higher than the present derived result for @xmath1er .
this diversity is probably due to application of different interpretations of where exactly the phase transition takes place . since the shapes of the heat - capacity curves of @xmath145er and @xmath1er are practically identical , we assume that the calculated critical temperature for quenching of pair correlations in @xmath145er coincides with the experimental findings for @xmath1er
. relativistic hartree - bcs calculations @xcite also find the proton and neutron pairing gaps to vanish around @xmath147 mev of temperature in @xmath1er and @xmath148er .
the heat capacity shows two peaks corresponding to these events , characteristic of second order phase transitions from superfluid to normal phase @xcite .
thus theoretically , the pair quenching is calculated to be a phase transition of second order .
provided that negative heat capacity is a sufficient signal , the observed breaking of nucleon pairs in the micro - canonical ensemble appears like first - order phase transitions .
the process of breaking one nucleon pair does however not lead to a total quenching of the pair correlations , and may not be described by the same physics as the pairing - phase transition .
it is not clear how a series of phase transitions should be interpreted physically .
the @xmath0-strength function is governed by different multipolarities of electric and magnetic character .
we assume however , that the @xmath0 decay in the continuum is dominated by dipole transitions and will try to model the @xmath0-strength function theoretically .
it is commonly adopted that the e1 strength is determined by the giant electric dipole resonance ( gedr ) at high @xmath0 energies .
more doubtful is the assumption that the tail of the gedr governs the e1 strength at low @xmath0 energies when the tail approaches zero @xcite .
a model @xcite taking into account the energy and temperature dependence of the gedr width is often utilized to describe the experimental data @xcite .
the m1 strength function also plays an important role governing the @xmath0 emission for low @xmath0 energies .
experiments indicate the existence of an m1 giant resonance due to spin - flip excitations in the nucleus @xcite . in fig .
[ fig : sffit ] , the experimental @xmath0-strength function is fitted by a theoretical strength function taking into account both the giant electric dipole resonance and the spin - flip resonance .
in addition , a weaker resonance at lower energies is needed in order to fit the experimental data .
because of the much lower strength compared to the gedr , the resonance is denoted the pygmy resonance . to account for the e1 radiation ,
the model @xcite @xmath149 is adopted .
the values for the giant electric dipole resonance parameters @xmath150 , @xmath151 and @xmath152 are taken from @xcite .
the temperature parameter @xmath96 is utilized as a constant fit parameter .
the m1 radiation is described by @xmath153 where @xmath154 , @xmath155 and @xmath156 are the giant magnetic dipole resonance parameters , which are taken from @xcite .
the pygmy resonance is here described with a similar lorentzian function @xmath157 as eq .
( [ eq : m1 ] ) , where the pygmy - resonance strength @xmath158 , width @xmath159 and centroid @xmath160 have been fitted in order to adjust the total theoretical strength function @xmath161 to the experimental data .
the resulting theoretical @xmath0-strength functions are shown as solid lines in fig .
[ fig : sffit ] .
the obtained values of the fitting parameters for the pygmy resonance , the normalization constant @xmath162 and the temperature @xmath96 are given in table [ tab : fit ] .
the low - energetic resonance has its centroid at @xmath163 mev in @xmath1er and at @xmath164 mev in @xmath2er . in @xmath10 reactions a pygmy resonance
is observed @xcite in the energy region @xmath165 mev for nuclei of mass 160 - 176 .
both the centroid and the width of the resonance is found @xcite to increase gradually with neutron number . in @xmath1er
the centroid , width and strength of the pygmy resonance are somewhat lower than in @xmath2er , fitting very well into the systematics of @xcite .
the enhanced @xmath0 strength around @xmath166 mev in @xmath48dy and @xmath49yb @xcite also fit into this systematics , indicating that the @xmath8 mev bump after ( @xmath3he,@xmath9 ) and ( @xmath3he,@xmath3he ) reactions probably is of the same origin as the pygmy resonance reported in @xcite .
the pygmy resonance has been explained by the enhancement of the e1 @xmath0-strength function @xcite .
the possibility that it on the contrary is of m1 character can not be excluded . at an excitation energy around @xmath166 mev , there is a concentration of orbital m1 strength in the weakly collective scissors mode @xcite
. the scissors mode was first observed in electron - scattering experiments @xcite , and is confirmed by the @xmath167 reaction @xcite .
a systematic survey of centroids of observed m1 strength distributions in even - even nuclei in the @xmath168 mass region is given in ref .
@xcite , where the scissors mode of @xmath1er is found to be centered at @xmath169 mev , in perfect agreement with the pygmy resonance found here .
figure [ fig : sffit ] also shows the predicted individual contributions from the giant electric dipole resonance @xmath170 , the giant magnetic dipole resonance @xmath171 and the pygmy resonance @xmath157 to the @xmath0-strength function .
the strength function is generally dominated by e1 radiation .
the m1-strength function @xmath171 is always @xmath172% lower than the e1-strength function @xmath170 .
moreover , the pygmy resonance contributes considerably around @xmath166 mev .
the normalization constant @xmath162 is close to 1 for @xmath2er ( table [ tab : fit ] ) , showing that the adopted theoretical model reproduces the absolute values of the @xmath0-strength functions very well . the deviation from @xmath136
can be explained by uncertain values of the gedr parameters for the investigated nuclei and uncertainties in the experimental normalization of the @xmath0-strength functions .
also the @xmath162 value for @xmath1er is reasonable , provided that the adopted value for the neutron - resonance radiative width is reliable .
levels in @xmath1er and @xmath2er in the excitation region up to the neutron separation energy were populated with the ( @xmath3he,@xmath9@xmath0 ) and ( @xmath3he,@xmath3he@xmath0 ) reaction , respectively . the level density and @xmath0-ray strength function of @xmath1er and @xmath2er are determined from their corresponding primary @xmath0-ray spectra .
thermodynamic observables are deduced from the level density and display signatures of phase - like transitions within the micro - canonical and the canonical ensemble , interpreted as the transition from a strongly pair - correlated phase to an uncorrelated phase .
micro - canonical thermodynamics give the possibility of investigating the successive breaking of nucleon pairs in detail , information which is hidden in the canonical approach . the canonical ensemble on the other hand , reveals the average properties of the pairing transition .
in addition , the canonical ensemble provides an excellent opportunity to study the different mechanisms governing the thermodynamic properties of odd and even systems .
the increase in the heat capacity with temperature is much steeper in @xmath1er than in @xmath2er , probably due to the tendency of the odd neutron to block higher quasiparticle excitations .
various estimates of the critical temperature for the pairing - phase transition is performed , giving the result @xmath5 mev .
the experimental @xmath0-strength function is fitted by a theoretical strength function , assuming that the @xmath0 decay in the continuum is governed by dipole transitions .
the contribution of electric and magnetic dipole radiation to the @xmath0-strength function is recognized .
a bump is observed in the @xmath0-strength function at @xmath173 and @xmath174 mev in the @xmath1er and @xmath2er , respectively , and is probably of the same origin as the pygmy resonance found in the @xmath10 reaction .
iljinov , m.v .
mebel , n. bianchi , e. de sanctis , c. guaraldo , v. lucherini , v. muccifora , e. polli , a.r .
reolon , and p. rossi , nucl .
* a543 , 517 ( 1992 ) , and references therein . r.b .
firestone , and v.s .
shirley , _
table of isotopes _ , @xmath175 edition , vol .
ii ( new york : john wiley & sons , inc . , 1996 ) .
dietrich and b.l .
berman , at .
data nucl .
data tables * 38 , 199 ( 1988 ) .
f. bev , p. cejnar , j. honztko , k. konen , i. tomandl , and r.e .
chrien , phys . rev .
c * 52 * , 1278 ( 1995 ) .
a. schiller , l. bergholt , m. guttormsen , e. melby , j. rekstad , and s. siem , nucl .
instrum .
methods phys .
a * 447 * 498 ( 2000 ) .
j. kopecky and m. uhl , phys .
c * 41 * , 1941 ( 1990 ) .
a. voinov , m. guttormsen , e. melby , j. rekstad , a. schiller , and s. siem , nucl - ex0009018 .
m. igashira , h. kitazawa , m. shimizu , h. komano , and n. yamamuro , nucl .
* a457 , 301 ( 1986 ) . m. guttormsen , a. atac , g. lvhiden , s. messelt , t. ramsy , j. rekstad , t.f .
thorsteinsen , t.s .
tveter , and z. zelazny , phys .
* t32 , 54 ( 1990 ) .
m. guttormsen , t.s .
tveter , l. bergholt , f. ingebretsen , and j. rekstad , nucl .
instrum .
methods phys
. res .
a * 374 , 371 ( 1996 ) .
m. guttormsen , t. ramsy , and j. rekstad , nucl .
instrum .
methods phys .
a * 255 * , 518 ( 1987 ) .
l. henden , l. bergholt , m. guttormsen , j. rekstad , and t.s .
tveter , nucl .
* a589 , 249 ( 1995 ) . d.m .
brink , ph.d .
thesis , oxford university ( 1955 ) .
p. axel , phys .
rev . * 126 * , 671 ( 1962 ) .
tveter , l. bergholt , m. guttormsen , e. melby , and j. rekstad , phys .
* 77 , 2404 ( 1996 ) .
e. melby , l. bergholt , m. guttormsen , m. hjorth - jensen , f. ingebretsen , s. messelt , j. rekstad , a. schiller , s. siem , and s.w .
degrd , phys .
lett . * 83 , 3150 ( 1999 ) .
a. schiller , m. guttormsen , e. melby , j. rekstad , and s. siem , phys .
c * 61 , 044324 ( 2000 ) .
t. von egidy , h.h .
schmidt , and a.n .
behkami , nucl .
* a481 * , 189 ( 1988 ) . m. guttormsen , a. bjerve , m. hjorth - jensen , e. melby , j. rekstad , a. schiller , s. siem , and a. beli , phys .
c * 62 * 024306 ( 2000 ) .
a. schiller , doctorial thesis , university of oslo ( 2000 ) .
blatt and v.f .
weisskopf , _ theoretical nuclear physics _ ( john wiley & sons , new york , 1952 ) .
handbook for calculations of nuclear reaction data , iaea , vienna , iaea - tecdoc-1024 ( 1998 ) .
m. dagostino _ et al .
_ , phys .
b * 473 * , 219 ( 2000 ) .
a. bohr and b.r .
mottelson , _ nuclear structure _ ,
i ( benjamin , new york , 1969 ) 169 .
a. schiller , m. guttormsen , m. hjorth - jensen , e. melby , j. rekstad , and s. siem , nucl - ex/0007009 .
gross and e.v .
votyakov , eur .
j. b * 15 * , 115 ( 2000 ) .
gross , phys .
rep . 279 , 119 ( 1997 ) , and references therein .
a. schiller , a. bjerve , m. guttormsen , m. hjorth - jensen , f. ingebretsen , e. melby , s. messelt , j. rekstad , s. siem , and s.w .
degrd , nucl - ex/9909011 .
s. liu and y. alhassid , nucl - th/0009006 .
egido , l.m .
robledo , and v. martin , phys .
* 85 * , 26 ( 2000 ) .
agrawal , tapas sil , j.n . de , and s.k .
samaddar , phys .
c * 62 * , 044307 ( 2000 ) .
kadmenski , v.p .
markushev , and v.i .
furman , yad .
fiz . * 37 , 277 ( 1983 ) [ sov .
j. nucl .
phys . * 37 , 165 ( 1983 ) ] .
j. kopecky and r.e .
chrien , nucl .
* a468 * , 285 ( 1987 ) .
a. richter , nucl .
phys . * a507 , 99c ( 1990 ) .
d. bohle , a. richter , w. steffen , a.e.l .
dieperink , n. lo iudice , f. palumbo , and o. scholten , phys .
b * 137 , 27 ( 1984 ) .
n. pietralla , p. von brentano , r .- d .
herzberg , u. kneissl , n. lo iudice , h. maser , h.h .
pitz , and a. zilges , phys .
c * 58 , 184 ( 1998 ) , and references therein . * * * * * * * * * * * * * * er ( left panel ) and @xmath2er ( right panel ) between the arrows to known levels at low excitation energy ( histograms ) and to the fermi - gas level density ( @xmath1er ) or to the level density calculated from neutron - resonance spacings ( @xmath2er ) at the neutron - separation energy ( squares ) .
the lines are the fermi - gas approximations.,width=642 ] er ( left ) and @xmath2er ( right ) compared to @xmath51dy and @xmath52yb , and @xmath55dy and @xmath56yb , respectively .
the dy isotopes are multiplied by 0.1 and the yb isotopes by 10 for better visualization .
all nuclei are popolated through the ( @xmath3he,@xmath9 ) reaction except for @xmath2er , which is populated through the ( @xmath3he,@xmath3he ) reaction . , width=642 ] -ray strength function of @xmath1er ( left ) compared to @xmath51dy and @xmath52yb , and @xmath2er ( right ) compared to @xmath55dy and @xmath56yb .
the dy isotopes are multiplied by 0.1 and the yb isotopes by 10 for better visualization .
, width=642 ] er ( left ) and @xmath2er ( right ) as a function of temperature .
the dashed lines are the fermi - gas approximations @xmath177 .
the local maxima relative to the fermi - gas expressions are marked with arrows , while the dashed - dotted lines describe the estimates of eq .
( [ eq : cvt ] ) , where @xmath133 is recognized as the critical temperature and marked with the vertival lines.,width=642 ] -ray strength function ( data points ) of @xmath1er ( left ) and @xmath2er ( right ) .
the solid line is the fit to the data by the theoretical model .
the dashed lines are the respective contributions of the gedr , the gmdr , and the pygmy resonance to the total theoretical strength function.,width=642 ] | the primary @xmath0-ray spectra of @xmath1er and @xmath2er are deduced from the ( @xmath3he,@xmath4 ) and ( @xmath3he,@xmath3he@xmath0 ) reaction , respectively , enabling a simultaneous extraction of the level density and the @xmath0-ray strength function .
entropy , temperature and heat capacity are deduced from the level density within the micro - canonical and the canonical ensemble , displaying signals of a phase - like transition from the pair - correlated ground state to an uncorrelated state at @xmath5 mev .
the @xmath0-ray strength function displays a bump around @xmath6 mev , interpreted as the pygmy resonance . 2 |
there is the following naive question from the early days of differential topology : given two closed smooth manifolds @xmath2 and @xmath3 , when does an isomorphism @xmath5 of integral cohomology rings imply that @xmath2 and @xmath3 are diffeomorphic ? this is generally regarded as an unlikely case , as in the 20th century topologists discovered many important series of manifolds for which the cohomology ring , or even the homotopy type , does not determine the diffeomorphism class .
three - dimensional lens spaces , milnor s exotic spheres and donaldson s four - dimensional manifolds are prominent examples of different level of complexity . we say that a family of closed smooth manifolds is _ cohomologically rigid _ if a cohomology ring isomorphism @xmath5 implies a diffeomorphism @xmath6 for any two manifolds in the family .
in this paper we establish cohomological rigidity for two particular families of manifolds of dimension @xmath7 and @xmath8 , respectively .
each of these families arises from a particularly important class of combinatorial polytopes , which we refer to as the _ pogorelov class _
it consists of simple @xmath7-dimensional polytopes which are flag and do not have 4-belts .
in particular , polytopes in @xmath0 do not have triangular and quadrangular facets .
the class @xmath0 includes combinatorial fullerenes , i.e. simple 3-polytopes with only pentagonal and hexagonal facets . by the results of pogorelov
@xcite and andreev @xcite , the class @xmath0 consists precisely of combinatorial @xmath7-polytopes which can be realised in lobachevsky space @xmath9 with right angles between adjacent facets ( _ right - angled @xmath7-polytopes _ for short ) .
the conditions specifying the pogorelov class also feature as the `` no @xmath10-condition '' and `` no @xmath11-condition '' in gromov s theory of hyperbolic groups @xcite .
our first family consists of _ hyperbolic @xmath7-manifolds of lbell type_. these manifolds were introduced and studied by vesnin @xcite .
they arise from right - angled realisations of polytopes from the pogorelov class @xmath0 ( see the details in subsection [ sechyp ] ) .
each hyperbolic @xmath7-manifold @xmath12 of lbell type is composed of @xmath13 copies of a polytope @xmath14 and is a _ small cover _ of @xmath1 in the sense of davis and januszkiewicz @xcite .
we prove ( in theorem [ 3cohrigreal ] ) that two such manifolds @xmath12 and @xmath15 are diffeomorphic if and only if their @xmath16-cohomology rings are isomorphic .
our second family arises from toric topology : it consists of quasitoric ( or topological toric ) manifolds whose quotient polytopes are in the class @xmath0 .
these are @xmath8-dimensional smooth manifolds acted on by a @xmath7-torus @xmath17 with quotient @xmath14 .
we show ( in theorem [ 3cohrig ] and corollary [ coro:5 - 1 ] ) that this family is cohomologically rigid , i.e. two manifolds @xmath2 and @xmath3 in the family are diffeomorphic if and only if their cohomology rings are isomorphic . in general a diffeomorphism between quasitoric manifolds @xmath2 and @xmath3 does not imply that the corresponding polytopes @xmath1 and @xmath4 are combinatorially equivalent , but this is the case when the quotient polytopes are in the class @xmath0 ( see theorem [ 3cohrig ] ) .
our proofs use both combinatorial and cohomological techniques of toric topology , and build upon recent important results of fan , ma and wang on cohomological rigidity of moment - angle manifolds @xcite , @xcite .
these authors come to the class @xmath0 independently , by considering the property of @xmath18-rigidity for simple polytopes ( see section [ cohrig ] ) .
the class @xmath0 also featured in the work of buchstaber and erokhovets @xcite on combinatorial classification of fullerenes .
we note that cohomological rigidity is open for the whole family of toric or topological toric manifolds .
in fact , we find it quite surprising that no counterexamples to the `` toric cohomological rigidity problem '' were found up to date .
in real dimension @xmath8 the families of quasitoric and topological toric manifolds coincide and contain strictly the family of toric manifolds ( smooth complete toric varieties ) .
the family of quasitoric ( or topological toric ) manifolds whose quotient polytopes are in the class @xmath0 is large enough , as there is at least one quasitoric manifold over any simple @xmath7-polytope .
indeed , the four colour theorem implies that any simple @xmath7-polytope admits a `` characteristic function '' ( see proposition [ 4cprop ] ) ; this remarkable observation was made by davis and januszkiewicz in @xcite .
toric manifolds whose associated polytopes are in @xmath0 are fewer , but still abundant ; many concrete examples were produced recently by suyama @xcite .
on the other hand , there are no _ projective _ toric manifolds whose associated polytopes are in @xmath0 .
this follows from a result of c.delaunay @xcite that a delzant 3-polytope must have at least one triangular or quadrangular face .
our results on cohomological rigidity of toric manifolds chime with the problem of diffeomorphism classification for simply connected oriented manifolds , which is a classical subject of algebraic and differential topology .
the foundations of this classification in dimensions @xmath19 were laid in the works of browder and novikov ( see @xcite , @xcite ) .
novikov @xcite showed that for a given simply connected oriented manifold @xmath2 of dimension @xmath19 there are only finitely many manifolds @xmath3 for which there exists a homotopy equivalence @xmath20 preserving the pontryagin classes .
important classification results in dimension @xmath8 were obtained in the works of wall @xcite , jupp @xcite and zhubr @xcite .
toric , quasitoric or topological toric manifolds @xmath2 are simply connected , and their cohomology rings @xmath21 are generated by @xmath22-dimensional classes . according to the general results mentioned above , two such manifolds in dimension @xmath8 are diffeomorphic if there is an isomorphism of their cohomology rings preserving the first pontryagin class @xmath23 .
therefore , the toric cohomological rigidity problem in dimension @xmath8 reduces to establishing the invariance of @xmath23 under cohomology isomorphisms .
this turns out to be a purely combinatorial and linear algebra problem , see details in section [ class6 ] .
however , we were not able to prove directly the invariance of @xmath23 under cohomology isomorphisms for toric manifolds over simple @xmath7-polytopes from the class @xmath0 .
one of our main results ( theorem [ 3cohrig ] ) can be interpreted as a classification result for a particular family of simply connected @xmath8-dimensional manifolds , and its proof is independent of the general classification results of wall , jupp and zhubr .
the authors are grateful to * feifei fan for introducing us to the results of the works @xcite and @xcite on the cohomological rigidity of moment - angle complexes ; * alexandr gaifullin for drawing our attention to siebennman s no @xmath11-condition in gromov s theory of hyperbolic groups , which is equivalent to the absence of @xmath24-belts in the case of polytopes ; * sergei novikov for stimulating discussions on the diffeomorphism classification of simply connected manifolds and invariance of pontryagin classes ; * andrei vesnin for his help in linking our results to hyperbolic geometry ; in particular , to hyperbolic @xmath7-manifolds of lbell type and the theorem of pogorelov on right - angled @xmath7-polytopes .
we thank the institute for mathematical science , national university of singapore , and the organisers of the program on combinatorial and toric homotopy , where we were able to exchange ideas which laid the foundation of this work .
we also thank the jsps program for advancing strategic international networks to accelerate the circulation of talented researchers based on osaka city university advanced mathematical institute ( ocami ) , which gave us an opportunity to meet and work together on this project .
here we collect the necessary information about toric varieties , quasitoric manifolds and moment - angle manifolds ; the details can be found in @xcite .
let @xmath25 be an @xmath26-dimensional euclidean space with the scalar product @xmath27 .
a _ convex polytope _
@xmath1 is a nonempty bounded intersection of finitely many half - spaces in some @xmath25 : @xmath28 where @xmath29 and @xmath30 .
we often fix a presentation by inequalities alongside with the polytope @xmath1 .
we assume that @xmath1 is @xmath26-dimensional , that is , the dimension of the affine hull of @xmath1 is @xmath26 .
we also assume that each inequality @xmath31 in is not redundant , that is , can not be removed without changing @xmath1 .
then @xmath1 has @xmath32 _ facets _ @xmath33 , where @xmath34 each facet is a polytope of dimension @xmath35 .
a _ face _ of @xmath1 is a nonempty intersection of facets .
zero - dimensional faces are _ vertices _ , and one - dimensional faces are _
edges_. we refer to @xmath26-dimensional polytopes simply as
_ @xmath26-polytopes .
_ two polytopes @xmath1 and @xmath36 are _ combinatorially equivalent _ ( @xmath37 ) if there is a bijection between their faces preserving the inclusion relation .
combinatorial polytope _ is a class of combinatorially equivalent polytopes .
a polytope @xmath1 is _ simple _ if exactly @xmath26 facets meet at each vertex of @xmath1 .
a simple polytope @xmath1 is called a _ flag polytope _ if every collection of its pairwise intersecting facets has a nonempty intersection
. an _ @xmath26-simplex _
@xmath38 is not flag for @xmath39 .
an _ @xmath26-cube _
@xmath40 is flag for any @xmath26 .
we denote by @xmath41 the vertex - edge graph of a polytope @xmath1 , and refer to it simply as the _ graph of @xmath1_. a graph is _ simple _ if it has no loops and multiple edges . a connected graph
@xmath42 is _ @xmath7-connected _ if it has at least @xmath8 edges and deletion of any one or two vertices with all incident edges leaves @xmath42 connected .
the following classical result describes the graphs of 3-polytopes .
[ s - theorem ] a graph @xmath42 is the graph of a @xmath7-polytope if and only it is simple , planar and @xmath7-connected . for @xmath43 , a _
@xmath44-belt _ in a simple polytope @xmath1 is a cyclic sequence @xmath45 of facets in which only two consecutive facets ( including @xmath46 ) have nonempty intersection . for a _
@xmath7-belt _ we assume additionally that @xmath47 .
a @xmath7-polytope @xmath1 with a triangular facet has a 3-belt around it , unless @xmath48 .
a simple 3-polytope @xmath49 is flag if and only if it does not contain 3-belts .
a _ fullerene _ is a simple 3-polytope with only pentagonal and hexagonal facets .
a simple calculation with euler characteristic shows that the number of pentagonal facets in a fullerene is @xmath50 .
the number of hexagonal facets can be arbitrary except for @xmath51 ( see ( * ? ? ? * proposition 2 ) ) .
also , any fullerene is a flag polytope without @xmath24-belts ( see @xcite and ( * ? ? ?
* corollary 3.16 ) ) .
toric variety _ is a normal complex algebraic variety @xmath52 containing an algebraic torus @xmath53 as a zariski open subset in such a way that the natural action of @xmath53 on itself extends to an action on @xmath52 .
we only consider nonsingular complete ( compact in the usual topology ) toric varieties , also known as _
toric manifolds_. there is a bijective correspondence between the isomorphism classes of complex @xmath26-dimensional toric manifolds and complete nonsingular fans in @xmath25 .
a _ fan _ is a finite collection @xmath54 of strongly convex cones @xmath55 in @xmath25 such that every face of a cone in @xmath56 belongs to @xmath56 and the intersection of any two cones in @xmath56 is a face of each .
a fan @xmath56 is _ nonsingular _ ( or _ regular _ ) if each of its cones @xmath57 is generated by part of a basis of the lattice @xmath58 ( we choose the standard lattice for simplicity ) .
in particular , each one - dimensional cone of @xmath56 is generated by a primitive vector @xmath59 .
a fan @xmath56 is _ complete _ if the union of its cones is the whole @xmath25 .
projective toric varieties are particularly important . a projective toric manifold @xmath52 is defined by a _ lattice delzant polytope _
given a simple @xmath26-polytope @xmath1 with vertices in the lattice @xmath60 , the _ normal fan _
@xmath61 has @xmath26-dimensional cones @xmath62 corresponding to the vertices @xmath63 of @xmath1 , where @xmath62 is generated by the primitive inside - pointing normals to the facets of @xmath1 meeting at @xmath63 . the polytope @xmath1 is _ delzant _ whenever its normal fan @xmath61 is nonsingular .
the fan @xmath61 defines a projective toric manifold @xmath64 .
different lattice delzant polytopes with the same normal fan produce different projective embeddings of the same toric manifold .
irreducible torus - invariant subvarieties of complex codimension one in @xmath52 correspond to one - dimensional cones of @xmath56 .
when @xmath52 is projective , they also correspond to the facets of @xmath1 .
we assume that there are @xmath32 one - dimensional cones ( or facets ) , denote the corresponding primitive vectors by @xmath65 , and denote the corresponding codimension-1 subvarieties by @xmath66 .
[ cohomtoric ] let @xmath52 be a toric manifold of complex dimension @xmath26 , with the corresponding complete nonsingular fan @xmath56 .
the cohomology ring @xmath67 is generated by the degree - two classes @xmath68 $ ] dual to the invariant submanifolds @xmath69 , and is given by @xmath70/\mathcal i,\qquad\deg v_i=2,\ ] ] where @xmath71 is the ideal generated by elements of the following two types : * @xmath72 such that @xmath73 do not span a cone of @xmath56 ; * @xmath74 , for any vector @xmath75 .
it is convenient to consider the integer @xmath76-matrix @xmath77 whose columns are the vectors @xmath78 written in the standard basis of @xmath60 .
then the ideal ( b ) of theorem [ cohomtoric ] is generated by the @xmath26 linear forms @xmath79 corresponding to the rows of @xmath80 .
the quotient of a projective toric manifold @xmath64 by the action of the compact torus @xmath81 is the polytope @xmath1 .
when a toric manifold @xmath52 is not projective , the quotient @xmath82 has a face structure of a _ manifold with corners_. this face structure locally looks like that of a simple convex polytope , but globally may fail to be so even combinatorially . in the case
@xmath83 , however , the quotient @xmath84 is combinatorially equivalent to a simple 3-polytope , by steinitz s theorem ( theorem [ s - theorem ] ) .
in their 1991 work @xcite davis and januszkiewicz suggested a topological generalisation of projective toric manifolds , which became known as quasitoric manifolds . a _
quasitoric manifold _ over a combinatorial simple @xmath26-polytope @xmath1 is a topological manifold @xmath2 of dimension @xmath85 with a locally standard action of @xmath86 and a projection @xmath87 whose fibres are the orbits of the @xmath86-action .
( an action of @xmath86 on @xmath2 is _ locally standard _ if every point @xmath88 is contained in a @xmath86-invariant neighbourhood equivariantly homeomorphic to an open subset in @xmath89 with the standard coordinatewise action of @xmath86 twisted by an automorphism of the torus .
the orbit space of a locally standard action is a manifold with corners . for a quasitoric manifold @xmath2 ,
the orbit space @xmath90 is homeomorphic to @xmath1 . ) not every simple polytope can be the quotient of a quasitoric manifold .
nevertheless , quasitoric manifolds constitute a much larger family than projective toric manifolds , and enjoy more flexibility for topological applications .
let @xmath91 be the set of facets of @xmath1 .
each @xmath92 is a quasitoric submanifold of @xmath2 of codimension 2 , called a _
characteristic submanifold_. the characteristic submanifolds @xmath93 are analogues of the invariant divisors @xmath69 on a toric manifold @xmath52 .
each @xmath94 is fixed pointwise by a closed one - dimensional subgroup ( a subcircle ) @xmath95 and therefore corresponds to a primitive vector @xmath96 defined up to a sign . choosing a direction of @xmath97 is equivalent to choosing an orientation for the normal bundle @xmath98 or , equivalently , choosing an orientation for @xmath94 , provided that @xmath2 itself is oriented .
an _ omniorientation _ of a quasitoric manifold @xmath2 consists of a choice of orientation for @xmath2 and each characteristic submanifold @xmath99 .
the vectors @xmath97 are analogues of the generators @xmath78 of the one - dimensional cones in the fan corresponding to a toric manifold @xmath52 , or analogues of the normal vectors to the facets of @xmath1 when @xmath52 is projective . however , the vectors @xmath97 need not be the normal vectors to the facets of @xmath1 in general . there is an analogue of theorem [ cohomtoric ] for quasitoric manifolds : [ cohomqtoric ] let @xmath2 be an omnioriented quasitoric manifold of dimension @xmath85 over a simple @xmath26-polytope @xmath1 .
the cohomology ring @xmath100 is generated by the degree - two classes @xmath68 $ ] dual to the oriented characteristic submanifolds @xmath94 , and is given by @xmath101/\mathcal i,\qquad\deg v_i=2,\ ] ] where @xmath71 is the ideal generated by elements of the following two types : * @xmath72 such that @xmath102 in @xmath1 ; * @xmath103 , for any vector @xmath75 .
we record a simple corollary for the latter use .
[ viprod ] in the notation of theorem [ cohomqtoric ] , * the product @xmath104\cdots[v_{i_n}]$ ] of @xmath26 different classes is a generator of @xmath105 if @xmath106 and is zero otherwise ; * for @xmath107 , we have @xmath68[v_j]=0 $ ] if and only if @xmath108 . by analogy with
, we consider the integer _
characteristic matrix _ @xmath109 whose columns are the vectors @xmath97 written in the standard basis of @xmath60 . the matrix @xmath80 has the following property : @xmath110 note that the ideal ( b ) of theorem [ cohomqtoric ] is generated by the @xmath26 linear forms @xmath111 corresponding to the rows of @xmath80 .
a map @xmath112 , @xmath113 , satisfying is called a _ characteristic function _ for a simple @xmath26-polytope @xmath1 .
one can produce a characteristic matrix @xmath80 from a characteristic function @xmath114 by fixing an order of facets .
characteristic pair _
@xmath115 consists of a simple polytope @xmath1 and its characteristic matrix @xmath80 .
a quasitoric manifold @xmath2 defines a characteristic pair @xmath115 . on the other hand ,
each characteristic pair gives rise to a quasitoric manifold as follows .
[ canqt ] let @xmath115 be a characteristic pair .
for each facet @xmath116 of @xmath1 we denote by @xmath117 the circle subgroup of @xmath118 corresponding to the @xmath119th column @xmath96 of the characteristic matrix @xmath80 . for each point @xmath120 ,
define a torus @xmath121 assuming that @xmath122 if there are no facets containing @xmath123 .
property implies that @xmath124 embeds as a subgroup in @xmath86 .
then define @xmath125 where the equivalence relation @xmath126 is given by @xmath127 whenever @xmath128 and @xmath129 .
one can see that @xmath130 is a quasitoric manifold over @xmath1 .
change of basis in the lattice results in multiplying @xmath80 from the left by a matrix from @xmath131 .
changing the orientation of the @xmath119th characteristic submanifold @xmath94 in the omniorientation data results in changing the sign of the @xmath119th column of @xmath80 .
a combinatorial equivalence between polytopes @xmath1 and @xmath4 allows us to identify their sets of facets @xmath132 and @xmath133 and therefore identify their characteristic functions .
these observations lead us to the following definition .
[ defcharpair ] two characteristic pairs @xmath115 and @xmath134 are _ equivalent _ if * @xmath1 and @xmath4 are combinatorially equivalent , and * @xmath135 , where @xmath136 and @xmath18 is a diagonal @xmath137-matrix with @xmath138 on the diagonal .
quasitoric manifolds @xmath130 and @xmath139 corresponding to equivalent pairs are equivariantly homeomorphic ( in the weak sense ) .
the latter means that there is a homeomorphism @xmath140 satisfying @xmath141 for any @xmath142 and @xmath143 , where @xmath144 is the automorphism of the torus given by the matrix @xmath145 .
furthermore , we have there is a one - to - one correspondence between equivariant homeomorphism classes of quasitoric manifolds and equivalence classes of characteristic pairs .
in particular , for any quasitoric manifold @xmath2 over @xmath1 with characteristic matrix @xmath80 , there is an equivariant homeomorphism @xmath146 .
both @xmath2 and @xmath130 were defined as topological manifolds in @xcite .
the manifold @xmath130 can be endowed with a canonical smooth structure by defining it as the quotient of the moment - angle manifold @xmath147 by a smooth torus action , see @xcite and subsection [ zpm ] . nevertheless , for a smooth quasitoric manifold @xmath2 , the existence of a _
@xmath146 is a delicate issue , see the discussion in @xcite . on the other hand , in the case of @xmath8-dimensional quasitoric manifolds ( which is our main concern in this paper ) ,
such a diffeomorphism follows from the classification results of wall and jupp discussed in section [ class6 ] . in dimensions
@xmath148 , there are simple @xmath26-polytopes @xmath1 which do not admit any characteristic matrix @xmath80 , see @xcite
. such a polytope can not be the quotient of a quasitoric manifold .
on the other hand , we have the following observation by davis and januszkiewicz , whose proof remarkably uses the four colour theorem : [ 4cprop ] any simple @xmath7-polytope admits a characteristic matrix @xmath80 .
given a 4-colouring of the facets of @xmath1 , we assign to a facet of @xmath119th colour the @xmath119th basis vector @xmath149 for @xmath150 and the vector @xmath151 for @xmath152 .
the resulting @xmath153-matrix @xmath80 satisfies , as any three of the four vectors @xmath154 form a basis of @xmath155 .
a projective toric manifold is a quasitoric manifold .
a non - projective toric manifold @xmath52 may fail to be quasitoric , as the quotient manifold with corners @xmath82 is not necessarily a simple polytope , even combinatorially .
first examples of this sort appear in dimension @xmath156 , see @xcite .
all complex 3-dimensional toric manifolds , even non - projective ones , are quasitoric by the steinitz theorem ( theorem [ s - theorem ] ) . replacing the torus @xmath86 in the definition of a quasitoric manifold by the subgroup @xmath157 ( @xmath26 commuting involutions )
, one obtains the definition of a small cover @xcite .
a _ small cover _ of a simple @xmath26-polytope @xmath1 is a manifold @xmath12 of dimension @xmath26 with a locally standard action of @xmath158 and a projection @xmath159 whose fibres are the orbits of the @xmath158-action .
the set of real points of a projective toric manifold @xmath64 ( i.e. the set of points fixed under complex conjugation ) is a small cover of @xmath1 ; it is sometimes called a _
real toric manifold_. the theory of small covers parallels that of quasitoric manifolds , and we just outline the most crucial points .
[ cohomsc ] let @xmath12 be a small cover of a simple @xmath26-polytope @xmath1 .
the cohomology ring @xmath160 is generated by the degree - one classes @xmath68 $ ] dual to the characteristic submanifolds @xmath161 , and is given by @xmath162/\mathcal i,\qquad\deg v_i=1,\ ] ] where @xmath71 is the ideal generated by elements of the following two types : * @xmath72 such that @xmath102 in @xmath1 ; * @xmath103 , for any vector @xmath163 .
the characteristic matrix @xmath80 corresponding to a small cover @xmath12 has entries in @xmath16 and satisfies the same condition .
the equivalence of @xmath16-characteristic pairs is defined in the same way as in the quasitoric case , with @xmath164 replaced by @xmath165 .
a small cover @xmath12 of @xmath1 is equivariantly homeomorphic to the `` canonical model '' @xmath166 with the equivalence relation @xmath126 defined as in the quasitoric case .
note that @xmath167 is composed of @xmath168 copies of the polytope @xmath1 , patched together along their facets . reducing a @xmath169-characteristic matrix @xmath170
we obtain a @xmath16-characteristic matrix .
the following question is open .
assume given a @xmath16-characteristic pair @xmath115 consisting of a simple @xmath26-polytope @xmath1 and an @xmath171-matrix @xmath80 with entries in @xmath16 satisfying .
can @xmath80 be obtained by reduction @xmath170 from an integer matrix satisfying the same condition ?
the answer to the above problem is positive for @xmath7-polytopes : [ lambda2 ] every @xmath16-characteristic pair @xmath115 with @xmath7-dimensional @xmath1 is the @xmath172 reduction of a @xmath169-characteristic pair .
one can check that any @xmath173-matrix with entries @xmath174 or @xmath51 and determinant @xmath175 has determinant @xmath138 when viewed as an integer matrix .
ineed , such a matrix either has a column with @xmath22 zeros , or is @xmath176 up to permutation of rows and columns .
the required property is then verified directly .
a particularly important class of @xmath7-dimensional small covers are _ hyperbolic @xmath7-manifolds of lbell type _ , introduced and studied by vesnin in @xcite .
let @xmath1 be a polytope in @xmath7-dimensional lobachevsky space @xmath9 with right angles between adjacent facets ( a _ right - angled @xmath7-polytope _ for short ) .
denote by @xmath177 the group generated by reflections in the facets of @xmath1 .
right - angled coxeter group _ given by the presentation @xmath178 where @xmath179 denotes the reflection in the facet @xmath116 .
the reflections in adjacent facets commute because of the right - angledness .
there are no relations between the reflections in non - adjacent faces , as the corresponding reflection hyperplanes do not intersect in @xmath9 .
the group @xmath177 acts on @xmath9 discretely with finite isotropy subgroups and with fundamental domain @xmath1 .
now consider an epimorphism @xmath180 . as was observed in (
* lemma 1 ) , the subgroup @xmath181 does not contain elements of finite order if and only if the images of the reflections in any three facets of @xmath1 that have a common vertex are linearly independent in @xmath182 . in this case
the group @xmath183 acts freely on @xmath9 .
the quotient @xmath184 is called a _ hyperbolic @xmath7-manifold of lbell type_. it is composed of @xmath185 copies of @xmath1 and has a riemannian metric of constant negative curvature .
furthermore , such a manifold @xmath12 is aspherical ( has the homotopy type of eilenberg mac lane space @xmath186 ) , as its universal cover @xmath9 is contractible . since the abelianisation of @xmath177 is @xmath187 , the epimorphism @xmath188 factors as @xmath189 , where @xmath80 is a linear map .
the above condition for the freeness of the action of @xmath183 on @xmath9 is equivalent to that @xmath80 satisfies ( i.e. @xmath80 is given by a @xmath16-characteristic matrix ) .
we therefore can identify @xmath184 with the small cover @xmath167 .
which combinatorial @xmath7-polytopes have right - angled realisations in @xmath9 ?
pogorelov stated necessary conditions in 1967 , which were later shown by andreev to be sufficient : a combinatorial @xmath7-polytope can be realised as a right - angled polytope in lobachevsky space @xmath9 if and only if it is simple , flag and does not have @xmath24-belts .
furthermore , such a realisation is unique up to isometry .
we refer to the class of simple flag @xmath7-polytopes without @xmath24-belts as the _ pogorelov class _ @xmath0 .
it will feature prominently throughout the rest of our paper .
a polytope from the class @xmath0 does not have triangular or quadrangular facets . according to a result of doli @xcite
( see also ( * ? ? ?
* corollary 3.16 ) ) , the pogorelov class contains all fullerenes .
we summarise the constructions and results above as follows .
a small cover @xmath167 of a @xmath7-polytope @xmath1 from the pogorelov class @xmath0 has the structure of a hyperbolic @xmath7-manifold of lbell type @xmath190 , with the epimorphsim @xmath188 given by the composition @xmath191 .
furthermore , such a @xmath7-manifold @xmath167 is aspherical . the conditions specifying the pogorelov class @xmath0 also feature in gromov s theory of hyperbolic groups . namely , the `` no @xmath10-condition '' from ( * ?
* e ) for a simplicial complex @xmath192 is the absence of missing @xmath22-faces , while the `` no @xmath11-condition '' is the absence of chordless @xmath24-cycles .
when @xmath192 is the dual complex of a simple polytope , these two conditions translate to the absence of @xmath7- and @xmath24-belts , respectively .
the relationship between small covers and hyperbolic manifolds was also mentioned in the work of davis and januszkiewicz @xcite , although the criterion for right - angled realisation of a polytope was stated there incorrectly .
a toric manifold is not necessarily a quasitoric manifold and a quasitoric manifold is also not necessarily a toric manifold .
however , both toric and quasitoric manifolds are examples of _ topological toric manifolds _ introduced in @xcite . recall that a toric manifold admits an algebraic action of @xmath53 with an open dense orbit .
it has local charts equivariantly isomorphic to a sum of complex one - dimensional _ algebraic _ representations of @xmath53 .
a topological toric manifold is a compact smooth @xmath85-dimensional manifold with an effective smooth action of @xmath53 having an open dense orbit and covered by finitely many invariant open subsets each equivariantly diffeomorphic to a sum of complex one - dimensional _ smooth _ representation spaces of @xmath53 .
( the latter condition automatically follows from the existence of a dense orbit in the algebraic category , but not in the smooth category . ) the cohomology ring of a topological toric manifold is described similarly to the toric or quasitoric case ; there is an analogue of theorems [ cohomtoric ] or [ cohomqtoric ] , see ( * ? ? ?
* proposition 8.3 ) .
let @xmath192 be an ( abstract ) _ simplicial complex _ on the set @xmath193=\{1,\ldots , m\}$ ] , i.e. @xmath194 is a collection of subsets @xmath195 $ ] such that for any @xmath196 all subsets of @xmath197 also belong to @xmath192 .
we always assume that the empty set @xmath198 and all one - element subsets @xmath199 $ ] belong to @xmath192 .
we refer to @xmath200 as a _ simplex _ ( or a _ face _ ) of @xmath192 .
every abstract simplicial complex @xmath192 has a _ geometric realisation _
@xmath201 , which is a polyhedron in a euclidean space ( a union of convex geometric simplices ) .
a _ non - face _ of @xmath192 is a subset @xmath195 $ ] such that @xmath202 .
a _ missing face _ ( or a _ minimal non - face _ ) of @xmath192 is an inclusion - minimal non - face of @xmath192 , that is , a subset @xmath203 $ ] such that @xmath197 is not a simplex of @xmath192 , but every proper subset of @xmath197 is a simplex of @xmath192 .
a simplicial complex @xmath192 is called a _ flag complex _ if each of its missing faces consists of two vertices .
equivalently , @xmath192 is flag if any set of vertices of @xmath192 which are pairwise connected by edges spans a simplex .
every flag complex @xmath192 is determined by its 1-skeleton @xmath204 , and is obtained from the graph @xmath204 by filling in all complete subgraphs by simplices .
let @xmath1 be a simple @xmath26-polytope with @xmath32 facets @xmath33 .
then @xmath205\colon f_{i_1}\cap\cdots\cap f_{i_k}\ne\varnothing\bigr\}\ ] ] is a simplicial complex on @xmath193 $ ] , called the _ dual complex _ of @xmath1 .
the vertices of @xmath206 correspond to the facets of @xmath1 , and the empty simplex @xmath198 corresponds to @xmath1 itself .
geometrically , @xmath207 is an @xmath208-dimensional sphere simplicially subdivided as the boundary of the _ dual polytope _ of @xmath1 .
the definitions of flag polytopes and complexes agree : @xmath1 is a flag polytope if and only if @xmath206 is a flag complex .
a @xmath44-belt in @xmath1 with @xmath43 corresponds to a _
chordless @xmath44-cycle _ in the graph @xmath209 .
we fix a commutative ring @xmath210 with unit .
the _ face ring _ of @xmath192 ( also known as the _ stanley reisner ring _ ) is defined as the quotient of the polynomial algebra @xmath211 $ ] by the square - free monomial ideal generated by non - simplices of @xmath192 : @xmath212=\k[v_1,\ldots , v_m]\big/\bigl(v_{i_1}\cdots v_{i_k}\ \colon \{i_1,\ldots , i_k\}\notin\sk\bigr).\ ] ] as @xmath213 $ ] is the quotient of the polynomial ring by a monomial ideal , it has a grading or even a multigrading ( a @xmath214-grading ) .
we use an even grading : @xmath215 and @xmath216 , where @xmath217 is the @xmath119th standard basis vector .
note that when @xmath218 for a simple polytope @xmath1 , the ring @xmath219 $ ] coincides with the quotient of @xmath220 $ ] by the relations ( a ) in theorem [ cohomtoric ] or in theorem [ cohomqtoric ] .
a simplicial complex @xmath192 is flag if and only if its face ring @xmath213 $ ] is a _ quadratic algebra _ , i.e. the quotient of @xmath211 $ ] by an ideal generated by quadratic monomials ( which have degree 4 in our grading ) .
let @xmath192 be a simplicial complex on the set @xmath193 $ ] , and let @xmath221 denote the pair of a disc and its boundary circle
. for each simplex @xmath222 , set @xmath223 the _ moment - angle complex _ is defined as @xmath224 if @xmath201 is homeomorphic to a sphere @xmath225 , then @xmath226 is a topological manifold . if @xmath192 is the boundary of a convex simplicial polytope or is a starshaped sphere ( the underlying complex of a complete simplicial fan ) , then @xmath226 has a smooth structure . in the polytopal case
there is an alternative way to define @xmath226 in terms of the dual simple polytope @xmath1 .
namely , assume given a presentation of a convex @xmath26-dimensional polytope @xmath1 by inequalities .
define the map @xmath227 so @xmath228 . also , define the map @xmath229 then define the space @xmath147 by the pullback diagram @xmath230 the space @xmath147 can be written as an intersection of @xmath231 hermitian quadrics in @xmath232 , and this intersection is nondegenerate precisely when the polytope @xmath1 is simple . in the latter case , @xmath147 is a smooth @xmath233-dimensional manifold .
furthermore , the manifold @xmath147 is diffeomorphic to the moment - angle complex @xmath234 . in particular
, the diffeomorphism type of @xmath147 depends only on the combinatorial type of @xmath1 .
we shall therefore not distinguish between @xmath147 and @xmath234 and refer to it as the _ moment - angle manifold _ corresponding to a simple polytope @xmath1 .
the details of these constructions can be found in ( * ? ? ?
* chapter 6 ) .
the standard coordinatewise action of the @xmath32-torus @xmath235 on @xmath236 or @xmath232 induces the _ canonical _ @xmath235-action on @xmath226 or @xmath147 .
we consider ( co)homology with coefficients in @xmath210 .
denote by @xmath237 $ ] the exterior algebra on @xmath32 generators over @xmath210 which satisfy the relations @xmath238 and @xmath239 .
the _ koszul complex _ ( or the _ koszul algebra _ ) of the face ring @xmath213 $ ] is defined as the differential @xmath240-graded algebra @xmath241\otimes\k[\sk],d)$ ] , where @xmath242 cohomology of @xmath241\otimes\k[\sk],d)$ ] is the _
@xmath244}(\k[\sk],\k)$ ] .
it also inherits a @xmath240-grading .
[ zkcoh]there are isomorphisms of ( multi)graded commutative algebras @xmath245}\bigl(\k[\sk],\k\bigr)\\ & \cong h\bigl(\lambda[u_1,\ldots , u_m]\otimes\k[\sk],d\bigr).\end{aligned}\ ] ] the cohomology of @xmath226 therefore acquires a multigrading , with the multigraded and ordinary graded components of @xmath246 given by @xmath247}\bigl(\k[\sk],\k\bigr ) , \quad h^\ell({\mathcal z_{\mathcal k}})=\bigoplus_{-i+2|j|=\ell}h^{-i,2j}({\mathcal z_{\mathcal k}}),\ ] ] where @xmath248 and @xmath249 .
the koszul algebra @xmath241\otimes\k[\sk],d\bigr)$ ] is infinitely generated as a @xmath210-module .
we define its quotient algebra @xmath250\otimes\k[\sk]\bigr/(v_i^2=u_iv_i=0,\ ; 1{\leqslant}i{\leqslant}m)\ ] ] with the induced multigrading and differential .
note that @xmath251 has a finite @xmath210-basis . passing to @xmath251
does not change the cohomology .
this can be proved either algebraically ( * ? ? ?
* lemma 3.2.6 ) or using the following topological interpretation : [ cellappr ] the algebra @xmath251 coincides with the cellular cochains of @xmath226 for the appropriate cell structure .
in particular , there is an isomorphism of cohomology algebras @xmath252 the multigraded component @xmath253 is zero unless all coordinates of the vector @xmath254 are @xmath174 or @xmath51 , and the same is true for the multigraded cohomology @xmath255 .
we can identify subsets @xmath256 $ ] with vectors @xmath257 .
given @xmath258 $ ] , we denote by @xmath259 the monomial @xmath260 $ ] , and similarly consider exterior monomials @xmath261 $ ] .
we also use the notation @xmath262 for the monomial @xmath263 in the koszul algebra @xmath237\otimes\k[\sk]$ ]
. then @xmath251 has a finite @xmath210-basis consisting of monomials @xmath262 where @xmath256 $ ] , @xmath196 and @xmath264 .
given @xmath256 $ ] , define the corresponding _ full subcomplex _ of @xmath192 as @xmath265 consider simplicial cochains @xmath266 with coefficients in @xmath210 .
let @xmath267 be the basis cochain corresponding to an oriented simplex @xmath268 ; it takes value @xmath51 on @xmath269 and vanishes on all other simplices .
define a @xmath210-linear map @xmath270 where @xmath271 is the sign given by @xmath272 and @xmath273 if @xmath274 is the @xmath275th element of the set @xmath256 $ ] written in increasing order . [ htor ] the maps combine to an isomorphism of cochain complexes @xmath276 and induce an isomorphism @xmath277}^{-i,2j } \bigl(\k[\sk],\k\bigr),\ ] ] where @xmath278 denotes the @xmath44th reduced simplicial cohomology group of @xmath279 .
[ prodfsc ] there are isomorphisms of @xmath210-modules @xmath280 } \widetilde h^{\ell-|j|-1}(\sk_j).\ ] ] these isomorphisms combine to form a ring isomorphism @xmath281 } \widetilde h^*(\sk_j),\ ] ] where the ring structure on the right hand side is given by the canonical maps @xmath282 which are induced by the simplicial maps @xmath283 for @xmath284 and zero otherwise .
[ 3dimcoh ] the @xmath7-dimensional cohomology @xmath285 is freely generated by the cohomology classes @xmath286=[u_jv_i]$ ] corresponding to pairs of vertices @xmath287 such that @xmath288 . if @xmath218 for a simple polytope @xmath1 , then these @xmath7-dimensional cohomology classes correspond to pairs of non - adjacent facets @xmath289 .
[ exacochain ] let @xmath290 be the union of two segments .
then nontrivial integral cohomology groups of @xmath226 are given below together with the cocycles in the algebra @xmath251 representing generators : @xmath291 cochains in @xmath292 are functions on the vertices of @xmath192 , and cocycles are functions which are constant on the connected components of @xmath192 . in our case ,
the cocycle @xmath293 represents a generator of @xmath294 .
it is mapped by to the cocycle @xmath295 representing a generator of @xmath296 .
moment - angle complexes @xmath226 may have nontrivial triple massey products of 3-dimensional cohomology classes .
first examples ( found by baskakov @xcite ) appear already for moment - angle manifolds corresponding to 3-polytopes ( see also @xcite ) . a complete description of the triple massey product @xmath297 is given by the following result of denham and suciu : [ desu ] the following are equivalent
: * there exist cohomology classes @xmath298 for which the massey product @xmath299 is defined and non - trivial ; * the graph @xmath204 contains an induced subgraph isomorphic to one of the five graphs in figure [ 5gr ] .
( 120,20 ) ( 0,0 ) ( 10,5 ) ( 15,5 ) ( 10,10 ) ( 15,10 ) ( 10,20 ) ( 0,0)(3,1)15 ( 0,0)(2,1)10 ( 0,0)(1,1)10 ( 0,0)(1,2)10 ( 10,5)(1,0)5 ( 10,5)(0,1)15 ( 10,10)(1,0)5 ( 15,10)(-1,2)5 ( 15,5)(0,1)5 ( 25,0 ) ( 35,5 ) ( 40,5 ) ( 35,10 ) ( 40,10 ) ( 35,20 ) ( 25,0)(3,1)15 ( 25,0)(2,1)10 ( 25,0)(1,2)10 ( 35,5)(1,0)5 ( 35,5)(0,1)15 ( 35,10)(1,0)5 ( 40,10)(-1,2)5 ( 40,5)(0,1)5 ( 50,0 ) ( 60,5 ) ( 65,5 ) ( 60,10 ) ( 65,10 ) ( 60,20 ) ( 50,0)(3,1)15 ( 50,0)(2,1)10 ( 50,0)(1,1)10 ( 50,0)(1,2)10 ( 60,5)(1,0)5 ( 60,5)(0,1)15 ( 65,10)(-1,2)5 ( 65,5)(0,1)5 ( 75,0 ) ( 85,5 ) ( 90,5 ) ( 85,10 ) ( 90,10 ) ( 85,20 ) ( 75,0)(3,1)15 ( 75,0)(2,1)10 ( 75,0)(1,2)10 ( 85,5)(1,0)5 ( 85,5)(0,1)15 ( 90,10)(-1,2)5 ( 90,5)(0,1)5 ( 100,0 ) ( 110,5 ) ( 115,5 ) ( 110,10 ) ( 115,10 ) ( 110,20 ) ( 100,0)(2,1)10 ( 100,0)(1,2)10 ( 110,5)(1,0)5 ( 110,5)(0,1)15 ( 115,10)(-1,2)5 ( 115,5)(0,1)5 let @xmath1 be a simple @xmath26-polytope with the dual simplicial complex @xmath206 . the existence of a characteristic matrix for @xmath1 is equivalent to a choice of @xmath26 linear forms @xmath300 such that @xmath301 $ ] is a finitely generated free module over @xmath302 $ ] . such @xmath303 form a linear _ regular sequence _ in @xmath301 $ ] .
this implies that @xmath304 $ ] is a _
cohen macaulay ring _ over any @xmath210 , but
the condition is actually stronger , as it assumes the existence of a _ linear _ regular sequence over @xmath169 ( and hence over any finite field ) .
given a characteristic matrix ( or a linear regular sequence in @xmath301 $ ] ) , one can define the corresponding homomorphism of tori @xmath305 .
its kernel @xmath306 is an @xmath307-dimensional subtorus in @xmath235 that acts _ freely _ on @xmath147 .
the quotient @xmath308 can be identified with the quasitoric manifold @xmath130 from construction [ canqt ] . as @xmath147 is a smooth intersection of quadrics and
the action of torus is smooth , we obtain a canonical smooth structure on @xmath130 as in @xcite .
we say that @xmath86-manifolds @xmath2 and @xmath3 are _ weakly equivariantly diffeomorphic _ if there is a diffeomorphism @xmath309 and an automorphism @xmath310 such that @xmath311 for any @xmath312 and @xmath142 .
the following result is immediate .
[ qteqd ] if characteristic pairs @xmath115 and @xmath134 are equivalent , then the corresponding quasitoric manifolds @xmath130 and @xmath139 are weakly equivariantly diffeomorphic .
the general homological properties of regular sequences imply yet another description of the cohomology of @xmath147 : [ zkcohred ] let @xmath1 be a simple @xmath26-polytope with @xmath32 facets , and assume there exists a linear integral regular sequence
. denote by @xmath313 the ideal in @xmath220 $ ] generated by @xmath303 .
then there is an isomorphism of cohomology rings @xmath314/\mathcal j}\bigl(\z[\sk_p]/\mathcal j,\z\bigr).\ ] ] note that @xmath301/\mathcal j$ ] is the cohomology ring of the quasitoric manifold @xmath130 , see theorem [ cohomqtoric ] .
the theorem above implies that the spectral sequence of the principal @xmath315-fibration @xmath316 degenerates at the @xmath317 term .
we continue to consider cohomology with coefficients in a commutative ring with unit @xmath210 .
when @xmath210 is not specified explicitly , we assume @xmath318 .
[ defcohorig ] we say that a family of closed manifolds is _ cohomologically rigid _ over @xmath319 if manifolds in the family are distinguished up to homeomorphism by their cohomology rings with coefficients in @xmath319 . that is , a family is cohomologically rigid if a graded ring isomorphism @xmath320 implies a homeomorphism @xmath321 whenever @xmath322 and @xmath323 are in the family .
there is a homotopical and smooth versions of cohomological rigidity , with homeomorphisms replaced by homotopy equivalences and diffeomorphisms , respectively . in toric topology , cohomological rigidity
is studied for ( quasi)toric manifolds and moment - angle manifolds .
we refer to @xcite , @xcite and @xcite for a more detailed survey of related results and problems .
the main question here is as follows .
[ crqtm ] let @xmath322 and @xmath323 be two toric manifolds with isomorphic cohomology rings .
are they homeomorphic ?
in other words , is the family of toric manifolds cohomologically rigid ?
one can ask the same question for quasitoric and topological toric manifolds , and with homeomorphisms replaced by diffeomorphisms . the problem is solved positively for some particular families of toric and quasitoric manifolds , such as cohomologically trivial bott towers ( whose corresponding polytopes are combinatorial cubes ) @xcite , @xmath324-cohomologically trivial bott towers @xcite , @xmath16-cohomologically trivial bott towers @xcite , bott towers of real dimension up to 8 @xcite , quasitoric manifolds over a product of two simplices @xcite and over some dual cyclic polytopes @xcite . the problem is open for general bott towers , and for ( quasi)toric manifolds of real dimension 6 ( over 3-dimensional polytopes ) .
the latter case is the subject of this paper : we give a solution for a particular class of 3-polytopes .
there is also a cohomological rigidity problem for real toric objects , such as real toric manifolds , small covers , and real topological toric manifolds @xcite , with @xmath16-cohomology rings .
this problem is solved positively for real bott towers @xcite , @xcite , but negatively in some other cases @xcite .
cohomological rigidity is also open for moment - angle manifolds , in both graded and bigraded versions : [ crmam ] let @xmath325 and @xmath326 be two moment - angle manifolds whose ( bigraded ) cohomology rings are isomorphic .
are they homeomorphic ?
in other words , is the family of moment - angle manifolds cohomologically rigid ?
a homeomorphism of two quasitoric manifolds over @xmath327 and @xmath328 , or a homeomorphism of moment - angle manifolds @xmath325 and @xmath326 does not imply that the polytopes @xmath327 and @xmath328 are combinatorially equivalent , as shown by the next example .
[ vtpol ] a _ vertex truncation
_ operation ( * ? ? ?
* construction 1.1.1 ) can be applied to a simple polytope @xmath1 to produce a new simple polytope @xmath329 with one more facet .
if one applies this operation iteratively , then the combinatorial type of the resulting polytope depends , in general , on the choice and order of truncated vertices .
for example , applying this operation three times to a 3-simplex one can produce three combinatorially different polytopes @xmath330 , @xmath150 , with @xmath331 facets each ( their dual simplicial polytopes are known as _ stacked _ ) . the corresponding moment - angle manifolds @xmath332 are diffeomorphic , see @xcite .
the polytopes @xmath330 have delzant realisations such that the correponding toric manifolds @xmath333 are obtained from @xmath334 by blowing it up three times in three different ways .
each of @xmath333 is diffeomorphic to a connected sum of 4 copies of @xmath334 ( see the details in @xcite ) .
one can look for classes of simple polytopes @xmath1 whose combinatorial type is determined by the cohomology ring of any ( quasi)toric manifold over @xmath1 or by the cohomology ring of the moment - angle manifold @xmath147 .
this leads to the following two notions of rigidity for simple polytopes , considered in @xcite and @xcite respectively .
a simple polytope @xmath1 is said to be _ c - rigid _ if any of the two conditions hold : * there are no quasitoric manifolds @xmath2 over @xmath1 ( equivalently , there are no linear regular sequences in @xmath301 $ ] ) , or * whenever there exist a quasitoric manifold @xmath2 over @xmath1 and a quasitoric manifold @xmath3 over another polytope @xmath4 with a cohomology ring isomorphism @xmath5 , there is a combinatorial equivalence @xmath335 .
we say that a property of simple polytopes is _ c - rigid _ if for any ring isomorphism @xmath5 , both @xmath1 and @xmath4 either have or do not have the property . a simple polytope @xmath1 is said to be _ b - rigid _ if any cohomology ring isomorphism @xmath336 of moment - angle manifolds implies a combinatorial equivalence @xmath335 .
we say that a property of simple polytopes is _ b - rigid _ if for any ring isomorphism @xmath336 , both @xmath1 and @xmath4 either have or do not have the property .
according to example [ vtpol ] , a truncated simplex with at least 3 truncations ( the dual to a stacked polytope with at least 3 stacks ) is neither c - rigid nor b - rigid .
previously known examples of c - rigid polytopes include products of simplices and their single vertex truncations @xcite , as well as a product of a simplex and a polygon @xcite .
also , c - rigidity was determined in @xcite for all simple 3-polytopes with @xmath337 facets .
the following relation between the two notions of rigidity can be extracted from the results of @xcite : [ bcrig ] if a simple polytope @xmath1 is b - rigid , then it is c - rigid
. assume that we have a cohomology ring isomorphism @xmath338 for quasitoric manifolds @xmath2 over @xmath1 and @xmath3 over @xmath4 .
we need to show that it implies a ring isomorphism @xmath339 , as the latter would give @xmath335 by b - rigidity .
let @xmath313 and @xmath340 denote the corresponding ideals in @xmath301 $ ] and @xmath341 $ ] , respectively , generated by the linear regular sequences .
then we have a ring isomorphism @xmath342/\mathcal j\stackrel\cong\longrightarrow\z[\sk_{p'}]/\mathcal j'$ ] .
we need to show that this isomorphism gives rise to a ring isomorphism @xmath343/\mathcal j}\bigl(\z[\sk_p]/\mathcal j,\z\bigr)\stackrel\cong\longrightarrow { \mathop{\rm tor}\nolimits}_{\z[v_1,\ldots , v_{m'}]/\mathcal j'}\bigl(\z[\sk_{p'}]/\mathcal j',\z\bigr),\ ] ] as the latter is nothing but an isomorphism @xmath344 according to theorem [ zkcohred ] .
this is is proved in ( * ? ? ?
* lemma 3.7 ) .
namely , the isomorphism @xmath342/\mathcal j\stackrel\cong\longrightarrow\z[\sk_{p'}]/\mathcal j'$ ] can be extended to a commutative diagram @xmath345/\mathcal j \ar[r]^{\cong } \ar[d ] & \z[v_1,\ldots , v_{m'}]/\mathcal j'\ar[d]\\ \z[\sk_p]/\mathcal j \ar[r]_\cong^\varphi & \z[\sk_{p'}]/\mathcal j ' , } \ ] ] implying in particular that @xmath346 .
the commutative diagram above gives rise to an isomorphism by the standard properties of @xmath243 .
more specifically , the isomorphism @xmath188 gives an isomorphism of the koszul algebras @xmath347/\mathcal j\otimes\z[\sk_p]/\mathcal j , d\bigr)\stackrel\cong\longrightarrow \bigl(\lambda[u'_1,\ldots , u'_m]/\mathcal j'\otimes\z[\sk_{p'}]/\mathcal j',d\bigr),\ ] ] where the ideals in the exterior algebras are defined by the same linear forms as in the face rings .
then is obtained by passing to the cohomology .
the argument above is essentially ( * ? ? ?
* lemma 3.7 ) .
the term `` b - rigidity '' was introduced in the last section of @xcite .
however , the implication of proposition [ bcrig ] was erroneously stated there in the opposite direction : `` if @xmath1 is c - rigid , then it is b - rigid '' .
this was a confusion .
it is not known whether c - rigidity is equivalent to b - rigidity , and it is unlikely to be true in general .
recall that the pogorelov class @xmath0 consists of simple @xmath7-polytopes @xmath1 which are flag and do not have @xmath24-belts ( or , equivalently , of simple @xmath7-polytopes @xmath49 without @xmath7- and @xmath24-belts ) . in this section
we consider combinatorial properties of polytopes @xmath14 and cohomological properties of the corresponding moment - angle manifolds @xmath147 .
the key statements here are theorem [ flagrig ] , theorem [ 4beltrig ] and lemma [ rig3cl ] ; they will be used in the proof of the main results in the next section .
the first property is straightforward : in a polytope @xmath14 , there are no @xmath7-gonal or @xmath24-gonal facets .
[ 3flag ] a simple @xmath7-polytope @xmath1 is flag if and only if each of its facets @xmath348 is surrounded by a @xmath44-belt with @xmath43 .
note that @xmath44 is the number of edges in @xmath348 .
assume that @xmath1 is flag .
let @xmath349 , @xmath43 , be the sequence of facets adjacent to @xmath348 , written in a cyclic order .
the claim is that it forms a @xmath44-belt .
indeed , otherwise there is a pair of facets @xmath289 with @xmath350 and @xmath351 .
then each pair of facets in the triple @xmath352 has nonempty intersection , but @xmath353 .
this contradicts the assumption that @xmath1 is flag .
now assume that @xmath1 is not flag .
then either @xmath48 , or @xmath1 has a @xmath7-belt , and none of the facets in this @xmath7-belt is surrounded by a belt . [ 2facets ] for any two facets @xmath116 and @xmath354 in a polytope @xmath14 , there is a vertex @xmath355 .
take any facet @xmath356 different from @xmath116 and @xmath354
. then @xmath356 has at most two common vertices with @xmath116 and at most two common vertices with @xmath354 . on the other hand
, @xmath356 has at least 5 vertices by the previous corollary .
thus , at least one vertex of @xmath356 does not lie in @xmath357 . in a flag @xmath7-polytope @xmath1 , for any facet @xmath116
there is a facet @xmath354 such that @xmath358 . by lemma [ 3flag ] the facet @xmath116 is surrounded by a @xmath44-belt @xmath359 with @xmath43 .
then @xmath360 consists of two connected components : one of them is the interior of @xmath116 , and the other contains the interior of a facet @xmath354 that we look for .
now we consider cohomology of moment - angle manifolds @xmath147 with coefficients in @xmath169 .
we recall from proposition [ 3dimcoh ] that @xmath361 has a basis of cohomology classes @xmath286=[u_jv_i]$ ] corresponding to pairs of non - adjacent facets @xmath289 .
[ prod3pol ] let @xmath1 be a simple @xmath7-polytope with @xmath32 facets and let @xmath218 be its dual simplicial complex . in the notation of theorem [ prodfsc ]
, we have @xmath362 \widetilde h^{2}(\sk)=\z&\text{for } \ell = m+3,\\ 0&\text{otherwise.}\\ \end{cases}\ ] ] in particular , @xmath363 does not have torsion .
furthermore , all nontrivial products in @xmath363 are of the form @xmath364 or @xmath365\setminus i})\to \widetilde h^2(\sk)=\z.\ ] ] for the multigraded components of @xmath363 , these two cases correspond to @xmath366\setminus i)}({\mathcal z_p})\to h^{-(m-3),\,2[m]}({\mathcal z_p})=\z,\end{aligned}\ ] ] where the latter is the poincar duality pairing .
this follows from theorem [ zkcoh ] , theorem [ htor ] and theorem [ prodfsc ] .
an element in a graded ring is called _ decomposable _ if it can be written as a sum of nontrivial products of elements of nonzero degree .
[ proddeco ] let @xmath1 be a flag @xmath7-polytope and @xmath192 its dual simplicial complex .
then the ring @xmath367 } \widetilde h^*(\sk_j)$ ] is multiplicatively generated by @xmath368 } \widetilde h^0(\sk_j)$ ] . to prove this lemma it is enough to show that each nontrivial cohomology class in @xmath369 is decomposable or , equivalently
, the product map @xmath370 is surjective .
this proof is quite technical .
we include it in appendix [ proofproddeco ] for the reader s convenience .
[ proddeco1 ] a simple @xmath7-polytope @xmath49 with @xmath32 facets is flag if and only if any nontrivial cohomology class in @xmath371 is decomposable . in particular , if @xmath372 then either @xmath1 is flag or @xmath48 .
suppose that @xmath1 is not flag .
since @xmath49 , it has a @xmath7-belt @xmath373 .
equivalently , the dual complex @xmath192 has a missing 3-face @xmath374 .
it gives a nonzero cohomology class @xmath375 .
consider the poincar duality pairing @xmath376 which specifies to @xmath377\setminus j)}({\mathcal z_p})\otimes h^{-1,2j}({\mathcal z_p})\to h^{-(m-3),\,2[m]}({\mathcal z_p})=\z\ ] ] ( see proposition [ prod3pol ] ) .
take @xmath378\setminus j)}({\mathcal z_p})\subset h^{m-2}({\mathcal z_p})$ ] such that @xmath379 is a generator of @xmath380}({\mathcal z_p})=\z$ ] . by theorem [ prodfsc ]
, @xmath381\setminus j)}({\mathcal z_p})=\widetilde h^0(\sk_{[m]\setminus j})$ ] , and any element of @xmath382\setminus j})$ ] is indecomposable by proposition [ prod3pol ] .
we have therefore found an indecomposable element @xmath383 .
now suppose that @xmath1 is flag . by proposition [ prod3pol ] , @xmath384 consider the poincar duality pairing @xmath385\setminus i})\to\z$ ] . since @xmath192 is flag , @xmath386\setminus i})=0 $ ] for @xmath387 ( as there are no missing @xmath22-faces ) .
hence , @xmath388=0 by poincar duality , and @xmath389
. then each nonzero element of @xmath371 is decomposable by lemma [ proddeco ] .
[ flagrig ] let @xmath1 be a flag @xmath7-polytope , and assume given a ring isomorphism @xmath336 for another simple @xmath7-polytope @xmath4 .
then @xmath4 is also flag . in other words , the property of being
a flag @xmath7-polytope is @xmath18-rigid .
this follows from lemma [ proddeco1 ] and the fact that @xmath390 is b - rigid .
theorem [ flagrig ] also follows from ( * ? ? ?
* theorem 6.6 ) . [ massey ]
let @xmath1 be a simple @xmath7-polytope .
* the product @xmath391 is trivial if and only if @xmath1 does not have @xmath24-belts . *
the triple massey product @xmath392 is trivial if @xmath1 does not have @xmath24-belts .
we first prove ( a ) .
suppose @xmath1 has a @xmath24-belt @xmath393 .
it corresponds to a chordless 4-cycle @xmath394 in @xmath218 , i.e. a cycle with @xmath395 and @xmath396 .
hence , we have a nontrivial product @xmath397 , and a nontrivial product @xmath391 .
now suppose there is a nontrivial product @xmath398 .
we have @xmath399 .
elements of @xmath400 are indecomposable .
an element of @xmath401 with @xmath402 can be decomposed into a product if and only if @xmath197 can be split into two pairs of non - adjacent vertices , which means that @xmath197 is a chordless 4-cycle .
it corresponds to a 4-belt in @xmath1 . to prove ( b ) , assume that there is a nontrivial massey product @xmath403 . then , by theorem [ desu ] , the graph @xmath209 contains an induced subgraph isomorphic to one of the five graphs in figure [ 5gr ] . by inspection ,
each of these five graphs has a chordless 4-cycle ( the outer cycle for the first four graphs , and the left cycle for the last one ) .
hence , the polytope @xmath1 has a 4-belt .
it is not known whether moment - angle manifolds of polytopes from the pogorelov class @xmath0 have nontrivial massey products of cohomology classes of dimension @xmath404 or of order @xmath404 , or whether these moment - angle manifolds are formal in the sense of rational homotopy theory . for general polytopes @xmath1 ,
there are examples of nontrivial massey products of any order in @xmath363 , see @xcite .
[ 4beltrig ] let @xmath1 be a simple @xmath7-polytope without @xmath24-belts , and assume given a ring isomorphism @xmath336 for another simple @xmath7-polytope @xmath4 .
then @xmath4 also does not have @xmath24-belts .
it other words , the property of being a @xmath7-polytope without @xmath24-belts is @xmath18-rigid .
this follows from proposition [ massey ] ( a ) .
recently fan , ma and wang proved that any polytope @xmath14 is @xmath18-rigid , see ( * ? ? ?
* theorem 3.1 ) .
the proof builds upon the following crucial lemma : [ rig3cl ] consider the set of cohomology classes @xmath405\in h^3({\mathcal z_p}),\quad f_i\cap f_j=\varnothing\}.\ ] ] if @xmath14 , then for any cohomology ring isomorphism @xmath339 , we have @xmath406 .
note that the lemma above does not hold for all simple @xmath7-polytopes .
for example , if @xmath1 is a 3-cube with the pairs of opposite facets @xmath407 , @xmath408 , @xmath409 , then @xmath410 and there is an isomorphism @xmath411 which maps @xmath412 $ ] to @xmath412+[u_2v_5]$ ] .
we include the proof of lemma [ rig3cl ] in appendix [ proofrig3cl ] for the reader s convenience , and also because some details were missing in the original argument .
note that this proof uses theorem [ flagrig ] and theorem [ 4beltrig ] .
here we prove the cohomological rigidity for small covers and quasitoric manifolds over @xmath7-polytopes from the pogorelov class @xmath0 .
we start with a crucial lemma .
[ crid2cl ] in the notation of theorem [ cohomqtoric ] , consider the set of cohomology classes @xmath413\in h^2(m),\quad i=1,\ldots , m\}.\ ] ] if @xmath1 is a @xmath7-polytope from the pogorelov class @xmath0 and @xmath3 is a quasitoric manifold over @xmath4 , then for any cohomology ring isomorphism @xmath414 we have @xmath415 .
moreover , the set @xmath416 is mapped bijectively under @xmath188 .
the idea is to reduce the statement to lemma [ rig3cl ] .
the ring isomorphism @xmath188 is determined uniquely by the isomorphism @xmath417 of free abelian groups .
let @xmath418)=\sum_{j=1}^m a_{ij}[v'_j]$ ] for some @xmath419 , @xmath420 .
the elements @xmath421 are not defined uniquely as there are linear relations between the classes @xmath422 $ ] in @xmath423 .
to get rid of this indeterminacy , one can choose a vertex @xmath424 of @xmath1 and a vertex @xmath425 of @xmath4 .
then the complementary cohomology classes @xmath68 $ ] with @xmath426 form a basis in @xmath427 and the cohomology classes @xmath428 $ ] with @xmath429 form a basis in @xmath423 , so we have @xmath430)=\sum_{p\notin\{p_1,p_2,p_3\}}^m b_{ip}[v'_p ] , \qquad i\in[m]\setminus\{i_1,i_2,i_3\},\ ] ] with uniquely defined @xmath431 , @xmath432\setminus\{i_1,i_2,i_3\}$ ] , @xmath433\setminus\{p_1,p_2,p_3\}$ ] .
as we have seen in the proof of proposition [ bcrig ] , the isomorphism @xmath188 gives an isomorphism @xmath434 , which is obtained from by passing to the cohomology .
we write as @xmath435 .
this isomorphism is defined on the exterior generators @xmath436 and the polynomial generators @xmath437 of the koszul algebra @xmath438 by the same formulae as @xmath188 .
now take a cohomology class @xmath286\in h^3({\mathcal z_p})$ ] . by lemma [ rig3cl ]
, it is mapped under @xmath439 to an element @xmath440\in h^3(\mathcal z_{p'})$ ] , @xmath441 .
choose vertices @xmath424 of @xmath1 and @xmath442 of @xmath4 such that @xmath443 and @xmath444 ( see lemma [ 2facets ] ) .
we use the vertices @xmath123 and @xmath445 to choose bases in the groups @xmath427 and @xmath423 as described in the first paragraph of the proof .
then we have @xmath446=\sum_{p , q\notin\{p_1,p_2,p_3\}}b_{ip}b_{jq}[u_p'v_q'].\ ] ] on the other hand , we have @xmath447=\varepsilon[u_k'v_l']$ ] by lemma [ rig3cl ] .
hence , @xmath448 is a coboundary in @xmath449 , so there exists @xmath450 we have @xmath451 comparing this with the expression for @xmath452 we obtain the following relations between the coefficients : @xmath453 from the third equation of we have , for any @xmath454\setminus\{p_1,p_2,p_3\}$ ] , either @xmath455 or @xmath456 .
the first equation of implies that for @xmath457 the vectors @xmath458 and @xmath459 are linearly dependent .
hence , for @xmath457 , either one of the vectors @xmath460 and @xmath461 is zero , or both vectors are nonzero and have a zero entry on the same place . from the second equation
@xmath462 we see that both vectors @xmath463 and @xmath464 are nonzero .
if there is a nonzero vector @xmath465 for some @xmath466 , then by considering the pairs @xmath467 and @xmath468 we see that both @xmath469 and @xmath470 have zero on the same place , which contradicts the second equation of .
it follows that @xmath471 for any @xmath466 , and @xmath472 since all entries are integer , we have @xmath473 and @xmath474
. then gives @xmath418)\in\{\pm[v_k'],\pm [ v_l']\}$ ] .
it follows that @xmath475 .
it remains to show that the set @xmath416 is mapped bijectively under @xmath188 .
for this we note that @xmath68\ne\pm[v_j]$ ] in @xmath427 for any @xmath107 .
indeed , by lemma [ 2facets ] we can choose a vertex @xmath355 .
then both @xmath68 $ ] and @xmath476 $ ] belong to a basis of @xmath427 .
now , since @xmath188 is an isomorphism , we also have @xmath418)\ne\pm\varphi([v_j])$ ] in @xmath423 . thus , both sets @xmath416 and @xmath477 consist of @xmath478 elements and @xmath415 .
it follows from the steinitz theorem that any toric manifold of complex dimension @xmath7 is a quasitoric manifold .
also , the family of quasitoric manifolds agrees with that of topological toric manifolds in real dimension @xmath8 if we forget the actions .
now we state the first main result .
[ 3cohrig ] let @xmath479 and @xmath480 be quasitoric manifolds over @xmath7-dimensional simple polytopes @xmath1 and @xmath4 , respectively .
assume that @xmath1 belongs to the pogorelov class @xmath0 .
then the following conditions are equivalent : * there is a cohomology ring isomorphism @xmath481 ; * there is a diffeomorphism @xmath6 ; * there is an equivalence of characteristic pairs @xmath482 .
the implication ( b)@xmath483(a ) is obvious .
the implication ( c)@xmath483(b ) follows from proposition [ qteqd ] .
we need to prove ( a)@xmath483(c ) . by lemma [ crid2cl ] , @xmath418)=\pm[v'_{\sigma(i)}]$ ] , where @xmath484 is a permutation of the set @xmath193 $ ] . renumbering the facets and multiplying the matrix @xmath80 from the right by a matrix @xmath18 as in definition [ defcharpair ]
, we may assume that @xmath418)=v_i'$ ] ; this does not change the equivalence class of the pair @xmath115
. then @xmath485=[v'_iv'_j]$ ] .
by corollary [ viprod ] , @xmath486=0 $ ] in @xmath21 if and only if @xmath487 and @xmath488=0 $ ] in @xmath21 if and only if @xmath489 in @xmath1 , and the same holds for @xmath490 and @xmath4 .
it follows that @xmath206 is isomorphic to @xmath491 .
hence , @xmath1 and @xmath4 are combinatorially equivalent .
now consider the @xmath492-matrices @xmath80 and @xmath493 .
first , by changing the order of facets in @xmath1 and @xmath4 if necessary we may assume that @xmath494 in @xmath1 and @xmath495 in @xmath4 .
then , by multiplying the matrices @xmath80 and @xmath493 from the left by appropriate matrices from @xmath496 we may assume that @xmath497 this does not change the equivalence class of pairs @xmath115 and @xmath134 .
now the entries @xmath498 , @xmath499 , are the coefficients in the expression of @xmath476 $ ] , @xmath500 , via the basis @xmath501,\ldots,[v_m]$ ] of @xmath427 .
the same holds for the @xmath502 . since @xmath418)=v_i'$ ]
, it follows that @xmath503 .
thus , the pairs @xmath115 and @xmath134 are equivalent . any smooth structure on a quasitoric manifold @xmath2 over a polytope @xmath14 is equivalent to the standard one defined on the canonical model @xmath130 via proposition [ qteqd ] .
this follows from the general classification results for @xmath8-dimensional manifolds , see corollary [ homeodif ] .
[ coro:5 - 1 ] toric , quasitoric and topological toric manifolds over polytopes from the pogorelov class @xmath0 are cohomologically rigid .
the family of quasitoric ( or topological toric ) manifolds over @xmath7-polytopes from the pogorelov class @xmath0 is large enough , as there is at least one quasitoric manifold over any such polytope by proposition [ 4cprop ] ( recall that it uses the four colour theorem ) .
there are fewer toric manifolds in this family .
in fact , there are no _ projective _ toric manifolds over combinatorial flag 3-polytopes without 4-belts .
the reason is that a delzant 3-polytope must have at least one triangular or quadrangular face by the result of c.delaunay @xcite ( see also @xcite ) . on the other hand , there are many non - projective toric manifolds in this family , see @xcite . in general , a ( non - equivariant ) diffeomorphism @xmath6 does not imply an equivalence of characteristic pairs .
for example consider the family of toric manifolds ( _ hirzebruch surfaces _ )
@xmath504 , where @xmath505 is the @xmath44th power of the canonical line bundle over @xmath506 , @xmath507 is a trivial line bundle , and @xmath508 denotes the complex projectivisation .
then each @xmath509 is a quasitoric manifold over a combinatorial 4-gon with characteristic matrix @xmath510 manifolds @xmath509 with even @xmath44 are all diffeomorphic to @xmath511 , but the characteristic matrices corresponding to different positive @xmath44 are not equivalent .
similar examples exist in all dimensions , see e.g. @xcite .
our second main result is about small covers ( or hyperbolic @xmath7-manifolds ) .
[ 3cohrigreal ] let @xmath512 and @xmath513 be small covers of @xmath7-dimensional simple polytopes @xmath1 and @xmath4 , respectively .
assume that @xmath1 belongs to the pogorelov class @xmath0 , so @xmath12 is a hyperbolic @xmath7-manifold of lbell type .
then the following conditions are equivalent : * there is a cohomology ring isomorphism @xmath514 ; * there is a diffeomorphism @xmath515 ; * there is an equivalence of @xmath16-characteristic pairs @xmath482 .
we only need to prove the implication ( a)@xmath483(c ) .
using proposition [ lambda2 ] we upgrade @xmath115 and @xmath134 to @xmath169-characteristic pairs and consider the corresponding quasitoric manifolds @xmath479 and @xmath480 . since the cohomology ring @xmath516 is obtained from @xmath160 by doubling the grading ( see theorem [ cohomsc ] ) , we have an isomorphism @xmath517 . now
the equivalence of characteristic pairs follows from theorem [ 3cohrig ] ( with coefficients in @xmath16 ) .
the classification of smooth simply connected @xmath8-dimensional manifolds with torsion - free homology was done in the works of wall @xcite and jupp @xcite .
they also stated a result in the topological category , whose proof was corrected in the work of zhubr @xcite .
the latter work also treated the case of homology with torsion .
we only give the following result , which will be enough for our purposes ( the cohomology is with integer coefficients , unless otherwise specified ) .
[ class1 ] let @xmath518 be an isomorphism of the cohomology rings of smooth closed simply connected @xmath8-dimensional manifolds @xmath519 with @xmath520 .
suppose that * @xmath521 , where @xmath522 is the second stiefel whitney class ; * @xmath523 , where @xmath524 is the first pontryagin class .
then the manifolds @xmath12 and @xmath15 are diffeomorphic .
the following lemma is proved using steenrod squares : [ steen ] suppose that the ring @xmath160 is generated by @xmath525 for some @xmath526 .
then any cohomology ring isomorphism @xmath527 preserves the total stiefel whitney class , i.e. @xmath528 . lemma [ steen ] applies to toric or quasitoric manifolds , whose cohomology is generated in degree two . from theorem [ class1 ]
we obtain let @xmath529 be an isomorphism of second cohomology groups of @xmath8-dimensional smooth quasitoric manifolds .
suppose that * @xmath188 preserves the cubic form @xmath530 given by the cohomology multiplication ; * @xmath188 preserves the first pontryagin class
. then the manifolds @xmath2 and @xmath3 are diffeomorphic . from the topological invariance of rational pontryagin classes ( proved in general by s.p .
novikov ) we obtain [ homeodif ] let @xmath2 and @xmath3 be @xmath8-dimensional smooth quasitoric manifolds . if @xmath2 and @xmath3 are homeomorphic , then they are diffeomorphic
. the characteristic classes of quasitoric manifolds are given as follows : in the notation of theorem [ cohomqtoric ] , the total stiefel whitney and pontryagin classes of a quasitoric manifold @xmath2 are given by @xmath531 in particular , @xmath532 , and @xmath533 .
a family of @xmath8-dimensional quasitoric manifolds is cohomologically rigid if any cohomology ring isomorphism between manifolds from the family preserves the first pontryagin class .
this reduces cohomological rigidity for @xmath8-dimensional quasitoric manifolds @xmath2 to a problem of combinatorics and linear algebra , as both the cohomology ring @xmath21 and the first pontryagin class @xmath533 are defined entirely in terms of the characteristic pair @xmath115 .
our result on cohomological rigidity for quasitoric manifolds over polytopes from the pogorelov class @xmath0 ( theorem [ 3cohrig ] ) gives a complete classification for this particular class of simply connected @xmath8-manifolds , and its proof is indepenent of the general classification results of wall and jupp .
the invariance of the first pontryagin class for quasitoric manifolds over polytopes from the pogorelov class follows directly from lemma [ crid2cl ]
. it would be interesting to find a direct ( combinatorial ? ) proof of this fact .
bott towers ( of any dimension ) form another family of toric manifolds for which the invariance of pontryagin classes is known , see @xcite
. in dimension @xmath24 we have the identity @xmath534\rangle=3\mathop{\mathrm{sign}}(m)$ ] , where @xmath535\in h_4(m)$ ] is the fundamental class and @xmath536 is the signature of @xmath2 .
therefore , @xmath23 is invariant under cohomology ring isomorphisms .
when @xmath2 is a toric manifold , the signature is equal to @xmath537 , where @xmath32 is the number of vertices in the quotient polygon @xmath1 ( see e.g. ( * ? ? ?
* example 9.5.3 ) ) . the identity @xmath534\rangle=3\mathop{\mathrm{sign}}(m)$ ] then becomes @xmath538\rangle=12 - 3m,\ ] ] which can be seen directly from theorem [ cohomtoric ] .
here we give proofs of two combinatorial lemmata on belts in flag 3-polytopes , originally due to @xcite and @xcite respectively .
these proofs are included mainly for the sake of completeness , but we also fill in some details missing in the original works .
lemma [ fab ] is used in the proof of the product decomposition lemma in appendix [ proofproddeco ] , while lemma [ fabc ] is used in the proof of b - rigidity of the set of canonical generators of @xmath361 in appendix [ proofrig3cl ] .
recall that a belt in a simple polytope @xmath1 corresponds to a chordless cycle in the dual simplicial complex @xmath206 , or to a full subcomplex @xmath539 isomorphic to the boundary of a polygon .
[ fab ] let @xmath1 be a flag simple @xmath7-polytope .
then for every three facets @xmath116 , @xmath540 , @xmath541 with @xmath542 , there exists a belt @xmath543 such that @xmath544 and @xmath545 .
we reformulate this lemma in the dual notation ; this is how the lemma was stated and proved in @xcite : [ fabd ] let @xmath192 be a flag triangulation of the disk @xmath546 with @xmath32 vertices , and let @xmath547 be the set of vertices of the boundary @xmath548 .
assume that @xmath549 .
then for every missing face @xmath550 of @xmath192 there exists a subset @xmath203 $ ] such that @xmath551 and @xmath552 is a chordless cycle ( boundary of a polygon ) . to obtain lemma [ fab ] from lemma [ fabd ] we take @xmath553\setminus\{k\}}$ ] , the simplicial complement to the vertex of @xmath206 corresponding to the facet @xmath554
. then @xmath192 is a flag triangulation of @xmath546 ( as a full subcomplex in the flag complex @xmath206 ) , and @xmath549 because @xmath206 is flag .
lemma [ fabd ] gives a chordless cycle @xmath197 in @xmath555 , which corresponds to the required belt in @xmath1 .
the _ star _ and _ link _ of a vertex @xmath556 are the subcomplexes @xmath557 we use the induction on @xmath32 , the number of vertices of @xmath192 . since @xmath192 is flag , @xmath548 has at least @xmath24 vertices , that is , @xmath558 and @xmath559 .
if @xmath560 , then @xmath561 and @xmath192 is the cone over a square
. then we can take @xmath562 , as every missing face of @xmath192 is contained in the chordless boundary cycle @xmath563 .
now assume that the statement holds for simplicial complexes with @xmath564 vertices .
if both vertices @xmath119 and @xmath565 lie in @xmath548 , then @xmath562 is the required chordless cycle . hence , we only need to consider the case @xmath566 .
if @xmath567 , then there is no missing face @xmath550 such that @xmath566 .
hence , @xmath568 .
for a vertex @xmath569 , denote by @xmath570 the number of vertices in @xmath571 .
then @xmath572 for any @xmath569 , since @xmath563 .
i. suppose that there is a vertex @xmath573 such that @xmath574 .
then the set of vertices of @xmath571 is @xmath575 , where @xmath576 and @xmath577 , see figure [ fig : star-4-vertex ] .
( 0,0)(1,-1)(2,0)(1,1)cycle ; ( 1,-1)(1,1 ) ; ( -0.3,0 ) node ; ( 1,-1.3 ) node ; ( 1,1.3 ) node ; ( 2.3,0 ) node ; ( 0 + 4,0)(1 + 4,-1)(2 + 4,0)(1 + 4,1)cycle ; ( 4,0)(6,0 ) ; ( -0.3 + 4,0 ) node ; ( 1 + 4,-1.3 ) node ; ( 1 + 4,1.3 ) node ; ( 2.3 + 4,0 ) node ; \(i ) if there is no vertex @xmath578 such that @xmath579 is an edge of @xmath192 , then the simplicial complex @xmath580\setminus\{j\}}$ ] satisfies the hypothesis of the lemma .
hence , there is a subset @xmath581 of @xmath193\setminus \{j\}$ ] such that @xmath582 and @xmath583 is a chordless cycle .
then @xmath584 is the required set , as @xmath585 .
\(ii ) now assume that there exists a vertex @xmath578 such that @xmath579 is an edge in @xmath192 .
let @xmath586 be the simplicial complex obtained from @xmath192 by applying a bistellar @xmath51-move at @xmath571 , see figure [ fig : star-4-vertex ] .
then @xmath587\setminus\{j\}}$ ] satisfies the hypothesis of the lemma . by induction
, there is a subset @xmath588 of @xmath193\setminus\{j\}$ ] such that @xmath589 and @xmath590 is a chordless cycle . if @xmath591 or @xmath592 is not in @xmath588 , then @xmath593 is the required set .
if both @xmath591 and @xmath592 are in @xmath588 , then @xmath594 is the required set .
suppose that @xmath595 for every @xmath573 .
let @xmath596 , ordered counterclockwise , and assume that @xmath597 .
let @xmath598 denote the set of vertices of @xmath599 , so @xmath600 , for @xmath601 .
note that if @xmath602 , then @xmath603 and @xmath604 .
\(i ) assume that , for some @xmath605 , there is no edge @xmath579 in @xmath192 such that @xmath606 then @xmath607\setminus\{j_p\}}$ ] satisfies the hypothesis of the lemma , so we can find the required subset @xmath197 of @xmath193\setminus\{j_p\}$ ] .
\(ii ) assume that , for every @xmath605 , there is an edge @xmath608 in @xmath192 satisfying @xmath609 for @xmath610 and @xmath611 .
we shall lead this case to a contradiction .
set @xmath612 .
then @xmath613 divides @xmath192 into two simplicial complexes @xmath614 and @xmath615 , where @xmath614 has boundary vertices @xmath616 , and @xmath615 has boundary vertices @xmath617 , see figure [ fig : complex from i_1 ] .
( 1.75,0.75)(2,0.5)(3,0)(3,1.5)(3,3)(2,2.5)(1.75,2.25 ) ; ( 1.5,2)(1.5,1 ) ; ( 3,3)(4,2.5)(4.25,2.25 ) ; ( 4.5,2)(4.5,1 ) ; ( 3,0)(4,0.5)(4.25,0.75 ) ; ( 3,0 ) circle(2pt ) ; ( 3,3 ) circle(2pt ) ; ( 3,1.5 ) circle(2pt ) ; ( 2,0.5 ) circle(2pt ) ; ( 4,0.5 ) circle(2pt ) ; ( 2,2.5 ) circle(2pt ) ; ( 4,2.5 ) circle(2pt ) ; ( 3,-0.3 ) node ; ( 3,3.3 ) node ; ( 3.3,1.5 ) node ; ( 2.25,0.75 ) node ; ( 3.75,0.75 ) node ; ( 2,2.25 ) node ; ( 4,2.25 ) node ; since @xmath566 , either @xmath618 or @xmath619 . without loss of generality , assume that @xmath618 .
then @xmath603 for @xmath620 .
by the flagness of @xmath192 and the condition for the existence of an edge satisfying @xmath609 , there is no vertex @xmath621\setminus s$ ] such that @xmath44 is connected to the vertices @xmath622 and @xmath623 for @xmath624 .
this implies in particular that @xmath625 .
now consider the path from @xmath626 to @xmath627 and to @xmath628 .
if @xmath629 , then we may assume that @xmath630 .
otherwise , @xmath627 must be contained in the simplicial complex @xmath614 . in either case ,
the the path @xmath631 is contained in the simplicial subcomplex @xmath614 with boundary vertices @xmath616 .
proceeding inductively , we obtain that the path @xmath632 is contained in the simplicial subcomplex whose boundary vertices are @xmath633 , see figure [ fig : path ] .
it follows that @xmath634 .
eventually we obtain @xmath635 such that @xmath636 , so the vertex @xmath637 is connected to the vertices @xmath638 and @xmath623 .
this is a contradiction .
( 0.5,2)(1,2.5)(2,3)(4,1.5)(2,0)(1.5,0.25 ) ; ( 0.5,1.25)(1,0.5 ) ; ( 1,2.5 ) circle(2pt ) ; ( 2,3 ) circle(2pt ) ; ( 4,1.5 ) circle(2pt ) ; ( 2,0 ) circle(2pt ) ; ( 0.7,2.5 ) node ; ( 2,3.3 ) node ; ( 4,1.8 ) node ; ( 2,-.3 ) node ; from i and ii , the lemma is proved .
[ fabc ] let @xmath1 be a flag @xmath7-polytope without @xmath24-belts .
then for every three different facets @xmath116 , @xmath540 , @xmath541 with @xmath639 there is a belt @xmath640 such that @xmath641 , @xmath545 , and @xmath541 does not intersect at least one of the two connected components of @xmath642 .
we work with the dual simplicial complex @xmath218 , which is a triangulated 2-sphere .
we need to find a subset @xmath203\setminus\{k\}$ ] such that @xmath551 , @xmath552 is a chordless cycle , and @xmath643 . by lemma [ fabd ] ,
there is a subset @xmath644 of @xmath193\setminus\{k\}$ ] such that @xmath645 and @xmath646 is a chordless cycle .
we construct the required subset @xmath197 by modifying @xmath644 . 0.3 ( 0,0 )
circle(1 cm ) ; ( 0.86,0.5 ) circle(2pt ) ; ( 0.5,0.86 ) circle(2pt ) ; ( -0.5,0.86 ) circle(2pt ) ; ( -0.86,0.5 ) circle(2pt ) ; ( -0.86,-0.5 ) circle(2pt ) ; ( -0.5,-0.86 ) circle(2pt ) ; ( 0.5,-0.86 ) circle(2pt ) ; ( 0,1 ) circle(2pt ) ; ( 0,-1 ) circle(2pt ) ; ( 0,1.2 ) node@xmath119 ; ( 0,-1.2 ) node@xmath565 ; ( 1,0 ) circle(2pt ) ( 0.86,-0.5 ) circle(2pt ) ( -1,0 ) circle(2pt ) ; ( -1.5,0.5 ) node@xmath646 ; 0.3 ( 0,0 ) circle(1 cm ) ; ( 0,0 ) circle(2pt ) ; ( 0.86,0.5 ) circle(2pt ) ; ( 0.5,0.86 ) circle(2pt ) ; ( -0.5,0.86 ) circle(2pt ) ; ( -0.86,0.5 ) circle(2pt ) ; ( -0.86,-0.5 ) circle(2pt ) ; ( -0.5,-0.86 ) circle(2pt ) ; ( 0.5,-0.86 ) circle(2pt ) ; ( 0,1 ) circle(2pt ) ; ( 0,-1 ) circle(2pt ) ; ( 1,0 ) circle(2pt ) ( 0.86,-0.5 ) circle(2pt ) ( -1,0 )
circle(2pt ) ; ( 0,1.2 ) node@xmath119 ; ( 0,-1.2 ) node@xmath565 ; ( -0.5,0.86 ) arc(120:240:1 cm ) ; ( 0.5,-0.86 ) arc(-60:60:1 cm ) ; ( -1.5,0 ) node@xmath647 ; ( 1.5,0 ) node@xmath648 ; ( 0,-0.2 ) node@xmath44 ; ( 0.5,0.3 ) node@xmath649 ; ( 1.5,1 ) node@xmath650 ; 0.3 ( 0,0 ) circle(1 cm ) ; ( 0,0 ) circle(2pt ) ; ( 0.86,0.5 ) circle(2pt ) ; ( 0.5,0.86 ) circle(2pt ) ; ( -0.5,0.86 ) circle(2pt ) ; ( -0.86,0.5 ) circle(2pt ) ; ( -0.86,-0.5 ) circle(2pt ) ; ( -0.5,-0.86 ) circle(2pt ) ; ( 0.5,-0.86 ) circle(2pt ) ; ( 0,1 ) circle(2pt ) ; ( 0,-1 ) circle(2pt ) ; ( 1,0 ) circle(2pt ) ( 0.86,-0.5 ) circle(2pt ) ( -1,0 ) circle(2pt ) ; ( 0,1.2 ) node@xmath119 ; ( 0,-1.2 ) node@xmath565 ; ( -1.5,0 ) node@xmath651 ; ( 1.5,0 ) node@xmath652 ; ( -0.86,0.5 ) circle(5pt ) ; ( -0.86,-0.5 ) circle(5pt ) ; ( 0.86,0.5 ) circle(5pt ) ; ( 1,0 ) circle(5pt ) ; ( 0.5,-0.86 ) circle(5pt ) ; ( 0,-0.2 ) node@xmath44 ; ( 0,0 ) edge ( 0.86,0.5 ) edge ( -0.86,0.5 ) edge ( -0.86,-0.5 ) edge ( 0.5,-0.86 ) edge ( 1,0 ) ; since @xmath646 is a cycle , it divides @xmath192 into two polygons ( triangulated discs ) @xmath649 and @xmath650 with the common boundary @xmath646 .
assume that the vertex @xmath44 is contained in @xmath649 .
the vertices @xmath119 and @xmath565 divide the cycle @xmath646 into two arcs , and we denote by @xmath647 and @xmath648 the sets of vertices in @xmath653 contained in these arcs , so @xmath654 .
we set @xmath655 and @xmath656 , see figure [ fig : i0 ] . if either @xmath651 or @xmath652 is empty , then @xmath657 , so @xmath658 is the required subset . in what follows
we assume that both @xmath651 and @xmath652 are nonempty .
we consider the links of all @xmath659 in @xmath650 . since @xmath646 is a chordless cycle , every such link has at least three vertices , that is , there is a vertex in @xmath660 which is not in @xmath644 . to simplify notation , for @xmath661 $ ] , we write @xmath662 instead of @xmath663 . now define @xmath664 we take the outermost path @xmath665 between @xmath119 and @xmath565 in @xmath666 with respect to the vertex @xmath44 , so that all vertices of @xmath666 not in @xmath665 are on the side of @xmath44 , see figure [ fig : outermost ] .
let @xmath667 be the vertex set of @xmath665 . 0.3 ( 0,0 )
circle(1 cm ) ; ( 0,1 ) coordinate ( v ) ( -0.34,0.93 ) coordinate ( u1 ) ( -0.64,0.76 ) coordinate(u2 ) ( -0.86,0.5 ) coordinate ( u3 ) ( -0.98,0.17 ) coordinate ( u4 ) ( -0.98,-0.17 ) coordinate ( u5 ) ( -0.86,-0.5 ) coordinate ( u6 ) ( -0.64,-0.76 ) coordinate(u7 ) ( -0.34,-0.93 ) coordinate ( u8 ) ( 0,-1 ) coordinate ( vp ) ( 0,0 ) coordinate ( w ) ( -0.8,1.2 ) coordinate ( a ) ( -1.5,0.8 ) coordinate ( b ) ( -1.3,0 ) coordinate ( d ) ( -1.5,-0.9 ) coordinate ( c ) ( -0.17,0.98 ) coordinate ( u0 ) ; ( 0.86,0.5 ) coordinate ( up2 ) ( 0.5,0.86 ) coordinate ( up1 ) ( 0.5,-0.86 ) coordinate ( up6 ) ( 0.86,-0.5 ) coordinate ( up5 ) ( 0.98,-0.17 ) coordinate ( up4 ) ( 0.98,0.17 ) coordinate ( up3 ) ( 0.5,0 ) coordinate ( e ) ( 0.5,-0.5 ) coordinate ( f ) ( 0.75,-0.25 ) coordinate ( g ) ; ( u1 ) circle ( 2pt ) ( u2 ) circle ( 2pt ) ( u3 ) circle ( 2pt ) ( u4 ) circle ( 2pt ) ( u5 ) circle ( 2pt ) ( u6 ) circle ( 2pt ) ( u7 ) circle ( 2pt ) ( a ) circle ( 2pt ) ( b ) circle ( 2pt ) ( c ) circle ( 2pt ) ( d ) circle ( 2pt ) ( up6 ) circle ( 2pt ) ( up5 ) circle ( 2pt ) ( up4 ) circle ( 2pt ) ( up3 ) circle ( 2pt ) ( up2 ) circle ( 2pt ) ( up1 ) circle ( 2pt ) ( u0 ) circle(2pt ) ; ( u8 ) circle ( 2pt ) ( v ) circle ( 2pt ) ( vp ) circle ( 2pt ) ; ( 0,0 ) circle(2pt ) ; ( 0,0 ) circle ( 1 cm ) ; ( 0,1.2 ) node@xmath119 ; ( 0,-1.2 ) node@xmath565 ; ( 0,-0.2 ) node@xmath44 ; ( 0,0 ) edge ( u2 ) edge ( u7 ) edge ( up2 ) edge ( up6 ) ; ( a ) edge ( u0 ) edge ( u1 ) edge ( u2 ) edge ( b ) ; ( b ) edge ( u2 ) edge ( u3 ) edge ( d ) edge ( c ) ; ( c ) edge ( u8 ) edge ( u7 ) edge ( u6 ) edge ( d ) ; ( d ) edge ( u3 ) edge ( u4 ) edge ( u5 ) edge ( u6 ) ; plot coordinates ( u1 ) ( a ) ( b ) ( u3 ) ; plot coordinates ( u6 ) ( c ) ( u8 ) ; 0.3 ( 0,0 ) circle(1 cm ) ; ( 0,1 ) coordinate ( v ) ( -0.34,0.93 ) coordinate ( u1 ) ( -0.64,0.76 ) coordinate(u2 ) ( -0.86,0.5 ) coordinate ( u3 ) ( -0.98,0.17 ) coordinate ( u4 ) ( -0.98,-0.17 ) coordinate ( u5 ) ( -0.86,-0.5 ) coordinate ( u6 ) ( -0.64,-0.76 )
coordinate(u7 ) ( -0.34,-0.93 ) coordinate ( u8 ) ( 0,-1 ) coordinate ( vp ) ( 0.5,-0.86 ) coordinate ( up5 ) ( 0.86 , -0.5 ) coordinate ( up4 ) ( 0,1 ) coordinate ( up3 ) ( 0.86,0.5 ) coordinate ( up2 ) ( 0.5,0.86 ) coordinate ( up1 ) ( -0.8,1.2 ) coordinate ( a ) ( -1.5,0.8 ) coordinate ( b ) ( -1.3,0 ) coordinate ( d ) ( -1.5,-0.9 ) coordinate ( c ) ( -0.17,0.98 ) coordinate ( u0 ) ; ( u1 ) circle ( 2pt ) ( u3 ) circle ( 2pt ) ( u4 ) circle ( 2pt ) ( a ) circle ( 2pt ) ( b ) circle ( 2pt ) ( c ) circle ( 2pt ) ; ( u5 ) circle ( 2pt ) ( u6 ) circle ( 2pt ) ( u8 ) circle ( 2pt ) ( u0 ) circle ( 2pt ) ( u2 ) circle ( 2pt ) ( u7 ) circle ( 2pt ) ; ( 0,1 ) circle(2pt ) ; ( 0,-1 ) circle(2pt ) ; ( 0,1 ) arc ( 90:270:1 cm ) ; ( u2 ) edge ( a ) edge ( b ) ; ( u7 ) edge ( c ) ; ( 0,1.2 ) node@xmath119 ; ( 0,-1.2 ) node@xmath565 ; ( a ) edge ( u0 ) edge ( u1 ) edge ( b ) ; ( b ) edge ( u3 ) edge ( c ) ; ( c ) edge ( u8 ) edge ( u6 ) ; 0.3 ( 0,0 ) circle(1 cm ) ; ( 0,1 ) coordinate ( v ) ( -0.34,0.93 ) coordinate ( u1 ) ( -0.64,0.76 ) coordinate(u2 ) ( -0.86,0.5 ) coordinate ( u3 ) ( -0.98,0.17 ) coordinate ( u4 ) ( -0.98,-0.17 ) coordinate ( u5 ) ( -0.86,-0.5 ) coordinate ( u6 ) ( -0.64,-0.76 ) coordinate(u7 ) ( -0.34,-0.93 ) coordinate ( u8 ) ( 0,-1 ) coordinate ( vp ) ( -0.8,1.2 ) coordinate ( a ) ( -1.5,0.8 ) coordinate ( b ) ( -1.3,0 ) coordinate ( d ) ( -1.5,-0.9 ) coordinate ( c ) ( -0.17,0.98 ) coordinate ( u0 ) ; ( u1 ) circle ( 2pt ) ( u3 ) circle ( 2pt ) ( u4 ) circle ( 2pt ) ( a ) circle ( 2pt ) ( b ) circle ( 2pt ) ( c ) circle ( 2pt ) ; ( u5 ) circle ( 2pt ) ( u6 ) circle ( 2pt ) ( u8 ) circle ( 2pt ) ; ( 0,-1 ) circle(2pt ) ( v ) circle ( 2pt ) ( 0,0 ) circle ( 2pt ) ( u0 ) circle ( 2pt ) ; ( u8 ) arc ( -110:-90:1 cm ) ; ( u2 ) circle ( 2pt ) ( u7 ) circle ( 2pt ) ; ( 0,1 ) arc ( 90:270:1 cm ) ; ( u2 ) edge ( a ) edge ( b ) ; ( u7 ) edge ( c ) ; ( 0,1.2 ) node@xmath119 ; ( 0,-1.2 ) node@xmath565 ( 0,0)[below ] node@xmath44 ; ( a ) edge ( u0 ) edge ( u1 ) edge ( b ) ; ( v ) arc(90:100:1 cm ) ; ( v ) arc(90:100:1 cm ) ; ( a ) edge ( u0 ) edge ( b ) ; ( a ) edge ( u0 ) edge ( b ) ; ( b ) edge ( u3 ) edge ( c ) ; ( b ) edge ( c ) ; ( b ) edge ( c ) ; ( c ) edge ( u8 ) edge ( u6 ) ; ( c ) edge ( u8 ) ; ( c ) edge ( u8 ) ; ( 0,0 ) coordinate ( w ) ( 0.86,0.5 ) coordinate ( up2 ) ( 0.5,0.86 ) coordinate ( up1 ) ( -0.5,0.86 ) coordinate ( u1 ) ( -0.86,0.5 ) coordinate ( u2 ) ( -1,0 ) coordinate ( u3 ) ( -0.86,-0.5 ) coordinate ( u4 ) ( -0.5,-0.86 ) coordinate ( u5 ) ( 0.5,-0.86 ) coordinate ( up6 ) ( 0.86,-0.5 ) coordinate ( up5 ) ( 0.98,-0.17 ) coordinate ( up4 ) ( 0.98,0.17 ) coordinate ( up3 ) ( 0,1 ) coordinate ( v ) ( 0,-1 ) coordinate ( vp ) ( 0.5,0 ) coordinate ( a ) ( 0.5,-0.5 ) coordinate ( b ) ( 0.75,-0.25 ) coordinate ( c ) ; * claim . * the full subcomplex of @xmath192 induced on @xmath667 is the path @xmath665 , i.e. , @xmath668 .
suppose to the contrary that there is a subset @xmath669 of @xmath670 such that @xmath671 is a triangle .
consider the intersection @xmath672 .
note that @xmath673 because @xmath674 is a part of a chordless cycle @xmath646 . we have the following cases , shown in figure [ fig : no - triangle ] . .18 ( 0,0 ) circle ( 1 cm ) ; ( -0.34,0.93 )
coordinate ( uip ) ( -1,0 ) coordinate ( y ) ( -0.86,0.5 ) coordinate ( x ) ( -0.5,0.86 ) coordinate ( u1 ) ( -1.2,1.5 ) coordinate ( z ) ( -0.9,1 ) coordinate ( c ) ; ( x ) node[below]@xmath635 ( y ) node[below]@xmath675 ( z ) node[above]@xmath275 ( uip ) node[below]@xmath123 ; ( z ) edge ( y ) edge ( x ) edge ( uip ) ; ( x ) circle(2pt ) ( y ) circle(2pt ) ( z ) circle(2pt ) ( uip ) circle(2pt ) ; ( uip ) circle ( 5pt ) ; .18 ( 0,0 ) circle ( 1 cm ) ; ( -1,0 ) coordinate ( uip ) ( -0.86,0.5 ) coordinate ( x ) ( -0.5,0.86 ) coordinate ( uiq ) ( -1.4,0.8 ) coordinate ( y ) ( -0.9,1.2 ) coordinate ( z ) ; ( x ) node[below]@xmath635 ( y ) node[above]@xmath675 ( z ) node[above]@xmath275 ( uip ) node[below]@xmath123 ( uiq ) node[below]@xmath445 ; ( x)(y)(z)cycle ; ( x ) circle(2pt ) ( y ) circle(2pt ) ( z ) circle(2pt ) ( uiq ) circle(2pt ) ( uip ) circle(2pt ) ; ( y)(uip ) ; ( z)(uiq ) ; ( uip ) circle ( 5pt ) ; ( uiq ) circle ( 5pt ) ; .18 ( 0,0 ) circle ( 1 cm ) ; ( -0.34,0.93 ) coordinate ( uiq ) ( -1,0 ) coordinate ( x ) ( -0.86,0.5 ) coordinate ( uip ) ( -0.5,0.86 ) coordinate ( u1 ) ( -1.2,1.5 ) coordinate ( y ) ( -0.9,1 ) coordinate ( z ) ; ( x ) node[below]@xmath635 ( y ) node[above]@xmath675 ( z ) node[above]@xmath275 ( uiq ) node[below]@xmath123 ; ( x)(y)(z)cycle ; ( x ) circle(2pt ) ( y ) circle(2pt ) ( z ) circle(2pt ) ( uiq ) circle(2pt ) ; ( y)(uiq ) ; ( z)(uiq ) ; ( uiq ) circle ( 5pt ) ; .18 ( 0,0 ) circle ( 1 cm ) ; ( -0.34,0.93 ) coordinate ( uiq ) ( -1.5,0.5 ) coordinate ( y ) ( -0.86,0.5 ) coordinate ( uir ) ( -1,0 ) coordinate ( uip ) ( -0.9,1.5 ) coordinate ( z ) ( -1.2,0.8 ) coordinate ( x ) ; ( x ) node[below]@xmath635 ( y ) node[below]@xmath675 ( z ) node[above]@xmath275 ( uip ) node[below]@xmath445 ( uiq ) node[below]@xmath676 ( uir ) node[below]@xmath123 ; ( x ) circle(2pt ) ( y ) circle(2pt ) ( z ) circle(2pt ) ( uiq ) circle ( 2pt ) ( uip ) circle(2pt ) ( uir ) circle(2pt ) ; ( x)(y)(z)cycle ; ( z ) edge ( uiq ) ; ( y ) edge ( uip ) ; ( x ) edge ( uir ) ; ( uip ) circle ( 5pt ) ; ( uiq ) circle ( 5pt ) ; ( uir ) circle ( 5pt ) ; .18 ( 0,0 ) circle ( 1 cm ) ; ( -0.34,0.93 ) coordinate ( uiq ) ( -0.8,1.2 ) coordinate ( y ) ( -0.86,0.5 ) coordinate ( uir ) ( -1,0 ) coordinate ( uip ) ( -0.9,1.5 ) coordinate ( z ) ( -1.2,0.8 ) coordinate ( x ) ; ( x ) node[below]@xmath635 ( y ) node[below]@xmath675 ( z ) node[above]@xmath275 ( uiq ) node[below]@xmath445 ( uir ) node[below]@xmath123 ; ( x ) circle(2pt ) ( y ) circle(2pt ) ( z ) circle(2pt ) ( uiq ) circle ( 2pt ) ( uip ) ( uir ) circle(2pt ) ; ( x)(y)(z)cycle ; ( z ) edge ( uiq ) ; ( y ) edge ( uiq ) ; ( x ) edge ( uir ) ; ( uir ) circle ( 5pt ) ; ( uiq ) circle ( 5pt ) ; 1 . if @xmath677 , say @xmath678 , then @xmath635 and @xmath675 are consecutive vertices in @xmath651 , and @xmath275 is in @xmath660 for some @xmath659 .
then , @xmath635 or @xmath675 is on the side of @xmath44 in @xmath679 .
this is a contradiction .
2 . if @xmath680 , say @xmath681 , then @xmath682 and @xmath683 for some @xmath684 .
if @xmath685 , then @xmath635 must be on the side of @xmath44 in @xmath679 , which contradicts the assumption that @xmath686 .
if @xmath128 , then either @xmath675 or @xmath275 is on the side of @xmath44 in @xmath679 , and we obtain a contradiction again .
3 . if @xmath687 , then there are @xmath688 in @xmath651 such that @xmath689 , @xmath690 , and @xmath691 .
since @xmath692 are in the outermost path @xmath665 , the case @xmath693 is impossible .
hence , we may assume that @xmath685 or @xmath694 .
if @xmath688 are all distinct , then one of @xmath635 , @xmath675 , and @xmath275 must be on the side of @xmath44 in @xmath679 , which contradicts the assumption that @xmath692 are on @xmath665 and @xmath665 is the outermost path with respect to @xmath44 .
if @xmath695 , then either @xmath675 or @xmath275 is on the side of @xmath44 in @xmath679 .
this final contradiction finishes the proof of the claim .
we return to the proof of lemma [ fabc ] .
the endpoints of the path @xmath696 are @xmath697 and there is no edge connecting @xmath44 and @xmath667 .
therefore , if @xmath698 is a chordless cycle , then @xmath699 is the required set @xmath197 .
suppose that @xmath700 has a chord .
then the chord must be an edge in @xmath650 .
note that since @xmath192 has no chordless @xmath24-cycles , there is no edge connecting @xmath701 and @xmath652 .
we consider the vertices @xmath702 and @xmath703 that are closest to @xmath119 and @xmath565 , respectively , on the arc containing @xmath647 .
similarly , consider the vertices @xmath704 and @xmath705 that are closest to @xmath119 and @xmath565 , respectively , on the arc containing @xmath648 .
denote by @xmath706 the subset of vertices in @xmath647 lying strictly between @xmath119 and @xmath707 .
define the subsets @xmath708 , @xmath709 and @xmath710 similarly .
see figure [ fig : case1 ] , left . 0.3 ( 0,0 )
circle(1 cm ) ; ( 0,1 ) coordinate ( v ) ( -0.34,0.93 ) coordinate ( u1 ) ( -0.64,0.76 ) coordinate(u2 ) ( -0.86,0.5 ) coordinate ( u3 ) ( -0.98,0.17 ) coordinate ( u4 ) ( -0.98,-0.17 ) coordinate ( u5 ) ( -0.86,-0.5 ) coordinate ( u6 ) ( -0.64,-0.76 ) coordinate(u7 ) ( -0.34,-0.93 ) coordinate ( u8 ) ( 0,-1 ) coordinate ( vp ) ( 0,0 ) coordinate ( w ) ( -0.8,1.2 ) coordinate ( a ) ( -1.5,0.8 ) coordinate ( b ) ( -1.3,0 ) coordinate ( d ) ( -1.5,-0.9 ) coordinate ( c ) ; ( 0.86,0.5 ) coordinate ( up2 ) ( 0.5,0.86 ) coordinate ( up1 ) ( 0.5,-0.86 ) coordinate ( up6 ) ( 0.86,-0.5 ) coordinate ( up5 ) ( 0.98,-0.17 ) coordinate ( up4 ) ( 0.98,0.17 ) coordinate ( up3 ) ( 0.5,0 ) coordinate ( e ) ( 0.5,-0.5 ) coordinate ( f ) ( 0.75,-0.25 ) coordinate ( g ) ( -0.17,0.98 ) coordinate ( u0 ) ; ( u1 ) circle ( 2pt ) ( u2 ) circle ( 2pt ) ( u3 ) circle ( 2pt ) ( u4 ) circle ( 2pt ) ( u5 ) circle ( 2pt ) ( u6 ) circle ( 2pt ) ( u7 ) circle ( 2pt ) ( a ) circle ( 2pt ) ( b ) circle ( 2pt ) ( c ) circle ( 2pt ) ( d ) circle ( 2pt ) ( up6 ) circle ( 2pt ) ( up5 ) circle ( 2pt ) ( up4 ) circle ( 2pt ) ( up3 ) circle ( 2pt ) ( up2 ) circle ( 2pt ) ( up1 ) circle ( 2pt ) ( e ) circle ( 2pt ) ( f ) circle ( 2pt ) ( g ) circle ( 2pt ) ; ( u8 ) circle ( 2pt ) ( v ) circle ( 2pt ) ( vp ) circle ( 2pt ) ( u0 ) circle ( 2pt ) ; ( 0,0 ) circle(2pt ) ; ( 0,0 ) circle ( 1 cm ) ; ( 0,1.2 ) node@xmath119 ; ( 0,-1.2 ) node@xmath565 ; ( 0,-0.2 ) node@xmath44 ; ( 0,0 ) edge ( u2 ) edge ( u7 ) edge ( up2 ) edge ( up6 ) ; ( a ) edge ( u0 ) edge ( u1 ) edge ( u2 ) edge ( b ) ; ( b ) edge ( u2 ) edge ( u3 ) edge ( d ) edge ( c ) ; ( c ) edge ( u8 ) edge ( u7 ) edge ( u6 ) edge ( d ) ; ( d ) edge ( u3 ) edge ( u4 ) edge ( u5 ) edge ( u6 ) ; ( u2 ) node[left]@xmath707 ( u7 ) node[below]@xmath711 ; ( up2 ) node[right]@xmath712 ( up6 ) node[below]@xmath713 ; ( e ) edge ( w ) edge ( f ) edge ( g ) edge ( up2 ) edge ( up3 ) ; ( f ) edge ( w ) edge ( up6 ) edge ( up5 ) edge ( g ) ; ( g ) edge ( up3 ) edge ( up4 ) edge ( up5 ) ; plot coordinates ( up2 ) ( e ) ( g ) ( up4 ) ; plot coordinates ( up4 ) ( g ) ( f ) ( up6 ) ; plot coordinates ( up3 ) ( g ) ( up5 ) ; 0.3 ( 0,0 ) circle(1 cm ) ; ( 0,1 ) coordinate ( v ) ( -0.34,0.93 ) coordinate ( u1 ) ( -0.64,0.76 ) coordinate(u2 ) ( -0.86,0.5 ) coordinate ( u3 ) ( -0.98,0.17 ) coordinate ( u4 ) ( -0.98,-0.17 ) coordinate ( u5 ) ( -0.86,-0.5 ) coordinate ( u6 ) ( -0.64,-0.76 ) coordinate(u7 ) ( -0.34,-0.93 ) coordinate ( u8 ) ( 0,-1 ) coordinate ( vp ) ( 0.5,-0.86 ) coordinate ( up5 ) ( 0.86 , -0.5 ) coordinate ( up4 ) ( 0,1 ) coordinate ( up3 ) ( 0.86,0.5 ) coordinate ( up2 ) ( 0.5,0.86 ) coordinate ( up1 ) ( -0.8,1.2 ) coordinate ( a ) ( -1.5,0.8 ) coordinate ( b ) ( -1.3,0 ) coordinate ( d ) ( -1.5,-0.9 ) coordinate ( c ) ( -0.17,0.98 ) coordinate ( u0 ) ; ( 0,1 ) circle(2pt ) ; ( 0,-1 ) circle(2pt ) ; ( 0,-1 ) arc ( -90:90:1 cm ) ; ( 0,0 ) coordinate ( w ) ( 0.86,0.5 ) coordinate ( up2 ) ( 0.5,0.86 ) coordinate ( up1 ) ( -0.5,0.86 ) coordinate ( u1 ) ( -0.86,0.5 ) coordinate ( u2 ) ( -1,0 ) coordinate ( u3 ) ( -0.86,-0.5 ) coordinate ( u4 ) ( -0.5,-0.86 ) coordinate ( u5 ) ( 0.5,-0.86 ) coordinate ( up6 ) ( 0.86,-0.5 ) coordinate ( up5 ) ( 0.98,-0.17 ) coordinate ( up4 ) ( 0.98,0.17 ) coordinate ( up3 ) ( 0,1 ) coordinate ( v ) ( 0,-1 ) coordinate ( vp ) ( 0.5,0 ) coordinate ( a ) ( 0.5,-0.5 ) coordinate ( b ) ( 0.75,-0.25 ) coordinate ( c ) ; ( up1 ) circle(2pt ) ( up2 ) circle(2pt ) ( up6 ) circle ( 2pt ) ( a ) circle(2pt ) ( b ) circle ( 2pt ) ( c ) circle ( 2pt ) ( up3 ) circle ( 2pt ) ( up4 ) circle ( 2pt ) ( up5 ) circle ( 2pt ) ; ( v ) node[above]@xmath119 ( vp ) node[below]@xmath565 ( up2 ) node[right]@xmath712 ( up6 ) node[below]@xmath713 ; ( a ) edge ( b ) edge ( c ) edge ( up2 ) edge ( up3 ) ; ( b ) edge ( c ) edge ( up6 ) edge ( up5 ) ; ( c ) edge ( up4 ) edge ( up5 ) edge ( up3 ) ; 0.3 ( 0,0 ) circle(1 cm ) ; ( 0,1 ) coordinate ( v ) ( -0.34,0.93 ) coordinate ( u1 ) ( -0.64,0.76 ) coordinate(u2 ) ( -0.86,0.5 ) coordinate ( u3 ) ( -0.98,0.17 ) coordinate ( u4 ) ( -0.98,-0.17 ) coordinate ( u5 ) ( -0.86,-0.5 ) coordinate ( u6 ) ( -0.64,-0.76 ) coordinate(u7 ) ( -0.34,-0.93 ) coordinate ( u8 ) ( 0,-1 ) coordinate ( vp ) ( -0.8,1.2 ) coordinate ( a ) ( -1.5,0.8 ) coordinate ( b ) ( -1.3,0 ) coordinate ( d ) ( -1.5,-0.9 ) coordinate ( c ) ( -0.17,0.98 ) coordinate ( u0 ) ; ( a ) circle ( 2pt ) ( b ) circle ( 2pt ) ( c ) circle ( 2pt ) ; ( u8 ) circle ( 2pt ) ; ( 0,-1 ) circle(2pt ) ( v ) circle ( 2pt ) ( 0,0 ) circle ( 2pt ) ( u0 ) circle ( 2pt ) ; ( u8 ) arc ( -110:-90:1 cm ) ; ( v ) arc ( 90:100:1 cm ) ; ( 0,1.2 ) node@xmath119 ; ( 0,-1.2 ) node@xmath565 ( 0,0)[below ] node@xmath44 ; ( a ) edge ( u0 ) edge ( b ) ; ( b ) edge ( c ) ; ( c ) edge ( u8 ) ; ( 0,-1 ) arc ( -90:90:1 cm ) ; ( 0,-1 ) arc ( -90:-60:1 cm ) ; ( 0.86,0.5 ) arc ( 30:90:1 cm ) ; ( 0,0 ) coordinate ( w ) ( 0.86,0.5 ) coordinate ( up2 ) ( 0.5,0.86 ) coordinate ( up1 ) ( -0.5,0.86 ) coordinate ( u1 ) ( -0.86,0.5 ) coordinate ( u2 ) ( -1,0 ) coordinate ( u3 ) ( -0.86,-0.5 ) coordinate ( u4 ) ( -0.5,-0.86 ) coordinate ( u5 ) ( 0.5,-0.86 ) coordinate ( up6 ) ( 0.86,-0.5 ) coordinate ( up5 ) ( 0.98,-0.17 ) coordinate ( up4 ) ( 0.98,0.17 ) coordinate ( up3 ) ( 0,1 ) coordinate ( v ) ( 0,-1 ) coordinate ( vp ) ( 0.5,0 ) coordinate ( a ) ( 0.5,-0.5 ) coordinate ( b ) ( 0.75,-0.25 ) coordinate ( c ) ; ( up1 ) circle(2pt ) ( up2 ) circle(2pt ) ( up6 ) circle ( 2pt ) ( a ) circle(2pt ) ( b ) circle ( 2pt ) ( c ) circle ( 2pt ) ( up3 ) circle(2pt ) ( up4 ) circle(2pt ) ( up5 ) circle ( 2pt ) ; ( v ) node[above]@xmath119 ( vp ) node[below]@xmath565 ( up2 ) node[right]@xmath712 ( up6 ) node[below]@xmath713 ; ( a ) edge ( b ) edge ( c ) edge ( up2 ) edge ( up3 ) ; ( b ) edge ( c ) edge ( up6 ) edge ( up5 ) ; ( c ) edge ( up4 ) edge ( up5 ) edge ( up3 ) ; ( 0,0 ) edge ( up2 ) edge ( a ) edge ( b ) edge ( up6 ) ; plot coordinates ( up2 ) ( a ) ( b ) ( up6 ) ; we consider two cases .
there is no edge connecting @xmath667 and @xmath714 in @xmath650 . *
we define @xmath715 we take the innermost path @xmath716 connecting @xmath119 and @xmath565 in @xmath717 with respect to @xmath44 , see figure [ fig : case1 ] , and let @xmath718 be the vertex set of @xmath716
. then @xmath719 by the same argument as the claim above , and @xmath720 is the required subset @xmath197 . 0.3 ( 0,0 ) circle ( 1 cm ) ; ( 0,0 ) circle(1 cm ) ; ( 0,0 ) coordinate ( w ) ( 0,1 ) coordinate ( v ) ( -0.34,0.93 ) coordinate ( u1 ) ( -0.64,0.76 ) coordinate(u2 ) ( -0.86,0.5 ) coordinate ( u3 ) ( -0.98,0.17 ) coordinate ( u4 ) ( -0.98,-0.17 ) coordinate ( u5 ) ( -0.86,-0.5 ) coordinate ( u6 ) ( -0.64,-0.76 ) coordinate(u7 ) ( -0.34,-0.93 ) coordinate ( u8 ) ( 0.34,-0.93 ) coordinate ( up7 ) ( 0.64,-0.76 ) coordinate ( up6 ) ( 0.86,-0.5 ) coordinate ( up5 ) ( 0.98,-0.17 ) coordinate ( up4 ) ( 1,0 ) coordinate ( up3 ) ( 0.86,0.5 ) coordinate ( up2 ) ( 0.5,0.86 ) coordinate ( up1 ) ; ( 0,0.3 ) coordinate ( a ) ( 0,0.6 ) coordinate ( x ) ( -1.2,1.2 ) coordinate ( g ) ( -1.9,1.4 ) coordinate ( h ) ( -1.5,0.3 ) coordinate ( j ) ( -0.3,1.4 ) coordinate ( y ) ( 0.5,-0.3 ) coordinate ( b ) ( -1,-1.2 ) coordinate ( c ) ; ( u1 ) circle ( 2pt ) ( u2 ) circle ( 2pt ) ( u3 ) circle ( 2pt ) ( u4 ) circle ( 2pt ) ( a ) circle ( 2pt ) ( b ) circle ( 2pt ) ; ( u5 ) circle ( 2pt ) ( u6 ) circle ( 2pt ) ( u7 ) circle ( 2pt ) ( u8 ) circle ( 2pt ) ( v ) circle ( 2pt ) ( vp ) circle ( 2pt ) ( up1 ) circle ( 2pt ) ( up2 ) circle ( 2pt ) ( up3 ) circle ( 2pt ) ( up4 ) circle ( 2pt ) ( up5 ) circle ( 2pt ) ( w ) circle ( 2pt ) ( up6 ) circle ( 2pt ) ( up7 ) circle ( 2pt ) ( c ) circle ( 2pt ) ; ( g ) circle ( 2pt ) ( h ) circle ( 2pt ) ( j ) circle ( 2pt ) ( x ) circle ( 2pt ) ( y ) circle ( 2pt ) ; ( u4 ) node[left]@xmath707 ; ( v ) node[above]@xmath119 ( vp ) node[below]@xmath565 ( up3 ) node[right]@xmath712 ( w ) node[below]@xmath44 ( u7 ) node[below]@xmath711 ( up6 ) node[below]@xmath713 ; ( b ) edge ( up3 ) edge ( up4 ) edge ( up5 ) edge ( w ) edge ( up6 ) ; ( c ) edge ( u6 ) edge ( u7 ) edge ( u8 ) ; ( w ) edge ( a ) edge ( up3 ) edge ( u4 ) edge ( u7 ) edge ( up6 ) ; ( a ) edge ( up3 ) edge ( up2 ) edge ( up1 ) edge ( u2 ) ; ( g ) edge ( u1 ) edge ( u2 ) edge ( u3 ) edge ( h ) ; ( h ) edge ( u3 ) edge ( u4 ) edge ( j ) ; ( j ) edge ( u4 ) edge ( u5 ) ; ( x ) edge ( v ) edge ( u1 ) edge ( u2 ) edge ( up1 ) edge ( a ) ; ( y ) edge ( v ) edge ( g ) edge ( u1 ) ; plot [ smooth ] coordinates ( h ) ( -1,2 ) ( 0,2.3 ) ( 0.6,1.8 ) ( up2 ) ; plot [ smooth ] coordinates ( h ) ( -1,2 ) ( 0,2.3 ) ( 0.6,1.8 ) ( up2 ) ; 0.3 ( 0,0 ) circle ( 1 cm ) ; ( 0,0 ) circle(1 cm ) ; ( 0,0 ) coordinate ( w ) ( 0,1 ) coordinate ( v ) ( 0.34,-0.93 ) coordinate ( up7 ) ( 0.64,-0.76 ) coordinate ( up6 ) ( 0.86,-0.5 ) coordinate ( up5 ) ( 0.98,-0.17 ) coordinate ( up4 ) ( 1,0 ) coordinate ( up3 ) ( 0.86,0.5 ) coordinate ( up2 ) ( 0.5,0.86 ) coordinate ( up1 ) ; ( 0,0.3 ) coordinate ( a ) ( 0,0.6 ) coordinate ( x ) ( 0.5,-0.3 ) coordinate ( b ) ; ( a ) circle ( 2pt ) ( b ) circle ( 2pt ) ; ( v ) circle ( 2pt ) ( vp ) circle ( 2pt ) ( up1 ) circle ( 2pt ) ( up2 ) circle ( 2pt ) ( up3 ) circle ( 2pt ) ( up4 ) circle ( 2pt ) ( up5 ) circle ( 2pt ) ( up6 ) circle ( 2pt ) ( up7 ) circle ( 2pt ) ; ( x ) circle ( 2pt ) ; ( v ) node[above]@xmath119 ( vp ) node[below]@xmath565 ( up3 ) node[right]@xmath712 ( up6 ) node[below]@xmath713 ; ( b ) edge ( up3 ) edge ( up4 ) edge ( up5 ) edge ( up6 ) ; ( a ) edge ( up3 ) edge ( up2 ) edge ( up1 ) ; ( x ) edge ( v ) edge ( up1 ) edge ( a ) ; plot coordinates ( v ) ( x ) ( a ) ( up3 ) ( b ) ( up6 ) ; ( vp ) arc ( -90:-50:1 cm ) ; 0.3 ( 0,0 ) circle ( 1 cm ) ; ( 0,0 ) circle(1 cm ) ; ( 0,0 ) coordinate ( w ) ( 0,1 ) coordinate ( v ) ( -0.34,0.93 ) coordinate ( u1 ) ( -0.64,0.76 ) coordinate(u2 ) ( -0.86,0.5 ) coordinate ( u3 ) ( -0.98,0.17 ) coordinate ( u4 ) ( -0.98,-0.17 ) coordinate ( u5 ) ( -0.86,-0.5 ) coordinate ( u6 ) ( -0.64,-0.76 ) coordinate(u7 ) ( -0.34,-0.93 ) coordinate ( u8 ) ; ( -1.2,1.2 ) coordinate ( g ) ( -1.9,1.4 ) coordinate ( h ) ( -1.5,0.3 ) coordinate ( j ) ( -0.3,1.4 ) coordinate ( y ) ( -1,-1.2 ) coordinate ( c ) ; ( u1 ) circle ( 2pt ) ( u2 ) circle ( 2pt ) ( u3 ) circle ( 2pt ) ( u4 ) circle ( 2pt ) ; ( u5 ) circle ( 2pt ) ( u6 ) circle ( 2pt ) ( u7 ) circle ( 2pt ) ( u8 ) circle ( 2pt ) ( v ) circle ( 2pt ) ( vp ) circle ( 2pt ) ( c ) circle ( 2pt ) ; ( g ) circle ( 2pt ) ( h ) circle ( 2pt ) ( j ) circle ( 2pt ) ( y ) circle ( 2pt ) ; ( u4 ) node[left]@xmath707 ;
( v ) node[above]@xmath119 ( vp ) node[below]@xmath565 ( u7 ) node[below]@xmath711 ; ( c ) edge ( u6 ) edge ( u7 ) edge ( u8 ) ; ( g ) edge ( u1 ) edge ( u2 ) edge ( u3 ) edge ( h ) ; ( h ) edge ( u3 ) edge ( u4 ) edge ( j ) ; ( j ) edge ( u4 ) edge ( u5 ) ; ( y ) edge ( v ) edge ( g ) edge ( u1 ) ; plot coordinates ( v ) ( y ) ( g ) ( h ) ( j ) ( u5 ) ; plot coordinates ( u6 ) ( c ) ( u8 ) ; ( u5 ) arc ( 190:210:1 cm ) ; ( vp ) arc ( -90:-110:1 cm ) ; * case 2 .
there is an edge connecting @xmath667 and @xmath721 or @xmath722 in @xmath650 . *
suppose that @xmath667 is connected by an edge in @xmath650 to only one of @xmath721 and @xmath722 , say to @xmath721 .
we define @xmath723 we take the innermost path @xmath724 connecting @xmath119 and @xmath565 in @xmath649 with respect to the vertex @xmath44 , and let @xmath725 be the vertex set of @xmath724
. see figure [ fig : case-2 ] , middle .
then @xmath726 by the same reason as the claim above .
if @xmath727 is a chordless cycle , then @xmath728 is the required subset @xmath197 .
if @xmath729 has a chord , then it must be an edge in @xmath649 connecting @xmath730 and @xmath731 . in this case
we modify @xmath667 as follows .
we define @xmath732 we take the outermost path @xmath733 connecting @xmath119 and @xmath565 in @xmath734 with respect to the vertex @xmath44 , see figure [ fig : case-2 ] , right .
let @xmath735 be the vertex set of @xmath733 .
then @xmath736 by the same argument as the claim above , and we can see that @xmath737 is the required subset @xmath197 . indeed , we only need to check that there is no edge connecting @xmath738 and @xmath652 in @xmath650 .
this is because there is an edge connecting @xmath667 and @xmath721 .
it remains to consider the case when @xmath667 is connected to both @xmath721 and @xmath722 by edges in @xmath650 .
here the same argument as above works if we consider @xmath739 and @xmath740 instead of @xmath741 and @xmath734 , respectively .
here we give a proof which is different from the original proof of @xcite .
it uses a reformulation of the description of product in the cohomology of a moment - angle complex ( theorem [ prodfsc ] ) in terms of the polytope @xmath1 .
a detailed description of this approach can be found in @xcite .
we need to prove that the product map @xmath742 is surjective for any flag @xmath7-polytope @xmath1 and @xmath203 $ ] .
we first restate this in terms of the polytope @xmath1 rather than its dual simplical complex @xmath192 .
the decomposition of @xmath743 into facets @xmath33 defines a cell decomposition of @xmath743 which is poincar dual to the simplicial decomposition @xmath192 .
the two decompositions have the same barycentric subdivision , @xmath744 .
we identify the set of facets @xmath745 with @xmath193 $ ] , and for each @xmath203 $ ] define @xmath746 note that @xmath747 is the combinatorial neighbourhood of @xmath748 in @xmath586 , so there is a deformation retraction @xmath749 .
we have poincar duality isomorphisms @xmath750 where the boundary @xmath751 consists of points @xmath752 such that @xmath753 for some @xmath754 .
topologically , @xmath747 is a disjoint union of several discs with holes , and @xmath751 is a disjoint union of edge cycles .
the cellular homology groups @xmath755 have the following description .
let @xmath756 be the decomposition into connected components .
then * @xmath757 is a free abelian group with basis of homology classes @xmath758=\sum_{i\in i^k}[f_i]$ ] , @xmath759 ; * @xmath760 , where @xmath761 is a free abelian group of rank one less the number of cycles in @xmath762 .
a basis of @xmath761 is given by any set of edge paths in @xmath763 connecting one fixed boundary cycle with the other boundary cycles . as the product map is stated in terms of the reduced cohomology groups
@xmath764 , we introduce the corresponding `` reduced '' homology groups @xmath765\bigr ) , & i=2 .
\end{cases}\ ] ] then we can rewrite as @xmath766 with this interpretation in mind , we can rewrite the product map as the `` intersection pairing '' @xmath767\otimes[p_{i_2^q } ] & \mapsto[p_{i_1^p}\cap p_{i_2^q}]=[\gamma_1]+\cdots+[\gamma_r ] , \end{aligned}\ ] ] where @xmath768 is a connected component of @xmath769 , @xmath770 is a connected component of @xmath771 , and @xmath772 are edge paths in @xmath1 which form the connected components of the intersection @xmath773 .
( there is a sign involved in the transition from to , but it does not affect our subsequent considerations . ) to see that is surjective for a flag 3-polytope @xmath1 , we recall that @xmath774 and consider for each connected component @xmath763 of @xmath747 the decomposition @xmath775 into boundary cycles .
we may assume that @xmath776 , as otherwise @xmath763 is a disc and @xmath777 .
for each pair of boundary cycles @xmath778 and @xmath779 among @xmath780 , we shall decompose the generator @xmath781 of @xmath782 corresponding to an edge path from @xmath778 to @xmath779 into a product of elements of @xmath783 and @xmath784 , @xmath785 .
this will prove the surjectivity of .
we choose facets @xmath786 and @xmath787 in @xmath788 adjacent to @xmath778 and @xmath779 respectively , see figure [ a - b - belt ] . by lemma [ fab ] ,
there is a belt @xmath789 with @xmath790 and @xmath791 , where @xmath792 .
let @xmath793 be a part of the belt between @xmath786 and @xmath787 ( there are two such parts , and we can take any of them ) .
the complement @xmath794 is a union of two open discs .
we denote the closures of these discs by @xmath795 and @xmath796 ; each of them is a union of facets not in @xmath640 .
now set @xmath797\in
\widehat{h}_2(p_{i_1},\partial p_{i_1 } ) , & & b=\sum\limits_{f_j\in p_{i^k}\cap \mathcal{w}_1}[f_j]\in \widehat{h}_2(p_{i_2},\partial p_{i_2}).\end{aligned}\ ] ] then @xmath798+\cdots+[\gamma_s]\in h_1(p_i,\partial p_i)$ ] , where each @xmath799 is an edge path in @xmath763 that begins at some boundary cycle @xmath800 and ends at @xmath801 .
we may assume that @xmath802 begins at @xmath778 and @xmath803 ends at @xmath779 ( where @xmath804 is the pair of boundary cycles chosen above ) .
the homology class @xmath805+\cdots+[\gamma_s]\in\widehat h_1(p_i,\partial p_i)$ ] is then equal to the chosen generator @xmath781 of @xmath806 corresponding to an edge path from @xmath778 to @xmath779 .
we have therefore decomposed @xmath781 into a product @xmath807 , as needed .
the proof uses the combinatorial result of lemma [ fabc ] and an algebraic `` annihilator lemma '' of fan , ma and wang . [ annlemma ]
let @xmath1 be a @xmath7-polytope from the pogorelov class @xmath0 , with the dual complex @xmath218 .
let @xmath810 , where @xmath319 is a field . in the notation of lemma [ rig3cl ] , consider a @xmath319-linear combination of elements of @xmath811 , @xmath812\ ] ] with at least two nonzero @xmath813 .
then , for any @xmath814 such that @xmath815 , @xmath816>\dim\ann_r\alpha.\ ] ] take a complementary subspace @xmath818 to @xmath819 $ ] in @xmath808 , so that @xmath820=r$ ] . for any @xmath821 we have @xmath822\ne0 $ ]
. furthermore , we can choose @xmath818 respecting the multigrading , so that the @xmath197th multigraded component of @xmath818 is a complementary subspace to @xmath819\cap\widehat h_*(p_i,\partial p_i)$ ] in @xmath823 .
then we can write @xmath824\setminus\{k , l\}}\beta_i$ ] , where @xmath825 denotes the @xmath197th multigraded component of @xmath821 . ( note that @xmath826 whenever @xmath827 , as such @xmath825 would annihilate @xmath828 $ ] . )
we can choose @xmath203\setminus\{k , l\}$ ] such that @xmath829\ne0 $ ] .
now consider @xmath830 $ ] .
we claim that the @xmath831th multigraded component of @xmath832 consists of @xmath829 $ ] only .
indeed , for any other component @xmath833 of @xmath834 with @xmath835 and any summand @xmath836 $ ] of @xmath837 , we have @xmath838 , as @xmath839\setminus\{k , l\}$ ]
. then @xmath840\ne0 $ ] .
hence , @xmath841 , which implies that @xmath842{\geqslant}\dim\ann_r\alpha$ ] . in order to show that the strict inequality holds , we shall find an element @xmath843 $ ] such that @xmath844 . take a summand @xmath845 $ ] of @xmath837 different from @xmath846 $ ] .
that is , @xmath847 and @xmath848 .
we can assume without loss of generality that @xmath849 .
by lemma [ fabc ] , there is a belt @xmath543 in @xmath1 such that @xmath850 , @xmath851 , and @xmath852 does not intersect one of the two connected components @xmath853 and @xmath854 of @xmath855 , say @xmath853 . in the dual language , there is a chordless cycle @xmath856 in @xmath206 such that @xmath857 , @xmath858 , and the vertex @xmath859 is not joined by an edge to any vertex of the connected component @xmath860 of @xmath861 .
now we observe that @xmath861 is a full subcomplex of @xmath206 and take @xmath862 to be the cohomology class in @xmath863 given by a generator of @xmath864 .
such a generator is represented by the 0-cocycle @xmath865 ( see example [ exacochain ] ) .
we have @xmath866=0 $ ] because we can write @xmath867 $ ] ( see example [ exacochain ] ) and @xmath868 for any @xmath869 by the choice of the cycle @xmath856 . on the other hand , the product @xmath870 $ ] corresponds to a generator of @xmath871 .
therefore , @xmath843 $ ] and @xmath872 ( the latter is because the multigraded component of @xmath873 corresponding to @xmath856 is @xmath874\ne0 $ ] ) .
take @xmath824\setminus\{k , l\}}\beta_i\in l_{kl}\setminus\{0\}$ ] and choose @xmath203\setminus\{k , l\}$ ] such that @xmath875\ne0 $ ] , as in the beginning of the proof .
the multigrading of @xmath862 does not contain @xmath859 , so we have @xmath876 $ ] for some @xmath877 $ ] .
now , @xmath878=0 $ ] because @xmath879 $ ] and @xmath868 for any @xmath869 , as @xmath119 and @xmath859 are not joined by an edge .
hence , @xmath880 .
thus , @xmath881 and we have proved that @xmath844 .
this implies that @xmath842>\dim\ann_r\alpha$ ] .
we are given a @xmath7-polytope @xmath1 from the pogorelov class @xmath0 and a ring isomorphism @xmath882 .
we defined the set @xmath405\in
h^3({\mathcal z_p}),\quad f_i\cap f_j=\varnothing\},\ ] ] and the corresponding set for @xmath4 , @xmath883\in h^3(\mathcal z_{p'}),\quad f'_i\cap f'_j=\varnothing\}.\ ] ] we need to show that @xmath406 , in other words , @xmath884)=\pm[u'_rv'_s]$ ] .
we first use theorems [ flagrig ] and [ 4beltrig ] to conclude that @xmath4 also belongs to the class @xmath0 .
now suppose that @xmath884)=\alpha'=\sum r_{ij}[u'_iv'_j]$ ] with at least two nonzero @xmath885 .
we are then in the situation of lemma [ annlemma ] , which we can apply to @xmath4 .
we obtain that @xmath886 $ ] for any nonzero summand @xmath887 $ ] of @xmath888 .
considering the inverse isomorphism @xmath889 , we can choose @xmath890 $ ] such that @xmath891)=\alpha=\sum r_{ab}[u_av_b]$ ] where @xmath892 $ ] appears in the latter sum .
as an isomorphism preserves the dimension of the annihilator subspace , we obtain @xmath893=\dim\ann_{r'}\alpha'<\dim\ann_r[u'_kv'_l]= \dim\ann_r\alpha\\<\dim\ann_r[u_pv_q],\end{gathered}\ ] ] which is a contradiction .
it follows that @xmath884)$ ] is a multiple of a single @xmath894 $ ] .
since @xmath439 is an isomorphism over @xmath895 , we have @xmath884)=\pm[u'_rv'_s]$ ] .
baskakov , ilia .
_ massey triple products in the cohomology of moment - angle complexes
. _ uspekhi mat .
nauk * 58 * ( 2003 ) , no . 5 , 199200 ( russian ) ; russian math .
surveys * 58 * ( 2003 ) , no . 5 , 10391041 ( english translation ) . baskakov , ilia ; buchstaber , victor ; panov , taras . _ cellular cochain algebras and torus actions_. uspekhi mat . nauk * 59 * ( 2004 ) , no . 3 , 159160 ( russian ) ; russian math . surveys * 59 * ( 2004 ) , no . 3 , 562563 ( english translation ) . buchstaber , victor . _
lectures on toric topology_. in _ proceedings of toric topology workshop kaist 2008_. information center for mathematical sciences , kaist .
trends in math.new series * 10 * ( 2008 ) , no . 1 , 164 . buchstaber , victor ; erokhovets , nikolay .
_ truncations of simple polytopes and applications_. trudy mat .
inst . steklova * 289 * ( 2015 ) , 115144 ( russian ) ; proc .
steklov inst .
* 289 * ( 2015 ) , no . 1 , 104133 ( english translation ) . buchstaber , victor ; panov , taras .
_ torus actions , combinatorial topology and homological algebra_. uspekhi mat .
nauk * 55 * ( 2000 ) , no . 5 , 3106 ( russian ) .
russian math . surveys * 55 * ( 2000 ) , no . 5 , 825921 ( english translation ) .
masuda , mikiya ; panov , taras .
_ semifree circle actions , bott towers , and quasitoric manifolds_. mat . sbornik * 199 * ( 2008 ) , no . 8 , 95122 ( russian ) ; sbornik math .
* 199 * ( 2008 ) , no . 78 , 12011223 ( english translation ) .
masuda , mikiya ; suh , dong youp .
_ classification problems of toric manifolds via topology_. in : _ toric topology _ , m. harada _
et al . _ ,
contemp . math .
soc . , providence , ri , 2008 , pp . 273286 .
zhubr , alexei v. _ closed simply connected six - dimensional manifolds : proofs of classification theorems .
i analiz * 12 * ( 2000 ) , no .
4 , 126230 ( russian ) ; st .
petersburg math .
j. * 12 * ( 2001 ) , no . 4 , 605680 ( english translation ) . | we consider the class @xmath0 of 3-dimensional simple polytopes @xmath1 which are flag and do not have 4-belts of facets .
it includes fullerenes , i.e. simple 3-polytopes with only 5-gonal and 6-gonal facets . according to a theorem of pogorelov , any polytope from @xmath0
admits a right - angled realisation in lobachevsky 3-space , and such a realisation is unique up to isometry .
we study two families of smooth manifolds associated with polytopes from the class @xmath0 .
the first family consists of 6-dimensional quasitoric manifolds over polytopes from @xmath0 .
the second family consists of 3-dimensional small covers of polytopes from @xmath0 or , equivalently , hyperbolic 3-manifolds of lbell type .
our main result is that both families are cohomologically rigid , i.e. two manifolds @xmath2 and @xmath3 from either of the families are diffeomorphic if and only if their cohomology rings are isomorphic .
we also prove that if @xmath2 and @xmath3 are diffeomorphic , then their corresponding polytopes @xmath1 and @xmath4 are combinatorially equivalent . these results are intertwined with the classical subjects of geometry and topology , such as combinatorics of 3-polytopes , the four colour theorem , diffeomorphism classification of 6-manifolds and invariance of pontryagin classes .
the proofs use techniques of toric topology . |
understanding the behavior of qcd matter under different conditions of temperature and density is the subject of intense experimental and theoretical work in nuclear physics .
the transition from hadronic matter to a quark - gluon plasma at high temperature is expected to be achieved in high energy heavy ion collisions .
the hadronization of partons in vacuum or cold nuclear matter ( i.e. in a nucleus ) is also of keen interest , and represents a nonperturbative and dynamic qcd process .
the formation and interaction of heavy quarkonia ( for example @xmath0 mesons ) in vacuum , cold nuclear matter and hot nuclear matter present an excellent laboratory for gaining insights on these transformations .
recent results from the relativistic heavy ion collider reveal a significant suppression of the final @xmath0 yield in central ( small impact parameter ) @xmath7 reactions at @xmath1=200 gev , relative to expectations scaled from @xmath8 reactions at the same energy @xcite .
a possible source of this suppression is the screening of the attractive interaction between the quark - antiquark pair in the hot nuclear medium , as temperatures are expected to be above the critical temperature for a quark - gluon plasma transition .
larger @xmath0 suppression is observed at forward rapidity than at midrapidity , which contradicts models with only color screening of quarkonia proportional to the local energy density .
produced @xmath9 pairs must pass through the remaining nuclear material from the incident cold nuclei , in addition to surviving any hot medium environment .
the so - called cold nuclear matter effects @xcite , including modification of initial parton distribution functions ( shadowing , gluon saturation , anti - shadowing , emc effect , etc . ) , initial and final state partonic multiple scattering , and related initial state parton energy loss need to be accounted for before firm conclusions can be drawn about the effect of the hot medium thought to be created .
in fact , these various cold nuclear matter effects are interesting in their own right , notably in terms of hadronization time scales , parton energy loss in matter , and the various initial state effects mentioned above .
this paper presents a new analysis of the modification of @xmath0 production in deuteron - gold ( @xmath10 ) collisions relative to proton - proton ( @xmath8 ) collisions at @xmath1=200 gev and the implications for understanding the @xmath7 and @xmath11 data at the same energy .
the phenix experiment has previously published a result using @xmath8 and @xmath10 data taken in 2003 @xcite .
a modest @xmath0 suppression was observed at forward rapidity ( i.e. in the deuteron moving direction ) , which is a possible indication of shadowing of low-@xmath3 gluons in the gold nucleus . a substantially larger ( more than an order of magnitude ) @xmath8 data set was recorded in 2005 with the @xmath0 results published in @xcite , and has been used as the baseline for recent @xmath7 and @xmath11 nuclear modification factors @xcite .
the same @xmath8 data set is used in the analysis presented here to determine the @xmath10 nuclear modifications more accurately and in a fully consistent way with those in the @xmath7 and @xmath11 cases .
in addition , during the two years between the analyses of the 2003 and 2005 data sets significant improvements in the reconstruction software and signal extraction method were achieved , as well as an overall better understanding of the detector performance .
these improvements have been included in this analysis , allowing maximal cancellation of systematic errors when using the 2005 @xmath8 data sample to form the @xmath0 nuclear modification factor .
we first describe the updated analysis , then present the new nuclear modification factors and their implications .
the phenix apparatus is described in @xcite .
it consists of two sets of spectrometers referred to as the central arms , which measure particles emitted at midrapidity ( @xmath12 ) , and the muon arms , measuring particles emitted at backward and forward rapidity ( @xmath13 and @xmath14 ) . at midrapidity ,
@xmath0 particles are measured via their decay into two electrons .
electrons are identified by matching tracks reconstructed with drift chambers ( dc ) and pad chambers ( pc ) to clusters in the electromagnetic calorimeters ( emcal ) and hits in the ring imaging cerenkov counters ( rich ) . in @xmath10 collisions ,
a charged track is identified as an electron candidate by requiring at least three matching rich phototube hits within a certain radius with respect to the the center defined by the track projection at the rich , a position matching of @xmath15 standard deviations between the emcal cluster and the reconstructed track , and a cut on the ratio of energy to momentum . in @xmath8 collisions
the electron identification cuts are the same except that only two matching rich phototube hits are required . at forward and backward rapidity
, @xmath0 particles are measured via their decay into two muons .
muons are identified by matching tracks measured in cathode strip chambers ( referred to as the muon tracker , or mutr ) to hits in alternating planes of iarocci tubes and steel absorbers ( referred to as the muon identifier , or muid ) .
each muon arm is located behind a thick copper and iron absorber that is meant to absorb most hadrons produced during the collisions , so that the measured muons must penetrate 8 to 11 interaction lengths of material in total .
the @xmath10 data used for this analysis were recorded in 2003 using a minimum bias trigger that required hits in each of the two beam - beam counters ( bbc ) located at positive and negative rapidity ( @xmath16 ) , and represent integrated luminosities for the different spectrometers ranging from 1.4 nb@xmath17 to 1.7 nb@xmath17 ( or equivalently 2.7 to 3.4 billion interactions ) .
this trigger covers @xmath18% of the total @xmath10 inelastic cross section of 2260 mb @xcite . for the electrons ,
an additional trigger was used that required one hit above threshold in the emcal and a matching hit in the rich . for the muons ,
two additional triggers were used at different times during the data - taking period .
the muon triggers are based on information from the muid , which has five active detector layers between the steel absorbers .
for the first part of the data - taking period , one of the tracks was required to reach the fourth muid plane , while the other was only required to reach the second muid plane . for the latter part ,
the trigger required at least two tracks to reach the fourth muid plane of iarocci tubes .
the bbcs are also used to determine the centrality of the @xmath10 collisions by measuring the energy deposited in the counters located at negative rapidity ( in the gold - going direction ) . for a given centrality bin ,
the average number of equivalent nucleon - nucleon collisions ( @xmath19 ) is derived from this energy using a glauber calculation @xcite coupled to a simulation of the bbc .
the centrality bins used in this analysis and the corresponding number of collisions are listed in table [ table : centrality ] . to ensure that the centrality categories are well defined , collisions are required to be within @xmath20 cm of the center of the interaction region . [
cols="^,^,^,^,^,^ " , ] the modified nuclear pdfs from eks and ndsg are constrained from other experimental measurements such as deep inelastic scattering from various nuclear targets and the resulting @xmath21 structure functions .
a geometric parametrization of these pdfs based on the path of the parton through the nucleus is described in @xcite and @xcite .
one can test this geometric dependence by comparison with the @xmath10 nuclear modification factors as a function of @xmath19 . using this geometric dependence ,
the most probable @xmath22 is calculated independently in three rapidity ranges ( see table [ tab : sigma_breakup ] )
. the corresponding nuclear modification values and their one standard deviation bands are shown as a function of @xmath19 in figure [ fig : model_comparisons_dau_ncoll ] . the two calculations with eks and ndsg nuclear pdfs yield almost identical bands since the same geometric dependence is used in both cases .
however , each band represents a different balance of modification due to the nuclear pdf and the breakup cross section . for both the eks and the ndsg pdfs ,
the values of @xmath22 extracted from the overall rapidity dependence of @xmath2 and from the @xmath19 dependence of @xmath2 within the different rapidity ranges are consistent within the large systematic uncertainties . it should be noted that though the @xmath23p - value for the best fit is poor at backward rapidity ( as can be seen in fig .
[ fig : model_comparisons_dau_ncoll ] ) , there is still a well defined maximum in the likelihood function for @xmath22 . a future higher precision @xmath2 measurement as a function of centrality will be crucial to constrain the exact geometric dependence .
one can also utilize this model to do a consistent calculation of the contribution from cold nuclear matter effects that should be present in @xmath11 and @xmath7 collisions .
these contributions , using the best fit value of @xmath22 and their one standard deviation values extracted from the data in figure [ fig : model_comparisons_dau ] for each of the two shadowing models , are shown in figures [ fig : model_comparisons_cucu ] and [ fig : model_comparisons_auau ] . in the @xmath11 case ,
@xmath0 production is not suppressed beyond cold nuclear matter effects at midrapidity or at forward rapidity , within the limits of the large error bands , and the midrapidity data in the @xmath7 case is similarly inconclusive .
however , there is a significant suppression in the data at forward rapidity , beyond the uncertainties in both the data and the projection .
it should be noted that the uncertainty band at forward and midrapidities are entirely correlated , as they reflect only the uncertainty in the @xmath22 .
there is no systematic uncertainty included for the choice of modified nuclear pdf model , which is the only way to change the relative suppression between forward and midrapidities within the context of this calculation .
the more data - driven calculation described later in this section , however , is performed independently at different rapidities and does not suffer the same stipulation .
it should also be noted that the theoretical calculations yield @xmath24 as a continuous function of the number of participants , whereas the data points are at discrete values representing a convolution of the modification factor with the @xmath25 distribution within a particular centrality category . a glauber simulation combined with a monte carlo of the phenix experimental trigger and centrality selection is utilized to convert the continuous theory predictions into discrete predictions in the simulated phenix centrality categories .
thus , the results shown in the figures are in fact predictions for the matched event selection categories of the experimental data points . in order to explore the cold nuclear matter constraints further , an alternative data - driven method proposed in @xcite
this approach assumes that there is a single modification factor parameterizing all cold nuclear matter effects that is a simple function of the radial position in the nucleus .
this computation has the advantage of not having to assume a specific shadowing scheme and a specific breakup cross section , but instead relies only on the measured impact parameter dependence .
it assumes that the cold nuclear matter effects suffered by a @xmath0 in a @xmath7 collision at a given rapidity are the product of the modifications measured in @xmath10 collisions at the same rapidity and the modifications measured at the opposite rapidity ( or equivalently in a @xmath26 collision ) .
this assumption holds for the two effects considered so far , namely shadowing and subsequent breakup .
it also assumes that the same parton distributions are sampled by the @xmath0 particles observed in the ( wide ) rapidity range in @xmath7 and @xmath10 collisions .
note that since this model implicitly includes any possible modified nuclear pdfs , the modification factors may have an @xmath3-dependence that is accounted for by considering the backward , mid and forward rapidity @xmath10 data .
the different rapidity regions are sensitive to the initial - state partons in the gold nucleus in three broad ranges of @xmath3 , corresponding to @xmath27 0.002 - 0.01 , 0.01 - 0.05 , and 0.05 - 0.2 , as determined from pythia .
a glauber monte carlo and a simulation of the bbc detector used for centrality determination and triggering are done . the resulting four centrality categories ( 0 - 20% , 20 - 40%
, 40 - 60% , 60 - 88% ) in @xmath10 collisions are characterized by a distribution in the number of binary collisions , as shown in the top panel of figure [ fig : nagle_glauber ] .
in addition , the distribution of radial positions @xmath28 in the @xmath29 nucleus of binary collisions is calculated and shown in the bottom panel of figure [ fig : nagle_glauber ] .
the procedure is to use the forward , mid , and backward rapidity centrality - dependent @xmath2 to constrain the modification factor @xmath30 for three broad regions of initial parton @xmath3 ( @xmath31 , @xmath32 , @xmath33 respectively ) .
then one can use these parameterizations to project the cold nuclear matter effect in the @xmath7 case .
the current @xmath10 data are insufficient to constrain the functional form of @xmath30 . as a simplifying case
, @xmath30 is assumed to be linear in @xmath28 and to be fixed at @xmath34 at the edge of the gold nucleus .
thus , the only free parameter is the slope ( or equivalently the magnitude of the modification factor at @xmath35 ) .
other functions were tried and essentially differ by their extrapolation to lower and higher radial positions , since the data are not precise enough to constrain the shape .
this has a particularly strong impact on the most peripheral collisions for which our assumption that @xmath36 adds a significant constraint to the shape . for all possible slope parameters , consistency with the experimental data is checked using the procedure detailed in @xcite , which utilizes the full statistical and systematic uncertainties .
the range of parameters within one standard deviation of the uncertainties is determined separately for backward , mid , and forward rapidity . using this range of parameters , the cold nuclear matter suppression expected in @xmath7 collisions
is projected as a function of collision centrality and for mid and forward rapidity .
note that the forward rapidity @xmath7 @xmath0 production is sensitive to the low-@xmath3 partons in one gold nucleus and the high-@xmath3 partons in the other gold nucleus .
thus , in the monte carlo , for every binary collision at @xmath37 and @xmath38 ( the radii with respect to the center of each nucleus ) the expected modification is @xmath39 .
the midrapidity @xmath7 @xmath0 production is predominantly sensitive to the mid-@xmath3 partons from both gold nuclei and therefore the expected modification is @xmath40 .
the total modification expected is calculated by taking the average over all correlated @xmath41 and @xmath42 positions for binary collisions within overall @xmath7 collisions in each @xmath7 centrality class .
the results of these calculations matched to the experimentally measured @xmath7 centrality bins are shown in figure [ fig : nagle ] .
it is notable that the midrapidity cold nuclear matter extrapolation agrees within the uncertainty of the experimental data at midrapidity .
thus , it is not possible within the current constraints to determine the potential extent of hot nuclear matter effects .
this conclusion is qualitatively similar to that reached from the previous model calculations as shown in figure [ fig : model_comparisons_auau ] .
however , at forward rapidity , this method projects a somewhat larger range of possible cold nuclear matter effects than the previous models . neither the predictions of cold nuclear matter effects in heavy ion collisions based on fitting of the @xmath10 data with theoretical curves ( figures [ fig : model_comparisons_cucu ] and [ fig : model_comparisons_auau ] ) , nor those obtained directly from the @xmath10 data points ( figure [ fig : nagle ] ) are well enough constrained to permit quantitative conclusions about additional hot nuclear matter effects .
results from a glauber model monte carlo including simulation of the @xmath10 centrality selection and triggering based on the phenix bbc .
the top panel shows the distribution of the number of binary collisions for events in each of the four centrality classes 0 - 20% , 20 - 40% , 40 - 60% , 60 - 88% .
the distribution for radial impact points in the gold nucleus of binary collisions is shown in the lower panel . ]
predictions of the data driven method @xcite constrained by the @xmath2 as a function of collision centrality for the au+au @xmath24 for midrapidity ( top ) and at forward rapidity ( bottom).,title="fig : " ] predictions of the data driven method @xcite constrained by the @xmath2 as a function of collision centrality for the au+au @xmath24 for midrapidity ( top ) and at forward rapidity ( bottom).,title="fig : " ]
a new analysis of @xmath0 production in d+au collisions at @xmath43 gev has been presented using the phenix 2003 d+au data set .
cuts and analysis techniques that are consistent with the previously published results for @xmath8 , @xmath11 and @xmath7 collisions at the same energy @xcite are used .
the new analysis also benefits from the significantly larger p+p data set from run-5 .
a statistical comparison of these new results to theoretical calculations has been performed with a detailed handling of the experimental uncertainties to estimate a @xmath0 ( or precursor ) breakup cross section in cold nuclear matter on top of models for the modifications of the parton distribution functions in the nucleus .
using eks ( ndsg ) shadowing , a breakup cross section of @xmath44 ( @xmath5 ) mb is obtained .
these breakup cross - section values are consistent within large uncertainties with the @xmath6 mb determined at lower energies at the cern - sps @xcite .
the measured values are then used to predict the expected cold nuclear matter effects on @xmath0 production in @xmath11 and @xmath7 collisions , and these are compared to the measured nuclear modification factors for those systems .
these predictions are found to be similar to those from a less model - dependent and more data - driven method based on the variation of the nuclear modification factor measured in @xmath10 collisions as a function of both rapidity and centrality @xcite .
it is notable that the latter method yields a somewhat larger possible suppression in the forward rapidity case . in all cases the large error bars associated with
the extrapolation prevent making firm quantitative statements on any additional @xmath0 suppression in @xmath7 collisions beyond that expected from cold nuclear matter effects . a @xmath10 data set with much improved statistical precision is needed to both reduce the statistical uncertainties and permit better control over the systematic uncertainties .
we thank the staff of the collider - accelerator and physics departments at brookhaven national laboratory and the staff of the other phenix participating institutions for their vital contributions .
we also thank ramona vogt for useful discussions and for the calculations used to set the level of the breakup cross sections .
we acknowledge support from the office of nuclear physics in the office of science of the department of energy , the national science foundation , abilene christian university research council , research foundation of suny , and dean of the college of arts and sciences , vanderbilt university ( u.s.a ) , ministry of education , culture , sports , science , and technology and the japan society for the promotion of science ( japan ) , conselho nacional de desenvolvimento cientfico e tecnolgico and fundao de amparo pesquisa do estado de so paulo ( brazil ) , natural science foundation of china ( people s republic of china ) , ministry of education , youth and sports ( czech republic ) , centre national de la recherche scientifique , commissariat lnergie atomique , and institut national de physique nuclaire et de physique des particules ( france ) , ministry of industry , science and tekhnologies , bundesministerium fr bildung und forschung , deutscher akademischer austausch dienst , and alexander von humboldt stiftung ( germany ) , hungarian national science fund , otka ( hungary ) , department of atomic energy ( india ) , israel science foundation ( israel ) , korea research foundation and korea science and engineering foundation ( korea ) , ministry of education and science , rassia academy of sciences , federal agency of atomic energy ( russia ) , vr and the wallenberg foundation ( sweden ) , the u.s .
civilian research and development foundation for the independent states of the former soviet union , the us - hungarian nsf - otka - mta , and the us - israel binational science foundation . | we present a new analysis of @xmath0 production yields in deuteron - gold collisions at @xmath1 = 200 gev using data taken by the phenix experiment in 2003 and previously published in [ s.s .
adler _
et al . _ ,
phys .
rev .
lett 96 , 012304 ( 2006 ) ] .
the high statistics proton - proton @xmath0 data taken in 2005 is used to improve the baseline measurement and thus construct updated cold nuclear matter modification factors ( @xmath2 ) . a suppression of @xmath0 in cold nuclear matter
is observed as one goes forward in rapidity ( in the deuteron - going direction ) , corresponding to a region more sensitive to initial state low-@xmath3 gluons in the gold nucleus .
the measured nuclear modification factors are compared to theoretical calculations of nuclear shadowing to which a @xmath0 ( or precursor ) breakup cross section is added .
breakup cross sections of @xmath4 ( @xmath5 ) mb are obtained by fitting these calculations to the data using two different models of nuclear shadowing .
these breakup cross - section values are consistent within large uncertainties with the @xmath6 mb determined at lower collision energies . projecting this range of cold nuclear matter effects to copper - copper and gold - gold collisions
reveals that the current constraints are not sufficient to firmly quantify the additional hot nuclear matter effect . |
figure [ fig : trap_schematic ] is a schematic of the trap used in these experiments .
it was originally designed to have microwaves applied to all three axial electrodes , allowing the creation of a microwave magnetic field gradient , while maintaining a field null at the trap s centre . however , in this work we take advantage of the ddms scheme , which allows us to use a simpler `` un - nulled '' configuration , with the rsb and bsb applied to a single electrode and , correspondingly , a non - zero field amplitude at the trap s centre .
m above the trap s surface .
the @xmath59 static field lies in the trap plane , perpendicular to the trap s axis .
the two outer axial electrodes carry the r.f . trapping voltage .
the six radial electrodes carry d.c .
voltages which , in combination with the d.c .
bias applied to the axial electrodes , create the static trapping potential .
the three axial electrodes may be driven with microwaves .
the microwave field produced by the right ( centre ) axial electrode lies at an angle of @xmath60 ( @xmath61 ) to the b - field [ 17,43 6-6.2 ] . ] to minimise off - resonant excitation of the @xmath62-polarized qubit transition , we apply the rsb and bsb microwaves to the trap s central axial electrode , which produces a predominantly @xmath26-polarized field ( the @xmath26-polarized spectator transitions are detuned by @xmath6350 from the qubit ) , whose gradient has a strong @xmath64-polarized component .
the ddms carrier drive is applied to the right axial electrode , whose field has a much larger @xmath64 component ( figure 4 ) .
the microwave magnetic field gradient created by each sideband is @xmath637t / m .
this gradient is accompanied by a field amplitude of @xmath63780@xmath65 t ( inferred from the measured value of @xmath11 , see below ) , which drives the qubit ( @xmath26 spectator transitions ) with a rabi frequency of 370khz ( 14mhz ) [ 17,43 6 - 6.2 ] .
we reiterate that , despite these large rabi frequencies , the data in figure 3 demonstrate that we are able to suppress errors due to off - resonant rabi flopping ( spin flips ) to @xmath55 .
0.75cyy & + transition & rsb & bsb + @xmath66 & @xmath67 & @xmath68 + @xmath69 & @xmath70 & @xmath71 + @xmath72 & @xmath73 & @xmath74 + @xmath75 & @xmath76 & @xmath77 + @xmath78 & @xmath79 & @xmath80 + total & -47 & + 68 + the a.c .
zeeman shifts arising from off - resonant excitation of the various transitions connecting to the qubit states are given in table 2 .
these values include the bloch - siegert correction , which is significant due to the high degree of cancellation between the various shifts .
accordingly , we calculate the shift @xmath81 on the qubit due to off - resonant excitation of a transition with frequency @xmath82 by a microwave field with frequency @xmath83 and rabi frequency on that transition @xmath84 as @xmath85 where @xmath86 for the rsb ( @xmath87 ) and bsb ( @xmath88 ) , with the values of @xmath11 and @xmath3 used in our experiment ( see the main text ) . for these calculations , we have assumed the @xmath89 microwave polarization angle previously measured in [ 43 ] .
we determine the overall field amplitude from the measured differential a.c .
zeeman shift @xmath11 , assuming equal amplitudes for the rsb and bsb fields .
we expect this to give a more accurate estimate of the field than [ 43 ] ( based on the technique described in [ 41 ] ) , which would predict a field of 1.0mt for our gradient .
we do not attempt to include the second - order correction due to the change in @xmath82 arising from a.c .
zeeman shifts .
instead , we use the bare ( un - shifted ) values of each @xmath82 in our calculations .
the simulations shown in figure [ fig : time_scans ] demonstrate the advantages of the dynamically decoupled mlmer - srensen(ddms ) gate over the single sideband ( ssb ) gate .
they were performed by numerical integration of the von neumann equation for the hamiltonian @xmath90 including 2 spin states and a single motional mode , truncated to 30 harmonic oscillator fock states . for the ssb gate simulations ,
the bsb rabi frequency is set to zero , while the rsb rabi frequency is double that used in the ddms gate simulations .
figure [ fig : time_scans ] ( a ) shows the spin - state dynamics for both gate schemes . while the ddms dynamics are identical to those for a standard mlmer - srensen(ms ) gate , the ssb gate is more complicated , featuring characteristic wiggles .
these more complicated dynamics are likely to make the ssb gate more difficult to set up and optimize in practice .
figure [ fig : time_scans ] ( b ) shows the gate error as a function of carrier rabi frequency , assuming a constant uncompensated a.c
. zeeman shift of @xmath91 .
we see that the ddms gate does not have a significant error as long as @xmath92 .
in contrast , the ssb gate is only able achieve low errors at particular values of @xmath27 , making it more difficult to set up and more susceptible to noise in @xmath27 .
this sensitivity to noise in @xmath27 is likely to mean that a second - order carrier drive is required to achieve high fidelities with the ssb gate .
we note that the error on the ddms curve is still lower than that for the standard ms gate when @xmath93 ( @xmath94 , compared with @xmath95 ) , demonstrating that the refocussing @xmath26-pulse provides some robustness to a.c .
zeeman shifts , even without the carrier drive
. this may be useful in experiments with only small uncompensated a.c .
zeeman shifts . | we demonstrate a two - qubit logic gate driven by near - field microwaves in a room - temperature microfabricated ion trap .
we measure a gate fidelity of 99.7(1)% , which is above the minimum threshold required for fault - tolerant quantum computing .
the gate is applied directly to `` atomic clock '' qubits ( coherence time @xmath0 ) using the microwave magnetic field gradient produced by a trap electrode .
we introduce a dynamically - decoupled gate method , which stabilizes the qubits against fluctuating a.c .
zeeman shifts and avoids the need to null the microwave field .
laser - cooled trapped atomic ions are a promising platform for the development of a general - purpose quantum computer @xcite . in common with other technologies
, the present challenge is performing all elementary logic operations with the fidelity necessary for quantum error correction , whilst using techniques which can be scaled to the number of qubits required to perform a useful computation .
trapped - ion qubits are based on either optical @xcite or hyperfine @xcite atomic transitions .
hyperfine qubits lie in the convenient microwave domain , and have exhibited minute - long memory coherence times @xcite .
nevertheless , they are usually manipulated via optical ( raman ) transitions , firstly because of the convenience of addressing individual ions with tightly - focussed laser beams @xcite , and secondly because the short optical wavelength allows efficient multi - qubit logic gates based on coupling the ions spin and motional degrees of freedom @xcite .
microwave methods have been proposed @xcite , and recently demonstrated , both for individual qubit addressing @xcite , and for multi - qubit logic gates @xcite .
this offers the prospect of performing all coherent operations using purely electronic methods , making phase control significantly easier , and replacing lasers with cheaper , smaller , more stable microwave devices .
microwave elements can also be integrated into trapping structures more easily than their optical counterparts for improved scalability @xcite .
furthermore , microwave gates can theoretically attain higher fidelities as they are not fundamentally limited by photon scattering @xcite .
two distinct microwave methods are being pursued : using far - field microwaves in combination with a local static magnetic field gradient ; and , using a local near - field microwave magnetic field gradient .
microwave - driven two - qubit gates have previously been reported in a single experiment using the near - field method ( with 76% fidelity @xcite ) , and in two experiments using the far - field method ( with 70% fidelity in a 3-ion chain and , very recently , 98.5% for a pair of ions @xcite ) .
beyond quantum information processing , microwave quantum logic techniques are also applicable to metrology and high - resolution spectroscopy , for example for the study of systems without accessible optical transitions @xcite . in this letter , we report a near - field microwave two - qubit gate with fidelity exceeding the @xmath1 minimum threshold required for fault - tolerant quantum error correction @xcite , and comparable to that of the best reported fidelities achieved with lasers or other qubit technologies @xcite .
we estimate the main sources of error in the gate .
we also set a limit on the errors induced by the gate fields on an `` idle '' memory qubit .
the two - qubit gate operation is implemented with the same qubit states , and in the same device , which we have previously used in demonstrating high - fidelity ( @xmath2 ) single - qubit state preparation , gates , memory and readout @xcite .
the trap is a lithographically - defined two - dimensional surface - electrode design , incorporating integrated microwave waveguides and resonators , and is operated at room temperature @xcite .
surface traps are especially promising for scaling up to large numbers of trap zones , as proposed for a `` quantum ccd '' architecture @xcite .
this work was performed using the intermediate - field atomic clock qubit described in @xcite .
the qubit is formed from a pair of hyperfine states in the ground - level , separated by a 3.200 transition ( figure [ fig : ms_levels ] ) , whose frequency is first - order independent of magnetic field at a static field of 14.6 .
details of the laser cooling , initialisation and measurement of this qubit may be found in @xcite .
is the unperturbed qubit frequency , @xmath3 is the motional mode frequency , @xmath4 is the gate detuning , and @xmath5 is the differential a.c .
zeeman shift produced by the ( strong ) sideband fields .
the ( weak ) carrier field used for dynamical decoupling is tuned to resonance with the shifted qubit transition at @xmath6 . ]
the two - qubit gate implemented in this work is an extension of the ideas of mlmer and srensen ( ms ) , ospelkaus _ et al .
_ , and bermudez _ et al . _
@xcite ; it is a gate driven by a microwave near - field gradient , which is robust to what would otherwise be the largest source of experimental error in our system , _ viz .
_ fluctuating a.c .
zeeman shifts arising from the microwave fields .
a standard ms gate is implemented with a bichromatic field with frequencies near the first red and blue sideband transitions for one of the ions normal modes of motion , resulting in dynamics described by the hamiltonian @xmath7 here , @xmath8 and @xmath4 are the gate rabi frequency and detuning respectively ( see figure [ fig : ms_levels ] ) , and is the collective spin operator , where @xmath9 is the pauli operator acting on ion @xmath10 and the sign is positive ( negative ) if the ions normal motions are in phase ( anti - phase ) .
the ions motion is driven by the spatial gradient of the microwave magnetic field .
in general , this gradient will be accompanied by a non - zero field amplitude at the ions equilibrium positions .
( the field can be made small by nulling with additional microwave electrodes , or with specific trap geometries @xcite , but in practice it will always be present at some level . )
this unwanted field will drive off - resonant rabi flopping ( spin flips ) on any hyperfine transition connecting to the qubit states , but the effect of this on the operation of the gate can be highly suppressed using pulse - shaping techniques @xcite .
the field also induces a differential a.c .
zeeman shift @xmath11 on the qubit transition ( figure [ fig : ms_levels ] ) , described by the hamiltonian @xmath12 a constant a.c
. zeeman shift may be treated as an effective change in the qubit frequency and compensated for by adjusting the microwave frequencies appropriately . however
, any fluctuations in @xmath11 will lead to qubit dephasing , which can be a significant source of error .
henceforth , we assume that the bulk of the a.c .
zeeman shift has been compensated , and use @xmath13 to represent the residual fluctuations ( which we take to be slowly varying compared with the gate s duration ) .
if @xmath14 acted in the @xmath15 basis , it would commute with @xmath16 and this error could be suppressed by performing the gate inside a spin - echo sequence @xcite .
however , @xmath17 gates are not straightforward to implement with microwaves @xcite .
instead , we take advantage of the fact that @xmath14 does commute with a carrier drive of the qubit transition , provided that the carrier phase is chosen to produce rotations about the same axis of the bloch sphere as @xmath14 .
the corresponding hamiltonian is @xmath18 this carrier drive dynamically decouples the qubit from @xmath19 , as can be seen by considering the total system hamiltonian , , in the interaction picture with respect to @xmath20 : @xmath21 if @xmath22 the summed terms in oscillate rapidly and may be disregarded .
the rotating - frame hamiltonian then reduces to @xmath14 .
furthermore , setting @xmath23 for gate time @xmath24 and integer @xmath25 ensures that the rotating frame coincides with the lab frame at @xmath24 , so that an error - free ms gate is achieved in both frames . the requirement that may be avoided by using a composite gate sequence , with an additional @xmath26-pulse on each ion mid - way through the gate to refocus any partially complete carrier rabi oscillations @xcite ( figure [ fig : results ] ) . for this to work ,
the gate must be composed of an even number of phase - space loops , so that the @xmath26-pulse is applied while the ions spins are disentangled from their motion . in this case
, the gate is not sensitive to the exact value of @xmath27 , provided that the applied pulse area is the same for each half .
this sequence has the added benefit of being insensitive to transient a.c .
zeeman shifts at the beginning and end of each each half . moreover , if the @xmath26-pulse phase is chosen to give a rotation about the @xmath28-axis , errors due to drifts in the motional mode frequency are also suppressed @xcite . the dynamically - decoupled ms ( ddms ) gate described above is closely related to the `` single - sideband '' ( ssb ) gate proposed and demonstrated in @xcite , which uses only one of the red or blue sideband fields in combination with a carrier drive .
the ssb gate was originally introduced for use with lasers , where it has the advantage that , unlike the ms gate , it does not require interferometric stability between optical fields .
this advantage is inconsequential for microwave gates due to the relative ease of accurately controlling microwave phases . in their original proposal for the ssb gate
, the authors noted that their carrier drive technique could be extended to the standard ms gate @xcite .
our work develops this idea , identifying the importance of the relationship between the carrier and sideband phases ( which is not significant for the ssb gate ) , and providing numerical modelling ( supplementary material ) , as well as an experimental demonstration .
the principle advantage of the ddms gate is that , unlike the ssb gate and other `` dressed - state '' schemes @xcite , the carrier drive is merely used to suppress noise , rather than forming a fundamental part of the gate mechanism . as a result , when @xmath29 , the ddms gate exactly reproduces the ms dynamics at all times and for all values of @xmath27 .
this is not true for the ssb gate , which is consequently very susceptible to noise in @xmath30 ( see supplementary material ) .
this is a significant limitation of the ssb gate , potentially requiring the use of second - order driving fields to achieve high fidelities @xcite .
additionally , the ddms gate requires half the total microwave power to achieve a given gate speed , reducing the power dissipated in the ion trap chip .
-pulses , leaving all microwave frequencies and pulse durations set as for the two - qubit gate . for the red dataset ( squares )
, we repeat the measurement without the dynamical decoupling or the refocussing @xmath26-pulse .
( a ) we scan the phase @xmath31 of the second @xmath32-pulse to obtain a fringe .
lines show maximum - likelihood fits to the data , giving fringe contrasts of @xmath33 and @xmath34 .
( b ) we set @xmath31 to the fringe s peak and monitor fluctuations in @xmath35 over time
. the average of the blue data points gives @xmath36 , indicating the gate fields induce @xmath37 loss of qubit coherence .
the right - hand ordinate gives , for the red points , the simulated ms gate error corresponding to a given fringe contrast , assuming normally - distributed shot - to - shot a.c .
zeeman shift fluctuations .
all data have been corrected for spam errors . ] for the experimental implementation , we confine a pair of ions 75 above the surface - electrode ion trap described in @xcite .
we perform the gate on one of the ions radial rocking ( out - of - phase ) modes , whose secular frequency is 3.255 .
this mode was chosen over the 3.286 centre of mass ( com ) mode because of its lower heating rate , which was measured to be @xmath38 ( compared with @xmath39 for the com mode ) .
we generate the paul trap radiofrequency drive ( 38.3 , @xmath40 amplitude ) using a home - built high - stability source , which reduces fluctuations in the motional mode frequency to @xmath41 r.m.s .
we further suppress errors due to the residual mode frequency fluctuations by cooling the rocking mode close to its ground state with raman sideband cooling @xcite .
we also ground - state cool the spectator rocking mode to minimise dephasing due to cross - phase modulation @xcite . in future experiments
, ground - state cooling could be achieved using microwave sideband cooling @xcite .
we generate microwaves by upconverting r.f . at @xmath42300 from a commercial direct digital synthesis ( dds )
source lo . ] . the rsb and bsb are generated and amplified separately , before being combined on a quadrature hybrid .
after the hybrid , a custom cavity filter removes noise ( which had been observed to excite microwave spectator transitions during the gate ) from the signal before it is fed to one of the trap s microwave electrodes . to minimise the effect of off - resonant spin - flips
, we turn the rsb and bsb on / off adiabatically with a rise / fall time of 3 .
additionally , we pre - distort the sideband pulse envelope to compensate for slow ( @xmath43 ) power transients during the gate .
we use a slow digital feedback loop based on an ic power detector to reduce long - term drifts in the sideband power .
finally , we apply a 200 zero - peak linear ramp to the rsb and bsb dds frequencies during the gate to compensate for an observed `` chirp '' in the radial mode frequency ( which may originate from thermal transients in the trap caused by the microwaves ) . using 2 in each sideband , we achieve a gate rabi frequency of @xmath44 . the resulting differential a.c .
zeeman shift is @xmath45 . for the carrier drive
, we apply 3 to a second trap electrode , giving @xmath46 .
further details about the microwave fields are given in the supplement .
the gate sequence is shown in figure [ fig : results](a ) .
the gate consists of 4 loops in motional phase - space , with a @xmath28-axis @xmath26-pulse ( @xmath47 duration ) mid - way through .
the total gate time is @xmath48 .
we measure the fidelity of the bell state produced by the gate using the standard tomographic method described in @xcite .
the population after the gate is @xmath49 . combining this with the parity measurement shown in figure [ fig : results](b )
, we calculate a fidelity of 99.7(1)% . here , we have corrected for the independently - measured state - preparation and measurement ( spam ) error of 0.34(3)% per qubit @xcite . a significant source of spam error ,
not present in @xcite , is two - ion fluorescence detection using a photomultiplier , which could be significantly reduced using a camera @xcite . to estimate the benefit of the ddms gate over the basic ms scheme , we perform the single - ion ramsey experiment shown in figure [ fig : ac_zeeman_noise ] . here ,
the rsb and bsb are set up as for a gate on the two - ion rocking mode , leaving them @xmath50 detuned from the nearest single - ion motional mode . as a result
, they create an a.c .
zeeman shift without coupling to the ion s motion .
the fluctuations in this a.c .
zeeman shift are determined from the resulting loss of fringe contrast . without the dynamical decoupling and refocussing pulse
, we measure a fringe contrast of 0.924(6 ) . assuming normally - distributed shot - to - shot fluctuations in the a.c .
zeeman shift , this corresponds to @xmath51 r.m.s .
, which would give a ms gate error of 5.6(5)% ( figure 3(b ) ) . with the dynamical decoupling and refocussing pulse
, we find no loss of fringe contrast at the level of the measurement s sensitivity .
this experiment also implies that the ddms gate fields would introduce @xmath37 error on `` idle '' memory qubits stored in neighbouring trap zones in a multi - zone architecture .
the measured two - qubit gate error is consistent with the @xmath52 error expected from the rocking mode heating rate and the @xmath53 error expected from the independently - measured @xmath54 r.m.s .
fluctuations in the rocking - mode frequency .
we infer from the data in figure [ fig : ac_zeeman_noise ] that errors due to off - resonant excitation and a.c .
zeeman shift contribute @xmath55 error .
similarly , from the agreement between theory and data in figure [ fig : results](a ) , we estimate the error due to systematic mis - calibration in the sideband rabi frequencies or gate time to be @xmath55 .
.summary of errors in elementary qubit operations achieved in the present experimental apparatus for the ( , ) hyperfine `` atomic clock '' qubit .
the readout error could be reduced to the @xmath56 level measured in @xcite using spatially - resolved fluorescence detection @xcite .
addressed single - qubit gates with @xmath57 error were demonstrated for hyperfine qubits using microwave techniques in a similar surface trap by warring _ et al .
_ @xcite . [ cols="^,^,^",options="header " , ] in conclusion , we have introduced a dynamically - decoupled two - qubit gate scheme for trapped - ions , which we have implemented with @xmath58 fidelity using near - field microwave techniques in a room - temperature microfabricated surface trap .
the gate was applied to hyperfine qubits , for which state - of - the - art single - qubit performance was previously demonstrated in the same apparatus ( table [ t : summary ] ) .
present limits to the gate speed and fidelity are purely technical , and could be improved significantly in future experiments .
heating rates can be decreased using surface cleaning techniques @xcite or cryogenic operation @xcite .
the gate speed could be substantially increased , thereby also reducing its sensitivity to heating and motional mode frequency fluctuations , by moving the ion closer to the trap electrodes or increasing the microwave power .
off - resonant excitation and a.c .
zeeman shifts could be reduced by nulling the microwave field using multiple electrodes @xcite or improved trap geometries @xcite .
the dynamical decoupling demonstrated here should prove to be particularly effective when used in combination with these more advanced trap designs , as it significantly reduces the level of field suppression that must be achieved .
it may also be useful in mitigating the effects of crosstalk between different trap zones in a multi - zone architecture , arising from the large microwave fields required for multi - qubit gates .
finally , we note the ddms gate may be useful for laser - driven gates on optical or hyperfine qubits , where a.c .
stark shifts can present a significant experimental complication @xcite .
laser - driven two - qubit gates with comparable fidelity have recently been implemented in a surface trap for hyperfine qubits at sandia national laboratories @xcite .
we thank a. m. steane and d. n. stacey , and members of the nist ion storage group , in particular t. r. tan and d. slichter , for helpful discussions , and a. bermudez for comments on the manuscript .
this work was supported by the u.k .
epsrc `` networked quantum information technology '' hub and the u.s .
army research office ( ref . w911nf-14 - 1 - 0217 ) . |
the solar wind ends in a termination shock ( e.g. * ? ? ?
* ) , where the pressure of the heliosphere balances the ram pressure of the surrounding interstellar medium ( ism ) .
massive stars with more energetic winds generate much stronger shocks . in cases where the relative motion between the star driving the wind and the ambient ism is large ,
the shock will be bent back around the star .
if the relative velocity is supersonic , the ambient ism gas is swept into a second shock , forming an arc - shaped `` bowshock '' that is separated from the termination shock by a contact discontinuity .
stellar - wind bowshocks have been reported for a variety of sources , including nearby runaway o stars @xcite , high - mass x - ray binaries @xcite , ll ori - type stars @xcite , radio pulsars @xcite , galactic center o stars @xcite , and mass - losing red giants @xcite .
recently , an infrared ( ir ) bowshock has been observed around the young a - type star @xmath1 vel @xcite .
cometary regions also resemble bowshocks , due either to density gradients in the ambient gas or to motion of the ionizing source with respect to the interstellar surroundings @xcite . both the direction of a bowshock and its `` standoff distance '' from the star generating the wind are determined by the velocity of the star with respect to the surrounding medium . in the case of runaway o stars ,
this is dominated by the high space motion of the star .
we report the detection of three mid - ir bowshocks in each of two massive star formation regions : m17 and rcw49 .
two of the bowshocks in m17 are around known o stars .
we will demonstrate that the other bowshocks are also around likely o stars .
since these stars are in or near expanding regions , we find that the direction of the bowshock is determined principally by the flow of the ism rather than the space motion of the star .
the galactic legacy infrared mid - plane survey extraordinaire ( glimpse ; * ? ? ?
* ) programs have mapped the inner galactic midplane ( @xmath2 ) using the irac instrument on the _ spitzer space telescope _ ( 2 resolution ; * ? ? ?
rcw49 , located at @xmath3 , was observed as part of the validation of the glimpse observing strategy .
the same area of the sky was imaged 10 times with 2-second exposures , and the combined data were mosaicked together to produce high resolution images ( 0.6 pixels ) of this region in the four irac bands : 3.6 , 4.5 , 5.8 , and 8.0 @xmath4 m .
an overview of these observations was given by @xcite .
m17 , at @xmath5 , was included in the glimpse survey area , with two visits on the sky combined to make mosaics with 0.6 pixels .
@xcite studied the diffuse emission morphology of m17 at multiple wavelengths from ir to radio , constraining the distance to 1.41.9 kpc .
we will assume the widely - adopted value of 1.6 kpc for the m17 distance , following @xcite .
the winds and radiation of the o stars in the ngc 6618 cluster have excavated a cavity in the center of the m17 region .
the cavity is filled with hot , x - ray - emitting gas from shocked stellar winds @xcite .
@xcite noted the presence of 3 apparent stellar - wind bowshocks along the northern wall of the cavity , along with a prominent `` elephant trunk , '' or pillar , all oriented in the direction of the central ring of 7 o stars in ngc 6618 . these structures are highlighted in figure [ m17 ] , a glimpse image of m17 with an enlargement of the region containing the bowshocks .
the bowshocks stand out as yellow - orange features in the image because they are faint at 3.6 @xmath4 m ( colored _ blue _ ) compared to the other 3 irac bands .
we designate these stellar - wind bowshocks in figure [ m17 ] with the name of the region followed by an identification number , in order of increasing galactic longitude ( for example , m17-s1 ) .
ir fluxes for all of the bowshock driving stars are given in table 1 , and background - subtracted ir fluxes from aperture photometry of the bowshocks are listed in table 2 .
crrcccccc m17-s1 & 15.07486 & -0.64607 & 196 & 214 & 186 & 96 & 67 & + m17-s2 & 15.08126 & -0.65699 & 427 & 534 & 495 & 257 & 181 & 120 + m17-s3 & 15.10325 & -0.64867 & 103 & 200 & 238 & 151 & @xmath6 & @xmath7 + rcw49-s1 & 284.07646 & -0.43228 & 24 & 40 & 46 & 28 & 21 & 15 + rcw49-s2 & 284.30107 & -0.37121 & 10 & 18 & 20 & 13 & @xmath8 & @xmath9 + rcw49-s3 & 284.33999 & -0.28269 & 83 & 97 & 90 & 47 & 32 & 27 + crrcccc m17-s1 & 15.0744 & -0.6465 & @xmath10 & @xmath11 & @xmath12 & @xmath13 + m17-s2 & 15.0791 & -0.6613 & @xmath14 & @xmath15 & @xmath16 & @xmath17 + m17-s3 & 15.1026 & -0.6503 & @xmath18 & @xmath19 & @xmath20 & @xmath21 + rcw49-s1 & 284.0775 & -0.4305 & @xmath22 & @xmath23 & @xmath24 & @xmath25 + rcw49-s2 & 284.3018 & -0.3712 & @xmath26 & @xmath27 & @xmath28 & @xmath29 + rcw49-s3 & 284.3388 & -0.2829 & @xmath30 & @xmath31 & @xmath32 & @xmath33 + & + & & & & & + rcw49-s1 & & @xmath34 & @xmath35 & @xmath36 & @xmath37 + & + & & & & & + rcw49-s1 & & @xmath38 & @xmath39 & @xmath40 & @xmath41 + spectral types of two of the driving stars have been determined photometrically @xcite and spectroscopically @xcite .
m17-s1 is associated with cen 16 , an o9b2 star .
the larger bowshock , m17-s2 , is driven by cen 18 , an earlier - type star ( o7o8 ) .
both bowshocks were detected at 10.5 and 20 @xmath4 m by @xcite .
these observations did not resolve the arc shapes of the bowshocks , and @xcite attributed the excess ir emission to circumstellar disks and classified cen 16 and irs 9 ( the star visible just to the right of the arrowhead in figure [ m17 ] ) as massive protostars . m17-s3 lies outside of the field analyzed by @xcite , and the driving star does not appear in any catalog of the region .
it is not found in any glimpse sourcelist , because the bright , spatially variable diffuse background prevented the automatic extraction of the point source .
we measured the flux this source manually using a 5 aperture . using the spectral energy distribution ( sed ) fitting tool of @xcite , we fit the ir fluxes of this star ( table 1 ) with @xcite stellar atmosphere models . following the method described by @xcite ,
we scale the models to the 1.6-kpc distance of m17 and estimate a spectral type of o7 v for the star . carrying out the same analysis on 7 other o stars in m17 with independently known spectral types @xcite
, we estimate that our spectral typing is accurate to within 2 subclasses .
rcw49 presents a more complicated morphology than m17 .
@xcite discussed the structure and spectrum of the diffuse emission .
like m17 , rcw49 is filled with x - ray gas @xcite .
the interstellar structures are dominated by two large cavities .
the first , blown out to the west , contains the massive young cluster westerlund 2 , and the second is an enclosed bubble around the wolf - rayet star wr 20b ( figure [ rcw49 ] ) .
the distance to rcw49 remains disputed , with recent estimates placing the region as close as 2.8 kpc @xcite or as far as 8.0 kpc @xcite , with a kinematic distance estimate of @xmath42 kpc @xcite .
we consider distances of both 4.2 kpc @xcite and 6 kpc in our analysis .
we have found three bowshocks associated with rcw49 , shown in three separate insets in figure [ rcw49 ] .
rcw49-s1 is unique among our sample , because it lies relatively far from the region , but we believe it is associated with rcw49 for the following reasons : ( 1 ) the rcw49 distance is consistent with the luminosity of the driving star being an o star ( see below ) ; and ( 2 ) the bowshock points ( approximately ) toward the central cluster , westerlund 2 . because it is far from any other bright ir source , rcw49-s1 appears as a point source in both the _ midcourse space experiment _ ( _ msx _ ; * ? ? ?
* ) and _ infrared astronomical satellite _ ( _ iras _ ; * ? ? ?
* ) point source catalogs , and these fluxes are also given in table 2 .
rcw49-s2 is oriented away from westerlund 2 and appears to be influenced primarily by the nearby wr 20b .
rcw49-s3 is oriented in the general direction of westerlund 2 .
none of the rcw49 bowshocks points directly back toward the central cluster , while all three m17 bowshocks do .
perhaps the expanding bubbles driven by westerlund 2 and the wolf - rayet stars are interacting turbulently , producing non - radial components to the flows .
it is also possible that the bowshock driving stars have large orbital motions relative to the dynamic interstellar medium .
the three bowshock driving stars in rcw49 are of previously undetermined spectral type , so we again estimate the spectral type by fitting model seds to the broadband fluxes of table 1 , scaled to both the 4.2 kpc and 6 kpc distances .
the results are given in table 3 .
lccccccc[ht ] + m17-s1 & o9b2 v & @xmath43 & 0.03 & 1.4 & @xmath44 + m17-s2 & o7o8 v & 0.52.5 & 0.11 & 0.9 & 1022 + m17-s3 & o7 v & 0.52.5 & 0.05 & 1.7 & 2147 + + rcw49-s1 & o6 v & @xmath45 & 0.16 & 16.2 & 11 + rcw49-s2 & o9 v & @xmath46 & 0.04 & 2.46 & 17 + rcw49-s3 & o5 v or o9 iii & 2.53.2 & 0.096 & 6.13 & 2528 + + rcw49-s1 & o5 iii & @xmath47 & 0.23 & 23.2 & 26 + rcw49-s2 & o6 v & @xmath45 & 0.06 & 3.52 & 28 + rcw49-s3 & o3 v or o6.5 iii & @xmath48 & 0.14 & 8.8 & @xmath49 + all three stars are plausibly o stars . assuming 4.2 kpc , the driving star of rcw49-s1 is fit as o6 v , rcw49-s2 as o 9 v , and rcw49-s3 as o5 v ( or o9 iii ) .
if we increase the distance to 6 kpc , the fits become o 5 iii , o6 v , and o5 v ( or o6.5 iii ) , respectively ( see * ? ? ?
* for an explanation of the degeneracy between luminosity classes ) .
the spectral types at 6 kpc seem improbably luminous .
highly luminous and windy stars dominate the dynamics of their local ism .
we do not observe ir bowshocks around any of the earliest - type stars in either rcw49 or m17 , because they have blown large , evacuated cavities in the centers of the regions .
little or no ambient material remains close to the stars to produce a bowshock .
apart from the wolf - rayet systems , one of the earliest stars in rcw49 is g284.2642 - 00.3156 . using optical spectroscopy ,
@xcite classified this star as o4 v(f ) and derived a spectrophotometric distance of @xmath50 kpc .
using ir sed fitting , we confirm that the luminosity of this star is consistent with an o4 v star at 3.2 kpc . yet
this distance falls on the low end of the range of published distance estimates for rcw49 , and we do not adopt it here . like cen 1 in m17 , g284.2642 - 00.3156 may be an unresolved , equal - mass o4o4 binary .
doubling the luminosity moves the spectrophotometric distance from 3.2 to 4.5 kpc , in agreement with a distance of 4.2 kpc but inconsistent with 6 kpc .
the standoff distance @xmath51 of a bowshock from its driving star is the point where the momentum flux of the stellar wind balances the momentum flux of the ambient medium : @xmath52 .
following @xcite , we normalize the stellar wind mass loss rate , @xmath53 m@xmath54 yr@xmath55 , stellar wind velocity @xmath56 cm s@xmath55 , ambient hydrogen particle density of @xmath57 @xmath58 , and use a mean ism gas mass per hydrogen atom @xmath59 g. assuming a spherically symmetric stellar wind with a mass - loss rate given by @xmath60 , the velocity of the star with respect to the ambient ism can be written as @xmath61,\ ] ] where @xmath62 is the terminal velocity of the stellar wind
. values of @xmath63 and @xmath51 for each bowshock are presented in table 3 .
the standoff distance can be measured only as @xmath64 on the sky , where @xmath65 is the inclination ( or viewing angle ) made by the line connecting the star with the apex of the bowshock against the plane of the sky ( @xmath66 if the line joining the star to the bowshock apex lies in the plane of the sky ) . because a bowshock oriented at high @xmath65 will not produce an arc morphology , it is likely that @xmath67 , and hence @xmath68 will not differ greatly from unity in our measurements . for @xmath69 is 0.9 . ]
the distance from each bowshock to the likely source of the large - scale ism flow , measured on the sky as @xmath70 , are also presented for reference in table 3 .
the mass - loss rates and stellar wind velocities in equation 1 suffer from high dispersion as a function of spectral type @xcite , a factor of 2 or even greater , and this is compounded by a comparable level of uncertainty in the spectral types .
the uncertainty on our measurements of @xmath51 ranges from @xmath71 for the largest bowshocks ( m17-s2 and rcw49-s1 ) to @xmath72 for the smaller bowshocks that are barely resolved by irac .
we estimate that @xmath73 is uncertain by a factor of 2 in m17 and up to a factor of 3 in rcw49 , where the spectral types of the driving stars are less constrained .
these uncertainties , reflected in the range of values for @xmath73 in table 3 , are dominated by the uncertainty in the stellar wind properties .
we have neglected the effects of turbulent pressure in our calculation of the momentum flux balance of the bowshocks .
this is a potentially significant contributor to the total ism pressure held off by the bowshocks .
the effect of turbulent pressure would be to systematically decrease the standoff distance @xmath51 , causing us to overestimate @xmath74 .
in reality , for the fast winds of early - type stars , the observed bowshock is displaced from the standoff distance by a significant amount .
this happens because the cooling timescale of the shocked stellar wind is very long .
the result is a thick layer of hot gas intervening between the wind terminal shock at the standoff distance and the thin , dense layer of interstellar gas and dust forming the observed bowshock .
the numerical simulations of @xcite predict that the bowshock should be located at twice the standoff distance from the driving star . in this case ,
our derived values of @xmath73 in table 3 would be underestimated by the same factor of 2 .
this systematic correction is comparable to the intrinsic uncertainties in our estimates of @xmath73 , and it partially compensates for the effects of neglecting turbulent pressure in the ambient ism .
therefore , the assumption that the observed distance of the bowshocks from the driving stars corresponds to the standoff distance @xmath51 should not have a large impact on our results , and we find that the cautious application of equation 1 yields reasonable results .
orbital velocities of o stars in massive clusters are typically @xmath75 .
the expansion speed of ionized gas in regions is generally comparable to the sound speed of @xmath76 , and this appears to be true in m17 @xcite .
most of the bowshocks ( the exception being rcw49-s1 ) are apparently located within the ionized gas of the radio regions .
the likely explanation for the bowshock emission in the ir is that dust in the regions @xcite is swept - up by the bowshocks .
the observed average electron density in the northern bar of the m17 region is @xmath77 @xmath58 @xcite , so the values of @xmath78 listed in table 3 are likely to be close to the actual relative velocities of the stars and the ism for most of the bowshocks .
the values of @xmath63 calculated for m17-s1 and -s2 are in good agreement .
m17-s3 , however , presents a different picture .
m17-s3 appears to be associated with a `` teardrop '' structure ( figure 1 ) in the photodissociation region ( pdr ) , near the ionization front .
the ambient density surrounding this star could be significantly higher than the density within the region .
if @xmath74 for this bowshock is comparable to that of the other 2 bowshocks in m17 , then @xmath79 @xmath58 in this location
. this value agrees with the measurements of electron density in the dense clumps of ionized gas in m17 @xcite .
the unusual morphology of the diffuse ir emission associated with m17-s3 suggests that the star may have recently emerged from an evaporating globule on the edge of the pdr ; a larger , more evolved analog of the nearby pillar structure seen in figure 1 . in rcw49 , all 3 bowshocks are found in very different locations , but we note that 2 of the bowshocks appear to be similar in size , color , and ambient environment ( figure 2 ) .
the largest bowshock , rcw49-s1 , is different , since it is located relatively far ( 16.2 pc at 4.2 kpc ) from westerlund 2 , outside the region .
the presence of rcw49-s1 , along with its orientation , indicates that rcw49 vents a large - scale flow of gas through the cavity opening to the west .
this flow likely originates in the combined winds of the westerlund 2 cluster and thus should be much more diffuse than the region gas @xcite . assuming a density of 1 @xmath58 in the flow from westerlund 2 , the standoff distance of rcw49-s1 at 4.2 kpc gives a flow velocity of @xmath80 .
such a high value of @xmath74 is reasonable , given that the gas must move supersonically relative to the star to produce a shock , and the sound speed in the hot , rarefied gas of the flow streaming away from the region is @xmath81 . because rcw49-s1 was detected by _ msx _ and _ iras _ in addition to glimpse , we can construct the sed of the bowshock from 4.5 @xmath4 m to 100 @xmath4 m ( table 2 ) .
we computed sed models of rcw49-s1 using a 3-d radiative equilibrium code @xcite modified to include very small grain ( vsg ) and pah emission @xcite .
we used the canonical mass fraction for vsg / pah grains of 5% @xcite .
we modeled the bowshock geometry as a paraboloid with the apex offset by @xmath82 pc from the o6 v star ( assuming @xmath66 and the 4.2 kpc distance ) . for models that reproduced the observed images , the sed shape was insensitive both to the thickness of the shock and to the radial density profile of the dust ( with the total mass scaled to match the observed sed ) .
the black line in figure [ model ] shows the sed for a model with a density varying as @xmath83 from the 0.16 pc standoff distance out to a radius of 1.5 pc .
the observed sed shortward of 30 @xmath4 m is well - matched by the model . for radial density exponents from @xmath84 to 0 , the mass of this material ranges from @xmath85 m@xmath86 to 2 m@xmath86 , respectively , assuming a dust - to - gas mass ratio of 0.01 .
the optical depth in all models is low , with @xmath87 within 1 pc . to match the _ iras _ data at 60 and 100 @xmath4 m requires low - density material farther from the star .
the _ green _ line in figure [ model ] is a model including a shell 23 pc from the star .
this material added mid - ir pah emission , so we lowered the vsg / pah mass fraction to 3% to continue to match the mid - ir sed , and the image still matches the data well .
the @xmath88 through the bowshock in this model is 0.25 , most of it due to the outer shell . along the line of sight to the star , @xmath89 , so the bowshock and shell contribute a negligible fraction of the interstellar extinction .
these models show that dust distributed in a bowshock geometry matches both the images and the sed reasonably well .
the ir emission from the bowshock can be explained by reprocessed stellar radiation without any additional dust heating by the shock .
the structure of rcw - s1 is probably significantly different from that of the other 5 bowshocks in our sample , because it is located in a distinct interstellar environment . because of the high temperature and low density in the flow outside the region , the shocked interstellar gas can not cool quickly enough to form a dense layer behind the bowshock , as likely happens in the other 5 cases .
the shock is approximately adiabatic , remaining very hot and only moderately compressed ( by a factor of @xmath90 ) as it flows past the star , forming a relatively thick layer .
hence , while rcw - s1 is the only bowshock in our sample observed at enough different mid - ir wavelengths to allow us to create a meaningful model of the emission , it may not be appropriate to draw strong conclusions about the remaining bowshocks based upon this model .
we have observed 6 prominent ir bowshocks in m17 and rcw49 .
these objects appear to be produced by the winds of individual o stars colliding with large - scale interstellar gas flows in the regions .
one bowshock , m17-s3 , may be the leading edge of an evaporating globule containing a newly - formed and previously undiscovered o star in the well - studied m17 region .
all three bowshocks associated with rcw49 lead us to identify new candidate o stars . our stellar classifications also suggest that the true distance to rcw49 is less than the kinematic distance of 6 kpc .
the bowshocks are bright at ir wavelengths due to emission from dust swept up from the ambient ism and heated by radiation from the bowshock driving stars . as @xcite note , ir excess emission from a bowshock could be attributed to the presence of a circumstellar disk , particularly when the bowshock morphology is not spatially well - resolved .
this can be a pitfall for observational studies of accreting massive protostars .
the collective winds of the most luminous stars in young , massive clusters produce overlapping large - scale flows that hollow out thermally hot cavities in the parent molecular cloud @xcite .
the largest bowshock presented here , rcw49-s1 , is evidence that the combined winds of the ionizing stars in westerlund 2 have escaped the region , creating a flow of hot gas moving at a few @xmath0 that extends at least 16 pc away from rcw49 .
the driving stars of the other 5 bowshocks are surrounded by ionized gas and dust of their natal regions , where the density of the ambient medium ( @xmath91 @xmath58 ) is sufficiently high to produce the observed bowshocks with a relative velocity of only 1020 .
the winds of the bowshock driving stars do not directly encounter the @xmath92 winds from the most massive stars in the cluster .
instead , the bowshocks are shaped by the expansion of the ionized gas in the regions relative to the orbital motions of the stars .
eventually , supernova explosions will produce high velocity shock waves that heat and disperse the original gas cloud . in star forming regions like m17 and rcw49 that have not yet been disrupted by supernovae , ir bowshocks serve as interstellar `` weather vanes , '' indicating the speed and direction of large - scale gas flows at points within and around giant regions .
we thank the anonymous referee for incisive and very useful suggestions that helped us improve this work .
we are grateful to joseph cassinelli , eric pellegrini , john raymond , ellen zweibel , heidi gneiser , and don cox for useful conversations while preparing this paper .
m. s. p. thanks the members of the `` dissertator club , '' kathryn devine , k. tabetha hole , and nicholas murphy , for their helpful comments .
this work was supported by nsf grant ast-030368 ( e. b. c. ) and nasa / jpl contracts 1282620 and 1298148 .
additional support was provided by the nasa theory program ( nng05gh35 g ; b. a. w. ) .
r. i. acknowledges support from a spitzer fellowship at the time that these data were analyzed , and from jpl rsa1275467 .
arthur , s. j. , & hoare , m. g. 2006 , , 165 , 283 ascenso , j. , alves , j. , beletsky , y. , & lago , m. t. v. t. 2007 , , 466 , 137 bally , j. , odell , c. r. , & mccaughrean , m. j. 2000 , , 119 , 2919 beichman , c. a. , neugebauer , g. , habing , h. j. , clegg , p. e. , & chester , t. j. , eds .
1988 , iras catalogs and atlases : explanatory supplement , nasa rp-1190 benjamin , r.a .
2003 , , 115 , 953 brown , d. , & bomans , d. j. 2005 , , 439 , 183 bumgardner , t. e. 1992 , m.s .
thesis , ohio state univ . ,
columbus chini , r. , elssser , h. , & neckel , t. 1980 , , 91 , 186 ( cen ) churchwell , e. , bieging , j. h. , van der hucht , k. a. , williams , p. m. , spoelstra , t. a. th . , & abbott , d. c. 1992 , , 393 , 329 churchwell , e. et al .
2004 , , 154 , 322 comern , f. & kaper , l. 1998 , , 338 , 273 comern , f. & pasquali , a. 2007 , , 467 , l23 dame , t. m. 2007 , , 665 , l163 decker , r. b. et al .
2005 , science , 309 , 2020 draine , b. t. & li , a. 2007 , , 657 , 810 fazio , g. g. et al .
2004 , , 154 , 10 felli , m. , churchwell , e. , & massi , m. 1984 , , 136 , 53 france , k. , mccandliss , s. r. , & lupu , r. e. 2007 , , 655 , 920 fullerton , a. w. , massa , d. l. , & prinja , r. k. 2006 , , 637 , 1025 gaensler , b. m. , & slane , p. o. 2006 , , 44 , 17 gspr , a. , su , k. y. l. , rieke , g. h. , balog , z. , kamp , i. , martnex - galarza , j. r. , & stapelfeldt , k. 2008 , , 672 , 974 geballe , t.r . , rigaut , f. , roy , j .-
, & draine , b.t .
2004 , , 602 , 77 geballe , t. r. , najarro , f. , rigaut , f. , & roy , j .-
2006 , , 652 , 370 hanson , m. m. , howarth , i. d. , & conti , p. s. 1997 , , 489 , 698 huthoff , f. , & kaper , l. 2002 , , 383 , 999 kaper , l. , van loon , j. t. , augusteijn , t. , goudfrooij , p. , patat , f. , waters , l. b. f. m. , & zijlstra , a. a. 1997 , , 475 , l37 kurucz , r. 1993 , atlas9 stellar atmosphere programs and 2 km / s grid .
kurucz cd - rom no . 13 .
cambridge , ma : smithsonian astrophysical observatory martin , d , c. et al , 2007 , nature , 448 , 780 nielbock , m. , chini , r. , jtte , m. , & manthey , e. 2001 , , 273 , 284 noriega - crespo , a. , van buren , d. , & dgani , r. 1997 , , 113 , 780 pellegrini , e. w. et al .
2007 , , 658 , 1119 povich , m. s. , stone , j. m. , churchwell , e. , zweibel , e. g. , wolfire , m. g. , babler , b. l. , indebetouw , r. , meade , m. r. , & whitney , b. a. 2007 , , 660 , 346 povich , m. s. et al .
2008 , , submitted price , s. d. , egan , m. p. , carey , s. j. , mizuno , d. r. , and kuchar , t. a. 2001 , , 121 , 2842 rauw , g. , manfroid , j. , gosset , e. , naz , y. , sana , h. , de becker , m. , foellmi , c. , & moffat , a. f. j. 2007 , , 463 , 981 robitaille , t. p. , whitney , b. a. , indebetouw , r. , & wood , k. 2007 , , 169 , 328 townsley , l. k. , feigelson , e. d. , montmerle , t. , broos , p. s. , chu , y .- h . , & garmire , g. p. 2003
, , 593 , 874 townsley , l. , feigelson , e. , montmerle , t. , broos , p. , chu , y .- h .
garmire , g. , & getman , k. 2005 , x - ray and radio connections , eds.l.o .
sjouwerman & k.k dyer ( nrao : santa fe , nm ) uzpen , b. et al .
2005 , , 629 , 512 van buren , d. & mccray , r. 1988 , , 329 , 93 van buren , d. , mac low , m. m. , wood , d. o. s. , & churchwell , e. 1990 , , 353 , 570 van buren , d. , noriega - crespo , a. , & dgani , r. 1995 , , 110 , 2914 vink , j. s. , de koter , a. , & lamers , h. j. g. l. m. 2001 , , 369 , 574 watson , c. , povich , m. s. , churchwell , e. b. , babler , b. l. , chunev , g. hoare , m. , indebetouw , r. , meade , m. r. , & whitney , b. a. 2008 , , 681 , 1341 ( astro - ph 0806.0609 ) whitney , b. a. , wood , k. , bjorkman , j. e. , & wolff , m. j. 2003 , , 591 , 1049 wood , k. , whitney , b. a. , robitaille , t. p. , & draine , b. t. 2008 , , in press ( astro - ph 0807.2398 ) | we report the discovery of six infrared stellar - wind bowshocks in the galactic massive star formation regions m17 and rcw49 from _ spitzer _ glimpse ( galactic legacy infrared mid - plane survey extraordinaire ) images .
the infrared array camera ( irac ) on the _ spitzer space telescope _ clearly resolves the arc - shaped emission produced by the bowshocks .
we combine _ two micron all - sky survey ( 2mass ) , _ _ spitzer , _ _ msx , _ and _ iras _ observations to obtain the spectral energy distributions ( seds ) of the bowshocks and their individual driving stars .
we use the stellar seds to estimate the spectral types of the three newly - identified o stars in rcw49 and one previously undiscovered o star in m17 .
one of the bowshocks in rcw49 reveals the presence of a large - scale flow of gas escaping the region at a few @xmath0 .
radiation - transfer modeling of the steep rise in the sed of this bowshock toward longer mid - infrared wavelengths indicates that the emission is coming principally from dust heated by the star driving the shock .
the other 5 bowshocks occur where the stellar winds of o stars sweep up dust in the expanding regions . |
+ davide girolami , tommaso tufarelli , and gerardo adesso we refer to @xcite for a summary of the relevant properties of the skew information which constitute the main ingredients of the proofs . in a bipartite system @xmath59 , classically correlated states @xmath113 with respect to measurements on @xmath8 , also known as @xmath8-classically correlated states or classical - quantum states , are states with zero quantum discord on @xmath8 . for these states there
exists at least one set of projectors @xmath114 such that @xmath115 .
the @xmath8-classically correlated states take in general the form @xmath116 with @xmath11 denoting an orthonormal basis for subsystem @xmath8 . to prove that @xmath8-classically correlated states have vanishing lqu @xmath117 , it is sufficient to define the observable @xmath118 where @xmath119 is diagonal in the basis defined by @xmath120 , to obtain @xmath121=0 $ ] which means @xmath122 . on the other hand ,
a vanishing lqu ensures the existence of a local observable @xmath123 such that @xmath124 .
hence @xmath125 commutes with the density matrix , and we can diagonalize them simultaneously . since the observable is assumed nondegenerate , its eigenvectors define a _ unique _ basis on @xmath8 ( up to phases ) , say @xmath126 . then , an eigenvector basis for @xmath127 will be simply @xmath128 , and the state must necessarily be of the form @xmath129 , which is a zero discord state .
this proves that @xmath130 vanishes if and only if @xmath9 is an @xmath8-classically correlated state .
let us now show that the lqu is invariant under local unitary transformations .
we have + @xmath131 + @xmath132 + @xmath133 + @xmath134 , + as minimizing over the local observables @xmath135 is obviously equivalent to do it over the ones rotated by @xmath136 .
we then note that the skew information @xmath137 is contractive under completely positive and trace - preserving maps @xmath138 on @xmath26 , @xmath139 .
consequently , the lqu inherits this property .
denoting as @xmath127 the most certain observable for @xmath9 , we have @xmath140 .
finally , for pure states @xmath141 , the lqu reduces to the variance of @xmath135 minimized over all local observables @xmath135 . in the next section we present a proof ( which can be of independent interest ) that such a quantity decreases monotonically under local operations and classical communication , so that the lqu , alias minimal local variance , reduces to an entanglement measure on pure states .
+ [ permutanda ] consider a @xmath142-dimensional density matrix @xmath9 , and the set @xmath143 of all observables with fixed spectrum @xmath144 .
then , the variance @xmath145 is minimised by an observable @xmath146 commuting with @xmath9
. working in the eigenbasis of the density matrix , one has the representation @xmath147 .
an observable in the considered set can the be written as @xmath148 , where @xmath149 is a unitary transformation .
the variance of @xmath12 on the state @xmath9 reads ( @xmath150 ) @xmath151 ^ 2\qquad p_{ij}\equiv p_i\lambda_j^2,\;q_{ij}\equiv p_i\lambda_j,\;b_{ij}\equiv|v_{ij}|^2.\label{boh}\ ] ] note that @xmath26 is a unistochastic matrix , and in fact , any unistochastic matrix is expressible as @xmath152 for some unitary @xmath149 .
hence , the problem of minimizing the variance can be equivalently formulated as a minimization of the right hand side of eq . over the set of unistochastic matrices . since
every unistochastic matrix is also bistochastic ( but not vice - versa ) , one has , in general @xmath153 ^ 2\right],\label{varbound}\ ] ] where @xmath154 is the set of all @xmath155 bistochastic matrices .
one can now exploit the birkhoff - von neumann theorem , and express a generic bistochastic matrix as a convex sum of permutations of the form @xmath156 , where the @xmath157 s are probabilities and @xmath158 is the set of permutation matrices in dimension @xmath142 , which has @xmath159 elements . then , @xmath160 ^ 2\right]&=\min_{\{q_k\}}\left[\sum_kq_k{\text{tr}}\{ps_k\}-\left(\sum_kq_k{\text{tr}}\{qs_k\}\right)^2\right]\geq\min_{\{q_k\}}\sum_kq_k\left[{\text{tr}}\{ps_k\}-\left[{\text{tr}}\{qs_k\}\right]^2\right]\nonumber\\ & \geq\sum_kq_k\left[{\text{tr}}\{ps_\text{min}\}-\left[{\text{tr}}\{qs_\text{min}\}\right]^2\right]={\text{tr}}\{p s_\text{min}\}-[{\text{tr}}\{qs_\text{min}\}]^2,\\end{aligned}\ ] ] where we have exploted the convexity of the square , and @xmath161 is a particular permutation that minimises the expression @xmath162 $ ] . such minimizing permutation
can always be found since @xmath158 is a finite set . noting that permutations are also unistochastic matrices , the above steps imply that the equality sign in eq .
can be always achieved : @xmath163 ^ 2=\sum_i p_i\lambda_{\mathcal{p}(i)}^2-\left(\sum_ip_i\lambda_{\mathcal{p}(i)}\right)^2,\ ] ] where @xmath164 indicates the permutation of the indices associated to the matrix @xmath161 .
this implies that the variance is minimised by an observable of the form @xmath165 , which clearly commutes with @xmath166 [ lem1 ] let @xmath167 .
suppose that @xmath168 .
under local operations on subsystem @xmath8 , a globally pure state @xmath169 evolves within a subspace @xmath170 , where @xmath171 is a @xmath172-dimensional subspace of @xmath173 .
we can suppose that @xmath169 is in schmidt form : @xmath174 clearly , @xmath175 , where @xmath171 is spanned by the @xmath172 orthonormal vectors @xmath176 .
a local operation on @xmath8 is described via kraus operators of the form @xmath177 . applying the operator on the state ,
one has : @xmath178 which is still a vector with support in @xmath170 .
[ lem2 ] suppose @xmath179 , and the ( non - degenerate ) spectrum of the @xmath8-observables is fixed as @xmath180 .
one has , @xmath181 where @xmath182 is the set of b - observables whose @xmath183 eigenvalues are non degenerate and are a _ subset _ of @xmath184 : @xmath185 .
we start by noting that , in general , @xmath186 .
in fact , by rotating @xmath169 to the schmidt form , we see that the variance of any observable @xmath187 is achieved by an operator @xmath38 on @xmath8 .
given @xmath188 such that @xmath189 , it is sufficient to choose @xmath135 such that @xmath190 .
the two operators clearly yield the same variance , since the labels @xmath191 do not affect its calculation .
note that it is always possible to pick @xmath135 in the above form since the operators on @xmath8 restricted to a @xmath183-dimensional subspace can assume the same form as any operator in @xmath182 .
we now show that the inequality @xmath192 is also verified , hence equality must hold .
the most certain observable on @xmath8 has to commute with the reduced state @xmath50 ( lemma [ permutanda ] ) .
hence , if the latter has eigenvalues @xmath193 , @xmath194 , there is an appropriate permutation @xmath164 such that : @xmath195 the latter equality is obtained by choosing @xmath196 diagonal in the same basis as @xmath197 , with eigenvalues @xmath198 , and by noting that @xmath50 and @xmath197 have the same eigenvalues . by lemma [ lem1 ] , we can suppose @xmath179 .
we already have invariance under local unitaries and contractivity under local operations on @xmath26 ( see appendix in the main text ) . to complete the proof we need to prove that , on average , the lqu of @xmath199 can not be increased under operations on @xmath8 .
let @xmath200 be the kraus operators on alice : @xmath201 .
the output ensemble is given by @xmath202 , where @xmath203 we want to demonstrate that @xmath204 .
suppose that @xmath205 is such that @xmath206 , as given by lemma [ lem2 ] .
@xmath207 in the first line , we used lemma [ lem2 ] . in the second line
, we have used that the variance is concave as a function of the state .
+ an interesting case study concerns the final state of the dqc1 ( discrete quantum computation with one bit ) model , a protocol designed for estimating the trace of a unitary matrix , say @xmath208 , applied on a @xmath51-qubit register @xcite .
discordlike correlations , but vanishing entanglement , are created between an ancillary qubit and the register in the output state @xcite .
the ancilla @xmath8 , in a state with arbitrary polarization @xmath209 , say @xmath210 , and the register @xmath26 , in a @xmath51-qubit maximally mixed state , i.e. , @xmath211 , are initially uncorrelated : @xmath212 .
the protocol returns the final state @xmath213 measuring the ancilla polarization in the output state yields an estimation of the trace of the unitary matrix : @xmath214\right ] , \langle\sigma_2\rangle_{\rho_a^{out}}=\text{im}\left[\text{tr}[u]\right]$ ] . for ` typical ' unitaries in high dimensions ( which have approximately zero trace @xcite ) , the entanglement between the ancilla and the @xmath51-qubit register is always negligible . on the other hand , we find the following .
we choose the basis @xmath216 on @xmath26 which diagonalizes @xmath208 : @xmath217 .
we may then rewrite eq . as @xmath218 , where @xmath219 and @xmath220 .
the square root of the density matrix can then be expressed as @xmath221 , where @xmath222 , where @xmath223 and the pair @xmath224 verify @xmath225 and @xmath226 .
both @xmath227 and @xmath228 do not depend on @xmath229 , while @xmath230 does .
the elements of the matrix @xmath44 are then given by @xmath231 now , we see that @xmath232 .
hence , @xmath233 substituting the explicit expressions for the components of @xmath230 , eq
. requires evaluation of the sums @xmath234 , @xmath235 , and @xmath236 .
we observe that for large @xmath51 and ` typical ' unitaries , where the phases @xmath237 are uniformly distributed @xcite , we can approximate those sums with integral averages of the trigonometric functions over the interval @xmath238 $ ] : @xmath239 . then , @xmath240 finally , the conditions given above on @xmath224 can be used to express @xmath227 in terms of the qubit initial polarization , as @xmath241 .
substituting this in eq .
yields the anticipated result of eq . .
as expected , the expression increases monotonically with the ancilla polarization and is independent of the number of qubits in the register .
this is in agreement with what predicted by using the quantum discord @xcite . | quantum mechanics predicts that measurements of incompatible observables carry a minimum uncertainty which is independent of technical deficiencies of the measurement apparatus or incomplete knowledge of the state of the system .
nothing yet seems to prevent a single physical quantity , such as one spin component , from being measured with arbitrary precision . here
we show that an intrinsic quantum uncertainty on a single observable is ineludible in a number of physical situations . when revealed on local observables of a bipartite system
, such uncertainty defines an entire class of _ bona fide _ measures of nonclassical correlations . for the case of @xmath0 systems
, we find that a unique measure is defined , which we evaluate in closed form .
we then discuss the role that these correlations , which are of the ` discord ' type , can play in the context of quantum metrology .
we show in particular that the amount of discord present in a bipartite mixed probe state guarantees a minimum precision , as quantified by the quantum fisher information , in the optimal phase estimation protocol . _
introduction._ in a classical world , error bars are exclusively due to technological limitations , while quantum mechanics entails that two noncommuting observables can not be jointly measured with arbitrary precision @xcite , even if one could access a flawless measurement device .
the corresponding uncertainty relations have been linked to distinctive quantum features such as nonlocality , entanglement and data processing inequalities @xcite .
remarkably , even a single quantum observable may display an intrinsic uncertainty as a result of the probabilistic character of quantum mechanics .
let us consider for instance a composite system prepared in an entangled state @xcite , say the bell state @xmath1 of two qubits .
this is an eigenstate of the global observable @xmath2 ( @xmath3 are the pauli matrices ) , so there is no uncertainty on the result of such a measurement . on the other hand ,
the measurement of _ local _ spin observables of the form @xmath4 ( where @xmath5 is a real vector ) is intrinsically uncertain .
indeed , the state @xmath6 , and in general any entangled state , can not be eigenstates of a local observable
. only uncorrelated states of the two qubits , e.g. @xmath7 , admit at least one completely ` certain ' local observable . .
an observer on subsystem @xmath8 is equipped with a _ quantum meter _ ,
a measurement device whose error bar shows the quantum uncertainty only ( note : in order to access such quantity , the measurement of other observables that are defined on the full bipartite system may be required , in a procedure similar to state tomography ) .
( a ) if @xmath9 is uncorrelated or contains only classical correlations ( brown shade ) , i.e. @xmath9 is of the form @xmath10 ( with @xmath11 an orthonormal basis for @xmath8 ) @xcite , the observer can measure at least one observable on @xmath8 without any intrinsic quantum uncertainty .
( b ) if @xmath9 contains a nonzero amount of quantum correlations ( yellow shade ) , as quantified by entanglement for pure states @xcite and quantum discord in general @xcite , any local measurement on @xmath8 is affected by quantum uncertainty .
the minimum quantum uncertainty associated to a single measurement on subsystem @xmath8 can be used to quantify discord in the state @xmath9 , as perceived by the observer on @xmath8 . in this letter
we adopt the wigner - yanase skew information @xcite to measure the quantum uncertainty on local observables.,width=207 ] extending the argument to mixed states , one needs to filter out the uncertainty due to classical mixing , i.e. , lack of knowledge of the state , in order to identify the genuinely quantum one .
we say that an observable @xmath12 on the state @xmath9 is ` quantum - certain ' when the statistical error in its measurement is solely due to classical ignorance . by adopting a meaningful quantitative definition of quantum uncertainty , as detailed later , we find that @xmath12 is quantum - certain if and only if @xmath13 , where @xmath14 is the density matrix of the state after the measurement of @xmath12 .
it follows that not only entangled states but also almost all ( mixed ) separable states @xcite can not admit any quantum - certain local observable .
the only states left invariant by a local complete measurement are those described within classical probability theory @xcite , i.e. , embeddings of joint probability distributions .
these are the states with zero _ quantum discord _ @xcite . the quantum uncertainty on local observables is then entwined to the notion of quantum discord ( see fig .
[ figurauno ] ) , a form of nonclassical correlation which reduces to entanglement on pure states , and is currently subject to intense investigations for quantum computation and information processing @xcite . in the following ,
an entire class of discordlike measures is defined , interpreted and analysed within the framework of local quantum uncertainty . _
skew information and local quantum uncertainty._ there are several ways to quantify the uncertainty on a measurement , and here we aim at extracting the truly quantum share .
entropic quantities or the variance , though employed extensively as indicators of uncertainty @xcite , do not fit our purpose , since they are affected by the state mixedness .
it has been proposed to isolate the quantum contribution to the total statistical error of a measurement as being due to the noncommutativity between state and observable : this may be reliably quantified via the _ skew information _
@xcite @xmath15 ^ 2\},\label{skewinfo}\end{aligned}\ ] ] introduced in @xcite and employed for studies on uncertainty relations @xcite , quantum statistics and information geometry @xcite .
referring to @xcite for the main properties of the skew information , we recall the most relevant ones : it is nonnegative , vanishing if and only if state and observable commute , and is convex , that is , nonincreasing under classical mixing .
moreover , @xmath16 is always smaller than the variance of @xmath12 , @xmath17 , with equality reached on pure states , where no classical ignorance occurs ( see fig . [ figurawerner ] ) .
hence , we adopt the skew information as measure of quantum uncertainty and deliver a theoretical analysi in which we convey and discuss its operational interpretation . as a central concept in our analysis
, we introduce the _ local quantum uncertainty _ ( lqu ) as the minimum skew information achievable on a single local measurement .
we remark that by ` measurement ' in the following we always refer to a complete von neumann measurement .
let @xmath18 be the state of a bipartite system , and let @xmath19 denote a local observable , with @xmath20 a hermitian operator on @xmath8 with spectrum @xmath21 .
we require @xmath21 to be nondegenerate , which corresponds to maximally informative observables on @xmath8 .
the lqu with respect to subsystem @xmath8 , optimized over all local observables on @xmath8 with nondegenerate spectrum @xmath21 , is then @xmath22 eq . ( [ gen ] ) defines a family of @xmath21-dependent quantities , one for each equivalence class of @xmath21-spectral local observables over which the minimum skew information is calculated . in practice , to evaluate the minimum in eq . , it can be convenient to parametrize the observables on @xmath8 as @xmath23 , where @xmath24 is varied over the special unitary group on @xmath8 . in this representation
, the ( fixed ) spectrum @xmath21 may be interpreted as a standard ruler " , fixing the units as well as the scale of the measurement ( that is , the separation between adjacent ` ticks ' ) , while @xmath24 defines the measurement basis that can be varied arbitrarily on the hilbert space of @xmath8 . in the following , we prove some general qualitative properties of the @xmath21-dependent lqus , which reveal their intrinsic connection with nonclassical correlations . in a werner state
@xcite @xmath25,$ ] of two qubits @xmath8 and @xmath26 .
the red line is the variance @xmath27 of the @xmath28 operator , which amounts to the total statistical uncertainty .
the blue dashed curve represents the local quantum uncertainty @xmath29 , which in this case is @xmath30 ( any local spin direction achieves the minimum for this class of states ) .
the green dotted curve depicts the ( normalized ) linear entropy @xmath31 of the global state @xmath9 , which measures its mixedness .
notice that the werner state is separable for @xmath32 but it always contains discord for @xmath33 .
, width=245 ] _ a class of quantum correlations measures_. what characterizes a discordant state is , as anticipated , the non - existence of quantum - certain local observables .
in fact , we find that each quantity @xmath34 defined in eq .
( [ lqu ] ) is not only an indicator , but also a full fledged _ measure _ of bipartite quantum correlations ( see fig .
[ figurauno ] ) @xcite , i.e. it meets all the known _ bona fide _
criteria for a discordlike quantifier @xcite .
specifically , in the supplemental material @xcite we prove that the @xmath21-dependent lqu ( for any non - degenerate @xmath21 ) is invariant under local unitary operations , is nonincreasing under local operations on @xmath26 , vanishes if and only if @xmath9 is a zero discord state with respect to measurements on @xmath8 , and reduces to an entanglement monotone when @xmath9 is a pure state .
if we now specialize to the case of bipartite @xmath0 systems , we further find that quantifying discord via the lqu is very advantageous in practice , compared to all the other measures proposed in the literature ( which typically involve formidably hard optimizations not admitting a closed formula even for two - qubit states ) @xcite . indeed , the minimization in eq .
( [ lqu ] ) can be expressed in closed form for arbitrary states @xmath35 of a qubit - qudit system defined on @xmath36 , so that @xmath37 admits a _
computable _ closed formula .
moreover notice that , when @xmath8 is a qubit , all the @xmath21-dependent measures are equivalent up to a multiplication constant @xcite .
we thus drop the superscript @xmath21 for brevity , and pick nondegenerate observables @xmath38 on the qubit @xmath8 of the form @xmath39 , with @xmath40 .
this choice corresponds to a lqu normalized to unity for pure , maximally entangled states .
eq . can then be rewritten as the minimization of a quadratic form involving the unit vector @xmath41 , yielding simply @xmath42 where @xmath43 denotes the maximum eigenvalue , and @xmath44 is a @xmath45 symmetric matrix whose elements are @xmath46 with @xmath47 .
it is easy to check that , for a pure state @xmath48 , eq . reduces to the linear entropy of entanglement , @xmath49 , where @xmath50 is the marginal state of subsystem @xmath8 .
qubit - qudit states represent a relevant class of states for applications in quantum information processing , and we present some pertinent examples in this letter .
the evaluation of the lqu for werner states of two qubits is displayed in fig .
[ figurawerner ] .
a case study of the discrete quantum computation with one bit ( dqc1 ) model of quantum computation @xcite is reported in the supplemental material @xcite , showing that our measure ( evaluated in the one versus @xmath51 qubits partition ) exhibits the same scaling as the canonical entropic measure of discord @xcite . beyond the practicality of having a closed formula ,
the approach adopted in this letter provides in general a nice physical interpretation of discord as the minimum quantum contribution to the statistical variance associated to the measurement of local observables in correlated quantum systems .
interestingly , the lqu in a general state @xmath35 of a @xmath36 system , can be reinterpreted geometrically as the minimum squared hellinger distance between @xmath35 and the state after a least disturbing root - of - unity local unitary operation applied on the qubit @xmath8 , in a spirit close to that adopted to define ` geometric discords ' based on other metrics @xcite .
let us recall that the squared hellinger distance between density matrices @xmath9 and @xmath52 is defined as @xmath53 @xcite .
observing that , for qubit @xmath8 , any generic nondegenerate hermitian observable @xmath54 is a root - of - unity unitary operation , which implies @xmath55 for any function @xmath56 , we have @xmath57 ; therefore , minimizing over the local observables @xmath58 yields the geometric interpretation of the lqu , analytically computed in eq .
( [ anale ] ) , in terms of hellinger distance .
the study of further connections between uncertainty on a single local observable and geometric approaches to nonclassicality of correlations , possibly in larger and multipartite systems , opens an avenue for future investigations . of a bipartite system @xmath59
is prepared , and a local unitary transformation depending on an unobservable parameter @xmath60 acts on subsystem @xmath8 , transforming the global state into @xmath61 . by means of a suitable measurement at the output one can construct an ( unbiased ) estimator @xmath62 for @xmath60 .
the quality of the estimation strategy is benchmarked by the variance of the estimator . for a given probe state @xmath9 , the optimal measurement at the output returns an estimator @xmath63 for @xmath60 with the minimum allowed variance given by the inverse of the qfi @xmath64 , according to the quantum cramr - rao bound @xcite . in the prototypical case of optical phase estimation , the present scheme corresponds to a mach - zender interferometer . restricting to pure inputs , research in quantum metrology @xcite
has shown that in this case entangled probes allow to beat the shot noise limit @xmath65 ( @xmath51 being the input mean photon number ) and reach ideally the heisenberg scaling @xmath66 .
however , recent investigations have revealed how in presence of realistic imperfections the achieved precision quickly degrades to the shot noise level @xcite .
for mixed bipartite probes , we show that the qfi is bounded from below by the amount of quantum correlations in the probe state @xmath9 as quantified by the lqu.,width=321 ] _ applications to quantum metrology._ we now discuss the operative role that discord , as quantified by the lqu , can play in the paradigmatic scenario of phase estimation in quantum metrology @xcite .
we focus here on an ` interferometric ' setup employing bipartite probe states , as sketched in fig .
[ metrofiga ] . given a ( generally mixed ) bipartite state @xmath9 used as a probe
, subsystem @xmath8 undergoes a unitary transformation ( specifically , a phase shift ) so that the global state changes to @xmath67 , where @xmath68 , with @xmath69 a local hamiltonian on @xmath8 , which we assume to have a nondegenerate spectrum @xmath70 .
the goal is to estimate the unobservable parameter @xmath60 .
the protocol , which has wide - reaching applications , from gravitometry to sensing technologies @xcite , can be optimized by picking the best probe state @xmath9 and the most informative measurement at the output .
it is known that the latter optimization can be solved in general by choosing , for any probe state @xmath9 , the measurement strategy which saturates asymptotically the quantum cramr - rao bound , @xmath71 $ ] @xcite , where the quantum fisher information ( qfi ) @xmath64 sets then the precision of the optimal estimation , and @xmath72 denotes the number of times the experiment is repeated ( @xmath73 is assumed ) .
we will denote by @xmath63 the estimator obtained from the optimal measurement strategy , so that @xmath74 $ ] . recall that the qfi can be written as @xcite @xmath75 , with @xmath76 being the symmetric logarithmic derivative defined implicitly by @xmath77 .
we focus therefore on the optimization of the input state . in practical conditions ,
e.g. when the engineering of the probe states occurs within a thermal environment or with a reduced degree of control , it may not be possible to avoid some degree of mixing in the prepared probe states .
it is then of fundamental and practical importance to investigate the achievable precision when the phase estimation is performed within specific noisy settings @xcite . here
we assess whether and how quantum correlations in the ( generally mixed ) state @xmath9 play a role in determining the sensitivity of the estimation .
notice that the remaining steps of the estimation process are assumed to be noiseless ( the unknown transformation @xmath78 is unitary and the output measurement is the ideal one defined above ) .
the key observation stems from the relation between the wigner - yanase and the fisher metrics @xcite , which implies that the skew information of the hamiltonian is majorized by the qfi @xcite .
as @xmath69 is not necessarily the most certain local observable with spectrum @xmath79 , the @xmath79-lqu itself fixes a lower bound to the qfi : @xmath80 then , for probe states with any nonzero amount of discord , and for @xmath81 repetitions of the experiment , the optimal detection strategy which asymptotically saturates the quantum cramr - rao bound produces an estimator @xmath63 with necessarily limited variance , scaling as @xmath82 hence , we established on rigorous footings that the quantum correlations measured by lqu , though not necessary @xcite , are a sufficient resource to ensure a guaranteed upper bound on the smallest possible variance with which a phase @xmath60 can be measured with mixed probes .
we now provide a simple example to clarify the above general discussion .
suppose system @xmath8 is a spin-@xmath83 particle undergoing a phase rotation @xmath84 , where @xmath85 is the third spin component , and @xmath60 the phase to be estimated . in this case
the estimation precision is bounded by the so - called _ heisenberg limit _
@xcite . a typical scheme achieving this limit
can be outlined as follows .
assume that system @xmath26 is simply a qubit with states @xmath87 .
the @xmath59 system is initially prepared in the product state @xmath88 , where @xmath89 are the eigenstates of @xmath85 with eigenvalues @xmath90 , and @xmath91 .
then , a ` control - flip ' operation @xmath92 is applied , so that the system evolves to @xmath93 .
one can see that the entangled state @xmath94 used as a probe achieves the heisenberg limit .
our general treatment allows us to study quantitatively the effect of noise on the estimation power of the bipartite state @xmath94 .
suppose now that the probe state , ideally @xmath94 , is prepared in a noisy environment , which induces partial dephasing in the basis @xmath89 .
then , our probe state is in general given by @xmath95,$ ] where @xmath96 quantifies the degree of residual coherence , and @xmath97 .
as this is effectively a @xmath98-qubit state , we can restrict our analysis to a truncated @xmath99 hilbert space . here
the restriction of @xmath85 has the spectrum @xmath100 .
we can thus calculate the @xmath70-lqu in this effective @xmath99 hilbert space , obtaining @xmath101 . for any @xmath83 , notice that the discord is a monotonically increasing function of the coherence @xmath102 .
hence , from eq .
( [ premestruo ] ) one has @xmath103 as the spin number @xmath83 is increasing , this guarantees that the classical scaling @xmath104 ( i.e. the so - called _ shot noise limit _
@xcite ) can still be beaten provided that @xmath105 . the connection between the lqu and the sensitivity of parameter estimation
can also be appreciated in more abstract geometrical terms , without the need for invoking the fisher information . as shown by brody @xcite ,
the skew information @xmath106 of the hamiltonian @xmath69 determines the squared speed of evolution of the density matrix @xmath9 under the unitary @xmath68 .
this provides another geometric interpretation for the lqu : the observable @xmath38 which achieves the minimum in eq .
[ lqu ] is the local observable with the property that the resulting local unitary operation @xmath107 makes the given state @xmath9 of the whole system evolve as slowly as possible ( the observable @xmath38 is the least disturbing in this specific sense ) . since a higher speed of state evolution under a change in the parameter @xmath60 means a higher sensitivity of the given probe state to the estimation of the parameter , our result can be interpreted as follows : the amount of discord ( lqu ) in a mixed correlated probe state @xmath9 used for estimation of a parameter @xmath60 bounds from below the squared speed of evolution of the state under any local hamiltonian evolution @xmath108 , hence the sensitivity of the given probe state @xmath9 to a variation of @xmath60 , which is a general measure of precision for the considered metrological task . _
conclusions._ in this letter we studied the quantum uncertainty on single observables .
the exploration of this concept allowed us to define and investigate a class of measures of bipartite quantum correlations of the discord type @xcite , which are physically insightful and mathematically rigorous .
in particular , for qubit - qudit states a unique measure is defined ( up to normalization ) , and it is computable in closed form .
quantum correlations , in the form known as quantum discord @xcite , manifest in the fact that any single local observable displays an intrinsic quantum uncertainty .
discord in mixed probe states , measured by the local quantum uncertainty , is further proven to guarantee a minimum sensitivity in the protocol of optimal phase estimation @xcite .
we believe worthwhile to substantiate in future work the promising uncovered connections between quantum mechanics , information geometry and complexity science @xcite by addressing the role of quantum uncertainty , in particular induced by quantum correlations , in such contexts . + _ note added in proof .
_ very recently , an alternative measure of discord based on the quantum hellinger distance has been proposed in ref .
@xcite . _
acknowledgments._ d.g . and g.a .
thank v.p .
belavkin _ in memoriam _ for having shared quantum bits of his awesome talent and his encyclopedic knowledge of the laws of nature .
we warmly thank l. correa for extensively commenting on the manuscript , and a. balyuk , m. ahmadi , d. brody , j. calsamiglia , f. ciccarello , p. j. coles , a. datta , m. g. genoni , v. giovannetti , s. girolami , m. guta , f. illuminati , m. s. kim , s. luo , l. maccone , s. pascazio , m. piani , t. rudolph for fruitful discussions .
we acknowledge financial support from universitas 21 , the university of nottingham , the epsrc research development fund ( grant no .
pp-0313/36 ) , and the qatar national research fund ( grant no . nprp 4 - 554 - 1 - 084 )
. 99 w. heisenberg , ber den anschaulichen inhalt der quantentheoretischen kinematik und mechanik . z. phys . * 43 * , 172 ( 1927 ) .
j. oppenheim and s. wehner , the uncertainty principle determines the nonlocality of quantum mechanics .
science * 330 * , 1072 ( 2010 ) .
m. berta , m. christandl , r. colbeck , j. m. renes , and r. renner , the uncertainty principle in the presence of quantum memory .
nature phys .
* 6 * , 659 - 662 ( 2010 ) .
p. j. coles , r. colbeck , l. yu , and m. zwolak , uncertainty relations from simple entropic properties .
phys .
rev .
lett . * 108 * , 210405 ( 2012 ) .
r. horodecki , p. horodecki , m. horodecki , and k. horodecki , quantum entanglement .
rev .
mod .
phys . * 81 * , 865 ( 2009 ) .
a. ferraro , l. aolita , d. cavalcanti , f. m. cucchietti , and a. acn , almost all quantum states have nonclassical correlations .
phys .
rev .
a * 81 * , 052318 ( 2010 ) .
p. perinotti , discord and nonclassicality in probabilistic theories .
phys .
rev .
lett .
* 108 * , 120502 ( 2012 ) .
h. ollivier and w. h. zurek , quantum discord : a measure of the quantumness of correlations .
phys .
rev .
lett . * 88 * , 017901 ( 2001 ) .
l. henderson and v. vedral , classical , quantum and total correlations .
j. phys .
a. : math . gen . * 34 * , 6899 ( 2001 ) . k. modi , a. brodutch , h. cable , t. paterek , and v. vedral , the classical - quantum boundary for correlations : discord and related measures .
rev .
mod .
phys . * 84 * , 1655 ( 2012 ) .
a. datta , a. shaji , and c. m. caves , quantum discord and the power of one qubit .
phys .
rev .
lett * 100 * , 050502 ( 2008 ) .
m. piani , p. horodecki , and r. horodecki , no - local - broadcasting theorem for multipartite quantum correlations .
phys .
rev .
lett . * 100 * , 090502 ( 2008 ) .
b. daki _
et al . _
quantum discord as resource for remote state preparation .
nature phys .
* 8 * , 666 ( 2012 ) .
m. gu _
et al .
_ observing the operational significance of discord consumption
. nature phys .
* 8 * , 671 ( 2012 ) .
v. giovannetti , s. lloyd , and l. maccone , advances in quantum metrology .
nature photon .
* 5 * , 222 ( 2011 ) .
s. luo , wigner - yanase skew information and uncertainty relations .
phys .
rev .
lett .
* 91 * , 180403 ( 2003 ) .
e. p. wigner and m. m. yanase , information contents of distributions .
proc .
natl .
acad .
sci .
u.s.a . *
49 * , 910 ( 1963 ) .
s. luo , s. fu , and c. h. & oh , quantifying correlations via the wigner - yanase skew information .
phys .
rev .
a * 85 * , 032117 ( 2012 ) .
p. gibilisco and t. isola , wigner - yanase information on quantum state space : the geometric approach . j. math
. phys . * 44 * , 3752 ( 2003 ) .
d. c. brody , information geometry of density matrices and state estimation .
j. phys .
a : math . theor . * 44 * , 252002 ( 2011 ) .
f. hansen , metric adjusted skew information .
proc .
natl .
acad .
sci .
u.s.a . *
105 * , 9909 ( 2008 ) .
a. monras , g. adesso , s. m. giampaolo , g. gualdi , g. b. davies , and f. illuminati , entanglement quantification by local unitary operations .
phys .
rev .
a * 84 * , 012301 ( 2011 ) .
s. m. giampaolo , a. streltsov , w. roga , d. bruss , and f. illuminati , quantifying nonclassicality : global impact of local unitary evolutions .
phys .
rev .
a * 87 * , 012313 ( 2013 ) .
s. gharibian , quantifying nonclassicality with local unitary operations .
phys .
rev .
a * 86 * , 042106 ( 2012 ) . for an alternative relation between quantum correlations and local ( entropic ) uncertainty see p. j. coles , unification of different views of decoherence and discord .
phys .
rev .
a * 85 * , 042103 ( 2012 ) .
see the supplemental material at [ epaps ] for additional proofs and examples . the geometric discord based on hilbert - schmidt norm [ b. daki , v. vedral , and c. brukner , necessary and sufficient condition for nonzero quantum discord .
phys .
rev .
lett . * 105 * , 190502 ( 2010 ) ] is also computable for the same class of states , but it has been shown to be not a valid measure of quantum correlations [ m. piani , problem with geometric discord .
phys .
rev .
a * 86 * 034101 ( 2012 ) ] .
indeed , a nondegenerate qubit observable with fixed spectrum can be parametrized as @xmath109 , where @xmath110 are constants ( @xmath111 ) while @xmath41 varies on the unit sphere .
then @xmath112 , implying that all the @xmath37 s are proportional to each other .
e. knill and r. laflamme , power of one bit of quantum information .
phys .
rev .
lett .
* 81 * , 5672 ( 1998 ) .
i. bengtsson and k. zyczkowski , _ geometry of quantum states : an introduction to quantum entanglement _ ( cambridge university press , cambridge , 2006 ) .
s. luo and q. zhang , informational distance on quantum - state space .
phys .
rev .
a * 69 * , 032106 ( 2004 ) .
m. g. a. paris , quantum estimation for quantum technology .
int .
j. quant . inf .
* 7 * , 125 ( 2009 ) .
s. l. braunstein and c. m. caves , statistical distance and the geometry of quantum states .
phys .
rev .
lett . * 72 * , 3439 - 3443 ( 1994 ) .
d. petz and c. ghinea , introduction to quantum fisher information .
e print arxiv:1008.2417 ( 2010 ) .
u. dorner _
et al . _ optimal quantum phase estimation .
phys .
rev .
lett . * 102 * , 040403 ( 2009 ) .
r. demkowicz - dobrzanski , j. kolodynski , and m. guta , the elusive heisenberg limit in quantum - enhanced metrology , nature commun .
* 3 * , 1063 ( 2012 ) .
b. m. escher , r. l. de matos filho , and l. davidovich , general framework for estimating the ultimate precision limit in noisy quantum - enhanced metrology , nature phys .
* 7 * , 406 ( 2011 ) .
s. luo , wigner - yanase skew information vs. quantum fisher information .
proc .
amer .
math .
soc . * 132 * , 885 ( 2003 ) .
l. pezz and a. and smerzi , entanglement , nonlinear dynamics , and the heisenberg limit .
phys .
rev .
lett . * 102 * , 100401 ( 2009 ) .
k. modi , h. cable , m. williamson , and v. vedral , quantum correlations in mixed - state metrology .
phys .
rev .
x * 1 * , 021022 ( 2011 ) .
s. boixo , s. t. flammia , c. m. caves , and j. m. geremia , generalized limits for single - parameter quantum estimation .
phys .
rev .
lett . * 98 * , 090401 ( 2007 ) . c. vaneph , t. tufarelli and m. g. genoni , quantum estimation of a two - phases spin rotation arxiv:1211.7224v2 .
m. gu , k. wiesner , e. rieper , and v. vedral , quantum mechanics can reduce the complexity of classical models . nature commun . * 3 * , 762 ( 2012 ) . l. chang and s. luo , remedying the local ancilla problem with geometric discord .
phys .
rev .
a * 87 * , 062303 ( 2013 ) . |
tracing the buildup of galaxies at high redshifts requires reliable measurements of two fundamental quantities : stellar masses and star - formation rates ( sfr ) .
the integrated history of star - formation rates within galaxies provides a lower limit to their stellar mass growth .
stellar mass growth in excess of this contribution is to be ascribed to galaxy mergers .
most of the rise in the space density of massive galaxies takes place in the cosmic time interval between @xmath12 and @xmath13 ( e.g. , dickinson et al .
2003 ; rudnick et al . 2003 ; 2006 ; fontana et al .
2004 ; 2006 ; daddi et al . 2004a , d04a hereafter ; van dokkum 2006 ; franceschini et al .
2006 ; pozzetti et al . 2007 ; arnouts et al . 2007 ) .
it is still unclear , however , what is the major channel for galaxy growth during this period , and the quantitative role of mergers in building up the mass of galaxies .
violent starbursts have been found to be common in the @xmath2 universe ( d04a , daddi et al .
2005b , d05b hereafter ; shapley et al.2005 ; frster - schreiber et al .
2004a ) , but their relevance on the mass growth of galaxies critically depends on the effective duration of these vigorous star - formation events ( see d05b ) . to make substantial progress one needs to look at the distribution of star - formation rates in distant galaxies , in order to ascertain the relative frequency of these events among massive galaxies .
this requires accurate measurements of individual star - formation rates in large samples of distant galaxies .
the advent of the spitzer space telescope holds substantial promise to achieve these goals , as it now routinely allows one to detect distant galaxies with the multiband imaging photometer for spitzer ( mips ) at 4mu all the way to redshift of @xmath14 ( and even beyond ) .
this corresponds to the rest - frame spectral mid - ir region ( 512@xmath15 m ) , which tends to be dominated by polycyclic aromatic hydrocarbon features ( pah ) . in the local universe ,
the mid - ir luminosities of galaxies are known to correlate with the total infrared luminosities ( @xmath16 ; see chary & elbaz 2001 ; elbaz et al .
2002 ; dale & helou 2002 ; frster - schreiber et al .
2004b ) , a well established probe of the star - formation rate of galaxies .
problems however exist because the pah emission is known to be reduced at low metallicities ( engelbracht et al .
2005 ; madden et al .
spitzer observations show a substantial scatter in the local correlation of mid - ir luminosity vs. the total infrared luminosity @xmath16 ( dale et al .
2005 ; smith et al . 2006 ; armus et al .
in addition , a substantial contribution to the mid - ir light can come from the accretion disks of active galactic nuclei ( agns ) .
agns in distant galaxies can be identified by their x - ray emission , but this would not be detectable at @xmath5 in the case of large column densities of obscuring material even for luminous agn ( e.g. , as in the case of ngc 1068 ) .
several recent surveys have used the 4mu emission in @xmath17 galaxies in order to estimate their ongoing star - formation rates and related quantities ( e.g. , prez - gonzlez et al .
2005 ; reddy et al.2005 ; d05b , papovich et al . 2006 ;
caputi et al .
2006ab ; etc . ) .
this is generally accomplished by using prescriptions relating the mid - ir luminosity to @xmath16 , as calibrated in the spectral libraries from chary & elbaz ( 2001 ; ce01 hereafter ) and dale & helou ( 2002 ; dh02 hereafter ) , although sometimes other empirical conversions have been used ( see , e.g. , caputi et al .
however , it remains to be established to which extent these estimates are still accurate for @xmath2 galaxies .
for example , the impact of the silicate 9.7@xmath15 m absorption feature at high redshift has still to be quantified , and the relative strength of the pah emission as a function of star formation activity might be different at high redshift with respect to the local universe .
using multi - wavelength data from the northern field of the great observatories origins deep survey ( goods ) , d05b showed that the average sed of massive @xmath5 galaxies is consistent with that of present day ulirgs .
this implies , on average , a good agreement between star - formation tracers as calibrated locally and based on the radio , far - ir , mid - ir and even uv ( corrected for the extinction by dust ) and the x - ray .
this agreement is encouraging in view of our need to estimate the star - formation rates in distant galaxies , especially given the currently limitations in probing the luminosities of galaxies at @xmath2 at the peak of the far - ir emission at @xmath18 m in the rest - frame . in this regard
, the situation will eventually improve with herschel , and later on with alma . in d05b
a number of interesting features of high redshifts ulirgs were discussed , e.g. , that they may have fairly long duty cycles and that they appear to be relatively transparent to the uv photons .
these and other aspects are visited again here in more detail .
the main aim of this paper is to verify to which extent this agreement among the various star formation tracers persists individually for distant massive galaxies , hence to which extent sfrs can be unambiguously measured at high redshift . in particular , reliable sfrs are instrumental for relating to them other galaxy properties in a meaningful way .
for example , recent surveys have shown that a tight correlation exists at @xmath9 and at @xmath8 between the stellar mass and the sfrs of blue , actively star forming galaxies ( noeske et al . 2007 ;
elbaz et al .
it is not yet clear if such a fundamental correlation exists also at higher redshifts , when most of the stars in massive galaxies were formed . for this work we make use of some of the deepest multiwavelength datasets , available as part of goods .
in particular , compared to d05b , we use here a much deeper 1.4 ghz dataset over goods - north ( goods - n hereafter ) obtained with the very large array ( vla ) ( morrison et al . in preparation ) , and 0mu photometry from both goods fields from frayer et al . ( 2006 ) . also crucial to this work
is the large dataset of spectroscopic redshifts that is available to us , including the vlt / fors2 spectroscopic redshifts for the goods - south field ( goods - s hereafter ; vanzella et al .
2005 , 2006 ) and those from the ultradeep spectroscopy of the gmass survey ( kurk et al . , in preparation ) .
finally , we have used in this work the latest realization of galaxy formation models based on the millennium simulations ( springel et al .
2005 ) , in the form of mock lightcones presented by kitzbichler & white ( 2007 ) .
this allows us to constrain the models and gain insights into the nature of star forming galaxies at high redshifts .
the paper is organized as follows .
the sample selection of @xmath2 galaxies , with spectroscopic and photometric redshifts are presented in section [ sec : data ] .
the multiwavelength datasets used are presented in section [ sec : multi ] , with a discussion of known limitations of each dataset as a star formation indicator , especially for high redshift galaxies .
section [ sec : mips ] compares the mu rest - frame luminosities of galaxies ( estimated from their 4mu flux densities ) to those at other wavelengths , while section [ sec : uv ] focuses on the use of ultraviolet ( uv ) luminosities and on the reliability of extinction corrections .
a suggested recipe for estimating sfrs in @xmath5 galaxy samples is given in section [ sec : recipe ] .
we present the implications of our findings for the characterization of star formation in distant galaxies , galaxy growth and the comparison with theoretical models in section [ sec : impli ] .
our conclusions are summarized in section [ sec : end ] .
we assume a salpeter initial mass function ( imf ) from 0.1 and 100 @xmath19 , and a wmap3 cosmology with @xmath20 , and @xmath21[km s@xmath22 mpc@xmath22]@xmath24 ( spergel et al .
2007 ) . throughout the paper
we use magnitudes in the vega scale , unless stated otherwise .
the sample of galaxies analyzed in this study has been selected in the @xmath25-band , and applying the @xmath26 selection technique introduced by daddi et al .
( 2004b ; d04b hereafter ) , that allows us to define highly complete samples of galaxies in the redshift range @xmath0 . unless explicitly stated otherwise , we limit the analysis to @xmath26 star forming galaxies ( or @xmath27s , as opposed to the passive ones or @xmath28s ) , as we aim to focus on galaxies with active star formation . in the goods - n field
we have selected 273 galaxies to @xmath29 , corresponding to the 5-sigma detection limit of the data taken with the flamingos camera on the mayall 4 m noao telescope ( the dataset used here is slightly deeper than that used in d05b ) . in the goods - s field we selected 1018 sources detected with @xmath30 down to @xmath31 .
the goods - s field @xmath25-band vlt isaac data is not homogeneously deep over the whole field , but reasonably complete over the whole used area at these @xmath25-band limits .
we have excluded from the analysis all hard x - ray detected galaxies in both goods fields , using the catalogs of alexander et al .
( 2003 ) , and galaxies with power law seds over the 3.624 @xmath15 m wavelength range observed by the spitzer infrared array camera ( irac ) as likely agn - dominated sources .
only a handful of power - law galaxies remain undetected in our deep x - ray imaging data . also , in both fields we only consider galaxies whose irac counterpart is closer than 0.5@xmath32 from the k - band position .
we have empirically verified that this criterion is efficient at excluding sources where the irac bands photometry is substantially contaminated by blending due to the relatively poor irac psf of about 1.6@xmath32 .
statistically , this excludes some 10% of the galaxies , for the accurate irac detection procedures carried out in goods and based on mexhat kernel detections using sextractor ( dickinson et al .
, in preparation ) .
this criterion is required in order to have reliable photometry over the irac bands ( hence solid photometric redshift estimates ) , and to avoid problems with blending in the mips 4mu photometry . the 4mu flux densities ( chary et al .
, in preparation ) are in fact estimated using prior source positions from the irac catalogs . in both fields we used multiband photometry in the optical and near - ir
( see giavalisco et al .
2004 for datasets descriptions ) , updated to the latest releases and observations .
seds were derived using the method described , e.g. , in d04b and d05b .
photometric redshifts were determined in both fields for all @xmath26 selected @xmath2 galaxies by fitting to the empirical templates of coleman et al .
( 1980 ) , as described in d04ab .
however , for the goods - s field we find that the photometric redshifts measured by grazian et al .
( 2006 ) , which include deep u - band data similarly to what we have in goods - n , are in slightly better agreement with the spectroscopic redshifts and we decided to adopt them for the present analysis .
spectroscopic redshifts are available in goods for a fairly representative fraction of the galaxy sample considered in this paper . when limiting to spectroscopic redshifts with flags warranting reliability , redshifts are available for 175 galaxies in goods - s ( 17% of the total sample ) and @xmath33 in goods - n ( 21% of the sample ) . for goods - s , in addition to all sources of publicly available redshifts ( vvds : le fevre et al .
2004 ; goods / fors2 : vanzella et al .
2005 , 2006 ; k20 survey : mignoli et al . 2005 ;
cdfs survey : szokoly et al .
2004 ) we have most notably used results of redshift surveys developed as part of goods or in collaboration with goods , that explicitly targeted the @xmath2 galaxies .
this includes the galaxy mass assembly spectroscopic survey ( kurk et al .
2006 , and in preparation ) that targeted @xmath34 galaxies selected at 4.5@xmath15 m ( irac channel 2 ) to a limiting ab magnitude = 23 .
for the goods - n field major sources of @xmath11 spectroscopic redshifts are keck lr - blue and deimos observations carried out within goods ( stern et al .
, in preparation ) and the bm / bx samples of reddy et al . ( 2006b ) , in addition to the tkrs survey ( wirth et al . 2004 ) .
[ fig : zphot_n ] panels show the comparison of photometric and spectroscopic redshifts for our samples of @xmath26 selected galaxies .
the dispersion of the residual in the difference between spectroscopic and photometric redshifts is @xmath35 for goods - n and goods - s
. the larger number of outliers in the full samples , with respect to the @xmath26 samples , shows that the use of the @xmath26 selection does help in defining a clean sample of @xmath2 galaxies , which is substantially free from lower redshift contaminants .
we have identified a small sample of galaxies in both fields with strong discrepancies between spectroscopic and photometric redshifts . from a careful analysis of the multiwavelength sed of these outliers
, we found that in most cases the spectroscopic redshift appears to be the wrong one , and we have adopted the photometric redshift for the analysis .
this does not affect any of our major conclusions .
[ fig : zhist_n ] panels show the redshift distribution of the sources used for the analysis of @xmath2 sfr indicators in this paper .
we do not include in the analysis carried out in the next sections the objects in the tails of the distributions , with either @xmath36 or @xmath37 .
that can be detected with the various tracers used in this work .
we consider limits of 14.1@xmath15jy ( 3@xmath38 ) for radio 1.4ghz , 12@xmath15jy ( 3@xmath38 ) for mips 4mu , 1.8mjy for mips 0mu ( 3@xmath38 ) and 4mjy for 850@xmath15 m ( flux density of typical detections , pope et al .
2006 , some small regions in goods - n have actually somewhat fainter detections ) .
we use the model of ce01 to convert flux densities into @xmath16 as a function of redshift , except for radio where we use the radio - ir correlation . for the uv , the effective sfr limit in our samples ( formally converted to @xmath16 using eq .
1 ) is given in principle by the depth of the b - band imaging ( @xmath39 ab , 3@xmath38 ) , from which we derive the uv luminosity . in case of non - zero dust reddening
the reached sfr limit is less deep .
mips 4mu is by far the deepest long wavelength probe of sfr and ir luminosities at these high redshifts .
, width=332 ]
a major focus of this paper is the comparison between luminosities at various wavelengths , each of which is known to correlate with the sfr in galaxies . in order to put emission at different wavelengths onto a common footing for intercomparison
, we will adopt nominal conversions between these observed luminosities and fiducial quantities such as the total infrared luminosity ( @xmath16 , integrated from 8 to 1000@xmath15 m ) or an equivalent star formation rate ( sfr ) . the use of @xmath16 as a star formation indicator implies that all energy from star formation is absorbed by dust and re - radiated in the thermal infrared , and that there are no other significant sources for dust heating .
neither assumption is universally correct : on the one hand not all uv photons from young massive stars may be absorbed by dust , whereas emission from an agn can boost emission at various wavelengths , including the infrared luminosity @xmath16 .
thus , sfr estimates based only on @xmath16 would result in underestimates in the former case , and in overestimates of the true sfr in the latter .
starburst galaxies with the high sfrs ( many tens to hundreds of @xmath19 yr@xmath22 ) and infrared luminosities ( @xmath40 are called luminous infrared galaxies or lirgs ; then , for @xmath41 one refers to ultraluminous infrared galaxies , or ulirgs ; we will also refer to objects with @xmath42 as hyperluminous , or hylirgs ) characteristic of some the @xmath43 galaxies studied here .
values for @xmath16 and the sfr may differ substantially , depending on the specific tracer used to evaluate them . in this section
, we present the different indicators , and give a description of the techniques used to analyze the data , and of the recipes used to convert the observed emission to sfrs .
these recipes are generally based on empirical calibrations in the local universe . in the next section
, we will compare the derived equivalent @xmath16 or sfr values , looking for agreement , which would suggest coherent behavior from low to high redshift , or for discrepancies , which may suggest that other effects are coming into play .
following kennicutt ( 1998 ) , we may express the conversion between @xmath16 and sfr as : galaxies in our sample as a function of the k - band limiting magnitude ( cumulative fractions ) , for the goods - s ( solid line ) and goods - n ( dotted line ) fields .
red lines are for a detection limit of 100@xmath15jy , matched by most shallower surveys performed at 4mu with spitzer .
the very large detection fraction in goods up to @xmath44 is in line with the expectations based on fig .
[ fig : pl_limi ] .
, width=332 ] _ ir [ m_yr^-1 ] = 1.7310 ^ -10 l_ir [ l _ ] .
[ eq : lir ] this conversion depends on the adopted imf ; here we have adopted the imf of salpeter ( 1955 ) . given this relation , sfr@xmath45 and @xmath16 can be seen as equivalent quantities for star forming galaxies . in order to estimate the total sfr ,
we must also include the contribution of energetic emission from young , massive stars that is not absorbed and reprocessed by dust .
we estimate it from the rest - frame uv luminosity as in eq . 5 of d04b , with no reddening correction , so to have : _ tot = sfr_ir + sfr_uv , uncorrected .
[ eq : sfr ] we shall return to the conversion of uv light to sfr in section 3.6 .
the second term above is negligible for red galaxies , as for most of the @xmath26-selected @xmath43 sources with @xmath46 , but becomes more relevant at fainter magnitudes when bluer galaxies are more prevalent .
both goods - n and s have been observed with spitzer mips at 4mu for about 10 hours per sky pixels .
see d05b for a slightly more detailed description of these datasets and for the associations of 4mu flux densities to k - selected galaxies .
catalogs for the 4mu sources will be presented by chary et al .
( in preparation ) . for a @xmath5 galaxy
the 4mu emission corresponds to approximately an 8@xmath15 m rest - frame measurement . for each galaxy in the @xmath47 redshift range we have derived bolometric infrared luminosities from the observed flux density at 4mu and using the luminosity dependent sed libraries of chary & elbaz ( 2001 ) and dale & helou ( 2002 ) . in practice , this is done by interpolating @xmath16 over the template seds , sorting the value of @xmath16 that corresponds to the observed 4mu flux density .
this determination of the bolometric ir luminosity implies a large extrapolation , as the @xmath48 m rest - frame luminosity is generally a small fraction of @xmath16 , and therefore the result is template - dependent .
for this reason we also derived more direct 8@xmath15 m rest - frame luminosities from the 4mu flux density , using the model seds ( from ce01 and dh02 ) only to k - correct the measurement from 4mu observed to 8@xmath15 m rest - frame luminosities ( this k - correction is also template dependent but generally small in the probed redshift range , within a factor of 2 ) .
for the ce01 templates the relations between @xmath49 and @xmath16 is : log ( ) = 1.50log ( ) - 4.31 for @xmath50 , and log ( ) = 0.93log ( ) + 1.23 for @xmath51 ; while for the dh02 templates it is : log ( ) = 1.21log ( ) - 1.25 .
[ eq : l8 ] here , and in the remainder of the paper , @xmath52 is @xmath53 at @xmath54 .
the two relations give the same @xmath55 for @xmath56 , for higher @xmath57 the ce01 templates result in increasingly larger values of @xmath16 , up to a factor of 2.3 for @xmath58 . at @xmath59
the ce01 templates results in lower @xmath16 values with respect to dh02 . for the range of interest in this work , the differences are always well within a factor of 2 , also due to the two - slope dependence of the ce01 conversion .
the 0mu data used here are a combination of all of the goods - n and goods - s available to us from the spitzer programs id-3325 , -20147 , -81 , and -30948 .
the data were reduced following the techniques given in frayer et al .
these data represent 3 hours of integration over the full goods - n field and the central ( @xmath60 ) region of goods - s .
the outer regions of goods - s have an integration time of about 2 hours .
the typical @xmath61 limit of these data is 1.8mjy at 0mu in the deepest regions .
only 7 @xmath26 star forming @xmath62 galaxies are individually detected at 0mu ( 6 of which are in goods - n , a likely effect of small number statistics ) .
therefore , we used stacking techniques to measure average properties of galaxies at 0mu .
bright low - redshift sources were removed from the data to avoid biasing the stacking results .
we convert the 0mu flux densities into total ir luminosities again using the ce01 and dh02 template libraries .
being closer to the peak of ir seds , the conversion of observed 0mu flux densities into @xmath16 is much less critical than for 4mu , even at @xmath2 . for this work ,
we use new ultradeep 20 cm ( 1.4ghz ) imaging of the goods - n obtained at the vla in various array configurations for maximizing the data quality and the dynamic range ( morrison et al .
, in preparation ) .
coadded to the previous radio data ( mostly from richards et al .
2000 ) , the new data currently reach an rms of 4.5@xmath15jy per beam over the field .
we then assume a radio spectral index @xmath63 for deriving 1.4 ghz rest - frame radio luminosities from the observed flux density .
the local radio - ir correlation ( condon et al . 1992 ; yun , reddy & condon 2001 ) is used to estimate the bolometric ir luminosity from the radio luminosity : l_ir / l_= 3.510 ^ -12 l(1.4 ghz ) [ eq : radioir ] using radio luminosities to estimate @xmath16 for our sources implies assuming that the radio - ir correlation still holds at high redshift .
the dispersion of the local radio - ir correlation can reach a factor of 2 for the most luminous sources ( yun et al .
2001 ) , with luminosities comparable to those of the objects detectable at @xmath2 .
the introduction of a k - correction term , to derive 1.4 ghz rest - frame flux densities from the observed ones , can introduce extra noise ( or a bias ) if the radio slope is different from the assumed value .
this effect is likely smaller than the intrinsic dispersion seen locally , as a 10% error in the slope translate in a similar error in the intrinsic 1.4 ghz rest - frame flux density and therefore in the estimated @xmath16 . in order to reach deeper radio flux densities , we stacked individually undetected sources using the following technique .
we measured the average radio flux density level in a 1@xmath32 radius aperture and correlated this measure to the total radio flux density using simulated radio sources of various flux densities ( defining an aperture correction , about a factor of 1.4 ) . by measuring a large number of empty regions in the radio image we estimated the rms fluctuations of this measure in order to associate an error to each measure .
this allows us to recover radio flux densities of ultra - faint sources with an rms of 4.7@xmath15jy per source , very close to the noise per beam .
we used a 1@xmath32 radius aperture , very close to the radio beam size , after testing a variety of apertures from 0.5@xmath32 to 2@xmath32 , as it provides the best compromise between final s / n and the necessity to measure over a relatively large area in order to cope with possible coordinate misalignments . to statistically account for the coordinate mismatches ( of order of 0.10.2@xmath32 ) . by applying this procedure to bright ,
individually detected galaxies based on the k - band prior position , and comparing the result to radio flux densities derived by direct fit of the profiles , we found that the radio flux densities required an average additional correction of 10% .
all these measurements and errors take into account the primary beam correction for off - axis measurements . for obtaining stacking measurements of undetected sources we obtained weighted averages of the flux densities measured in this way .
the goods - n field has been observed with the submillimeter common user bolometer array ( scuba ) at 850@xmath15 m in a variety of programs , whose data have been merged together into a so - called super - map by borys et al .
( 2003 ) and pope et al .
the depth of the submillimeter ( submm ) map is highly variable over the field . in d05b
, we found that only one goods - n @xmath26 galaxy with @xmath65 is individually detected at 850@xmath15 m in the catalog described in pope et al .
when considering @xmath26 galaxies down to @xmath66 , as in the present work , we detect five objects at 850@xmath15 m .
we note again here that we are excluding all submm galaxies ( smgs ) that are individually detected in hard x - rays .
we obtained scuba 850@xmath15 m stacking of individually undetected sources by estimating flux densities and errors from the super - map at the position of each program galaxy , and then proceeding as in d05b . using sub - mm flux densities at 850@xmath15 m to infer bolometric luminosities ( @xmath16 )
relies on assumptions for the long wavelength shape of the sed , i.e. , on dust temperatures .
for this work , we continue to use the sed shape templates of ce01 and dh02 , derived from local correlations .
the @xmath67 to @xmath16 conversion is to first approximation independent of redshift for the @xmath0 range explored here .
however , the conversion is not perfectly linear . according to the ce01 templates
, a ulirg ( @xmath68 ) would have a flux density of about 1 mjy at these redshifts , raising to about 4.6 mjy for a hylirg ( @xmath69 ) .
dh02 templates give @xmath70% higher flux densities at a given @xmath16 , over the same regime of luminosities .
the non linearity is due to the luminosity - temperature trend observed in the local universe ( less luminous galaxies harbor colder dust ) .
we use the catalog of x - ray detected sources of alexander et al.(2003 ) .
all sources detected in the hard x - ray band were excluded from the sfr analysis in this paper as likely containing agns .
sfrs can be derived from the soft band 0.52 kev fluxes ( mapping closely to the 28 kev rest - frame band ) at @xmath5 ( see e.g. , ranalli et al .
2004 , persic et al .
2004 , hornschemeier et al .
2005 ) , provided that no substantial agn contamination is present at those energies .
we have performed x - ray stacking of @xmath26 samples over the two fields .
results are presented in , as an assessment of the presence of agns is required for interpreting the x - ray stacking results .
ongoing sfrs in @xmath2 galaxies are estimated using the uv luminosity , following the prescriptions in d04b .
the 1500 rest - frame luminosity is derived from the observed @xmath71-band flux density , and a k - correction term is derived based on the redshift of each source ( photometric or spectroscopic ) and its uv spectral slope .
then l(1500 ) is converted into a sfr using eq .
5 of d04b , i.e. : ( m_^-1 ) = l_1500 [ erg s^-1 hz^-1 ] / ( 8.8510 ^ 27 ) .
[ eq : madau ] the amount of dust reddening is estimated adopting a a calzetti et al .
( 2000 ) reddening law ( see also meurer , heckman & calzetti 1998 ) , and using the relation from d04b : e(b - v)=0.25(b - z+0.1)_ab .
[ eq : redde ] an attenuation of @xmath72 mag is then adopted to obtain dust - corrected sfrs from eq .
8 . with this procedure
we obtain reddening corrected sfrs consistent to what derived with sed fitting ( d04b ) for the bulk of @xmath25-selected @xmath2 galaxies .
clearly , this technique is affected by several limitations . in applying a reddening correction ( which median is a factor of @xmath73 for @xmath65 , @xmath5 galaxies ) , the technique assumes that the uv spectral slope is entirely due to reddening , rather then to the presence of evolved stellar populations .
therefore we expect it should only work for actively star forming galaxies .
even more importantly , it is well known that strongest starbursts are opaque to uv radiation and its total sfr activity can not be reliably estimated solely from the uv , even after reddening corrections ( e.g. , goldader et al.2002 ; buat et al .
things can be different though at higher redshifts , as shown later in the paper ( see also flores et al .
2004 ; d05b ; reddy et al .
2005 ) .
it is worth mentioning to which extent the analysis presented in this paper depends on the adopted imf , i.e. , a salpeter ( 1955 ) imf from 0.1 to 100 @xmath19 .
all of the luminosities that we use as sfr indicators are sensitive to the population of high mass stars ( @xmath74 ) , but the bulk of the mass in stars that is actually formed is dominated by lower mass stars .
therefore , changing the slope of the low - mass end of the imf , e.g. , adopting the imf of kroupa ( 2001 ) or chabrier ( 2003 ) that are in better agreement with current observations , would just imply lower overall sfrs , i.e. , different normalizations in and different stellar masses ( by nearly the same factor ) .
we prefer to quote numbers for salpeter imf for consistency with much of the literature in the field .
however , changing the intermediate or high - mass slope could have a significant effect on the relative conversions between different sf indicators , since stars in different mass ranges can be responsible for varying contributions to the different sf indicators .
for example , the radio emission is thought to be powered by higher mass stars than the ir or the uv emission ( e.g. , condon et al .
1992 ) , and a steeper imf slope at the high masses would systematically increase the ir / uv over radio luminosities .
timescale issues , e.g. , observing a starburst very close to the beginning ( roussel et al .
2003 ) or a long time after the sfr peak , can also alter the luminosity ratios .
conceivably , imf and star formation timescale are partly degenerate for some of the effects that we discuss later in the paper .
we neglect timescale effects in the following of the paper .
[ fig : pl_limi ] shows the limiting sensitivity to bolometric ir luminosity from the various tracers using the recipes described in the previous sections ( deep radio and scuba photometry are available only in goods - n ) .
it is clear that the mips 4mu flux density is the most sensitive ir tracer by at least an order of magnitude , allowing in goods to detect galaxies with @xmath75 for most of the @xmath0 redshift range considered here .
this is illustrated in fig .
[ fig : det_rate ] , showing that the fraction of @xmath2 galaxies in our sample detected at 4mu remains very high , over 60% even at the @xmath31 limit of the goods - s field .
even lower sfrs can in principle be derived from the uv over the whole redshift range , provided there is no or just small dust obscuration .
an unreddened galaxy at @xmath5 with @xmath39 ab ( 3@xmath38 limit ) would have sfr@xmath76 m@xmath77 yr@xmath22 .
a more practical limit is given by the faintest galaxy detectable in the @xmath25-band ( @xmath44 ) having no dust reddening , which would have sfr@xmath78 yr@xmath22 for @xmath5 , similar to the depth reached at 4mu .
in the case of galaxies with non zero dust reddening we can measure the sfrs only above higher limits .
for example , a galaxy with @xmath39 and @xmath79 ( the median value at the bright @xmath65 limits ) would have sfr@xmath80 m@xmath77 yr@xmath22 at @xmath5 ( fig .
[ fig : pl_limi ] ) .
it is therefore quite interesting to compare uv and 4mu based sfr and @xmath16 estimates , given that these are the most powerful tracers , and available for nearly all of the @xmath2 galaxies in our goods sample .
the dependence of @xmath16 with redshift is generally steep , and varying as a function of redshift itself , except for scuba 850@xmath15 m photometry . in turn ,
photometric redshift errors propagate into @xmath16 and sfr errors differently when different tracers are used .
we note , however , that this is mitigated when considering the ratios of @xmath16 derived from two independent tracers , as is generally done in this paper . even in this case , however , one can commit significant errors in cases of catastrophic redshift failures , e.g. , if a galaxy at @xmath81 is mistaken to be at @xmath82 . in this case , the @xmath16 derived from the 4mu or 0mu data could be strongly overestimated relative to that derived from , e.g. , 20 cm data or from the uv , due to the stronger k - corrections for the emission in the mips bandpasses .
we take these effects into account when comparing the sfrs from different tracers . in general , the high quality of the adopted photometric redshift ( see fig . [
fig : zphot_n ] ) should limit this problem to a small minority of cases .
in addition , we use the extensive spectroscopic redshift sample to consolidate any conclusion first derived using the full combination of spectroscopic and photometric redshifts .
galaxies in the goods - n field , of the total ir luminosity ( @xmath16 ) as inferred from radio ( using the local radio-@xmath16 correlation ; condon 1992 ; yun et al .
2001 ) to the 8@xmath15 m rest - frame luminosity ( derived from the 4mu observed flux density ) .
black and red symbols are as in fig .
[ fig:24umlir_n ] .
the green points show the results of stacking radio undetected sources in 3 bins of 8@xmath15 m luminosity .
the blue points show the average trend versus 8@xmath15 m luminosity , including both radio detected and undetected sources .
the solid lines show the expected correlation based on the ce01 and dh02 templates .
hard x - ray detected agns are excluded .
, width=332 ] m inferred @xmath16 for @xmath2 objects to the 4mu based luminosities .
for the 5 smgs ( red squares ) we have photometric redshifts @xmath83 .
the green points show stacking of 850@xmath15 m undetected sources .
hard x - ray detected agns are excluded .
, width=332 ]
fig . [ fig:24umlir_n ] shows the flux density versus redshift relation of mips 4mu detected sources in the two goods fields , to the corresponding @xmath25-band limit in each field .
[ fig:24um_n ] shows the bolometric ir luminosity inferred from the 4mu flux densities using the ce01 templates . from both sets of figures
it is apparent that moving to higher redshifts sources with brighter and brighter @xmath16 are found , implying ( after accounting for volume effects ) a rapid brightening of the ir luminosity function ( as already emphasized in the literature ; e.g. , le floch et al . 2005 ; caputi et al .
however , it is somewhat surprising that the rise extends all the way to redshift @xmath84 , and sources with @xmath85 and even @xmath86 appear in the goods fields ( see peres gonzales et al .
2004 ; papovich et al .
2006 ; caputi et al .
2006 for similar results ) . if due to star - formation
, these extreme luminosities would imply obscured sfrs up to 10000 @xmath19 yr@xmath22 .
such extremely high sfrs as inferred from 4mu flux densities are somewhat suspicious , suggesting that the mid - ir light may not be entirely powered by star formation , or may be reflect different intrinsic mid - ir properties of these @xmath5 star forming galaxies , with respect to local sources .
bolometric luminosities derived from the 4mu flux densities are compared to those derived from radio in fig .
[ fig : xlum_n ] , for goods - n @xmath2 galaxies . for radio detected sources
( about 21% of the total sample to @xmath87 ) the average galaxy has @xmath88 and there is overall agreement between 4mu and radio derived luminosities , for both ce01 and dh02 templates .
however , a large scatter is present and there are galaxies with 4mu to radio inferred @xmath16 ratios much different from unity in both directions .
we will come back to these sources later in this section . for 94% ( 46/49 ) of the galaxies with @xmath89 , the radio detection ( or limit )
implies a higher @xmath16 estimated from 4mu with respect to radio , using either ce01 or dh02 templates .
we used radio stacking in order to investigate trends for radio - undetected sources .
green symbols in fig .
[ fig : xlum_n ] show the result for radio - undetected sources , stacked as a function of the 8@xmath15 m rest - frame luminosity , in turn derived from the 4mu flux density .
stacked flux densities are in the range of 37@xmath15jy , comparable to or larger than the single beam noise , and the signal to noise of the detections is larger than 3 in all cases .
the stacking emphasizes the existence of a substantial population of sources with @xmath89 where the @xmath16 estimated from 4mu exceeds on average by an order of magnitude the same quantity derived from radio .
when considering together radio detections and non - detections , we see a good agreement at low @xmath52 luminosities and an overall increasing discrepancy of higher @xmath90 with respect to both ce01 and dh02 templates . for @xmath91 [ @xmath92 ,
the average overestimate is about a factor of 3 .
it can be noticed from fig .
[ fig : xlum_n ] that radio limits move to higher @xmath16 in the right part of the plot .
this corresponds to the fact that most of these putative hylirgs ( from 4mu ) tend to lie at higher redshift than average .
there are eight galaxies , spread over a range of mid - ir and radio luminosities , that on the contrary show a radio - derived bolometric luminosity that exceeds by a factor @xmath93 ( and up to @xmath94 ) that inferred from @xmath52
. this could be due to additional radio emission from an agn ( despite non detection at x - ray wavelengths ) , as suggested also by donley et al .
( 2007 ) , mostly in lower redshift sources .
some sources might have 4mu _ deficit _ , as expected in the case of very low metallicity ( engelbracht et al .
2005 ; madden et al .
2006 ) , or in the case of a strong 9.7 @xmath15 m silicate absorption entering the 4mu bandpass .
the latter case can happen for @xmath95 , and we do find that 6 out of the 8 galaxies are indeed at @xmath96 .
we have cross - checked this result using additional @xmath16 indicators .
[ fig : xlum70_n ] panels show the comparison with luminosities inferred from the mips 0mu flux densities from the survey of frayer et al .
( 2006 ; and in preparation ) . in the area with available deep 0mu imaging from spitzer , about 200 arcmin@xmath97 from the two fields , there are only a handful of individually detected galaxies at @xmath98 .
the 0mu imaging has reached much shallower @xmath16 limits than the 4mu data ; however the non - detections are meaningful for the sources with higher @xmath16 as inferred from the 4mu data and imply that @xmath16 from 4mu is overestimated in those sources . similarly to radio , we have performed stacking of galaxies at 0mu in both goods fields , binning the individually undetected galaxies as a function of their 8@xmath15 m rest - frame luminosities ( see green boxes in fig . [
fig : xlum70_n ] ) .
again , the most notable feature is that sources with the brightest 8@xmath15 m rest - frame luminosities are not correspondingly brighter at 0mu , as would be expected on the basis of local templates .
the average 0mu to 4mu flux density ratio is @xmath99 for sources with @xmath100 , and of@xmath14 for those with @xmath89 .
templates from ce01 and dh02 predict instead that the 0mu to 4mu flux density ratio at @xmath2 should steeply increase with intrinsic @xmath16 .
a similar trend is found by papovich by al .
( 2007 ) in @xmath5 galaxies .
notice that contrarily to the radio case , for a given @xmath16 the 0mu to 4mu flux density ratio is fairly independent on the redshift , for a given template , in the range considered including @xmath101 , due to more similar k - corrections .
therefore these conclusions are relatively robust against uncertainties coming from photometric redshifts . ) .
the latter is defined as the difference between the dust - corrected and the uncorrected sfr estimated from the 1500 rest - frame luminosities .
the blue points show the behavior of the median galaxy , for the usual 3 bins of @xmath102 m rest - frame luminosity .
hard x - ray detected agns are excluded .
, width=332 ] finally , we have used the scuba super - map from borys et al .
( 2003 ) in order to investigate the comparison of 850@xmath15 m versus 4mu inferred @xmath16 values for @xmath2 galaxies .
only 5 galaxies in the sample are individually detected , all at the highest luminosities . for 4 of them
the bolometric luminosities inferred from the 4mu and 850@xmath15 m flux densities are in good agreement with expectations based on local templates , while in a single case the 4mu flux density lead to overpredicting @xmath16 by over an order of magnitude compared to scuba .
it is interesting to notice that this result is different from the findings by pope et al .
( 2006 ) for 850@xmath15 m selected galaxies , where for typical galaxies the 850@xmath15 m flux density was in excess of what could have been predicted from the 4mu flux density .
our 5 sub - mm detected @xmath26 galaxies represents actually a small part of the pope et al .
( 2006 ) sample .
the most likely explanation of the differences is that the sub - mm selection favors galaxies with relatively cold dust , as emphasized also in pope et al .
( 2006 ) .
we stacked 850@xmath15 m undetected sources as a function of their @xmath103 m luminosities .
we detect the intermediate and highest luminosity bins only , both with s / n=4 and @xmath104mjy .
this is consistent with the result presented in d05b and knudsen et al.(2005 ) for scuba stacking of distant @xmath2 sources .
this also confirms the result that for the most luminous 8@xmath15 m rest - frame galaxies , the 8@xmath15 m luminosity is in excess of what the local templates would predict given the @xmath16 of the galaxies estimated from other means . as to first order
the 850@xmath15 m flux density is independent of redshifts for ir luminous galaxies within the probed redshift range , finding the same 850@xmath15 m for the two bins shown in fig .
9 while their average @xmath102 m luminosity differers by a factor of @xmath6 testifies again for a large scatter in @xmath103 m to bolometric @xmath16 luminosity ratio for high-@xmath105 lirgs and ulirgs in our k - selected sample .
d05b found that the uv estimated sfrs , averaged over the full sample of @xmath26 star forming galaxies with @xmath65 , was well within a factor of 2 of the average of the estimates derived from the radio , mid - ir , x - rays and sub - mm data .
this suggested that at least on average uv is a a good tracer of star - formation for massive @xmath5 galaxies , as advocated e.g. , also by reddy et al .
we investigate here to which extent this remains valid when considering individual galaxies .
[ fig : xxx_uv_s ] shows the uv estimate of the _ obscured _ sfr as a function of the 8@xmath15 m rest - frame luminosities , for the goods - s field only where we reach much fainter star - forming galaxies .
this quantity is defined as : _ uv , obscured = sfr_uv , corr - sfr_uv , uncorr , where the correction is for dust extinction , as described in section 3.6 .
the term sfr@xmath106 is equivalent to sfr@xmath45 in .
moreover , entering with this quantity into eq . 1 and solving for @xmath16 one gets the expected infrared luminosity as due to the partial obscuration of the star formation and the reprocessing of its uv photons , i.e. , @xmath107 , also plotted in fig .
[ fig : xxx_uv_s ] . for this comparison
we limit to sources detected at 4mu and to galaxies where the @xmath108 color is well estimated , with total error below 0.4 mag ( corresponding to a maximum acceptable error of a factor of 2.5 in the reddening correction to the sfr ) . at low 8@xmath15 m rest - frame luminosities
, there is a fairly good agreement between ir light ( and sfr ) estimated from uv and from the 4mu flux density , which is within a factor of 2 for most galaxies .
the two estimates start to diverge for @xmath109 , and by @xmath89 the uv underestimates the ir light by more than a factor of @xmath110 compared to the mid - ir . however , given the discussion in the previous section , it is not clear how often this is due to an underestimate of the obscuration of the uv light , or to an excess emission at 4mu . in order to further check the reliability uv - derived sfrs we compared uv - based estimates to radio observations . for radio detected sources ( 20% of the sample ) we find that the median ratio between the ir luminosity estimated from the radio to that from the uv is about a factor of 2 , as shown in fig .
[ fig : rad_lum_n ] .
this is likely the effect of the limiting flux density in the radio , such that only the sources with brightest continuum are detected ( i.e. , the malmquist bias ) . when stacking radio - undetected sources as a function of uv inferred @xmath16 , and considering together radio - detected and undetected sources , fig .
[ fig : rad_lum_n ] shows that the ratio of uv to radio @xmath16 s appear to be very close to unity at all uv luminosities , within the errors .
this implies that the uv - based sfr , corrected for dust reddening , is on average a good tracer of sfr even for the case of ulirgs with @xmath111 . on the other hand , there are also evidences of extreme obscuration at uv wavelengths . for 2 - 3 sources in the sample
the uv underestimates @xmath16 by a factor up to 1030 ( some of these objects might be radio loud agn , despite their non detection in the x - ray ) . for the most luminous sources with @xmath112 based on radio ( a limit above which we should have a complete sample over goods - n ) the median galaxy has uv underestimating @xmath16 by a factor of 2.7 . for the typical ( or median ) ulirg in our sample , however , the uv appears to provide a fairly good estimate of @xmath16 , with no systematic trend / offset with respect to the radio estimates .
this is quite different from local ulirgs , where the uv is known to underestimate @xmath16 by factors @xmath113 ( trentham et al . 1999 ; goldader et al .
2002 ) even after correcting for dust extinction .
similarly , for @xmath2 smgs the uv underestimates sfrs by an average factor of 120 ( chapman et al .
for the 5 galaxies in goods - n to @xmath87 that are detected by scuba at 850@xmath15 m , the 3 objects where the uv slope can be estimated with relatively good s / n the ratio of radio- to uv - derived sfr is in the range between 3 and 35 . for the other two objects , with a poorer quality
uv - derived sfr , this ratio is also @xmath7 .
as we are considering here a subsample of smgs with bright near - ir magnitudes , it is likely that we are limiting to the least obscured sources in their class , hence these results appear to be consistent with chapman et al .
( 2005 ) .
all in all , there seems to be evidence that the typical @xmath2 ulirg is fairly transparent to uv light , the word meaning that we can estimate its sfr given its rest - frame uv 1500 luminosity and uv slope .
however , at much brighter bolometric luminosities , well in excess of @xmath114 , the uv starts to saturate , as inferred from the brightest radio or sub - mm galaxies .
there appears to be a shift of uv saturation level from about @xmath115 at @xmath9 to @xmath16 greater than a few @xmath114 at @xmath5 .
this is in line with the overall rise of the sfr density of the universe during these epochs .
thus , it appears that the uv is on average performing well for the @xmath116 @xmath5 galaxies .
however , for the discrepant cases with @xmath117 ( see again fig . [
fig : xxx_uv_s ] ) it remains unclear whether the discrepancy is due to excess obscuration or to an intrinsic 4mu excess . to disentangle this ambiguity
we have again used the radio as a diagnostic tool . in fig .
[ fig : who_right ] we show the total sfr [ from eq . 2 using @xmath118 as input into eq . 1 ] of the radio - derived sfrs , both normalized to the reddening corrected sfr from the uv . for the sources whose sfr derived from the 4mu flux density exceeds by more of a factor of 3 that derived from the uv , we find that only @xmath11915% of them have a similar or larger excess also in the radio ( and thus could be represent ulirgs with anomalously high obscuration ) . for the majority of 4mu excess objects , however , we find again that radio sfrs agree with uv sfrs , suggesting again that it is 4mu flux density to be genuinely in excess also with respect to the uv luminosity .
color , as defined in d04b ) .
we classify galaxies above the dashed horizontal line as _ mid - ir excess _ galaxies .
the nature of such sources is investigated in detail in .
, width=332 ] the discrepancies that have emerged from the comparisons in the previous sections lead to the question of how well we can correct uv luminosities for dust reddening , given the adopted reddening law , and if the mid - ir excess is somewhat related to dust reddening and uv slopes .
[ fig : ebv_s ] shows the ratio of `` total '' sfr as from eq . 2 estimated from the mid - ir using the 4mu flux densities over the uv - corrected sfr , as a function of the estimated reddening from eq .
. there appears to be no appreciable trend of a larger such ratio as a function of reddening .
this implies that if excess obscuration is present over what corrected using the calzetti et al .
( 2000 ) law , then the uv slope has to be very poorly correlated with such an excess obscuration .
a _ grey _ obscuration would work , but all known dust absorbers are not grey . on the other hand , for the average galaxy , correcting the observed uv luminosities based on the uv slope and using the calzetti et al .
reddening law , allows us to obtain fairly robust sfr estimates , as indeed illustrated by fig .
[ fig : ebv_s ] .
there are a few galaxies where the reddening corrected sfr from the uv exceeds that derived from the mid - ir by more than a factor of 3 . as opposed to 4mu excess objects , fig .
[ fig : who_right ] shows that in most of these cases it is the uv that is overestimating sfrs .
these latter ones are mostly confined to the sources with reddest colors , and are likely galaxies in a post - starburst phase that are on their way to turn passive / quiescent . in these cases , the redness of the uv continuum is due to a dearth of young / massive stars rather than to dust reddening , and the sfr inferred from the uv , corrected for dust extinction following the recipes of section 3.6 , is obviously overestimated .
these sources are intermediate cases between passive @xmath26 galaxies and star - forming ones , and similar to the two blue systems discussed in daddi et al .
( 2005a ) .
the fraction of these systems is fairly small , implying a rapid transition from the star - forming stage to the passive one .
this is related to the possible presence inside @xmath27 samples of sources with passive or quiescent stellar populations , as emphasized by grazian , nonino & gallozzi ( 2007 ) and quadri et al .
the 25% discrepancy between average radio and uv estimates of @xmath16 for the most luminous uv galaxies ( see the blue points in fig . [
fig : rad_lum_n ] ) is completely solved by excluding these nearly quiescent sources ( about 20% of the radio undetected galaxies in that bin ) .
clearly , mips 4mu photometry is required to identify these likely _ post starburst _ galaxies , where the uv corrected for dust extinction provides an incorrect estimate of the ongoing sfr activity .
the results of the previous sections have clearly singled out the existence of a population of @xmath2 galaxies that we call mid - ir excess galaxies , for which the mips 4mu flux densities would imply ir bolometric luminosities and sfr in excees to those implied by all other available tracers of star formation activity .
we emphasize that the existence of these mid - ir excess galaxies is independent on the classes of templates used or on the particular recipe to convert an observed 4mu flux density at @xmath2 into a measure of sfr/@xmath16 .
the result shown in can be read as evidence that at a fixed bolometric luminosity around @xmath120 there is a very large scatter of @xmath49 , and one can objectively define mid - ir excess galaxies as the galaxies with the largest observed @xmath49 for a givel @xmath16 .
understanding and investigating the nature of these sources and the origin of the mid - ir excess has potentially far reaching implications .
these aspects are directly developed in full in . in such paper
we provide evidence indicating that mid - ir excess galaxies contain heavily obscured , often compton thick , agns .
we then argue that the agn energetic input is the most likely cause of the mid - ir excess in these galaxies .
we summarize the work described in the previous sections by suggesting a recipe for obtaining the best estimates of sfrs for samples of @xmath62 galaxies , by exploiting a variety of multiwavelength sfr indicators as available in goods ( and many of the current surveys , although to less deep levels , such as e.g. , in cosmos , scoville et al .
this procedure assumes that galaxies hosting relatively unobscured and luminous agn have been eliminated from the sample , e.g , . by means of deep x - ray imaging .
* compare uv - derived sfrs ( sfr@xmath121 ) , obtained as described in section [ uv_s ] , to the ones obtained from the 24 @xmath15 m flux density , using the ce01 or dh02 templates .
for the sources where sfr@xmath122 and sfr@xmath121 agree within a factor of @xmath123 ( fig .
[ fig : who_right ] ) , use the 24 @xmath15 m based sfr plus the uv - based sfr uncorrected for reddening ( i.e. , sfr@xmath122 ) as the best estimate for the galaxy sfr .
these sources are called _ mid - ir normal _ star forming galaxies . * for the sources where sfr@xmath122 is lower than about one third of sfr@xmath121 , use only the 24@xmath15 m based sfr ( sfr@xmath124 ) as the best estimate . these sources are not likely to have the uv / optical light dominated by young massive stars , and are probably quiescent or post starbursts ( fig .
[ fig : who_right ] ) . * in the cases where sfr@xmath122 is larger than about 3 times sfr@xmath121 , i.e. , for _ mid - ir excess _
galaxies , use sfr@xmath121 as the best sfr indicator , as it is likely that there is a contribution to the 24 @xmath15 m flux density from other than star formation .
availability of deep radio imaging and/or submm / mm imaging , can allow one to identify the subsample of galaxies among these that are actually _ opaque _ to uv light , and longer wavelength estimates of sfr are to be used in the latter . to our understanding
these recipes makes the best possible use of the multiwavelength database currently available , in order to cope with the limitations of each individual star formation indicator .
in particular , they are meant to single out those sources where a particular method is bound to fail , e.g. , because the uv slope reflects age rather than reddening effects , or the mid - ir output is partly powered by an agn . in principle , the similarity in the trends observed in fig .
710 could suggest to define modified , highly non linear relations between 8@xmath15 m luminosities and @xmath16 for @xmath5 galaxies , as suggested also by papovich et al .
( 2007 ) , for recovering corrected @xmath16 for all galaxies . however , fig .
[ fig : corre ] shows that in practice any such relation would be dependent on the sample properties itself , e.g. shifting as a function of @xmath25-band magnitude , and therefore not well defined in general .
in this section , we exploit the results presented so far , and in particular the improved sfrs for our sample of star - forming galaxies at @xmath2 , in order to address a number of issues which are relevant for our understanding of the nature of star - formation and mass assembly in such galaxies . as already shown by d05b ( see also papovich et al .
2006 ; reddy et al . 2006a ; caputi et al .
2007 ) , the typical massive ( @xmath125 yr@xmath22 ) star - forming galaxy at @xmath5 is a ulirg , i.e. , it appear to have @xmath88 or more . here
, we provide a more detailed estimate of the space density of @xmath5 ulirgs using multiwavelength information . in goods -
n there are 113 @xmath26 galaxies that are classified as ulirgs based on their 24 @xmath15 m flux densities , 58 based on uv sfrs , and 45 based on radio emission ( or up to 69 if considering the undetected radio sources with limits still consistent with the ulirg regime ) . similarly ,
in goods - s there are 155 @xmath26s classified as ulirgs from their 24 @xmath15 m flux densities , and 50 from the uv sfrs . at luminosities of @xmath68
the mid - ir starts to be systematically in excess of radio ( fig .
[ fig : xlum_n ] ) , and therefore ulirg densities derived from mips 4mu data are likely to be overestimated .
we notice however that the number of ulirgs inferred from mips 24@xmath15 m emission does not increase appreciably when moving from the sample with @xmath29 in goods - n to the one limited at @xmath126 in goods - s . on the other hand , there can be some sources that are not classified as ulirgs in the uv due to obscuration . also
, the radio flux density limits over this redshift range are very close to the ulirg detection threshold , and hence a radio - based estimate could be somewhat incomplete . therefore , we average the uv / radio and mips estimates to derive ulirg space densities at @xmath2 . combining goods - n ( 154 arcmin@xmath97 ) and goods - s ( 140 arcmin@xmath97 used in this work ) , and correcting the ulirg counts to compensate for the fact that we are including only sources not blended in the irac images , we finally derive a sky density of ulirgs of 0.6 arcmin@xmath127 . using the comoving volume within @xmath0 ( fig . [
fig : zhist_n ] ) , this implies a space density of @xmath128 mpc@xmath129 .
the uncertainty in this density is order of 0.2 dex , when accounting for the scatter in the ulirg counts from the different @xmath16 tracers , as mentioned above .
we note that we have excluded galaxies with hard x - ray detections from this analysis as likely hosting agns . if we use the 24@xmath15 m emission for the x - ray agns to estimate @xmath16 , the sample of ulirgs increases by 15 - 20% , thus still within the range of uncertainties estimated above .
the space density of ulirgs at @xmath5 is a factor of 1000 larger than in the local universe ( e.g. , sanders er al .
2003 ) , confirming the previous estimate of d05b .
and @xmath8 inferred from goods ( red filled circles ) is compared to the local density from sanders et al .
( 2003 ; filled green circle ) and to predictions of star - forming galaxy density from the mock lightcones based on the millennium simulation ( kitzbichler & white 2007 ) .
the rate of 120@xmath19 yr@xmath22 corresponds to a ulirg in the case of a kroupa ( 2001 ) or chabrier ( 2002 ) imf .
, width=332 ] in a similar way , we derive the space density of ulirgs in goods at @xmath8 , taking galaxies with @xmath130 ( using photometric and spectroscopic redshifts ) and estimating @xmath16 using 4mu flux densities .
we find only 7 such sources in the combined goods , which implies a space density of @xmath131 , an order of magnitude lower than that at @xmath5 .. ] in fig .
[ fig : ulirgs_mille ] we compare these results to predictions from the millennium simulation of galaxy formation ( springel et al .
we use the lightcones based on semianalitic models presented by kitzbichler & white ( 2007 ) . for a kroupa
( 2001 ) or chabrier ( 2003 ) imf the ulirg regime corresponds to a star - formation rate of 120@xmath19yr@xmath22 , lower than the rate reported in eq . 1 for a salpeter imf .
we considered all simulated galaxies with sfr above this threshold as ulirgs . we explicitly quote here results for the kroupa ( 2001 ) and chabrier ( 2003 ) imfs because these are favored by observations in the local universe and therefore more appropriate for comparison to predictions of sfrs in galaxies .
the model underpredicts the space density of ulirgs at both @xmath8 and @xmath5 by at least an order of magnitude . only at higher redshifts , @xmath132
, do the millennium models predict a ulirg space density as large as that which we observe at @xmath133 .
this suggests that star - formation in massive galaxies occurs too early in the simulations , or was terminated too early due to the adopted truncation of sfr from agn feedback ( e.g. , croton et al .
2006 ) . however , we reach similar conclusion when comparing to simulated galaxy catalogs based on the model of oppenheimer & dav ( 2006 ) , which also predicts a comparably large space density of ulirgs only at @xmath134 .
counts from the two fields of the large area survey of kong et al . ( 2006 ) .
predictions obtained from the mock lightcones based on the millennium simulation ( kitzbichler & white 2007 ) are also shown , for galaxies with @xmath11 and sfr@xmath135 and 10 @xmath19yr@xmath22 .
, width=332 ] one possible means for reconciling the models with the observations would be to delay star - formation until later epochs .
therefore , it is important to compare observations and model predictions for the abundance of passive , non - star - forming galaxies at @xmath133 .
in fact , the feedback from agns has been implemented in the millennium models in order to allow them to reproduce the existence of passive and massive galaxies both at high and low redshifts ( de lucia et al .
the oppenheimer & dav ( 2006 ) models , lacking the crucial mechanism of agn feedback , predict very low space density of passive galaxies at @xmath13612 , much below the observed density . in fig .
[ fig : pbzk_mille ] we show the number counts of _ passive _ @xmath26 ( @xmath28 ) galaxies from kong et al .
these objects , selected according to the criteria defined in d04b , are expected to provide a fair census of the population of @xmath11 passively evolving systems ( see , e.g. , daddi et al 2005a ; arnouts et al .
2007 ) . in goods the b - band data are not deep enough to reliably distinguish passive from star - forming @xmath26 galaxies using the color definition of d04b . however , we may approximate this by using the number counts of goods @xmath26 galaxies that have sfr@xmath137yr@xmath22 based on the 4mu measurements . as we have shown in fig .
[ fig : pl_limi ] , the mips 4mu observations are very sensitive for detecting low rates of star formation , and in particular to identify galaxies with red uv slopes but relatively low sfrs ( fig .
[ fig : who_right ] ) .
the counts of galaxies with sfr@xmath138yr@xmath22 from the combined goods n&s fields are shown in fig .
[ fig : pbzk_mille ] .
these are lower than the @xmath28 counts of kong et al .
( 2006 ) , most likely due to cosmic variance , as these sources are very strongly clustered ( kong et al .
. the larger cosmos survey may provide more robust statistics on the abundance of passive galaxies , although we note that the 4mu data in cosmos ( sanders et al .
2007 ) are shallower and can not set such tight constraints on low sfrs at @xmath133 .
overall , the millennium simulations predictions , shown in fig .
[ fig : pbzk_mille ] , seem to reproduce the data reasonably well , given the large uncertainties , perhaps somewhat overestimating the counts of galaxies with relatively low sfrs .
the millennium models might under - predict the ulirg space density if the star formation episodes were short lived , so that the peak sfr is missed caught when averaging over the simulation time steps ( 250 myr at @xmath5 for the evolution of the dark matter , with much finer steps for physical processes computed in the semi - analytic models ) .
it is therefore important to try to constrain the duration of the phase at high redshifts .
this is also relevant in order to evaluate the contribution from this star - formation mode to the assembly of stars in massive galaxies . in d05b
it was suggested that duty cycles are high , based on the very large detection rate and brightness of the most massive galaxies at 4mu .
caputi et al .
( 2006 ) agree with this conclusion , based on their 4mu study of @xmath5 galaxies in goods .
we investigate this point in more detail here , in the light of our new results on the comparison of the different star formation estimators . in order to do this
, we consider a mass and volume - limited sample of @xmath26 galaxies , and evaluate the fraction of ulirgs in the sample . if ulirgs stopped forming stars , they would remain in the mass - selected sample as passive or slowly star - forming galaxies .
the fraction of ulirgs in this mass - selected sample is therefore a lower limit to their duty cycle , as the ulirg phase could have started well before the galaxies reached the required mass threshold , or at an earlier era , before the redshift boundary of the sample considered here . to build a mass - limited sample of @xmath2 galaxies
, we consider all of the @xmath11 @xmath26 galaxies , including both star forming and passive galaxies .
the latter were excluded for the analyses in previous parts of the paper .
we use the empirical mass estimate described in d04b that is based on detailed sed fitting results with the bruzual & charlot ( 2003 ) models performed by fontana et al .
this calibration allows us to associate a mass to each galaxy based on its k - band magnitude , with a correction for varying @xmath139 ratio based on the observed @xmath140 color .
maraston et al .
( 2006 ; 2007 ) have shown that for passive galaxies , where the overall sed is dominated by @xmath1411 gyr old stars , the use of the maraston ( 2005 ) model libraries results in masses that are smaller by 0.20.3 dex ( see also bruzual 2007 ) . for star - forming @xmath26 galaxies , where the young stellar component dominates the light , we find ( maraston et al .
, in preparation ) that the maraston models provide @xmath139 ratios that are consistent on average with those from bruzual & charlot ( 2003 ) . in the @xmath142 sample constructed in this manner
, we find 34 galaxies at @xmath2 in goods - s and @xmath143 in goods - n .
only 1/34 of the goods - s galaxies above this mass threshold has @xmath144 , which suggests that the sample is fairly complete also in the goods - n to this limit .
the space density of massive galaxies with @xmath145 is about @xmath146 mpc@xmath129 .
this estimate is within a factor of @xmath147 of the one reported by van dokkum et al .
( 2006 ) using the larger music survey , despite the different methods for estimating stellar masses . in order to estimate sfrs , we use both mips 4mu and the uv luminosities corrected for reddening . while we have shown that sfrs inferred from 4mu are suspect if they are substantially larger those derived from the uv data , the mips data are also very efficient for measuring relatively low sfrs in red galaxies where uv - reddening estimates would be in turn overestimated ( see fig .
[ fig : who_right ] ) .
therefore , the mips data are ideal for isolating the most quiescent galaxies even in the case that some low level residual sfrs place these sources in the star forming @xmath26 region . by taking the minimum sfr for each object , estimated from either the 4mu flux density or from the extinction corrected uv luminosity , we find that the fraction of ulirgs among the most massive galaxies at @xmath133 is @xmath148% ( poisson counting error only ) .
this very conservative estimate sets a similar constraint on the duty cycle of the ulirg phase in massive high-@xmath105 galaxies , i.e. , @xmath14940% .
the cosmic time span elapsing during the redshift range @xmath0 is about 2 gyr , with only 1 gyr per galaxy available on average ( given the reasonably flat redshift distribution within these limits ) .
this implies that these ulirg phase will last for at least some 400 myr , one order of magnitude longer than what is currently estimated for smgs at similar redshifts , or for local ulirgs ( see , e.g. , greve et al .
2005 ; tacconi et al . 2006 ; although see also swinbank et al .
2006 ) . , but here simulated galaxies at @xmath150 are plotted , taken from the millennium lightcones of kitzbichler & white ( 2007 ) .
the cyan line is a power law with slope of 0.65 , and the blue solid line is the same of the one in fig .
[ fig : sfr - mass ] , showing the average trend for the goods @xmath26 sample at the same redshift .
, width=332 ] we can gain further insights on the nature of star - formation at @xmath5 by examining correlations between galaxy mass and star formation rates .
recently , noeske et al . ( 2007 ) and elbaz et al . ( 2007 ) have shown that star formation and stellar mass define a tight correlation ( 0.2 dex scatter ) in galaxies at @xmath151 , with rough proportionality ( logarithmic slope of 0.9 ) .
similar proportionality is also seen at @xmath9 in data from the sloan digital sky survey ( elbaz et al .
2007 ) , although with a lower normalization reflecting the overall decline in cosmic sfr density with time .
[ fig : sfr - mass ] ( left panel ) shows the sfr- stellar mass relation for galaxies at @xmath133 in goods - s using the uv to derive sfrs . in this plot , we show only those sources with mips detection , with 4mu - derived sfr not less than 1/2 of the uv one ( to avoid estimating sfr from uv for evolved galaxies , see fig .
[ fig : who_right ] ) and with a well - defined uv slope for obtaining an accurate reddening estimate .
we find that a proportionality with a logarithmic slope of 0.9 ( blue line in fig .
[ fig : sfr - mass ] ) , normalized to the median sfr / m@xmath152 ratio of the sample ( 2.4 gyr@xmath22 ) , provides a good fit to the data , with semi interquartile range of only 0.16 dex in the sfr / m ratio .
we have cross - checked the validity of this relation in goods - n using the deep radio observations .
we have divided the galaxies ( limited to @xmath87 , a smaller range than what plotted in fig .
[ fig : sfr - mass ] ) in 3 independent mass ranges and obtained the average sfr in the bin using the radio , averaging together radio detections and non detections .
the green points in fig .
[ fig : sfr - mass ] show that the star formation results from the stacked radio emission agree very well with the relation that we derive from the uv light .
this check is important because it is based on an indicator that should , in principle , be more solid than uv , and also because it allows us to include many of the massive galaxies for which uv sfr could not be estimated due to their faintness in the optical bands . due to the substantial presence of mid - ir excess objects at @xmath133
, we do not expect that this relation can be accurately recovered using the 4mu inferred sfrs .
this explains why caputi et al .
( 2006b ) find a much looser correlation between sfr and masses for @xmath5 galaxies .
if we use 4mu inferred sfr only for those galaxies where this quantity is within a factor of 3 of the uv one ( the mid - ir normal objects ) , we obtain a sfr - mass correlation fully consistent with that based on uv ( fig .
[ fig : sfr - mass ] , right panel ) .
the @xmath5 correlation appears to have a slope similar to that at lower redshifts .
instead , the normalization at @xmath5 is a factor of 3.7 larger than that at @xmath8 , and 27 times larger than at @xmath9 ( elbaz et al . 2007 ; noeske et al . 2007 ) .
at fixed stellar mass , star forming galaxies were much more active on average in the past .
this is most likely due to a larger abundance of gas , depleted with passing of time .
the inferred correlation is quite tight , with semi interquartile range of only 0.16 dex in the dispersion of specific sfrs .
we caution that , having used mainly the uv as a sfr estimator , we can not reliably rule out the presence of a larger number of outliers at low stellar masses , for which we might strongly underestimate the sfr from the uv . these can not be reliably identified to meaningful depths with radio data ( due to the flux density limits of current observations ) , nor at 4mu ( due to the existence of mid - ir excess sources ) .
we will have to wait for herschel and alma to address this issue accurately .
submillimeter - selected galaxies are strong outliers to this trend , however .
tacconi et al .
( 2006 ) estimate that smgs in their sample have typically @xmath69 and dynamical masses @xmath153 ( see also greve et al . 2005 for similar results ) . for a given stellar mass ( assuming that most of the dynamical mass in the central regions of smgs is stellar )
, smgs are forming stars at a 10x or larger rate respect to ordinary massive star forming galaxies .
their space density is also approximately an order of magnitude smaller .
smgs at @xmath5 appear to be like lirgs and ulirgs at @xmath9 , i.e. , relatively rare objects and outliers of the mass - sfr correlation ( see elbaz et al .
2007 ) . to this regard ,
it is not surprising that much shorter star formation duty cycles and lifetimes have been inferred for smgs and local ulirgs ( @xmath154myr ; greve et al .
2005 ; solomon & vanden bout 2005 ) .
these represent short - lived stages of the life of galaxies , due , e.g. , to ongoing mergers or some temporary perturbations ( see , e.g. , dannerbauer et al .
2006 for a discussion ) .
ulirgs and smgs also generally have smaller physical sizes ( tacconi et al .
2006 ) than those of more ordinary massive , star - forming galaxies at @xmath155 ( d04a ; ravindranath et al . 2007 , in preparation ) , suggesting that the smgs are in more advanced merger states .
we have used the mock lightcones from kitzbichler et al .
( 2007 ) , based on the millennium simulations , to explore the comparison of mass and sfr at @xmath5 in these models ( fig .
[ fig : sfr - mass - mille ] ) . as emphasized already by finlator et al .
( 2006 ) theorethical simulations quite naturally predict the existence of correlations between galaxies sfrs and stellar masses .
however , we find that at fixed stellar masses , the model galaxies are forming stars at about 1/4 of the observed rate for galaxies with @xmath156 .
the correlation is also substantially tilted , with decreasing specific sfr at larger masses .
it seems that a major change required for models would be to increase the star - formation efficiency ( and thus the typical sfr ) at all masses for star - forming galaxies at redshifts @xmath157 , while still keeping the current proportion of massive galaxies in a passive / quiescent state .
interestingly , we find that if we reproduce fig .
[ fig : sfr - mass - mille ] plotting simulated millennium galaxies at @xmath158 instead of @xmath159 , we find that simulated galaxies match the @xmath5 goods galaxies quite accurately , with the same sfr versus mass normalization , slope and with similar scatter .
this again reinforces the idea that sfr and mass growth happens too early in the current version of the simulations .
the observed mass - sfr correlation defines a more basic dichotomy of galaxy properties than those based on colors , as emphasized also by elbaz et al .
( 2007 ) . in future papers
, we will investigate other physical properties of @xmath5 galaxies ( like morphology ) as a function of their distance from the correlation .
finally , we note that the trend of sfr increasing with galaxy mass , and the fact that star formation is sustained with long duty cycles , formally imply very rapid growth of the stellar mass during the period spanned by the observations .
we confirm the conclusions of d05b , that the epoch within @xmath0 corresponds to the major build - up period of the most massive galaxies in the local universe .
we have investigated the properties of k - selected , star forming galaxies at @xmath43 in the goods fields , by means of over 1000 galaxies with @xmath160 , several hundred of which have measured spectroscopic redshifts .
we have used the deepest observations at a variety of wavelengths , including radio , spitzer 4mu and 0mu , scuba 850@xmath15 m and uv light , in order to derive and compare multiwavelength luminosities of galaxies , and to investigate the reliability of these luminosities for measuring star formation rates .
our findings can be summarized as follow : * for most galaxies with moderate mid - ir luminosities , @xmath161 , spitzer 4mu measurements can be used to estimate total infrared luminosities and star formation rates of galaxies at @xmath2 .
the ratios of the mid - ir luminosities to those at other wavelengths ( radio , 0mu , 850@xmath15 m and even reddening - corrected uv light ) are consistent with expectations from local correlations , provided that their mid - ir luminosities is not too large .
* however , most galaxies with @xmath162@xmath163 show a mid - ir excess with respect to all other luminosities , compared to that expected from local correlations , and on ratios observed in less luminous @xmath5 galaxies .
we therefore report the discovery of a large population of @xmath2 galaxies with intrinsic mid - ir excess .
we investigate the nature of these sources in a companion paper ( daddi et al . 2007 ; hereafter ) , concluding that these objects contain heavily obscured , often compton thick , agn . *
the presence of galaxies with seemingly extreme @xmath164 @xmath19 yr@xmath22 estimated from 4mu flux densities is nearly always due to the existence of these mid - ir excess sources
. a large proportion of 4mu sources at @xmath43 detected in spitzer surveys shallower than goods will be mid - ir excess objects , and their star formation rates and infrared luminosities may be greatly overestimated if bolometric corrections from local starburst galaxy templates are applied .
the true sfr of these objects rarely exceeds a few hundred @xmath19 yr@xmath22 .
* for typical @xmath5 galaxies , observed at fainter 24@xmath15 m flux densities , the local correlations among the sfr tracers that we have probed here still hold within the uncertainties . * the observed uv luminosity , corrected for dust reddening based on the observed uv slope using a calzetti et al .
( 2000 ) extinction law , allows one to obtain reliable estimate of the sfr activity for a large majority of @xmath5 galaxies .
this is demonstrated in comparison with mid - ir estimates ( except for sources with intrinsic mid - ir excess ) , and with radio , 0mu and 850@xmath15 m observations through stacking in bins of uv and mid - ir luminosity .
the typical ulirg at @xmath5 is transparent to uv light , contrary to present
day ulirgs and to @xmath5 smgs .
* as exceptions , we do find a minority of @xmath2 galaxies which appear to be opaque to uv emission from star formation . even after correction for dust reddening ,
the sfr derived from the uv light in these objects is underestimated by large amounts .
these objects are similar to smgs , and given the depths of current goods multiwavelength data , they are best recognized as highly obscured outliers by comparing their uv and radio properties . *
similarly , there is a small population of k - selected sources that are classified as star forming galaxies based on optical - uv colors ( e.g. , with the @xmath26 diagram ) but for which the uv estimates of sfr greatly overpredict the true amount of sfr .
these are probably post - starburst galaxies , whose red uv colors are not only due to dust reddening , but also have important contributions from intrinsically cooler , lower - mass stars . also these objects can be identified from the comparison of uv and mid - ir based sfrs . * exploiting all available @xmath5 sfr estimators , we have derived a measure of the space density of ulirgs at @xmath2 in the goods fields .
we find a space density of about @xmath165 mpc@xmath129 and a sky density of 0.6 arcmin@xmath127 at @xmath166 .
this is a factor of 10 larger than the abundance predicted by semianalitic models based on the millennium simulations . *
similarly , we have used our improved understanding of @xmath5 sfr tracers to investigate the abundance of @xmath2 massive galaxies with low ongoing sfr ( below 10 @xmath19 yr@xmath22 ) .
we have computed number counts of these galaxies as a function of their k - band magnitude , and find that the counts are lower than that of passive ( @xmath28 ) objects selected by kong et al .
( 2006 ) , likely due to cosmic variance and possibly to some contamination of @xmath28 samples from @xmath167 dusty galaxies .
predictions based on the millennium simulations can reasonably well account for the space density of these galaxies .
+ * we have built a mass - limited sample of galaxies with @xmath145 at @xmath2 . by estimating the distribution of sfrs in these galaxies
, we derive lower limits to the duty cycle of ulirgs .
this is defined as the fraction of ulirgs inside the mass - limited sample .
we conclude that this duty cycle is at least 0.4 .
this is much larger than the duty cycle estimated for local ulirgs , and also for @xmath5 smgs .
the relatively large duty cycles imply sfr duration at ulirg levels of order of 0.5 gyr or more .
this also implies that large gas reservoirs have to exist in most of the massive @xmath5 galaxies , or that gas must be accreted very efficiently over time .
* we have found that a relatively tight stellar mass - sfr relation is already in place at @xmath5 , for star forming galaxies detected at 4mu , with a scatter ( semi - interquartile range ) of only 0.16 dex in the specific star formation rate .
this has been derived mainly basing on the uv estimates of sfrs , and confirmed through radio stacking of galaxies in mass bins and using sfrs derived from the 4mu flux density , once mid - ir excess objects are excluded from the sample . the typical star forming galaxy at @xmath5 is forming stars more rapidly by factors of 3.7 and 27 respect to an object with similar mass at @xmath8 and @xmath9 .
current numerical simulations are able to reproduce the mass - sfr correlation with similar slope and similarly small scatter to what observed , but with much lower normalization than the observed one . *
we conclude by suggesting that current realizations of galaxy formation models could come to fairly good agreement with the observations available for @xmath5 galaxies if , by some means , the major periods of star formation activity could be delayed until later epochs than in current implementations .
+ we thank the rest of the gmass team for allowing us to use the still unpublished spectroscopic redshifts , and the the many other members of the goods team , who have helped to make these observations possible .
we are grateful to emily macdonald , daniel stern and hy spinrad for collecting some of the redshifts used in this work , and to manfred kitzbichler , romeel dav and collaborators for the galaxy catalogs based on their model predictions .
ed gratefully acknowledges nasa support ( at the beginning of this work ) through the spitzer fellowship program , award 1268429 .
dma thanks the royal society for funding .
jk acknowledges financial support from the german science foundation ( dfg ) under contract sfb-439 . support for this work , part of the spitzer space telescope legacy science program , was provided by nasa , contract number 1224666 issued by the jpl , caltech , under nasa contract 1407 .
alexander , d. m. , et al .
2003 , , 125 , 383 armus l. , charmandaris v. , bernard - salas j. , et al . , 2006 , astro - ph/0610218 arnouts s. , walcher c.j . , le fevre o. , et al . ,
2007 , submitted to a&a ( arxiv:0705.2438 ) borys c. , chapman s. , halpern m. , scott d. , 2003 , mnras , 344 , 385 bruzual , g. & charlot , s. 2003 , , 344 , 1000 bruzual g. , 2007 , proceedings of the meeting `` from stars to galaxies : building the pieces to build up the universe '' , eds .
a. vallenari , r. tantalo , l. portinari , and a. moretti , asp conf .
( in press ) ( astro - ph/0702091 ) buat v. , et al . , 2005 , apj , 619 , l51 caputi k.i . , et al . , 2006a , apj , 637 , 727 caputi k.i . , et al . ,
2006b , apj , 454 , 143 caputi k.i .
, et al . , 2007 ,
apj , in press ( astro - ph/0701283 ) calzetti d. , armus l. , bohlin r. c. , et al . , 2000 , apj , 533 , 682 chabrier g. , 2003 , apj , 586 , l133 chapman , s. c. , blain , a. w. , smail , i. , & ivison , r. j. 2005 , , 622 , 772 chary r. , elbaz d. , 2001 , apj , 556 , 562 ( ce01 ) coleman , g. d. , wu , c .- c . , & weedman , d. w. 1980 , , 43 , 393 condon j. j. , 1992 , ara&a , 30 , 575 croton d. j. , et al . , 2006 , mnras , 365 , 11 daddi e. , cimatti a. , renzini a. , et al .
, 2004a , apj , 600 , l127 ( d04a ) daddi e. , cimatti a. , renzini a. , et al . , 2004b , apj , 617 , 746 ( d04b ) daddi e. , et al .
, 2005a , apj , 626 , 680 ( d05a ) daddi e. , dickinson m. , chary r. , et al , 2005b , apj , 631 , l13 ( d05b ) daddi e. , alexander d.m .
, dickinson m. , et al .
, 2007 , apj in press ( arxiv:0705.2832 ) ( ) dale d.a .
, helou g. , 2002 , apj , 576 , 159 ( dh02 ) dale d.a . , et al .
, 2005 , apj 633 857 dannerbauer h. , et al . , 2006 ,
apj , 637 , l5 de lucia g. , springel v. , white s. d. m. , croton d. , kauffmann g. , 2006 , mnras , 366 , 499 dickinson m. , papovich c. , ferguson h. c. , budavri t. , 2003 , apj , 587 , 25 elbaz d. , cesarsky c.j . , chanial p. , et al . , 2002 ,
a&a , 384 , 848 elbaz d. , daddi e. , le borgne d. , et al . , 2007 ,
a&a in press ( astro - ph/0703653 ) engelbracht c. w. , et al . , 2005 , apj , 642 , l127 finlator k. , dav r. , papovich c. , hernquist l. , 2006 , apj , 639 , 672 flores h. , hammer f. , elbaz d. , cesarsky c. j. , liang y. c. , fadda d. , gruel n. , 2004 , a&a , 415 , 885 fontana a. , et al .
, 2004 , a&a , 424 , 23 fontana a. , et al . , 2006 , a&a , 459 , 745 frster schreiber n. m. , roussel h. , sauvage m. , charmandaris v. , 2004 , a&a , 419 , 501 frster schreiber n. m. , van dokkum p. g. , franx m. , et al . , 2004 , apj , 616 , 40 franceschini a. , et al . , 2006 , a&a , 453 , 397 frayer d.t . , huynh m.t . , chary r. , et al . , 2006 ,
apj , 647 , l9 giavalisco m. , ferguson h.c .
, koekemoer a. , et al . , 2004 ,
600 , l93 goldader , j. d. , meurer , g. , heckman , t. m. , seibert , m. , sanders , d. b. , calzetti , d. , & steidel , c. c. 2002 , , 568 , 651 grazian a. , fontana a. , de santis c. , et al . , 2006 ,
a&a , 449 , 951 grazian a. , nonino m. , gallozzi s. , 2007 , a&a in press ( astro - ph/0701233 ) greve t. r. , et al .
, 2005 , mnras , 359 , 1165 hornschemeier a. e. , heckman t. m. , ptak a. f. , tremonti c. a. , colbert e. j. m. , 2005 , aj , 129 , 86 kennicutt r. c. , 1998 , ara&a , 36 , 189 kitzbichler m.g . ,
white s.d.m . , 2007 , mnras , 376 , 2 knudsen k. k. , et al .
, 2005 , apj , 632 , l9 kong x. , daddi e. , arimoto n. , et al . , 2006 , apj , 638 , 72 kroupa p. , 2001 , mnras , 322 , 231 kurk j. d. , cimatti a. , daddi e. , et al . , 2006 , to appear in the proceedings of `` infrared diagnostics of galaxy evolution '' , pasadena , november 2005 ( astro - ph/0604132 ) le fvre o. , et al . , 2004 ,
a&a , 428 , 1043 le floch e. , et al . , 2005 , apj , 632 , 169 madden s. c. , galliano f. , jones a. p. , sauvage m. , 2006 , a&a , 446 , 877 maraston c. , 2005 , mnras , 362 , 799 maraston c. , daddi e. , renzini a. , cimatti a. , dickinson m. , papovich c. , pasquali a. , pirzkal n. , 2006 , apj , 652 , 85 maraston c. , daddi e. , renzini a. , cimatti a. , dickinson m. , papovich c. , pasquali a. , pirzkal n. , 2007 , apj , 656 , 1241 ( erratum ) meurer g. r. , heckman t. m. , calzetti d. , 1999 , apj , 521 , 64 mignoli m. , et al .
, 2005 , a&a , 437 , 883 noeske k.g .
, weiner b.j . ,
faber s.m . , et al .
, 2007a , apj letters , in press ( astro - ph/0701924 ) noeske k.g .
, faber s.m .
, weiner b.j . , et al .
, 2007b , apj letters , in press ( astro - ph/0703056 ) oppenheimer b. d. , dav r. , 2006 , mnras , 373 , 1265 papovich c. , et al . , 2006 , apj , 640 , 92 papovich c. , rudnick g. , le floch e. , et al . , 2007 , apj , in press ( arxiv:0706.2164 ) prez - gonzlez p.g . , rieke g.h . , egami e. , et al . , 2005 , apj , 630 , 82 persic m. , rephaeli y. , braito v. , et al . , 2004 , a&a , 419 , 849 pope a. , borys c. , scott d. , conselice c. , dickinson m. , mobasher b. , 2005 , mnras , 358 , 149 pope a. , scott d. , dickinson m. , 2006 , et al . , mnras , 370 , 1158 pozzetti l. , bolzonella m. , lamareille f. , et al . ,
2007 , submitted to a&a ( arxiv:0704.1600 ) quadri r. , et al .
, 2007 , apj , 654 , 138 ranalli p. , comastri a. , setti g. , 2003 , a&a , 399 , 39 reddy n.a . , et al . , 2005 , apj , 633 , 748 reddy n.a . , et al . ,
2006a , apj , 644 , 792 reddy n.a .
, et al . , 2006b , apj , 653 , 1004 richards e. a. , 2000 , apj , 533 , 611 roussel h. , helou g. , beck r. , et al . , 2003 , apj , 593 , 733 rudnick g. , et al . , 2003 ,
apj , 599 , 847 rudnick g. , et al . , 2006 , apj , 650 , 624 sanders d. b. , mazzarella j. m. , kim d .- c . , et al . , 2003 ,
aj , 126 , 1607 sanders d. b. , et al . , 2007 , apjs in press ( astro - ph/0701318 ) scoville n. , benson a. , blain a.w . , et al . , 2007 , apjs in press ( astro - ph/0612306 ) shapley a. , et al . , 2005 ,
apj , 626 , 698 smith j.d.t . ,
draine b.t . ,
dale d.a . , et al . , 2006 ,
apj in press astro - ph/0610913 solomon p. m. , vanden bout p. a. , 2005 , ara&a , 43 , 677 spergel d.n . , et al . , 2007 ,
apj , in press ( astro - ph/0603449 ) springel v. , white s.d.m .
, jenkins a. , et al . , 2005 , nature , 435 , 629 steidel , c. c. , shapley , a. e. , pettini , et al . , 2004 , , 604 , 534 swinbank a. m. , chapman s. c. , smail i. , lindner c. , borys c. , blain a. w. , ivison r. j. , lewis g. f. , 2006 , mnras , 371 , 465 szokoly g. p. , bergeron j. , hasinger g. , et al .
, 2004 , apjs , 155 , 271 tacconi l. j. , et al . , 2006 , apj , 640 , 228 trentham , n. , kormendy , j. , & sanders , d. b. 1999 , , 117 , 2152 van dokkum p.g . , et al . , 2006 ,
apj , 638 , l59 vanzella e. , et al . , 2005 , a&a , 434 , 53 vanzella e. , et al . , 2006 ,
a&a , 454 , 423 wirth g. d. , et al . , 2004 ,
aj , 127 , 3121 yun m. s. , reddy n. a. , condon j. j. , 2001 , apj , 554 , 803 | examining a sample of massive galaxies at @xmath0 with @xmath1 from the great observatories origins deep survey , we compare photometry from spitzer at mid- and far - ir , to submillimeter , radio and rest - frame ultraviolet wavelengths , to test the agreement between different tracers of star formation rates ( sfrs ) and to explore the implications for galaxy assembly . for @xmath2 galaxies with moderate luminosities ( @xmath3 ) ,
we find that the sfr can be estimated consistently from the multiwavelength data based on local luminosity correlations . however , 2030% of massive galaxies , and nearly all those with @xmath4 , show a mid - ir excess which is likely due to the presence of obscured active nuclei , as shown in a companion paper .
there is a tight and roughly linear correlation between stellar mass and sfr for 4mu - detected galaxies . for a given mass ,
the sfr at @xmath5 was larger by a factor of @xmath6 and @xmath7 relative to that in star forming galaxies at @xmath8 and @xmath9 , respectively .
typical ultraluminous infrared galaxies ( ulirgs ) at @xmath5 are relatively transparent to ultraviolet light , and their activity is long lived ( @xmath10 myr ) , unlike that in local ulirgs and high redshift submillimeter - selected galaxies .
ulirgs are the common mode of star formation in massive galaxies at @xmath5 , and the high duty cycle suggests that major mergers are not the dominant trigger for this activity .
current galaxy formation models underpredict the normalization of the mass - sfr correlation by about a factor of 4 , and the space density of ulirgs by an order of magnitude , but give better agreement for @xmath11 quiescent galaxies . |
the production of heavy @xmath1 quarkonia has the potential of offering valuable insights into qcd dynamics , complementary to those given by open heavy flavor production . in both cases
, the creation of the heavy quark pair requires an initial parton collision of hardness @xmath2 .
most of the time the heavy quarks hadronize independently of each other and are incorporated into separate hadrons .
the qcd factorization theorem exploits the conservation of probability in the hadronization process to express the total heavy quark production cross section in terms of target and projectile parton distributions and a perturbative subprocess cross section such as @xmath3 .
the quarkonium cross section is a small fraction of the open flavor one and is thus not constrained by the standard qcd factorization theorems @xcite .
nevertheless , it is plausible that the initial creation of the heavy quarks is governed by the usual parton distributions and hard subprocess cross sections , with the invariant mass of the @xmath1 pair constrained to be close to threshold . before the quarkonium emerges in the final state there can , however , be further interactions which , due to the relatively low binding energy , can either `` make or break '' the bound state .
quarkonium studies can thus give new information about the environment of hard production , from the creation of the heavy quark pair until its `` freeze - out '' .
the quantum numbers of the quarkonium state furthermore impose restrictions on its interactions .
thus states with negative charge conjugation ( @xmath4 , @xmath5 ) or total spin @xmath6 ( @xmath7 ) require the @xmath1 pair to interact at least once after its creation via @xmath8 .
despite an impressive amount of data on the production of several quarkonium states with a variety of beams and targets we still have a poor understanding of the underlying qcd dynamics .
thus quarkonia can not yet live up to their potential as ` thermometers ' of @xmath9 collisions , where @xmath10 , hadron or nucleus .
rather , it appears that we need simultaneous studies and comparisons of several processes to gain insight into the production dynamics .
we will now summarize the successes and failures of the color singlet model @xcite , which we consider as a guideline for understanding the nature of the quarkonium production dynamics .
in the color singlet model ( csm ) , the @xmath1 pair is directly prepared with the proper quantum numbers in the initial hard subprocess and further interactions are assumed to be absent .
the quarkonium production amplitude is then given by the overlap of the non - relativistic wave function with that of the @xmath1 pair .
this model at nlo correctly predicts the normalization and momentum dependence of the @xmath4 photoproduction rate @xcite .
while the absolute normalization of the csm prediction is uncertain by a factor of @xmath11 there appears to be no need for any additional production mechanism , for longitudinal momentum fractions @xmath12 and @xmath13 .
the comparison with leptoproduction data @xcite is less conclusive since only lo csm calculations exist .
the csm underestimates the directly produced @xmath4 and @xmath5 hadroproduction rates by more than an order of magnitude .
this is true both at low @xmath14 ( fixed target ) @xcite and at high @xmath15 ( collider ) @xcite data .
similar discrepancies for the @xmath16 states @xcite indicate that the anomalous enhancement does not decrease quickly with increasing quark mass .
the inelastic cross section ratio @xmath17 is similar in photoproduction @xcite and hadroproduction @xcite and consistent with the value @xmath18 expected in the csm @xcite .
the ratio does not depend on @xmath19 in the projectile fragmentation region and is independent of the nuclear target size in @xmath20 collisions .
the csm thus underestimates the @xmath4 and @xmath5 hadroproduction cross sections , as well as that of the @xmath7 @xcite , by similar large factors .
the quantum numbers of these charmonium states require final - state gluon emission in the csm , @xmath21
. this emission is not required for the @xmath22 where the csm cross section @xmath23 is only a factor @xmath24 below the hadroproduction data @xcite . in the csm , @xmath25 photoproduction is suppressed by a power of @xmath26 compared to the @xmath27 and @xmath5 production rates .
one indeed observes a smaller value of the @xmath28 ratio in photoproduction @xcite than in hadroproduction @xcite .
the analysis of agreements and discrepancies between the csm and quarkonium data led to the comover enhancement scenario of quarkonium production @xcite .
hadroproduced @xmath1 pairs are created within a comoving color field and form @xmath4 , @xmath5 and @xmath7 through gluon _ absorption _ rather than emission , enhancing the cross section relative to the csm since the pair gains rather than loses energy and momentum .
the @xmath22 cross section is not as strongly influenced since no gluon needs to be absorbed or emitted .
most importantly , such a mechanism is consistent with the success of the csm in photoproduction since no color fields are expected in the photon fragmentation region , @xmath29 . in qed ; ( b ) hadroproduction of a @xmath30 pair , _ e.g. _ @xmath31 ; and ( c ) photoproduction of a @xmath30 pair , @xmath32 . the creation of a comoving color field is specific to hadroproduction , ( b).,width=226 ] the effect of a comoving color field is illustrated in fig .
[ comofield ] .
light charged particles carry gauge fields which are radiated in high energy annihilations into a heavy particle pair . in @xmath33 annihilations ,
the photon fields pass through each other and materialize as forward bremsstrahlung , fig .
[ comofield](a ) . in @xmath0 , on the other hand , the self - interaction of the color field can also result in the creation of a gluon field at intermediate rapidities , fig .
[ comofield](b ) . hadroproduced @xmath1 pairs
thus find themselves surrounded by a color field .
we postulate that interactions between the @xmath1 pair and this comoving field are important in quarkonium hadroproduction . in direct photoproduction
, the incoming photon does not carry any color field and the @xmath1 pair is left in a field - free environment after the collision , fig .
[ comofield](c ) .
the proposed rescattering thus does not affect photoproduction .
the importance of rescattering effects in hadroproduction as compared to photoproduction is also suggested by data on open charm production . hadroproduced @xmath34 pairs are almost uncorrelated in azimuthal angle @xcite , at odds with standard qcd descriptions .
photoproduced pairs on the other hand , emerge nearly back - to - back @xcite , following the charm quarks of the underlying @xmath35 process .
since the hardness of the gluons radiated in the creation process increases with quark mass , the rescattering effect persists for bottomonium . due to the short timescale of the radiation
the heavy quark pair remains in a compact configuration during rescattering and overlaps with the quarkonium wave function at the origin .
the successful csm result for @xmath17 @xcite is thus preserved .
the @xmath1 pair may also interact with the more distant projectile spectators after it has expanded and formed quarkonium .
such spectator interactions are more frequent for nuclear projectiles and can cause the breakup ( absorption ) of the bound state .
this conventional mechanism of quarkonium suppression in nuclei is thus fully compatible with , but distinct from , interactions with the comoving color field .
we have investigated the consequences of the ces using pqcd to calculate the interaction between the @xmath1 and the comoving field . while we find consistency with data , quantitative predictions depend on the structure of the comoving field .
hence tests of the ces must rely on its generic features which we discuss below .
the ces distinguishes three proper timescales in quarkonium production : * @xmath36 , the @xmath1 pair production time ; * @xmath37 , the dglap scale over which the comoving field is created and interacts with the @xmath1 pair ; * @xmath38 , while rescattering with comoving spectators may occur . in the following we will consider quarkonium production at @xmath39 . in quarkonium production at @xmath40 , a large @xmath41 parton is first created on a timescale @xmath42 , typically through @xmath43 .
the virtual gluon then fragments , @xmath44 , in proper time @xmath45 . thus high @xmath41 quarkonium production
is also based on the ces @xcite .
the @xmath1 pair is created in a standard parton subprocess , typically @xmath0 , at a time scale @xmath36 .
this first stage is common to other theoretical approaches such as the color evaporation model ( cem ) @xcite and non - relativistic qcd @xcite , here referred to as the color octet model ( com ) .
the momentum distribution of the @xmath1 is determined by the product of projectile @xmath46 and target @xmath47 parton distributions , such as @xmath48 where @xmath49 may be a lepton , photon , hadron or nucleus .
this production process is consistent with the quarkonium data . according to pqcd the @xmath1 is dominantly produced in a color octet , @xmath50 configuration , close to threshold .
such a state can obtain the quarkonium quantum numbers through a further interaction which flips a heavy quark spin and turns the pair into a color singlet .
the amplitude for processes of this type are suppressed by the factor @xmath51 where @xmath52 is the momentum scale of the interaction .
the various theoretical approaches differ in the scale assumed for @xmath52 .
csm : here @xmath53 .
thus @xmath4 production proceeds via the emission of a hard gluon in the primary process , @xmath54 .
the @xmath22 is produced without gluon emission , @xmath55 , through a subdominant @xmath56 color singlet production amplitude .
com : the @xmath1 quantum numbers are changed via gluon emission at the bound state momentum scale @xmath57 .
this corresponds to an expansion in powers of the bound state velocity @xmath58 , introducing nonperturbative matrix elements that are fit to data .
cem : here @xmath59 .
soft interactions are postulated to change the @xmath1 quantum numbers with probabilities that are specific for each quarkonium state but independent of kinematics , projectile and target .
ces : the quantum numbers of the @xmath1 are changed in perturbative interactions with a comoving field at scale @xmath60 , as described below .
the scale @xmath37 refers to the time in which collinear bremsstrahlung , the source of qcd scaling violations , is emitted in the heavy quark creation process @xcite .
thus @xmath61 characterizes the effective hardness of logarithmic integrals of the type @xmath62 where @xmath63 is the factorization scale .
we stress that @xmath61 is an intermediate but still perturbative scale , @xmath64 , which grows with @xmath65 .
the fact that the @xmath1 pair acquires the quarkonium quantum numbers over the perturbative timescale @xmath37 is a feature of the ces and distinguishes it from other approaches . at this time , the pair is still compact and couples to quarkonia via the bound state wavefunction at the origin or its derivative(s ) . thus no new parameters are introduced in this transition . however , the interactions of the @xmath1 pair depend on the properties of the comoving color field such as the intensity and polarization
. quantitative predictions in the ces are only possible when the dependence on the comoving field is weak
. ratios of radially excited quarkonia , such as @xmath17 , are insensitive to the comoving field and are thus expected to be process - independent when absorption on spectators at later times can be ignored , see below .
the fact that this ratio is observed to be roughly universal @xcite is one of the main motivations for the ces .
even the measured variations of the ratio in different reactions agree with expectations , see ref .
@xcite for a discussion of its systematics in elastic and inelastic photoproduction , leptoproduction and hadroproduction at low and high @xmath41 .
the ratio @xmath66 is measured to be @xmath67 in pion - induced @xcite and @xmath68 in proton - induced @xcite reactions .
the csm underestimates this ratio @xcite as well as that of @xmath69 @xcite .
the rescattering contribution increases @xmath70 and @xmath71 , enhancing the above ratios .
the quarkonium cross section can be influenced by rescattering effects in both the target and projectile fragmentation regions . for definiteness
, we assume the charmonium is produced in the projectile fragmentation region , @xmath72 . the nuclear target dependence is usually parametrized as @xmath73 .
it turns out that @xmath74 and obeys feynman scaling , _ ie .
_ , @xmath75 decreases with @xmath19 rather than with the momentum fraction @xmath76 of the target parton @xcite .
the comparison with lepton pair production in the drell - yan process shows that the @xmath4 nuclear suppression can not be attributed to shadowing of parton distributions in the nucleus @xcite .
the @xmath77-dependence is thus difficult to explain in the csm , com and cem approaches . in the feynman scaling regime
, we may assume that the @xmath1 pair energy is high enough to remain compact while traversing the target .
the relative transverse momentum of the @xmath78 and @xmath79 could increase as a result of rescattering in the target , thus suppressing the binding probability .
however , this explanation is unlikely in view of the absence of nuclear suppression in photoproduction @xcite . in the ces ,
the nuclear target suppression is ascribable to absorption of the comoving color field in the target nucleus .
this field is emitted by a projectile parton with transverse size @xmath37 , larger than the size , @xmath80 , of the @xmath1 pair .
due to lorentz time dilation , the field is emitted long before reaching the target and reinteracts with the @xmath1 long after passing the target .
absorption of the comoving field in the target implies suppression of @xmath4 production in the ces . at high energies
, we have @xmath81 , which explains the feynman scaling of this effect .
moreover , as @xmath19 increases , less energy is available to be radiated to the gluon field which therefore becomes softer , further increasing its absorption in the target and thus the nuclear suppression .
this explanation is consistent with the fact that the nuclear target suppression of the @xmath4 and the @xmath5 is found to be the same for @xmath82 @xcite .
it also predicts that the suppression will be similar for @xmath7 production . on the other hand ,
the nuclear target suppression should be _ reduced _ for @xmath22 since a substantial fraction is directly produced without a comoving field .
a measurement of @xmath83 in the projectile fragmentation region would thus constitute an important test of the ces . in a glauber picture of the nuclear suppression ,
a relatively large value for the absorption cross section is required , @xmath84 .
we interpret this value as the joint cross section of the @xmath1 pair and the comoving field , thus of order @xmath85 .
since @xmath61 scales with @xmath65 , we expect less nuclear absorbtion for the @xmath16 states than for the @xmath86 , as observed experimentally @xcite . by the time , @xmath87 fm , that the @xmath1 pair encounters comoving projectile spectators , the pair has already expanded and is distributed according to the quarkonium wave function .
the spectator rescattering effects are thus independent of the quarkonium formation process .
larger and more loosely bound charmonia are more easily broken up by secondary scattering .
hence the @xmath25 and @xmath5 cross sections should be depleted compared to that of the @xmath4 .
likewise bottomonium is generally less affected by spectator interactions than charmonium .
spectator interactions at large enough @xmath19 are likely to be unimportant for hadron projectiles judging from the approximate universality of the @xmath17 ratio in photo- and hadroproduction .
the lower ratio seen in nucleus - nucleus collisions @xcite , on the other hand , is most naturally explained by absorption on spectators .
it would be important to confirm this by also measuring the ratio in @xmath88 scattering .
data on quarkonium production have proved challenging for qcd models .
the richness of the observed phenomena indicates that quarkonium cross sections are indeed sensitive to the environment of the hard qcd scattering .
we may , however , only be able to decipher its message through systematic experimental and theoretical studies of several species of quarkonia produced with a variety of beams and targets in a range of kinematic conditions .
it is thus useful to begin by asking fairly general questions about the production mechanism before fitting detailed models to data .
this is the spirit of the comover enhancement scenario that we have discussed here , motivated by apparent regularities of the data and general theoretical constraints .
s. p. would like to thank cristina volpe for very stimulating discussions .
we also thank ramona vogt for her careful reading of our manuscript . during this work ,
n. m. and s. p. were supported by the cnrs grant number 11240 within the 2002 cnrs / snf cooperation program .
h. was employed by nordita , copenhagen until 31 july 2002 .
this work was completed during a visit to nordita , and has been partially supported also by the european commission under contract hprn - ct-2000 - 00130 .
j. c. collins , d. e. soper and g. sterman , adv .
high energy phys . * 5 * ( 1988 ) 1 .
e. l. berger and d. l. jones , phys . rev .
d * 23 * ( 1981 ) 1521 .
r. baier and r. ruckl , phys .
b * 102 * ( 1981 ) 364 .
m. kramer , prog .
* 47 * ( 2001 ) 141 [ arxiv : hep - ph/0106120 ] . c. adloff
_ et al . _
[ h1 collaboration ] , eur .
j. c * 25 * ( 2002 ) 25 [ arxiv : hep - ex/0205064 ] . c. adloff
_ et al . _
[ h1 collaboration ] , eur .
j. c * 25 * ( 2002 ) 41 [ arxiv : hep - ex/0205065 ] .
d. m. kaplan [ e789 collaboration ] , int .
j. mod .
a * 12 * ( 1997 ) 3827 [ arxiv : hep - ex/9610003 ] .
f. abe _ et al . _
[ cdf collaboration ] , phys .
* 75 * ( 1995 ) 4358 .
t. alexopoulos _ et al . _
[ e771 collaboration ] , phys .
b * 374 * ( 1996 ) 271 .
t. affolder _ et al . _
[ cdf collaboration ] , phys .
* 84 * ( 2000 ) 2094 [ arxiv : hep - ex/9910025 ] .
a. bertolin [ h1 and zeus collaborations ] , _ prepared for international europhysics conference on high - energy physics ( eps - hep 99 ) , tampere , finland , 15 - 21 jul 1999_. published in ` tampere 1999 , high energy physics ' 580 - 582 .
t. alexopoulos _ et al . _
[ e-771 collaboration ] , phys .
d * 55 * ( 1997 ) 3927 . c. lourenco , nucl .
a * 610 * ( 1996 ) 552c [ arxiv : hep - ph/9612222 ] .
m. vnttinen , p. hoyer , s. j. brodsky and w. k. tang , phys .
d * 51 * ( 1995 ) 3332 [ arxiv : hep - ph/9410237 ] .
p. roudeau [ na14 collaboration ] , nucl .
. suppl .
* 7b * ( 1989 ) 273 .
l. antoniazzi _ et al .
_ [ e705 collaboration ] , phys . rev .
* 70 * ( 1993 ) 383 .
l. antoniazzi _ et al . _
[ e705 collaboration ] , phys .
d * 49 * ( 1994 ) 543 .
p. hoyer and s. peign , phys .
d * 59 * ( 1999 ) 034011 [ arxiv : hep - ph/9806424 ] .
p. hoyer , nucl .
suppl . * 75b * ( 1999 ) 153 [ arxiv : hep - ph/9809362 ] .
n. marchal , s. peign and p. hoyer , phys .
d * 62 * ( 2000 ) 114001 [ arxiv : hep - ph/0004234 ] .
e. m. aitala _ et al .
_ [ e791 collaboration ] , eur .
j. c * 4 * ( 1999 ) 1 [ arxiv : hep - ex/9809029 ] .
l. frabetti _ et al . _
[ e687 collaboration ] , phys .
b * 370 * ( 1996 ) 222 .
h. fritzsch , phys .
b * 67 * ( 1977 ) 217 .
f. halzen , phys .
b * 69 * ( 1977 ) 105 .
m. gluck , j. f. owens and e. reya , phys .
d * 17 * ( 1978 ) 2324 .
g. t. bodwin , e. braaten and g. p. lepage , phys .
d * 51 * ( 1995 ) 1125 [ erratum - ibid .
d * 55 * ( 1997 ) 5853 ] [ arxiv : hep - ph/9407339 ] .
m. beneke and i. z. rothstein , phys .
d * 54 * ( 1996 ) 2005 [ erratum - ibid .
d * 54 * ( 1996 ) 7082 ] [ arxiv : hep - ph/9603400 ] .
p. hoyer and s. peign , phys .
d * 61 * ( 2000 ) 031501 .
v. koreshev _ et al . _
[ e672-e706 collaborations ] , phys . rev .
* 77 * ( 1996 ) 4294 .
t. alexopoulos _ et al . _
[ e771 collaboration ] , phys .
d * 62 * ( 2000 ) 032006 [ arxiv : hep - ex/9908010 ] .
p. hoyer , m. vnttinen and u. sukhatme , phys .
b * 246 * ( 1990 ) 217 .
m. j. leitch _ et al . _
[ fnal e866/nusea collaboration ] , phys .
* 84 * ( 2000 ) 3256 [ arxiv : nucl - ex/9909007 ] .
p. amaudruz _ et al .
_ [ new muon collaboration ] , nucl .
b * 371 * ( 1992 ) 553 .
d. m. alde _ et al .
_ , phys .
* 66 * ( 1991 ) 2285 .
m. j. leitch _ et al . _
[ e772 and e789 collaboration ] , nucl . phys .
a * 544 * ( 1992 ) 197c . | quarkonium data suggest an _ enhancement _ of the hadroproduction rate from interactions of the heavy quark pair with a comoving color field generated in the hard @xmath0 subprocess .
we review the motivations and principal consequences of this comover enhancement scenario ( ces ) . |
this model is essentially equivalent to the problem of distributing @xmath31 coins among @xmath21 boxes .
the total area is @xmath21 , each box has area 1 , and @xmath33 . from eq .
( [ eq:12 ] ) and eq .
( [ eq:13 ] ) one immediately obtains @xmath34 the second moment @xmath35 of @xmath24 can also be computed and we obtain for @xmath36 the following equation , @xmath37 note that as @xmath38 , @xmath39 and @xmath40 .
this means that when the coin density is very low , the clusters have just one coin and the distribution is a @xmath41-function ( @xmath42 ) . in the @xmath43 limit , @xmath44 ,
increases with @xmath22 .
the behavior of the quantity of interest to us , @xmath19 , is show , in the case of model i , in fig.(1 ) .
in the present context percolation means that the average area of the cluster can not be constant : it must increase with density . in the low density limit , we can accept that model i is correct , i.e. , @xmath45 however , in the @xmath43 limit , percolation requires @xmath46 or , @xmath47 . in this case , @xmath24 is dominated by big size clusters .
percolation implies that we have a @xmath41-function distribution , not only in the @xmath38 limit , @xmath48 , but a @xmath41-function distribution , as well , in the @xmath49 limit , @xmath50 .
+ in order to satisfy eq .
( [ eq:16 ] ) and eq .
( [ eq:17 ] ) we choose for @xmath27 , as a function of @xmath22 , the function where @xmath52 represents a percolation function , @xmath53 means no percolation , @xmath33 , @xmath54 means full percolation , with the heavyside function approximated by a smoother function , because of the finite size of the systems in simulation,(see last paper in ref .
@xcite ) @xmath55 with this function we fit the fraction of occupied area by all string in the transverse plane and we obtain for @xmath56 and @xmath57 this parameter is the percolation threshold . + in fig.(2 ) we show @xmath27 as a function , of @xmath22 for model i and ii . from eqs .
( [ eq:12 ] ) , ( [ eq:13 ] ) and ( [ eq:18 ] ) one immediately obtains @xmath58 + 1 ) \label{eq:19}\ ] ] with limits @xmath39 as @xmath38 , as in eq . , and
@xmath59 , as @xmath43 ( the average cluster , asymptotically , contains all the strings ) . + in order to estimate the second moment @xmath35 , or the variance , one has to keep in mind that : @xmath60 @xmath61 @xmath62 by making use of eqs .
( [ eq:20 ] ) we write @xmath63 } } { 1-e^{-\eta}}\,\ , { \langle n \rangle}^2\;. \label{eq:21}\ ] ] it is easily seen that eq .
( [ eq:21 ] ) satisfies the constraints of eqs .
( [ eq:20 ] ) . + in fig.(1 )
we also show the quantity @xmath19 for model ii .
it is clear what is the essential difference between model i ( no percolation ) and model ii ( percolation ) . while in model
i , @xmath64 always increases with @xmath22 , thus originating increasing @xmath0 fluctuations , in model ii , as @xmath43 , @xmath64 starts to decrease , thus originating decreasing @xmath0 fluctuations . before comparing models i and ii with data , by making use of eq .
( [ eq:8 ] ) , we would like to give estimates for the values of @xmath21 and @xmath65 .
+ in string percolation the variable @xmath22 , eq .
( [ eq:11 ] ) , is defined as @xmath66 where @xmath67 is the transverse radius of the string ( we assume @xmath68 fm ) and @xmath69 is the average interaction radius .
it is natural to identify @xmath21 , which has the meaning of an area , with @xmath70 .
we then have @xmath71 where @xmath72 is the radius of the nucleus with , for gold , @xmath73 fm .
we thus obtain the bounds @xmath74 where the left hand side represents the limit of pp collision . concerning the parameter @xmath65 , eq .
( [ eq:10 ] ) , the quantity @xmath75 , if we admit , as in schwinger model , that the particle @xmath76 distribution of the string is gaussian . the quantity @xmath8 , the string particle density , was in ref .
@xcite estimated to be @xmath77 , for valence strings , and @xmath78 times smaller for sea strings .
we obtain the bounds , @xmath79 note that @xmath65 should be a decreasing function of energy , as the fraction of sea strings increases with energy .
+ the number of produced strings is obtained from the number of participants by making use of the approximate usual formula @xmath80 where @xmath81 is the average number of strings in pp collision , @xmath81 is assumed to be equal @xmath82 at @xmath83 gev , which agree with ref .
@xcite , and @xmath84 .
+ in fig.(3 ) we compare our models i and ii with data published in ref .
. the parameters @xmath21 and @xmath65 were , within bounds , inequalities eq .
( [ eq:24 ] ) and eq .
( [ eq:25 ] ) , adjusted separatly for each one of the models .
+ in conclusion , data clearly favour model ii , with percolation .
+ we would like to thank elena ferreiro , carlos pajares and roberto ugoccioni for advice and discussions .
m. tannenbaum , phys .
b * 498 * , 24 ( 2001 ) ; + h. appelshauser _
et al . _ ,
na49 collaboration , phys .
b * 459 * , 679 ( 1999 ) ; + h. appelshauser _
et al . _ ,
ceres collaboration , nucl .
a * 698 * , 253c ( 2002 ) .
n. armesto , m. a. braun , e. g. ferreiro and c. pajares , phys .
* 77 * , 3736 ( 1996 ) ; + m. nardi and h. satz , phys .
b * 442 * , 14 ( 1998 ) ; + j. dias de deus , r. ugoccioni and a. rodrigues , eur .
j. c * 16 * , 537 ( 1996 ) . | transverse momentum fluctuations can be understood as resulting from clustering of strings or partons .
data allows to distinguish clustering without percolation , from clustering with percolation .
percolation is clearly favored by data .
event by event fluctuations of total transverse momentum , @xmath0 , in high energy hadronic and heavy ion collisions , may give relevant information on the thermodynamical properties of dense hadronic matter , and , in particular , on the existence of a phase transition @xcite .
the problem has been discussed in a series of experimental papers @xcite .
+ in this paper we argue that the occurrence of @xmath0 fluctuations is not necessarily a signal of a thermal phase transition .
it may rather be a geometrical effect , due to clustering and percolation of strings @xcite .
+ let us suppose that in hadron - hadron and nucleus - nucleus collisions one always produces a cluster of @xmath1 independent identical strings . for the @xmath0 distribution one can then write @xmath2 from eq .
( [ eq:1 ] ) it follows that @xmath3 and @xmath4 such that @xmath5 , \label{eq:4}\ ] ] as expected for independent strings .
+ however , if one considers fluctuations in the number @xmath1 of strings per cluster , one obtains @xmath6{\langle p^2 \rangle}_{1}+ { \langle n \rangle}\left [ { \langle p^2 \rangle}_{1}-{\langle p \rangle}_{1}^{2}\right ] \label{eq:5}\ ] ] note that if one assumes that particles from the single string are emitted independently ( poisson like distribution ) , one arrives at : @xmath7 where @xmath8 is the string average multiplicity and @xmath9 refers to the particle transverse momentum . + introducing now the quantity @xmath10 ( see , for instance , refs .
@xcite ) , @xmath11 with @xmath12 , @xmath13 and @xmath14 , such that @xmath15 for random emission , we obtain @xmath16 where @xmath17 is the normalized variance of the cluster distribution , and @xmath18 is characteristic of the string particle production mechanism .
the inclusion of color summation factors when forming clusters of strings , not considered here , may affect the value of c , eq .
( [ eq:10 ] ) . in any case , physics is mostly determined by @xmath19 .
+ experimentally , data on @xmath10 , ref .
@xcite , show that this quantity increases as matter density ( number of participant nucleons ) increases up to some point starting to decreases for larger densities . making use of two very simple string clustering models ,
model i without percolation and model ii with percolation , we show that only model ii , with percolation , is consistent with data .
+ let us consider the ( two dimensional ) problem of distributing @xmath20 strings with a definite transverse area , in a total area @xmath21 ( the impact parameter transverse interaction area ) .
a relevant parameter is the transverse area density @xmath22 : @xmath23 + let us now assume that in a collision clusters of strings are formed , @xmath24 being the probability of having a @xmath1-cluster(cluster with @xmath1 strings ) .
+ in principle - by using monte carlo simulations - it is possible to calculate @xmath25 , the average number of clusters , @xmath26 the average number of strings per cluster , higher moments of @xmath24 , and the average area occupied by the clusters @xmath27 .
+ there are two sum - rules which have , in general , to be satisfied : @xmath28@xmath29the first sum - rule expresses the conservation of number of strings and the second is simply the area sum - rule : the right hand side of this sum - rule is the overall area occupied by the clusters , refs .
@xcite . note that eq .
( [ eq:12 ] ) and eq .
( [ eq:13 ] ) are invariants , independent of percolation .
percolation affects @xmath27 , @xmath26 and @xmath30 , but not @xmath31 and @xmath32 .
+ in the next two sections we introduce the two models that illustrate our arguments . |
a _ real structure _ on an oriented @xmath0-manifold ( respectively @xmath1-manifold ) @xmath2 with boundary ( possibly empty ) is defined as an involution @xmath3 on @xmath2 which is orientation preserving if @xmath4 is even and orientation reversing if @xmath4 is odd and the fixed point set of which is of dimension @xmath4 ( respectively @xmath5 ) , if it is not empty .
hence , if @xmath6 is the oriented boundary of an oriented @xmath0-manifold @xmath2 , a real structure on @xmath2 restricts to a real structure on @xmath6 .
we call a manifold together with a real structure a _ real manifold _ and the fixed point set of the real structure the _ real part_. if a symplectic structure ( or a contact structure ) pulls back to minus itself under a real structure @xmath7 , such structures are called @xmath7-real and those manifolds are called @xmath7-real symplectic ( or @xmath7-real contact respectively ) manifolds . real algebraic varieties and links of real algebraic isolated singularities supplied with natural structures
are examples for real symplectic and contact manifolds .
every closed oriented 3-manifold admits an open book decomposition @xcite and a positive contact structure @xcite .
furthermore the work of e. giroux points out that there is a one to one correspondence between open book decompositions up to positive stabilization and positive contact structures up to isotopy @xcite ( see e.g. @xcite or @xcite for a careful discussion and for the definitions of open book decompositions and contact structures ) .
however not every 3-manifold admits a real structure ; in fact the ones which do not admit a real structure are in abundance @xcite .
nevertheless , once a real structure on a 3-manifold is given we can make appropriate definitions similar to the ones above and show that there are analogous interrelations .
for instance , a _
real open book decomposition _ is an open book with the real structure preserving the page structure and leaving exactly two pages invariant .
a _ @xmath8-real ( or characteristic ) heegaard splitting _ is a @xmath9-invariant heegaard surface with the two handlebodies being mapped to each other by @xmath9 .
it is known that every real 3-manifold @xmath10 admits a characteristic heegaard splitting ( * ? ? ?
* proposition 2.4 ) . in section [ rak ]
we present examples of real open books and real heegaard decompositions .
we state some observations related to real heegaard decompositions and real open books on lens spaces in section [ mercek ] .
we describe some real lens spaces which can not have a real heegaard splitting with genus 1 or 2 .
we also observe that there are ( infinitely many ) closed oriented real 3-manifolds which can not have a _ maximal _ real heegaard decomposition ; in a maximal real heegaard decomposition , the real part of the real structure has exactly genus@xmath11 connected components ( proposition [ maksimal ] ) .
a _ positive real stabilization _ of a real open book is performed by attaching a 2-dimensional 1-handle to the page of an abstract open book and modifying the real structure and monodromy appropriately ( see section [ rstab ] ) .
this boils down to taking connected sum with a real tight contact @xmath12 , which is unique up to equivariant contact isotopy @xcite . with this definition ,
we prove a giroux correspondence in the existence of a real structure .
we show that there is a one to one correspondence between the @xmath9-real contact structures on a real 3-manifold @xmath10 up to equivariant contact isotopy and the @xmath9-real open books on @xmath6 up to positive real stabilization and equivariant isotopy ( theorem [ jiru ] ) . in the sequel , instead of using the term @xmath7-real
, we usually drop the reference to @xmath7 whenever the real structure is understood
. * acknowledgements .
* the second author would like to thank e. giroux for generously sharing his deep knowledge on the subject ; she would also like to thank j.y .
welschinger and a. degtyarev for fruitful discussions .
the project leading to this article is supported by the scientific and technological research council of turkey [ tubitak - ardeb-109t671 ] .
the second author is supported by the european community s seventh framework programme ( [ fp7/2007 - 2013 ] [ fp7/2007 - 2011 ] ) under grant agreement no [ 258204 ] .
let @xmath13 be an open book decomposition of a real 3-manifold @xmath14 .
we say that @xmath13 is a _ real open book decomposition _ , ( or shortly , a _ real open book _ ) if @xmath15 where @xmath16 is a reflection .
in particular , we have @xmath17 . an _ isomorphism _ between two real open book decompositions of @xmath10 is a pair of orientation preserving equivariant diffeomorphisms of @xmath10 and @xmath18 commuting with the projections . as orientation reversing reflections on @xmath19 form a single class , we fix , once for all , @xmath20 pour @xmath21 . by definition ,
any real open book decomposition of @xmath14 has two pages , @xmath22 , which are invariant under the action of the real structure @xmath9 .
the restrictions @xmath23 yield real structures on @xmath24 and , respectively , @xmath25 .
these pages together with the inherited real structures are called _
real pages_. one of the fundamental property of the monodromy @xmath26 of a real open book is that @xmath27 .
or equivalently , @xmath28 as well as @xmath29 .
( related to the definition above and the argumentation in the sequel , see @xcite . ) therefore , a real open book decomposition of @xmath14 gives rise to a triple @xmath30 where @xmath31 is one of the real pages , @xmath7 is the inherited real structure on @xmath31 and @xmath26 is the monodromy satisfying @xmath32id and @xmath33 . by definition ,
@xmath7 acts on @xmath34 , which is the disjoint union of finitely many circles .
hence either @xmath7 acts on a connected component of @xmath34 as reflection or @xmath7 swaps a pair of connected components , reversing the induced orientations as boundary . [ yayilanc ]
let @xmath35 be as above .
then the real structure @xmath36 extends to a real structure on @xmath6 .
_ proof : _ here we explain how a real structure on a page extends to a real structure on @xmath37 , a 3-manifold described by the abstract open book @xmath38 .
( as @xmath37 is diffeomorphic to @xmath6 , the real structure obtained on @xmath37 can be pulled to @xmath6 by means of a chosen diffeomorphism between @xmath37 and @xmath6 . )
first we describe the real structure on the mapping torus . instead of the usual description ,
consider the mapping torus as @xmath39 for all @xmath40 in @xmath31 ( see figure [ boru ] ) . here
@xmath41 $ ] , @xmath42 $ ] .
note that the manifold constructed has monodromy @xmath43 .
now consider the map @xmath44 which acts as identity between the cylinders , i.e. @xmath45 in @xmath46 ( resp .
@xmath47 ) is sent to @xmath48 ( resp .
@xmath49 ) . after identification @xmath50 induced a reflection on @xmath51 . restricted to each torus component of @xmath52 ,
the map @xmath53 is either a rotation by @xmath54 fixing exactly 4 points or it interchanges two tori components .
the various ways to extend @xmath53 as a real structure over the solid torus neighborhoods of binding components are dictated by the behavior of @xmath7 on @xmath34 .
either a pair of solid tori are mapped to each other by @xmath53 or @xmath53 extends over a solid torus as rotation by @xmath54 . in each case
the extension is unique up to isotopy ( see e.g. ( * ? ? ?
* lemma 4.4 ) ) and the extended map preserves the page structure .
therefore , we obtain a real structure on @xmath37 , described by @xmath55 , preserving the page structure . @xmath56 + thus , we have the following definition .
let @xmath38 be an abstract open book , where @xmath31 is a compact surface with boundary and @xmath57 is the monodromy so that @xmath26 is the identity on @xmath34 .
suppose that @xmath7 is a real structure on @xmath31 , i.e. an orientation reversing involution .
an _ abstract real open book _ is a triple @xmath55 with @xmath26 satisfying @xmath58 .
an _ isomorphism _ of abstract open books is an orientation preserving diffeomorphism of the surface commuting both with the monodromy and with the real structure .
note after the proof of lemma 1 that whenever we represent a real manifold with an embedded open book as an real abstract open book , we must allow an equivariant isotopy .
note also that once we adopt the model in figure [ boru ] so that @xmath7 is placed on the _ left _ , an ( embedded ) real open book determines an abstract real open book uniquely up to the equivalence above .
recall that a positive stabilization of an abstract open book @xmath38 is the abstract open book with page @xmath59 a 1-handle and with monodromy @xmath60 where @xmath61 is the extension of @xmath26 over the 1-handle trivially and @xmath62 is a right - handed dehn twist along a curve @xmath63 in @xmath64 that intersects the co - core of the 1-handle exactly once .
the type of stabilizations that we need should lead a real structure on the new abstract open book . here is the definition that will provide .
let @xmath55 be an abstract real open book .
let @xmath65 where * either @xmath66 is a 1-handle with its attaching region @xmath7-invariant , ( in particular if the attaching region is neighborhoods of a pair of real points , then we impose the condition that the real points belong to the same real component ) ; * or @xmath67 where @xmath68 and @xmath69 are 1-handles with their attaching regions interchanged by @xmath7 . in such cases
, @xmath7 extends uniquely over @xmath66 to a real structure , say , @xmath70 on @xmath64 .
let @xmath61 denote the extension of @xmath26 over @xmath64 with @xmath71id .
we consider , in the former case , a simple closed curve @xmath63 such that @xmath72 and that @xmath63 intersects the co - core of @xmath66 once ( existence of such an invariant curve @xmath63 is guaranteed by the imposed condition ) , while in the latter case , a pair of simple closed curves @xmath73 such that @xmath63 and hence @xmath74 intersects the co - core of @xmath75 and , respectively , of @xmath76 once . depending on the succeeding cases ,
let @xmath77 denote either the dehn twist @xmath62 along @xmath63 or the product @xmath78 .
then a _ positive real stabilization _ of the abstract real open book @xmath55 is defined as the open book @xmath79 .
first note that @xmath80 and @xmath81 since @xmath70 is orientation reversing .
then it is easy to see that @xmath82 is really an abstract real open book .
in fact , @xmath83 furthermore the real manifold @xmath84 is equivariantly diffeomorphic to @xmath85 .
let us see this in the case where @xmath66 is a single 1-handle ( the other case is similarly treated ) .
observe that a sufficiently small closed neighborhood @xmath86 of the core @xmath87 of @xmath66 in @xmath88 is a real 3-ball , which is known to be unique up to equivariant isotopy .
the local model for @xmath89 can be taken as depicted in figure [ topmodelleri ] ( see e.g @xcite , section 1 ) .
if @xmath70 fixes @xmath87 pointwise then the model is as in figure [ topmodelleri](a ) or else @xmath70 acts on @xmath87 as reflection and we have the model as in figure [ topmodelleri](b ) . .
] a positive stabilization in the non - abstract sense is nothing but excising a sufficiently small closed neighborhood @xmath89 of @xmath87 and gluing back a model for @xmath86 depending on the behavior of the real structure on @xmath87 .
thus the uniqueness of the model in each case guarantees that the ( equivariant ) identity diffeomorphism from @xmath90 to @xmath91 uniquely extends ( equivariantly ) to a diffeomorphism from @xmath89 to @xmath86 . in figure [ stablar ]
, we depict all possible choices for the attachment(s ) of the 1-handle(s ) and the extension the real structure through the handle(s ) . in the first six cases , the real structure acts as a reflection on the boundary component(s ) of @xmath31 , while in the last three cases
, the real structure swaps two boundary components . a positive real stabilization of type i , ii , v , vii is a connected sum with real @xmath12 equipped with the real open book ( defined by @xmath92 , @xmath93 ) whose binding is the positive hopf link @xmath94 , where @xmath95 .
this open book admits two real structures : @xmath96 and @xmath97 . in the case of @xmath96 ,
the real structure acts on each binding component and the real pages are as depicted at the top of figure [ hopfstablari ] . in the other case , however , the real structure takes one binding circle to the other and acts on the real pages as shown at the bottom of figure [ hopfstablari ] . in @xcite
, we showed that there is a unique tight contact 3-ball ; therefore , as in the non - real case , if there exists a real contact structure _ supported _ ( in the sense of definition [ supported ] ) by a real open book , a positive real stabilization of the open book coincides with the real contact connected sum of the manifold with real @xmath12 .
other types ( iii , iv , vi , viii , ix ) of positive real stabilizations can be seen as connected sums with a pair of @xmath12 exchanged by the real structure . whose binding is the positive hopf link . ] here we present the notion of a real heegaard decomposition which is closely related to a real open book .
let @xmath98 be an oriented hypersurface in a closed oriented manifold @xmath6 .
if the topological closure of @xmath99 is a disjoint union of two handlebodies diffeomorphic to each other , then the pair @xmath100 is called an _ embedded heegaard decomposition_. let @xmath101 be an oriented handlebody with nonempty boundary @xmath66 and @xmath102 be an orientation reversing diffeomorphism .
the pair @xmath103 is called an _ abstract heegaard decomposition _ of the manifold @xmath104 where @xmath105 is a copy of @xmath101 .
it is obvious that an abstract heegaard decomposition determines an embedded one immediately . to see the converse ,
consider an embedded heegaard decomposition @xmath106 .
if @xmath107 is an orientation preserving diffeomorphism such that @xmath108 and @xmath109 is orientation reversing on @xmath98 , then the pair @xmath110 defines an abstract heegaard decomposition .
let @xmath10 be a real manifold and @xmath106 be an embedded heegaard decomposition .
if @xmath111 and @xmath112 is orientation reversing , then we call the triple @xmath113 an _ embedded real heegaard decomposition_. observe that for an embedded real heegaard decomposition @xmath113 with handlebodies @xmath114 ( @xmath115 ) , if @xmath116 is orientation reversing then @xmath117 is an abstract heegaard decomposition for @xmath6 .
it was proven by t. a. nagase that every real manifold admits an embedded real heegaard decomposition @xcite .
recall that an embedded ( respectively abstract ) open book decomposition describes an embedded ( respectively abstract ) heegaard decomposition .
let @xmath118 be an embedded real open book for the real manifold @xmath10 .
denote by @xmath119 and @xmath120 the diffeomorphisms obtained by restricting @xmath9 to the two invariant pages @xmath121 and @xmath25 .
note that @xmath122 .
then @xmath123 is an embedded real heegaard decomposition and @xmath124 , is an abstract real heegaard decomposition for @xmath10 .
here we give a series of examples of real open book decompositions and real ( embedded ) heegaard decompositions on @xmath12 .
it is well - known that on @xmath12 , up to isotopy , there is a unique real structure with nonempty real part ; its fixed point set is an unknot @xcite ( cf .
@xcite ) . let us fix the real structure @xmath125 on @xmath12 , considered as one point compactification of @xmath126 , as the one induced from the rotation on @xmath126 by @xmath54 along the @xmath127-axis .
we consider the open book on @xmath12 with binding @xmath128-axis @xmath129 and with pages topologically disks .
abstractly , this is the open book of @xmath12 with page @xmath130 a disk and monodromy @xmath131 the identity map .
there are two invariant pages on which the real structure acts as the reflection , @xmath132 , with respect to the @xmath127-axis .
the corresponding heegaard decomposition has the splitting surface @xmath133 a sphere and the gluing map @xmath134 between the two handlebodies is a reflection on @xmath133 fixing an _
equator_. let us perform a positive real stabilization on the real open book we have .
at this point it is important to note that we will describe the stabilization on an abstract real open book .
however , one should keep in mind that abstract positive real stabilization can be seen as an _ equivariant _ connected sum with @xmath135 equipped with a positive hopf real open book .
now , consider the open book @xmath136 , we make a positive stabilization of type i ( see figure [ stablar ] ) to obtain the page @xmath137 an annulus .
the monodromy @xmath138 is the dehn twist @xmath139 along the core @xmath140 of @xmath137 .
the real structure @xmath141 on @xmath142 is the reflection with respect to the core @xmath143 composed by the dehn twist along @xmath143 , while the real structure @xmath144 on the opposite real page acts as the reflection with respect to @xmath143 ; the corresponding heegaard surface @xmath145 is a torus ; the real structure on @xmath145 is @xmath146 which is exactly the gluing map @xmath147 between the two handlebodies .
note that @xmath148 is conjugate to the involution which interchanges the meridian and the longitude . in the next steps
we perform several positive real stabilizations of type viii so that we obtain a sequence of real open books and real heegaard decompositions of odd genus for @xmath12 : given odd @xmath149 , there is a real heegaard splitting , see figure [ stabor1 ] , with surface @xmath150 of genus @xmath151 and with the real structure and the gluing map given by @xmath152 on @xmath153 where @xmath154 with @xmath155 , the real structure on @xmath156 obtained by extending @xmath132 to the added handles . now starting from the real open book @xmath157 of @xmath12 with disk pages
, we perform a sequence of positive real stabilizations of type iii .
thus we get a real heegaard splitting , see figure [ stabor2 ] , with surface @xmath150 of genus @xmath158 and with the real structure and the gluing map @xmath159on @xmath150 where @xmath160 is the real structure on @xmath161 .
note that in this case the real part of the real structure @xmath162 is a separating curve .
finally , starting from the open book of @xmath12 with disk pages again , we perform twice positive real stabilization of type ii and then several positive real stabilizations of type iii . we obtain a real heegaard splitting , see figure [ stabor3 ] , with surface @xmath153 of even genus @xmath158 and the gluing map @xmath159 , where @xmath163 .
we want to make some observations on real heegaard decompositions on lens spaces .
the classification of real structures on lens spaces up to isotopy is known @xcite . in op.cit .
the real structures of type @xmath164 and @xmath165 on a lens space are already in the form of real heegaard decompositions of genus 1 . however on most of the lens spaces the real structures of type @xmath166 , @xmath167 and @xmath168
can not be presented as a real heegaard decomposition of genus 1 while we know that there _ is _ a real heegaard decomposition @xcite .
of course , in that case the higher genera decompositions will have the real structure act on the heegaard surface not maximally .
we will call a real heegaard decomposition _
maximal _ if the real part of the real structure has genus@xmath11 connected components ( equivalently , if the number of real components is maximal with respect to harnack inequality ) .
this leads to the following observation .
[ maksimal ] ( through an observation in @xcite ) there are infinitely many real closed oriented 3-manifolds which can not have a maximal real heegaard decomposition .
several more observations are in order : * all open book decompositions with annuli pages can be made real ( since the monodromy can only be a power of the dehn twist along the belt circle and any power of the dehn twist can be written as a product of two real structures ) .
thus , all lens spaces admitting an open book with annuli pages can be made real and hence admit a real heegaard splitting of genus 1 .
for example , in the link of @xmath169 singularities ( the lens spaces @xmath170 so that the coefficients of the continued fractions of @xmath171 are only @xmath172 s ) the _ horizontal _ open book obtained from the plumbing ( see e.g. @xcite for the construction ) has annuli pages .
* the lens spaces @xmath173 where the coefficients of the continued fraction expansion of @xmath171 are all @xmath172 except one which is @xmath174 have an open book decomposition with page thrice punctured sphere and with monodromy a product of certain powers of the dehn twists around the boundary components . since with respect to any real structure on the page the boundary components are invariant , these open books are always real for any choice of real structure .
hence , this type of lens spaces have a real genus-2 heegaard splitting . * for an example to above consider @xmath175 on which there are 3 real structures ; they are of types @xmath166 , @xmath167 and @xmath168 .
hence there is no genus-1 real heegaard decomposition of @xmath175 .
meanwhile , on a sphere with three punctures there are only two real structures .
thus , with respect to at least one real structure the corresponding heegaard decomposition must have genus greater than @xmath176 . following this discussion ,
we ask a natural question : * question . *
given a real closed 3-manifold @xmath10 , what is the _
@xmath9-real heegaard genus _
, i.e. what is the minimum genus among all real heegaard decompositions representing @xmath9 ? to finish this section , let us display the question we raised in the introduction .
* question .
* does every real 3-manifold @xmath10 admit a @xmath9-real open book ?
note that if the answer is affirmative then with theorem [ jiru ] below , this would imply that every real 3-manifold admits a real contact structure .
in this section we prove ( _ real giroux correspondence _ ) [ jiru ] let @xmath10 be a real 3-manifold . then there is a one to one correspondence between the real contact structures on @xmath6 up to equivariant contact isotopy and the real open books on @xmath6 up to positive real stabilization and equivariant isotopy .
first we explain how to construct a real contact form on the real manifold @xmath177 built - up as before from the abstract real open book decomposition @xmath178 .
the construction is the real version of the one of w. p. thurston and h. winkelnkemper @xcite and it motivates the following definition . _
proof : _ let @xmath183 be a 1-form on @xmath31 such that @xmath184 near @xmath34 and @xmath185 is a positive volume form on @xmath31 .
it is elementary to show that such a form exists .
an argument can be found in @xcite .
we now consider the 1-forms @xmath186 on @xmath187 and @xmath188 on @xmath189 so that we define @xmath190 which , by definition , induces a 1-form on the mapping torus @xmath191 however , this form is in general not real . thus let us set @xmath192 . by definition @xmath193
is a real form .
as the action of @xmath50 is the identity on @xmath194 , we get @xmath195 we claim that for large values of @xmath196 , @xmath193 is a real contact form on @xmath197 . to prove this claim we only need to show that @xmath198 is positive on @xmath197 for sufficiently large @xmath196 where positivity is determined with respect to the orientation defined by the form @xmath199
we have @xmath200\\ & & + dt\wedge [ ( f \circ c)^*\beta\wedge c^*\beta+ \beta \wedge ( ( f\circ c)^*\beta - c^*\beta)]\\ & & + [ ( 1-t)c^*\beta+t(f \circ c)^*\beta+\beta]\\ & & \;\;\;\ ; \wedge[(1-t)dc^*\beta+td(f\circ c)^*\beta - d\beta ] .
\end{array}\ ] ] the last term evaluates zero on @xmath197 .
this is because both @xmath183 and @xmath185 are forms on the surface @xmath31 , so are their pull - backs by a diffeomorphism of @xmath31 .
thus any wedge product of the form @xmath201 evaluated on the 3-manifold @xmath197 is zero .
in addition , as @xmath7 is orientation reversing , both @xmath202 and @xmath203 are volume forms on @xmath31 defining the same orientation as @xmath185 .
thus , the first term @xmath204 $ ] which is equal to @xmath205 \wedge dt$ ] is positive on @xmath197 and is dominant when @xmath196 is large enough . for the extension of @xmath193 over the solid tori , we need to consider two separate cases which are distinguished by the action of @xmath206 on @xmath34 .
as discussed above , either @xmath206 acts on a boundary component as reflection or it switches two boundary components .
we consider @xmath208 , \theta \in [ -\pi , \pi ] \}$ ] where @xmath209 . on the other hand ,
we take @xmath210 , r\in [ 0,1 ] \}$ ] together with the real structure @xmath211 and the 1-form @xmath212 .
note that , @xmath213 is a real form for all @xmath214 and is a contact form for those @xmath214 satisfying @xmath215 .
we identify @xmath216\times s^1 \subset s^1\times d^2 $ ] with @xmath217 by an equivariant orientation preserving diffeomorphism @xmath218 defined as @xmath219\times s^1 & \to & \nu(s^1_{\partial } ) \times s^1\\ ( \vartheta , r , \varphi)&\mapsto&(1-r-\epsilon , \vartheta , \varphi ) .
\end{array}\ ] ] thus , we get @xmath220 , $ } \\
2e^{1-r-\epsilon } d\vartheta-2kd\varphi & \textrm{if $ \varphi \in [ -\pi,0 ] .$ } \end{array}\right.\ ] ] since we require the extended 1-form to be positive on the binding , and to match with @xmath193 on @xmath216\times s^1 $ ] , it is enough to find smooth @xmath221 and @xmath222 such that @xmath223 , \end{array}\ ] ] and it is easy to see that there exist @xmath224 satisfying the above conditions .
case 2 : let @xmath225 and @xmath226 be two boundary components of @xmath31 such that @xmath227 .
we consider two contact solid tori @xmath228 with contact structures @xmath229 and @xmath230 . as before ,
we identify @xmath231 \times s^1 \subset s^1\times d^2 $ ] with @xmath232 by @xmath218 then the identification @xmath233 of @xmath234 is determined by the symmetry @xmath235\times s^1 \ar[r]^{\upsilon } \ar[d]_{c_{s^1\times d^2 } } & \nu(s^1_{\partial_{1}})\times s^1 \ar[d]^{c_{s_{f } } } \\ s^1\times [ 1-\epsilon , 1]\times s^1 \ar[r]^{\upsilon ' } & \nu(s^1_{\partial_{2}})\times s^1 .}\ ] ] let @xmath241 , \varphi , \vartheta \in [ -\pi , \pi]\}$ ] be a neighborhood of binding , and @xmath242 a function such that @xmath243 near 0 , and @xmath244 near @xmath245 and @xmath246 .
define @xmath247 .
it can be easily checked that all @xmath248 for @xmath249 and @xmath250 are contact and real , thus they define ( equivariantly ) isotopic real contact forms .
proof follows from lemma [ hucreleme ] and lemma [ kurdele aksi ] .
first we construct a _ real contact cell decomposition _ on @xmath6 ( lemma [ hucreleme ] ) and then using that we build the real open book ( lemma [ kurdele aksi ] ) .
let @xmath254 be the real contact manifold .
a _ real contact cell decomposition _ over @xmath6 is a cell decomposition of @xmath6 with each @xmath255-skeleton @xmath9-invariant ( @xmath256 ) ; each 2-cell is convex with tw@xmath257 ; and @xmath258 is tight when restricted to each 3-cell . _
proof : _ we cover @xmath6 with a @xmath9-invariant , finite set of darboux balls .
we then choose a real cell decomposition such that each 3-cell lies in the interior of a darboux ball .
note that every real manifold has an equivariant cell decomposition @xcite .
we now turn this equivariant cell decomposition into a contact one .
first , we make the 1-skeleton legendrian .
note that the real structure acts on a 1-cell as the identity or a reflection , or it swaps two 1-cells . by definition ,
the real part is legendrian .
if the real structure exchanges two 1-cells , those cells are disjoint .
if their closure are not disjoint , by possiblly subdividing the cell decompostion equivariantly , we can make sure that they have a single common end point which is real .
as the condition of being legendrian is satisfied at a real point , we can make such a 1-cell pair legendrian symmetrically .
if the real structure acts on a 1-cell @xmath259 as reflection then the two halves of @xmath259 can be perturbed legendrian symmetrically relative their common real end point , at which the legendrian condition is satisfied .
as each 2-cell remains inside the darboux balls , ( weak ) bennequin inequality applies , so tw@xmath260 for each 2-cell @xmath261 .
hence we can perturb each @xmath261 to make it convex ( * ? ? ?
* proposition 3.1 ) .
the real structure can act on a 2-cell as a reflection or a rotation , or it switches two 2-cells . in the latter case ,
@xmath261 and @xmath262 are disjoint ; closure of them may have a common boundary piece which is real
. therefore each such 2-cell @xmath261 can be perturbed relative its boundary to make it convex and @xmath262 can be perturbed symmetrically .
if the real structure acts on @xmath261 as reflection , then as before the two halves can be perturbed symmetrically relative their boundary to make them convex .
if @xmath9 acts on a 2-cell @xmath261 as rotation , then the characteristic field on @xmath261 is an anti - symmetric vector field tangent to the boundary .
the work in @xcite shows that among all such vector fields , the structural stable ones are dense .
thus we conclude that @xmath261 can be perturbed convex , employing similar final steps in the proof of proposition 3.3 in @xcite .
once every 2-cell is made convex , we subdivide those 2-cells with tw@xmath263 to get 2-cells with correct twisting .
the subdivision can be made symmetrically .
this is true even if the 2-cell is @xmath8-invariant since the dividing set can be chosen invariant ( * ? ? ?
* theorem 3.1 ) . by the legendrian realization principle
@xcite each arc added to 2-cells during the subdivision can be made legendrian keeping the 2-cells convex .
therefore , we obtain a @xmath9-equivariant contact cell decomposition on @xmath6 . @xmath56 + given @xmath254 take a real contact cell decomposition .
the _ real ribbon _ of the 1-skeleton @xmath264 is a compact surface @xmath265 with boundary such that ( i ) @xmath9 acts on @xmath265 reversing the orientation , ( ii ) @xmath265 retracts onto @xmath264 and ( iii ) @xmath265 is tangent to @xmath266 if and only if @xmath267 .
it is easy to see that any invariant legendrian graph has a real ribbon .
_ proof : _ note that the second claim is independent from the real structure and a proof has been given in @xcite .
moreover , since it is known that the ribbon is a page of an open book supporting @xmath258 ( see @xcite or @xcite ) , we only need to show that the open book having real ribbon as a page can be constructed equivariantly . to this end ,
let us first choose an invariant neighborhood @xmath269 of the real ribbon @xmath265 such that @xmath270 and that the boundary @xmath271 is a convex surface with dividing curve @xmath272 .
we can identify @xmath269 with @xmath273 $ ] quotient by @xmath274 , for @xmath275 and @xmath276 $ ] on which real structure acts as @xmath277 .
then the projection @xmath278/_{\sim}\to [ -1 , 1]$ ] defines one half " of the open book . by the construction of the contact cell decomposition
each 2-cell @xmath89 of the cell decomposition is convex and has boundary with twisting @xmath279 relative @xmath89 . since @xmath265 twists with the contact structure along the
1-skeleton , each 2-cell @xmath89 intersects @xmath272 at exactly two points .
thus , we can extend the fibration in @xmath269 , in an obvious way , in a neighborhood of each 2-cell . in figure [ 2hucre ] ( left ) , we depict a 2-cell foliated by the pages of the open book .
each page traces an arc and all arcs meet at points touching @xmath272 . as for
the extension of the real structure , let us first assume that the 2-cell @xmath89 is invariant , so the boundary of the 2-cell admits an involution which is a reflection or a rotation . in the case of reflection ,
the involution fixes the points touching @xmath272 .
since any two involutions on a circle fixing the same pair of points can be extended to an involution of a disk in such way that the extension is unique up to isotopy relative to boundary , we can extend the real structure on a neighborhood of the 2-cell in such a way that it commutes with the fibration ( figure [ 2hucre ] , middle ) . in the case of rotation the points touching @xmath272 interchanged by the real structure , so similar idea applies ( figure [ 2hucre ] , right ) .
the case when the real structure swaps two 2-cells can also be similarly treated .
now , if we trace the dividing curve on the boundary of each ( 3-cell@xmath280int@xmath281 ) , we obtain a connected closed curve @xcite .
therefore , we can extend the open book over 3-cells in such a way that the dividing curve becomes a part of the binding and the pages foliate the 3-cells as shown in figure [ 3top ] .
in fact , an involution on @xmath282 with the prescribed image of a fixed closed curve can be extended to a unique ( up to isotopy relative boundary ) involution of @xmath283 preserving the fibration described above . _
proof : _ the proof follows the idea of the non - real case proposed by e. giroux @xcite .
it is enough to show that any real open book supporting @xmath258 comes from a real contact cell decomposition after possibly a number of positive real stabilizations .
then the rest of the proof will follow similar lines of the original proof .
let @xmath13 be a real open book of @xmath14 supporting @xmath258 .
we consider the heegaard decomposition @xmath284 defined by @xmath54 with the heegaard surface @xmath285 .
let @xmath7 be the diffeomorphism @xmath286 where @xmath287 is the inherited real structure on the page @xmath288 , @xmath289 . recall that @xmath290 where @xmath26 is the monodromy of the open book .
note that @xmath66 is convex ; in fact , consider a self - indexing real morse function @xmath291 $ ] with @xmath292 and @xmath293)$ ] , @xmath294)$ ] .
then its gradient field is by definition contact and transversal to @xmath66 , rendering @xmath66 convex .
moreover @xmath167 is an invariant dividing curve on @xmath66 .
let @xmath98 be an abstract surface diffeomorphic to a page .
we consider the abstract handlebody @xmath295/_{(x , t)\sim(x , t')}$ ] for all @xmath296 and @xmath297 $ ] .
note that @xmath298 gets a real structure @xmath299 from @xmath24 ( keeping the same notation for the real structures on @xmath24 and on @xmath161 ) .
we define a cell decomposition on @xmath300 as follows .
consider a @xmath299-invariant set @xmath166 of disjoint proper arcs on @xmath161 dividing it into disks .
each such arc @xmath87 defines a disk @xmath301/_{\sim } \subset w$ ] . by the choice of the set @xmath166 , the set of all @xmath302 s cuts @xmath300 into 3-balls ,
so we have a cell decomposition on @xmath300 whose 1-skeleton is two copies of @xmath166 ( one on @xmath161 , the other , say @xmath303 on @xmath304 ) and whose 2-skeleton consists of the disks @xmath305 and the disks on @xmath306 .
let @xmath164 denote a @xmath299-invariant set of closed curves on @xmath161 forming a core for @xmath161 .
we may impose that @xmath164 contains the closed real curves ( if exist ) lying on @xmath161 .
such a core can always be chosen .
now , we add @xmath164 to the 1-skeleton of the cell decomposition described above to get a cell decomposition on @xmath300 in the sense of l. siebenmann @xcite .
the cell decomposition on @xmath300 constructed above gives a cell decomposition on @xmath6 ( by means of the identifications of @xmath300 with @xmath307 ) whose 1-skeleton contains @xmath308 on @xmath161 and @xmath309 on @xmath310 .
note first that , the 1-skeleton is @xmath9-invariant .
namely , we have @xmath311 , as @xmath164 and @xmath166 are chosen @xmath299-invariant .
meanwhile , @xmath312 ; thus , @xmath309 is @xmath120-invariant .
it is easy to see that by construction , 2- and 3-skeletons are also @xmath9-invariant .
thus , we get a real cell decomposition on @xmath14 .
we now perturb 1-skeleton equivariantly as before to make it legendrian . for 1-cells on @xmath161
legendrian realization principle applies , so we make them legendrian keeping @xmath161 convex and the 1-cells @xmath119-invariant .
then we perturb the 1-cells on @xmath310 legendrian keeping them equivariant .
note that we are not required to reside on @xmath310 nor to keep @xmath310 convex here .
again as before , each 2-cell has twisting less than @xmath279 , so for each 2-cell @xmath261 we can arrange tw@xmath313 to be @xmath279 by subdividing @xmath261 .
each arc of the subdivision can be made legendrian keeping @xmath261 convex . by construction
each 3-cell is away from a core of a page , thus it lies in a tight manifold .
now , for each arc in the 1-skeleton not lying on @xmath161 , we can perform a stabilization to include it to the page @xmath161 .
such a stabilization exists since by the construction of the cell decomposition , each such arc together with an arc on @xmath161 bounds a disk @xmath89 with tw@xmath314 .
hence , we perform a finite number of positive real stabilizations so that the stabilized @xmath161 becomes a real ribbon of the 1-skeleton .
@xmath56 + it is worth mentioning that in the proof of lemma [ hucreleme ] , it is possible to start with an equivariant cell decomposition where the real part lies entirely in the 1-skeleton .
furthermore , it is easy to observe that if given an equivariant cell decomposition , one can always get the real part lie entirely in the 1-skeleton .
then by turning such a decomposition into a contact cell decomposition and by defining the real ribbon , we indeed obtain a real open book with real part lying entirely on a page .
however , if we restrict ourselves to such class of equivariant cell decompositions , then the proof for proposition [ aynikitap ] would not work . more precisely , it is not true that given a real open book there is a series of positive real stabilizations which produces a real open book with the real part lying entirely on a page .
for example , the real hopf open book depicted on the top of figure [ hopfstablari ] can not be positively equivariantly stabilized so that the real part becomes a subset of a page .
in fact , it is not hard to see that if a page contains more than one piece of the same connected component of the real part , then there is no way to get the real part lie on a page by positive real stabilizations .
etnyre , j. b. lectures on open book decompositions and contact structures .
floer homology , gauge theory , and low - dimensional topology , 103141 , clay math .
proc . , 5 , amer .
soc . , providence , ri , 2006 .
giroux , e. gomtrie de contact : de la dimension trois vers les dimensions suprieures , in proceedings of the international congress of mathematicians ( beijing , 2002 ) , vol .
ii , higher education press , beijing ( 2002 ) , 405414 . | a real 3-manifold is a smooth 3-manifold together with an orientation preserving smooth involution , called a real structure . in this article
we study open book decompositions on smooth real 3-manifolds that are compatible with the real structure .
we call them real open book decompositions . we show that giroux correspondence applies to real manifolds ; namely we prove that there is a one to one correspondence between the real contact structures on a 3-manifold up to equivariant contact isotopy and the real open books up to positive real stabilization . |
a key factor in the efficiency of a peer - to - peer overlay network is the level of collaboration provided by each peer .
this paper takes a first step towards quantifying the level of collaboration that can be expected from each participant , by proposing a model to evaluate the cost each peer incurs for being a part of the overlay .
such a cost model has several useful applications , among which , ( 1 ) providing a benchmark that can be used to compare between different proposals , complementary to recent works comparing topological properties of various overlays @xcite , ( 2 ) allowing for predicting disincentives , and designing mechanisms that ensure a protocol is _ strategyproof _
@xcite , and ( 3 ) facilitating the design of load balancing primitives .
this work is not the first attempt to characterize the cost of participating in a network .
jackson and wolinsky @xcite proposed cost models to analyze formation strategies in social and economic networks .
more recent studies @xcite model ( overlay ) network formation as a non - cooperative game .
these studies assume that each node has the freedom to choose which links it maintains , whereas we assume that the overlay topology is constrained by a protocol
. moreover , our approach extends previously proposed cost models @xcite , by considering the load imposed on each node in addition to the distance to other nodes and degree of connectivity . in the remainder of this paper
, we introduce our proposed cost model , before applying it to several routing geometries used in recently proposed distributed hash tables ( dht ) algorithms @xcite .
we conclude by discussing some open problems this research has uncovered .
the model we propose applies to any peer - to - peer network where nodes request and serve items , or serve requests between other nodes .
this includes peer - to - peer file - sharing systems @xcite , ad - hoc networks @xcite , peer - to - peer lookup services @xcite , peer - to - peer streaming systems @xcite , or application - layer multicast overlays @xcite , to name a few examples . to simplify the presentation
, we assume a dht - like structure , defined by quadruplet @xmath0 , where @xmath1 is the set of vertices in the network , @xmath2 is the set of edges , @xmath3 is the set of keys ( items ) in the network , and @xmath4 is the hash function that assigns keys to vertices .
we denote by @xmath5 the set of keys stored at node @xmath6 .
we have @xmath7 , and we assume , without loss of generality , that the sets @xmath8 are disjoint .
we characterize each request with two independent random variables , @xmath9 and @xmath10 , which denote the node @xmath11 making the request , and the key @xmath12 being requested , respectively .
consider a given node @xmath6 .
every time a key @xmath13 is requested in the entire network , node @xmath14 is in one of four situations : 1 .
node @xmath14 does not hold or request @xmath13 , and is not on the routing path of the request .
node @xmath14 is not subject to any cost .
2 . node @xmath14 holds key @xmath13 , and pays a price @xmath15 for serving the request .
we define the _ service cost _ @xmath16 incurred by @xmath14 , as the expected value of @xmath15 over all possible requests .
that is , @xmath17 \ .\ ] ] 3 .
node @xmath14 requests key @xmath13 , and pays a price to look up and retrieve @xmath13 .
we model this price as @xmath18 , where @xmath19 is the number of hops between @xmath14 and the node @xmath20 that holds the key @xmath13 , and @xmath21 is a ( positive ) proportional factor .
we define the _ access cost _ suffered by node @xmath14 , @xmath22 , as the sum of the individual costs @xmath18 multiplied by the probability key @xmath23 is requested , that is , @xmath24 \ , \label{eq : access}\ ] ] with @xmath25 if there is no path from node @xmath14 to node @xmath20 , and @xmath26 for any @xmath14 .
4 . @xmath14 does not hold or request @xmath13 , but has to forward the request for @xmath13 , thereby paying a price @xmath27 .
the overall _ routing cost _ @xmath28 experienced by node @xmath14 is the average over all possible keys @xmath13 , of the values of @xmath27 such that @xmath14 is on the path of the request .
that is , we consider the binary function @xmath29 and express @xmath28 as @xmath30 \pr[y = k ] \chi_{j , l}(i)\ .
\label{eq : routing}\ ] ] in addition , each node keeps some state information so that the protocol governing the dht operates correctly . in most dht algorithms ,
each node @xmath14 maintains a neighborhood table , which grows linearly with the out - degree @xmath31 of the node , resulting in a _ maintenance cost _ @xmath32 given by @xmath33 where @xmath34 denotes the cost of keeping a single entry in the neighborhood table of node @xmath14 .
last , the _ total cost _ @xmath35 imposed on node @xmath14 is given by @xmath36 which can be used to compute the total cost of the network , @xmath37 .
the topology that minimizes @xmath38 , or `` social optimum , '' is generally not trivial . in particular ,
the social optimum is the full mesh only if @xmath39 for all @xmath14 , and the empty set only if @xmath40 for all @xmath41 .
we next apply the proposed cost model to a few selected routing geometries . we define a routing geometry as in @xcite , that is , as a collection of edges , or topology , associated with a route selection mechanism . unless otherwise noted , we assume shortest path routing , and distinguish between different topologies . we derive the various costs experienced by a node in each geometry , before illustrating the results with numerical examples .
we consider a network of @xmath42 nodes , and , for simplicity , assume that , for all @xmath14 and @xmath13 , @xmath43 , @xmath44 , @xmath45 , and @xmath46 . for the analysis in this section , we also assume that each node holds the same number of keys , and that all keys have the same popularity . as a result , for all @xmath14 , @xmath47 = \frac{1}{n } \ , \ ] ] which implies
@xmath48 regardless of the geometry considered .
we also assume that requests are uniformly distributed over the set of nodes , that is , for any node @xmath14 , @xmath49 = \frac{1}{n } \ .\ ] ] last , we assume that no node is acting maliciously . [ [ star - network ] ] star network + + + + + + + + + + + + the star frequently appears as an equilibrium in network formation studies using cost models based on graph connectivity @xcite .
we use @xmath50 to denote the center of the star , which routes all traffic between peripheral nodes .
that is , @xmath51 for any @xmath52 ( @xmath53 , @xmath54 ) . substituting in eqn .
( [ eq : routing ] ) , we get @xmath55 the center node is located at a distance of one hop from all @xmath56 other nodes , thus @xmath57 in addition , @xmath58 , which implies that the cost incurred by the center of the star , @xmath59 , is @xmath60 peripheral nodes do not route any traffic , i.e. , @xmath61 for all @xmath62 , and are located at a distance of one from the center of the star , and at a distance of two from the @xmath63 other nodes , giving @xmath64 furthermore , @xmath65 for all peripheral nodes .
thus , @xmath66 , and the total cost imposed on nodes @xmath67 is @xmath68 the difference @xmath69 quantifies the ( dis)incentive to be in the center of the star . as expressed in the following two theorems , there is a ( dis)incentive to be in the center of the star in a vast majority of cases . if the number of nodes @xmath70 ( @xmath71 ) is variable , @xmath72 unless @xmath73 .
[ theo : star - asymmetry - x1 ] assume that @xmath74 . because @xmath75 , @xmath74 is equivalent to @xmath76 . using the expressions for @xmath59 and @xmath35 given in eqs .
( [ eq : c0-star ] ) and ( [ eq : ci - star ] ) , and rewriting the condition @xmath76 as a polynomial in @xmath70 , we obtain @xmath77 we can factor the above by @xmath63 , and obtain @xmath78 a polynomial in @xmath70 is constantly equal to zero if and only if all of the polynomial coefficients are equal to zero .
thus , eqn .
( [ eq : polynomial ] ) holds for _ any _ value of @xmath70 if and only if : @xmath79 the solutions of the above system of equations are @xmath80 .
hence , @xmath74 for any @xmath70 only when nodes only pay an ( arbitrary ) price for serving data , while state maintenance , traffic forwarding , and key lookup and retrieval come for free . if the number of nodes @xmath70 ( @xmath42 ) is held fixed , and at least one of @xmath81 , @xmath82 , or @xmath83 is different from zero , @xmath84 only if @xmath85 or @xmath86 , where @xmath87 is a positive integer that must satisfy : @xmath88 additionally , @xmath72 for any @xmath89 if @xmath90 and @xmath91 .
[ theo : star - asymmetry - x2 ] recall from the proof of theorem [ theo : star - asymmetry - x1 ] , that @xmath92 is equivalent to eqn .
( [ eq : polynomial ] ) . clearly , setting @xmath93 satisfies eqn .
( [ eq : polynomial ] ) for all values of @xmath94 , @xmath82 , and @xmath81 .
assuming now that @xmath95 , to have @xmath92 , we need to have @xmath96 since at least one of @xmath81 , @xmath82 , or @xmath83 is not equal to zero , eqn .
( [ eq : degree2 ] ) has at most two real solutions .
we distinguish between all possible cases for @xmath81 , @xmath82 , and @xmath83 such that at least one of @xmath81 , @xmath82 , and @xmath83 is different from zero . * if @xmath97 , and @xmath98 , eqn .
( [ eq : degree2 ] ) reduces to @xmath99 , which implies @xmath73 , thereby contradicting the hypothesis that at least one of @xmath81 , @xmath82 , and @xmath83 is different from zero .
therefore , eqn .
( [ eq : degree2 ] ) does not admit any solution , i.e. , there is a ( dis)incentive to be in the center of the star regardless of @xmath70 . * if @xmath97 and @xmath100 , the only solution to eqn .
( [ eq : degree2 ] ) is @xmath101 note that if @xmath102 , @xmath103 which is not feasible .
( the number of nodes has to be positive . ) * if @xmath104 , then eqn . ( [ eq : degree2 ] ) admits two real roots ( or a double root if @xmath105 ) , given by @xmath106 however , because @xmath107 , and @xmath108 , @xmath109 so that the only potentially feasible @xmath87 is given by @xmath110 combining eqs .
( [ eq : degree1-sol ] ) and eqs .
( [ eq : degree2-sol ] ) yields the expression for @xmath87 given in eqn .
( [ eq : n0 ] ) .
note that the expression given in eqn .
( [ eq : n0 ] ) is only a necessary condition .
in addition , @xmath87 has to be an integer so that we can set the number of nodes @xmath70 to @xmath111 .
[ [ de - bruijn - graphs ] ] de bruijn graphs + + + + + + + + + + + + + + + + de bruijn graphs are used in algorithms such as koorde @xcite , distance - halving @xcite , or odri @xcite , and are extensively discussed in @xcite . in a de bruijn graph , any node @xmath14 is represented by an identifier string @xmath112 of @xmath113 symbols taken from an alphabet of size @xmath114 .
the node represented by @xmath112 links to each node represented by @xmath115 for all possible values of @xmath116 in the alphabet .
the resulting directed graph has a fixed out - degree @xmath114 , and a diameter @xmath113 .
denote by @xmath117 the set of nodes such that the identifier of each node in @xmath117 is of the form @xmath118 .
nodes in @xmath117 link to themselves , so that @xmath119 for @xmath120 . for nodes
@xmath121 , the maintenance cost @xmath32 is @xmath122 .
the next two lemmas will allow us to show that the routing cost at each node also depends on the position of the node in the graph . with shortest - path routing , nodes @xmath120
do not route any traffic , and @xmath61 .
[ lemma : lowerbnd - deb - routes ] ( by contradiction . )
consider a node @xmath120 with identifier @xmath118 , and suppose @xmath14 routes traffic from a node @xmath20 to a node @xmath13 .
the nodes linking to @xmath14 are all the nodes with an identifier of the form @xmath123 , for all values of @xmath116 in the alphabet .
the nodes linked from @xmath14 are all the nodes of the form @xmath124 for all values of @xmath125 in the alphabet .
therefore , there exists @xmath126 and @xmath127 such that traffic from node @xmath20 to node @xmath13 follows a path @xmath128 . because , in a de bruijn graph , there is an edge between @xmath129 and @xmath130 , traffic using the path @xmath131 between @xmath20 and @xmath13 does not follows the shortest path .
we arrive to a contradiction , which proves that @xmath14 does not route any traffic .
the number of routes @xmath132 passing through a given node @xmath14 is bounded by @xmath133 with @xmath134 the bound is tight , since it can be reached when @xmath135 for the node @xmath136 .
[ lemma : upperbnd - deb - routes ] the proof follows the spirit of the proof used in @xcite to bound the maximum number of routes passing through a given edge . in a de bruijn graph , by construction
, each node maps to an identifier string of length @xmath113 , and each path of length @xmath13 hops maps to a string of length @xmath137 , where each substring of @xmath113 consecutive symbols corresponds to a different hop @xcite .
thus , determining an upper bound on the number of paths of length @xmath13 that pass through a given node @xmath14 is equivalent to computing the maximum number , @xmath138 , of strings of length @xmath137 that include node @xmath14 s identifier , @xmath139 , as a substring . in each string of length
@xmath137 corresponding to a paths including @xmath14 , where @xmath14 is neither the source nor the destination of the path , the substring @xmath140 can start at one of @xmath141 positions @xmath142 .
there are @xmath114 possible choices for each of the @xmath13 symbols in the string of length @xmath137 that are not part of the substring @xmath140 . as a result ,
@xmath143 with shortest path routing , the set of all paths going through node @xmath14 include all paths of length @xmath137 with @xmath144 $ ] .
so , @xmath145 we improve the bound given in eqn .
( [ eq : intermed - pi - bound ] ) by considering the strings of length @xmath146 that are of the form @xmath147 , where @xmath148 is a string of length @xmath113 .
strings of the form @xmath147 denote a cycle @xmath149 , and can not be a shortest path in a de bruijn graph .
hence , we can subtract the number of the strings @xmath147 from the bound in eqn .
( [ eq : intermed - pi - bound ] ) . because @xmath150 is a substring of @xmath147 of length @xmath113
, @xmath148 has to be a circular permutation of @xmath140 , for instance @xmath151 . since @xmath14 does not route any traffic when @xmath14 is the source of traffic , @xmath152 .
thus , there are only @xmath153 possibilities for @xmath148 , and @xmath153 strings @xmath147 .
subtracting @xmath153 from the bound in eqn .
( [ eq : intermed - pi - bound ] ) yields @xmath154 . from lemmas [ lemma : lowerbnd - deb - routes ] and
[ lemma : upperbnd - deb - routes ] , we infer that , in a de bruijn graph , for any @xmath14 , @xmath20 and @xmath13 , @xmath155 \leq l_{\max}/n^2 $ ] . because @xmath156 is a binary function , @xmath157 = e[\chi_{i , j}]$ ] , and we finally obtain @xmath158 with @xmath159 we next compute upper and lower bounds on the access cost . to derive a tight upper bound on @xmath22 , consider a node @xmath120 .
node @xmath14 links to itself and has only @xmath160 neighbors .
each neighbor of @xmath14 has itself @xmath114 neighbors , so that there are @xmath161 nodes @xmath13 such that @xmath162 . by iteration and substitution in eqn .
( [ eq : access ] ) , we get , after simplification , @xmath163 , with @xmath164 and @xmath165 for nodes in @xmath117 .
now , consider that each node @xmath14 has at most @xmath114 neighbors .
then , node @xmath14 has at most @xmath166 nodes at distance 2 , at most @xmath167 nodes at distance 3 , and so forth .
hence , there are at least @xmath168 nodes at the maximum distance of @xmath113 from node @xmath14 .
we get @xmath169 which reduces to @xmath170 , with @xmath171 it can be shown that @xmath172 for the node @xmath173 when @xmath135 .
note that , the expressions for both @xmath174 and @xmath175 can be further simplified for @xmath176 , that is , when the identifier space is fully populated . [
[ d - dimensional - tori ] ] @xmath113-dimensional tori + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we next consider @xmath113-dimensional tori , as in can @xcite , where each node is represented by @xmath113 cartesian coordinates , and has @xmath146 neighbors , for a maintenance cost of @xmath177 for any @xmath14 .
routing at each node is implemented by greedy forwarding to the neighbor with the shortest euclidean distance to the destination .
we assume here that each node is in charge of an equal portion of the @xmath113-dimensional space . from @xcite ,
we know that the average length of a routing path is @xmath178 hops .. ] because we assume that the @xmath113-dimensional torus is equally partitioned , we conclude by symmetry , that for all @xmath14 , @xmath179 to determine the routing cost @xmath28 , we compute the number of routes passing through a given node @xmath14 , or _ node loading _
, as a function @xmath180 of the dimension @xmath113 . with our assumption that the @xmath113-torus is equally partitioned , @xmath180 is the same for all @xmath14 by symmetry .
we next compute @xmath180 by induction on the dimension @xmath113 .
[ [ base - case - d1 . ] ] * base case @xmath181 . * for @xmath182 , the @xmath113-torus is a ring , as depicted in figure [ fig : can - proof - ring ] for @xmath183 .
each of the diagrams in the figure corresponds to a case where the source of all requests , represented by a black node , is held fixed .
the numbers in each node ( @xmath184 ) represent the node coordinate , the different line styles represent the different routes to all destinations , and the numbers in parentheses denote the number of routes originating from the fixed source that pass through each of the other nodes . as shown in the figure , shifting the source of all requests from 0 to @xmath185 only results in shifting the number of routes that pass through each node .
hence , the node loading @xmath186 at each node @xmath14 , is equal to the sum of the number of routes passing through _ each _ node when the source is held fixed . in the figure , for @xmath183
, we have for any @xmath14 , @xmath187 . more generally , for @xmath70 odd , the sum of the number of routes passing through each node is equal to @xmath188 and for @xmath70 even , is given by @xmath189 we can express eqs .
( [ eq : can - l1-odd ] ) and ( [ eq : can - l1-even ] ) in a more compact form , which holds for any @xmath70 , @xmath190 routing in 3-torus .
coordinates are corrected one at a time , first along the horizontal axis , then along the vertical axis , and finally along the diagonal axis . , scaledwidth=80.0% ] [ [ general - case - d1 . ] ] * general case @xmath191 . *
the key observation to compute the number of routes @xmath180 passing through each node @xmath14 for @xmath192 , is that there are several equivalent shortest paths along the cartesian coordinates , because the coordinates of two consecutive nodes in a path can not differ in more than one dimension .
consider for instance , for @xmath193 , going from node ( 0,0 ) to node ( 1,1 ) : both @xmath194 and @xmath195 are equivalent shortest paths .
therefore , we can always pick the path that corrects coordinates successively , starting with the first coordinate , i.e. , @xmath196 in the above example .
consider a @xmath113-torus , as represented for @xmath197 in figure [ fig : can - proof - cube ] , where each of the @xmath198 nodes is represented by a dot .
the figure illustrate how requests are routed by correcting coordinates successively , with the example of three different paths , @xmath199 , @xmath200 , and @xmath201 . for any node @xmath13 ,
we compute the number of routes passing through @xmath13 .
we denote the source of the route as node @xmath14 , and the destination of the route as node @xmath20 .
we have @xmath202 .
we further denote the coordinates of @xmath14 , @xmath20 , and @xmath13 by @xmath112 , @xmath203 , and @xmath204 .
we distinguish between the only three possibilities for @xmath13 that are allowed by the routing scheme that corrects coordinates one at a time : 1 .
node @xmath13 has the same @xmath113-th coordinate as both the source @xmath14 and the destination @xmath20 , i.e. , @xmath205 . in other words , the route @xmath206 is entirely contained within a @xmath153-torus
this case is illustrated in the figure for the route represented by a solid line going from @xmath207 to @xmath208 through @xmath209 .
the corresponding @xmath153-torus containing @xmath14 , @xmath20 and @xmath13 is denoted by the shaded box . by definition of the node loading ,
the node loading resulting from all possible paths @xmath131 contained in a @xmath153-torus is equal to @xmath210 .
there are @xmath211 different such @xmath153-tori in the @xmath113-torus under consideration , one for each possible value of @xmath205 .
so , the total load incurred on each node by all paths which remain contained within a @xmath153-torus is equal to @xmath212 .
nodes @xmath14 , @xmath20 and @xmath13 all differ in their @xmath113-th coordinate , i.e. , @xmath213 .
because coordinates are corrected one at a time , for any @xmath214 , we must have @xmath215 .
this case is illustrated in the figure for the route represented by a dashed line , going from @xmath216 to @xmath217 , through @xmath218 .
since @xmath215 , nodes @xmath20 and @xmath13 belong to the same ring where only the @xmath113-th coordinate varies .
such a ring is represented in the figure by the dotted curve . from node
@xmath13 s perspective , routing traffic from @xmath14 to @xmath20 is equivalent to routing traffic between nodes @xmath219 and @xmath20 , where node @xmath219 satisfies @xmath220 for any @xmath214 , and @xmath221 .
( in the figure , the coordinates of node @xmath219 are @xmath222 . ) from our hypothesis @xmath223 , we have @xmath224 , which implies @xmath225 .
therefore , computing the number of routes passing through node @xmath13 coming from @xmath14 is equivalent to computing the number of routes passing through node @xmath13 and originating from node @xmath219 .
summing over all possible destination nodes @xmath20 , the computation of all routes passing through @xmath13 and originating from all nodes @xmath219 in the same ring as @xmath20 and @xmath13 is identical to the computation of the node loading in the base case @xmath182 .
so , the load imposed on node @xmath13 is equal to @xmath186 .
now , summing over all possible nodes @xmath14 is equivalent to summing over all possible rings where only the @xmath113-th coordinate varies .
there are are @xmath226 such rings in the @xmath113-torus .
we conclude that the total load incurred on each node @xmath13 by the paths going from all @xmath14 to all @xmath20 satisfying @xmath213 is equal to @xmath227 .
node @xmath13 has the same @xmath113-th coordinate as node @xmath14 , and a @xmath113-th coordinate different from that of the destination @xmath20 . in other words , @xmath228 , @xmath229
this situation is illustrated in the figure for the route going from @xmath230 to @xmath231 and passing through @xmath232 , and represented by a thick dotted line . in this configuration , there are @xmath233 possible choices for the destination node @xmath20 such that @xmath234 , and @xmath235 for @xmath236 .
there are @xmath237 possible choices for the source node @xmath14 such that @xmath229 and @xmath238 .
hence , in this configuration , there is a total of @xmath239 routes passing through each node @xmath13 . summing the node loadings obtained in all three possible cases above
, we obtain @xmath240 replacing @xmath186 by the expression given in eqn .
( [ eq : can - l1 ] ) , using @xmath241 , and removing the recursion in the above relationship , we obtain , for any node @xmath14 , @xmath242 for all @xmath14 , @xmath28 immediately follows from @xmath180 with @xmath243 [ [ plaxton - trees ] ] plaxton trees + + + + + + + + + + + + + we next consider the variant of plaxton trees @xcite used in pastry @xcite or tapestry @xcite .
nodes are represented by a string @xmath112 of @xmath113 digits in base @xmath114 .
each node is connected to @xmath244 distinct neighbors of the form @xmath245 , for @xmath246 , and @xmath247 .
, this geometry reduces to a hypercube . ]
the resulting maintenance cost is @xmath248 . among the different possibilities for the remaining coordinates @xmath249
, the protocols generally select a node that is nearby according to a spatial proximity metric .
we here assume that the spatial distribution of the nodes is uniform , and that the identifier space is fully populated ( i.e. , @xmath250 ) , which enables us to pick @xmath251 .
thus , two nodes @xmath14 and @xmath20 at a distance of @xmath13 hops differ in @xmath13 digits .
there are @xmath252 ways of choosing which digits are different , and each such digit can take any of @xmath253 values .
so , for a given node @xmath14 , there are @xmath254 nodes that are at distance @xmath13 from @xmath14 . multiplying by the total number of nodes @xmath250 , and dividing by the total number of paths @xmath255 , we infer that , for all @xmath14 , @xmath20 , and @xmath13 , we have @xmath256 = \frac{{d\choose k}(\delta - 1)^k}{n } \ . \label{eq : tij - lpm}\ ] ] now , for any @xmath14 and @xmath20 such that @xmath257 , because routes are unique , there are exactly @xmath258 different nodes on the path between @xmath14 and @xmath20 .
so , the probability that a node @xmath13 picked at random is on the path from @xmath14 to @xmath20 is @xmath259 = \frac{l-1}{n } \ . \label{eq : chi - knowing}\ ] ] the total probability theorem tells us that @xmath260 = \sum_{l=1}^{d}\pr[\chi_{i , j}(k ) = 1 | t_{i , j } = l]\cdot\pr[t_{i , j}=l ] \ .\ ] ] substituting with the expressions obtained for @xmath261 $ ] and @xmath262 $ ] in eqs .
( [ eq : tij - lpm ] ) and ( [ eq : chi - knowing ] ) gives : @xmath260 = \frac{1}{n^2}\sum_{l=1}^{d } ( l-1){d\choose l}(\delta - 1)^l \
, \label{eq : plaxton - x0}\ ] ] which can be simplified as follows .
we write : @xmath263 and rewrite the right - hand term as a function of the derivative of a series , @xmath264 or , equivalently , @xmath265 the binomial theorem allows us to simplify the above to : @xmath266 which , making the partial derivative explicit , becomes , @xmath267 and reduces to @xmath268 substituting in eqn .
( [ eq : plaxton - x0 ] ) gives : @xmath260 = \frac{\delta^{d-1}(d(\delta -1)-\delta)+1}{n^2 } \ , \ ] ] which we multiply by @xmath82 to obtain @xmath269 to compute the access cost @xmath22 , we use the relationship @xmath270 $ ] .
we have @xmath271 = \sum_{k=1}^{d } k \pr[t_{i , j } = k ] \ , \ ] ] which , using the expression for @xmath272 $ ] given in eqn .
( [ eq : tij - lpm ] ) , implies @xmath271 = \sum_{k=0}^{d } k \frac{{d\choose k}(\delta - 1)^k}{n } \ , \ ] ] and can be expressed in terms of the derivative of a classical series : @xmath271 = \frac{\delta-1}{n}\frac{\partial}{\partial\delta}\left(\sum_{k=0}^{d } { d\choose k}(\delta - 1)^k\right)\ .\ ] ] using the binomial theorem , the series on the right - hand side collapses to @xmath273 , which yields @xmath271 = \frac{\delta-1}{n}\frac{\partial(\delta^d)}{\partial\delta } \ .\ ] ] we compute the partial derivative , and obtain @xmath271 = \frac{d\delta^{d-1}(\delta-1)}{n } \ .\nonumber\ ] ] multiplying by @xmath83 to obtain @xmath22 , we eventually get , for all @xmath14 , @xmath274 which can be simplified , using @xmath250 : @xmath275 [ [ chord - rings ] ] chord rings + + + + + + + + + + + in a chord ring @xcite , nodes are represented using a binary string ( i.e. , @xmath276 ) .
when the ring is fully populated , each node @xmath14 is connected to a set of @xmath113 neighbors , with identifiers @xmath277 for @xmath278 .
an analysis identical to the above yields @xmath28 and @xmath22 as in eqs .
( [ eq : routing - lpm ] ) and ( [ eq : access - lpm ] ) for @xmath276 .
note that eqn .
( [ eq : access - lpm ] ) with @xmath276 is confirmed by experimental measurements @xcite . .
[ tab : debruijn ] asymmetry in costs in a de bruijn graph @xmath279 [ cols="^,^,^,^,^,^,^",options="header " , ] we illustrate our analysis with a few numerical results . in table
[ tab : debruijn ] , we consider five de bruijn graphs with different values for @xmath114 and @xmath113 , and @xmath11 and @xmath12 i.i.d .
uniform random variables .
table [ tab : debruijn ] shows that while the access costs of all nodes are comparable , the ratio between @xmath280 and the second best case routing cost , over all nodes but the @xmath114 nodes in @xmath117 for which @xmath61 . ]
@xmath281 , is in general significant .
thus , if @xmath282 , there can be an incentive for the nodes with @xmath283 to defect .
for instance , these nodes may leave the network and immediately come back , hoping to be assigned a different identifier @xmath284 and incurring a lower cost .
additional mechanisms , such as enforcing a cost of entry to the network , may be required to prevent such defections .
we graph the access and routing costs for the case @xmath285 , @xmath286 and @xmath287 in figure [ fig : debruijn ] .
we plot the access cost of each node in function of the node identifier in figure [ fig : debruijn](a ) , and the routing cost of each node in function of the node identifier in figure [ fig : debruijn](b ) .
figure [ fig : debruijn ] further illustrates the asymmetry in costs evidenced in table [ tab : debruijn ] , by exhibiting that different nodes have generally different access and routing costs .
therefore , in a de bruijn graph , there is potentially a large number of nodes that can defect , which , in turn , may result in network instability , if defection is characterized by leaving and immediately rejoining the network .
next , we provide an illustration by simulation of the costs in the different geometries .
we choose @xmath276 , for which the results for plaxton trees and chord rings are identical .
we choose @xmath288 for the @xmath113-dimensional tori , and @xmath289 for the other geometries .
we point out that selecting a value for @xmath113 and @xmath114 common to all geometries may inadvertently bias one geometry against another .
we emphasize that we only illustrate a specific example here , without making any general comparison between different dht geometries .
we vary the number of nodes between @xmath290 and @xmath291 , and , for each value of @xmath70 run ten differently seeded simulations , consisting of 100,000 requests each , with @xmath11 and @xmath12 i.i.d .
uniform random variables .
we plot the access and routing costs averaged over all nodes and all requests in figure [ fig : all ] .
the graphs show that our analysis is validated by simulation , and that the star provides a lower average cost than all the other geometries .
in other words , a centralized architecture appears more desirable to the community as a whole than a distributed solution .
however , we stress that we do not consider robustness against attack , fault - tolerance , or potential performance bottlenecks , all being factors that pose practical challenges in a centralized approach , nor do we offer a mechanism creating an incentive to be in the center of the star .
while the cost model proposed here can be used to quantify the cost incurred by adding links for a higher resiliency to failures , we defer that study to future work .
we proposed a model , based on experienced load and node connectivity , for the cost incurred by each peer to participate in a peer - to - peer network .
we argue such a cost model is a useful complement to topological performance metrics @xcite , in that it allows to predict disincentives to collaborate ( peers refusing to serve requests to reduce their cost ) , discover possible network instabilities ( peers leaving and re - joining in hopes of lowering their cost ) , identify hot spots ( peers with high routing load ) , and characterize the efficiency of a network as a whole .
we believe however that this paper raises more questions than it provides answers .
first , we only analyzed a handful of dht routing geometries , and even omitted interesting geometries such as the butterfly @xcite , or geometries based on the xor metric @xcite . applying the proposed cost model to deployed peer - to - peer systems such as gnutella or fasttrack
could yield some insight regarding user behavior .
furthermore , for the mathematical analysis , we used strong assumptions such as identical popularity of all items and uniform spatial distribution of all participants . relaxing these assumptions
is necessary to evaluate the performance of a geometry in a realistic setting .
also , obtaining a meaningful set of values for the parameters @xmath292 for a given class of applications ( e.g. , file sharing between pcs , ad - hoc routing between energy - constrained sensor motes ) remains an open problem .
finally , identifying the minimal amount of knowledge each node should possess to devise a rational strategy , or studying network formation with the proposed cost model are other promising avenues for further research .
k. gummadi , r. gummadi , s. gribble , s. ratnasamy , s. shenker , and i. stoica . the impact of dht routing geometry on resilience and proximity . in _ proceedings of acm sigcomm03
_ , pages 381394 , karlsruhe , germany , august 2003
. m. f. kaashoek and d. karger .
koorde : a simple degree - optimal distributed hash table . in _ proceedings of the 2nd international workshop on peer - to - peer systems ( iptps03 ) _ ,
pages 323336 , berkeley , ca , february 2003 .
d. loguinov , a. kumar , v. rai , and s. ganesh .
graph - theoretic analysis of structured peer - to - peer systems : routing distances and fault resilience . in _ proceedings of acm sigcomm03 _ ,
pages 395406 , karlsruhe , germany , august 2003 .
p. maymounkov and d. mazires .
kademlia : a peer - to - peer information system based on the xor metric . in _ proceedings of the 1st international workshop on peer - to - peer systems ( iptps02 ) _ ,
pages 5365 , cambridge , ma , february 2002 . c. ng , d. parkes , and m. seltzer .
strategyproof computing : systems infrastructures for self - interested parties . in _ proceedings of the 1st workshop on the economics of peer - to - peer systems _ , berkeley , ca , june 2003 . a. rowston and p. druschel .
pastry : scalable , decentralized object location and routing for large scale peer - to - peer systems . in _ proceedings of the 18th ifip / acm international conference on distributed systems platform ( middleware01 ) _ , pages 329350 ,
heidelberg , germany , november 2001 . | in this paper , we model the cost incurred by each peer participating in a peer - to - peer network . such a cost model allows to gauge potential disincentives for peers to collaborate , and provides a measure of the `` total cost '' of a network , which is a possible benchmark to distinguish between proposals .
we characterize the cost imposed on a node as a function of the experienced load and the node connectivity , and show how our model applies to a few proposed routing geometries for distributed hash tables ( dhts ) .
we further outline a number of open questions this research has raised . |
with effective temperatures @xmath5 2200 k , the cool atmospheres of l and t dwarfs generate complex spectra that are rich in molecular features , especially at near - infrared ( nir ) wavelengths where ro - vibrational transitions of many molecules dominate .
fortunately , these cool , very low luminosity objects are also brightest in the nir . until recently ,
most infrared spectroscopic investigations of l and t dwarfs have concentrated on the identification of strong , broad spectral features , useful for the establishment of spectral classification , and have employed resolving powers of @xmath6 ( burgasser et al .
2002 , 2004 , 2006 ; cushing et al . 2003 , 2005 ;
geballe et al . 1996 , 2002 ; jones et al .
1994 ; leggett et al .
2000 , 2001 ; mclean et al . 2000a , 2001 , 2003a ; reid et al .
2000 ; testi et al .
observations with significantly higher spectral resolution are potentially very important because line - blending from molecular transitions is reduced and weak features are resolved . higher resolution spectra are more useful for constraining models of the complex molecular chemistry of brown dwarf atmospheres and for characterizing properties such as gravity and metallicity ( mohanty et al .
for example , less massive brown dwarfs and younger brown dwarfs have smaller surface gravities which results in less pressure broadening and a different line shape .
furthermore , spectra with @xmath7 20,000 ( @xmath515 km s@xmath1 ) are required for the measurement of radial and rotational velocities , and to search for radial velocity variability associated with brown dwarf spectroscopic binaries .
obtaining high signal - to - noise observations with an increase in spectral resolution of a factor of ten is difficult because brown dwarfs are so faint .
basri et al .
( 2000 ) successfully resolved the resonance absorption lines of cs and rb in the far - red visible regime for a sample of m and l dwarfs using the hires echelle spectrograph on the keck 10-m telescope and derived effective temperatures through comparison with available model atmospheres .
reid et al . (
2002 ) also used high - resolution optical echelle spectroscopy to study 39 dwarfs with spectral types between m6.5 and l0.5 .
however , because brown dwarf fluxes are significantly less in visible light , high - resolution observations of fainter l dwarfs and of the even dimmer t dwarfs are not tenable in the optical and require infrared observations . with the advent and development of sensitive large - format ir array detectors , ir spectroscopy with the requisite spectral resolution is now possible ( mclean et al .
1998 , 2000b ) .
in this paper we present the first well - sampled spectral sequence of late m , l and t dwarfs observed at high resolution ( r @xmath0 20,000 ) in the nir .
this work is part of the nirspec brown dwarf spectroscopic survey ( bdss ) being carried out at the keck observatory ; preliminary results were presented in mclean et al .
( 2003b ) .
the goals of the bdss are to obtain a significant sample of nir spectra of low - mass stars and brown dwarfs of differing ages , surface gravities , and metallicities at both medium ( r @xmath0 2,000 ) and high spectral resolution for spectral classification studies and comparisons with model atmospheres .
mclean et al . (
2003a ) , hereafter m03 , describes the lower resolution part of the survey ; spectra from that study are available online.mclean/bdssarchive/ ] here , we investigate the @xmath2-band using ten times higher spectral resolution than in m03 .
the @xmath2-band ( defined as 1.15 - 1.36 @xmath3 m in this paper ) is important because this region contains four strong lines of neutral potassium ( k i ) that are both temperature and gravity - sensitive , and which persist throughout the m , l , and t dwarf sequence . in
2 we describe our observations and data reduction procedures .
3 provides a discussion of the rich , spectral morphology .
in addition to atomic k i , there are lines of al i , fe i , mn i , na i and ti i , and transitions of molecular species such as crh , feh , and h@xmath4o that can provide a unique resource for improving model atmospheres at these low temperatures .
we show that the sudden disappearance of the al i lines critically defines the m - l boundary at these resolutions .
concentrating on the strongly pressure - broadened k i lines , we look for correlations between spectral type and equivalent widths , velocity widths ( fwhm ) , and residual intensity .
the relation between molecular line strengths and spectral type is also investigated .
the effects of rotation , surface gravity and metallicity are explored in 4 .
a summary of the overall results and concluding remarks is given in 5 .
targets for the initial survey , the bdss ( m03 ) , were selected primarily from well - known m dwarfs and from l and t dwarfs identified in the two micron all sky survey ( 2mass ; kirkpatrick et al .
1997 , 1999 , 2000 , 2001 ; burgasser et al .
2000 , 2002 ; reid et al .
2000 ; wilson et al . 2001 ) , augmented with discoveries from the deep near - infrared survey of the southern sky ( denis ; delfosse et al .
1997 , 1999 ) , the sloan digital sky survey ( sdss ; leggett et al .
2000 ; geballe et al .
2002 ) , and other investigations ( becklin & zuckerman 1988 ; ruiz , leggett and allard 1997 ) . to ensure high signal - to - noise spectra for the high - resolution part of the survey ,
a subset of 12 of the brightest objects ( @xmath8 ) , spanning the spectral type range from m6 to t6 , was selected . of these , only 2mass 0140 + 27 was not part of the initial survey .
the m2.5 star g196 - 3a was also observed along with its l2 companion g196 - 3b , both examples of objects substantially younger than 1 gyr ( rebolo et al .
1998 ) and therefore most likely to exhibit gravity effects ( mcgovern et al . 2004 ) .
in addition , the peculiar t6 dwarf 2mass 0937 + 29 ( burgasser et al .
2002 ) was added to the list because of its apparent lack of k i features in lower resolution spectra . another late t dwarf ( 2mass 2356@xmath915 , t5.5 )
was observed after completion of the initial set for comparison to the 2mass 0937 + 29 . although the signal - to - noise ratio was sufficient to establish the presence of k i absorption in this t5.5 dwarf , the fainter magnitude ( @xmath10 ) and stronger 2o absorption made quantitative analysis too difficult so the spectrum is not shown .
table 1 provides the complete list of 16 targets and the observing log .
shorthand names such as 2mass 1507@xmath916 are used in the text for simplicity , but the full designations are listed in table 1 .
two targets were known visual doubles at the time of observing , 2mass 0746 + 20 ( l0.5 ) and denis 0205@xmath911 ( l7 ) , but in neither case did we have sufficient angular resolution to separate the components .
subsequent to making our observations , denis 0205 - 11 was reported as a possible triple brown dwarf system ( bouy et al .
2005 ) based on hubble telescope images .
burgasser et al .
( 2005 ) subsequently found sdss 0423@xmath904 to be double , the average spectral type of t0 being due to an l6 and t2 combination .
even more recently , the binary nature of kelu-1 , a 0@xmath1129 pair , was revealed using laser guide star adaptive optics on the keck telescope ( liu & leggett 2005 ; gelino et al .
2006 ) . again , in neither case were these targets resolved in our nirspec observations .
all of the observations were made using the nirspec cryogenic spectrometer on the keck ii 10-m telescope on mauna kea , hawaii .
detailed descriptions of the design and performance of this ucla - built instrument are given elsewhere ( mclean et al .
1998 ; 2000b ) . for this study ,
nirspec was used in its cross - dispersed echelle mode .
high resolution spectra are dispersed across the 1024@xmath121024 insb detector at 0@xmath11143 per pixel while the spatial scale in the cross - dispersed direction is 0@xmath1119 per pixel .
an independent slit - viewing camera with a scale of 0@xmath1118 per pixel is available for centering and guiding . with the gratings used in nirspec ,
the relationship between the blaze wavelength ( @xmath13 ) and echelle order number ( m ) is @xmath14 = 76.56 @xmath3 m ; together with the free spectral range ( see below ) , this equation gives the order location of a given wavelength .
the spectrometer was set up with the nirspec-3 order - sorting filter and specific echelle and cross - dispersion grating angles to record 11 echelle orders ( @xmath15 = 66 to @xmath15 = 56 ) covering the wavelength range from 1.151.36@xmath3 m , corresponding approximately to the standard @xmath2-band .
the free spectral range ( @xmath16 ) at 1.255 @xmath3 m ( order 61 ) is @xmath0206 , but the effective dispersion is 0.179 / pixel , allowing for only 183 ( 89% ) of this order to be captured by the detector .
in fact , the captured wavelength range varies from 171 ( 94% ) in order 65 to 192 ( 84% ) in order 58 .
thus , because the spectral interval captured by the detector is slightly smaller than the free spectral range in each order , there are small gaps , increasing with wavelength , in the total spectral coverage .
table 2 summarizes the spectral range for each order used in the subsequent analysis . in practice , for an entrance
slit 0@xmath1143 ( 3 pixels ) wide , the final spectral resolution in the reduced data is @xmath17 , ( or 15 km s@xmath1 ) , compared to the theoretical value of @xmath18 .
the average value of one spectral resolution element is @xmath00.625 ( equivalent to 3.6 pixels ) over most of the @xmath2-band region .
spectroscopic observations were made as nodded pairs .
typically , integrations of 600 s each were taken with the object placed at two positions , designated a and b , separated by @xmath07@xmath19 on the @xmath012@xmath19 long entrance slit of nirspec .
shorter exposure times were used for brighter objects .
exposures of 300 s per nod position were used for 2mass 0746 + 2000ab , kelu-1ab and 2mass 1507 - 1627 , 120 s for wolf 359 and 60 s per nod position for g196 - 3a .
total integration times per object ranged from a few minutes to 1.5 hours depending on the apparent @xmath2 magnitude .
signal - to - noise ratios were typically greater than 20 ( 5% ) per resolution element over most orders , and sometimes greater than 100 ( 1% noise ) .
seeing conditions were @xmath20 and therefore a slit width of 0@xmath1143 ( 3 pixels ) was used for all observations , except in the case of 2mass 1507@xmath916 , for which we used a @xmath21 ( 4 pixels ) slit because of poorer seeing .
v stars were observed at an airmass very close to that of the target object to calibrate for absorption features caused by the earth s atmosphere .
arc lamp spectra , taken immediately after each observation , and oh night sky lines in the observed spectra , were used for wavelength and dispersion calibration .
a white - light spectrum and a corresponding dark frame were obtained for flat - fielding .
for the data reduction we used redspec , an idl - based software package developed at ucla for nirspec by s. kim , l. prato , and i. mclean . for each echelle order
, redspec uses the position of the two - dimensional spectra on the nirspec array and the calibration line spectra to construct spatial and spectral maps necessary to transform the raw data onto a uniform grid .
if the target spectrum itself is too faint to provide the spatial rectification , then the a0v star observed with the same set up was used instead .
although four arc lamps are available , it is often the case that there are too few well - distributed lines per echelle order for good spectral rectification .
consequently , oh night sky lines were also used .
the dispersion was more than adequately fit by a second order polynomial of the form @xmath22 where @xmath23 [email protected] / pixel and @xmath25 7 x 10@xmath26 / pixel@xmath27 .
to extract spectra free from atmospheric background and uneven detector response , the difference of an a / b image pair was formed and flat - fielded .
the flat - fielded difference frame was then rectified using the spatial and spectral maps and the raw spectrum produced by summing 5@xmath910 rows from each trace in the rectified image .
the extracted traces ( one positive , one negative ) are subtracted again to produce a positive spectrum with residual night sky emission line features removed , unless a line was saturated . in the @xmath2-band , none of the night sky emission lines are saturated .
a0 v star spectra were reduced in the same way , interpolating over the intrinsic pa@xmath28 hydrogen absorption line at 1.28@xmath3 m in the @xmath2-band spectra .
the raw target spectrum was then divided by the raw a0 v star spectrum to remove telluric features .
the true slope of the target spectrum was restored by multiplication with a blackbody spectrum of t@xmath29 k for an a0 v star ( tokunaga 2000 ) .
finally , the spectra reduced from multiple a / b pairs were averaged together to improve the signal - to - noise ratio .
for each of the 16 targets we have extracted 8 echelle orders ( see table 2 ) yielding a total of 128 spectra . before examining and interpreting the new spectra in detail ,
it is very useful to have a broad overview of the basic spectral features present and an awareness of the general trends that occur in the high - resolution @xmath2-band data as a function of spectral type . a convenient way of doing this is to select a representative source for a few spectral types and present all eight echelle orders on the same plot , thus enabling the entire @xmath2-band to be viewed at a glance .
figures 1@xmath96 show the reduced spectra of wolf 359 ( m6 ) , 2mass 0140 + 27 ( m9 ) , 2mass 0345 + 25 ( l0 ) , 2mass 1507@xmath916 ( l5 ) , sdss 0423@xmath904ab ( t0 ) and 2mass 0559@xmath914 ( t4.5 ) .
the double nature of sdss 0423@xmath904ab means that we do not have a true t0 spectrum , but lower resolution studies ( m03 ) show that j - band spectral variations are relatively weak from l6 to t2 and therefore this binary remains a useful proxy for a t0 dwarf . in these plots ,
echelle orders 58 through 65 are shown together ; the remaining orders at the edges of the @xmath2-band are too contaminated by strong atmospheric absorption to be useful .
for ease of comparison , all spectra are shown in the laboratory reference frame and vacuum wavelengths are used throughout ; radial velocities and searches for radial velocity variations will be reported and discussed in a separate forthcoming paper ( prato et al .
2006 in prep . ) .
each order is normalized to unity at the same wavelength .
comparison of the spectra in these six figures , all of which have excellent signal to noise ratios ( at least 20:1 per pixel ) , shows that the region is densely populated with numerous weak absorption features and a few stronger lines .
we will show that the fine - scale spectral structure is real and repeatable , and that it is mainly attributable to feh or h@xmath4o .
the strongest atomic features are the doublets of k i that occur in orders 61 and 65 .
these lines persist from m6-t4.5 but clearly change their character with spectral type . in later sections the k
i lines will be singled out for closer inspection . for reference ,
table 3 summarizes the main spectral transitions observed in the @xmath2-band over the spectral type range from m6-t4.5 , including the energy levels of the atomic transitions . as shown in figure 1 , the m6 dwarf wolf 359 has at least one distinguishing feature in each order .
atomic lines of al i at 1.31270 and 1.31543 @xmath3 m appear in order 58 .
there is a moderately strong line of mn i at 1.29033 @xmath3 m in order 59 , plus some weaker lines of ti i at 1.28349 and 1.28505 @xmath3 m .
a weak unresolved na i doublet is seen at 1.26826 @xmath3 m in order 60 . the first pair of strong k i lines at 1.24357 and 1.25256 @xmath3
m appears in order 61 .
multiple weak absorption features occur in both orders 62 and 63 , the most notable grouping being the set of lines around 1.222 @xmath3 m .
several of the stronger features have been identified with feh from lower resolution studies ( jones et al .
1996 ; cushing et al . 2003 ; m03 ) .
note however , that a major feh band head at 1.24 @xmath3 m is just off the detector at the short wavelength edge of this order . in order 64
there is a pair of strong fe i lines at 1.18861 and 1.18873 @xmath3 m and another fe i line at 1.19763 @xmath3 m .
order 65 contains the second set of strong k i lines , one at 1.169342 @xmath3 m and the close pair at 1.177286 , 1.177606 @xmath3 m . in general , the lines are relatively sharp and well - resolved .
wolf 359 is a bright source and therefore the signal - to - noise ratio in this spectrum is at least 100:1 .
following these spectral features order by order through figures 1 - 6 reveals certain general trends as a function of spectral type . comparing the m9 object ( figure 2 ) with the m6 source ( figure 1 ) we see that the al i lines at 1.3127 and 1.3154 @xmath3 m in order 58 are somewhat weaker at m9 and then suddenly they are no longer present at l0 ( figure 3 ) , or in any later spectral types ( figures 4 - 6 )
this is an important observation that relates to the m - l transition and we will discuss the al i lines in the next section . throughout order 58
there are other weaker spectral features , the so - called fine - scale spectral structure .
this spectral structure becomes more pronounced at m9 ( figure 2 ) , seems broader in the l0 and l5 objects ( figures 3 and 4 ) , weakens at t0 or more accurately from l6 to t2 ( figure 5 ) and then completely changes character by t4.5 ( figure 6 ) .
the most likely interpretation of this spectral sequence is that it represents changes in the physical structure of these cool atmospheres ( temperature , pressure , chemistry ) . in subsequent sections we compare opacity data for different molecular species to identify the primary absorbers at each spectral class . in order 59 the sharp mn
i line at 1.2903 @xmath3 m seen at m6 and m9 ( figures 1 and 2 ) broadens and disappears after l0 ( figure 3 ) .
the fine - scale spectral structure in this order is dominant until t0 composite type ( figure 5 ) when the spectrum becomes remarkably smooth . here again the high resolution data reveal a striking effect , this time at the transition from l to t dwarfs .
new spectral structure develops in this order between types t0 at t4.5 ( figure 6 ) but , as was the case for order 58 , the pattern is different , indicating different atmospheric conditions .
the weak na i line at 1.2683 @xmath3 m detected in order 60 in the m6 object ( figure 1 ) is already absent in the m9 object ( figure 2 ) .
otherwise , the behavior of the fine - scale structure follows a pattern similar to order 59 becoming remarkably weak at t0 ( figure 5 ) and leading to a smoother appearance for these spectra near the l - t transition .
order 61 contains one of the pairs of strong k i doublets located at 1.2436 and 1.2525 @xmath3 m .
these lines deepen and widen slightly from m6 to m9 , and then become increasingly broader and shallower from l0 to t4.5 .
nirspec spectral order 61 also has many fine - scale features attributable to molecular transitions .
two features , one at 1.24637 and the other at 1.24825 @xmath3 m have been identified previously with feh ( cushing et al . 2003 ) .
these feh lines strengthen slightly from m6 to m9 ( figures 1 and 2 ) , become much broader in the l dwarfs ( figures 3 and 4 ) and then vanish completely in the t dwarfs ( figures 5 and 6 ) to leave , once again , a remarkably smooth continuum between the k i lines . comparing orders 62 and 63 in figures 1 - 6 ,
the known feh features in these spectral bands strengthen from m6 to m9 , broaden markedly at l0 and remain strong and broad through l5 before becoming weaker in the t0 and t4.5 . as with the atomic lines ,
the individual feh lines seem to broaden significantly at the transition from spectral types m9 to l0 .
evidence of weak feh absorption is still present around 1.222 @xmath3 m at spectral type t0 , and possibly even at t4.5 , as shown in figures 5 and 6 , but this molecular species is clearly not dominant in t dwarfs . for the m6 dwarf ( figure 1 ) , order 64 is characterized by a pair of strong lines of fe i at 1.18861 and 1.18873 @xmath3 m that are easily resolved , and another fe i line at 1.19763 @xmath3 m that is blended with feh .
these features remain strong at m9 ( figure 2 ) and persist into the l - dwarf range , becoming broader at l0 ( figure 3 ) , and then undetectable by l5 ( figure 4 ) . from t0 to t4.5 ( figures 5 and 6 ) , order 64 becomes increasingly chopped up by new spectral features , some of which are quite sharp and deep . in section 3.3
we show that these features are caused by absorption by 2o . finally , there is order 65 , which contains the second k i doublet and exhibits some of the largest changes with spectral type .
the slightly weaker k i companion line at 1.17728 @xmath3 m , only 3.3 from the longer wavelength member of the doublet is easily resolved in the m9 object ( figure 2 ) , already blended from line broadening in the l0 ( figure 3 ) , barely discernable at l5 ( figure 4 ) , and completely washed out by line broadening and numerous molecular features at t4.5 ( figure 6 ) .
order 65 , being close to the short wavelength edge of the @xmath2-band where terrestrial water vapor absorption is expected , also contains many strong intrinsic transitions of hot h@xmath4o , for example , the feature at 1.175 @xmath3 m .
figure 7 provides a more detailed view of the al i doublet in order 58 . in this plot ,
the spectra for the m6 , m9 and l0 objects shown in figures 1 - 3 are expanded and overlaid .
evidently , there is significant spectral structure in this part of the spectrum making it difficult to identify a true continuum level .
all three spectral types show consistent features , in particular the wide depression containing the shorter wavelength al i line .
the equivalent width of the al i lines clearly decreases from m6 to m9 , but the change over these three spectral types is only about 25% . because this region of spectrum is contaminated by 2o absorption , it is difficult to obtain accurate equivalent widths .
a pseudo - equivalent width over a 4.2 interval centered on each line was obtained relative to the local continuum in the troughs where the al i lines are found .
for the stronger line of the pair at 1.3127 @xmath3 m , the measured values of equivalent width for the m6 and m9 dwarfs respectively are 420@xmath2420 m and 300@xmath2440 m . at l0
, however , the pseudo - equivalent width of this line is @xmath3040 m .
clearly , at the transition from m9 to l0 , both al i lines vanish completely .
although only three objects bridging this transition were observed at high resolution , the conclusions given here are supported fully by the results of our low resolution bd spectroscopic survey ( m03 ) where two objects of every spectral type from m6 to l5 was included . as shown in table 3 ,
these lines arise from absorption from an energy level at 3.14 ev .
interestingly , the na i line at 1.268 @xmath3 m in order 60 is already absent in the m9 , and careful inspection shows that a somewhat broadened mn i line at 1.290 @xmath3 m in order 59 persists through l0 .
the na i line is excited from a high energy state at 3.6 ev whereas the mn i line comes from a state at only 2.1 ev .
thus , the sequence in which the lines disappear is at least qualitatively consistent with thermal excitation .
but the abrupt loss of al i lines at the classical m - l boundary is too great to be explained by boltzmann factors alone .
for example , for a temperature change from 2850 k from the m6 to about 2400 k for the m9 say , the population of excited atoms in the upper level would drop by 51% and the equivalent width of the line might change from 420 m to about 210 m if all other factors remain the same . from m9 to l0 the change would be a further 13% assuming a change in effective temperature of 150 k. thus , the line should still be measurable with an equivalent width of 140 @xmath2440 m , or about one - third its value at m6 .
yet , both lines disappear abruptly .
it is likely , from the models of lodders ( 2002 ) , that aluminum has been sequestered in compounds such as hibonite ( caal@xmath31o@xmath32 ) and that this abrupt change in absorption line strength is really caused by the sudden depletion of aluminum as an absorber due to a significant change in atmospheric chemistry , rather than simply a drop in effective temperature .
gas temperatures typical of this transition are near 2000 k ( lodders 2002 ) .
it is also curious that the intensity ratio of the components of the al i doublet is closer to 3:2 than the expected 2:1 ratio based on their statistical weights .
however , as shown in figure 7 , this spectral region is highly complex with many overlapping transitions which makes it difficult to determine the true continuum level for each line .
alternatively , the peculiar line ratios may be a non - lte effect , or the result of line blending .
another element that is also important for understanding the temperature structure of these cool , dust - forming objects is iron .
as previously mentioned , order 64 contains a remarkably strong pair of fe i lines at 1.1886 and 1.1976 @xmath3 m .
the shorter wavelength fe i line is a resolved double with a separation of 1.2 in the m6 and m9 objects , but appears as a single broad feature at l0 . by l5 the fe lines
are completely absent .
lower resolution studies ( m03 ) also suggested that fe disappeared around l2 or l3 .
these fe i lines arise from low - lying energy levels near 2.2 ev , and gas phase iron requires temperatures above about 1700 k ( burrows et al .
2001 ) . combining the results that al disappears at l0 and fe
is no longer present by l3 , and using the chemistry temperature scale , suggests that there is about a 300 k temperature change from l0 to l3 , which is shallower but still consistent with the interval of about 140 k per spectral type derived by burgasser et al .
( 2002 ) as well as the effective temperature scale of golimowski et al .
( 2004 ) . as noted by burgasser et al .
( 2002 ) , temperatures derived from condensation chemistry tend to be systematically cooler by about 500 k than those derived from empirical determinations of t@xmath33 using objects with known parallax . these conclusions are not necessarily inconsistent if different spectral features probe a range of optical depths in the atmosphere .
finally , we note the presence of several weak lines of ti i that arise from energy states near 1.4 ev , even lower than those of the strong potassium lines .
a strong ti multiplet at 0.97 has also been seen in the spectra of m dwarfs up to at least m9 ( cushing et al .
unfortunately , these weak lines are impossible to trace after m9 . the astronomical @xmath2-band is bounded by 2o absorption bands from terrestrial water vapor .
it is therefore no surprise that high - temperature 2o ( hot steam ) transitions intrinsic to m , l and t dwarfs encroach far into the @xmath2-band from both the short and long wavelength ends .
these so - called infrared water bands are difficult to model because millions of transitions are needed ( partridge & schwenke 1997 ) .
typically , models over - estimate the depth of the infrared water bands .
in addition to 2o , some of the stronger non - atomic transition features are known to be attributable to feh from lower resolution studies ( jones et al.1996 ; mclean et al .
these features occur at 1.2091 , 1.2113 , 1.2135 and 1.2221 @xmath3 m . cushing et al .
( 2003 ) verified the features at 1.1939 and 1.2389 @xmath3 m as the 01 and 12 band heads of the f@xmath34
x@xmath34 system of feh , and attributed a blended feature described by mclean et al .
( 2000 ) at 1.2221 @xmath3 m as the f@xmath35 x@xmath35
q - branch .
these authors also listed 24 other relatively strong features lying within the @xmath2-band . in the cushing et al .
list , no feh features were identified in the wavelength interval covered by our order 65 , which includes the strong shorter - wavelength doublet of k i , and only one feature ( at 1.2464 @xmath3 m ) was tabulated for order 61 , where the other k i doublet dominates . to identify many more of the complex fine - scale features seen in the spectral sequences of figures 16
, we analyzed opacity ( cross - section ) data for both feh and h@xmath4o , ( r. freedman 2003 , private communication ) and utilized the feh line list and transition catalog by phillips et al .
we are also grateful to adam burrows who provided crh opacity data ( burrows et al .
2002 ) and linda brown who provided new opacity calculations for 4 in the @xmath2-band ( l. brown 2004 , private communication ) .
figure 8 is a detailed view of order 62 ( 1.221 - 1.239 @xmath3 m ) for the m9 object 2mass 0140 + 27 .
superimposed on the m9 dwarf spectrum is a normalized and scaled feh opacity plot for a temperature of 2000 k and a pressure of 1 bar .
we use this plot to assist in identifying spectral features .
this interesting region of the @xmath2-band contains a feature which is seen in lower resolution spectra as a broad , flat - bottomed line at 1.2221 @xmath3 m ( mclean et al 2000a ) .
each transition in the opacity data can be correlated with either 01 or 12 transitions of the f@xmath34
x@xmath34 system tabulated by phillips et al .
( 1987 ) . as shown in figure 8 ,
the feature observed at 1.2221 @xmath3 m and attributed to the q - branch by cushing et al .
( 2003 ) , is actually a composite of four q - branch and three p - branch transitions plus one r - branch transition of feh .
thus , at our higher resolution , the broad flat - bottomed feature seen in lower resolution spectra is completely resolved into eight separate transitions .
these transitions are the following : q(0 - 1 ) at 1.22137 and 1.221383 @xmath3 m blended , 1.221934 , 1.222504 and 1.22305 @xmath3 m ; p(0 - 1 ) at 1.22166 , 1.22219 and 1.22244 @xmath3 m ; r(0 - 1 ) at 1.22218 @xmath3 m .
r - branch transitions tend to correspond closely to p - branch lines .
for example , the p(0 - 1 ) 1.22219 @xmath3 m line is only 0.1 from the r(0 - 1 ) transition just given .
also , the q - branch line at 1.222504 @xmath3 m is blended with the nearby p - branch transition at 1.22244 @xmath3 m .
the remainder of order 62 is dominated by 5 r - branch lines and about 30 p - branch transitions ; all are members of the 01 band of f@xmath34 x@xmath34 system . in this one order alone ,
there are now more identified feh transitions than previously known for the entire @xmath2-band .
summing up across all nirspec orders , we identify over 200 matches to the feh opacity data base and the tables by phillips et al .
there is little doubt that feh is a dominant molecular absorber in late m and early l spectral types .
interestingly , there are several features that can not be identified with feh transitions in the given opacity tables .
for example , a sharp line at 1.2227 @xmath3 m contaminates the q - branch feature , and there are other isolated groups near 1.225 , 1.228 , 1.231 and 1.234 @xmath3 m which also are not attributable to feh . because these features have a dependence on spectral type that is similar to feh , they may be unknown feh transitions or transitions of crh . to explore the latter possibility we also compared our spectra to crh opacity calculations by burrows et al .
although the crh opacity data had a lower resolution than our spectra , there was good coincidence for the features at 1.225 , 1.228 and 1.231 @xmath3 m , but the features near 1.234 @xmath3 m could not be identified with crh .
although transitions of crh contaminate those of feh , the strongest fine - scale features in this part of the @xmath2-band are attributable to feh rather than crh .
stronger crh bands occur at shorter wavelengths ( kirkpatrick et al .
1999 ) .
we also investigated the fine - scale structure in orders 61 and 65 which contain the important k i doublets .
figure 9 provides a plot of the feh opacity ( at 2000 k and 1 bar pressure ) across order 61 , normalized for convenience , smoothed to match the resolution of our spectra , and over - plotted with the spectra of our m9 and l5 dwarfs . essentially every feature can be traced to feh , with a few notable exceptions .
for example , absorption features at 1.2448 and 1.2458 @xmath3 m are clearly associated with feh , but two significant absorptions bands in between these limits do not correlate with feh .
note also that the l5 spectral features are broader than the equivalent features in the m9 object . as shown in figures 1 - 6 ,
there is a general increase in the strength of feh from mid - m until about l4 and then a decrease in feh features towards later spectral types .
figure 10 shows a plot for order 61 in which the l5 and the t4.5 dwarfs in our sample are compared . in this figure
h@xmath4o opacity curves ( at 1000 k and 1 bar ) are over - plotted .
the t4.5 spectrum lacks feh , and its shallow depressions and small features are consistent with the h@xmath4o opacity .
the feh feature near 1.245 @xmath3 m appears strongest in the early to mid ls , while the 2o transition at 1.1752 @xmath3 m seems to gain in strength systematically toward later spectral types . in order 65 ,
no transitions for feh are listed in the opacity files at the shorter wavelengths .
therefore , in figure 11 , we plot the h@xmath4o opacity ( at 1000 k and 1 bar ) for order 65 , and over - plot with the spectrum of the l3.5 and t4.5 from our sample .
this pair of spectral types provides distinct morphological samples .
most of the features in the t4.5 dwarf evidently correspond to h@xmath4o .
there is a particularly strong 2o feature at 1.1752 @xmath3 m in the t4.5 dwarf , which is also present in the l3.5 object at a weaker level .
in fact , weak h@xmath4o absorption is already present even in late m objects .
we also looked for transitions associated with the @xmath36 bands of tio ( galehouse 1980 ; phillips 1973 ) .
these bands have been detected in young , low - gravity sub - stellar objects by mcgovern et al .
( 2004 ) , but are not usually apparent in low resolution spectra of older field dwarfs . a model spectrum kindly provided by d. saumon shows the location of numerous @xmath36-band transitions in order 61 , with noticeable band heads for the @xmath37=-1 , ( 0 - 1 ) and ( 1 - 2 ) transitions at 1.25644 and 1.24239 @xmath3 m respectively .
absorption features do occur at these wavelengths from m6-l5 , but these also coincide with feh features and are less likely to be tio because none of the other tio transitions of the @xmath36 band are seen . finally , using new opacity data for 4 ( l. brown 2004 , private communication ) , we searched for evidence of methane transitions in the t dwarfs .
the strongest features should occur in order 65 which also contains the very broad shorter - wavelength k i doublet .
unfortunately , this spectral region is already heavily blanketed by 2o absorption . even in the high signal - to - noise spectra of 2mass 0559 - 1404 ( t4.5 ) , there are no transitions that can be uniquely attributed to 4 . because the k i doublets dominate orders 61 and 65 , and because these strong lines clearly persist throughout almost the entire m , l and t range , we plot 12 of the objects from table 1 together in figures 12 and 13 to provide a more continuous sequence for these features .
for this figure , the l4 gd165b has been left out because the l3.5 2mass 0036 + 18 , with better signal - to - noise ratio , has been included ; gd165b will be discussed separately .
the other two objects omitted from these plots are the peculiar l2 , g196 - 3b , and the peculiar t6 dwarf , 2mass 0937 + 29 ; again , these objects are discussed separately later .
each spectrum in figures 12 and 13 has been normalized to a continuum value of one at a common wavelength and shifted for clarity by an additive constant along the y - axis .
the spectra have been ordered according to their published classification ( optical types for l - dwarfs : kirkpatrick et al .
1999 , 2000 , 2001 ; nir for t dwarfs : burgasser et al . 2006 ) and aligned at the k i rest wavelengths in vacuum . as mentioned already , radial velocity determinations will be presented in a separate paper ( prato et al .
2006 , in prep . ) .
as given in table 3 , the shorter wavelength k i lines ( in order 65 ) correspond to the multiplet designation 4p @xmath38p@xmath39@xmath93d @xmath38d , and the order 61 pair to the 4p @xmath38p@xmath39@xmath95s @xmath38s multiplet .
all are transitions between states at [email protected] ev .
the pair of k i lines in order 65 have almost equal intensity at the line center , whereas in order 61 , the 1.2436 @xmath3 m line is always weaker than the 1.2526 @xmath3 m line . for both k i doublets ,
the line ratios are similar throughout the spectral sequence .
one of the most prominent results , evident in both figures 12 and 13 , is the significant k i line broadening in the later type objects . primarily because of its higher temperature , the m2.5 dwarf manifests very narrow lines with almost no wings .
there is a slight contamination of the shortest wavelength k i line in order 65 from an fe i line for spectral types before l4 .
for the m2.5 , m6 and m9 dwarfs , separate peaks are clearly discernable for the longer wavelength line of the k i doublet in order 65 at 1.17761 @xmath3 m and the secondary k i line at 1.17728 @xmath3 m , but for later spectral types the 1.17761 @xmath3 m lines are heavily blended and not discernable as separate features . in the earlier type objects , the pair of lines near 1.1786 @xmath3 m in order 65 are attributable to ti and fe
to quantify the changes in the k i lines illustrated in figures 12 and 13 as a function of spectral type we calculate three quantities , a line depth , a line strength ( equivalent width ) and a line width .
line depth is defined in terms of the measured flux ( @xmath40 ) at the line center compared to the average value for the continuum . for convenience
we construct the line depth as @xmath41 at the line center ; a weak line has almost the same flux in the line as in the continuum and its line depth measure is therefore almost zero .
deep lines have a depth index approaching unity . determining the continuum or pseudo - continuum level introduces the largest uncertainty into this ratio .
the continuum is estimated by fitting a sloping line across the feature between the highest points on either side lying within @xmath24 50 of the line center .
multiple trials obtained by varying the positions by a few angstroms provide a mean value and an estimate of the uncertainty from fluctuations in the continuum level .
equivalent width ( in ) is also determined by fitting a continuum line between two points on either side of the line , summing the residual intensities and multiplying by the width of a pixel . the same range and method is used as for the line depth .
line width ( in km s@xmath1 ) is defined as the full width of the line at the intensity level halfway between the apparent local continuum , defined above for the line depth and equivalent width , and the minimum line intensity at the line center ( fwhm ) .
several sources contribute to the uncertainty in the derived quantities .
the continuum level is difficult to identify , photon noise reduces the signal - to - noise ratio for the fainter sources , and contamination by numerous weak features is dependent on spectral type .
molecular line contamination results in additional fluctuations in the measured k i quantities and gives larger uncertainties , despite the good signal - to - noise ratio of most of the spectra .
our final plotted uncertainties are based on repetitive trial fits and photon noise estimates .
table 4 and table 5 provide the measured quantities for the 1.2525 and 1.2436 @xmath3
m k i lines respectively .
this pair of k i lines ( order 61 ) is preferred because the shorter wavelength doublet is too heavily contaminated by h@xmath4o absorption , and one of the k i lines is itself a blend ( 3 ) .
figures 14 , 15 , and 16 show , respectively , the line depth , equivalent width and velocity line width as a function of spectral type for the 1.2525 @xmath3 m line .
several of the trends mentioned in the previous section are confirmed .
the k i line depth increases from a weak line in early m dwarfs to a strong deep line in the early / mid l dwarfs before decaying towards later spectral types .
interestingly , there is considerable scatter in line depth among similar spectral types . the equivalent width ( designated by w )
is better behaved .
variations in defining a consistent continuum may have a large effect on the line depth quantity , whereas the growth of the line wings may overcome such variations when calculating w and fwhm .
equivalent width increases to a broad maximum around l2-l4 and then decreases again , remaining essentially constant from l5 - t4.5 .
these results are consistent with our more extensive lower resolution studies ( m03 ) . in contrast
, the velocity line width of the 1.2525 @xmath3 m k i line ( figure 16 ) increases almost monotonically and steeply with spectral type .
the change in velocity width is dramatic , ranging from @xmath0 60 km s@xmath1 at m6 to almost 500 km s@xmath1 at t4.5 .
this change does not result from rotational broadening because the line develops extensive wings characteristic of pressure broadening as discussed in the next section .
one l - dwarf appears to stand out from the others in this sequence and is indicated by a star symbol .
this object is kelu-1ab , and will be discussed separately in the next section .
the behavior of the 1.2436 @xmath3 m k i line ( table 5 ) is similar .
the changing behavior of the k i lines can be explained as follows .
as the temperature decreases towards later spectral types , the transition levels become less populated and the line should become weaker .
however , with decreasing temperature , dust grains settle below the photosphere and the transparency of the atmosphere at wavelengths where gas opacity is weak improves .
thus , line formation can be observed at much greater depths and pressures ( saumon et al .
higher pressures cause the development of the broad line wings through collision broadening , primarily with h@xmath4 molecules ( e.g. burrows & voloyubev 2003 ) , and hence an increase in the fwhm of the lines .
the greater column depth of k i in these transparent regions also serves to increase their equivalent widths .
the behavior has been observed with the k i and na i fundamental doublets ( reid et al . 2000 ) , and is one of the unique properties of brown dwarf atmospheres . the pressure broadening of the k i lines at j - band follow quantitative expectations as well .
the observed ratio between the k i line widths in the t4.5 and l5 spectra is about 2:1 . as the t@xmath33 decreases , the pressure at @xmath42=2/3 increases as already described . according to models of cloudy atmospheres , kindly provided by d. saumon ( 2003 ; private communication ) , the change in the photospheric pressure at these wavelengths is a factor of 5 from 1600 k to 1000 k. these effective temperatures were chosen to illustrate a prediction of the models over the typical range of l / t spectral types .
this implies that the average kinetic energy of the molecules of h@xmath4 is five times greater , or that the average speeds are @xmath02.24 times larger .
pressure broadened line widths should be proportional to the average velocity of the disturbing atoms . thus , the observed pressure broadening from l5 to t4.5 is consistent with the expected change .
in contrast to the k i lines , figures 17 and 18 show the changing trends with spectral type for two representative feh and h@xmath4o features that fall within the orders containing the k i doublets ; data are provided in table 6 .
for feh we plot the line depth index of the 1.245 @xmath3 m feature in order 61 , and for h@xmath4o we use the strong 1.175 @xmath3 m feature in order 65 .
the feh line depth reaches a peak around l3-l4 , whereas the h@xmath4o line depth increases more - or - less monotonically from m6 to t4.5 .
once again kelu-1ab stands out , this time because the feh bands are not as deep as expected for an l2 . in the previous section we found that the k i line width ( fwhm ) and the 2o line depth provided the best correlations with spectral type .
the exception to the fwhm correlation is kelu-1ab , which _ is _ known to be a rapid rotator ( basri et al .
2000 ) , and which was recently discovered to be a @xmath003 visual binary ( liu & leggett 2005 ; gelino et al . 2006 ) , although the orbital velocity of 3 km s@xmath1 is only a fraction of the @xmath43 .
as will be shown below , the k i lines are not really suitable for @xmath43 studies .
interestingly , kelu-1ab is not abnormal in its 2o ratio .
the one point in the 2o plot of figure 18 that does seem slightly high is due in fact to gd165b ( l4 ) ; this result is consistent with m03 . as the companion to a white dwarf , gd165b is most likely to be an old l dwarf with a higher gravity .
burgasser et al .
2003 has shown that late - type subdwarfs , which are also presumably older and have higher surface gravities , can show stronger 2o compared to equivalently classified disk dwarfs .
the presence of a binary companion can impact the observed line properties discussed above in several ways . for example , a spectroscopic companion only partially resolved in velocity space might produce a spectrum with anomalously broadened lines at certain epochs .
a spectroscopic companion of similar mass to the primary object should not only manifest itself in broadened lines , observed at a favorable phase , but also should almost double the expected brightness of the system .
fainter companions , although interesting in their own right , should not significantly impact the primary dwarf spectrum if the mass ratio , @xmath44 , is less than @xmath00.5 ( prato et al .
visual companions at separations too great to noticeably effect the spectrum will also cause the observed spectrum to appear brighter than expected for the distance and age of the target system . four targets in table 1 are visual binaries : 2mass 0746 + 20 ( l0.5 ) , kelu-1 ( l2 ) , denis 0205@xmath911 ( l7 ) and sdss 0423@xmath904 ( t0 ) .
signatures of velocity variations in these spectra will be addressed in a forthcoming publication .
spectral features in kelu-1ab appear to be significantly broader than other field dwarfs of similar temperature by a factor of 2 - 3 , consistent with its high rotational velocity of @xmath45 km / s ( basri et al .
2000 ) . in figure 19
we compare the l2 dwarf , kelu-1ab , with the l0.5 and l3.5 dwarfs in our sample . because no feature is completely free of blending at this spectral resolution and wavelength throughout the late m to t sequence , measuring precise rotational velocities is challenging . indeed , because of the increasing complexity of molecular features with cooler effective temperatures , most if not all of the lines in the mid - l to t objects may be blends .
however , the wealth of fine - scale spectral structure attributable to feh and 2o throughout the @xmath2-band does provide some possibility of detecting trends in @xmath43 , especially because pressure broadening should not be a large effect for molecular transitions . examining the rich feh structure in order 61 ( figure 12 ) , as well as orders 62 and 63 ( figures 1 - 4 ) where there are no strong alkali lines
, it is clear that there is a sudden broadening of the feh lines at the transition from m9 to l0 , and that all of the feh features remain broad after that transition .
in fact , if the m9 spectra in orders 62 and 63 are smoothed by a factor of 3 using a simple moving average , the resulting spectrum is a very good match for the observed l0 spectrum .
this sudden broadening could be caused by either increased pressure broadening or an increase in rotational velocities .
examination again of figure 13 shows that many of the 2o spectral features that develop in the t dwarfs are relatively sharp , indicating that they are not formed at the same depth and pressure as the potassium line wings .
therefore , in the coolest objects , we are seeing evidence of vertical structure in the atmosphere .
like the feh lines in the late m and early l dwarfs , the 2o features provide a better estimate of rotational velocities than the alkali lines .
we can de - convolve the instrumental profile by using the observed widths of unresolved oh night sky lines and the very narrow features seen in the m2.5 object .
if we interpret the change in line width at the m9-l0 boundary as a rotational velocity , then after removal of the instrumental profile , all of the early l dwarfs appear to be rapid rotators with @xmath43 @xmath46 30 km s@xmath1 .
rotational velocities have been reported for some of the objects in our target list .
basri et al .
( 2000 ) give @xmath43 = 60@xmath245 km s@xmath1 for kelu-1ab and 22@xmath245 km s@xmath1 for denis 0205@xmath911ab and @xmath53 km s@xmath1 for wolf 359 ( gl 406 ) .
the l3.5 dwarf 2mass 0036 + 18 is reported to have a @xmath43 of @xmath015 km s@xmath1 by schweitzer et al .
several of our targets were also observed by zapatero osorio et al .
( 2006 ) at high resolution , enabling them to measure the @xmath43 of 2mass 0036 + 18 , sdss 1254 - 01 , and 2mass 0559 - 14 .
we note with interest that the rotational velocities they measure for these sources , as well as for another dozen dwarfs , are all very close to 30 km s@xmath1 within the uncertainties .
we will analyze in detail the rotational velocities of our sample in a future paper ; however , we comment here that our figures 12 and 13 suggest that the t0 and t4.5 dwarfs in our sample , sdss 0423 - 04 and 2mass 0559 - 14 , appear to reflect larger values of @xmath43 .
if there is a tendency towards more rapid rotation among the t dwarfs , it is not apparent in the measurements of zapatero osorio et al .
( 2006 ) .
clarke et al .
( 2003 ) report time resolved spectroscopy of kelu-1ab used to search for variability in photospheric molecular species .
they confirm the short rotation period of the system and find variable h@xmath47 profiles .
no evidence for a spectroscopic companion was detected and it appears to be a normal l dwarf apart from the high rotation rate ( clarke et al .
the recently discovered 0.29binary contributes only about 3 km s@xmath1 in velocity shift and therefore has no impact on the measured @xmath43 .
however , it is interesting to ask whether or not both components have the same @xmath43 . in the binary system gl569b ( zapatero osorio et al .
2004 ) , line broadening is 2 - 3 times greater in one component than the other , although this has been attributed to a nested spectroscopic binary ( simon et al . 2006 ) .
it is possible that the higher rotation rate of kelu-1ab is the result of age .
a younger l dwarf just past the accretion phase might have a higher rotational velocity . on the other hand , as discussed in the next section , the young l2 dwarf g196 - 3b does not have such a rapid rotation , or is being viewed at a special angle .
as shown by models ( e.g. burrows et al .
2001 ) , the youngest brown dwarfs are expected to be hotter and more luminous .
as a sub - stellar mass object cools over the first 100 myr , its radius will contract by a factor of 2 - 10 , depending on initial mass , and then remain almost constant at a value close to that of jupiter s radius .
consequently , a very young low - mass brown dwarf could be observed with a much earlier spectral type than it will have when older than 1 gyr , and its surface gravity ( @xmath48 ) will be less than that of an old brown dwarf with the same observed spectral type .
a lower surface gravity implies less pressure broadening ( p@xmath0g / k@xmath49 , where k@xmath49 is the rosseland mean opacity ) and therefore one expects the k i lines to be narrower in such objects .
the effect of surface gravity on the k i lines is significant and has already been observed at lower resolution ( mcgovern et al .
2004 ; kirkpatrick et al .
2006 ) . high resolution infrared spectroscopy of brown dwarfs provides a means of measuring surface gravities and hence estimating mass .
because of their greater column depth the k i lines are much more sensitive to pressure and therefore to surface gravity , whereas the weaker feh and 2o lines formed high in the atmosphere provide a better probe of rotational broadening .
g196 - 3b , the companion to the m2.5 g196 - 3a , is classified as an l2 dwarf ( kirkpatrick et al .
the g196 - 3ab system is believed to be @xmath020 - 300 myr old , rather than the 1 gyr thought to be typical of field l dwarfs ( rebolo et al . 1998 ) .
figure 20 shows the order 61 spectrum of g196 - 3b plotted together with the l0.5 dwarf ( 2mass 0746 + 20ab ) and the l3.5 dwarf ( 2mass 0036 + 18 ) . clearly , although the signal - to - noise ratio is poorer , all of the spectral features of g196 - 3b are much narrower than expected for an l2 dwarf . using high resolution far - red optical spectra , basri et al .
( 2000 ) measured 10 km s@xmath1 for the rotational velocity of g196 - 3b .
the infrared k i lines have cores only half as wide as those of the l0.5 and l3.5 dwarfs , and much less pronounced pressure - broadened line wings .
this behavior suggests that the line profiles result primarily from lower surface gravity , consistent with the age of this object .
burgasser et al .
( 2002 ) noted that the t6 dwarf 2mass 0937 + 29 is peculiar because , despite having characteristics common to its spectral class , it appears to have no k i lines in a low - resolution @xmath2-band spectrum .
if the k i lines were shallow and very broad , or sharp and weak , then they might escape detection at low spectral resolution .
as shown in figure 21 however , where we compare our high resolution spectrum of this t6 to that of the t4.5 dwarf 2mass 0559@xmath914 , the k i lines are neither weak nor broad , they are completely absent . given the good signal - to - noise ratio of this high resolution spectrum , there is no possibility that the k i lines were simply too weak to detect .
disappearance of the k i lines is unexpected because other t6 dwarfs still show these features ( burgasser et al .
2002 , m03 ) .
in fact , we have observed the t5.5 2mass 2356@xmath915 which exhibits broad k i lines similar to those of a t5 .
strong 2o absorption features are clearly present in order 65 in the t6 .
these features show good correlation with the same features in the t4.5 but there is clear variation in the individual features .
one possible reason for the absence of the k i lines in 2mass 0937 + 29 is that this object has a different metallicity from other field dwarfs .
alternatively , the lines could be veiled by dust , but the trend at these spectral types is for dust to settle below the photosphere . in any case , there are other features in 2mass 0937 + 29 that do not appear to be veiled .
2mass 0937 + 29 was also classified as peculiar because of its extremely blue nir colors and because it also has a very red optical spectrum for its type , as determined from ch@xmath50 and 2o strengths .
this combination of attributes led burgasser et al .
( 2002 ) to propose that 2mass 0937 + 29 is an old , metal - poor brown dwarf in which enhanced collision - induced h@xmath4 absorption in the @xmath51-band gives the unusual blue nir color .
if this is an old t dwarf then it is also probably a fairly high - mass brown dwarf that has cooled to the temperature of a t6 , and consequently it has a higher gravity , higher pressure atmosphere .
the higher gravity could also result in increased pressure broadening of the 2o lines in 2mass 0937 + 29 which would explain why these features seem muted and broader . in a recent study , burgasser , burrows and kirkpatrick ( 2006 )
have found that the high surface gravity can result in a cooler effective temperature than equivalent t6 dwarfs .
it is known ( e.g. , m03 ) that k i lines disappear around t7 or t8 , hence it is possible that 2mass 0937 + 29 is depleted in k i because of its low temperature rather than a low metallicity .
because the high - resolution spectra presented here show no residual trace of the k i lines , and because the 2o lines are unusually broad , we contend that 2mass 0937 + 29 must be cooler and/or more metal poor than a normal t6 dwarf . +
using a sequence of high - resolution infrared @xmath2-band spectra ( r@xmath020,000 or 15 km s@xmath1 ) obtained with nirspec on the keck ii telescope , we have studied the spectral morphology of objects from m6 to t6 .
the principal results of the survey are as follows : \(1 ) hundreds of small - scale spectral features are identified to be either feh or h@xmath4o absorption features . over ten times as many feh features as previously identified in brown dwarf spectra are now confirmed .
a few features of crh are also identified , but no convincing transitions of tio or 4 at j - band are found in this sample .
feh features attain maximum strength in the mid - l dwarfs , while h@xmath4o absorption becomes steadily stronger towards later spectral types .
\(2 ) feh and 2o line widths are typically @xmath020 km s@xmath1 for the late m dwarfs , but broaden abruptly by over a factor of two at the m to l transition .
we interpret this effect as evidence for increased rotational velocities in l dwarfs .
\(3 ) the doublet of al i at 1.31270 and 1.31543 @xmath3 m is shown to be very sensitive to spectral type .
this doublet weakens through m9 and then vanishes abruptly between m9 and l0 .
we suggest that this sudden disappearance is more a consequence of a transition in atmospheric chemistry than a simple decrease in atomic population levels resulting from a change in effective temperature .
that is , the aluminum atoms are suddenly bound up into molecules and grains .
\(4 ) the wings and line widths of the k i doublets at 1.16934 , 1.17761 @xmath3 m and at 1.24357 , 1.25256 @xmath3 m increase systematically , while line depth weakens with later spectral type .
the equivalent width ( w ) of the k i features reaches a maximum in the mid - l dwarfs , decreases and then remains almost constant through t4.5 .
the k i line profiles begin to exhibit pressure broadened wings as early as late m. line widths ( fwhm ) range from @xmath050 km s@xmath1 at m5 to almost 500 km s@xmath1 at t4.5 .
this effect is consistent with the much greater depth that is probed in cool t dwarf atmospheres at @xmath2-band .
\(5 ) as shown in figure 12 , a characteristic of the transition from l to t dwarfs is the decay of feh spectral structure , resulting in a smooth spectrum at high resolution for late l dwarfs and early t dwarfs , before 2o dominant spectral structure develops .
\(8 ) finally , the peculiar t6 dwarf 2mass 0937 + 29 displays a complete absence of potassium , in contrast to other late t objects .
we interpret this as either a metallicity effect ( depletion of k atoms ) or a cooler t@xmath33 for this high surface gravity object .
although the sample of objects of different spectral types is relatively small , these high - resolution , high signal - to - noise spectra of m , l and t dwarfs should provide an important benchmark for the detailed development and improvement of model atmospheres .
+ the authors wish to thank the staff of the keck observatory for their outstanding support .
i.s.m . acknowledges the staff of the ucla infrared laboratory and colleagues james graham ( ucb ) , james larkin ( ucla ) and eric becklin ( ucla ) for their support throughout the development of the nirspec instrument .
we thank adam burrows , katharina lodders , linda brown , didier saumon , richard freedman , travis barman and mark marley for helpful discussions and opacity data .
finally , we thank the anonymous referee for a careful and complete critique of this paper . work by s.s.k .
was supported by the astrophysical research center for the structure and evolution of the cosmos ( arcsec ) of korea science and engineering foundation through the science research center ( src ) program .
a.j.b . acknowledges support by nasa through hubble fellowship grant hst - hf-01137.01 awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . ,
for nasa , under contract nas 5 - 26555 .
this research has made use of the nasa / ipac infrared science archive , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
this publication makes use of data from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center , funded by the national aeronautics and space administration and the national science foundation .
finally , the authors wish to extend special thanks to those of hawaiian ancestry on whose sacred mountain we are privileged to be guests .
auman , j. , jr . , 1967 ,
apjs , 14 , 171 basri , g. 2000 , , 38 , 485 basri , g. , mohanty , s. , allard , f. , hauschildt , p. h. , delfosse , x. , martn , e. l. , forveille , t. , & goldman , b 2000 , apj , 538 , 363 becklin , e.e . , &
zuckerman , b. 1988 , nature , 336 , 656 bouy , h. , martn , e. l. , brandner , w. , & bouvier , j. 2005 , aj , 129 , 511 burgasser , a. j. , et al .
2000 , aj , 120 , 1100 burgasser , a. j. , et al .
2002 , apj , 564 , 421 burgasser , a. j. , mcelwain , m. w. , kirkpatrick , j. d. , cruz , k. l. , tinney , c. g. & reid , i. n. 2004 , aj , 127 , 2856 burgasser , a. j. , reid , i. n. , leggett , s. k. , kirkpatrick , j. d. , liebert , j. & burrows , a. 2005 , apj , 634 , l177 burgasser , a. j. , geballe , t. r. , leggett , s. k. , kirkpatrick , j. d. , & golimowski , d. a. 2006 , apj , 637 , 1067 burgasser , a. j. , burrows , a. , & kirkpatrick , j. d. 2006 , apj , 639 , 1095 burrows , a. , hubbard , w. b. , lunine , j. i. , & liebert , j. 2001 , rev .
phys . , 73 , 719 burrows , a. , ram , r.s . ,
bernath , p. , sharp , c.m .
& milsom , j.a .
2002 , apj , 577 , 986 burrows , a. & voloyubev , m. 2003 , apj , 583 , 985 .
clarke et al .
2003 , mnras , 341 , 239 cushing , m. c. , rayner , j. t. , davis , s. p. , & vacca , w. d. 2003 , apj , 582 , 1066 cushing , m. c. , rayner , j. t. , & vacca , w. d. 2005 , apj , 623 , 1115 delfosse , x. , et al .
1997 , a & a , 327 , l25 delfosse , x. , tinney , c. g. , forveille , t. , epchtein , n. , borsenberger , j. , foque , p. , kimeswenger , s. , tiphene , d. , 1999 , a & as , 135 , 41 galehouse , d. c. , davis , s. p. , & brault , j. w. 1980 , apjs , 42 , 241 geballe , t. r. , kulkarni , s. r. , woodward , c. e. & sloan , g. c. 1996 , apj , 467 , l101 geballe , t. r. , et al .
2002 , apj , 564 , 466 gelino , c. r. , kulkarni , s. r. , & stephens , d. c. 2006 , pasp , 118 , 611 golimowski , d. a. , et al .
2004 , aj , 127 , 3516 jones , h. r. a. , longmore , a. j. , jameson , r. f. , & mountain , c. m. 1994 , mnras , 267 , 413 jones , h. r. a. , longmore , a. j. , allard , f. , & hauschildt , p. h. 1996 , mnras , 280 , 77 kirkpatrick , j. d. , beichman , c. a. , & skrutskie , m. f. 1997 , apj , 476 , 311 kirkpatrick , j. d. , et al . 1999 ,
apj , 519 , 802 kirkpatrick , j. d. , et al .
2000 , aj , 120 , 447 kirkpatrick , j. d. , dahn , c. c. , monet , d. g. , reid , i. n. , gizis , j. e. , liebert , j. , & burgasser , a. j. 2001 , aj , 121 , 3235 kirkpatrick , j. d. , barman , t. s. , burgasser , a. j. , mcgovern , m. r. , mclean , i. s. , tinney , c. g. , & lowrance , p. j. 2006 , apj , 639 , 1120 leggett , s. k. , et al .
2000 , apj , 536 , l35 leggett , s. k. , allard , f. , geballe , t. , hauschildt , p. h. , & schweitzer , a. 2001 , apj , 548 , 908 liu , m.c . &
leggett , s.k .
, 2005 , apj , 634 , 616 lodders , k. 2002 , apj , 577 , 974 .
mcgovern , m. r. , kirkpatrick , j. d. , mclean , i. s. , burgasser , a. j. , prato , l. , & lowrance , p. j. 2004 , apj , 600 , 1020 mclean , i. s. , et al .
1998 , proc .
spie , 3354 , 566 mclean , i. s. , et al .
2000a , apj , 533 , l45 mclean , i. s. , graham , j. r. , becklin , e. e. , figer , d. f. , larkin , j. e. , levenson , n. a. , & teplitz , h. i .. 2000b , proc .
spie , 4008 , 1048 mclean , i. s. , prato , l. , kim , s. s. , wilcox , m. k. , kirkpatrick , j. d. , & burgasser , a. 2001 , apj , 561 , l115 mclean , i. s. , mcgovern , m. r. , burgasser , a. j. , kirkpatrick , j. d. , prato , l. , kim , s. s. 2003a , apj , 596 , 561 ( m03 ) mclean , i. s. , mcgovern , m. r. , prato , l. , burgasser , a. j. , & kirkpatrick , j. d. 2003b , in brown dwarfs , iau symp .
e. martn , asp conference series .
mohanty , s. , basri , g. , jayawardhana , r. , allard , f. , hauschildt , p. , & ardila , d. 2004 , apj , 609 , 854 partridge , h. , & schwenke , d. w. 1997 , j. chem .
phys . , 106 , 4618 phillips , j.g .
1973 , apjs , 26 , 313 phillips , j.g . ,
davis , s.p . , lindgren , b. , & balfour , w.j .
1987 , apjs , 65,721 prato , l. , simon , m. , mazeh , t. , mclean , i.s . , norman , d. & zucker , s. 2002 , apj , 569 , 863 rebolo , r. , zapatero osorio , m. r. , madruga , s. , bejar , v. j. s. , arribas , s. , licandro , j. , 1998 , science , 282,1309 reid , i. n. , kirkpatrick , j. d. , gizis , j. e. , dahn , c. c. , monet , d. g. , williams , r. j. , liebert , j. , & burgasser , a. j. 2000 , aj , 119 , 369 reid , i. n. , kirkpatrick , j. d. , liebert , j. , gizis , j. e. , dahn , c. c. , & monet , d. g. , 2002 , aj , 124 , 519 ruiz , m. t. , leggett , s. k. , & allard , f. 1997 , apj , 491 , l107 saumon , d. , marley , m. s. , lodders , k. , & freedman , r. s.2003 , in brown dwarfs , iau symp .
e. martin , asp conference series .
schweitzer , a. , gizis , j. e. , hauschildt , p. h. , allard , f. , & reid , i. n. 2001 , apj , 555 , 368 simon , m. , bender , c. , & prato , l. 2006 , apj , 644 , 1183 testi , l. , et al .
2001 , apj , 522 , l147 tokunaga , a. t. 2000 , in allen s astrophysical quantities , ed .
a. n. cox ( 4th ed .
; new york : springer ) , 151 wilson , j. c. , kirkpatrick , j. d. , gizis , j. e. , skrutskie , m. f. , monet , d. g. , & houck , j. r. 2001 , aj , 122 , 1989 zapatero osorio , m. r. , lane , b. f. , pavlenko , ya . , martin , e. l. , britton , m. , & kulkarni , s. r. 2004 , apj , 615 , 958 zapatero osorio , m. r. , martn , e. l. , bouy , h. , tata , r. , deshpande , r. , & wainscoat , r. j. 2006 , apj , 647 , 1405 g196 - 3a & 10 04 21.0 & 50 23 06 & m2.5 & [email protected] & 2002 apr 23 + wolf 359 ( gj 406 ) & 10 56 28.9 & 07 00 53 & m6 & [email protected] & 2002 apr 23 + 2massw j0140026 + 270150 & 01 40 02.6 & 27 01 50 & m9 & [email protected] & 2000 dec 4 + 2masp j0345432 + 254023 & 03 45 43.2 & 25 40 23 & l0 & [email protected] & 2000 dec 4 + 2massi j0746425 + 200032 & 07 46 42.6 & 20 00 32 & l0.5 & [email protected] & 2002 jan 1 + kelu-1 & 13 05 40.2 & @xmath925 41 06 & l2 & [email protected] & 2003 may 12 + g196 - 3b & 10 04 21.0 & 50 23 06 & l2 & [email protected] & 2002 apr 23 + 2massw j0036159 + 182110 & 00 36 16.2 & 18 21 10 & l3.5 & [email protected] & 2000 dec 4 + gd165b & 14 24 39.1 & 09 17 10 & l4 & [email protected] & 2003 may 13 + 2massw j1507476@xmath9162738 & 15 07 47.7 & @xmath916 27 39 & l5 & [email protected] & 2000 apr 25 + denis - p j0205.4@xmath91159 & 02 05 29.4 & @xmath911 59 30 & l7 & [email protected] & 2001 oct 9 + sdssp [email protected] & 04 23 48.6 & @xmath904 14 04 & t0 & [email protected] & 2001 oct 9 + sdssp [email protected] & 12 54 53.9 & @xmath901 22 47 & t2 & [email protected] & 2003 may 14 + 2mass j05591914@xmath91404488 & 05 59 19.1 & @xmath914 04 48 & t4.5 & [email protected] & 2001 oct 9 + 2massi j2356547@xmath9155310 & 23 56 54.8 & @xmath915 53 11 & t5.5 & [email protected] & 2005 july 19 + 2massi j0937347 + 293142 & 09 37 34.7 & 29 31 42 & t6p & [email protected] & 2003 may 12 + + 58 & 1.304471.32370 & 192.3 & 0.188 & 84.5 + 59 & 1.282621.30151 & 188.9 & 0.184 & 85.9 + 60 & 1.261371.27999 & 186.2 & 0.182 & 87.5 + 61 & 1.240811.25913 & 183.2 & 0.179 & 89.0 + 62 & 1.220931.23899 & 180.6 & 0.176 & 90.7 + 63 & 1.201681.21938 & 177.0 & 0.173 & 91.7 + 64 & 1.182931.20011 & 171.8 & 0.168 & 91.9 + 65 & 1.164961.18207 & 171.1 & 0.167 & 94.4 + + lccc al i & 1.3127007 & 4s @xmath38s@xmath53 - 4p @xmath38p@xmath54 & 3.143 - 4.087 + al i & 1.3154345 & 4s @xmath38s@xmath53 - 4p @xmath38p@xmath53 & 3.143 - 4.085 + crh & 1.18 ? & 6 bands of a@xmath55 @xmath56 - x@xmath55 @xmath56 & + fe i & 1.1693174 & a@xmath57p@xmath58 - z@xmath57d@xmath39@xmath58 & 2.223 - 3.283 + fe i & 1.1786490 & b@xmath59p@xmath4 - z@xmath59d@xmath39@xmath60 & 2.831 - 3.884 + fe i & 1.1886098 & a@xmath57p@xmath4 - z@xmath57d@xmath39@xmath60 & 2.198 - 3.241 + fe i & 1.1887337 & a@xmath57p@xmath58 - z@xmath57d@xmath39@xmath4 & 2.223 - 3.266 + fe i & 1.1976325 & a@xmath57p@xmath60 - z@xmath57d@xmath39@xmath50 & 2.176 - 3.211 + feh & 1.1939 band head & 0 - 1 band of f@xmath61 @xmath62-x@xmath61 @xmath62 & + feh & 1.2389 band head & 1 - 2 band of f@xmath61 @xmath62-x@xmath61 @xmath62 & + h@xmath4o & 1.135 & @xmath63 & + h@xmath4o & 1.331 & 2@xmath64 band & + k i & 1.1693420 & 4p @xmath38p@xmath53 - 3d @xmath38d@xmath54 & 1.610 - 2.670 + k i & 1.1772861 & 4p @xmath38p@xmath54 - 3d @xmath38d@xmath54 & 1.617 - 2.670 + k i & 1.1776061 & 4p @xmath38p@xmath54 - 3d @xmath38d@xmath65 & 1.617 - 2.670 + k i & 1.2435675 & 4p @xmath38p@xmath53 - 5s @xmath38s@xmath53 & 1.610 - 2.607 + k i & 1.2525560 & 4p @xmath38p@xmath54 - 5s @xmath38s@xmath53 & 1.617 - 2.607 +
mn i & 1.290329 & a@xmath55d@xmath66 - z@xmath55p@xmath67 & 2.114 - 3.075 + na i & 1.268261 & 3d @xmath38d@xmath68 - 5f @xmath38f@xmath65 & 3.617 - 4.595 + na i & 1.268269 & 3d @xmath38d@xmath54 - 5f @xmath38f@xmath65 & 3.617 - 4.595 + ti i & 1.1896124 & b@xmath59f@xmath4 - z@xmath59d@xmath39@xmath58 & 1.430 - 2.472 + ti i & 1.2674567 & b@xmath59f@xmath4 - z@xmath59f@xmath39@xmath60 & 1.430 - 2.408 + ti i & 1.2834947 & b@xmath59f@xmath4 - z@xmath59f@xmath39@xmath4 & 1.430 - 2.396 + ti
i & 1.2850544 & b@xmath59f@xmath60 - z@xmath59f@xmath39@xmath60 & 1.443 - 2.408 + m2.5 & g196 - 3a & [email protected] & [email protected] & 39@xmath244 + m6 & wolf 359 & [email protected] & [email protected] & 64@xmath246 + m9 & 2mass j0140 + 27 & [email protected] & [email protected] & 110@xmath2411 + l0 & 2mass j0345 + 25 & [email protected] & [email protected] & 180@xmath2418 + l0.5 & 2mass j0746 + 20ab & [email protected] & [email protected] & 210@xmath2421 + l2 & kelu-1ab & [email protected] & [email protected] & 320@xmath2432 + l3.5 & 2mass j0036 + 18 & [email protected] & [email protected] & 240@xmath2424 + l4 & gd165b & [email protected] & [email protected] & 230@xmath2423 + l5 & 2mass j1507@xmath916 & [email protected] & [email protected] & 240@xmath2424 + l7 & denis j0205@xmath911ab & [email protected] & [email protected] & 290@xmath2429 + t0 & sdss j0423@xmath904ab & [email protected] & [email protected] & 240@xmath2424 + t2 & sdss j1254@xmath901 & [email protected] & [email protected] & 270@xmath2427 + t4.5 & 2mass j0559@xmath914 & [email protected] & [email protected] & 490@xmath2450 + m2.5 & g196 - 3a & [email protected] & [email protected] & 47@xmath245 + m6 & wolf 359 & [email protected] & [email protected] & 65@xmath247 + m9 & 2mass j0140 + 27 & [email protected] & [email protected] & 78@xmath248 + l0 & 2mass j0345 + 25 & [email protected] & [email protected] & 220@xmath2422 + l0.5 & 2mass j0746 + 20ab & [email protected] & [email protected] & 230@xmath2423 + l2 & kelu-1ab & [email protected] & [email protected] & 320@xmath2431 + l3.5 & 2mass j0036 + 18 & [email protected] & [email protected] & 290@xmath2429 + l4 & gd165b & [email protected] & [email protected] & 150@xmath2415 + l5 & 2mass j1507@xmath916 & [email protected] & [email protected] & 270@xmath2427 + l7 & denis j0205@xmath911ab & [email protected] & [email protected] & 390@xmath2439 + t0 & sdss j0423@xmath904ab & [email protected] & [email protected] & 300@xmath2430 + t2 & sdss j1254@xmath901 & [email protected] & [email protected] & 260@xmath2426 + t4.5 & 2mass j0559@xmath914 & [email protected] & [email protected] & 460@xmath2446 + m2.5 & g196 - 3a & [email protected] & [email protected] + m6 & wolf 359 & [email protected] & [email protected] + m9 & 2mass j0140 + 27 & [email protected] & [email protected] + l0 & 2mass j0345 + 25 & [email protected] & [email protected] + l0.5 & 2mass j0746 + 20ab & [email protected] & [email protected] + l2 & kelu-1ab & [email protected] & [email protected] + l3.5 & 2mass j0036 + 18 & [email protected] & [email protected] + l4 & gd165b & [email protected] & 0.35 @xmath240.05 + l5 & 2mass j1507@xmath916 & [email protected] & [email protected] + l7 & denis j0205@xmath911ab & [email protected] & [email protected] + t0 & sdss j0423@xmath904ab & [email protected] & [email protected] + t2 & sdss j1254@xmath901 & [email protected] & [email protected] + t4.5 & 2mass j0559@xmath914 & [email protected] & [email protected] + m for the m9 dwarf 2mass 0140 + 27 is over - plotted on a scaled feh opacity plot smoothed to r@xmath020,000 .
the region near 1.222 @xmath3 m previously attributed to the q - branch of the f@xmath61 @xmath69 - x@xmath61 @xmath69 system of feh is resolved , and almost every other feature is identified with p - branch ( dotted vertical lines not otherwise labeled ) or r - branch ( labeled ) lines of the 01 band .
lines not attributed to feh are discussed in the text . ]
16 ) dwarfs from table 1 .
dashed lines show the location of the feh transitions and the darker solid lines correspond to a resolution of r@xmath020,000 .
although the features are broader in the l5 dwarf , almost all of the weaker features in order 61 can be attributed to feh .
lines not attributed to feh are discussed in the text .
note the broadening of the strong k i lines from m9 to l5 . ]
o opacity at 1000 k and a pressure of 1 bar , is overlaid with the order 61 spectra of the l5 ( 2mass 1507@xmath916 ) and t4.5 ( 2mass 0559@xmath914 ) dwarfs from table 1 .
dashed lines show the location of the 2o transitions and the solid lines correspond to a resolution of r@xmath020,000 .
h@xmath4o features do not dominate either spectrum .
spectral features in the l5 that were not consistent with feh can not be attributed to h@xmath4o .
feh is almost completely absent from the t4.5 . ]
o opacity at 1000 k and a pressure of 1 bar , overlaid with the order 65 nirspec high - resolution spectra of the l3.5 ( 2mass 0036 + 18 ) and t4.5 ( 2mass 0559@xmath914 ) dwarfs from table 1 . as in the previous figure
, the solid lines correspond to a resolution of r@xmath020,000 .
no feh lines are listed in the opacity tables for this order .
all of the strong , sharper features in the t4.5 spectrum correlate well with h@xmath4o transitions . for the l3.5
the correlation is weaker but still present .
note the broadening of the k i lines from l3.5 to t4.5 . ]
m ( order 61 ) of the l2 dwarf kelu-1ab compared to l0.5 and l3.5 dwarfs ( 2mass 0746 + 20ab and 2mass 0036 + 18 , respectively ) .
the k i lines are significantly broader and shallower in kelu-1ab , even compared to the binary l0.5 .
spectra are normalized to unity and offset by a constant . each spectrum has been shifted to the laboratory ( vacuum ) rest frame . ] | we present a sequence of high resolution ( r@xmath020,000 or 15 km s@xmath1 ) infrared spectra of stars and brown dwarfs spanning spectral types m2.5 to t6 .
observations of 16 objects were obtained using eight echelle orders to cover part of the @xmath2-band from 1.165 - 1.323 @xmath3 m with nirspec on the keck ii telescope . by comparing opacity plots and line lists , over 200 weak features in the @xmath2-band
are identified with either feh or h@xmath4o transitions .
absorption by feh attains maximum strength in the mid - l dwarfs , while h@xmath4o absorption becomes systematically stronger towards later spectral types .
narrow resolved features broaden markedly after the m to l transition .
our high resolution spectra also reveal that the disappearance of neutral al lines at the boundary between m and l dwarfs is remarkably abrupt , presumably because of the formation of grains .
neutral fe lines can be traced to mid - l dwarfs before fe is removed by condensation .
the neutral potassium ( k i ) doublets that dominate the @xmath2-band have pressure broadened wings that continue to broaden from @xmath0 50 km s@xmath1 ( fwhm ) at mid - m to @xmath0 500 km s@xmath1 at mid - t . in contrast however , the measured pseudo - equivalent widths of these same lines reach a maximum in the mid - l dwarfs .
the young l2 dwarf ,
g196 - 3b , exhibits narrow potassium lines without extensive pressure - broadened wings , indicative of a lower gravity atmosphere .
kelu-1ab , another l2 , has exceptionally broad infrared lines , including feh and h@xmath4o features , confirming its status as a rapid rotator .
in contrast to other late t objects , the peculiar t6 dwarf 2mass 0937 + 29 displays a complete absence of potassium even at high resolution , which may be a metallicity effect or a result of a cooler , higher - gravity atmosphere . |
information discovery is central to many activities in life , from finding restaurants while attending a conference to making strategic decisions in a big company . in science ,
information discovery is , for example , used to stay up - to - date , doing literature research , and it is crucial in the process of scientific research itself . in a world
where online information is generated 24 hours a day , 7 days a week , this journey of discovery can easily become a daunting task .
we need powerful discovery tools to help us on this journey .
general search engines support a low level of information retrieval , sufficient to get a general idea , but when you are looking for technical material in a rich metadata environment , you need specialized digital libraries .
the sao / nasa astrophysics data system ( ads ) is such a digital library ( @xcite , @xcite ) .
it has very successfully served the astronomy and physics community for almost 20 years , free of charge . in order to support a richer and more efficient information discovery experience
, we created the ads labs environment , in which we expose our users to new search paradigms and tools , to better support our community s research needs . in the following
we will argue that the ads labs environment will prove to be a useful tool for people involved in science education and outreach .
general search engines return so many results that it quickly consumes excessive amounts of time to parse all these results and determine if any of these are relevant in a science education environment . as a result
, instructors often return to resources they have been using time and time again , while missing out on a wealth of new material continuously being added through a broad spectrum of resources .
a specialized digital library is a highly efficient tool to locate these resources . through its contents and functionality
, ads labs will prove to be such tool .
in addition to publications relevant for scientific research , the ads repository also contains a set of journals that are directly relevant to people involved in science education .
examples of such journals are : _ astronomy education review _ , _ american journal of physics _ ; _ the science teacher _ ; _ journal of science education and technology _ ; _ journal of science teacher education _ ;
_ international journal of science education _ ; _ research in science education _ ; _ science & education _ ; _ spark , the aas education newsletter_. a large portion of the astronomical research of the 19th and early 20th centuries was reported in publications written and published by individual observatories .
many of these collections were not widely distributed and complete sets of these volumes are now , at best , difficult to locate . this material is often requested by amateur astronomers and researchers , because the observatory report is the only published record of the research and observations .
this makes these publications a great resource for classes that have a component dealing with the history of astronomy .
we have a large number of historical publications ( mostly from microfilm ) in our repository of scanned literature ( @xcite ) .
the functionality that makes ads labs such a powerful tool is a combination of being able to specify ahead of time what kind of results you are interested in , and the ability to efficiently filter the results afterwards , using facets .
we will illustrate this using an example .
imagine you are looking for publications describing or on discussing extrasolar planets in a classroom environment .
figure [ screenshot : streamlined ] shows how you could start this search , using the `` streamlined search '' of ads labs . in the search box
we specify `` extrasolar planets classroom '' , as search scope we select `` all '' ( next to the search button ) , to search the entire ads repository , and we specify `` most relevant '' for sorting .
this sorts on a combination of several indicators , including date , position of the query words in the document , position of the author in the author list , citation statistics and usage statistics ( this is how many popular search engines rank ) .
this example query generates the results list shown in figure [ screenshot : results ] .
the results page consists of a list of publications with a panel of facets on the left , which offer an efficient way of further filtering the results . besides serving as a filter
, the author facet summarizes the people active in the field defined by the query .
the author facet also shows all the spelling variants by which author names occur in the ads repositories ( by clicking on the author name ) , which can assist in filtering out different authors with the same initials ( where e.g. the first name is spelled out ) .
the diagram below the facets shows the number of publications as a function of year , providing a measure for the activity in a field . in this way the facets , besides serving as filters , also provide you with valuable information .
the `` data '' facet provides you with an overview of available data products , if available .
every entry in the results list has a potential option to look inside the publication .
if an abstract is available , the `` matches in abstract '' link will open a snippet showing the matches in the abstract of the query terms .
the link `` matches in fulltext '' will show these matches in the full text version of the publication ( which could be the preprint version from arxiv ) .
the results page also provides the menu `` more '' ( in the upper right corner ) that contains tools to further explore the results .
the `` author network '' allows one to visualize collaborations between authors and further filter the search results by selecting within the nodes in the network .
the paper network allows one to visualize the relationships between papers and further filter the search results by selecting within the nodes in the network .
if the publications in the results list have astronomical objects identified in them , you can visualize these using the `` sky map '' option .
this uses google sky .
the sky map can also be used for further filtering .
the `` word cloud '' visualizes the most relevant terms found in the list of results and further filter the list by selecting interesting terms from the cloud .
the word cloud is only indirectly based on word frequencies : the word frequencies in the astronomy corpus are subtracted , so that relatively frequent words stand out .
the `` metrics '' page provides an overview of bibliometric indicators , based on the publications in the results list .
this overview can be exported in excel format .
the next level is the abstract page ( see figure [ screenshot : abstract ] ) .
the principal function of this page is to provide the user with a concentrated description of an article , sufficient for the user to decide whether or not to download and read it .
this page provides a view on the basic metadata ( authors , affiliations , title , abstract ) .
it also provides links to the full text ( including open access versions ) and opportunities to share the abstract on various social media sites , or save it in an ads private library ( available for users with an ads account , see below ) .
when sufficient data is available , a set of `` suggested articles '' is provided , which is generated by a recommender system ( @xcite ) .
depending on availability , the abstract view provides access to the bibliography of the publication ( those papers cited by the publication , for which there is a record in the ads ) , an overview of publications citing the publication , a list of papers most frequently read in conjunction with this publication ( `` co - reads '' ) and a list of papers that are similar to this publication , based on word similarity ( `` similar articles '' ) .
figure [ screenshot : streamlined ] has a area on the right with two tabs ( `` myads articles '' and `` recently viewed articles '' ) .
these relate to the existence of user accounts in the ads .
the ads offers the option to create a login , providing its users with the possibility to personalize the service .
this includes access to a powerful alert service called `` myads '' ( see @xcite ) , the creation of `` private libraries '' ( which are essentially baskets in which users can store , and annotate , links to publications ) and a way to specify a `` library link server '' , allowing access to full text using institutional subscriptions .
this makes the ads portable , because it provides you with online access to the full text of articles from anywhere in the world .
when a user is logged in to their account , visiting the streamlined search will result in displaying the most recent content for their `` daily myads '' service and an overview of their most recently viewed records . as an illustration of finding historical material ,
consider the following example : you are interested in the history and use of an instrument called the `` mural circle '' .
when you run the query `` mural circle " 1800 - 1850 '' in the streamlined search , you will find about two dozen results .
these results show that in this period there were such instruments at the madras observatory , royal greenwich observatory , armagh observatory , u.s .
naval observatory and cape observatory .
most publications in this results list are available in pdf format .
through the publications in its holdings and the user - friendly , intuitive streamlined search , the ads is a useful instrument in the tool box of search engines for professionals involved in science education and outreach .
we do realize that this is a group of users with requirements and needs that , in some aspects , differ significantly from those that have traditionally been using the ads .
we would love to get feedback and suggestions to help us optimize the search experience for all our users .
our users are a big part of our curation efforts , so if you encounter material in our database relevant for science education , but not flagged as such , we would love to hear from you as well .
feedback should be sent to [email protected] .
henneken , e. , kurtz , m. j. , eichhorn , g. , et al .
2007 , library and information services in astronomy v , 377 , 106 henneken , e. a. , kurtz , m. j. , & accomazzi , a. 2011 , arxiv:1106.5644 henneken , e. a. , kurtz , m. j. , accomazzi , a. , et al .
2011 , astrophysics and space science proceedings , 1 , 125 kurtz , m. j. , eichhorn , g. , accomazzi , a. , grant , c. s. , murray , s. s.,watson , j. m. ( 2000 ) .
astronomy and astrophysics supplement series 143 , 41 - 59 .
thompson , d. m. , accomazzi , a. , eichhorn , g. , et al .
2007 , library and information services in astronomy v , 377 , 102 | the sao / nasa astrophysics data system ( ads ) is an open access digital library portal for researchers in astronomy and physics , operated by the smithsonian astrophysical observatory ( sao ) under a nasa grant , successfully serving the professional science community for two decades .
currently there are about 55,000 frequent users ( 100 + queries per year ) , and up to 10 million infrequent users per year .
access by the general public now accounts for about half of all ads use , demonstrating the vast reach of the content in our databases .
the visibility and use of content in the ads can be measured by the fact that there are over 17,000 links from wikipedia pages to ads content , a figure comparable to the number of links that wikipedia has to oclc s worldcat catalog .
the ads , through its holdings and innovative techniques available in ads labs ( http://adslabs.org ) , offers an environment for information discovery that is unlike any other service currently available to the astrophysics community . literature discovery and review are important components of science education , aiding the process of preparing for a class , project , or presentation .
the ads has been recognized as a rich source of information for the science education community in astronomy , thanks to its collaborations within the astronomy community , publishers and projects like compadre .
one element that makes the ads uniquely relevant for the science education community is the availability of powerful tools to explore aspects of the astronomy literature as well as the relationship between topics , people , observations and scientific papers .
the other element is the extensive repository of scanned literature , a significant fraction of which consists of historical literature . |
figure [ fig : experiment ] shows a depiction of our experimental setup , which has been previously described @xcite .
the setup is based on a horizontal self - built inverted microscope , where the sample is illuminated by a white lamp while the image is captured by a ccd camera .
an infrared diode laser ( @xmath109 , lumics , @xmath110 maximum power ) coupled in a single - mode fiber ( avanex , 1998plm 3cn00472ag highpower @xmath111 ) is highly focused by a high numerical aperture ( na ) immersion oil objective @xmath112 ( nikon , cfi pl fl 100@xmath113 na 1.30 ) to create the optical potential . prior to entering the objective , the optical beam is expanded by lenses @xmath114 ( focal length@xmath115 ) and @xmath116 ( focal length@xmath117 ) to overfill the input pupil of the objective .
laser controller ( arroyo instruments 4210 ) allows the management of the optical power at a maximum rate of @xmath118 using an external voltage @xmath119 .
since the trap stiffness @xmath27 depends linearly on the optical power , @xmath27 can be controlled with at the same rate as the external voltage @xcite .
polystyrene beads ( g. kisker - products for biotechnology , pps-1.0 , diameter @xmath120 ) are diluted in milli - q water to a final concentration of a few microspheres per ml .
the solution is injected into a custom - made fluid chamber , which is placed in a holder whose position in the three axes can be controlled with picomotors ( newport , 8752 ) .
afterwards , the chamber is mechanically sealed to avoid fluxes and contamination , allowing us to work several days with the same solution .
polystyrene beads have an inherent charge in polar liquids which allows us to apply deterministic forces into our trapped microsphere .
we add two aluminum electrodes at the two ends of the chamber to apply a controllable voltage ( @xmath121 ) to the sample . both @xmath119 and @xmath121
are controlled by the same signal generator ( tabor electronics , ww5062 ) run by labview software . in the case of @xmath121 ,
the output signal of the signal generator is amplified @xmath122 times with a high - voltage power amplifier ( trek , 623b ) .
the particle is tracked using an additional @xmath123 laser collimated by a microscope objective ( @xmath11310 , na 0.10 ) and sent through the trapping objective ( @xmath112 ) .
the light scattered by the trapped object is collected by the objective @xmath124 ( olympus , 40@xmath113 , na 0.75 ) and projected into a quadrant photo detector ( qpd , newfocus 2911 ) .
the maximum acquisition frequency of the qpd is @xmath125 . a @xmath123 pass filter ( @xmath126 ) blocks additional scattered light .
the signal is transferred through an analog - to - digital conversion card ( national instruments pci-6120 ) and recorded with labview software .
the calibration of the nanodetection is obtained from the analysis of the thermal fluctuations of the bead within a static trap at room temperature . from the study of the power spectral density of the trajectories , both voltage - to - nanometers conversion factor , @xmath127(nm / v ) , and @xmath27
are obtained .
all experiments are done with the beads trapped 20 @xmath128 m above the coverslip surface , in order to avoid surface effects in the friction coefficient @xmath129 @xcite .
the input voltage controls the noise intensity and can be linked to the effective temperature of the particle as @xmath130 , where @xmath131(k / v@xmath132 ) is the calibration factor .
all calibrations are repeated each time a new bead is trapped . in the experiments presented here , @xmath27 is calibrated as a function of @xmath119 ( data not shown ) , @xmath119 being of the order of @xmath133 , and noise amplitudes of the order of thousands of @xmath134 which led to values of @xmath135 up to thousands of kelvins .
note that , although we do not know the actual value of the electric field in our chambers , it is not needed for our calculations . .
the stiffness of the trap is modified by a power supply that controls the intensity of the laser .
a noisy electric potential @xmath121 is applied to the electrodes in the fluid chamber to control the kinetic temperature of the trapped particle .
a red laser ( 532 nm ) is used to track the position of the particle , by recording the forward scattered light collected by the objective @xmath136 into a quadrant photodiode ( qpd ) . a led and a ccd camera are used for visualization .
, width=340 ]
the measurement of instantaneous velocity of a brownian particle requires , in principle , to sample the position of the particle with acquisition rates of the order of the momentum relaxation frequency @xmath137 @xcite or use a different system where the friction is not as strong as it is in water @xcite .
we make use of another recent result described in ref .
@xcite that allows one to measure the mean squared instantaneous velocity at any time @xmath49 during a quasistatic process , @xmath138 from the mean squared time averaged velocity , @xmath139 , being the latter obtained from low frequency samplings ( @xmath140 or @xmath141 ) , @xmath142 $ ] . for an underdamped brownian particle of mass @xmath57 trapped with a quadratic potential of time - dependent stiffness @xmath143 , @xmath144 and immersed in a thermal bath at temperature @xmath0 , the mean squared instantaneous velocity equals to @xmath145 where @xmath146 is a correction factor that depends on the acquisition frequency and on the physical parameters of the system at time @xmath49 : @xmath147^{-1 } , \label{eq : factorbdec_trapped}\ ] ] where @xmath148 , @xmath149 and @xmath150 . at the beginning ( @xmath151 , blue squares ) and at the end ( @xmath152 , red circles ) of the pseudo - adiabatic process .
the data of @xmath153 is obtained from a time window @xmath154 $ ] with @xmath155 for @xmath151 ( beginning ) and @xmath156 ( end ) for an ensemble of @xmath55 cycles of duration @xmath77 .
the rescaling is done using @xmath146 calculated using eq . .
we also show the maxwell - boltzmann equilibrium velocity distribution ( black curve).,width=245 ] in a quasistatic process , @xmath59 is a gaussian variable with zero mean and variance @xmath157 . since @xmath158 ,
the time - averaged velocity is also gaussian distributed , as the sum of gaussian variables is also a gaussian variable .
its mean is trivially zero , @xmath159 . to fully specify the distribution we must therefore only give its second momentum , which is @xmath160 with the correcting factor @xmath146 as given in . as a result
, @xmath161 has the same distribution as the instantaneous velocity @xmath2 , or equivalenty , @xmath162 the latter being described by the equilibrium maxwell - boltzmann velocity distribution .
we tested the theoretical prediction given by eq . with data from the quasistatic processes described in the main text , as well as in the two control experiments ( isothermal and isochoric processes ) . in fig .
[ fig : pdfv ] we show the probability density function of @xmath163 for @xmath151 and @xmath152 for the pseudo - adiabatic process , obtained from @xmath55 cycles .
both distributions coincide with maxwell - boltzmann distribution ( black curve in fig .
[ fig : pdfv ] ) with high accuracy for two orders of magnitude .
the tails of the reconstructed experimental distributions deviate from the theoretical distribution due to statistical under sampling .
this deviation might decrease when using data from a larger number of cycle repetitions .
let us consider a brownian particle of mass @xmath57 that is immersed in a fluid at temperature @xmath0 .
we assume that the particle moves in one dimension , and its position and velocity at time @xmath49 are denoted as @xmath48 and @xmath59 , respectively .
the particle is trapped with a potential @xmath164 that can be changed in time via a control parameter @xmath165 , and the temperature might change in time as @xmath166 .
the dynamics of such a brownian particle is described by the underdamped langevin equation @xcite @xmath167 where @xmath129 is the friction coefficient of the particle in the fluid .
thermal fluctuations are modeled by a gaussian white noise with @xmath168 and @xmath169 , @xmath31 being boltzmann s constant .
the term @xmath170 accounts for any external forces that can be directly exerted to the particle .
we consider quasistatic thermodynamic processes where the control parameter changes slower than any relaxation time of the system .
in such a case , the phase space density of the system can be described by gibbs distribution throughout the process .
if the trapping potential is quadratic , @xmath143 being the stiffness of the potential , @xmath171 , and the energy of the particle can be described by the hamiltonian @xmath172 at time @xmath49 , the system is described by a canonical state @xmath173/z_t$ ] , where @xmath174 and @xmath175 is the partition function , which is equal to @xmath176 the free energy of the particle at time @xmath49 is @xmath177 the entropy of the particle @xmath178 satisfies @xmath179 . therefore in a quasistatic adiabatic process @xmath180 , or equivalently , @xmath181 . the same analysis can be carried out using the overdamped description , i.e. , neglecting the velocity degree of freedom .
in such a case , it can be shown that @xmath182 , which implies that entropy in the position degree of freedom can be conserved when @xmath183 . [ cols="<,<,<,<",options="header " , ] we now study thermodynamic processes where time runs in the interval @xmath184 $ ] . during such processes , the position and the velocity of the brownian particle describe a trajectory @xmath185 . for systems described by an underdamped langevin equation , the work exerted on the particle and the heat transferred from the thermal bath in an interval of time @xmath186
$ ] are , respectively : @xmath187 where @xmath188 denotes the stratonovich product @xcite .
notice that using , the heat can be decomposed in two terms , @xmath189 the first term being the kinetic energy change @xmath190 and the second the heat transferred to the position of the particle , @xmath191 . for a brownian particle trapped with a harmonic potential of stiffness @xmath143 , equipartition theorem implies that @xmath192 , and therefore , @xmath193 along the process , where the brackets denote ensemble average . in such a case , one can calculate the work transferred to the particle averaged over many realizations , @xmath194 and the heat transferred to the position , @xmath195 notice that equipartition theorem implies that @xmath196 $ ] , @xmath197 being the internal energy change .
the total energy change satisfies @xmath198 $ ] . in table
[ tab : firstlaw ] we show the values of the quasistatic ensemble averages for selected thermodynamic processes : isothermal ( @xmath199 ) , isochoric ( @xmath200 ) and adiabatic processes . in the latter case ,
we distinguish between the pseudo - adiabatic process , where @xmath183 , which yields @xmath201 but @xmath202 , and the actual adiabatic process , where @xmath203 and the total average heat vanishes @xmath204 .
in this section , we calculate the energy change , heat and work distributions in quasistatic process in which a brownian particle whose position is denoted as @xmath1 is trapped with a quadratic potential of stiffness @xmath27 , @xmath3 .
we consider both the overdamped description where only the position degree of freedom is taken into account , and the full underdamped description including the velocity . along the protocol of duration @xmath205 , the temperature changes from @xmath206 to @xmath207 and the stiffness from @xmath208 to @xmath209 .
we assume that the process is quasistatic and therefore the distribution of the position at any time @xmath49 during the process is the equilibrium ( gaussian ) distribution @xmath210 where @xmath211 and @xmath212 is the partition function for the @xmath1 degree of freedom . in a quasistatic process
, the work distribution is peaked at its mean value @xcite @xmath213 taking into account the first law of thermodynamics , @xmath214 , the heat distribution is equal to the distribution of the internal energy change centered in @xmath215 @xmath216 if the initial position of the particle is @xmath217 and the final position is @xmath218 , the internal energy change is @xmath219 we calculate the distribution of @xmath197 for a quasistatic process where the temperature and stiffness change from @xmath220 to @xmath221 .
the probability distribution of @xmath197 to be @xmath222 $ ] is equal to @xmath223 , where @xmath224 where the integration is done from @xmath225 to @xmath226 unless we specify different integration limits .
we now do the following change of variables @xmath227 for @xmath228 .
the equilibrium distribution of the random variable @xmath229 is @xmath230 where the @xmath231function in the integrand can be rewritten as @xmath232 .
\label{eq : deltasum}\end{aligned}\ ] ] by replacing and in , and taking into account that @xmath229 can only take positive values , we obtain @xmath233 where @xmath234 is the step function evaluated at @xmath229 , and @xmath235 . the distribution of the internal energy change can be now expressed in terms of the new variables @xmath236 integrating over @xmath237 , @xmath238 which yields , using the expression for the distribution @xmath239 , @xmath240 for @xmath241 , the integral in is equal to @xmath242 where @xmath98 is the zeroth - order modified bessel function of the second kind . for @xmath243 , @xmath244 then , for any value of @xmath245 , we obtain @xmath246 the heat distribution is obtained from the distribution of the internal energy change , @xmath247 which is eq .
( 1 ) in the main text .
the distribution is asymmetric with respect to @xmath248 except for the isothermal case ( @xmath249 ) .
we start by computing the distribution of the energy change for a general non - isothermal quasistatic process .
the total internal energy change is given by @xmath250 let us first derive the equilibrium distribution of the energy of state @xmath251 , @xmath252 : @xmath253 where the partition functions for the position and velocity degrees of freedom read @xmath254 and @xmath255 .
we can first simplify the integral using the change of variables @xmath256 : @xmath257 transforming the integral to polar coordinates @xmath258 and rewriting dirac s delta as in , we finally obtain an exponential distribution for the energy as expected for an equilibrium state : @xmath259 the step function @xmath260 reflects the fact that total energy is necessarily positive .
the distribution for the energy change can be computed as @xmath261 which after using the delta to eliminate one integral can be readily shown to yield @xmath262 using the first law and taking into account that work is delta distributed in a quasistatic process , we find that heat distribution is exponentially distributed , according to @xmath263 since the final result only depends on starting with an equilibrium distribution for the energy , the total energy change distribution coincides with the work distribution for an adiabatic process in a classical hamiltonian system as the one referred to in the text , where the heat is exactly zero for every trajectory : @xmath264
total entropy production @xmath265 can be expressed as the sum of the system entropy change @xmath108 plus the entropy change in the environment , @xmath266 , where @xmath17 is the total heat absorbed by the system . in the quasistatic limit ,
total entropy production vanishes and is delta distributed , @xmath267 we might therefore consider fluctuations of system and environment entropies , which in this case satisfy @xmath268 , and therefore , @xmath269 let us consider the fluctuations of the system entropy change in a quasistatic thermodynamic process of duration @xmath205 .
the system entropy change from @xmath151 to @xmath270 is a state function and its value only depends on the initial and final micro state of the system , described by @xmath271 and @xmath272 , respectively , @xmath273 for a quasistatic process , initial and final distributions are canonical , yielding , @xmath274 + \ln \frac{z_\tau}{z_0 } , \label{eq : sparts}\ ] ] which finally gives @xmath275 \nonumber .
\label{eq : ssys}\end{aligned}\ ] ] in the above formula , the first term is deterministic while the second is stochastic .
the distribution of system entropy change is the distribution of the second term shifted by the value of the first term .
computing the distribution of @xmath276 follows exactly the same lines of the previous section computation for @xmath277 , except @xmath278 factors have to be absorbed in the change of variables .
consequently , @xmath279 also follows an exponential distribution : @xmath280 hence , system entropy is distributed according to @xmath281 being @xmath282 the deterministic part of the entropy change . for the overdamped case , @xmath283 and the derivation is analogous to that for the potential energy change @xmath197 , also absorbing the @xmath278 factors in the change of variables .
thus , the entropy change for a process only considering the position degree of freedom follows a distribution given by @xmath284 with @xmath285 .
since both distributions are even , the deterministic parts coincide with the mean value in both cases . when considering the distribution of system entropy shifted by its mean value @xmath286 , eqs . and yield eqs .
( 2 ) and ( 3 ) of the main text , respectively . | we experimentally realize quasistatic adiabatic processes using a single optically - trapped microsphere immersed in water whose effective temperature is controlled by an external random electric field .
a full energetic characterization of adiabatic processes that preserve either the position distribution or the full phase space volume is presented .
we show that only in the latter case the exchanged heat and the change in the entropy of the particle vanish when averaging over many repetitions .
we provide analytical expressions for the distributions of the fluctuating heat and entropy , which we verify experimentally .
we show that the heat distribution is asymmetric for any non - isothermal quasistatic process .
moreover , the shape of the distribution of the system entropy change in the adiabatic processes depends significantly on the number of degrees of freedom that are considered for the calculation of system entropy .
stochastic energetics @xcite and the fluctuation theorems @xcite have been developed as the theoretical framework that studies thermodynamics at small scales , thus establishing the emerging field of stochastic thermodynamics . in parallel , recent advances on micromanipulation and force - sensing techniques @xcite have allowed to measure the dynamics and energy changes in physical systems where thermal fluctuations are relevant @xcite and to test theoretical results derived from stochastic thermodynamics @xcite .
as a major application , miniaturization of thermodynamic engines to single - molecule devices has been possible for the case of stirling engine @xcite or a variety of maxwell s demons @xcite . until now , the design of microscopic heat engines has been restricted to those cycles formed by isothermal processes or instantaneous temperature changes @xcite , where the validity of a heat fluctuation theorem has been tested @xcite .
recent works have shown that exerting random forces on a microscopic particle one can accurately tune the effective kinetic temperature of the particle both under equilibrium @xcite and nonequilibrium driving @xcite . however
, the application of such a technique to implement non - isothermal processes has not been fully exploited yet @xcite . among all the non - isothermal processes , adiabatic processes are of major importance in thermodynamics since they are the building blocks of the carnot engine @xcite .
_ microadiabaticity _ , i.e. adiabaticity at the microscopic scale , can not be realized for single - trajectories due to the unavoidable heat flows between microscopic systems and their surroundings .
however , a process where no net heat transfer is obtained when averaged over many trajectories could in principle be realized .
although several theoretical proposals are available @xcite , their experimental implementation is still lacking .
is immersed in a thermal bath at temperature @xmath0 , moves in one dimension @xmath1 with velocity @xmath2 and is trapped with a harmonic potential @xmath3 .
the blue solid circle @xmath4 represents the ensemble of microstates described by a hamiltonian @xmath5 with a given energy @xmath6 .
the units of position and velocity are normalized by their standard deviation from equipartition theorem , @xmath7 and @xmath8 .
the red dashed ellipse @xmath9 is the microstate set at the same energy but after two different adiabatic processes : ( a ) pseudo - adiabatic process , where @xmath10 and @xmath11 ( i.e. , @xmath12 ) ; ( b ) adiabatic process , where @xmath13 and @xmath14 ( i.e. , @xmath15 ) .
the arrows indicate the direction in which the process occurs .
notice that the area of phase space that satisfies @xmath16 is conserved only along the adiabatic process .
[ fig : huevoscoloridos],width=226 ] in this letter , we report on the realization of quasistatic adiabatic processes with an optically - trapped microparticle immersed in water whose kinetic temperature is controlled by means of an external noisy electric field @xcite .
we provide a complete characterization of the thermodynamics of such adiabatic processes . the contributions due to the heat transferred to the momentum degree of freedom are also considered @xcite , thus adopting the full , underdamped description of the system .
interestingly , we show that doing so does not constitute a trivial extension of the overdamped description , but distinct features clearly arise .
in particular , we discuss the shape of the distribution of the fluctuations of heat and entropy in both descriptions , concluding that asymmetries in the heat distributions are a fingerprint of quasistatic non - isothermal processes . in classical hamiltonian systems , the total heat @xmath17 transferred in quasistatic adiabatic processes vanishes , and the heat distribution is @xmath18 .
the work , @xmath19 , @xmath20 being the internal energy , is exponentially distributed . in the microscopic regime
, one can attain processes where @xmath21 , @xmath22 denoting average over many realizations in the quasistatic limit . in the latter case , at odds with the hamiltonian case ,
the work is delta distributed , @xmath23 @xcite whereas the heat is exponentially distributed with the same distribution as @xmath24 , as shown in the supplemental material @xcite .
microadiabatic processes are those where the phase space volume is conserved @xcite . in the overdamped limit , where changes in the momentum degree of freedom are neglected , such condition is met by keeping the position distribution constant @xcite .
however , as we discuss below , the overdamped approximation is incomplete when dealing with non - isothermal processes , and a full underdamped description is mandatory .
a process where the position distribution is conserved is therefore a _ pseudo - adiabatic _ , since an unavoidable amount of heat is transferred due to the kinetic energy change @xcite .
in contrast , in the actual _ adiabatic _ process , the full phase space ( position and momentum ) volume is conserved and no net heat is transferred to the particle @xcite .
figure [ fig : huevoscoloridos ] illustrates the difference between the evolution of the phase space along both quasistatic pseudo - adiabatic and adiabatic processes for a brownian particle trapped with a harmonic potential .
notice that only in the adiabatic process the phase space volume enclosed by the energy surface defined by the system s energy at every moment @xmath25 is conserved , as required for a quasistatic and adiabatic change of parameters @xcite .
our system of study is a microparticle of radius @xmath26 immersed in water trapped by an optical harmonic optical potential @xmath3 , where @xmath27 is the stiffness of the trap and @xmath1 the position of the particle with respect to the trap center .
the key capability of our setup is the independent control of the kinetic temperature of the trapped bead @xmath28 and @xmath27 , thus allowing one to design a large variety of different thermodynamic processes @xcite .
both parameters can be electronically synchronized in order to fullfill any desired protocole with high time resolution , of the order of @xmath29 .
@xmath28 is defined from the application of equipartition theorem to the fluctuations of the position of the bead in the trap as follows .
these fluctuations obey equipartition theorem , @xmath30 , @xmath31 being boltzmann s constant and @xmath0 the temperature of the sample @xcite .
applying to the particle an external random force characterized by a gaussian white noise process of amplitude @xmath32 , we can mimic the kicks of the solvent molecules to the bead in a higher temperature reservoir @xmath28 . from equipartition theorem ,
the kinetic temperature of the particle depends on its mean squared displacement @xmath33 , @xmath34 , @xmath35 being the stokes friction on a sphere of radius @xmath36 in a fluid with kinetic viscosity @xmath37 far away from a surface .
these two parameters can be easily controlled , since @xmath27 is proportional to the intensity of the trapping laser @xcite and @xmath28 increases linearly with the square of the amplitude of a noisy voltage applied to a pair of electrodes in the fluid chamber @xcite .
see fig .
[ fig : protocols](a ) for a sketch of the experimental system . of a micro particle in an optical trap of stiffness @xmath27
is controlled with a noisy electric field .
( b ) pseudo - adiabatic protocol .
kinetic temperature from the mean squared displacement , @xmath38 ( left axis , blue solid line ) , kinetic temperature from the calibration ( left axis , blue dashed line ) and stiffness of the trap ( right axis , red dash - dot line ) as functions of time .
( c ) the same for the adiabatic process . notice the larger fluctuations as @xmath39 increases.,width=321 ] following the usual approach , we first implement a pseudo - adiabatic protocol where the entropy is conserved in the overdamped approximation , i.e. where @xmath40 ( see supplemental material @xcite for a proof ) , as the one shown in fig .
[ fig : protocols](b ) .
an actual microadiabatic process is achieved making @xmath41 ( see supplemental material @xcite ) and is implemented as shown in fig .
[ fig : protocols](c ) .
all the protocols presented here have a duration of @xmath42 .
since the relaxation time of the particle in the trap @xmath43 is of the order of milliseconds @xcite , then @xmath44 and the processes can be considered as quasistatic .
the latter is confirmed in figs . [
fig : protocols](b - c ) , where we show that the measured kinetic temperature fluctuates around the value prescribed by the protocol . after defining and implementing the desired protocols
, we calculate the thermodynamic quantities from measurements of the position of the trapped bead the stiffness of the trap and the kinetic temperature of the bead , the latter being obtained from standard calibration procedures @xcite .
the data acquisition frequency was @xmath45 .
the work done on the particle in the time interval @xmath46 $ ] is calculated as @xmath47 , @xmath48 being the position of the particle at time @xmath49 @xcite .
the heat transferred from the thermal bath to the position of the particle is calculated as @xmath50 .
the internal energy change is measured as the sum of the heat and the work transferred to the particle , @xmath51 . in the limit @xmath52 , the cumulative sum up to time @xmath49 of our definitions of heat and work
return sekimoto s expressions @xmath53 and @xmath54 @xcite .
ensemble averages and probability distributions are calculated from datasets of @xmath55 repetitions of each process .
we estimate the kinetic energy changes following the technique described in @xcite .
the sampling frequency in our experiment is far below the momentum relaxation frequency @xmath56 , @xmath57 being the mass of the bead @xcite .
therefore , we can only measure time averaged velocities @xmath58 rather than instantaneous velocities @xmath59 . in the quasistatic limit , we can obtain the mean squared instantaneous velocity from the mean squared time averaged velocity , @xmath60 , where @xmath61 is a function of the sampling frequency as well as of the parameters of the system at time @xmath49 ( stiffness , mass , friction coefficient ) @xcite .
the ensemble average kinetic energy change can be therefore calculated as @xmath62= \frac{m}{2 } [ \mathcal{l}_{t+\delta t } \langle \overline{v}_{t+\delta t}^2\rangle - \mathcal{l}_t \langle \overline{v}_t^2\rangle]$ ] .
in addition , we can assess the distribution of the instantaneous velocity from the distribution of the time averaged velocity .
the latter is gaussian with zero mean and the variance is related to that of the velocity distribution by @xmath63 , allowing to reconstruct the velocity distribution as @xmath64 , as shown in the supplemental material @xcite .
the velocity distributions can be exploited to determine the entropy of the system in the underdamped description at any time @xmath49 , @xmath65 , or equivalently , the average system entropy change in the interval @xmath46 $ ] , @xmath66 , where @xmath67 is the ensemble average system entropy at time @xmath49 @xcite .
we also consider the overdamped system entropy change obtained when we neglect the velocity degree of freedom , @xmath68 , @xmath69 being the entropy of the system in the position degree of freedom .
( blue ) , @xmath70 ( red ) , @xmath71 ( green ) , @xmath72 ( cyan ) , @xmath73 ( magenta ) and @xmath74 ( magenta ) .
( b ) system entropy as a function of time for the pseudo - adiabatic process , @xmath75 ( blue ) and @xmath76 ( red ) .
( c ) energetics of the adiabatic process .
( d ) system entropy change in the adiabatic process .
@xmath75 ( blue ) and @xmath76 ( red ) .
ensemble averages are obtained from @xmath55 repetitions of cycles of duration @xmath77 using a sampling rate of @xmath78 .
dashed curves are the theoretical values of the thermodynamic quantities obtained in the quasistatic limit . , width=321 ] using the aforementioned definitions of energetic quantities and entropy , we can now characterize the two types of microadiabatic processes .
let us first analyse the pseudo - adiabatic process .
figure [ fig : termoadiabatics](a ) shows ensemble averages of the cumulative sum of work , heat , kinetic energy , internal energy and total energy , which coincide with the expected values from equilibrium thermodynamics .
the average heat transferred to the position degree of freedom vanishes within experimental errors , yielding a net positive total value of the heat @xmath79>0 $ ] .
the average overdamped entropy change vanishes along the protocol , @xmath80 , as shown in fig .
[ fig : termoadiabatics ] ( b ) .
the pseudo - adiabatic nature of the protocol is revealed as a positive full system entropy change , @xmath81 [ red curve in fig .
[ fig : termoadiabatics ] ( b ) ] . for the adiabatic protocol [ fig .
[ fig : protocols](c ) ] , the ensemble average of the _ total _ heat transferred to the particle vanishes within experimental errors , @xmath82 , as shown in fig .
[ fig : termoadiabatics ] ( c ) . as a result
, the system entropy change vanishes along the adiabatic process @xmath83 despite entropy is reduced in the position degree of freedom , as shown in fig .
[ fig : termoadiabatics ] ( d ) .
let us now consider the fluctuations of the measured quantities .
again , their values fluctuate around the predictions obtained in the quasistatic limit .
the shape of the these distributions reveals qualitative differences between the considered processes .
first , in fig .
[ fig : heatdistributions ] we show the experimental probability density function of @xmath84 ( symbols ) for the two adiabatic processes considered .
we also include the distributions for two control processes : an isothermal process where @xmath85 and @xmath27 changes linearly in time from @xmath86 to @xmath87 , and an isochoric processes where @xmath88 and the kinetic temperature changes linearly from @xmath89 to @xmath90 @xcite . for different thermodynamic processes : isothermal ( blue circles ) , isochoric ( red squares ) , pseudo - adiabatic ( green stars ) and adiabatic ( black crosses ) .
the distributions are obtained from @xmath55 cycles of duration @xmath91 each .
the lines are theoretical distributions obtained from eq . using the initial and final values of kinetic temperature and stiffness used in the experiments , for the different processes : isothermal ( solid line ) , isochoric ( dashed line ) , pseudo - adiabatic ( dotted line ) and adiabatic ( dashed - dotted line).,width=207 ] remarkably , we notice that the heat distribution is asymmetric around its mean for all the non - isothermal processes .
the measured heat distributions can be well described by @xmath92 k_0 \left[\overline{\beta}|q_x+\langle w\rangle|\right ] , \label{eq : pdfq}\ ] ] where @xmath93 , @xmath94 , @xmath95 , @xmath96 , @xmath97 , @xmath98 is zeroth order modified bessel function of the second kind and @xmath99 is the ensemble average of the work in the quasistatic limit @xcite .
equation was obtained with the only assumption of quasistaticity along the process , and proves that the asymmetry of the distribution of @xmath84 around @xmath100 is a consequence of the non - isothermal character of the process , and not of any nonequilibrium constraint of the system , as suggested in ref .
@xcite .
for the isothermal case , @xmath101 and we recover the symmetric distribution @xmath102 $ ] firstly derived by imparato _ et al _ @xcite . the asymmetry observed in the heat fluctuations is not present in the distribution of the entropy .
the distribution of the overdamped entropy change along the whole process , @xmath103 is symmetric around around its mean value for both pseudo - adiabatic and adiabatic cases , as shown in fig .
[ fig : distr_entropy ] .
both distributions fit well to the expected value for general quasistatic non - isothermal processes @xcite , @xmath104 we also calculate the distribution of the full system entropy change along the whole process , @xmath105 in both pseudo - adiabatic and adiabatic processes ( see fig . [
fig : distr_entropy ] ) .
system entropy change is distributed symmetrically around its mean value but presents a different qualitative behavior , in this case described by @xcite : @xmath106 notice that in the case of the full system entropy change , the agreement with the theory extends over one order of magnitude less than in the overdamped description , @xmath107 .
this mismatch is caused by the poor estimation of the tails of the distribution of the instantaneous velocity from the distribution of the time averaged velocity .
( blue squares ) and of the total system entropy change @xmath108 ( red circles ) in the pseudo - adiabatic ( open symbols ) and adiabatic ( closed symbols ) processes shown in fig .
[ fig : protocols ] ( b ) and ( c ) .
the distributions are obtained from @xmath55 cycles of duration @xmath91 each .
theoretical distributions for @xmath107 [ blue solid curve , eq . ] and @xmath108 [ red dashed curve , eq .
] are also shown .
all the quantities are shifted by their mean such that the mean of the represented quantities is zero.,width=207 ] to summarize , we have realized quasistatic adiabatic processes with a single microparticle trapped with optical tweezers .
we have studied the difference between the pseudo - adiabatic ( position distribution conserving ) and adiabatic ( phase space volume conserving ) processes , showing that only the latter are such that the average total heat vanishes in the ensemble average .
the fluctuations of the heat transferred to the position of the particle have been shown to be asymmetric for any non - isothermal ( equilibrium or nonequilibrium ) thermodynamic process .
the description of the dynamics of the system with full or limited information affects not only to the average values of the entropy but also to the fluctuations , showing a different qualitative behavior .
the microadiabatic protocols studied in the present work could be used to design a microscopic - sized carnot engine by a cyclic sequence of isothermal and adiabatic processes , thus extending our understanding of micro and nano electromechanical systems towards new and efficient engines @xcite .
we acknowledge enlightening theoretical discussions with j. m. r. parrondo .
i.a.m .
, e.r .
, d.p . and r.a.r .
acknowledge financial support from the fundaci privada cellex barcelona , generalitat de catalunya grant 2009-sgr-159 , and from the spanish ministry of science and innovation ( micinn fis2011 - 24409 ) .
e.r . and l.d .
acknowledge financial support from enfasis ( spanish government ) .
the initial ideas of this work were conceived by prof .
d. petrov , leader of the optical tweezers group at icfo , who passed away on 3rd february 2014 .
41ifxundefined [ 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 _ _ , vol .
( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) _ _ , @noop ph.d . thesis , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * ( ) |
the gamma ray inverse problem ( @xcite ) is as follows : from a given number of observed parameters we have to reconstruct the intrinsic spectra .
this will be estimated at some other number of points .
the relation between the measured parameters and the intrinsic spectra is given by the detector response matrix ( drm ) .
in the case of the batse mer data , the spectra has to be approximated at @xmath0 intervals from @xmath1 measured count rates .
+ the inverse problem corresponds to the discretized version of the fredholm integral equation of the first kind .
@xmath2 here @xmath3 is the so - called kernel function , in our case the detector s response matrix .
@xmath4 is the measured count rate , @xmath5 is the quantity to be determined .
this is known to be an ill - posed and underdetermined problem . in the case of batse mer @xmath6
is a @xmath1 element vector , @xmath7 a @xmath0 element vector and @xmath8 is a @xmath9 matrix .
data were taken from the batse database ( ftp://cossc.gsfc.nasa.gov ) .
we only used the medium energy resolution data type .
background fits were made using intervals by norris ( http:// cossc.gsfc.nasa.gov/ docs/ cgro/ batse/ batseburst/ sixtyfour_ms/ bat_files.revamp_join ) for the _ cat64 _ data .
we have subtracted @xmath10 degree polynomials from all the @xmath1 channels of mer data .
throughout the literature there are articles which discuss in detail the inverse problem in astronomy ( @xcite ) and other sciences ( @xcite ) .
gamma - ray bursts spectra are most commonly reconstructed by the method parameter fitting also knows as forward folding @xcite . this is done by postulating spectral forms with some adjustable parameters . once multiplied with the drm we get a solution for the count rates and we then compare them to the measured set of data .
we accept the set of parameters which minimizes the @xmath11 of the solution and the measured counts .
although widely used and a well established method , still it is obliging when it comes to the shape of the spectra and it consists of estimating the parameters and not the spectra itself @xcite .
we decompose the response matrix using this well known mathematical method .
we solve the grip by multiplying with the generalised inverse obtained through svd .
spectra obtained from simply multiplying with the general inverse tend to be noisy . to avoid this we can use filters .
this method introduces some _ a priori _ information to reconstruct the spectra ( @xmath12 ) .
this is done by fitting a band grb spectral form @xcite .
we propose to maximize the expression : @xmath13 where @xmath14 is a parameter , @xmath12 is the _ a priori _ spectra ( from model fitting or svd ) .
it is with respect to this that we measure the entropy of the intrinsic spectra ( @xmath7 ) .
the expression of the entropy is as follows : @xmath15 .
when @xmath16 has a maximum then @xmath17 . from this
we derive : @xmath18 ( d - r f)$ ] , where @xmath19 $ ] is a diagonal matrix with elements of the @xmath20 vector and @xmath8 is the response matrix . since @xmath21 results from the linear combination of the @xmath22 s columns ,
we can search for the solution in the form of : @xmath23 . here
@xmath24 is a @xmath1 element vector .
this ensures that @xmath7 will always be positive , and since @xmath24 has a smaller scale , it is easier to find it as a result of a minimizing process .
the choice of @xmath14 can be made with the so - called cross - validation technique or another method based on a bayesian approach .
we have applied the method to batse mer data .
first results suggest that the band grb function provides a good enough fit .
its main advantage is that it is a linear method(@xcite ) , data in the intrinsic spectra is reconstructed as a linear combination of the measured count rates : @xmath25 and @xmath26 .
we call @xmath27 the resolution function .
our aim is to approximate the resolution function with dirac - delta function as well as possible . by doing this
we minimize for @xmath28 .
the other issue at hand is the variance of the spectra .
this is dealt with by minimizing @xmath29 also for @xmath28 .
these two expressions are combined into a single one depending on the parameter @xmath30 in the following manner : @xmath31 . by minimizing @xmath24 with respect to @xmath28
we get the backus gilbert inverse ( fig.[fig:1 ] ) .
( @xmath32 is merely a factor to get the two components to the same order of magnitude ) we will provide only a brief description of the method .
like at the maximum entropy method , we are also in need here of an _ a priori _ spectrum . this information is used to reduce the characteristic length scale of the problem .
we introduce the sum - of - squares of the spectrum s @xmath33-th derivative ( @xmath34 ) .
we propose to minimize some linear combination of the @xmath35 and the @xmath34 . as an _ a priori _ spectrum we fit a band grb function .
we find that there is only insignificant gain in the function to be minimized with respect to the band grb function .
having applied several of the above presented methods to invert batse mer spectra , we conclude that the methods give a consistent estimation of the incident spectrum each with regard to the minimizing function they use .
the svd method is the fastest , the backus gilbert requires the most time , but it is less sensitive to noise .
we have also found that the band grb function provides a good fit to the spectra we analyzed .
this research was supported by otka grant t048870 and by a grant from swedish wenner - gren foundations ( a.m. ) . | the gamma ray inverse problem is discussed .
four methods of spectral deconvolution are presented here and applied to the batse s mer data type .
we compare these to the band spectra .
[ 1999/12/01 v1.4c il nuovo cimento ] |
the recent discovery of superconductivity at @xmath440 k in 2 @xcite has opened new prospects for the understanding of the microscopic origin of high-@xmath2 superconductivity .
in contrast to the complex crystal structures , multicomponent nature and complications due to magnetism and strong electron - electron correlations of the high-@xmath2 cuprates , 2 has a simple hexagonal crystal structure ( alb@xmath5-type ) comprising close - packed mg@xmath6 layers alternating with graphite - like boron layers and @xmath7 electrons involved in the superconducting process .
several attempts @xcite have been made to introduce dopants in the mg layers in order to explore the relationship between @xmath2 , crystal structure and doping level .
the reported experiments invariably lead to a decrease in @xmath2 .
successful substitution at the b sites can be also achieved by carbon doping @xcite , which results in a significant contraction of the @xmath8 lattice parameter , but affects little the interlayer separation .
@xmath2 also decreases with increasing doping .
experimental studies of the superconducting properties of 2 show deviations from those calculated with the standard bcs theory @xcite .
various experiments , including scanning tunneling microscopy ( stm ) @xcite , point - contact spectroscopy @xcite , specific heat measurements @xcite , optical @xcite and raman spectroscopy @xcite suggested the existence of a secondary superconducting gap , implying that the simple one - band approach for 2 superconductivity must be extended .
on the basis of the electronic structure , the existence of multiple gaps has been invoked in order to explain the magnitude of @xmath2 in 2 @xcite .
in addition , solution of the full eliashberg equations at low temperature yields different gap values for the different parts of the fermi surface ( @xmath9 mev for 3d sheets and @xmath10 mev for 2d sheets ) @xcite .
although there is much support for the applicability of the multiband description to 2 , there is still some debate in the literature , in particular since some tunneling and nmr measurements show only a single gap @xcite .
measurements have also made significant contributions to the understanding of the superconducting properties of 2 , mainly through the precise determination of the penetration depth , @xmath11 from the zero - temperature extrapolated value of the @xmath12 spin depolarisation rate , @xmath13 .
early experiments @xcite proposed that the low temperature magnetic penetration depth of 2 shows a quadratic temperature dependence and were interpreted in terms of unconventional superconductivity with an energy gap that has nodes at certain points in the @xmath14 space . however , a systematic study by niedermayer _
et al . _
@xcite reported that the temperature evolution of the depolarisation rate , @xmath1 in polycrystalline 2 could be well interpreted in terms of a two - gap model . in the present paper
we employ the transverse - field ( tf ) variant of the technique to characterise the superconducting properties of carbon - doped 2 systems ( , @xmath0= 0.02 , 0.04 , 0.06 ) .
analysis of the experimental data in terms of a two - gap model reveals the effect of electron doping on the superconducting gaps and the anisotropic properties of the compounds .
polycrystalline ( @xmath0= 0.02 , 0.04 , 0.06 ) samples were prepared by reaction of mg , amorphous b and carbon black at 900@xmath15c for 2 h , as described elsewhere @xcite .
the measurements were performed with the gps spectrometer on the @xmath16m3 muon beamline at the paul scherrer institute ( psi ) , switzerland . pressed sample pellets were attached with low - temperature varnish on a ag sample - holder placed on the stick of a he continuous flow cryostat operating down to 1.8 k. after cooling the sample in an external field , @xmath17 to temperatures below @xmath2 in order to induce a homogenous flux line lattice , positive muons ( 100% spin - polarised ) with their initial spin polarisation transverse to the external field were implanted in the solid sample .
the implanted muons come to rest at an interstitial site and act as highly sensitive local magnetic probes . in the presence of local fields , @xmath18 , the @xmath12 spin undergoes larmor precession with frequency , @xmath19 , where @xmath20=13.553 khz / g is the muon gyromagnetic ratio .
the time evolution of the @xmath12 spin polarisation , @xmath21 is measured by monitoring the positrons preferentially emitted along the @xmath12 spin direction at the instant of muon decay . for type ii superconductors , @xmath21 is an oscillatory function with decreasing amplitude and the damping of the @xmath12 precession signal provides a measure of the inhomogeneity of the magnetic field , @xmath22 in the vortex state and hence of the magnetic penetration depth , @xmath11 .
for polycrystalline samples in the vortex state , the tf @xmath12 spin polarisation function is approximately gaussian ( @xmath23 ) and the depolarisation rate , @xmath1 is proportional to the second moment of the field distribution ( @xmath24@xmath25@xmath26@xmath27 ) . in the case of anisotropic ( @xmath28 4 )
type ii superconductors , @xmath1 is related to the in - plane magnetic penetration depth , @xmath29 via the relation @xcite : @xmath30=7.086\times10 ^ 4~\lambda_{ab}^{-2}[nm^{-2}]$ ] .
the anisotropy of the upper critical field in single crystalline 2 was reported as 6 ( at 5 k ) @xcite or 4.3 @xcite , while for polycrystalline samples and thin films , the reported values span a wide range .
the above equation holds in the london limit ( @xmath31 ) and in the absence of pinning - induced distortions of the vortex lattice .
the former restriction is valid for 2 ( @xmath32 7 nm @xcite , @xmath33 95 nm @xcite ) , while the latter is addressed below .
the london model predicts that the second moment of the magnetic field distribution of a perfect vortex lattice should be independent of @xmath17 for @xmath34 where @xmath35 is the distance between vortices . in order to check the influence of pinning on the depolarisation rate , @xmath1 ,
we have measured its field dependence at 5 k. our results show that @xmath36 increases almost linearly up to 50 mt , displays a peak at @xmath480 mt and then decreases rapidly , reaching a plateau at higher fields ( @xmath37 t ) in excellent agreement with the reported data for 2 @xcite .
we thus performed our measurements at an external field of 0.6 t in order to ensure formation of an ideal vortex lattice and avoid underestimation of @xmath38 .
[ f.1 ] presents the extracted temperature dependence of the tf- depolarisation rate at @xmath17= 0.6 t for the ( @xmath0= 0.02 , 0.04 , 0.06 ) samples .
the values of @xmath39 are derived after subtraction of the normal state temperature - independent depolarisation rate , @xmath40 ( @xmath41 ) . in the case of mgb@xmath42c@xmath43
, @xmath39 increases monotonically as the temperature decreases below @xmath435 k and reaches a plateau at @xmath44 6 k , remaining almost constant at lower temperatures .
analogous behaviour is observed for the other two compositions , except that as the doping level increases , the low - temperature plateau smooths out and the extrapolated value , @xmath45 shifts to lower values .
the experimental @xmath46 dependence is reproduced well by means of a two - gap model , while attempts to fit the experimental data with a @xmath47-law or an isotropic one - gap model led to unsatisfactory results ( inset in fig .
[ f.1 ] ) .
the two - gap model is based on the existence of two discrete superconducting gaps , @xmath48 and @xmath49 , at @xmath3= 0 k , both closing at @xmath2 and each associated with a different energy band . by assuming that the coupling between the two bands ( i.e. due to impurity or phonon scattering ) is sufficiently weak ( @xmath50 )
, @xmath46 can be expressed as @xcite : @xmath51 where @xmath52d\varepsilon\ ] ] @xmath53 is the fermi distribution of quasiparticles and @xmath54 the energy of the normal electrons relative to the fermi energy : @xmath55 each band is characterised by a partial sommerfeld constant , @xmath56 ( @xmath57 , where @xmath58 is the total sommerfeld constant ) . as the sommerfeld constant is proportional to the density - of - states at the fermi level , the ratios @xmath56/@xmath58 determine the partial @xmath59(@xmath60 ) for the two bands .
the temperature dependence of the band gaps is taken from bcs theory , i.e. @xmath61 where @xmath62 is the normalised bcs gap at the reduced temperature , @xmath63 @xcite .
the fitted parameters are summarised in table i. for comparison the corresponding values for 2 @xcite are also included .
the evolution of the depolarisation rate @xmath45 with @xmath2 and of the superconducting gap sizes , @xmath48(0 ) and @xmath49(0 ) with carbon content , @xmath0 are shown in fig .
@xmath45 decreases monotonically with increasing carbon content and decreasing @xmath2 ( d@xmath45/d@xmath2= @xmath640.25 @xmath65s@xmath66k@xmath66 ) .
the @xmath411% decrease in @xmath45 at @xmath0= 0.06 reflects an increase of @xmath46% in the in - plane magnetic penetration depth , @xmath67 .
in addition , as the carbon content increases , the two gap sizes at 0 k shift to smaller values at nearly the same rate ( d@xmath68/d@xmath0= @xmath6420.4 mev , @xmath69= 1,2 ) .
@xmath70/@xmath58 which is a measure of the partial @xmath71(@xmath72 ) for the band with the larger gap also shows a tendency to increase with increasing electron doping .
the superfluid plasma frequency in the @xmath73 plane , @xmath74 which is related to the charge density of the superfluid condensate at 0 k is also included in table i. @xmath13 & @xmath67 & @xmath48 & @xmath49 & @xmath70/@xmath58 & @xmath75 & @xmath76(@xmath77/@xmath78)@xmath66 + & ( k ) & ( @xmath65s@xmath66 ) & ( nm ) & ( mev ) & ( mev ) & & ( ev ) & ( @xmath7910@xmath80 @xmath81 ) + 0 & 38.3 & 7.9 & 94.7 & 6.0(0 ) & 2.6(2 ) & 0.6(2 ) & 2.08(1 ) & 3.15 + 0.02 & 37.2 & 7.5(1 ) & 97.2(6 ) & 5.8(4 ) & 2.0(2 ) & 0.7(2 ) & 2.03(1 ) & 2.99(4 ) + 0.04 & 36.1 & 7.3(1 ) & 98.5(7 ) & 5.2(2 ) & 1.5(1 ) & 0.8(1 ) & 2.00(1 ) & 2.91(4 ) + 0.06 & 34.8 & 7.0(1 ) & 100.6(7 ) & 4.8(3 ) & 1.3(1 ) & 0.8(2 ) & 1.96(1 ) & 2.79(3 ) + in the london model at the clean limit , the magnetic penetration depth , @xmath11 is related to the superconducting carrier density , @xmath76 and the effective mass , @xmath82 by @xmath83 where @xmath84 is the coherence length and @xmath85 the mean free path . by assuming that is in the clean - limit ( @xmath84= 50 , @xmath86 1000 ) , the term in brackets is close to unity allowing us to estimate the ratio , @xmath87 .
the calculated values of @xmath88 ( where @xmath89 is the effective mass enhancement in the b layers in units of @xmath78 ) for ( table i ) are larger than those obtained for high-@xmath2 superconductors and decrease quasilinearly with increasing doping .
in addition , @xmath76 can be converted first into a 2d carrier density , @xmath90 on the b planes by multiplying by the interlayer distance , @xmath91 and then into an effective fermi temperature , @xmath92 through the expression , @xmath93 .
@xmath92 decreases from @xmath43200 k in 2to @xmath42700 k in 6 resulting in @xmath94 0.01 .
uemura @xcite has established that in contrast to the conventional bcs superconductors in which @xmath950.01 , the cuprates , fullerides and other exotic superconductors have rather high @xmath2 with respect to @xmath92 with @xmath94 0.01 - 0.1 .
it is noticeable that the 2 superconductors also share this feature and are located in the uemura plot between conventional @xmath96-wave and unconventional high-@xmath2 materials .
the electronic band structure of 2 has been extensively investigated @xcite .
the valence band of 2 is made up predominantly of b 2@xmath97 states , which form two distinct sets of bands of @xmath98 and @xmath99 type whose @xmath14 dependence differs considerably .
the most pronounced dispersion for the b @xmath100 states is along the @xmath101-k direction of the brillouin zone ( bz ) , while a flat zone is formed in the @xmath102 direction ( @xmath101-a ) , reflecting the 2d character of the boron lattice .
these @xmath1 bands are partially unoccupied creating a hole - type conduction band which gives rise to two 2d light - hole and heavy - hole sheets forming coaxial cylinders along the @xmath101-a bz direction .
the strong coupling of these holes to the optical bond stretching modes drives the superconductivity in 2 @xcite .
in addition , b @xmath16 states dominate at the bottom of the conduction band , while the fermi surface associated with these bands consists of a 3d tubular network .
these bands exhibit maximum dispersion along the @xmath101-a direction of the bz .
theoretical calculations estimate that 44% of @xmath71(@xmath60 ) comes from the 2d @xmath1 cylindrical sheets and the rest from the 3d @xmath16 sheets @xcite . for the superconducting state of 2 ,
the gap is nonzero everywhere on the fermi surface and the gap values are grouped in two distinct regions @xcite .
the larger superconducting gap , @xmath48 on the 2d @xmath1 cylindrical sheets has an average value of @xmath46.8 mev , while the smaller one , @xmath49 associated with the 3d @xmath16-sheets an average value of @xmath41.8 mev .
although the electronic structure and the superconducting behaviour of 2 has been studied theoretically in detail , there is no systematic investigation concerning the effect of doping . in the framework of the rigid band model
, it is expected @xcite that doping of the b sublattice with c should lead to a shift of the fermi level to higher energies in the region of the dos minimum ( pseudogap ) .
a rigid band model estimate gives a value of @xmath40.16 electrons / cell in order to fill the @xmath1 bands @xcite .
electrons are also added to the @xmath16 bands that lie in the same energy range , leading to a total doping level @xmath40.25 electrons / cell .
thus it is expected that electron doping should lead to both superconducting gaps decreasing gradually .
however , the differences in the slope @xmath64dln@xmath2/d@xmath103 between high - pressure experiments ( @xmath40.28 @xmath104 ) on 2 @xcite and chemical substitution @xcite ( @xmath40.37 @xmath104 ) , imply that the effect of carbon doping is more complicated due to the b layer contraction and theoretical studies beyond the rigid band approximation are necessary .
carbon substitution should also lead to an increase in the interband impurity scattering , which should result in the size of the @xmath1- and @xmath16-gaps converging to the same value .
however , theoretical calculations @xcite have shown that the particular electronic structure of 2 results in extremely weak @xmath105 impurity scattering , even for low quality samples and in the presence of mg vacancies , mg - substitutional impurities and b - site substitutions by n or c. the dominant mechanism for impurity scattering is due to intraband scattering of the @xmath1 and @xmath16 bands with the scattering rate inside the @xmath16 bands greater than that of the @xmath1 bands .
intraband scattering does not change @xmath2 and the gap values but influences the penetration depth ( anderson s theorem ) @xcite .
our experimental data can be described well with a two - gap model in which impurity scattering is essentially ignored .
this indicates that interband impurity scattering is relatively weak and both superconducting gaps are preserved , at least up to @xmath0= 0.06 .
carbon doping also reduces both superconducting gaps with this effect relatively more pronounced for the smaller gap , @xmath49 associated with the 3d sheets .
the error in @xmath70/@xmath106 ( @xmath4@xmath107(@xmath72)/@xmath108(@xmath72 ) ) is quite large and hence it is difficult to extract reliably its dependence on doping .
nonetheless , there is a tendency for @xmath70/@xmath106 to increase with increasing doping level . in any case , systematic theoretical studies are necessary to shed more light on this issue .
in conclusion , analysis of the temperature dependence of the tf- depolarisation rate for ( @xmath0= 0.02 , 0.04 , 0.06 ) shows that increasing electron doping is accompanied by a decrease of the low - temperature @xmath1(0 ) and the spectral weight , @xmath109 , consistent with the decreasing density - of - states at the fermi level . within a two - gap model ,
both superconducting gap sizes decrease almost linearly with the smaller gap affected more on doping , while the interband scattering remains relatively weak up to at least @xmath0= 0.06 .
we thank c. niedermayer and i. i. mazin for helpful discussions , psi for provision of beamtime and a. amato , k. brigatti , a. lappas and i. margiolaki for help with the experiments .
we acknowledge support from the royal society and the marie curie fellowship programme of the eu `` improving the human research potential '' under contract numbers hpmf - ct-2001 - 01435 ( k. papagelis ) and -01436 ( j. arvanitidis ) . | the evolution of the superconducting properties of the carbon - doped 2 superconductors , ( @xmath0= 0.02 , 0.04 , 0.06 ) have been investigated by the transverse - field muon spin rotation ( tf- ) technique .
the low - temperature depolarisation rate , @xmath1(0 ) at 0.6 t which is proportional to the second moment of the field distribution of the vortex lattice decreases monotonically with increasing electron doping and decreasing @xmath2 .
in addition , the temperature dependence of @xmath1(@xmath3 ) has been analysed in terms of a two - gap model .
the size of the two superconducting gaps decreases linearly as the carbon content increases . |
effective field theory ( eft ) has become a powerful tool in the description of the strong interactions at low energies .
the central idea is due to weinberg @xcite : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` ... if one writes down the most general possible lagrangian , including all terms consistent with assumed symmetry principles , and then calculates matrix elements with this lagrangian to any given order of perturbation theory , the result will simply be the most general possible s matrix consistent with analyticity , perturbative unitarity , cluster decomposition and the assumed symmetry principles . ''
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the prerequisite for an effective field theory program is ( a ) a knowledge of the most general effective lagrangian and ( b ) an expansion scheme for observables in terms of a consistent power counting method . the application of these ideas to the interactions among the goldstone bosons of spontaneous chiral symmetry breaking in qcd results in mesonic chiral perturbation theory ( chpt ) @xcite ( see , e.g. , refs .
@xcite for an introduction and overview ) . in the following
we will outline some recent developments in devising a renormalization scheme leading to a simple and consistent power counting for the renormalized diagrams of a manifestly lorentz - invariant approach to baryon chpt @xcite .
the effective lagrangian relevant to the one - nucleon sector consists of the sum of the purely mesonic and @xmath1 lagrangians , respectively , @xmath2 which are organized in a derivative and quark - mass expansion @xcite .
for example , the lowest - order basic lagrangian @xmath3 , already expressed in terms of renormalized parameters and fields , is given by @xmath4 where @xmath5 , @xmath6 , and @xmath7 denote the chiral limit of the physical nucleon mass , the axial - vector coupling constant , and the pion - decay constant , respectively
. the ellipsis refers to terms containing external fields and higher powers of the pion fields .
when studying higher orders in perturbation theory one encounters ultraviolet divergences . as a preliminary step , the loop integrals are regularized , typically by means of dimensional regularization . for example , the simplest dimensionally regularized integral relevant to chpt is given by @xmath8+o(n-4),\end{aligned}\ ] ] where @xmath9 - 1 $ ] approaches infinity as @xmath10 .
the t hooft parameter @xmath11 is responsible for the fact that the integral has the same dimension for arbitrary @xmath12 . in the process of renormalization
the counter terms are adjusted such that they absorb all the ultraviolet divergences occurring in the calculation of loop diagrams .
this will be possible , because we include in the lagrangian all of the infinite number of interactions allowed by symmetries @xcite . at the end
the regularization is removed by taking the limit @xmath10 .
moreover , when renormalizing , we still have the freedom of choosing a renormalization prescription . in this context
we will adjust the finite pieces of the renormalized couplings such that renormalized diagrams satisfy the following power counting : a loop integration in @xmath12 dimensions counts as @xmath13 , pion and fermion propagators count as @xmath14 and @xmath15 , respectively , vertices derived from @xmath16 and @xmath17 count as @xmath18 and @xmath19 , respectively .
here , @xmath20 collectively stands for a small quantity such as the pion mass , small external four - momenta of the pion , and small external three - momenta of the nucleon .
the power counting does not uniquely fix the renormalization scheme , i. e. there are different renormalization schemes leading to the above specified power counting .
.,height=75 ] as an example , consider the one - loop contribution of fig .
[ fig : nucleonselfenergypionloop ] to the nucleon self energy .
after renormalization , we would like to have the order @xmath21 the application of the @xmath22 renormalization scheme of chpt @xcite indicated by `` r''yields @xmath23= { \cal o}(q^2),\ ] ] where @xmath24 is the lowest - order expression for the squared pion mass in terms of the low - energy coupling constant @xmath25 and the average light - quark mass @xmath26 @xcite .
the @xmath22-renormalized result does not produce the desired low - energy behavior which , for a long time , was interpreted as the absence of a systematic power counting in the relativistic formulation of chpt .
several methods have been suggested to obtain a consistent power counting in a manifestly lorentz - invariant approach .
we will illustrate the underlying ideas in terms of a typical one - loop integral in the chiral limit @xmath27[k^2+i0^+]},\ ] ] where @xmath28 is a small quantity . applying the dimensional counting analysis of ref .
@xcite the result of the integration is of the form @xmath29 where @xmath7 and @xmath30 are hypergeometric functions which are analytic for @xmath31 for any @xmath12 . in the infrared regularization of becher and leutwyler @xcite
one makes use of the feynman parametrization @xmath32 ^ 2}\ ] ] with @xmath33 and @xmath34 .
the resulting integral over the feynman parameter @xmath35 is then rewritten as @xmath36 where the first , so - called infrared ( singular ) integral satisfies the power counting , while the remainder violates power counting but turns out to be regular and can thus be absorbed in counter terms .
the central idea of the extended on - mass - shell ( eoms ) scheme @xcite consists of subtracting those terms which violate the power counting as @xmath10 .
since the terms violating the power counting are analytic in small quantities , they can be absorbed by counter term contributions . in the present case ,
we want the renormalized integral to be of the order @xmath37 .
to that end one first expands the integrand in small quantities and subtracts those integrated terms whose order is smaller than suggested by the power counting .
the corresponding subtraction term reads @xmath38[k^2+i0^+]}\right|_{p^2=m^2}\ ] ] and the renormalized integral is written as @xmath39 as @xmath10 . * using a suitable renormalization condition one obtains a consistent power counting in manifestly lorentz - invariant baryon chpt including , e.g. , vector mesons @xcite or the @xmath40 resonance @xcite as explicit degrees of freedom .
* the infrared regularization of becher and leutwyler @xcite has been reformulated in a form analogous to the eoms renormalization @xcite . *
the application of both infrared and extended on - mass - shell renormalization schemes to multi - loop diagrams was explicitly demonstrated by means of a two - loop self - energy diagram @xcite . in both cases
the renormalized diagrams satisfy a straightforward power counting .
using the reformulated infrared regularization @xcite we have calculated the nucleon mass up to and including order @xmath41 in the chiral expansion @xcite : @xmath42 in eq .
( [ h1:emff : massexp ] ) , @xmath5 denotes the nucleon mass in the chiral limit , @xmath43 is the leading term in the chiral expansion of the square of the pion mass , @xmath11 is the renormalization scale ; all the coefficients @xmath44 have been determined in terms of infrared renormalized parameters .
the results of ref .
@xcite represent the first complete two - loop calculation in manifestly lorentz - invariant baryon chpt .
our results for the renormalization - scheme - independent terms agree with the heavy - baryon chpt results of ref .
@xcite .
pion mass dependence of the term @xmath45 ( solid line ) for @xmath46 mev . for comparison
also the term @xmath47 ( dashed line ) is shown.,scaledwidth=45.0% ] the numerical contributions from higher - order terms can not be calculated so far since , starting with @xmath48 , most expressions in eq .
( [ h1:emff : massexp ] ) contain unknown low - energy coupling constants ( lecs ) from the lagrangians of order @xmath49 and higher .
the coefficient @xmath50 is free of higher - order lecs and is given in terms of the axial - vector coupling constant @xmath6 and the pion - decay constant @xmath7 . while the values for both @xmath6 and @xmath7 should be taken in the chiral limit , we evaluate @xmath50 using the physical values @xmath51 and @xmath52 mev . setting @xmath53 , @xmath54 mev , and @xmath55 mev
we obtain @xmath56 mev .
this amounts to approximately @xmath57% of the leading nonanalytic contribution at one - loop order , @xmath47 .
figure [ fig : nucleonmass ] shows the pion mass dependence of the term @xmath45 ( solid line ) in comparison with the term @xmath58 ( dashed line ) for @xmath46 mev . for @xmath59 the @xmath50 term is as large as the @xmath60 term .
imposing the relevant symmetries such as translational invariance , lorentz covariance , the discrete symmetries , and current conservation , the nucleon matrix element of the electromagnetic current operator @xmath61 can be parameterized in terms of two form factors , @xmath62u(p),\quad n = p , n,\ ] ] where @xmath63 , @xmath64 , and @xmath65 is the proton mass . at @xmath66 ,
the so - called dirac and pauli form factors @xmath67 and @xmath68 reduce to the charge and anomalous magnetic moment in units of the elementary charge and the nuclear magneton @xmath69 , respectively , @xmath70 the sachs form factors @xmath71 and @xmath72 are linear combinations of @xmath67 and @xmath68 , @xmath73 and , in the non - relativistic limit , their fourier transforms are commonly interpreted as the distribution of charge and magnetization inside the nucleon .
the description of the electromagnetic form factors of the nucleon presents a stringent test for any theory or model of the strong interactions .
calculations in lorentz - invariant baryon chpt up to fourth order fail to describe the proton and nucleon form factors for momentum transfers beyond @xmath74 @xcite . in ref .
@xcite it was shown that the inclusion of vector mesons can result in the re - summation of important higher - order contributions . in ref .
@xcite the electromagnetic form factors of the nucleon up to fourth order have been calculated in manifestly lorentz - invariant chpt with vector mesons as explicit degrees of freedom .
a systematic power counting for the renormalized diagrams has been implemented using both the extended on - mass - shell renormalization scheme and the reformulated version of infrared regularization .
the inclusion of vector mesons results in a considerably improved description of the form factors ( see fig . [ h1:emff : g : neu ] ) .
the most dominant contributions come from tree - level diagrams , while loop corrections with internal vector meson lines are small @xcite .
assuming isospin symmetry , the most general parametrization of the isovector axial - vector current evaluated between one - nucleon states is given by @xmath75 \frac{\tau^a}{2}u(p),\ ] ] where @xmath63 , @xmath64 , and @xmath76 denotes the nucleon mass .
@xmath77 is called the axial form factor and @xmath78 is the induced pseudoscalar form factor .
the value of the axial form factor at zero momentum transfer is defined as the axial - vector coupling constant , @xmath79 , and is quite precisely determined from neutron beta decay .
the @xmath80 dependence of the axial form factor can be obtained either through neutrino scattering or pion electroproduction .
the second method makes use of the so - called adler - gilman relation @xcite which provides a chiral ward identity establishing a connection between charged pion electroproduction at threshold and the isovector axial - vector current evaluated between single - nucleon states ( see , e.g. , ref.@xcite for more details ) .
the induced pseudoscalar form factor @xmath78 has been investigated in ordinary and radiative muon capture as well as pion electroproduction ( see ref .
@xcite for a review ) . in ref .
@xcite the form factors @xmath81 and @xmath82 have been calculated in manifestly lorentz - invariant baryon chpt up to and including order @xmath49 .
in addition to the standard treatment including the nucleon and pions , the axial - vector meson @xmath83 has also been considered as an explicit degree of freedom .
the inclusion of the axial - vector meson effectively results in one additional low - energy coupling constant which has been determined by a fit to the data for @xmath77 .
the inclusion of the axial - vector meson results in an improved description of the experimental data for @xmath81 ( see fig . [ h1_axff_gawith ] ) , while the contribution to @xmath82 is small .
left panel : axial form factor @xmath81 in manifestly lorentz - invariant chpt at @xmath49 including the axial - vector meson @xmath83 explicitly .
full line : result in infrared renormalization , dashed line : dipole parametrization .
the experimental data are taken from ref .
right panel : the induced pseudoscalar form factor @xmath82 in manifestly lorentz - invariant chpt at @xmath49 including the axial - vector meson @xmath83 explicitly .
full line : result with axial - vector meson ; dashed line : result without axial - vector meson .
one can clearly see the dominant pion pole contribution at @xmath84 . ]
left panel : axial form factor @xmath81 in manifestly lorentz - invariant chpt at @xmath49 including the axial - vector meson @xmath83 explicitly .
full line : result in infrared renormalization , dashed line : dipole parametrization .
the experimental data are taken from ref .
right panel : the induced pseudoscalar form factor @xmath82 in manifestly lorentz - invariant chpt at @xmath49 including the axial - vector meson @xmath83 explicitly .
full line : result with axial - vector meson ; dashed line : result without axial - vector meson .
one can clearly see the dominant pion pole contribution at @xmath84 . ]
the pion - nucleon form factor @xmath85 may be defined in terms of the pseudoscalar quark density @xmath86 and the average light - quark mass @xmath26 as @xcite @xmath87 where @xmath63 , @xmath64 , and @xmath88 is the corresponding interpolating pion field .
the pion - nucleon coupling constant is given by @xmath89 .
using the ( qcd- ) partially conserved axial - vector current ( pcac ) relation , @xmath90 , the pion - nucleon form factor is completely given in terms of the axial and the induced pseudoscalar form factors , @xmath91 this is an exact relation which holds true for any value of @xmath80 .
the result at @xmath49 is given by @xcite @xmath92 where @xmath93 denotes the goldberger - treiman discrepancy .
the chiral expansion of the pion - nucleon coupling constant can be found in ref .
@xcite .
diagrams contributing to the pion - nucleon vertex at @xmath49.,scaledwidth=50.0% ] finally , we would like to stress that one carefully has to distinguish between the pion - nucleon form factor @xmath85 of eq.([gpin ] ) on the one hand and the renormalized pion - nucleon vertex in chpt on the other hand .
the unrenormalized vertex , when evaluated between on - shell nucleon states , is of the form @xmath94 at @xmath49 one needs to calculate the diagrams shown in fig .
[ fig : pinnvertex ] . in general , the pion - nucleon vertex depends on the choice of the field variables in the effective lagrangian . only at @xmath95 , we expect the same coupling strength , since both @xmath96 and the field @xmath97 of eq .
( [ lpin1 ] ) serve as interpolating pion fields : @xmath98 on the other hand , for arbitrary @xmath99 @xmath100 in the present case , the pion - nucleon vertex is only an auxiliary quantity , whereas the `` fundamental '' quantity ( entering chiral ward identities ) is the matrix element of the pseudoscalar density .
both the infrared regularization and the eoms scheme allow for a simple and consistent power counting in manifestly lorentz - invariant baryon chiral perturbation theory .
we have discussed some results of a two - loop calculation of the nucleon mass .
the inclusion of vector and axial - vector mesons as explicit degrees of freedom leads to an improved phenomenological description of the electromagnetic and axial form factors , respectively .
work on the application to electromagnetic processes such as compton scattering and pion production is in progress .
i would like to thank d. djukanovic , t. fuchs , j. gegelia , g. japaridze , and m. r. schindler for the fruitful collaboration on the topics of this talk .
this work was made possible by the financial support from the deutsche forschungsgemeinschaft ( sfb 443 and sche 459/2 - 1 ) and the eu integrated infrastructure initiative hadron physics project ( contract number rii3-ct-2004 - 506078 ) . 99 -2pt s. weinberg , _ physica _ a 96 ( 1979 ) 327 j. gasser and h. leutwyler , _ annals phys . _ 158 ( 1984 ) 142 s. scherer , _ adv .
_ 27 ( 2003 ) 277 s. scherer and m. r. schindler , arxiv : hep - ph/0505265 j. bijnens , _ prog .
_ 58 ( 2007 ) 521 v. bernard , arxiv:0706.0312 [ hep - ph ] j. gasser , m. e. sainio , and a. varc , _ nucl .
_ b 307 ( 1988 ) 779 s. weinberg , _ the quantum theory of fields .
vol . 1 : foundations _ ( cambridge university press , cambridge , uk , 1995 ) j. gegelia , g. s. japaridze , and k. s. turashvili , _ theor
. math .
_ 101 ( 1994 )
1313 t. becher and h. leutwyler , _ eur .
j. _ c 9 ( 1999 ) 643 t. fuchs , j. gegelia , g. japaridze , and s. scherer , _ phys .
d _ 68 ( 2003 ) 056005 t. fuchs , m. r. schindler , j. gegelia , and s. scherer , _ phys .
_ b 575 ( 2003 ) 11 c. hacker , n. wies , j. gegelia , and s. scherer , _ phys . rev .
_ c 72 ( 2005 ) 055203 m. r. schindler , j. gegelia , and s. scherer , _ phys .
_ b 586 ( 2004 ) 258 m. r. schindler , j. gegelia , and s. scherer , _ nucl .
_ b 682 ( 2004 ) 367 m. r. schindler , d. djukanovic , j. gegelia , and s. scherer , _ phys .
_ b 649 ( 2007 ) 390 m. r. schindler , d. djukanovic , j. gegelia , and s. scherer , arxiv:0707.4296 [ hep - ph ] j. a. mcgovern and m. c. birse , _ phys .
_ b 446 ( 1999 ) 300 b. kubis and u .- g .
meissner , _ nucl .
phys . _ a 679 ( 2001 ) 698 t. fuchs , j. gegelia , and s. scherer , _ j. phys .
_ g 30 ( 2004 ) 1407 m. r. schindler , j. gegelia , and s. scherer , _ eur .
j. _ a 26 ( 2005 ) 1 j. friedrich and th .
walcher , _ eur .
j. _ a 17 ( 2003 ) 607 s. l. adler and f. j. gilman , _ phys .
_ 152 ( 1966 ) 1460 v. bernard , l. elouadrhiri , and u. g. meissner , _ j. phys .
g _ 28 ( 2002 ) r1 t. fuchs and s. scherer , _ phys .
_ c 68 ( 2003 ) 055501 t. gorringe and h. w. fearing , _ rev .
_ 76 ( 2004 ) 31 m. r. schindler , t. fuchs , j. gegelia , and s. scherer , _ phys .
c _ 75 ( 2007 ) 025202 | we will give a short introduction to the one - nucleon sector of chiral perturbation theory and will address the issue of a consistent power counting and renormalization .
we will discuss the infrared regularization and the extended on - mass - shell scheme . both allow for the inclusion of further degrees of freedom beyond pions and nucleons and the application to higher - loop calculations .
as applications we consider the chiral expansion of the nucleon mass to order @xmath0 and the inclusion of vector and axial - vector mesons in the calculation of nucleon form factors . |
the statistical analysis that combines the results of several independent is known as meta - analysis and it is used in clinical trails and behavioral sciences .
consider we have @xmath0 independent normal populations with means @xmath1 and variances @xmath2 .
also we have a random samples of sizes @xmath3 , @xmath4 from each one .
we denote these samples by @xmath5 @xmath6 @xmath7 , where @xmath8 and @xmath9 are constant .
the problem of interest is to combine the summary statistics of samples for statistical inference about the parameter @xmath10 .
the statistical analysis that combines the results of several independent used in clinical trails and behavioral sciences . if @xmath11 and @xmath12 then , @xmath13 and this problem is known as the common mean for several normal populations .
there are some inference for this problem in statistical literature .
for example see ; krishnamoorthy and lu ( 2003 ) , lin and lee ( 2005 ) .
if @xmath11 and @xmath14 then @xmath15 and this is equivalent to problem of common mean of several lognormal populations .
our interest in this paper is inference about this problem . for
the common lognormal mean , a few authors proposed approximate methods : ahmed et al ( 2001 ) proposed an estimator and approximate confidence interval for the common lognormal mean ; baklizi and ebrahem ( 2005 ) studied several types of large samples and bootstrap intervals ; gupta and li ( 2005 ) developed procedures for estimating the common mean and investigated the performance of the resulting confidence interval for two lognormal populations . in this paper , we first propose estimation of @xmath16 when the variances , @xmath17 are known .
then two methods are given that are applicable for both hypothesis testing and interval estimation for @xmath16 , based on the concepts of generalized @xmath18-value and generalized confidence interval .
these methods are based on extending the method of krishnamoorthy and lu ( 2003 ) and the method of lin and lee ( 2005 ) , which are used for the problem of common mean of several normal populations .
our methods also are applicable for the common mean of several lognormal for the interval mean of @xmath0 lognormal populations .
this cahpter also is devoted to a short review regarding the existing method for inference of the common lognormal mean and application of our two methods for this problem .
finally , we give a numerical example for the common lognormal mean and by monte carlo simulation , we compare the coverage probabilities , size and power of these methods for the common mean of two lognormal populations .
let @xmath19 , @xmath4 , @xmath20 , where @xmath8 @xmath9 are constants and @xmath2 s are known .
the estimator @xmath21 is umvue and mle for @xmath10 and @xmath22 the probability density function for @xmath23 is @xmath24 since the distribution of @xmath25 is from exponential family , in the form @xmath26 , then @xmath27 is umvue for @xmath28 and @xmath29 is umvue for @xmath10 ( see casella and berger , 1990 , page 263)@xmath30 it is easy to prove the rest of the theorem . if b=0 then @xmath31 is the best linear unbiased estimator for @xmath32 if @xmath33 i.e. @xmath34 is a lognormal variable , then @xmath35 is umvue for @xmath36 , but the mle of @xmath37 is @xmath38 if @xmath2 are unknown , then we can not find a closed form for mle s of @xmath10 ; we have to use a numerical approximation
suppose @xmath19 , @xmath4 , @xmath20 , where @xmath8 @xmath9 are constants . for the @xmath39th population ,
let @xmath40 be the sample mean and sample variance . in this section , by using the idea of generalized @xmath18-value and by extending ( i ) the method of krishnamoorthy and lu ( 2003 ) and ( ii ) the method of lin and lee ( 2005 ) , for the problem of common mean of normal populations , we give two generalized pivot variables for interval estimation and hypothesis testing for @xmath41 and we obtain two generalized @xmath18-values for testing hypothesis @xmath42 it is clear that @xmath43 , @xmath4 .
therefore , the generalized pivot variable for estimating @xmath10 based on the @xmath39th sample is @xmath44 where @xmath45 and @xmath46,@xmath47 is the observed value of @xmath48 , @xmath49 .
the generalized pivot variable for estimating @xmath2 based on the @xmath39th sample is given by @xmath50 where @xmath51 are independent @xmath52 random variables ( weerahandi , 1995 ) .
the generalized variable that we want to propose is a weighted average of the generalized pivot variables @xmath53 in ( [ eq2.2.2 ] ) .
the weights are inversely proportional to the generalized pivot variables @xmath54 in ( [ eq2.2.3 ] ) for the variances , and they are directly proportional to the sample sizes .
( see krishnamoorthy and lu , 2003 ) .
let @xmath55 and @xmath56 , with the observed values @xmath57 and @xmath58 , respectively .
then , the generalized variable can be expressed as @xmath59 } { a\sum\limits_{j=1}^{k}\dfrac{n_{j}v_{j}}{(n_{j}-1)s_{j}^{2}}}-\mu \\ & = & \sum\limits_{i=1}^{k}w_{i}t_{i}^{\ast } -\mu , \nonumber\end{aligned}\ ] ] where the weights are @xmath60 the distribution of @xmath61 is an increasing function with respect to @xmath10 .
therefore , the generalized @xmath18-value for ( [ eq2.2.2 ] ) is given by @xmath62 this generalized @xmath18-value can be well approximated by a monte carlo simulation using the following algorithm : [ alg2.1 ] for a given @xmath63 , @xmath64 and ( @xmath65 : for @xmath66 generate @xmath67 @xmath68 generate @xmath69 @xmath68 generate @xmath70 @xmath68 compute @xmath71 compute @xmath72 ( end @xmath73 loop ) let @xmath74 if @xmath75 , else @xmath76 . then @xmath77 is a monte carlo estimate of the generalized @xmath18-value for ( [ eq2.2.5 ] ) .
@xmath78 is a generalized pivot variable for @xmath10 and we can use that to obtain a generalized confidence interval for @xmath16 .
if @xmath11 and @xmath79 then @xmath80 } { \sum\limits_{j=1}^{k}\dfrac{n_{j}v_{j}}{(n_{j}-1)s_{j}^{2}}}-\mu\ ] ] and this generalized variable is introduced by krishnamoorthy and lu ( 2003 ) for inference on the common mean of several normal populations . from theorem 1 , we have @xmath81 we know that @xmath82 is a generalized pivot variable for @xmath83 @xmath6 where @xmath84 let @xmath55 and @xmath85 , with the observed values @xmath57 and @xmath86 , respectively .
we define a generalized variable for @xmath10 based on the umvue for @xmath10 in ( [ eq2.1.1 ] ) by @xmath87 the distribution of @xmath88 is an increasing function with respect to @xmath10 , and therefore the generalized @xmath18-value for testing ( [ eq2.2.1 ] ) is @xmath89 , \nonumber\end{aligned}\ ] ] where @xmath90 and @xmath91 is distribution function of the standard normal variable and expectation is taken with respect to chi - square random variables with @xmath92 @xmath6 degrees of freedom .
this generalized @xmath18-value can be well approximated by a monte carlo simulation like the algorithm [ alg2.1 ] .
@xmath93 in ( 2.10 ) is a generalized pivot variable for @xmath10 and we can use that to obtain a generalized confidence interval for @xmath16 . if @xmath11 and @xmath79 then @xmath94 which is a generalized variable , introduced by lin and lee ( 2005 ) , for the common mean of several normal populations . for testing the hypothesis of the form
@xmath95 the @xmath18-value is @xmath96 and @xmath97 can be rejected when @xmath98 .
consider independent @xmath34 with lognormal distribution , for @xmath4 , @xmath20 , and assume that @xmath99 , where @xmath100 , i.e. , the @xmath0 lognormal populations have common mean @xmath101 therefore , we have @xmath102 , where @xmath103 , and to find a confidence interval for @xmath104 , it is enough to have a confidence interval for @xmath16 , and a hypothesis test for @xmath104 is equivalent to a hypothesis test for @xmath16 .
for example the hypothesis test @xmath105 is equivalent to @xmath106 it is useful to review the existing methods for the problem of common lognormal mean .
let @xmath107 @xmath108 , @xmath109 then a combined sample estimate of @xmath110 is given by @xmath111 where @xmath112 @xmath113 , @xmath114 and @xmath115 the estimator @xmath116 is asymptotically normal with mean @xmath117 and asymptotic variance @xmath118 , which can be estimated by @xmath119 therefore , a @xmath120 confidence interval for @xmath117 is @xmath121 the acceptance set for all @xmath117 is @xmath122 this is a quadratic function in @xmath117 whose two roots can be found directly .
since the coefficient of @xmath123 in this expression is positive , it follows that the set of all values of @xmath117 between the two roots is the desired confidence interval .
let @xmath124 @xmath125 be a vector of parameters , where @xmath126 @xmath127 and @xmath128 is the common mean .
the joint log - likelihood function based on the log - transformed data of two independent log - normal populations is given by @xmath129 where @xmath130 let @xmath31 is mle for @xmath32 the asymptotic variance of @xmath31 is @xmath131 where @xmath132 and @xmath133 are mles for @xmath134 and @xmath135 a @xmath120 confidence interval for @xmath136 is @xmath137 in fact , the problem of common lognormal mean is a special case of our model when @xmath11 and @xmath138 thus , the generalized variable in ( [ eq2.2.4 ] ) becomes @xmath139 } { \sum\limits_{j=1}^{k}\dfrac{n_{j}v_{j}}{(n_{j}-1)s_{j}^{2}}}-\mu , \ ] ] and the generalized variable in ( [ eq2.2.7 ] ) becomes @xmath140
in this section , we give a numerical example and compare our methods with other methods for the problem of common lognormal mean .
the data come from the regenstrief medical record system ( rmrs ) ( mcdonald et al , 1988 ; zhou et al , 1997 ) on effects of race on medical charges of patients with type i diabetes who had received inpatient or outpatient care at least two occasions during the period from 1 january 1993 , through 30 june 1994 .
the data set consists of 119 african american patients and 106 white patients . the mean medical charges and their corresponding variance for the african american and white groups are given in table [ table2.1 ] .
gupta , r. c. and li , x. ( 2005 ) .
statistical inferences on the common mean of two log - normal distributions and some applications in reliability , appeared in _ computational statistics and data analysis_. krishnamoorthy , k. and mathew , t. ( 2003 ) .
inferences on the means of lognormal distributions using generalized p - values and generalized confidence interval , _ journal of statistical planning and inference _ , 115 , 103 - 121 . | a hypothesis testing and an interval estimation are studied for the common mean of several lognormal populations .
two methods are given based on the concept of generalized p - value and generalized confidence interval .
these new methods are exact and can be used without restriction on sample sizes , number of populations , or difference hypotheses . a simulation study for coverage probability , size and power
shown that the new methods are better than the existing methods .
a numerical example is given with some real medical data .
kewwords : lognormal population , common mean , generalized variable , generalized p - value , generalized confidence interval . |
the emission from active galactic nuclei ( agn ) is typically variable , with faster variability seen at shorter wavelengths .
this variability can be used as a tool to probe the surrounding structures , with reverberation mapping of the broad line region being an established technique .
however , the same techniques can be used to probe the structure of the accretion flow itself .
hard x - ray illumination of the disc should produce a lagged and smeared thermal reprocessing signal .
larger radii in the disc produce lower temperature emission , so this disc reprocessing picture predicts longer lags at longer wavelengths .
such differential lags are now starting to be seen ( sergeev et al .
2005 ; mchardy et al .
2014 ; edelson et al . 2015 ; fausnaugh et al .
2015 ; mchardy et al .
2016 ) confirming qualitatively that we are indeed seeing reprocessing from radially extended , optically thick material , as expected from a disc . however ,
quantitatively , the picture runs into difficulties .
it has long been known that the implied size - scales are larger by a factor of a few compared to the expected sizes from a shakura - sunyaev disc ( e.g. cackett , horne & winkler 2007 ) .
independent size - scale estimates from microlensing also imply that the optical / uv emission region is larger than expected by a similar factor ( e.g. morgan et al .
yet the optical / uv spectrum shows a strong rise to the blue , and can be fairly well fit by the emission expected from the outer disc regions in a shakura - sunyaev model ( e.g. jin et al . 2012 ; capellupo et al .
2015 ) , though there are discrepancies in detail ( e.g. davis , woo & blaes 2007 ) . thus qualitatively the disc reprocessing picture appears sound , yet quantitatively it fails to match the data .
this is perhaps not surprising for several reasons .
irradiation can change the structure of the disc e.g. by flaring it , as well as by changing the local heating ( cunningham 1976 ) .
secondly , the spectra of agn are clearly not simply a disc .
the hard x - ray corona itself must be powered by accretion , pointing to a change from a pure shakura - sunyaev disc structure in the inner regions ( e.g. done et al .
there is also the generic additional component seen in agn , the soft x - ray excess , which again points to some change in disc structure which is not captured by the simple shakura - sunyaev equations ( e.g. gierlinski & done 2004 ; porquet et al .
2004 ) .
here we use the unprecedented swift and hst lightcurves collected by the 2014 campaign on ngc 5548 , which spans 120 days with sampling of 0.5 days , across 9 continuum bands from v to hard x - rays ( edelson et al .
we use these to quantitatively test disc reprocessing models using a full model of illumination and reprocessing , with the aim to reproduce both the spectrum and variability of the source .
we first examine traditional disc reprocessing models and confirm that these can not explain the lag timescales of the optical / uv lightcurves of ngc 5548 .
however , these quantitative models reveal another more fundamental conflict , which is that the uv and optical lightcurves can not be produced by reprocessing of the observed hard x - ray flux . disc reprocessing smears the variability by a similar timescale to the lag .
the uv and optical are lagged by 1 - 2 days , but are much smoother than the hard x - ray lightcurve , with no sign of the rapid 1 - 2 day variability seen in the x - rays ( see also arevalo et al .
the fast hard x - ray variability is not seen by the outer disc , so we conclude that the hard x - rays are not the driver for reprocessing . instead
, we demonstrate that illumination by the fuv ( represented by the hst lightcurve ) , rather than direct illumination by the hard x - rays , produces a much better match to the observed smoothness of the optical / uv lightcurves of ngc 5548 .
this points to a geometry where the outer disc is shielded from the hard x - rays , so we investigate models where the soft x - ray excess is produced from a moderately thickened disc .
the observed difference between the fuv and soft x - ray lightcurves clearly shows that this is not a single component , so we assume that smaller radii of the thickened disc produce the soft x - ray part of the soft x - ray excess , which is also shielded from the outer disc , while the outer regions produce the fuv which can illuminate the outer thin disc .
however , this still predicts lags which are shorter than observed . increased flaring of the outer disc does not help because these large radii regions with the required long lag times are too cool to contribute significant optical flux due to their large area .
we explore the suggestion that the longer than expected lags come from the contribution of the classic blr ( h@xmath0 line ) to the optical and uv emission ( korista & goad 2001 ) , but this does not work either . since all known components fail , we reverse engineer a geometry which can fit both spectral and variability constraints .
we use the observed optical lags in the different wavelength bands to constrain the flux contributions of reprocessors at different distances and temperatures and hence constrain the area of the reprocessors .
these areas are far too small to be the outer disc , and are more consistent with optically thick clumps lifted above the outer disc .
these clouds are located inside the smallest inferred blr radii ( heii ) , so we propose that these are the extension of the blr clouds to smaller radii / higher densities .
they are too dense to emit substantial line emission and they instead reprocess the illuminating fuv continuum as mostly black body emission .
fundamentally , the time lags give us the wrong answers because the reverberation signal is not from a thin disc illuminated by hard x - rays , but from small clouds on the very inner edge of the blr / outer edge of the disc , illuminated by the fuv emission .
thus we are not probing the size - scale of the thin disc , but the size - scale of the innermost , densest clouds of the blr .
this picture would similarly explain the microlensing results , as the emission from the small dense clumps weights the emissivity to larger size - scales .
for all models we fix @xmath1 ( as used by edelson et al .
2015 , from pancoast et al . 2014 : see also their erratum 2015 ) .
this is similar to the denny et al .
( 2010 ) estimate of @xmath2 , though a factor of @xmath3 smaller than the bentz et al .
( 2010 ) estimate of @xmath4 .
we also fix distance @xmath5mpc , and spin @xmath6 , and assume an inclination angle of @xmath7 .
the unabsorbed , dereddened broadband spectrum of ngc 5548 from summer 2013 is shown in mehdipour et al .
the x - ray flux is very hard , with photon index @xmath8 , and the x - rays dominate the energy output of the source , peaking in @xmath9 at @xmath10 ergs @xmath11 s@xmath12 at 100 kev .
the simultaneous optical / uv spectrum looks similar to that expected from an outer thin disc ( though is shape is subtly different : mehdipour et al .
the flux in uvw1 ( at around 5 ev ) has @xmath13 ergs @xmath11 s@xmath12 .
we use the optxagnf model in xspec to find reasonable physical parameters for the accretion flow .
this assumes that the mass accretion rate is constant with radius , with fully relativistic novikov - thorne emissivity per unit area , @xmath14 , with dimensionless radius @xmath15 for @xmath16 , but that this energy is dissipated in a ( colour temperature corrected ) disc only down to some radius @xmath17 , with the remainder split between powering a soft x - ray excess component and the hard x - ray coronal emission .
this code calculates the angle averaged spectrum , so we boost the normalisation by a factor @xmath18 to roughly account for our assumed inclination angle .
we find that we can match the uvw1 and x - ray flux for @xmath19 with @xmath20 .
we assume that all the power within @xmath20 goes to make the hard x - ray corona ( @xmath21 ) .
this does not necessarily mean that the optically thick , geometrically thin disk itself is not present , only that the accretion power is not dissipated within this structure ( svensson & zdziarski 1994 ; petrucci et al 2010 ) .
however , there is additional information in the hard x - ray spectrum which does point to this conclusion .
an optically thick disc can not be present underneath an isotropically emitting corona in this object , as such a disc will intercept around half of the hard x - ray flux , giving a strong compton hump which is not present in the nustar data ( see spectrum in mehdipour et al.2015 and ursini et al .
thermalisation of the non - reflected emission also produces too many seed photons for the x - ray source to remain hard ( haardt & maraschi 1991 ; 1993 ; stern et al .
1995 ; malzac et al . 2005 ; petrucci et al .
together , these imply that either the x - ray source is very anisotropic , or the disc truly truncates for sources with hard x - ray spectra .
black hole binaries similarly show hard x - ray spectra and small compton hump in their low / hard state , but here there are clear limits on the possible anisotropy from comparing sources with different binary inclination angles ( heil , uttley & klein - wolt 2015 ) .
this argues strongly for true truncation of the accretion disc , as does the currently popular lense - thirring precession model for the origin of the low frequency qpo s seen in high inclination binary systems ( ingram , done & fragile 2009 ) .
hence we assume the disc is truly truncated in these low luminosity agn ( see also petrucci et al .
2013 ; noda 2016 ) .
the irradiation pattern strongly depends on the relative geometry of the hard x - ray source and disc . in the black hole binaries
, there is evidence from the complex pattern of energy dependent lags that the hard x - ray source is somewhat radially extended ( kotov , churazov & gilfanov 2001 ; ingram & done 2012 ) , as is also expected if it is some form of hot , radiatively inefficient accretion flow such as an advection dominated accretion flow ( narayan & yi 1995 ) .
we assume that the extended hard x - ray source has volume emissivity @xmath22 and neglect light - bending and red / blueshifts to work out the irradiating flux as in the appendix of zycki et al ( 1999 ) , i.e. @xmath23 where @xmath24 is the distance from the hard x - ray source to the disc surface element , @xmath25 is the angle between the source and normal to the disc , and @xmath26 is the fraction of coronal hard x - ray luminosity , @xmath27 which thermalises .
we set @xmath28 in this section in order to see the maximum irradiation flux .
we then assume that @xmath29 and model the resulting optically thick disc emission at this radius as a ( colour temperature corrected ) blackbody .
however , the colour temperature correction makes very little difference for the black hole mass , mass accretion rate and @xmath17 used here , as the disc peaks at 5.5 ev ( @xmath30 ) so is too cool for the colour temperature correction to be significant .
[ cols= " < " , ] separating the variability of the source into its different components and constraining the contribution of these components to different bands , has allowed us to constrain the spectra of the reprocessors . assuming these reprocessors are black body emitters , allows us to constrain their temperature and covering fraction .
the temperature of the reprocessors increases as their radial distance decreases , as expected .
the covering fraction of the reprocessors is tiny just @xmath31 for the reprocessor at @xmath32ld .
this further underlines why these reprocessors can not be the disc , which has a huge area at these radii .
this covering fraction can be interpreted as a scale - height ( @xmath33 ) above the disc , assuming @xmath34 for material launched from the plane of the disc . in which case , the scale - heights for the reprocessors located between @xmath32 and @xmath35ld range from @xmath31 @xmath36 , i.e. @xmath37ld .
kollatschny & zetzl ( 2013 ) find that the blr of ngc 5548 has a wedge - shaped disc - wind - like structure , with wind height increasing with increasing radius .
they find that the inner edge of the blr , where the he line is emitted , is located at a radius of @xmath38ld with a height above the midplain of @xmath39ld .
these black body reprocessors are therefore consistent with small - area material located just inside the inner edge of the broad line emitting region .
potentially they are clumps originating in a dust - driven disc wind just like the broad line emitting clouds ( czerny et al .
2015 ) , except that at these very small radii the clouds are too dense to emit broad lines so instead reprocess the illuminating flux as thermal emission .
fig.[fig9 ] summarises our proposed geometry for ngc 5548 .
we have built a full disc reprocessing model in an attempt to simulate the simultaneous multi - wavelength lightcurve data presented by edelson et al .
( 2015 ) , assuming the uv / optical variability seen in ngc 5548 is due to reprocessing of higher energy radiation by the accretion disc .
this higher energy radiation is traditionally assumed to be the hard x - ray power law , produced at small accretion flow radii .
we find that reprocessing of the hard x - rays by a standard accretion disc can not replicate the observed uv / optical lightcurves or their lags with respect to the illuminating x - rays .
specifically the simulated lightcurves reproduce too much of the hard x - ray high frequency power and the light travel lag times are too short . one obvious answer to increase the amount of smoothing and increase the light travel lag times would appear to be to make the accretion disc larger .
the first reason this is not an option is a constraint from the spectral energy distribution of ngc 5548 . we have limited the size of our model accretion disc to the self - gravity radius .
we could relax this condition and allow the disc to extend to larger radii , but this would cause the disc spectrum to extend to lower energies .
by contrast , the observed spectrum of ngc 5548 , as shown by mehdipour et al .
( 2015 ) , clearly does not show such emission .
the peak uv / optical emission occurs in the uvw1 bandpass and the emission below this energy rapidly drops off . with an outer disc radius at the self - gravity radius , we already slightly overpredict the lowest energy fluxes . however , we modelled this directly , and find that the radii required to smooth the reprocessed hard x - ray lightcurve to a level matching the data is so large that the resulting lag times are too long to be compatible with the observations .
we are then driven to a scenario where the reprocessor can not directly see the hard x - rays .
this requires some source of material with sufficient scale - height that it can block the optical emitting regions view of the hard x - rays .
the obvious candidate for this is the soft excess component .
there is much debate over the physical origin of the soft excess emission .
we envisage a scenario where it is produced in an optically thick comptonising region at larger radii than the central hard x - rays .
uv / fuv emission produced in this region lifts material out of the plane of the accretion flow , where it is illuminated by the hard x - ray flux , which overionises the material .
the resultant loss of uv opacity means it falls back down , resulting in a region with scale - height sufficient to prevent the hard x - rays illuminating the outer parts of the accretion flow .
the soft x - ray emission could come from the inner radii of this region , while the lowest energy fuv emission could come from its largest radii .
one problem with complete shielding is that the observed fuv and hard x - ray lightcurves _ are _ significantly correlated , though the correlation is quite poor .
one possibility to incorporate some feedback between the two regions is if hard x - rays illuminate the inner edge of the soft excess region and these fluctuations then dissipate outwards through the soft excess until they reach the outer fuv emitting regions , which then illuminate the disc .
it is then dissipation of the hard x - ray fluctuations through the soft excess component that causes the loss of high frequency power , not light travel time smoothing .
however , this dissipation process has to be extremely fast , though the viscous timescale in the soft x - ray excess section of the disc will be faster than in a standard shakura - sunyaev disc .
we show that a model where the fuv emission ( represented by the hubble band lightcurve ) provides the illumination gives a much better match to the shape of the observed optical lightcurves .
however , our model response is still too fast at the longest wavelengths .
the observed v band lag in ngc 5548 is @xmath38d behind the fuv emission , which requires the reprocessed v band flux to be emitted at radii @xmath40 for this source .
however disc annuli at these large radii are too cool to contribute significant flux to the v band , which is instead dominated by hotter emission from smaller disc radii with shorter lag times .
the heating effect of the illuminating flux makes very little difference , as large disc annuli have enormous area .
the illuminating flux is simply spread over too large an area , so barely changes the temperature of the annulus and certainly can not heat the annulus enough for it to contribute significant v band flux . in order for the illuminating flux to heat a reprocessor at these large radii enough to contribute to the v band
requires the reprocessor to have a small area .
we use the uv / optical lightcurves and cross correlation functions to constrain the amount of reprocessed flux with different lag times in the uvw1 and v bands and , assuming this reprocessed flux is black body emission , we then estimated the temperatures and covering factors of the black body reprocessors at different radii . for black body reprocessors located between @xmath41ld from the illuminating fuv
, we derive covering factors from @xmath42 , corresponding to a scale - height of @xmath43ld .
these tiny covering factors rule out the disc as the source of the reprocessed emission and are instead consistent with clumps of material just inside the blr .
we suggest they may be optically thick blr clouds at very small radii that are too dense to emit via line emission so instead reprocess the fuv flux as thermal blackbody emission .
we conclude that the uv / optical lightcurves of ngc 5548 are not consistent with reprocessing of the hard x - rays by an accretion disc , but can instead be explained by reprocessing of the fuv emission by optically thick broad line region clouds .
this provides an additional source of ( mostly ) continuum emission , though it may not completely thermalise ( h. netzer & m. mehdipour , private communication ) .
the model still requires the presence of a standard outer shakura - sunyaev accretion disc , but this is a mostly constant component .
the additional variable optical / uv emission component adds to this disc emission , and starts to dominate over the disc emission in the v band .
this could explain why the spectra of quasars are in general slightly redder than expected from standard disc models ( davis , woo & blaes 2007 ) , as well as explaining the microlensing size - scale results ( morgan et al .
2010 ) and perhaps also the additional red component required in decomposition of the spectral variability of other individual well studied objects such as 3c273 ( the @xmath44 component in fig.5 of soldi et al . 2008 ) . ultimately , whatever the origin of the lags in ngc 5548 , the datasets now available contain much more information than is encapsulated in a single lag time measurement .
we urge full spectral timing modelling of these data in order to extract all the new physical information on the structure and geometry of the accretion flow which is now within reach .
eg and cd acknowledges funding from the uk stfc grants st / i001573/1 and st / l00075x/1 .
cd acknowledges missagh mehdipour for multiple very helpful conversations and data , and hagai netzer for discussions about thermalisation . | the new multi - wavelength monitoring campaign on ngc 5548 shows clearly that the variability of the uv / optical lightcurves lags by progressively longer times at longer wavelengths , as expected from reprocessing on optically thick material , but that the timescales are longer than for a standard shakura - sunyaev accretion disc .
we build a full spectral - timing reprocessing model to simulate the uv / optical lightcurves of ngc 5548 .
we show that disc reprocessing of the hard x - ray lightcurve produces optical lightcurves with too much fast variability and the wrong lag time .
instead the data are consistent with reprocessing of the far uv ( fuv ) lightcurve .
this requires an intervening structure preventing the hard x - rays from illuminating the disc .
we propose this is the disc itself , due to uv line - pressure lifting the photosphere and increasing the scale - height .
this uv - emitting region is less dense , so less able to thermalise , making low temperature comptonised emission as required to produce the soft x - ray excess .
however , the reprocessing lags from the disc are still too short .
we reverse engineer a solution to match the observed lags , and infer the reprocessor geometry from the required luminosity and temperature of the lagged emission components .
we find the radii and covering factors are more consistent with optically thick clouds located just inside the broad line region ( blr ) .
we conclude the optical lags of ngc 5548 are not due to reprocessing of the hard x - rays by an accretion disc , but are instead due to reprocessing of the fuv by optically thick clouds on the inner edge of the blr .
= 1 = -0.5 cm [ firstpage ] black hole physics , accretion , x - rays : galaxies , galaxies : seyfert , galaxies : individual : ngc 5548 . |
the production of higher - dimensional black holes in collider @xcite is the important prediction arising from the large extra dimension scenario @xcite or the warped compactification ( rs1 ) scenario @xcite , where the fundamental ( higher - dimensional ) scale of gravitation can be set to the order of tev .
such phenomena would strongly suggest the presence of the extra dimensions , and we would also obtain the experimental method to seek for the quantum gravity .
hence , it is important to investigate the fundamental properties of higher - dimensional black holes such as stability , which is the subject of the present paper .
there is a variety of the higher - dimensional black objects with nontrivial topology unlike those in four - dimensional space - times where each component of stationary black hole is a topological two - sphere .
in fact , higher - dimensional vacuum einstein equation admits simple solutions called black branes which are direct product of the four - dimensional schwarzschild black hole and the flat space .
we are usually interested in the case where the flat dimensions in the black brane space - time are compactified into say a torus .
gregory and laflamme @xcite have found that the black branes are unstable under linear gravitational perturbations when the compactification scales are large compared with the schwarzschild radius . on the other hand , ishibashi and kodama @xcite
have shown that the higher - dimensional schwarzschild black holes in the spherically symmetric space - times are gravitationally stable as in the case with the four space - time dimensions .
an accurate way to compute the quasinormal frequencies of higher - dimensional schwarzschild black holes has been obtained via 6th order wkb approximation @xcite .
it is also important to consider the stability of the rotating black holes , since black holes produced in the collider generically have angular momenta and the production cross section is larger for more rapidly rotating black holes @xcite in contrast to the kerr - bound on four - dimensional rotating black holes , there are black hole solutions with an arbitrary large angular momentum for a fixed mass in higher - dimensional general relativity .
recently , the existence of the effective kerr - bound on such rapidly rotating black holes in higher dimension has been conjectured by emparan and myers @xcite .
they showed that the geometry of the event horizon of such rapidly rotating black holes in six or higher dimension behaves like the black membranes and argued that such rapidly rotating black holes therefore become unstable .
although we should analyze the gravitational perturbations of rotating black holes to confirm such instability , any formalism where the perturbative equations are separable is yet unknown .
here we just consider the perturbation of the massless scalar field propagating on the rotating black hole space - times .
even if we restrict ourselves to field perturbation on fixed background space - time , there exists an instability of massless field perturbation on rotating black branes and strings discovered by cardoso and lemos @xcite .
the equation of motion for the massless scalar field perturbation on the rotating black holes is known to be separable in any dimension @xcite , and authors examined the stability of this system in five - dimension as the simplest case in the previous work @xcite .
the present paper is the complementary work on the renewed interest in the instability ( or stability ) of the rapidly rotating black holes in six dimensions .
the @xmath0-dimensional myers - perry metric @xcite with only one nonzero angular momentum is given by @xmath1 where @xmath2 and @xmath3 denotes the standard metric of the unit @xmath4-sphere .
this metric describes a rotating black hole in asymptotically flat , vacuum space - time with the mass and angular momentum proportional to @xmath5 and @xmath6 , respectively .
hereafter , @xmath7 are assumed .
the event horizon is located at @xmath8 , such that @xmath9 , which is homeomorphic to @xmath10 . for @xmath11
, @xmath12 always has a positive root for arbitrary @xmath13 , namely regular black hole solutions exist with arbitrarily large angular momentum per unit mass . in the limit where @xmath14 , the coordinate radius of the horizon is approximated by @xmath15 .
hence , the horizon radius @xmath16 shrinks to zero when one take the limit where @xmath14 with fixed @xmath5 . to avoid a vanishing horizon radius , emparan and myers introduced the black membrane limit where @xmath14 and @xmath17 with fixed @xmath18 . in this limit
, the myers - perry metric behaves like a black membrane metric near the pole @xmath19 : @xmath20 where the new coordinate is defined by @xmath21 . in this paper
, we are concerned with the dimensionless kerr parameter @xmath22 .
the limit where @xmath23 with fixed @xmath16 is consistent to the membrane limit .
here we consider the perturbation of the massless scalar field on the background myers - perry metric ( [ eq : myersperry ] ) .
we put the scalar field configuration as @xmath24 where @xmath25 is the hyperspherical harmonics on the @xmath4-sphere with the eigenvalue @xmath26 , @xmath27 .
then we obtain the field equations with separated variables : @xmath28s & = & 0,\end{aligned}\ ] ] and @xmath29 ^ 2 \over \delta } -{j(j+n-1)a^2 \over r^2}-\lambda\right\}r & = & 0,\end{aligned}\ ] ] where @xmath30 .
we will solve the eigenvalue problem under the quasinormal boundary condition : @xmath31 where @xmath32 when @xmath33 , the eigenfunctions for the angular equation ( [ eq : angular ] ) are analytically given in terms of the hypergeometric functions : @xmath34 with the eigenvalues @xmath35 then @xmath36 behaves as the hyperspherical harmonics on @xmath10 , which belongs to the irreducible representation of the rotational group @xmath37 .
especially for @xmath38 , the hyperspherical function on @xmath39 with the eigenvalue @xmath40 belongs to the representation characterized by the dynkin index @xmath41 $ ] and has @xmath42-fold degeneracy .
when @xmath43 , @xmath44 symmetry of the space - time breaks down to @xmath45 and the degenerate mode partially split into the several modes with @xmath46-fold degeneracies , where @xmath47 and @xmath48 are constrained by @xmath49 , @xmath50 .
we use the continued fraction method to determine the resonant frequency @xmath51 and the separation constant @xmath52 .
we assume the following series expansion for @xmath53 : @xmath54 which automatically satisfies the regular boundary conditions at @xmath55 and converges at any time . substituting this expansion into the angular equation ( [ eq : angular ] )
, we obtain the three - term recurrence relations : @xmath56 where @xmath57 and the coefficients are given by @xmath58 , \\
\bar{\gamma}_k & = & -a_*^2\omega_*^2.\end{aligned}\ ] ] here @xmath22 and @xmath59 are the dimensionless quantities .
the limit where @xmath60 corresponds to the membrane limit as mentioned above .
we expand the radial function @xmath61 as @xmath62 which automatically satisfies the quasinormal boundary conditions and converges at any time .
when @xmath38 , substituting eq .
( [ eq : radexp ] ) into the radial equation ( [ eq : radial ] ) , we obtain the eight - term recurrence relations @xmath63 where @xmath64 and the coefficients are given by @xmath65 \ } \nonumber\\ & & + i ( a_*^2 + 3 ) \ { [ ( 14 k-9 ) a_*^4+(40 k-22 ) a_*^2 + 42 k-21 ] \sigma + 2 m a _ * ( a_*^2 + 3 ) ( 2 k-1 ) \ } , \\ \gamma_{k } & = & ( 27 a_*^6 + 117 a_*^4 + 89 a_*^2 - 129 ) \sigma^2 + 4 m a _ * ( a_*^2 + 3 ) ( 5 a_*^2 + 1 ) \sigma + 4 m^2 a_*^2 ( a_*^2 + 5 ) \nonumber\\ & & -(a_*^2 + 1 ) [ ( 9 k^2 - 26 k+24 ) a_*^4+(14 k^2 - 42 k+40 ) a_*^2 - 3 k^2 - 24 + 4 j ( j+1 ) a_*^2 ( 2 a_*^2 + 3 ) -12 \lambda ] \nonumber\\ & & -i\ { [ ( 30 k-42 ) a_*^6+(106 k-152 ) a_*^4+(58 k-114 ) a_*^2 - 114 k+108 ] \sigma + 4 m a _ * ( a_*^2 + 3 ) [ ( 3 k-4 ) a_*^2 - 3 ( k-1 ) ] \ } , \nonumber\\ & & \\ \delta_{k } & = & -(3 a_*^2 - 1 ) ( 5 a_*^4 + 6 a_*^2 - 15 ) \sigma^2 -8 m a _ * ( 2 a_*^2 + 3 ) ( a_*^2 - 3 ) \sigma -4 m^2 a_*^2 ( a_*^2 - 7 ) \nonumber\\ & & + ( a_*^2 + 1 ) [ ( 5 k^2 - 25 k+36 ) a_*^4 - 2 ( k-1 ) ( 3 k-2 ) a_*^2 - 3 k^2 + 3 k+24 + 4 j ( j+1 ) a_*^2 ( 3 a_*^2 + 2 ) -4 ( a_*^2 - 2 ) \lambda ] \nonumber\\ & & + i\ { [ ( 18 k-45 ) a_*^6+(6 k-55 ) a_*^4+(-42 k+25 ) a_*^2 + 66 k-141 ] \sigma + 4 m a _ * [ ( 2 k-5 ) a_*^4+(-4 k+5 ) a_*^2 + 6 k-12 ] \ } , \nonumber\\ & & \\ \epsilon_{k } & = & -2 ( 9 a_*^2 + 25 ) ( a_*^2 + 1)^2 \sigma^2 -8 m a _ * ( 2 a_*^4 + 3 a_*^2 - 3 ) \sigma -4 m^2 a_*^2 ( a_*^2 - 3 ) \nonumber\\ & & + ( a_*^2 + 1 ) [ ( 5 k^2 - 20 k+16 ) a_*^4+(14 k^2 - 60 k+64 ) a_*^2 + 5 k^2 - 24 k+40 -4 j ( j+1 ) a_*^2 ( 2 a_*^2 + 1 ) + 4 \lambda ] \nonumber\\ & & + i\ { [ ( 18 k-36 ) a_*^6+(78 k-164 ) a_*^4+(110 k-252 ) a_*^2 + 82 k-204 ] \sigma + 4 m a _ * [ ( 2 k-4 ) a_*^4+(4 k-9 ) a_*^2 + 6 k-15 ] \ } , \nonumber\\ & & \\ \zeta_{k } & = & ( a_*^2 + 1 ) \ { 4 ( 6 a_*^4 + 15 a_*^2 + 5 ) \sigma^2 + 20 m a _ * ( a_*^2 + 1 ) \sigma + 4 m^2 a_*^2 \nonumber\\ & & -(k-3 ) ( 9 k-28 ) a_*^4-(10 k^2 - 60 k+86 ) a_*^2-(k-2 ) ( 5 k-19 ) + 2 ( a_*^2 + 1 ) [ j ( j+1 ) a_*^2 + \lambda ] \ } \nonumber\\ & & -i ( a_*^2 + 1 ) \ { [ ( 30 k-93 ) a_*^4+(52 k-160 ) a_*^2 - 2 k+13 ] \sigma + 2 m a _ * [ ( 6 k-19 ) a_*^2 - 2 k+7 ] \ } , \\ \eta_{k } & = & ( a_*^2 + 1)^2 \ { -(9 a_*^2 + 13 ) \sigma^2 -4 m a _ * \sigma + ( k-4 ) [ ( 5 k-18 ) a_*^2+k-4 ] \ } \nonumber\\ & & + 2 i ( a_*^2 + 1)^2 \ { [ ( 7 k-27 ) a_*^2 + 7 k-28 ] \sigma + 2 m a _ * ( k-4 ) \ } , \\ \theta_{k } & = & ( a_*^2 + 1)^3 [ \sigma^2 - ( k-4 ) ( k-5 ) -i ( 2 k-9 ) \sigma].\end{aligned}\ ] ] after we repeatedly use the gaussian elimination as written in the appendix , the eight - term recurrence relations ( [ eq : eight - term ] ) are reduced to the three - term ones @xmath66 for given the three - term recurrence relations ( [ eq : angrec ] ) ,
the successive coefficients are obtained in two ways , the finite and infinite continued - fraction representations : @xmath67 where @xmath68 is an abbreviation for @xmath69 $ ] .
similarly for eq .
( [ eq : radrec ] ) , we obtain @xmath70 for fixed @xmath71 and @xmath72 , we obtain nonlinear algebraic equations ( [ eq : radcontfrac ] ) and ( [ eq : angcontfrac ] ) for two unknown complex numbers @xmath73 and @xmath52 .
we use a standard nonlinear root search algorithm provided by the minpack subroutine package to evaluate the numerical solutions of these equations .
the equations for different sets of @xmath74 are used as numerical check .
when the perturbation become unstable under the membrane limit , we expect that the behavior of such instability resembles one of the gregory - laflamme instability .
since they found that the @xmath75 wave is unstable , we consider the @xmath76 modes .
the critical value of the wavelength of the gregory - laflamme instability is of the order of @xmath16 . in the coordinate @xmath77 along the rotation plane
, the perturbation has the wave length @xmath78 for @xmath79-th mode .
hence , we expect that the possible instabilities occur for @xmath80 .
typical examples of quasinormal modes of the six - dimensional myers - perry black holes are plotted in figs .
[ fig : l0] [ fig : l5 ] . here
the real and imaginary parts of the resonant frequencies @xmath73 as functions of the parameter @xmath81 for various value of @xmath48 and @xmath79 are plotted .
although the kerr parameter @xmath82 is taken to have a sufficiently large value , each mode does not seem to have negative imaginary part in any case .
we investigated the massless scalar field equation in the six - dimensional myers - perry black hole background . under the quasinormal boundary condition ,
we have searched for the resonant modes . in the non - rotating case ( @xmath83 ) ,
the quasinormal modes with same @xmath84 are degenerate .
continuously varying the rotational parameter @xmath82 , these modes split into several modes characterized by @xmath85 .
the sequences of quasinormal modes we obtained have positive imaginary parts , thus they do not show any evidence for instability . recently , cardoso , siopsis and yoshida @xcite investigated the stability of the same system .
several quasinormal modes within our present work are also calculated by them in different way to expand the radial function and their results are consistent to our present results .
they also conclude that the stability of scalar field perturbations on myers - perry black holes are suggested .
the authors would like to thank g. w. kang , and h. kodama for useful discussions and comments .
y. m. is supported by a grant - in - aid for the 21st century coe `` center for diversity and universality in physics '' .
the eight - term recurrence relations ( [ eq : eight - term ] ) are reduced to the seven - term ones @xmath86 via the gaussian elimination : @xmath87 the similar procedures @xmath88
@xmath89 @xmath90 @xmath91 leads to the three - term relations @xmath92 | we have studied the scalar field perturbations on six - dimensional ultra - spinning black holes .
we have numerically calculated the quasinormal modes of rotating black holes .
our results suggest that such perturbations are stable . |
a @xmath0-dimensional symplectic ellipsoid is a region in standard euclidean space @xmath7 described by an inequality of the form @xmath8 , where @xmath9 is a positive definite quadratic form and @xmath10 .
since @xmath9 may be diagonalized by a linear change of coordinates , every ellipsoid may be written ( uniquely ) as @xmath11 where @xmath12 we denote an open ellipsoid by @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}$ ] and the ball @xmath14 by @xmath15 .
further , @xmath16 denotes @xmath17 with the rescaled form @xmath18 .
thus @xmath19 . throughout this paper
the word embedding " will be used to denote a symplectic embedding .
if @xmath17 embeds in @xmath20 we shall write @xmath21 .
this paper is concerned with the question of when it is possible to embed one @xmath0-dimensional ellipsoid into another .
there are two known obstructions to embedding @xmath11 into @xmath22 when @xmath23 are integers ; namely , if such an embedding exists , we must have \(i ) ( volume ) : @xmath24 ( since @xmath11 has volume @xmath25 ) ; and \(ii ) ( ekeland hofer capacities ) : @xmath26 . here
@xmath27 denotes the sequence obtained by arranging the numbers @xmath28 and @xmath29 in nondecreasing order ( with repetitions ) and @xmath26 means that every number in @xmath27 is no larger than the corresponding number in @xmath30 ; see @xcite . in particular , we must have @xmath31 as follows from gromov s nonsqueezing theorem .
for example , @xmath32 so that @xmath33 . since the volume inequality is also satisfied in this case the question arose as to whether @xmath34 embeds in the ball @xmath35 for all @xmath36 .
( by corollary [ cor : conn ] this is equivalent to asking if the interior @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(1,4)$ ] embeds into @xmath35 . )
when considering this embedding problem it is convenient to consider the maximal packing radius @xmath37 and the packing constant @xmath38 which is the ratio of the volume of the domain to that of the target .
note that both these numbers are scale invariant , i.e. do not change if all numbers @xmath39 are multiplied by the same constant @xmath40 .
we say that @xmath11 _ fully fills _
@xmath22 if @xmath41 . further we denote by @xmath42 the complex projective plane with its standard fubini study form , normalized so that the area of a line is @xmath40 .
it is obtained from the ball @xmath43 by collapsing its boundary sphere to a line .
[ thm:1 ] @xmath2 embeds in the open ball @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(\mu)$ ] if and only if @xmath4 disjoint balls @xmath45 embed in @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(\mu)$ ] .
the only if `` part of this statement was first observed by traynor @xcite .
it is very easy to prove using toric models , which make it immediately clear that @xmath2 contains @xmath4 disjoint open balls @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(1)$ ] : see fig .
[ fig:5 ] and lemma [ le : open ] .
however the if '' part requires more work .
the main idea is to cut the ellipsoid from the ball ( i.e. perform an orbifold blow up as in godinho @xcite ) and then to resolve the resulting orbifold singularities by further standard blow ups .
the necessary symplectic surgery techniques were developed by symington @xcite in her treatment of rational blowdowns .
[ cor:1 ] @xmath2 fully fills @xmath46 if and only if @xmath5 or @xmath6 .
moreover for @xmath47 the packing constant @xmath48 is : @xmath49 _ proof of corollary . _ in his foundational paper @xcite , gromov showed that @xmath50 and @xmath51 .
when @xmath52 or @xmath53 , the rest of the above statement follows from theorem [ thm:1 ] by mcduff polterovich @xcite .
the case @xmath54 is due to biran @xcite .
@xmath55 for explicit realizations of the ball packings at the integers @xmath56 , see karshon @xcite , traynor @xcite , schlenk @xcite , and wieck @xcite .
after this paper was written , i found out that opshtein s paper @xcite on maximal symplectic packings of @xmath1 contains a proof that @xmath2 fully fills @xmath1 when @xmath53 .
though not stated explicitly in his paper , this result follows immediately from his lemma 2.1 .
his argument has the virtue of providing a geometric recipe for constructing these packings . in (
* problem 15 ) cieliebak , hofer , latschev , and schlenk ask whether the two invariants listed above are the only obstructions to embedding one open ellipsoid into another . more formally , they asked if the ekeland hofer capacities @xmath27 together with the volume capacity @xmath57 generate the ( generalized ) symplectic capacities on the space @xmath58 of open @xmath0-dimensional ellipsoids . to understand what this means , consider the capacity given by embeddings into an open ball : @xmath59[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(\mu)\bigr\}.\ ] ] if this capacity were some combination of @xmath27 and @xmath57 for @xmath60[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m , n)$ ] then every time the inequalities given by these capacities are satisfied there would be an embedding of @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m , n)$ ] into @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(\mu)$ ] .
but this is not so .
for example , @xmath61[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(1,5 ) ) = \frac 52 $ ] because by corollary [ cor:1 ] the volume of the target ball must be at least @xmath62 times the volume of @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(1,5)$ ] . on the other hand ,
the volume and ekeland hofer capacities give no obstruction to the existence of an embedding of @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(1,5)$ ] into @xmath63[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(\sqrt 5)$ ] for all @xmath64 .
we conclude : [ cor:2 ] the ekeland
hofer capacities @xmath27 together with the volume capacity do not generate the ( generalized ) symplectic capacities on the space @xmath58 of open @xmath0-dimensional ellipsoids .
note that this follows from the easy ( only if ) part of theorem [ thm:1 ] .
our second set of results concern the problem of embedding one ellipsoid into another .
it is convenient to introduce the following terminology . given a positive integer @xmath4 and @xmath4 positive numbers @xmath65 the _ ( symplectic ) packing problem for @xmath4 balls with weights _
@xmath66 is the question of whether the @xmath4 disjoint ( closed ) balls @xmath67 embed into the open ball @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(1)$ ] .
[ thm : ell ] for any positive integers @xmath39 and any @xmath68 , there is an integer @xmath4 and weights @xmath69 such that the question of whether @xmath70 embeds into the open ellipsoid @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] is equivalent to the symplectic packing problem for @xmath4 balls with weights @xmath69 .
the following result ( which was proved for balls in @xcite ) is an easy consequence .
[ cor : conn ] let @xmath71 be any real positive numbers .
then : \(i ) the space of symplectic embeddings of @xmath72[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(a',b')$ ] is path connected whenever it is nonempty ; \(ii ) if @xmath73 embeds in @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(a',b')$ ] for all @xmath36 , then @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(a , b)$ ] also embeds in @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(a',b')$ ] .
we also work out two specific examples that answer a question raised by tolman @xcite .
[ prop : tol](i ) the question of whether @xmath74 embeds in @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(2,3)$ ] is equivalent to the packing problem with @xmath75 and @xmath76 .
hence the embedding exists iff @xmath77 .
\(ii ) the question of whether @xmath78 embeds in @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(2,3)$ ] is equivalent to the packing problem with @xmath79 and @xmath80 .
hence the embedding exists iff @xmath81 .
these packing problems arose in tolman s attempt to describe all @xmath82-dimensional hamiltonian @xmath83-manifolds @xmath84 with @xmath85 of rank @xmath86 .
in @xcite we use the ideas of the present paper to construct the two new manifolds among her list of four possibilities , thus completing her classification .
\(i ) the question of which weights @xmath87 correspond to a given packing problem is intimately related to standard ( rather than hirzebruch
jung ) continued fraction expansions : see remarks [ rmk : cf ] and [ rmk : sing ] .
\(ii ) our approach also gives a great deal of information about the function @xmath88 studied in schlenk @xcite and in @xcite .
this will be the subject of mcduff schlenk @xcite . * the nature of the ball packing problem .
* we now explain the results of @xcite that first convert the symplectic packing problem for balls into a question of understanding the symplectic cone of the @xmath4-fold blow up @xmath89 of @xmath1 , and then explain the structure of that cone .
the _ symplectic cone _ of an oriented manifold is the set of cohomology classes of @xmath84 with symplectic representatives compatible with the given orientation . by li
liu @xcite , when @xmath90 this is a disjoint union of ( connected ) subcones , each consisting of forms @xmath91 with a given canonical class @xmath92
. moreover , the group of diffeomorphisms of @xmath89 acts transitively on these subcones .
fix @xmath93 , where @xmath94 $ ] and @xmath95 are the exceptional divisors , and define @xmath96 to be the set of classes represented by symplectic forms with canonical class @xmath97 .
also , denote the poincar duals of @xmath98 by @xmath99 .
[ mcduff polterovich @xcite][prop : mp ] it is possible to embed @xmath4 balls with weights @xmath87 into @xmath100 or @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}^4(1)$ ] if and only if the class @xmath101 lies in @xmath102 .
therefore the problem is equivalent to understanding @xmath102 .
though not formulated in these terms , lemma 1.1 of @xcite in essence describes the closure of @xmath102 .
( the argument is sketched below . ) that paper solved the uniqueness question for symplectic packings . in @xcite , biran recast this lemma in terms of the symplectic cone , using it to solve the existence question for packings by equal balls .
however the cone itself , rather than its closure , was first described in li
liu @xcite .
they consider the set @xmath103 of classes @xmath17 with @xmath104 that can be represented by smoothly embedded @xmath105 spheres . by theorems b and c in li
li @xcite , given any symplectic form @xmath91 with canonical class @xmath97 , each @xmath106 can be represented by an @xmath91-symplectically embedded @xmath105 sphere .
thus the above set @xmath107 is the same as that used in @xcite .
[ li - liu , @xcite ] [ prop : m ] @xmath108 _ sketch of proof .
let @xmath91 be any symplectic form on @xmath89 with canonical class @xmath97 .
it follows from the work of kronheimer and mrowka @xcite on wall crossing in seiberg
witten theory that for all @xmath109 with @xmath110 the class @xmath111 has nontrivial seiberg
witten invariant for all sufficiently large integers @xmath112 .
therefore , by work of taubes @xcite , the poincar dual @xmath113 has a @xmath114-holomorphic representative for every @xmath91-tame @xmath114 .
moreover , if @xmath115 for all @xmath116 and @xmath114 is generic this representative is a connected and embedded submanifold . by inflating along this submanifold one can construct a family of symplectic forms @xmath117 with @xmath118 and such that @xmath119 $ ] converges to @xmath120 as @xmath121
therefore , the class @xmath120 is in the closure of @xmath102 ; see ( * ? ? ?
* lemma 1.1 ) .
a more careful version of this argument shows that @xmath120 must in fact be in @xmath102 provided that @xmath122 for all @xmath17 ; see @xcite . @xmath55 the arguments in 2 below
show explicitly how to use these ideas to construct full packings by the ellipsoids @xmath2 for @xmath123 and hence by equal balls . the general existence question is fully understood when @xmath56 since in that case @xmath124 is finite and is easily enumerated .
however , it is not always so easy to answer specific problems when @xmath6 since @xmath124 is more complicated .
we shall return to this question in@xcite .
the above results imply there are two obstructions to the existence of an embedding from one ellipsoid to another .
if @xmath125 is the class in @xmath89 of the corresponding packing problem , one needs : \(i ) @xmath126 , and \(ii ) @xmath127 for all @xmath106 .
the first condition is equivalent to the volume obstruction , while the second is a generalization of the condition used by gromov @xcite to find a packing obstruction when @xmath128 .
it is given by the rigid @xmath114-holomorphic spheres in @xmath89 , and hence should appear as a genus zero holomorphic trajectory in other contexts .
however , note that although the classes @xmath129 are rigid in the blow up , they correspond in @xmath1 to curves that satisfy some constraints , e.g. they might have to go through a certain number of points with given multiplicities .
moreover , in general these constraints may not be purely homological , but may involve descendents : cf . the blow down formulas in @xcite .
thus , for example , if one tried to understand the obstructions to embedding @xmath11 into @xmath22 by looking at the properties of the induced cobordism between their boundaries then these curves should appear in one of the higher dimensional ( but genus zero ) sft moduli spaces and hence should be visible , provided that one works in a context that takes these higher dimensional spaces into account .
* organization of the paper . *
2 considers the problem of embedding ellipsoids of the form @xmath2 into balls , where @xmath130 .
the case @xmath53 is particularly simple and is treated first .
theorem [ thm:1 ] is proved in 2.2 . in order to deal with general integral ellipsoids
one must understand exactly how to approximate them by chains of spheres .
this is the subject of 3.1 .
theorem [ thm : ell ] , corollary [ cor : conn ] and proposition [ prop : tol ] are proved in 3.2 .
* acknowedgements .
* i was inspired to think about this problem by discussions with hofer and guth , who in @xcite recently solved ( in the negative ) hofer s question about the existence of higher dimensional capacities .
i also thank tolman for giving me an advance copy of her paper @xcite , schlenk for his encouragement and useful comments , and the referree for pointing out various small inaccuracies .
we first give a direct geometric construction for embedding @xmath131 into a ball . we then prove theorem [ thm:1 ] .
this section proves the following result .
[ prop : d ] @xmath131 fully fills @xmath132 .
denote by @xmath133 the triangle with outward conormals @xmath134 and vertices at @xmath135 ; see figure [ fig:1 ] .
if @xmath136 are integers then the open ellipsoid @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m , n)\subset { { \mathbb c}}^2 $ ] is invariant under the obvious @xmath137 action and is taken by the moment map onto the interior " @xmath138 of @xmath133 , which we define to be the complement of its slanted edge .
the closed triangle @xmath133 is the moment polytope of a weighted projective space , but we sometimes think of it as the moment polytope of @xmath11 since it is the image of @xmath11 under the moment map . recall that if @xmath139 are the outward conormals to two successive edges ( ordered anticlockwise ) of a moment polytope in @xmath140 then their intersection is the image of an orbifold point of order @xmath4 iff @xmath141 in particular , it is smooth iff @xmath142 . we shall call a moment polytope _ smooth _ if all its vertices are smooth .
( for basic information on toric geometry in the present context see symington @xcite and traynor @xcite . )
; it does not include the slanted edge.,width=192 ] just as in the case of balls , if @xmath143 embeds in @xmath20 then one can cut it out and collapse the boundary along the characteristic flow to obtain an orbifold @xmath144 with an exceptional divisor @xmath145 ( the heavy line in figure [ fig:2 ] ) . in general ,
the two new vertices are singular points of @xmath144 . however , in the case of @xmath146 there is just one singular point on @xmath145 at @xmath147 . for more details ,
see @xcite .
the main idea of the current note is that instead of working with the orbifold @xmath144 we can cut away more ( i.e. blow up further ) in order to get a smooth manifold .
the case @xmath148 is particularly simple ; we simply need to blow up @xmath149 more times . in the toric picture
this amounts to cutting the polytope along lines with conormals @xmath150 .
the curve @xmath145 is then transformed into an exceptional curve in a smooth manifold . as we see below ,
it is possible to carry out this process omitting all mention of orbifolds .
we explain it first in the case @xmath151 . in the case
the four edges are named @xmath153 and their homology classes @xmath154.,width=384 ] when @xmath151 , this blow up process gives a configuration of @xmath0 spheres intersecting transversally , the @xmath105-sphere @xmath155 in class @xmath156 , and three @xmath157-spheres @xmath158 each in class @xmath159 : see figure [ fig:3 ] .
if we start off in a large ball or in a large @xmath1 , we can construct such a configuration .
namely , start with the first quadrant ; form @xmath160 by a cut with conormal @xmath161 and size @xmath162 .
( the size of the cut is given by the affine length of an edge @xmath163 of a moment polytope can be measured as follows .
take any affine transformation @xmath164 of @xmath140 that preserves the integer lattice and is such that @xmath165 lies along one of the axes , and then measure the euclidean length of @xmath165 . thus if @xmath163 has rational slope and endpoints on the integer lattice , @xmath166 where @xmath4 is the number of points of the integer lattice in the interior of @xmath163 .
note also that the divisor that is taken to @xmath163 by the moment map has symplectic area @xmath167 .
] of the resulting exceptional divisor , and hence in this case is given by @xmath168 , the height of the point where the cut meets the @xmath169 axis . ) then cut off most of @xmath160 by a cut with conormal @xmath170 and size @xmath171 .
make one more cut along @xmath172 of size @xmath173 .
then @xmath156 is what is left of @xmath145 .
it is easy to check that it has size @xmath174 , where @xmath175 .
let us denote by @xmath176 a small open neighborhood in the toric model of the final configuration @xmath177 of @xmath0 spheres .
thus @xmath177 is a configuration of @xmath0 symplectic spheres @xmath178 where @xmath179 intersects @xmath180 transversally for @xmath181 ( and there are no other intersection points ) , where @xmath155 has self - intersection @xmath105 and size @xmath182 and where the other @xmath179 have self - intersection @xmath157 and ( positive ) sizes @xmath183 and @xmath184 . in the following , we assume that @xmath185 and that @xmath186 .
[ le : nb ] any embedding of @xmath177 into @xmath20 is isotopic to one that extends to an embedding of @xmath176 . note that in any smooth toric manifold the spheres represented by @xmath187 edges meet orthogonally since we can put them on the axes by an affine transformation
. this may not be the case for the given embedding of @xmath177 .
however , one can slightly perturb this embedding so that the different spheres do meet orthogonally .
( a similar point occurs in the proof of ( * ? ? ?
* thm 9.4.7 ) .
see also ( * ? ? ?
9.4.8 ) . )
it then follows from the symplectic neighborhood theorem that the embedding extends to @xmath176 .
denote by @xmath188 the @xmath0-point blow up of the projective plane in which the line has symplectic area @xmath40 and we have blown up @xmath0 times by the amounts @xmath189 and @xmath190 .
if @xmath36 we can choose the @xmath191 so that each @xmath192 . because @xmath1 can be fully filled by @xmath0 balls of equal size , we can therefore construct @xmath193 for any @xmath36 and sufficiently small @xmath194 as above .
[ le:1 ] @xmath195 embeds into the interior of @xmath35 iff @xmath177 embeds into the complement of a line in @xmath193 for some small @xmath196 .
if @xmath195 embeds then this embedding extends to @xmath197 for some small @xmath198 .
hence @xmath195 has a standard neighborhood with a toric structure as above .
complete @xmath35 to @xmath199 and then blow up as explained above to get an embedding of @xmath177 into the complement of a line in @xmath193 .
conversely , suppose that @xmath177 embeds into the complement of a line in @xmath193 .
then , by lemma [ le : nb ] we may suppose that some standard neighborhood @xmath176 also embeds . by symington s discussion in @xcite of toric models for the rational blow down , we may then blow down " @xmath177 , i.e. perform a symplectic surgery along this nonsmooth divisor that adds the singular piece that was cut out when resolving the singularity . the resulting blow down manifold @xmath20 contains a symplectically embedded copy of @xmath200 of size @xmath187 and a disjoint copy of @xmath195 .
moreover , it is diffeomorphic to @xmath1
. therefore , by gromov s uniqueness result for symplectic forms on @xmath1 ( cf ( * ? ? ?
9 ) ) , @xmath20 can be identified with @xmath199 , so that @xmath201 is the interior of the standard ball @xmath35 .
this completes the proof .
[ rmk : rmk1]there is an obvious analog of this result for any @xmath4 . an appropriate definition of @xmath202 is explained in the proof of theorem 1.1 at the end of this section .
therefore we just need to embed @xmath177 into the complement of a line in @xmath193 .
this is possible for small @xmath203 .
we then will use the inflation process to increase @xmath203 . in this case , the construction can be done entirely explicitly : there is no need to use seiberg witten theory .
* proof of proposition [ prop : d ] .
* * step 1 : * _ explicit embedding of @xmath177 . _
start with @xmath199 .
let @xmath204 be a smooth conic , @xmath205 a line and @xmath206 a point on @xmath9 but not @xmath205 .
blow up at @xmath206 with size @xmath162 where @xmath203 is small and @xmath207 is tiny .
let @xmath208 be the intersection of the exceptional divisor @xmath160 with the proper transform @xmath209 of @xmath9 and blow up at @xmath208 with size @xmath210 to get an exceptional divisor @xmath211 .
now repeat this twice more , blowing up at @xmath212 by @xmath213 to get exceptional divisor @xmath172 and finally blowing up at @xmath214 by @xmath182 ( where @xmath215 ) to get @xmath155 in the @xmath0-fold blow up @xmath216 . for
@xmath217 denote by @xmath179 the proper transform of @xmath218 in the next blow up . then @xmath219 = e_i - e_{i+1}$ ] for @xmath220 and @xmath221 = e_4 $ ] .
thus we have constructed a copy of the configuration @xmath177 in @xmath193 .
denote the symplectic form on @xmath222 by @xmath223 .
note that if @xmath203 is sufficiently small we may assume that none of these blowups affect @xmath205 .
the conic @xmath9 becomes a curve @xmath224 in class @xmath225 and so has area @xmath226 . by construction @xmath224
meets @xmath155 once but not @xmath227 .
moreover , @xmath228 * step 2 : * _ the inflation process .
_ we will inflate @xmath229 along @xmath224 .
note that @xmath230 .
therefore we may identify a neighborhood @xmath231 with the product @xmath232 where @xmath233 is some area form on @xmath234 . by perturbing @xmath155 and the line @xmath205 and then shrinking @xmath231 if necessary
, we may assume that @xmath235 for @xmath220 ; @xmath236 is a flat disc @xmath237 ; @xmath238 is the union of two disjoint flat discs @xmath237 .
let @xmath239 be a nonnegative form on the two disc @xmath234 with support in its interior and @xmath240 .
define @xmath241 in @xmath231 and equal to the original symplectic form @xmath223 outside @xmath231 .
this is clearly symplectic everywhere .
then the integral of @xmath117 over @xmath224 and the @xmath242 , is constant while @xmath243 given any @xmath244 , choose @xmath245 so that @xmath246 and set @xmath247 .
define @xmath248 . by the uniqueness of symplectic forms on the blow ups of @xmath1 ( cf .
@xcite ) we may identify @xmath249 with @xmath250 , and may also identify the configuration with @xmath251 .
since the configuration is disjoint from a line by construction , we can now deduce the proposition from lemma [ le:1 ] .
* step 3 : * _ completion of the proof .
_ this argument extends immediately to the case @xmath252 : simply replace @xmath9 in the above construction by a degree @xmath253 curve in @xmath254 , and blow up @xmath255 times .
because the above argument uses symplectic inflation , it does not give a completely explicit geometric construction of the embedding .
it is not clear whether such a construction exists .
( the construction in opshtein @xcite gives an embedding into @xmath1 . ) the best results obtained so far by explicit construction are those of schlenk , who used a multiple symplectic folding technique to show that @xmath152 embeds in @xmath43 for @xmath40 approximately equal to @xmath256 , see ( * ? ? ?
3.3 ) .
we now prove theorem [ thm:1 ] .
embed in @xmath257.,width=240 ] [ le : open ] @xmath2 contains the disjoint union of @xmath4 open balls @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(1)$ ] . denote by @xmath258 the open triangle with vertices @xmath259 and @xmath260 . as is clear from figure [ fig:5 ] , there are integral affine transformations @xmath261 of @xmath140 such that the @xmath4 triangles @xmath262 embed in the moment polytope of @xmath2 . on the other hand @xmath263
is the moment polytope of the open subset @xmath264 which contains an embedded image of the open ball @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(1)$ ] by traynor @xcite .
* proof of theorem [ thm:1 ] . *
if @xmath2 embeds in @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(\mu)$ ] then @xmath265 also embeds for some @xmath198 .
but lemma [ le : open ] implies that @xmath265 contains @xmath4 disjoint closed balls @xmath45 .
hence so does @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(\mu)$ ] .
conversely , if @xmath4 copies of @xmath45 embed in @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(\mu)$ ] , choose @xmath266 so that their image is contained in @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(\mu_0)$ ] .
by proposition [ prop : mp ] the class @xmath267 has a symplectic representative with canonical class @xmath97 and so lies in @xmath102 .
hence , as in proposition [ prop : m ] , @xmath122 for all @xmath268 . without loss of generality
, we may suppose that @xmath269 and then choose @xmath112 so large that @xmath270 and @xmath113 has nontrivial gromov invariant .
now let @xmath202 be a configuration of @xmath4 symplectic spheres @xmath271 such that \(a ) @xmath272 iff @xmath273 , \(b ) @xmath274 , and @xmath275 , and \(c ) area@xmath276 if @xmath277 , area@xmath278 , and area@xmath279 , where @xmath280 and @xmath281 .
if @xmath203 and the @xmath191 are sufficiently small , the blow up construction in 2.1 shows that there is a symplectic form @xmath223 on @xmath89 with @xmath282 and such that the configuration @xmath202 embeds in the complement of a line @xmath205 in @xmath283 in such a way that @xmath179 lies in class @xmath284 for @xmath285 and @xmath286 = e_k$ ] .
( observe here that the class @xmath287 $ ] is determined by @xmath203 and the @xmath191 : it will not equal @xmath120 . )
moreover , @xmath202 has a toric neighborhood @xmath288 constructed just as in the case @xmath151 , and by lemma [ le : nb ] we may suppose that @xmath288 also embeds disjointly from @xmath205 . for simplicity , we shall identify @xmath288 with its image in @xmath289 .
denote by @xmath290 the set of @xmath91-tame @xmath114 for which @xmath202 and the line @xmath205 are holomorphic . because @xmath113 has nontrivial gromov invariant and @xmath291 for all @xmath129 , @xmath113
is represented by an embedded @xmath114-holomorphic curve @xmath9 for every @xmath91-tame @xmath114 that is sufficiently generic . since no smooth curve in class @xmath113 is represented entirely in @xmath292 , it follows from ( * ? ? ?
* ch 3 ) that we can take @xmath114 to be a generic element of @xmath290 .
then , by positivity of intersections , the fact that @xmath293 implies that @xmath294 .
further , we may perturb @xmath9 so that it intersects @xmath295 transversally @xmath112 times and @xmath205 transversally @xmath296 times .
now inflate along @xmath9 .
this construction gives a family @xmath297 of symplectic forms on @xmath89 lying in class @xmath287 + tqa$ ] that equal @xmath223 outside a small neighborhood of @xmath9 and restrict on @xmath295 ( resp .
@xmath205 ) to a symplectic form of area @xmath298 ( resp .
@xmath299 ) .
thus @xmath300 embeds in @xmath301 .
therefore , by lemma [ le:1 ] ( see also remark [ rmk : rmk1 ] ) , @xmath302 embeds in @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(1+qt\mu_0)$ ] for all @xmath303 .
hence @xmath304[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(1+qt\mu_0)}.\ ] ] since @xmath266 , @xmath305 for large @xmath306 .
hence the result .
we first show how to find the weights of the ball embedding problem that is equivalent to a given ellipsoidal embedding problem .
theorem [ thm : ell ] is proved in
[ ss : thm ] .
[ ss : approx ] we begin by discussing inner and outer approximations . throughout
, @xmath307 are mutually prime and @xmath308 . as in the case @xmath148 ,
it is possible to approximate the moment polytope @xmath133 of @xmath11 by a smooth polytope @xmath309 by blowing it up appropriately .
since every moment polytope is a blow up of @xmath310 ( up to scaling and integral affine transformation ) , one can equivalently start with @xmath311 and by a sequence of blow ups arrive at a polytope @xmath309 lying inside @xmath133 and with one conormal equal to @xmath307 .
one can then adjust the side lengths of this polytope to make its edge @xmath312 with conormal @xmath307 coincide with part of the corresponding edge of @xmath133 and also the region @xmath313 arbitrarily small ( in area ) .
we will call such @xmath309 an _ inner approximation _ to @xmath133 ; see figure [ fig:6](ii ) .
this is the relevant approximation when @xmath11 is the target of the embedding .
if @xmath11 is the source , then as in 2 one should look for _ outer approximations _
@xmath314 such that @xmath315 is small .
these are essentially the same as inner approximations to @xmath316 for @xmath317 : see remark [ rmk : cf ] ( i ) . for clarity ,
let us first concentrate on inner approximations .
there are many possible choices of approximation .
however , as we show below , there is a unique minimal sequence of blow ups of @xmath311 with exceptional divisors @xmath318 such that the conormal to the edge @xmath312 given by the last blow up is @xmath307 and so that none of the other new edges have self intersection @xmath105 .
, it would have to be disjoint from @xmath312 and hence could be blown down . ]
let us denote by @xmath319 the edge with conormal @xmath320 and by @xmath321 the edge created by the @xmath322th blow up .
further , denote by _ construction of the minimal blow up sequence for @xmath307 . _
the blow up process replaces the intersection of two adjacent edges with conormals @xmath336 by a new edge with conormal @xmath337 .
thus the three fractions @xmath338 are related by the identity @xmath339 hence they are adjacent terms in the _ farey sequence _ @xmath340 , @xmath341 .
( recall from hardy wright ( * ? ? ?
* ch.iii ) that @xmath340 is the finite sequence obtained by arranging the fractions @xmath338 , where @xmath342 are relatively prime and @xmath343 , in order of increasing magnitude . )
it is well known that @xmath340 can be constructed by starting with the fractions @xmath344 and @xmath345 and then repeatedly inserting the fractions @xmath346 with @xmath347 between any two neighbors @xmath348 .
since this precisely corresponds to the blow up procedure , it follows that one always can find a sequence of blow ups that starts from @xmath349 and @xmath350 and ends up with an edge @xmath312 with conormal @xmath307 .
moreover , to find a minimal sequence one should include @xmath346 in the sequence only if @xmath351 lies between the points @xmath352 and @xmath353 .
we denote the corresponding connected chain of adjacent edges by @xmath354 .
this chain , when ordered by increasing @xmath351 , starts with @xmath355 , includes an edge @xmath312 with conormal @xmath307 and ends with @xmath356 .
( note that these edges are numbered according to the order in which the blow ups are performed , not by adjacency . ) for any pair @xmath307 with @xmath357 we can construct a convex chain of edges @xmath354 of this form , such that all edges except for the last one @xmath312 are very short and so that @xmath312 is almost all of the slanted edge of @xmath133 . to do this , first place @xmath358 so that it meets the @xmath359 axis at @xmath360 for small @xmath361 and make the first cut @xmath362 so that it meets the @xmath169 axis at @xmath363 for small @xmath364 .
then perform all subsequent blow ups so that all the edges @xmath321 have positive length and so that @xmath312 coincides with part of the line @xmath365 : see figure [ fig:9 ] . note that if the conormal @xmath370 of @xmath321 has @xmath371 then the line of the corresponding cut must meet the @xmath169 axis at some point @xmath372 with @xmath373 , while if @xmath374 it meets the @xmath359-axis at @xmath375 where @xmath376 .
it follows that the edges for @xmath377 must be very short , while @xmath312 is almost all the slanted edge of @xmath133 .
this chain of edges @xmath354 lies in @xmath133 and bounds a smooth subpolytope @xmath378 of @xmath133 .
the above sequence of blow ups gives a way to construct the polytope @xmath378 from @xmath379 by performing a sequence of blow ups with conormals @xmath380 .
we shall denote by @xmath381 the region obtained after the @xmath322th blow up and by @xmath369 the new edge of @xmath382 formed by the @xmath322th blow up .
thus @xmath369 is an extension of @xmath321 .
notice that @xmath383 , where the triangle @xmath384 is equivalent under the action of @xmath385 to @xmath386 and where @xmath387 is the affine length of the cut edge @xmath388 .
it follows from this construction that the homology classes @xmath323 of the edges @xmath321 are given by @xmath389 as well as this minimal blow up sequence , we shall need an integral homology class @xmath390 with the property that @xmath391 ( this vector @xmath392 will give the weights of the corresponding ball embedding problem . ) for example , if @xmath393 then the conormals @xmath394 to the sequence of edges starting at @xmath350 and going to @xmath356 have slopes @xmath395 these edges are numbered @xmath396 according to the order of the blow ups .
they have classes @xmath397 as one can check by performing the relevant blow ups . hence , @xmath398 where @xmath399 note that @xmath400 .
this is a general fact , which is known in other contexts ; cf . remark [ rmk : cf ] .
however we include a proof here for completeness .
let us call the coefficient @xmath404 of @xmath218 in @xmath392 the _ label _ of the edge @xmath321 .
equation ( [ eq : hi ] ) implies that @xmath392 will satisfy ( [ eq : v ] ) if the labels @xmath405 are assigned as follows . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ set @xmath406 .
given @xmath407 , define @xmath404 to be the sum of the labels of the edges @xmath408 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ using induction , we shall show that for all @xmath307 such that @xmath413 then @xmath392 as defined above has the required properties .
moreover if @xmath414 is another pair with @xmath415 and if @xmath416 then @xmath417 the base case is @xmath418 . then @xmath419 and @xmath420 .
the required properties are easily verified .
suppose the result is known for @xmath421 and consider a pair @xmath422 with @xmath423 .
let @xmath424 be the neighbors of @xmath422 in the farey sequence @xmath425 , named so that @xmath426 .
then @xmath427 and also @xmath416 .
( note that this implies @xmath428 . ) to complete the inductive step , we will show that @xmath429 satisfies all the conditions required of @xmath430 , where the edge with conormal @xmath431 is called @xmath432 . to make sense of this formula , note that because @xmath426 , @xmath351 occurs as part of the farey sequence @xmath433 .
hence , because @xmath416 , @xmath351 and @xmath434 are adjacent in this farey sequence .
it follows that all the edges in @xmath435 occur in @xmath436 : indeed @xmath437 may be obtained from @xmath438 by repeatedly blowing up at the vertex of the edge @xmath312 closest " to @xmath434 .
( for example , if @xmath439 then one blows up at the vertex of @xmath312 closest to the @xmath359-axis . )
in particular the edge @xmath312 of @xmath438 , when considered as part of @xmath437 , is adjacent to @xmath440 .
therefore we may consider the classes @xmath218 that occur in @xmath441 to be a subset of those occurring in @xmath442 so that the above formula for @xmath443 makes sense .
moreover , if @xmath444 denotes the class in @xmath445 represented by @xmath321 then , for all @xmath446 , @xmath323 is the class obtained from @xmath444 by setting @xmath447 .
similarly , @xmath57 is the class obtained from @xmath448 by setting all @xmath449 to @xmath450 .
now observe that , because @xmath307 and @xmath414 are the neighbors of @xmath422 in @xmath425 , @xmath451 is obtained from @xmath437 by adding one extra edge @xmath432 between @xmath452 and @xmath440 .
hence , @xmath453 . further , if we denote the class of @xmath321 in @xmath454 by @xmath455 , @xmath456 it follows easily that if @xmath443 is defined by equation ( [ eq : v ] ) then the relations ( [ eq : v ] ) hold .
hence @xmath457 .
moreover by equation ( [ eq : vv ] ) @xmath458 it remains to check that @xmath459 and @xmath460 satisfy the analog of equation ( [ eq : vv ] ) .
this is left to the reader . to illustrate this ,
consider the case @xmath461 with farey neighbors @xmath462 .
the edges in @xmath463 have slopes @xmath464 and their classes are @xmath465 thus @xmath466 .
therefore , by ( [ eq : v7 ] ) , formula ( [ eq : v ] ) gives @xmath467 to @xmath133 for @xmath307 equal to ( i ) @xmath468 , ( ii ) @xmath469 and ( iii ) @xmath470 .
the points of the integral lattice are marked .
note that @xmath471 , @xmath472 and @xmath473 .
the sizes of the triangles in each decomposition are given by the coefficients of the @xmath218 in each @xmath392.,width=384 ] in the above discussion each edge @xmath321 was assumed to have positive length , although it was very short for @xmath377 .
now imagine performing these cuts so that these edges have zero length .
in other words , at each stage construct @xmath474 by cutting out a vertex of @xmath475 together with the whole of the shorter adjacent edge .
thus each cut @xmath476 has a vertex at @xmath477 or at @xmath478 and the end result is @xmath479 .
thus these cuts decompose the triangle @xmath480 into a union of triangles , each equivalent to a multiple @xmath481 of the standard triangle : see figure [ fig:8 ] .
we will think of this decomposition of @xmath482 as corresponding to a singular ( nonsmooth ) blow up of @xmath311 .
we now show that the multiplicities @xmath483 are precisely the weights @xmath404 .
since the area of the cut triangles is @xmath484 this shows that @xmath485 which gives a geometric explanation for the quadratic relation @xmath412 .
[ rmk : area]it turns out that this geometric blow up procedure is very closely related to the construction of an appropriate continued fraction .
this is easiest to see in the context of outer approximations : see remark [ rmk : cf ] .
the coefficient @xmath483 is just the affine length of the cut @xmath369 , i.e. it is the length of the corresponding edge of @xmath474 .
this edge represents the class @xmath489 in @xmath410 .
thus @xmath490 , the affine length of the slanted edge @xmath491 of @xmath133 .
one can now argue that @xmath492 for @xmath493 in turn .
the point is that the @xmath322th cut leaves an edge " @xmath321 in class @xmath323 of length @xmath450 .
but @xmath321 is the result of cutting @xmath369 by cuts of length @xmath494 .
therefore the result follows from formula ( [ eq : hi ] ) and the definition of the @xmath404 given in the proof of lemma [ le : ki ] . with this notation in hand
, we can now discuss inner approximations with more precision .
as explained abve , an inner approximation to @xmath133 is obtained by moving the edge of @xmath311 with conormal @xmath320 a little closer to the origin ( so that it meets the @xmath359-axis at the point @xmath498 for some @xmath361 ) , and then slightly adjusting the size of all the subsequent blow ups from @xmath404 to @xmath499 so as not to cut out quite all of an edge at each blow up .
( see figure [ fig:9 ] .
note that some of the @xmath191 may be negative ; cf .
the construction of @xmath177 in 2 . ) for suitable @xmath191 this will create a smooth polygonal arc @xmath500 whose edges @xmath501 have conormals @xmath327 as described above .
define @xmath502 where @xmath503 is poincar dual to @xmath218 , i.e. @xmath504 . since @xmath505 is the cohomology class of the symplectic form on @xmath506 , @xmath507 is the class of the symplectic form on @xmath325 obtained from @xmath508 by the blow up procedure explained above with the @xmath322th blow up of size @xmath499 .
hence the affine length @xmath326 of the edge @xmath153 is @xmath509 because @xmath510 is the poincar dual of @xmath392 , all the edges of @xmath500 are very short except for @xmath511 which has affine length nearly @xmath86
. note that condition ( ii ) implies that @xmath500 is a chain of edges with the same intersection properties as @xmath354 .
hence the slopes decrease as one moves along @xmath500 from @xmath517 to @xmath518 , so that it is a convex polygonal arc .
therefore because @xmath500 has endpoints @xmath519 and @xmath498 where @xmath520 , it lies outside @xmath521 .
therefore , to prove ( iii ) one must simply check that it lies inside @xmath133 . condition ( i ) in the definition obviously holds . to check ( ii ) ,
let us denote the edges of @xmath500 by @xmath524 then , if @xmath525 @xmath526 is a homogeneous linear function of @xmath196 and hence is positive for @xmath522 if it is positive for @xmath196 .
further @xmath527 a homogenous linear function of @xmath196 , and so again condition ( ii ) is satisfied by @xmath528 . to check ( iii ) , observe that the positions of the edges depend linearly on @xmath196 , i.e. for each @xmath322 there are constants @xmath529 and homogeneous linear functions @xmath530 such that the edge @xmath531 lies in the line @xmath532 moreover , by lemma [ le : v ] , the line @xmath533 goes through one of the points @xmath477 or @xmath478 .
hence ( iii ) holds iff @xmath534 for all @xmath322 .
the result follows .
there is an analogous discussion for outer approximations .
these approximate @xmath133 by a polytope with a concave chain @xmath535 of edges @xmath536 lying just outside the slanted edge of @xmath133 .
we did the case @xmath537 in 2 : @xmath538 ( which in the notation of 2 corresponds to the chain of spheres @xmath539 ) is illustrated in figure [ fig:3 ] . note that this chain of edges goes between the edges with conormals @xmath540 and @xmath541 but does not include them , so that the classes @xmath542 of the edges in @xmath543 are linear combinations of the exceptional divisors @xmath544 , with no mention of @xmath205 .
again we assume that @xmath545 is minimal , i.e. the last edge @xmath546 with conormal @xmath547 is the only edge whose class @xmath542 has self intersection @xmath105 .
we shall denote the analog of @xmath392 by @xmath548 .
thus @xmath549 is such that @xmath550 for example @xmath551 .
we leave the proof of the following statement to the reader .
[ rmk : cf]just as in the case of inner approximations , the conormals occurring in an outer approximation to @xmath133 give rise to a decomposition of a triangle , which this time is @xmath133 itself .
moreover , as we shall prove in @xcite , the sequence of labels @xmath557 that occur as coefficients in the vector @xmath548 can be obtained by the following version of the euclidean algorithm .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ first write down @xmath558 copies of @xmath559 where @xmath560 , then write down @xmath561 copies of @xmath562 where @xmath563 , and so on . at the @xmath322th step one
writes down @xmath564 copies of @xmath565 where @xmath566 .
the process stops as soon as some @xmath567 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ has labels @xmath568 with multiplicities @xmath569 .
note that in the diagram on the right one starts by expanding horizontally because the rectangle is wider than it is high ; at the second step one rotates by @xmath570 and then continues .
this rotation is equivalent to taking the reciprocal of the aspect ratio of the rectangle .
hence this expansion mirrors the continued fraction.,width=288 ] as pointed out dylan thurston the combinatorics of the resulting decomposition of @xmath133 are precisely the same as the combinatorics of one of the standard ways of getting the continued fraction expansion : see figure [ fig:53 ] . hence the multiplicities @xmath564 of the labels @xmath571 give the continued fraction expansion of @xmath351 .
for example , @xmath572 has labels @xmath573 with multiplicities @xmath574 , and @xmath575 note also that the multiplicative relation @xmath576 is obvious from this point of view .
since , as we show in theorem [ thm : ell2 ] below , the labels @xmath577 determine the weights of the corresponding ball embedding problem , this multiplicative relation corresponds to the geometric fact that the total volume of the balls corresponding to an ellipsoid @xmath17 must be the same as the volume of @xmath17 .
[ rmk : sing ] ( i ) we saw above that an inner approximation to @xmath133 is constructed from a decomposition of the triangle @xmath578 with outward conormals @xmath579 , while an outer approximation to @xmath133 is constructed from a decomposition of @xmath133 itself .
the matrix @xmath580 takes the conormals of @xmath482 to those of @xmath581 .
it is easy to check that this transformation takes the first decomposition into the second .
thus inner and outer approximations are essentially the same thing , though they are related in a slightly different way to the ambient triangle @xmath311 .
\(ii ) from the point of view of singularity theory , our construction of the inner ( or outer ) approximation to @xmath133 can be considered as a kind of joint resolution of the two singular points of the corresponding toric variety .
usually , one would resolve them separately , in which case , it is the hirzebruch - jung continued fractions ( with @xmath582 rather than @xmath583 signs ) that are relevant : see fulton @xcite . in the standard resolution of a single singularity one performs the blow ups near just one of the vertices getting half of our conormals . for example
fulton s method of resolving the vertex with outward conormals @xmath260 and @xmath584 ( where @xmath585 ) begins by a cut with conormal @xmath586 .
hence if we rotate his picture anticlockwise by @xmath587 we get the half of the inner approximation to @xmath133 near the vertex @xmath477 ; cf .
figure [ fig:6 ] .
note that the orbifold structure of this vertex has stabilizer of order @xmath588 , with generator @xmath589 acting on @xmath590 via @xmath591 where @xmath592 .
the other half of this inner approximation corresponds to the vertex @xmath478 which has stabilizer of order @xmath593 acting via @xmath594 where @xmath595 .
hence this corresponds to fulton s resolution of the vertex with outward conormals @xmath260 and @xmath596 where we choose @xmath4 so that @xmath597 ; i.e. we simply interchange the roles of @xmath593 and @xmath588 .
note that the interpretation of the coefficients of the continued fraction expansion is rather different in the two cases .
[ thm : ell2 ] suppose that each pair @xmath307 and @xmath414 is mutually prime , and let @xmath599 set @xmath600 then the question of whether @xmath143 embeds into the open ellipsoid @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] is equivalent to the symplectic packing problem for @xmath4 balls with weights @xmath69 .
suppose first that @xmath601 embeds into the open ellipsoid @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] . since @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')\subset { { \settowidth{\facewd}{b}\settoheight{\faceht}{b}\raisebox{1.1\faceht}{\makebox[0pt][l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(n')$ ] we may consider @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] as a subset of @xmath602 . by corollary [ cor : v ]
the complement of @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] in @xmath602 contains the @xmath324 open balls @xmath44[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}(k_i)$ ] ( cf .
lemma [ le : open ] . )
moreover these balls can be embedded disjointly from the line in @xmath602 represented by the edge of @xmath603 with conormal @xmath320 .
similarly , @xmath143 contains the @xmath604 open balls @xmath63[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}b}}({\widehat}k_j)$ ] .
hence , rescaling by @xmath605 , we see that the given symplectic packing problem has a solution with open balls .
however , the problem was formulated in terms of embedding closed balls . to deal with this ,
observe that when @xmath606[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] there is @xmath198 such that @xmath607 embeds in @xmath608[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] .
this means that the sizes of all the open balls can be slightly increased so that they contain closed balls of the correct size .
conversely , suppose that the ball packing problem has a solution , i.e. that there is a symplectic form on the @xmath4-fold blow up @xmath89 of @xmath1 in class @xmath609 since the space of symplectic forms is open , we may suppose without loss of generality that @xmath203 is rational .
moreover , it suffices to prove that @xmath610[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')\;\;\mbox { for all } { { \lambda}}_0<{{\lambda}}.\ ] ] before proceeding further , it is convenient to introduce some notation .
we will denote the divisors of @xmath89 by @xmath611 as usual ; hence the @xmath612 , correspond to the exceptional divisors @xmath544 associated to @xmath11 .
further @xmath196 will denote a tuple of small constants , whose length ( either @xmath613 or @xmath614 ) will depend on the context .
when it is necessary to be more specific we shall denote the @xmath615-tuple @xmath196 by @xmath616 .
further @xmath617 is admissible if its first @xmath618 components @xmath619 are admissible for @xmath22 while its last @xmath604 components @xmath552 are admissible for @xmath11 . given an inner approximation @xmath620 , we shall denote by @xmath621 the @xmath137 invariant open subset of @xmath622 whose moment image is the component of @xmath623 that lies in @xmath624 .
thus @xmath625[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] is a smooth approximation to @xmath22 .
we shall denote the chain of spheres corresponding to an inner approximation @xmath620 by @xmath626 and that corresponding to an outer approximation @xmath535 by @xmath627 . finally if @xmath628 is any subset of a symplectic manifold @xmath20 and @xmath629 we shall denote by @xmath630 the set @xmath628 provided with the form @xmath631 .
note that if @xmath632 is a symplectic form on @xmath89 such that the disjoint union @xmath633 embeds in @xmath634 , then the class @xmath635 $ ] is determined by @xmath196 and is close to @xmath636 .
* claim 1 : * _ if there is a symplectic form @xmath632 on @xmath89 such that @xmath637 embeds in @xmath634 for some admissible @xmath196 , then @xmath638[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] .
_ _ proof .
_ let @xmath639 be a @xmath137-invariant neighborhood of @xmath626 whose moment image is a neighborhood of @xmath620 .
as in lemma [ le : nb ] we may suppose that @xmath639 embeds in @xmath634 .
then a neighborhood of infinity in @xmath640 may be identified with a neighborhood of infinity in @xmath621 , where @xmath621 is as above .
( in fact , @xmath641 can be obtained from @xmath621 by further blowing up near the inverse image of @xmath642 . )
moreover @xmath643 contains a copy of @xmath644 and hence , as in lemma [ le:1 ] , blows down to an open set @xmath645 containing @xmath646 .
but @xmath647 by construction , and @xmath645 is symplectomorphic to @xmath621 at infinity .
hence , by the uniqueness of symplectic forms on starshaped subsets of @xmath648 that are standard near the boundary ( see ( * ? ? ?
9.4.2 ) ) , @xmath645 is symplectomorphic to @xmath621 .
therefore @xmath649 which proves the claim.@xmath55 * claim 2 : * _ if the ball packing problem has a solution with weights @xmath69 , then for all @xmath650 , there is a symplectic form @xmath632 on @xmath89 such that @xmath637 embeds in @xmath634 for some admissible @xmath196 . _
_ if @xmath651 then @xmath652 embeds linearly in @xmath13[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] and so , for small admissible @xmath196 there is a symplectic form @xmath653 on @xmath89 such that @xmath654 embeds in @xmath655 .
note that @xmath656 = a ' + r \,{\widehat}a + c({{\delta}})$ ] , where @xmath657 and @xmath658 is a homogeneous linear function of @xmath196 . by assumption
the class @xmath659 of equation ( [ eq : ala ] ) is rational and represented by a symplectic form .
therefore , as in the proof of proposition [ prop : m ] , the homology class @xmath660 has nontrivial gromov invariant for large @xmath112 .
choose an @xmath653 tame almost complex structure @xmath114 on @xmath89 such that both @xmath626 and @xmath661 are @xmath114-holomorphic .
if @xmath114 is sufficiently generic , then as in the proof of theorem [ thm:1 ] in 2 we may suppose that the class @xmath660 is represented by a connected @xmath114-holomorphic submanifold @xmath9 that intersects @xmath626 and @xmath662 transversally .
the inflation procedure gives a family of symplectic forms @xmath663 on @xmath89 that are nondegenerate on the two configurations of spheres and lie in class @xmath664 & = & [ { { \omega}}_{r,{{\delta } } } ] + tqa_{{\lambda}}\\ & = & ( 1+tq)\left(a'+ \bigl(\frac { r + { { \lambda}}tq}{1+tq}\bigr)\,{\widehat}a+c'({{\kappa}}')+{\widehat}c({\widehat}{{\kappa}})\right ) \end{aligned}\ ] ] where @xmath665 and @xmath666 are multiples of @xmath619 and @xmath667 .
and so are admissible by lemmas [ le : eede ] and [ le : outer ] .
observe that as @xmath121 the class @xmath668 $ ] converges to @xmath669 .
moreover , for appropriate @xmath196 , @xmath626 embeds in @xmath670 for all @xmath306 , while @xmath644 embeds in @xmath670 if @xmath671 . by equation ( [ eq : la0 ] )
this completes the proof .
* proof of corollary [ cor : conn ] .
* since the targets of the embeddings are open ellipsoids , an easy continuity argument implies that it suffices to prove these statements when @xmath71 are integers . part ( i )
is equivalent to saying that all deformation equivalent symplectic forms on @xmath672 are isotopic .
it can be proved in the same way as the uniqueness of symplectic forms on @xmath89 .
one just needs to inflate along curves @xmath9 that intersect @xmath626 and @xmath627 transversally , which is possible as in the proof of claim 2 above . for more details , see @xcite .
\(ii ) follows from ( i ) just as the analogous statement for balls follows from the fact that the space of embeddings of one ball into another is connected .
let @xmath673 be an increasing sequence with limit @xmath203 . from a sequence of embeddings @xmath674[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')\ ] ]
one first uses ( i ) with target ellipsoid @xmath675[l]{$\hspace{.3\facewd}\scriptscriptstyle\circ$}}e}}}(m',n')$ ] to construct a sequence @xmath676 such that @xmath677 . by using ( i ) again , this time with target ellipsoids @xmath678
, one makes a further adjustment so that @xmath679 restricts to @xmath676 on @xmath680 .
the result follows .
@xmath55 for simplicity ,
let us rename the blow up classes in @xmath683 as @xmath684 where @xmath685 for @xmath686 and @xmath687 for @xmath688 . by propositions [ prop : mp ] and [ prop : m ] , the second statement will follow if we show that the class @xmath689 takes positive values on all the elements in @xmath690 , where @xmath691 .
but @xmath690 is generated by classes of the form @xmath218 , @xmath692 together with classes that are equivalent to the following ( after permutation of indices ) : @xmath693 evaluating @xmath125 on @xmath694 we find that we need @xmath695 i.e. @xmath696 since the other curves in @xmath697 give weaker inequalities , the result follows.@xmath55 * proof of proposition [ prop : tol ] ( ii ) . * the first statement holds as before . to prove the second , recall that @xmath698 is generated by the classes in @xmath690 together with those of the following three forms : @xmath699 ( these structural results on @xmath690 and @xmath698 are classical and not hard to prove directly from the definition . )
one gets the sharpest inequality on @xmath203 from elements of the form ( v ) , which give @xmath700 .
hence the result.@xmath55 | we show how to reduce the problem of symplectically embedding one @xmath0-dimensional rational ellipsoid into another to a problem of embedding disjoint unions of balls into appropriate blow ups of @xmath1 .
for example , the problem of embedding the ellipsoid @xmath2 into a ball @xmath3 is equivalent to that of embedding @xmath4 disjoint equal balls into @xmath1 , and so can be solved by the work of gromov , mcduff polterovich and biran .
( here @xmath4 is the ratio of the area of the major axis to that of the minor axis . ) as a consequence we show that the ball may be fully filled by the ellipsoid @xmath2 for @xmath5 and all @xmath6 , thus answering a question raised by hofer .
= 1 |
we present a general framework for solving quasilinear wave equations with _ normally hyperbolic trapping _ , which extends the semilinear framework developed by the two authors in @xcite and the _ non - trapping _ quasilinear theory developed by the first author @xcite . for concreteness ,
let us describe the relevant features for the setup of our result on kerr - de sitter space .
the region of kerr - de sitter space we are interested in can be considered a ( non - compact ) @xmath0-dimensional manifold with boundary @xmath1 , with the boundary being an ` ideal boundary ' at infinity in the standard description of kerr - de sitter space .
@xmath1 is equipped with a lorentzian b - metric @xmath2 , recalled below , depending on three parameters @xmath3 ( the cosmological constant ) , @xmath4 ( the black hole mass ) and @xmath5 , the angular momentum , though we usually drop this in the notation .
, @xmath1 and @xmath5 are such that the non - degeneracy condition ( * ? ? ?
* ( 6.2 ) ) holds , which in particular ensures that the cosmological horizon lies outside the black hole event horizon .
] recall that on any @xmath6-dimensional manifold with boundary @xmath1 , the lie algebra of smooth vector fields tangent to the boundary is denoted by @xmath7 ; in local coordinates @xmath8 , with @xmath9 a boundary defining function , these are linear combinations of @xmath10 and @xmath11 with @xmath12 coefficients , and thus they are the space of all @xmath13 sections of a vector bundle , @xmath14 , with local basis @xmath10 , @xmath11 .
a b - metric is a non - degenerate symmetric bilinear form on the fibers of @xmath14 smoothly depending on the base point ; a lorentzian b - metric is one of signature @xmath15 . in order to set up our problem , we consider two functions @xmath16 , @xmath17 , with forward , resp .
backward , time - like differentials near their respective @xmath18-set @xmath19 , which are linearly independent at their joint @xmath18-set , and let @xmath20 , with @xmath21 compact , so @xmath21 is a compact manifold with corners with three boundary hypersurfaces @xmath22 and @xmath23 , all intersected with @xmath21 .
we are interested in solving the forward problem for wave - like equations in @xmath21 , i.e. imposing vanishing cauchy data at @xmath24 , which we assume is disjoint from @xmath25 ; initial value problems with general cauchy data can always be converted into an equation of this type . , the cauchy hypersurface @xmath24 and the hypersurface @xmath26 , which has two connected components which lie beyond the cosmological horizon and beyond the black hole event horizon , respectively .
the horizons at @xmath25 themselves are the projections to the base of the ( generalized ) radial sets @xmath27 , discussed below , each of which has two components , corresponding to the two horizons .
the projection to the base of the bicharacteristic flow is indicated near a point on @xmath28 ; near @xmath29 , the directions of the flowlines are reversed .
lastly , @xmath30 is the trapped set , and the projection of a trapped trajectory approaching @xmath30 within @xmath31 , discussed below , is indicated . ] in such a b - setting , one measures regularity with respect to @xmath7 , and for non - negative integer @xmath32 , one lets @xmath33 be the set of ( complex - valued ) @xmath34 such that @xmath35 for @xmath36 and @xmath37 . here
@xmath38 is the @xmath39-space with respect to any b - metric , such as the kerr - de sitter metric .
further , one introduces the weighted b - sobolev spaces @xmath40 .
sections of vector bundles in @xmath41 are defined by local trivializations ; the sobolev spaces on @xmath21 are defined by restriction .
we point out that the b - differential @xmath42 , defined locally by @xmath43 maps @xmath44 to @xmath45 , where @xmath46 is the dual bundle of @xmath14 .
we are typically interested in real - valued solutions ; @xmath47 , etc . , denotes these spaces , and @xmath42 preserves preserves reality .
the wave equations we consider include those of the form @xmath48 where @xmath49 , and for each @xmath50 , @xmath51 , depending smoothly on @xmath52 , and is only relevant if @xmath53 . ]
@xmath54 our central result in the form which is easiest to state is , with a more general version given below in theorems [ thmintrogeneralkg ] and [ thmintrogeneralwave ] : [ thmintrowave ] ( see corollaries [ corquasilinearkds ] and [ corquasilinearkgkds ] . ) on kerr - de sitter space with angular momentum @xmath55 , for @xmath56 sufficiently small , @xmath57 with sufficiently small @xmath58-norm , the wave equation @xmath59 , with @xmath60 as above with @xmath61 for all @xmath62 , has a unique smooth global forward solution of the form @xmath63 , @xmath64 , @xmath65 , @xmath66 identically @xmath67 near @xmath68 .
further , the analogous conclusion holds for the klein - gordon equation @xmath69 with @xmath70 sufficiently small , without the presence of the @xmath71 term , i.e. for @xmath72 sufficiently small , if @xmath73 with sufficiently small @xmath58-norm , @xmath74 has a unique smooth global forward solution @xmath75 .
in fact , for klein - gordon equations one can also obtain a leading term , analogously to @xmath71 , which now has the form @xmath76 , @xmath77 the resonance of @xmath78 with the largest imaginary part ; thus @xmath79 , so this is a decaying solution .
the _ only _ reason the assumption @xmath55 is made is due to the possible presence ( to the extent that we do not disprove it here ) of resonances in @xmath80 , apart from the @xmath18-resonance with constants as the resonant state , for larger @xmath5 .
below we give a general result in a form that makes it clear that this is the only remaining item to check indeed , this even holds in natural vector bundle settings .
having stated the result , we now explain _ why _ it holds . for this purpose
it is best to start with the underlying linear equation ; after all , the non - linearity is ` just ' a rather serious perturbation ! in general , analysis of b - differential operators ( locally finite sums of finite products of elements of @xmath7 ) , such as @xmath81 , has two ingredients , corresponding to the two orders , smoothness and decay , of the sobolev spaces : 1 .
b - regularity analysis .
this provides the framework for understanding pde at high b - frequencies , which in non - degenerate situations involves the b - principal symbol and perhaps a subprincipal term .
this is sufficient in order to control solutions @xmath82 in @xmath83 modulo @xmath84 , @xmath85 , i.e. modulo a space with higher regularity , but no additional decay . since for the inclusion @xmath86 to be compact one needs both @xmath87 and @xmath88
, this does not control the problem modulo relatively compact errors . 2 . normal operator analysis .
this provides a framework for understanding the decay properties of solutions of the pde .
the normal operator is obtained by freezing coefficients of the differential operator @xmath89 at @xmath68 to obtain a dilation - invariant b - operator @xmath90 .
one then mellin transforms the normal operator in the normal variable to obtain a family of operators @xmath91 , depending on the mellin - dual variable @xmath92 .
the b - regularity analysis , in non - degenerate situations , gives control of this family @xmath91 in a fredholm sense , uniformly as @xmath93 with @xmath94 bounded .
however , in any such strip , @xmath95 will still typically have finitely many poles @xmath96 ; these poles , called _ resonances _ , dictate the asymptotic behavior of solutions of the pde . in order to have a fredholm operator @xmath89 , one needs to work in spaces such as @xmath83 , where @xmath97 is such that there are no resonances @xmath96 with @xmath98 .
one can also work in slightly more general spaces , such as @xmath99 , @xmath100 , identified with a space of distributions via @xmath63 , @xmath64 , @xmath65 , corresponding to @xmath101 .
now , the b - regularity analysis for our non - elliptic equation involves the ( null)-bicharacteristic flow . in view of the version of hrmander s theorem on propagation of singularities in this setting , and in view of the a priori control on cauchy data at @xmath24 , what one would like is that all bicharacteristics tend to @xmath102 in one direction .
moreover , for the purposes of the adjoint problem , which effectively imposes cauchy data at @xmath26 , one would like that the bicharacteristics tend to @xmath103 in the other direction .
unfortunately , bicharacteristics within @xmath104 can never leave this space , and thus will not tend to @xmath102 .
this is mostly resolved , however , by the conormal bundle of the _ horizons _ at @xmath25 , which give rise to a bundle of saddle points for the bicharacteristic flow . since the flow is homogeneous , it is convenient to consider it in @xmath105 .
the characteristic set in @xmath106 has two components @xmath107 , with @xmath108 forward - oriented ( i.e. future oriented time functions increase along null - bicharacteristics in @xmath108 ) , @xmath109 backward oriented
. then the images of the conormal bundles of the horizons in the cosphere bundle are submanifolds @xmath110 of @xmath111 , with one - dimensional stable ( @xmath112)/unstable ( @xmath113 ) manifold @xmath114 transversal to @xmath115 .
( the flow within @xmath27 need not be trivial ; if it is , one has _
radial points _ , as in the @xmath116 de sitter - schwarzschild space .
however , for simplicity we refer to the @xmath27 estimates as _ radial point estimates _ in general . )
the realistic ideal situation , called a _
non - trapping _ one , then is if all ( null-)bicharacteristics in @xmath117 tend to @xmath118 in the backward direction , and @xmath119 in the forward direction , with a similar statement for @xmath108 , with backward and forward interchanged .
, there can not be non - trivial bicharacteristics in @xmath120 tending to @xmath28 in both the forward and the backward direction , since a bicharacteristic is either completely in @xmath111 , or completely outside it .
] in this non - trapping setting the only subtlety is that the propagation estimates through @xmath27 require that the differentiability order @xmath32 and the decay order @xmath97 be related by @xmath121 for a suitable @xmath122 ( dictated by the hamilton dynamics at @xmath27 ) , i.e. the more decay one wants , the higher the regularity needs to be .
this is still not the case in kerr - de sitter space , though it is true for neighborhoods of the static patch in de sitter space , and its perturbations .
the additional ingredient for kerr - de sitter space is normally hyperbolic trapping , given by smooth submanifolds @xmath123 . here
@xmath124 are invariant submanifolds for the hamilton flow , given by the transversal intersection of locally defined smooth , hamilton flow invariant , @xmath125 , with @xmath126 transversal to @xmath127 , and @xmath128 . combining results of @xcite ( which would work directly in a dilation invariant setting ) and @xcite
we show that for @xmath100 sufficiently small , one can propagate @xmath83 estimates through @xmath124 .
this suffices to complete the b - regularity setup if the non - trapping requirement is replaced by : all ( null-)bicharacteristics in @xmath129 tend to either @xmath130 in the backward direction , and @xmath131 in the forward direction , with the tending to @xmath132 allowed in only _
one _ of the forward and backward directions , with a similar statement for @xmath108 , with backward and forward interchanged . finally , this _ is _ satisfied in kerr - de sitter space , and also in its b - perturbations ( the whole setup is perturbation stable ) .
next , one needs to know about the resonances of the operator . for the wave operator ,
the only resonance with non - negative imaginary part is @xmath18 , with the kernel of @xmath91 one dimensional , consisting of constants .
since strips can only have finitely many resonances , there is @xmath100 such that in @xmath133 the only resonance is @xmath18 ; then @xmath134 works for our fredholm setup . for the klein - gordon equation with @xmath70 small
, the @xmath135 resonance at @xmath18 moves to @xmath136 inside @xmath137 , see @xcite .
thus , one can either work with @xmath138 where @xmath139 is sufficiently small ( depending on @xmath140 ) , or with @xmath134 , though with @xmath141 now identified with @xmath142 .
we now discuss the non - linear terms . here
the basic point is that @xmath143 is an algebra if @xmath144 , and thus for such @xmath32 , products of elements of @xmath83 possess even more decay if @xmath100 , but they become more growing if @xmath145 .
thus , one is forced to work with @xmath146 .
first , with the simplest semilinear equation , with no derivatives in the non - linearity @xmath60 ( so @xmath147 is replaced by @xmath53 ) , the regularity losses due to the normally hyperbolic trapping are in principle sufficiently small to allow for a contraction mapping principle ( picard iteration ) based argument . however
, for the actual wave equation on kerr - de sitter space , the @xmath18-resonance prohibits this , as the iteration maps outside the space @xmath134 .
thus , it is the semilinear klein - gordon equation that is well - behaved from this perspective , and this was solved by the authors in @xcite . on the other hand , if derivatives are allowed , with an at least quadratic behavior in @xmath148 , then the non - linearity annihilates the @xmath18-resonance .
unfortunately , since the normally hyperbolic estimate loses @xmath149 derivatives , as opposed to the usual real principal type / radial point loss of one derivative , the solution operator for @xmath150 will not map @xmath151 back into the desired sobolev space , preventing a non - linear analysis based on the contraction mapping principle .
fortunately , the nash - moser iteration is designed to deal with just such a situation . in this paper
we adapt the iteration to our requirements , and in particular show that semilinear equations of the kind just described are in fact solvable . in particular , we prove that all the estimates used in the linear problems are _
tame_. here we remark that klainerman s early work on global solvability involved the nash - moser scheme @xcite , though this was later removed by klainerman and ponce @xcite . in the present situation
the loss of derivatives seems much more serious , however , due to the trapping , so it seems unlikely that the solution scheme can be made more ` classical ' . however , we are also interested in quasilinear equations .
quasilinear versions of the above non - trapping scenario were studied by the first author @xcite , who showed the solvability of quasilinear wave equations on perturbations of de sitter space .
the key ingredient in dealing with quasilinear equations is to allow operators with coefficients with regularity the same kind as what one is proving for the solutions , in this case @xmath83-regularity .
all of the smooth linear ingredients ( microlocal elliptic regularity , propagation of singularities , radial points ) have their analogue for @xmath83 coefficients if @xmath32 is sufficiently large .
thus , in @xcite a picard - type iteration , @xmath152 was used to solve the quasilinear wave equations on de sitter space .
notice that @xmath153 has non - smooth coefficients ; indeed , these lie in a weighted b - sobolev space . in our kerr - de sitter situation
there is normally hyperbolic trapping .
however , notice that as we work in decaying sobolev spaces modulo constants , @xmath154 differs from a kerr - de sitter operator with smooth coefficients , @xmath155 , by one with _
decaying coefficients_. this means that one can combine the smooth coefficient normally hyperbolic theory , as in the work of dyatlov @xcite , with a tame estimate in @xmath83 with @xmath145 ; the sign of @xmath97 here is a crucial gain since for @xmath145 the propagation estimates through normally hyperbolic trapped sets behave in exactly the same way as real principal type estimates . in combination
this provides the required tame estimates for kerr - de sitter wave equations , and nash - moser iteration completes the proof of the main theorem .
we emphasize that our treatment of these quasilinear equations is systematic and general .
thus , quasilinear equations which at @xmath23 are modelled on a finite dimensional family @xmath156 , @xmath157 small corresponding to the zero resonances ( thus the family is @xmath18-dimensional without @xmath18-resonances ! ) , of smooth b - differential operators on a vector bundle with scalar principal symbol which has the bicharacteristic dynamics described above ( radial sets , normally hyperbolic trapping , etc . )
fits into it , _ provided two conditions hold for the normal operator _
( i.e. the dilation invariant model associated to @xmath89 at @xmath68 ) . of the operator
is irrelevant . ] 1 .
first , _ the resonances for the model @xmath158 _ have negative imaginary part , or if they have @xmath18 imaginary part , the non - linearity annihilates them .
second , the _ normally hyperbolic trapping estimates _ of dyatlov @xcite hold for @xmath91 ( as @xmath159 ) in @xmath160 for some @xmath161 . in the semiclassical rescaling , with @xmath162 , @xmath163 ,
this is a statement about @xmath164 , @xmath165 . by dyatlov s recent result @xcite
this indeed is the case if @xmath166 satisfies that _ at @xmath30 _ its skew - adjoint part , @xmath167 , for @xmath168 has semiclassical principal symbol bounded above by @xmath169 for some @xmath170 , where @xmath171 is the minimal expansion rate in the normal directions at @xmath30 ; see ( * ? ? ?
* theorem 1 ) and the remark below it ( which allows the non - trivial skew - adjoint part , denoted by @xmath172 there , microlocally at @xmath30 ) .
it is important to point out that in view of the decay of the solutions either to @xmath18 if there are no real resonance , or to the space of resonant states corresponding to real resonances , the conditions must be checked for at most a finite dimensional family of elements of the ` smooth ' algebra @xmath173 , and moreover there is no need to prove tame estimates , deal with rough coefficients , etc .
, for this point , and one is in a dilation invariant setting , i.e. can simply mellin transform the problem .
thus , in principle , solving wave - type equations on more complicated bundles is reduced to analyzing these two aspects of the associated linear model operator at infinity .
concretely , we have the following two theorems : [ thmintrogeneralkg ] let @xmath1 be a kerr - de sitter space with angular momentum @xmath174 that satisfies ( * ? ? ?
* ( 6.13 ) ) , and @xmath5 ensures non - trapping classical dynamics for the null - geodesic flow . ]
@xmath175 a vector bundle over it with a positive definite metric @xmath176 on @xmath175 , and let @xmath177 have principal symbol @xmath178 ( times the identity ) , and suppose that @xmath179 satisfies that 1 .
the large parameter principal symbol of @xmath180 , with the adjoint taken relative to @xmath181 , at the trapped set @xmath30 is @xmath182 as an endomorphism of @xmath175 , 2 .
@xmath183 has no resonances in @xmath80 .
then for @xmath56 sufficiently small , there exists , in particular , for the value of @xmath184 . ] @xmath185 such that the following holds : if @xmath73 has a sufficiently small @xmath186-norm , then the equation @xmath187 has a unique smooth global forward solution @xmath75 .
in particular , the conditions at @xmath30 for the theorem hold if @xmath55 , @xmath188 , @xmath189 the differential form dalembertian , or indeed if @xmath190 is a @xmath18th order operator , since hyperbolicity is shown in @xcite in the full stated range of @xmath5 , while for @xmath116 , @xmath191 can be computed explicitly at @xmath30 , with @xmath176 being the riemannian metric of the form @xmath192 near the projection of @xmath30 , where @xmath193 has the form @xmath194 , @xmath195 an appropriate boundary defining function on @xmath1 strictly away from the horizons .
thus , in this case the only assumption in the theorem remaining to be checked is the second one , concerning resonances .
[ thmintrogeneralwave ] let @xmath1 be a kerr - de sitter space with angular momentum @xmath174 that satisfies ( * ? ?
* ( 6.13 ) ) , @xmath175 a vector bundle over it with a positive definite metric @xmath176 on @xmath175 , and let @xmath177 have principal symbol @xmath178 ( times the identity ) .
suppose that @xmath179 is such that @xmath183 has a simple resonance at @xmath18 , with resonant states spanned by @xmath196 , and no other resonances in @xmath80 .
consider the family @xmath197 , @xmath198 with small enough norm .
suppose that 1 .
this family only has resonances at @xmath18 in @xmath80 , and these are given by @xmath199 , 2 .
@xmath30 is uniformly normally hyperbolic for @xmath200 for @xmath71 of small norm , 3 .
the large parameter principal symbol of @xmath180 , with the adjoint taken relative to @xmath181 , at the trapped set @xmath30 is @xmath182 , 4 .
@xmath201 for @xmath198
. then for @xmath56 sufficiently small , there exists is given in in the course of the proof of this theorem in section [ subsecgeneralproofs ] . ] @xmath185 such that the following holds : if @xmath57 has a sufficiently small @xmath202-norm , then the equation @xmath203 has a unique smooth global forward solution of the form @xmath63 , @xmath64 , @xmath204 , @xmath66 identically @xmath67 near @xmath68 . here
` uniformly normally hyperbolic ' in the theorem means that one has a smooth family @xmath205 of trapped sets , with a smooth family of stable / unstable manifolds , with uniform bounds ( within this family ) on the normal expansion rates for the flow , which ensures that the normally hyperbolic estimates are uniform within the family ( for small @xmath71 ) ; see the discussion around for details .
again , the conditions at @xmath30 for the theorem hold if @xmath55 , @xmath188 , if @xmath190 is a @xmath18th order operator , @xmath206 the differential form dalembertian , since the structurally stable @xmath97-normally hyperbolic statement is shown in @xcite ( which implies the uniform normal hyperbolicity required in the theorem ) , while for @xmath116 , @xmath191 can be computed explicitly at @xmath30 , as mentioned above , and upper bounds on this are stable under perturbations .
the uniform normal hyperbolicity condition at @xmath30 holds if @xmath207 , @xmath188 , @xmath189 the differential form dalembertian , with @xmath208 being a kerr - de sitter metric for @xmath198 with small norm since the hyperbolicity of @xmath30 was shown in this generality in @xcite .
however , the computation of @xmath191 is more involved .
we now discuss previous results on kerr - de sitter space and its perturbations .
there seems to be little work on non - linear equations in kerr - de sitter type settings ; indeed the only paper the authors are aware of is the earlier paper of the authors , @xcite , in which the semilinear klein - gordon equation was studied ( with small data well - posedness shown ) with non - linearity depending on @xmath82 only , so that the losses due to the trapping could still be handled by a contraction mapping argument .
in addition , the same paper also analyzed non - linearities depending on @xmath148 provided these had a special structure at the trapped set .
there is more work on the linear equation on perturbations of de sitter - schwarzschild and kerr - de sitter spaces : a rather complete analysis of the asymptotic behavior of solutions of the linear wave equation was given in @xcite , upon which the linear analysis of the present paper is ultimately based .
previously in exact kerr - de sitter space and for small angular momentum , dyatlov @xcite has shown exponential decay to constants , even across the event horizon ; see also the more recent work of dyatlov @xcite .
further , in de sitter - schwarzschild space ( non - rotating black holes ) bachelot @xcite set up the functional analytic scattering theory in the early 1990s , while later s barreto and zworski @xcite and bony and hfner @xcite studied resonances and decay away from the event horizon , dafermos and rodnianski in @xcite showed polynomial decay to constants in this setting , and melrose , s barreto and vasy @xcite improved this result to exponential decay to constants .
there is also physics literature on the subject , starting with carter s discovery of this space - time @xcite , either using explicit solutions in special cases , or numerical calculations , see in particular @xcite , and references therein . while it received more attention , the linear , and thus the non - linear , equation on kerr space ( which has vanishing cosmological constant ) does not fit directly into our setting ; see the introduction of @xcite for an explanation and for further references and @xcite for more background and additional references .
some of the key works in this area include the polynomial decay on kerr space which was shown recently by tataru and tohaneanu @xcite and dafermos , rodnianski and shlapentokh - rothman @xcite , after pioneering work of kay and wald in @xcite and @xcite in the schwarzschild setting .
andersson and blue @xcite proved a decay result for the maxwell system on slowly rotating kerr spaces ; see also the earlier work of bachelot @xcite in the schwarzschild setting . on the non - linear side ,
luk @xcite established global existence for semilinear wave equations on kerr space under a null condition .
( there was also recent work by marzuola , metcalfe , tataru and tohaneanu @xcite and tohaneanu @xcite on strichartz estimates , which are applied to the study of semilinear wave equations with power non - linearities , and by donninger , schlag and soffer @xcite on @xmath209 estimates on schwarzschild black holes , following @xmath209 estimates of dafermos and rodnianski @xcite , of blue and soffer @xcite on non - rotating charged black holes giving @xmath210 estimates , and finster , kamran , smoller and yau @xcite on dirac waves on kerr . )
the plan of this paper is the following . in section [ secnonsmoothcalc ]
we show that the non - smooth pseudodifferential operators of @xcite facilitate tame estimates ( operator bounds , composition , etc . ) , with section [ subsectamemicrolocal ] establishing tame elliptic estimates in section [ subsectameelliptic ] , tame real principal type and radial point estimates in section [ subsectamepropagation ] and tame estimates at normally hyperbolic trapping in section [ subsecnontrapping ] for @xmath145 . in section [ subsectrapping ] , we adapt dyatlov s analysis at normally hyperbolic trapping given in @xcite to our needs .
finally , in section [ secquasilinear ] we solve our quasilinear equations by first showing that the microlocal results of section [ subsectamemicrolocal ] combine with the high energy estimates for the relevant normal operators following from the discussion of section [ subsectrapping ] to give tame estimates for the forward propagator in section [ subsecforward ] , and then showing in section [ subsecsolve ] that the nash - moser iteration indeed allows for solving our wave equations .
section [ subsecsolvekg ] then explains the changes required for quasilinear klein - gordon equations .
finally , in section [ subsecgeneralproofs ] we show how our methods apply in the general settings of theorems [ thmintrogeneralkg ] and [ thmintrogeneralwave ] .
the authors are very grateful to semyon dyatlov for providing a preliminary version of his manuscript @xcite and for discussions about it , as well as for pointing out the reference @xcite .
they are also grateful to gunther uhlmann , richard melrose and rafe mazzeo for comments and interest in this project .
in this section we prove the basic tame estimates for the @xmath211-coefficient , or simply non - smooth , b - pseudodifferential operators defined in @xcite .
we start with the tame mapping estimate , proposition [ propmapping ] , which essentially states that for non - smooth pseudodifferential operators @xmath212 , a high regularity norm of @xmath213 can be estimated by a high regularity norm of @xmath212 times a low regularity norm of @xmath82 , plus a low regularity norm of @xmath212 times a high regularity norm of @xmath82 .
this is stronger than the a priori continuity estimate one gets from the bilinear map @xmath214 , which would require a product of high norms of both . in case @xmath212
is a multiplication operator , this is essentially a b - version of a ( weak ) moser estimate , see corollary [ corhbmodule ] .
we work on the half space @xmath215 with coordinates @xmath216 ; the coordinates in the fiber of the b - cotangent bundle are denoted @xmath217 , i.e. we write b - covectors as @xmath218 .
recall from @xcite the symbol class @xmath219 with the norm @xmath220 where @xmath221 denotes the mellin transform in @xmath9 and fourier transform in @xmath222 of @xmath5 .
left quantizations of such symbols , denoted @xmath223 , act on @xmath224 by @xmath225 also recall @xmath226 and @xmath227 . for brevity , we will use the following notation for sobolev , symbol class and operator class norms , with the distinction between symbolic and b - sobolev norms being clear from the context : @xmath228 if @xmath212 is a b - operator acting on an element of a weighted b - sobolev space with weight @xmath97 ( which will be apparent from the context ) , then @xmath229 is to be understood as @xmath230 , similarly for @xmath231 .
lastly , for @xmath232 , we write @xmath233 , by an abuse of notation , for an unspecified @xmath234-seminorm of @xmath212 . recall the notation @xmath235 for @xmath236 .
[ propmapping ] ( extension of ( * ? ? ?
* proposition 3.9 ) . )
let @xmath237 , @xmath238 , and suppose @xmath239 is such that @xmath240 , @xmath241 .
then @xmath212 defines a bounded map @xmath242 , and for all fixed @xmath243 with @xmath244 there is a constant @xmath245 such that @xmath246 observe that by the assumptions on @xmath32 and @xmath247 , the intervals of allowed @xmath243 are always non - empty ( since they contain @xmath248 and @xmath249 ) .
estimates of the form , called ` tame estimates ' e.g. in @xcite , are crucial for applications in a nash - moser iteration scheme .
we compute @xmath250 we split the inner integral into two pieces , corresponding to the domains of integration @xmath251 and @xmath252 , which can be thought of as splitting up the action of @xmath212 on @xmath82 into a low - high and a high - low frequency interaction .
we estimate @xmath253 and we claim that the integral which is the first factor on the right hand side is uniformly bounded in @xmath254 : indeed , if @xmath255 , then we use @xmath256 on the domain of integration , thus @xmath257 since @xmath258 ; if , on the other hand , @xmath259 , then @xmath260 gives @xmath261 since @xmath262 ; hence , from , the @xmath263 norm of the low - high frequency interaction in @xmath213 is bounded by @xmath264 .
we estimate the norm of high - low interaction in a similar way : we have @xmath265 if @xmath255 , the first inner integral on the right hand side is bounded by @xmath266 where we use @xmath240 , and this integral is finite in view of @xmath267 ; if @xmath259 , then @xmath268 which is finite in view of @xmath269 . in summary , we need @xmath270 and can then bound the @xmath263 norm of the high - low interaction by @xmath271 .
the proof is complete .
using @xmath272 , we obtain the following weak version of the moser estimate for the product of two b - sobolev functions : [ corhbmodule ] let @xmath144 , @xmath273 . if @xmath274 , then @xmath275 , and one has an estimate @xmath276 for fixed @xmath277 . in particular , for @xmath278 , @xmath279 for fixed @xmath258 .
we give a tame estimate for the norms of expansion and remainder terms arising in the composition of two non - smooth operators : [ propnonsmoothcomp ] suppose @xmath280 , @xmath281 are such that @xmath144 , @xmath282 and @xmath283 .
suppose @xmath284 , @xmath285 .
put @xmath286 then @xmath287 and @xmath288 , and for @xmath258 fixed , @xmath289 the statements about the @xmath290 follow from corollary [ corhbmodule ] .
for the purpose of proving the estimate for @xmath291 , we define @xmath292 where we write @xmath293 , similarly for @xmath294 .
notice that in particular @xmath295 .
then @xmath296 with @xmath297 by taylor s formula , hence @xmath298 which implies the claimed estimate for @xmath299 . for @xmath300
, we use a trick of beals and reed @xcite as in the proof of theorem 3.12 in @xcite to reduce the statement to the case @xmath299 : recall that the idea is to split up @xmath301 into a ` trivial ' part @xmath302 with compact support in @xmath254 and @xmath6 parts @xmath303 , where @xmath303 has support in @xmath304 , and then writing @xmath305 for some constants @xmath306 using the leibniz rule ; then what we have proved above for @xmath299 can be applied to the @xmath62-th summand on the right hand side , which we expand to order @xmath307 , giving the result .
we also need sharper bounds for reciprocals and compositions of b - sobolev functions on a compact @xmath6-dimensional manifold with boundary .
localizing using a partition of unity , we can simply work on @xmath215 .
[ prophbrec ] ( extension of ( * ? ? ?
* lemma 4.1 ) . )
let @xmath309 , @xmath310 , @xmath311 , and suppose that @xmath312 near @xmath313 .
then @xmath314 , and one has an estimate @xmath315 for any fixed @xmath316 and some @xmath32-dependent @xmath317 .
choose @xmath318 such that @xmath319 on @xmath313 , @xmath320 on @xmath321 , and such that moreover @xmath322 on @xmath323 .
then we have @xmath324 .
we now iteratively prove higher regularity of @xmath325 and an accompanying ` tame ' estimate : let us assume @xmath326 for some @xmath327 .
let @xmath328 be an operator with principal symbol @xmath329
. then @xmath330\frac{w}{a+u}\bigr\|_0\right ) \\ & \lesssim c_0^{-1}\biggl(\|w\|_{s'}+\bigl\|\frac{w}{a+u}\bigr\|_{s'-1}+\bigl\|\psi[\lambda_{s'},u]\frac{w}{a+u}\bigr\|_0\biggr ) , \end{split}\ ] ] where we used that the support assumptions on @xmath331 and @xmath332 imply @xmath333 , and @xmath334\in{\psi_{{{\mathrm{b}}}}}^{s'-1}$ ] .
hence , in order to prove that @xmath335 , it suffices to show that @xmath336\colon { h_{{{\mathrm{b}}}}}^{s'-1}\to{h_{{{\mathrm{b}}}}}^0 $ ] .
let @xmath337 .
since @xmath338 we have , by taking a first order taylor expansion of @xmath339 around @xmath340 , @xmath341v){\;\!\widehat{\ } \;\!}(\zeta)=\sum_{|\beta|=1}\int\left(\int_0 ^ 1{\partial}_\zeta^\beta \lambda_{s'}(\xi+t(\zeta-\xi))\,dt\right)({{}^{{{\mathrm{b}}}}\!d}_z^\beta u){\;\!\widehat{\ } \;\!}(\zeta-\xi)\hat v(\xi)\,d\xi,\ ] ] thus , writing @xmath342 , @xmath343v){\;\!\widehat{\ } \;\!}(\zeta)|\lesssim\int\left(\int_0 ^ 1{\langle}\xi+t(\zeta-\xi){\rangle}^{s'-1}\,dt\right)\big|\widehat{u'}(\zeta-\xi)\big||\hat v(\xi)|\,d\xi.\ ] ] to obtain a tame estimate for the @xmath344 norm of this expression , we again use the method of decomposing the integral into low - high and high - low components : the low - high component is bounded by @xmath345 the first inner integral , in view of @xmath346 , so the @xmath347 is bounded by @xmath348 , which cancels the corresponding term in the denominator , is finite for @xmath316 . for the high - low component , we likewise estimate @xmath349 and the first inner integral on the right hand side is bounded by @xmath350 because of @xmath351 , which is finite for @xmath267 .
we conclude that @xmath352v\|_0\leq c_{\mu\nu}(\|u\|_\mu\|v\|_{s'-1}+\|u\|_{s'}\|v\|_\nu),\ ] ] for @xmath353 . plugging this into yields @xmath354 where the implicit constant in the inequality is independent of @xmath355 and @xmath82 . using the abbreviations @xmath356 , @xmath357 , @xmath358 and fixing @xmath316 ,
this means @xmath359 with the implicit constant being independent of @xmath360 .
we will use this for @xmath361 with @xmath362 , and for @xmath363 , we will take @xmath364 , thus obtaining a tame estimate for @xmath365 . in more detail , for @xmath366
, we have @xmath367 which gives , with @xmath368 , @xmath369 using the bound @xmath370 . for @xmath371
, we have @xmath372 thus for integer @xmath373 with @xmath374 , @xmath375 where we used @xmath376 in the last inequality , thus proving the estimate in case @xmath32 is an integer ; in the general case , we just use @xmath377 for @xmath378 , in particular for @xmath379 , and use the above with @xmath380 replaced by @xmath381 . as in @xcite ,
one thus obtains regularity results for compositions , but now with sharper estimates . to illustrate how to obtain these ,
let us prove an extension of ( * ? ? ?
* proposition 4.5 ) .
let @xmath1 be a compact @xmath6-dimensional manifold with boundary , @xmath309 , @xmath382 .
[ propcompwithanalytic ] let @xmath383 .
if @xmath384 , @xmath385 , is holomorphic in a simply connected neighborhood @xmath21 of the range of @xmath82 , then @xmath386 , and @xmath387 for fixed @xmath316 . moreover , there exists @xmath170 such that @xmath388 depends continuously on @xmath389 , @xmath390 .
observe that @xmath391 is compact .
let @xmath392 denote a smooth contour which is disjoint from @xmath391 , has winding number @xmath67 around every point in @xmath391 , and lies within the region of holomorphicity of @xmath393 .
then , writing @xmath394 with @xmath395 holomorphic in @xmath21 , we have @xmath396 since @xmath397 is continuous by proposition [ prophbrec ] , we obtain , using the estimate , @xmath398 which implies in view of @xmath382 .
the continuous ( in fact , lipschitz ) dependence of @xmath399 on @xmath400 is a consequence of proposition [ prophbrec ] and corollary [ corhbmodule ] .
we also study compositions @xmath401 for @xmath402 and real - valued @xmath82 .
[ propcompwithsmooth ] ( extension of ( * ? ? ?
* proposition 4.7 ) . )
let @xmath402 , @xmath385 . then for @xmath403 , we have @xmath386 , and one has an estimate @xmath404 for fixed @xmath316 .
in fact , @xmath401 depends continuously on @xmath82 .
the proof is the same as in @xcite , using almost analytic extensions , only we now use the sharper estimate to obtain .
[ propcompwithsmooth2 ] ( extension of ( * ? ? ?
* proposition 4.8 ) . )
let @xmath402 , and @xmath405 ; put @xmath406 .
then @xmath407 , and one has an estimate @xmath408 for fixed @xmath316 and some @xmath317 .
in fact , @xmath401 depends continuously on @xmath82 .
the proof is the same as in @xcite , but now uses the sharper estimate .
when stating microlocal regularity estimates ( like elliptic regularity , real principal type propagation , etc . ) for operators with coefficients in @xmath409 , we will give two quantitative statements , one for ` low ' regularities @xmath410 , in which we will not make use of any tame estimates established earlier , and one for ` high ' regularities @xmath411 , in which the tame estimates will be used .
to concisely write down tame estimates , we use the following notation : the right hand side of a tame estimate will be a real - valued function , denoted by @xmath89 , of the form @xmath412 here , the @xmath413 and @xmath414 are continuous functions . in applications , @xmath415 will be a low / high regularity norm of the coefficients of a non - smooth operator , and @xmath416 will be a low / high regularity norm of a function that an operator is applied to .
the important feature of such functions @xmath89 is that they are linear in the @xmath416 , and all @xmath417 , corresponding to high regularity norms , only appear in the first power . concretely , we have the following quantitative elliptic estimate : [ proptameelliptic ] ( cf .
* theorem 5.1 ) . )
let @xmath418 and @xmath419 .
suppose @xmath420 has a homogeneous principal symbol @xmath421 .
moreover , let @xmath422 .
let @xmath423 , and suppose @xmath424 is elliptic at @xmath425 .
let @xmath426 be such that @xmath427 and @xmath428 , and suppose that @xmath429 satisfies @xmath430 then there exists @xmath431 elliptic at @xmath425 such that @xmath432 , and for @xmath433 , @xmath434 , the estimate @xmath435 holds . for @xmath436 , @xmath170 , there is a tame estimate @xmath437 let us point out that in our application of such an estimate to the study of nonlinear equations it will be irrelevant what exactly the low regularity norms in are ; in fact , it will be sufficient to know that there is _ some _ tame estimate of the general form , and this in turn is in fact clear without any computation , namely it follows directly from the fact that we have tame estimates for all ` non - smooth ' operations involved in the proof of this proposition .
the same remark applies to all further tame microlocal regularity results below .
the only point where the precise numerology does matter is when one wants to find an explicit bound on the number of required derivatives for the forcing term in theorems [ thmintrowave ] , [ thmintrogeneralkg ] and [ thmintrogeneralwave ] , as we will do .
we can assume that @xmath438 by conjugating @xmath439 by @xmath440 .
choose @xmath441 elliptic at @xmath425 such that @xmath442 is elliptic on @xmath443 .
let @xmath444 be a b - ps.d.o with full symbol @xmath445 independent of @xmath446 , whose principal symbol is @xmath447 , and define @xmath448 then by proposition [ prophbrec ] and corollary [ corhbmodule ] , we have @xmath449 put @xmath450 , then @xmath451 with @xmath452 ; by proposition [ propnonsmoothcomp ] , we have for @xmath453 @xmath454 now , since @xmath455 , we need to estimate the @xmath456-norms of @xmath457 , @xmath458 and @xmath459 , which we will do using proposition [ propmapping ] . in the low regularity regime , we have , for @xmath434 and @xmath433 , using and : @xmath460 giving . in the high regularity regime , in fact for @xmath461
, we have , for @xmath170 , @xmath462 giving .
the proof is complete .
there is a similar tame microlocal elliptic estimate for operators of the form @xmath463 with @xmath464 as above and @xmath465 , as in part ( 2 ) of ( * ? ? ?
* theorem 5.1 ) , where the tame estimate now also involves the @xmath466-norm of the ` smooth part ' @xmath467 of the operator for some ( @xmath32-dependent ) @xmath468 . will only depend on finitely many complex parameters , there is no need to prove an estimate which is also tame with respect to the @xmath466-norm of @xmath467 ; however , this could easily be done in principle . ]
tame estimates for real principal type propagation and propagation near radial points can be deduced from a careful analysis of the proofs of the corresponding results in @xcite .
the main observation is that the regularity requirements , given in the footnotes to the proofs of these results in @xcite , indicate what regularity is needed to estimate the corresponding terms : for example , an operator in @xmath469 with @xmath470 maps @xmath471 to @xmath472 under the condition @xmath473 , which is to say that one has a bound @xmath474 this means that the only places where one needs to use tame operator bounds for operators with coefficients of regularity @xmath32 are those where the condition for mapping properties etc . to hold reads @xmath475 where @xmath92 is the regularity of the target space , i.e. where @xmath92 is comparable to the regularity @xmath32 of the coefficients .
we again only prove the tame real principal type estimate in the interior ; the estimate near the boundary is proved in the same way , see also the discussion at the end of section [ subsectameelliptic ] .
[ proptamerealprtype ] ( cf .
* theorem 6.6 ) . )
let @xmath476 .
suppose @xmath477 has a real , scalar , homogeneous principal symbol @xmath442 , and let @xmath478 , @xmath479 .
let @xmath480 .
suppose @xmath32 and @xmath481 are such that @xmath482 and suppose @xmath483 satisfies @xmath484 .
suppose @xmath485 , and let @xmath486\to{{}^{{{\mathrm{b}}}}t}^*{{\overline{{\mathbb r}^n_+}}}{\setminus o}$ ] be a segment of a null - bicharacteristic of @xmath442 with @xmath487 , then @xmath488 for all @xmath489 $ ] .
moreover , for all @xmath490 elliptic at @xmath425 there exist @xmath491 elliptic at @xmath492 and @xmath493 elliptic on @xmath494)$ ] such that for @xmath495 , @xmath170 , @xmath496 moreover , for @xmath497 , @xmath170 , there is a tame estimate @xmath498 we follow the proof of the regularity result in @xcite and state the estimates needed to establish and along the way .
using the notation of the proof of ( * ? ? ?
* theorem 6.6 ) , but now calling the regularization parameter @xmath499 , in particular @xmath500 is the regularized commutant , which depends on a positive constant @xmath1 chosen below , and putting @xmath501 , we have , assuming @xmath502 and @xmath503 for now , @xmath504u , u{\rangle}\\ & = \frac{1}{2}{\langle}i(p_m - p_m^*)\check a_\delta u,\check a_\delta u{\rangle}- { \operatorname{re}}{\langle}i\check a_\delta\tilde f,\check a_\delta u{\rangle}+{\operatorname{re}}{\langle}i\check
a_\delta p_{m-1}u,\check a_\delta u{\rangle}\\ & \equiv i+ii+iii .
\end{aligned}\ ] ] for @xmath170 , we can bound the first term by @xmath505 the second one by @xmath506 where in turn @xmath507 for @xmath434 by proposition [ propmapping ] .
we estimate the third term by @xmath508u,\check a_\delta u{\rangle}|\ ] ] and further , with @xmath509 denoting a part of the expansion of @xmath510 $ ] as defined after ( * ? ? ?
* footnote 28 ) , @xmath511u,\check a_\delta u{\rangle}|&\leq c(m)\|p_{m-1}\|_{{h_{{{\mathrm{b}}}}}^{n/2 + 1+(m/2 - 1)_+ + { \epsilon}}{\psi_{{{\mathrm{b}}}}}^{m-1}}\|u\|_{\tilde s+m-3/2}^2
\\ & \hspace{3cm}+\|r_2 u\|_{-(m-1)/2}^2+\|\check a_\delta u\|_{(m-1)/2}^2 , \end{aligned}\ ] ] where @xmath512 therefore , we obtain , see ( * ? ? ?
* equation ( 6.24 ) ) , @xmath513+b_\delta^*b_\delta+m^2(\lambda\check a_\delta)^*(\lambda\check a_\delta)-e_\delta\bigr)u , u\right{\rangle}\\ & \hspace{2cm}\geq -|{\langle}e_\delta u , u{\rangle}|-\|\check a_\delta f\|_{-(m-1)/2}^2-l^2+\|b_\delta u\|_{l^2_{{\mathrm{b}}}}^2 , \end{split}\ ] ] where @xmath514 and @xmath89 is ` tame ' ; more precisely , for @xmath433 , @xmath434 , @xmath515 and for @xmath516 , @xmath517 next , in order to exploit the positive commutator of the principal symbols of @xmath518 and @xmath519 in the estimate , we introduce operators @xmath520 with principal symbols @xmath521 such that @xmath522 ; then @xmath523=\operatorname{op}(j^-\check a_\delta h_{p_m}\check a_\delta)+r_1+r_2+r_3+r_4,\ ] ] see ( * ? ? ?
* equation ( 6.27 ) ) , where @xmath524 and @xmath525 , hence @xmath526 thus , further following the proof in @xcite to equation ( 6.28 ) and beyond , it remains to bound @xmath527 from below , which is accomplished by @xmath528
lastly , for general @xmath529 , we rewrite the equation @xmath530 as @xmath531 with @xmath532 , @xmath533 , where @xmath534 ; hence , replacing @xmath439 by @xmath535 , @xmath82 by @xmath536 and @xmath140 by @xmath537 in the above estimates is equivalent to just replacing @xmath140 by @xmath537 in the b - sobolev norms of the coefficients of @xmath439 .
choosing @xmath538 as in @xcite then implies the estimates and with @xmath539 , @xmath540 an elliptic multiple of @xmath541 , and @xmath212 elliptic on the microsupport of @xmath542 . in a similar manner
, we can analyze the proof of the radial point estimate , obtaining , in the notation of @xcite : [ proptameradial ] let @xmath476 , @xmath56 .
let @xmath480 , where @xmath543 , @xmath544 , with @xmath545 and @xmath546 having real , scalar , homogeneous principal symbols @xmath421 and @xmath547 , respectively ; moreover @xmath548 , @xmath549 and @xmath550 with @xmath551 and @xmath552 .
suppose that the conditions ( 1)-(4 ) in @xcite hold for @xmath553 , and @xmath554 where @xmath555 is self - adjoint at every point .
finally , assume that @xmath32 and @xmath481 satisfy @xmath556 suppose @xmath483 is such that @xmath484 . 1
. [ enumthmradintobdy ] if @xmath557 , let us assume that in a neighborhood of @xmath27 , @xmath558 is disjoint from @xmath559 .
[ enumthmradfrombdy ] if @xmath560 , let us assume that a punctured neighborhood of @xmath27 , with @xmath27 removed , in @xmath561 is disjoint from @xmath559
. then in both cases , @xmath27 is disjoint from @xmath559 .
quantitatively , for every neighborhood @xmath562 of @xmath27 , there exist @xmath563 elliptic at @xmath27 , @xmath490 with microsupport in the respective a priori control region in the two cases above , with @xmath564 , @xmath17 , and @xmath565 , such for @xmath495 , @xmath170 , we have , with implicit dependence of the appearing constants on seminorms of the smooth operators @xmath566 and @xmath567 : @xmath568 moreover , for @xmath497 , @xmath170 , there is a tame estimate @xmath569 one detail changes as compared to the previous proof : while it still suffices to only assume microlocal regularity @xmath570 at @xmath27 , we now in addition need to assume local regularity @xmath571 , which is due to the use of elliptic regularity in the proof given in @xcite .
we now extend the proof of non - trapping estimates on weighted b - sobolev spaces at normally hyperbolically trapped sets given in ( * ? ? ?
* theorem 3.2 ) to the non - smooth setting . to set this up ,
let @xmath572 with @xmath573 where the adjoint is taken with respect to a fixed smooth b - density ; an example to keep in mind here and in what follows is @xmath574 for a smooth lorentzian b - metric @xmath193 on @xmath215 , considered a coordinate patch of kerr - de sitter space , in which case @xmath575 , and the threshold weight in theorem [ thmtamehyptr ] below is @xmath438 .
let @xmath576 be the principal symbol of @xmath577 .
let us use the coordinates @xmath578 on @xmath579 and write @xmath580 . with @xmath581 denoting the characteristic set of @xmath577 , we make the following assumptions : 1 .
[ enumntgamma ] @xmath582 is a smooth submanifold disjoint from the image of @xmath583 , so @xmath584 is elliptic near @xmath30 , 2 .
@xmath585 is a smooth submanifold of @xmath586 in a neighborhood @xmath587 of @xmath30 , 3 .
@xmath588 is a smooth submanifold of @xmath589 transversal to @xmath586 in @xmath587 , 4 .
@xmath585 has codimension @xmath590 in @xmath589 , @xmath588 has codimension @xmath67 , 5 .
@xmath585 and @xmath588 intersect transversally in @xmath589 with @xmath591 , 6 .
the vector field @xmath592 is tangent to both @xmath585 and @xmath588 , and thus to @xmath30 , 7 .
[ enumnttrapped ] @xmath585 is backward trapped for the hamilton flow , @xmath588 is forward trapped ; in particular , @xmath30 is a trapped set . in view of condition ( [ enumntgamma ] )
, we can take @xmath593 appropriately extended to @xmath594 , as the inverse of a boundary defining function of fiber infinity @xmath106 in @xmath594 . then
, let @xmath595 be the rescaled hamilton vector field of @xmath576 .
we make quantitative assumptions related to condition ( [ enumnttrapped ] ) : let @xmath596 be a defining function of @xmath585 in @xmath111 , and let @xmath597 be a defining function of @xmath588 .
thus , @xmath585 is defined within @xmath106 by @xmath598 .
let @xmath599 we then assume that 1 .
@xmath600 and @xmath601 satisfy @xmath602 with @xmath603 smooth near @xmath30 and @xmath604 smooth near @xmath30 .
this is consistent with the ( in)stability of @xmath588 ( @xmath585 ) , 2 .
@xmath9 satisfies @xmath605 which is consistent with the stability of @xmath588 , 3 .
[ enumntquant3 ] near @xmath30 , @xmath606 for some smooth @xmath607 , which holds in view of our choice of @xmath608 .
here we recall from ( * ? ? ?
* lemma 5.1 ) , see also ( * ? ? ?
* lemma 2.4 ) , that in the closely related semiclassical setting . ]
one can arrange for any ( small ) @xmath170 that @xmath609 where @xmath171 and @xmath610 are the minimal and maximal normal expansion rates ; see ( * ? ? ?
* equations ( 5.1 ) and ( 5.2 ) ) for the definition of the latter , with @xmath171 also given in below .
note that in these works of dyatlov our @xmath611 is denoted by @xmath612 .
in particular , if @xmath1 is replaced by @xmath613 , and if @xmath577 is dilation invariant , then the semiclassical and the b - settings are equivalent via the mellin transform and a rescaling , see e.g. ( * ? ? ? * section 3.1 ) ; since in our general case @xmath614 is what matters , we can replace @xmath577 by @xmath615 , and in particular applies , with the expansion rate calculated using @xmath616 .
we now perturb @xmath577 by a non - smooth operator @xmath617 , that is , we consider the operator @xmath618 where for some fixed @xmath56 , we have @xmath619 , @xmath620 , and @xmath621 .
we then have the following tame non - trapping estimate at @xmath30 : [ thmtamehyptr ] using the above notation and making the above assumptions , let @xmath622 be such that @xmath623 suppose @xmath483 is such that @xmath484 .
then for @xmath624 and for any neighborhood @xmath562 of @xmath30 , there exist @xmath625 elliptic at @xmath30 and @xmath626 with @xmath627 , @xmath628 , @xmath629 , and @xmath565 , such that the following estimate holds for @xmath495 , @xmath170 : @xmath630 moreover , for @xmath497 , @xmath170 , there is a tame estimate @xmath631 on the other hand , for @xmath632 and for appropriate @xmath633 with @xmath634 , the estimates and hold as well .
these estimates are understood in the sense that if all quantities on the right hand side are finite , then so is the left hand side , and the inequality holds .
the main part of the argument , in particular the choice of the commutant , is a slight modification of the positive commutator argument of ( * ? ? ?
* theorem 3.2 ) ; the handling of the non - smooth terms is a modification of the proof of the radial point estimate , ( * ? ? ?
* theorem 6.10 ) .
in particular , the positivity comes from differentiating the weight @xmath440 in the commutant . to avoid working in weighted b - sobolev spaces for the non - smooth problem
, we will conjugate @xmath439 by @xmath440 , giving an advantageous ( here meaning negative ) contribution to the imaginary part of the subprincipal symbol of the conjugated operator near @xmath30 .
_ throughout this proof , we denote operators and their symbols by the corresponding capital and lower case letters , respectively .
_ concretely , put @xmath635 , and define @xmath636 where @xmath637 then @xmath638 , and we must show a non - trapping estimate for @xmath639 on _ unweighted _ b - sobolev spaces .
a simple computation shows that @xmath640 but @xmath641 with @xmath642 near @xmath30 by , hence , using , @xmath643 with @xmath644 , where @xmath645 has principal symbol @xmath646 near @xmath30 , and @xmath647 .
notice that by assumption on @xmath97 , @xmath648 is elliptic on @xmath30 .
we now turn to the positive commutator argument : fix @xmath649 and define @xmath650 let @xmath651 for @xmath434 and @xmath652 for @xmath653 , further @xmath654 for @xmath655 to be chosen below , @xmath656 near @xmath18 , @xmath657 , and finally @xmath658 , @xmath320 near @xmath18 .
define for @xmath659 , specified later , @xmath660 on @xmath661 , we have @xmath662 , thus the argument of @xmath663 is bounded above by @xmath664 .
moreover , @xmath665 and @xmath666 , therefore @xmath5 is supported in any given neighborhood of @xmath30 if one chooses @xmath291 and @xmath667 small .
notice that @xmath5 is merely a _
symbol which does not grow at the boundary .
however , b - analysis for operators with conormal coefficients can easily be seen to work without much additional work , in fact , a logarithmic change of variables essentially reduces such a b - analysis on @xmath215 to the analysis of operators corresponding to uniform symbols on @xmath668 .
moreover , the proofs of composition results of smooth and non - smooth b-ps.d.os presented in @xcite go through without changes if one uses b-ps.d.os with non - growing conormal , instead of smooth , symbols .. then even though @xmath5 is not a smooth symbol of @xmath215 with the standard smooth structure , it becomes smooth if one changes the smooth structure of @xmath215 by blowing up the boundary to the @xmath669-th order , i.e. by taking @xmath670 as a boundary defining function , thus obtaining a manifold @xmath671 , which is @xmath215 as a topological manifold , but with a different smooth structure ; in particular , the function @xmath672 is smooth on @xmath671 in view of @xmath673 .
moreover , the blow - down map @xmath674 induces isomorphisms ( see e.g. @xcite ) @xmath675 therefore , one can continue to work on @xmath215 , tacitly assuming that all functions and operators live on , and all computations are carried out on , @xmath671 . ]
define the regularizer @xmath676 near @xmath30 , and put @xmath677 . put @xmath678 and define @xmath679 near @xmath30 by @xmath680 , @xmath681 .
then , with @xmath682 , we obtain , using - : @xmath683 where , writing @xmath684 in the second and third line , @xmath685 \end{aligned}\ ] ] note that in the definition of @xmath686 , by the choice of @xmath687 and using the fact that @xmath663 is bounded by a constant multiple of @xmath688 on its support , the constant being uniform for @xmath689 , the term @xmath690 dominates all other terms on the support of @xmath691 for @xmath291 and @xmath667 small enough , hence @xmath692 , and its contribution will be controlled by virtue of the sharp grding inequality . the term arising from @xmath693
will be controlled using the a priori regularity assumption of @xmath639 on @xmath588 , and @xmath694 , which is supported away from the characteristic set , will be controlled using elliptic regularity .
proceeding with the argument , we first make the simplification @xmath695 by replacing @xmath696 by @xmath697 , and we assume @xmath502 and @xmath503 for now .
then we have , as in the proof of ( * ? ? ?
* theorem 6.10 ) , @xmath698u_r , u_r{\rangle}+\left{\langle}\frac{1}{2i}(p_{0,r}-p_{0,r}^*)a_\delta u_r , a_\delta u_r\right{\rangle}\\ & = - \left{\langle}\frac{1}{2i}(\tilde p_{m , r}-\tilde p_{m , r}^*)a_\delta u_r ,
a_\delta u_r\right{\rangle}\\ & \hspace{4 cm } - { \operatorname{re}}{\langle}ia_\delta f , a_\delta u_r{\rangle}+{\operatorname{re}}{\langle}ia_\delta\tilde p_{m-1,r}u_r , a_\delta u_r{\rangle}. \end{aligned}\ ] ] estimating each term on the right hand side as in the proof of ( * ? ? ?
* theorem 6.10 ) and using , we obtain for any @xmath699 : @xmath700 + e_1+e_1'+b)a_\delta\bigr)u_r , u_r\big{\rangle}\geq -c_\mu - \mu\|a_\delta u_r\|_{(m-1)/2}^2.\ ] ] here and in what follows , we in particular absorb all terms involving @xmath701 into the constant @xmath702 . on the left hand side , the @xmath703-term can be dropped because of @xmath704 for sufficiently localized @xmath5 . moreover ,
the principal symbol of @xmath705 near @xmath30 is @xmath706 with @xmath60 an elliptic symbol of order @xmath707 , since , by assumption on @xmath97 , we have @xmath708 near @xmath30 .
therefore , we can write @xmath709 , where @xmath710 , @xmath711 , @xmath712 . again , the resulting term in the pairing involving @xmath713 can be dropped ; also , the term involving @xmath714 can be dropped at the cost of changing the constant @xmath702 , since @xmath715 .
hence , introducing @xmath716 , with real principal symbols , satisfying @xmath717 , we get @xmath718 we now plug the commutator relation into this estimate .
we obtain several terms , which we bound as follows : first , since @xmath719 uniformly , @xmath720 is a uniformly bounded family of maps @xmath721 ; thus , choosing @xmath722 with @xmath723 and with @xmath724 disjoint from @xmath725 , we conclude @xmath726 for some @xmath727 with @xmath629 .
next , the term @xmath728 is uniformly bounded , as detailed in the proof of ( * ? ? ?
* theorem 6.10 ) .
moreover , by the sharp grding inequality , see the argument in the proof of ( * ? ? ?
* theorem 6.6 ) , @xmath729 further , we obtain two terms involving @xmath730 ; introducing @xmath731 elliptic on @xmath732 , these can be bounded for @xmath699 by @xmath733 here , for the first estimate , we employ ( * ? ? ?
* theorem 3.12 ( 3 ) ) to obtain @xmath734 and @xmath735 is easily seen to be uniformly bounded from @xmath736 to @xmath472 , whereas @xmath737 , thus @xmath738 . for the second estimate ,
we simply use @xmath739 , and for the third estimate , we apply the peter paul inequality to the first pairing ; to bound the second pairing , we use the boundedness of @xmath740 . finally , including the terms @xmath741 into the estimate obtained from by making use of the above estimates , we obtain @xmath742 where @xmath743 is elliptic on @xmath732 with @xmath744 , and @xmath745
is identically @xmath67 near the projection of @xmath746 to the base @xmath1 .
since @xmath747 and @xmath748 have positive lower bounds near @xmath30 , we can absorb the terms on the right involving @xmath749 into the left hand side by choosing @xmath750 sufficiently small , at the cost of changing the constant @xmath702 ; likewise , @xmath751 has a positive lower bound near @xmath661 , hence the term on the right involving @xmath752 can be absorbed into the left hand side for small @xmath750 .
dropping the first two terms on the left hand side , we obtain the @xmath753-regularity of @xmath639 at @xmath30 , hence @xmath754 , and a corresponding tame estimate , which follows from a careful analysis of the above argument as in the proof of proposition [ proptamerealprtype ] .
next , we remove the restriction @xmath502 : let @xmath534 .
the idea , as before , is to rewrite @xmath530 as @xmath755 , where @xmath756 , with real principal symbols , satisfy @xmath757 .
we now have to be a bit careful though to not change the imaginary part of the subprincipal symbol of @xmath535 at @xmath30 . concretely ,
we choose @xmath758 self - adjoint with principal symbol @xmath759 near @xmath30 ; then @xmath760.\ ] ] clearly , @xmath761 , and the principal symbol of the second term is @xmath762)= -ih_{p_0}\lambda^+ = -ix(m_0-m)\rho^{m_0 - 1}c_f\ ] ] near @xmath30 by , hence , using , @xmath763 with @xmath764 and @xmath712 ; therefore , the first part of the proof with @xmath439 and @xmath82 replaced by @xmath535 and @xmath536 , respectively , applies .
the proof of the theorem in the case @xmath765 is complete .
when the role of @xmath585 and @xmath588 is reversed , there is an overall sign change , and we thus get a advantageous ( now meaning positive ) contribution to the subprincipal part of the conjugated operator @xmath766 for @xmath767 ; the rest of the argument is unchanged . complementing the results above on negatively weighted spaces , we recall results of dyatlov from @xcite on semiclassical estimates for smooth operators at normally hyperbolic trapping , which via the mellin transform correspond to estimates on non - negatively weighted spaces . here
we present the results in the semiclassical setting , then in section [ subsecforward ] we relate this to the solvability of linear equations with sobolev coefficients in theorem [ thmnhypsemiglobal ] and theorem [ thmexpansion ] .
we recall that prior to dyatlov s work , wunsch and zworski @xcite and nonnenmacher and zworski @xcite studied semiclassical estimates at normally hyperbolic trapping ; this was in turn much preceded by the work of grard and sjstrand @xcite in the analytic category .
the advantage of dyatlov s framework for us , especially as espoused in @xcite , is the explicit size of the ` spectral gap ' ( discussed below ) , which was also shown by nonnenmacher and zworski @xcite , the explicit inclusion of a subprincipal term of the correct sign , and the relative ease with which the parameter dependence can be analyzed .
we first recall dyatlov s semiclassical setting for @xmath768 both formally self - adjoint , with @xmath769 having non - negative principal symbol , @xmath770 elliptic in the standard sense .
in fact , dyatlov states the results in the special case @xmath135 , but by ellipticity of @xmath770 in the standard sense , it is straightforward to allow general @xmath140 ; see also the remark ( * ? ? ? *
bottom of p. 2 ) .
the main assumption , see @xcite , then is that @xmath771 has normally hyperbolic trapping semiclassically at @xmath772 compact , is the intersection of dyatlov s @xmath773 with the semiclassical characteristic set of @xmath439 , and similarly our @xmath774 are the intersection of dyatlov s @xmath775 with the characteristic set of @xmath439 . ] with all bicharacteristics of @xmath776 , except those in the stable ( @xmath112 ) and unstable ( @xmath113 ) submanifolds @xmath774 , entering the elliptic set of @xmath777 in the forward ( the exception being for only the @xmath112 sign ) , resp .
backward ( @xmath113 ) direction , and @xmath778 , where @xmath779 is the minimal normal expansion rate of the flow at @xmath780 , discussed above and in . if @xmath777 is microlocally in @xmath781 near @xmath780 , with @xmath782 having a non - negative principal symbol there , dyatlov shows that there is @xmath783 such that for @xmath784 , @xmath785 in view of @xmath780 lying in a compact subset of @xmath786 , the order @xmath32 is irrelevant in the sense that the estimate for one value of @xmath32 implies that for all other via elliptic estimates ; thus , one may just take @xmath787 , and even replace @xmath788 by @xmath18 , in which case this is an @xmath39-estimate , as stated by dyatlov .
suppose now that one has a family of operators @xmath789 depending on another parameter , @xmath790 , in a compact space @xmath791 , with @xmath792 depending continuously on @xmath790 , with values in @xmath793 , satisfying all of the assumptions listed above .
suppose moreover that this family satisfies the normally hyperbolic assumptions with @xmath794 continuously depending on @xmath790 in the @xmath13 topology , and uniform bounds for the normal expansion rates in the sense that both @xmath795 and the constant @xmath796 in @xmath797 with @xmath798 the unstable and stable normal tangent bundles at @xmath30 , can be chosen uniformly ( cf .
* equation ( 5.1 ) ) ) ; @xmath171 is then the @xmath347 of these possible choices of @xmath795 .
( note that since the trapped set dynamics involves arbitrarily large times , it is _ not _ automatically stable , unlike the dynamics away from the trapped set . ) in this case the implied constant @xmath799 in , as well as @xmath800 , is uniform in @xmath790 .
note that @xmath97-normal hyperbolicity for every @xmath97 implies the local ( hence global , in view of compactness ) uniformity of the normal dynamics by structural stability ; see ( * ? ? ?
* section 1 ) and ( * ? ? ?
* section 5.2 ) . to see this uniformity in @xmath799 ,
we first point out that in ( * ? ? ?
* lemma 5.1 ) the construction of @xmath801 can be done continuously with values in @xmath13 in this case . then in the proof of given in @xcite
, we only need to observe that the direct estimates provided are certainly uniform in this case for families @xmath792 , and furthermore for the main argument , using semiclassical defect measures , one can pass to an @xmath39-bounded subsequence @xmath802 such that @xmath803 , with @xmath804 for some @xmath805 in addition to @xmath806 converging to some @xmath807 . concretely ,
all of dyatlov s results in ( * ? ? ?
* section 2 ) are based on elliptic or ( positive ) commutator identities or estimates which are uniform in this setting . in particular , ( * ? ? ?
* lemma 2.3 ) is valid with @xmath808 , @xmath809 with convergence in @xmath810 .
( this uses that one can take @xmath811 in definition 2.1 , with @xmath812 , since the difference between @xmath811 and @xmath813 is bounded by a constant times the squared @xmath39-norm of @xmath802 times the operator norm bound of @xmath814 , with the latter going to @xmath18 . )
then with @xmath815 in place of @xmath816 , one still gets lemma 3.1 , which means that lemma 3.2 still holds with @xmath600 ( the limiting @xmath817 ) using lemma 2.3 .
then the displayed equation above ( * ? ? ?
* equation ( 3.9 ) ) still holds with the limiting @xmath818 , again by lemma 2.3 , and then one can finish the argument as dyatlov did . with this modification ,
one obtains the desired uniformity .
this in particular allows one to apply even if @xmath776 and @xmath777 depend on @xmath446 ( in a manner consistent with the other requirements ) , which can also be dealt with more directly using dyatlov s model form ( * ? ? ?
* lemma 4.3 ) .
it also allows for uniform estimates for families depending on a small parameter in @xmath141 , denoted by @xmath819 below , needed in section [ secquasilinear ] . allowing @xmath771 and @xmath777 depending on @xmath446 means , in particular , that we can replace the requirement on @xmath782 by the principal symbol of @xmath782 being @xmath820 , @xmath821 , and drop @xmath446 , so one has @xmath822 at this point it is convenient to rewrite this estimate , removing @xmath777 from the right hand side at the cost ( or benefit ! ) of making it microlocal . _ from here on it is convenient to change the conventions and not require that @xmath771 is formally self - adjoint ( though it is at the principal symbol level , namely it has a real principal symbol ) ; translating back into the previous notation , one would replace @xmath771 by its ( formally ) self - adjoint part , and absorb its skew - adjoint part into @xmath777 .
_ namely , we have [ thmnhypsemimic ] suppose @xmath771 satisfies the above assumptions , in particular the semiclassical principal symbol of @xmath823 being @xmath824 at @xmath780 .
, one has @xmath825 ; notice the minus sign on the right hand side .
] with @xmath826 analogous to theorem [ thmtamehyptr ] , with wave front set sufficiently close to @xmath780 , we have , for sufficiently small @xmath827 and for all @xmath468 and @xmath828 , @xmath829 note that the differential orders are actually irrelevant here due to wave front set conditions .
take @xmath830 with non - negative principal symbol such that @xmath831 is disjoint from @xmath832 , and so that all backward bicharacteristics from points not in @xmath833 , as well as forward bicharacteristics from points not in @xmath834 , reach the elliptic set of @xmath777 , and with @xmath835 elliptic on the complement of the elliptic set of @xmath777 .
let @xmath836 to be such that @xmath837 is disjoint from @xmath832 but @xmath838 .
let @xmath839 have wave front set near @xmath833 , with @xmath840 and with @xmath841 and with no backward bicharacteristic from @xmath832 reaching @xmath842 take @xmath843 elliptic on @xmath780 , with @xmath844 , again with non - negative principal symbol , with no backward bicharacteristic from @xmath845 reaching @xmath846 thus , all backward and forward bicharacteristics of @xmath771 reach the elliptic set of @xmath843 or @xmath777 .
see figure [ fignhypmic ] for the setup . and
@xmath588 , respectively , which intersect at @xmath30 ( large dot ) .
we use complex absorbing potentials @xmath777 ( with @xmath831 outside the large dashed circle ) and @xmath843 ( with @xmath845 inside the small dashed circle ) .
we obtain an estimate for @xmath847 by combining with microlocal propagation from the elliptic set of @xmath848 . ]
then @xmath849u+(i-\tilde a_+)[\tilde p_0,\tilde b_3]u - i\tilde q_0\tilde b_3u,\ ] ] so @xmath850u\\ & \qquad+\tilde b_0(\tilde p_0-i\tilde q_0)^{-1}(i-\tilde a_+)[\tilde p_0,\tilde b_3]u\\ & \qquad - i\tilde b_0(\tilde p_0-i\tilde q_0)^{-1}\tilde q_0\tilde b_3u+\tilde b_0(i-\tilde b_3)u , \end{aligned}\ ] ] and by , for @xmath851 , @xmath852 now , @xmath853 , so the corresponding terms in can be absorbed into @xmath854 .
on the other hand , since @xmath855)$ ] is disjoint from @xmath833 , the backward bicharacteristics from it reach the elliptic set of @xmath848 , and so we have the microlocal real principal type estimate for @xmath82 : @xmath856u\|_{h^{s - m}_{\hbar}}\lesssim h\|\tilde b_2 u\|_{h^{s-1}_{\hbar}}+\|\tilde b_1\tilde p_0 u\|_{h^{s - m}_{\hbar}}\ ] ] as @xmath857\in h{\psi_{{\hbar}}}^{m-1}(x)$ ] , so by , @xmath858u\|_{h^s_{\hbar}}\lesssim h^{-1}\|\tilde
b_2 u\|_{h^{s-1}_{\hbar}}+h^{-2}\|\tilde b_1\tilde p_0 u\|_{h^{s - m}_{\hbar}}.\ ] ] thus , follows if we can estimate @xmath859u\|_{h^s_{\hbar}}$ ] .
now , @xmath860)\cap\tilde\gamma_-=\emptyset$ ] by arrangement . in order to microlocalize , we now introduce a nontrapping model , @xmath861 .
we claim that @xmath862u-(\tilde p_0-i\tilde q_0)^{-1}\tilde a_+[\tilde p_0,\tilde b_3]u\ ] ] satisfies @xmath863 for all @xmath864 .
notice that for any @xmath865 one certainly has @xmath866 by plus its non - trapping analogue . to see , notice that @xmath867u,\ ] ] so by , with @xmath828 replaced by any @xmath868 ( since @xmath828 was arbitrary ) , and for any @xmath468 , @xmath869u\|_{h^{s'-m}_{\hbar}}\lesssim h^{n}\|u\|_{h^{s_0}_{\hbar}},\ ] ] since @xmath861 is non - trapping , hence @xmath870 propagates semiclassical wave front sets along forward bicharacteristics , and no backward bicharacteristic from @xmath845 can reach @xmath860)\subset{{\mathrm{wf}}_{{\hbar}}}'(\tilde a_+)\cap{{\mathrm{wf}}_{{\hbar}}}'(i-\tilde b_3)$ ] , proving the claim .
then , since backward bicharacteristics from @xmath832 do not encounter @xmath860)\cap\tilde\gamma_+$ ] before reaching the elliptic set of @xmath777 or @xmath843 , we conclude that @xmath871u\|_{h^s_{\hbar}}\\ & \qquad\leq \|\tilde b_0 ( \tilde p_0-i\tilde q_0-i\tilde q_1)^{-1}\tilde a_+[\tilde p_0,\tilde b_3]u\|_{h^s_{\hbar}}+\|\tilde b_0v\|_{h^s_{\hbar}}\\ & \qquad\lesssim h\|\tilde b_2 u\|_{h^s_{\hbar}}+\|\tilde b_1\tilde p_0 u\|_{h^{s - m+1}_{\hbar}}+h^n\|u\|_{h^{s_0}_{\hbar}}. \end{aligned}\ ] ] this proves , and thus the theorem .
we now generalize the setting considered in @xcite for the study of quasilinear equations on static asymptotically de sitter spaces to allow for normally hyperbolic trapping , as discussed in the previous section .
thus , working on a compact manifold @xmath1 with boundary @xmath25 , we assume that the operator @xmath439 is of the form @xmath872 , continuously depending on a small parameter @xmath873 , where @xmath874 for a smooth b - metric @xmath193 on @xmath1 that continuously depends on one real parameter ; here , @xmath56 .. ] we assume : 1 .
the characteristic set @xmath875 of @xmath577 has the form @xmath876 with @xmath107 a union of connected components of @xmath589 , 2 .
@xmath577 has normally hyperbolic trapping at @xmath123 for small @xmath819 , as detailed in assumptions ( [ enumntgamma])-([enumntquant3 ] ) in section [ subsectamepropagation ] , 3 .
@xmath577 has radial sets @xmath877 , which , in appropriate directions transverse to @xmath27 , are sources ( @xmath112)/sinks ( @xmath113 ) for the null - bicharacteristic flow within @xmath111 , with a one - dimensional stable / unstable manifold intersecting @xmath111 transversally ; for details , see @xcite . in particular , there are @xmath878 , @xmath879 , such that for a homogeneous degree @xmath880 boundary defining function @xmath608 of fiber infinity in @xmath594 and with @xmath881 , @xmath882 1 .
@xmath21 is compact , 2 .
putting @xmath886 , the @xmath19 intersect the boundary @xmath68 transversally , and @xmath24 and @xmath26 intersect only in the interior of @xmath1 , and they do so transversally , 3 .
the differentials of @xmath883 and @xmath884 have opposite timelike characters near their respective zero sets within @xmath21 ; more specifically , @xmath883 is future timelike , @xmath884 past timelike , 4 .
there is a boundary defining function @xmath9 of @xmath1 such that @xmath887 is timelike on @xmath888 with timelike character opposite to the one of @xmath883 , i.e. @xmath887 is past oriented , 5 .
the metric @xmath193 is _ non - trapping _ in the following sense : all bicharacteristics in @xmath889 from any point in @xmath890 flow ( within @xmath891 ) to @xmath131 in the forward direction ( i.e. either enter @xmath892 in finite time or tend to the radial set @xmath28 or the trapped set @xmath132 ) and to @xmath130 in the backward direction , and from any point in @xmath893 to @xmath894 in the forward direction and to @xmath895 in the backward direction , with tending to @xmath124 allowed in only one of the two directions . recall the space @xmath896 of distributions which are supported ( @xmath897 ) at the ` artificial ' boundary hypersurface @xmath24 and extendible ( @xmath112 ) at @xmath26 , and the other way around for @xmath898 .
then we have the following global energy estimates : this result does not rely on the dynamical structure of @xmath439 at the boundary , but only on the timelike nature of @xmath887 and of @xmath906 and @xmath907 near @xmath24 and @xmath26 , respectively , see also ( * ? ? ?
* remark 7.4 ) .
let us stress that we assume the parameter @xmath400 to be _ small _ so that in particular the skew - adjoint part of @xmath908 is small and does not affect the radial point and normally hyperbolic trapping estimates which are used in what follows ; the general case without symmetry assumptions on @xmath909 will be discussed in section [ subsecgeneralproofs ] . using a duality argument and
the tame estimates for elliptic regularity and the propagation of singularities ( real principal type , radial points , normally hyperbolic trapping ) given in propositions [ proptameelliptic ] , [ proptamerealprtype ] and [ proptameradial ] and theorem [ thmtamehyptr ] , we thus obtain solvability and higher regularity : [ lemmawaveglobalsolve ] ( cf .
* lemma 7.5 ) . )
let @xmath910 and assume @xmath911 , @xmath912 .
there exists @xmath900 such that for @xmath901 , there is @xmath245 with the following property : if @xmath913 , then there exists a unique @xmath914 such that @xmath530 , and @xmath82 moreover satisfies @xmath915 here , the implicit constant depends only on @xmath32 and @xmath916 for @xmath170 .
the proof proceeds as the proof given in the reference .
the tame estimate in particular is obtained by iterative use of the aforementioned microlocal regularity estimates ; the given bound for @xmath828 comes from an inspection of the norms in these estimates which correspond to the terms called @xmath917 in .
[ corwaveglobalcont ] let @xmath910 and assume @xmath911 , @xmath912 .
there exists @xmath900 such that for @xmath901 , there is @xmath245 with the following property : if @xmath914 is such that @xmath918 , then the estimate holds .
[ corwavepropsing ] let @xmath912 , @xmath919 , and assume @xmath911 ; moreover , let @xmath145 .
then there is @xmath245 such that the following holds : any @xmath920 with @xmath918 in fact satisfies @xmath914 , and obeys the estimate @xmath921 the proof of the two corollaries is as in the cited reference . for the radial point estimate involved in the proof of corollary [ corwavepropsing ] , we need the additional assumption @xmath922 , which however is automatically satisfied since @xmath346 and the @xmath347 is negative for @xmath145 .
we now note that the mellin transformed normal operator @xmath923 satisfies global large parameter estimates corresponding to the semiclassical microlocal estimates of theorem [ thmnhypsemimic ] . in order to state this
precisely we recall the connection between the b - structure , the normal operator ( and the large parameter algebra ) and the mellin transform of the latter . the weighted b - sobolev spaces @xmath924 are isometric to the large parameter sobolev spaces on @xmath25 on the line @xmath925 in @xmath141 via the mellin transform @xmath926 ; see ( * ? ? ? * equation ( 3.8 ) ) .
further , the latter can be described in terms of semiclassical sobolev spaces , namely the restriction @xmath927 to @xmath925 of the mellin transform identifies @xmath924 with @xmath928 see ( * ? ? ?
* equation ( 3.9 ) ) .
now , in order to relate b - microlocal analysis with semiclassical analysis , we first identify @xmath929 , @xmath930 , with @xmath931 . under the semiclassical rescaling ,
say by @xmath932 , one identifies the latter with @xmath163 , @xmath933 . in particular ,
if a conic set is disjoint from @xmath786 in @xmath934 , then its image under the semiclassical identification lies in a compact subset of @xmath786 .
thus , for @xmath935 dilation invariant , the large parameter principal symbol and wave front set of the mellin conjugate @xmath936 of @xmath645 are exactly those of @xmath645 under the above identification of @xmath929 , @xmath930 , with @xmath931 , and then the analogous statement also holds for @xmath937 considered as an element @xmath938 of @xmath810 under the semiclassical identification
. in particular , one has , for @xmath935 dilation invariant , with @xmath939 , that @xmath940 , with semiclassical wave front set in a compact subset of @xmath786 .
correspondingly , for any @xmath828 , @xmath941 now if @xmath942 , then @xmath615 is dilation invariant on @xmath613 , and its conjugate by the mellin transform is @xmath943 , whose rescaling @xmath944 is an element of @xmath793 .
further , with @xmath577 b - normally hyperbolic in the sense discussed above ( with the convention changed regarding formal self - adjointness , as stated before theorem [ thmnhypsemimic ] ) , @xmath771 is normally hyperbolic in the sense of dyatlov .
fix a smooth b - density on @xmath1 near @xmath25 , identified with @xmath945 as above ; we require this to be of the product form @xmath946 , @xmath795 a smooth density on @xmath25 ; we compute adjoints with respect to this density .
then for any @xmath947 , @xmath948 , see ( * ? ? ?
* section 3.3 ) for differential operators , and by a straightforward calculation using the mellin transform in general .
in particular , if @xmath949 , then @xmath950 for @xmath951 .
relaxing momentarily , we then assume that @xmath952 with @xmath171 the minimal normal expansion rate for the hamilton flow of the principal symbol of @xmath577 at @xmath953 , as above ; note that @xmath92 is elliptic on @xmath30 .
this gives that for @xmath951 , @xmath954 is order @xmath955 in the large parameter pseudodifferential algebra , so , defining @xmath956 , the semiclassical version gives @xmath957 with @xmath958 where @xmath780 is the image of @xmath30 under the semiclassical identification .
in particular , there is @xmath959 and @xmath960 such that if @xmath961 then @xmath962 with this background , under our assumptions on the dynamics , propagating estimates from the radial points towards @xmath26 , in particular through @xmath780 , and using the uniformity in parameters described above theorem [ thmnhypsemimic ] , we have : [ thmnhypsemiglobal ] let @xmath963 .
suppose @xmath964 satisfies at @xmath30 , @xmath771 is the semiclassical rescaling of @xmath943 , @xmath965 , @xmath966 , @xmath967 , @xmath959 as in .
then there is @xmath783 such that for @xmath851 , @xmath968 , @xmath969 with the implied constant and @xmath800 uniform in @xmath819 with @xmath970 .
this is immediate from piecing together the semiclassical propagation estimates from radial points ( which is where @xmath965 is used , see the corresponding statement in the b - setting given in ( * ? ? ?
* proposition 2.1 , footnote 20 ) ) through @xmath780 , using theorem [ thmnhypsemimic ] , which is where @xmath967 is used and where @xmath971 , rather than @xmath972 , is obtained for the right hand side , to @xmath973 , which is where @xmath966 is used . going back to the operator @xmath908 satisfying the conditions stated at the beginning of this section , and under the additional assumption of uniform normal hyperbolicity
as explained above , we can now obtain partial expansions of solutions to @xmath530 at infinity , i.e. at @xmath25 : [ thmexpansion ] ( cf .
* theorem 7.9 ) . )
let @xmath974 .
suppose @xmath439 has a simple rank @xmath67 resonance at @xmath18 with resonant state @xmath67 , and that all other resonances have imaginary part less than @xmath975 .
let @xmath911 , @xmath976 , , then the full condition on @xmath828 holds if we choose @xmath56 sufficiently small . ] and assume @xmath977 .
let @xmath978 .
then any solution @xmath979 of @xmath530 with @xmath980 satisfies @xmath981 with @xmath982 for @xmath145 and @xmath983 for @xmath100 , and the following tame estimate holds : @xmath984 the proof works in the same way as in the reference by an iterative argument that consists of rewriting @xmath530 as @xmath985 and employing a contour deformation argument , see ( * ? ? ?
* lemma 3.1 ) ( which uses high - energy estimates for the inverse normal operator family @xmath986 and the location of resonances , i.e. of the poles of this family ) , to improve on the decay of @xmath82 by @xmath987 in each step , but losing an order of differentiability as we are treating @xmath988 as an error term ; using tame microlocal regularity for the equation @xmath530 , corollary [ corwavepropsing ] , one can regain this loss .
we obtain @xmath989 after a finite number of iterations in case @xmath145 , , @xmath990 . ] and @xmath991 for all @xmath900 in case @xmath100 .
assuming we are in the latter case , the next step of the iteration gives a partial expansion @xmath992 with @xmath993 ( identified , as before , with @xmath994 , where @xmath995 is a smooth cutoff near the boundary ) and @xmath996 for any @xmath139 satisfying @xmath997 and @xmath998 ; here , we need @xmath999 so that the normally hyperbolic trapping estimate holds with @xmath1000 , with loss of two derivatives . if @xmath1001 , we can use this information to deduce @xmath1002 which implies that the expansion @xmath992 in fact holds with the membership @xmath1003 ; notice the improvement in the weight .
therefore , @xmath1004 , finishing the proof .
[ thmtamesolutionop ] under the assumptions of theorem [ thmexpansion ] with @xmath100 and @xmath1005 , define the space @xmath1006 then the operator @xmath1007 has a continuous inverse @xmath791 that satisfies the tame estimate @xmath1008 we continue to work in the setting of the previous section . with the tame forward solution operator constructed in theorem [ thmtamesolutionop ] in our hands , we are now in a position to use a nash - moser implicit function theorem to solve quasilinear wave equations .
we use the following simple form of nash - moser , given in @xcite : [ thmnashmoser ] let @xmath1009 and @xmath1010 be banach spaces for @xmath1011 with @xmath1012 and indeed @xmath1013 for @xmath1014 , likewise for @xmath1015 and @xmath1016 ; put @xmath1017 and similarly @xmath1018 .
assume there are smoothing operators @xmath1019 satisfying for every @xmath1020 , @xmath1021 and @xmath1022 : @xmath1023 let @xmath1024 be a @xmath1025 map , and assume that there exist @xmath1026 , @xmath1027 , @xmath1028 and constants @xmath1029 and @xmath1030 such that for any @xmath1031 , @xmath1032 moreover , assume that for every @xmath1033 with @xmath1034 there exists an operator @xmath1035 satisfying @xmath1036 and the tame estimate @xmath1037 for all @xmath1038 . then if @xmath1039 is sufficiently small depending on @xmath1040 and @xmath1041 , where @xmath1042 , there exists @xmath1033 such that @xmath1043 . to apply this in our setting
, we let @xmath1044 and @xmath1045 with the corresponding norms ; @xmath1046 will be the quasilinear equation , with implicit dependence on the forcing term .
we now construct the smoothing operators @xmath1047 ; we may assume , using a partition of unity , that @xmath21 is the closure of an open subset of @xmath215 , say @xmath1048 , where we let @xmath1049 . then there are bounded extension and restriction operators @xmath1050 for @xmath1011 ; the operator @xmath175 can be constructed such that @xmath1051 for @xmath1052 .
if we then define for @xmath1021 and @xmath1053 : @xmath1054 where @xmath1055 is a smoothing operator on @xmath215 with properties as in , then @xmath1056 satisfies in view of @xmath1057 being the identity on @xmath1058 if the norms on the left hand side are understood to be @xmath1059-norms .
however , note that @xmath1056 does not map @xmath1060 into itself , since smoothing operators such as @xmath1055 enlarge supports ; we will thus need to modify @xmath1056 below to obtain the operators @xmath1047 . in order to construct @xmath1055 on weighted b - sobolev spaces @xmath41
, it suffices by conjugation by the weight to construct it on the unweighted spaces @xmath308 ; then , by a logarithmic change of coordinates , we only need to construct the smoothing operator @xmath1061 on the standard sobolev spaces @xmath1062 , which we will do in lemma [ lemmasmoothing ] below . in order to deal with the issue of @xmath1056 enlarging supports
, we will define @xmath1061 such that @xmath1063 in particular , when one undoes the logarithmic change of coordinates , this implies @xmath1064 more generally , with @xmath1065 denoting dilations @xmath1066 on @xmath215 , we have @xmath1067 with the operator norm independent of @xmath1068 near @xmath67 .
now , in our application of theorem [ thmnashmoser ] , we will have @xmath1069 and correspondingly we will have forward solution operators @xmath332 going in the reverse direction , with all relevant constants being uniform in @xmath1070 . looking at the proof of theorem [ thmnashmoser ] in @xcite , one only uses the smoothing operator @xmath1071 with @xmath1072 in the @xmath176-th step of the iteration , with @xmath1073 chosen sufficiently large ; in our situation , where we have , we can therefore use the smoothing operator @xmath1074 in the @xmath176-th iteration step .
note that , for @xmath1073 large , we have @xmath1075 the solution @xmath82 to @xmath1043 , obtained as a limit of an iterative scheme ( see ( * ? ? ?
* lemma 1 ) ) , therefore is an element of @xmath1076 . taking the hyperbolic nature of the pde @xmath1043 into account once more
, it will then , in our concrete setting , be easy to conclude that in fact @xmath1077 .
choose @xmath1085 with @xmath1086 so that the fourier transform @xmath1087 is identically @xmath67 near @xmath18 ; put @xmath1088 and define the operator @xmath1089 .
then @xmath1090 with @xmath1091 , therefore holds for @xmath1092 in place of @xmath1061 with constants @xmath1093 since @xmath1087 decays super - polynomially , and holds for @xmath1092 in place of @xmath1061 with constants @xmath1093 since @xmath1094 vanishes at @xmath1095 with all derivatives .
next , let @xmath1096 be a smooth function depending only on @xmath1097 , i.e. @xmath1098 , so that @xmath1099 for @xmath1100 $ ] , @xmath1101 for @xmath1102 , and @xmath1103 .
put @xmath1104 , and define @xmath1105 condition is satisfied by the support assumption on @xmath332 .
let @xmath1106 and @xmath1107 .
to prove the other two conditions , we use the estimate @xmath1108 which we will establish below . taking this for granted
, we obtain for @xmath400 with @xmath1109 : @xmath1110 for @xmath1111 , which is the estimate ; and follows from @xmath1112 for @xmath1113 .
we now prove for @xmath1114 . for multiindices
@xmath1115 with @xmath1116 , we have for @xmath400 with @xmath1109 and for @xmath1117 , which in particular implies @xmath1118 : @xmath1119 thus @xmath1120 where @xmath1121 but @xmath1122 for all @xmath468 : indeed , this reduces to the statement that for a fixed @xmath1123 , one has @xmath1124 hence , we obtain , and the proof is complete .
we now combine theorem [ thmtamesolutionop ] , giving the existence of tame forward solution operators , with theorem [ thmnashmoser ] , in the extended form described above , to solve quasilinear wave equations .
we use the space @xmath1125 of real - valued elements of @xmath1126 .
[ thmquasilinearwave ] let @xmath1127 and @xmath1128 , @xmath1129 for @xmath1130 ; define the map @xmath1131 by @xmath1132 and assume that @xmath1133 satisfies the assumptions of section [ subsecforward ] and of theorem [ thmtamesolutionop ] .
moreover , define @xmath1134 where @xmath1135 then there exists @xmath1136 such that for all forcing terms @xmath1137 satisfying @xmath1138 , the equation @xmath1139 has a unique solution @xmath1140 .
if more generally @xmath1141 with @xmath1142 and @xmath1143 , then there exists @xmath1136 such that for all forcing terms @xmath1137 satisfying @xmath1144 , the equation @xmath1145 has a unique solution @xmath1140 .
we write @xmath1146 for the @xmath1126-norm and @xmath1083 for the @xmath41-norm .
, is @xmath1147 for varying @xmath1148 .
] we define the map @xmath1149 and check that it satisfies the conditions of theorem [ thmnashmoser ] with @xmath1150 . from the definition of @xmath154 and the tame
estimates for products , reciprocals and compositions , corollary [ corhbmodule ] and propositions [ prophbrec ] and [ propcompwithsmooth2 ] , we obtain @xmath1151 thus the first estimate of for @xmath1152 , @xmath1153 , @xmath1154 .
next , we have @xmath1155 , where the first order b - differential operator @xmath89 is of the form @xmath1156 with the second sum capturing one term of the linearization of terms @xmath1157 in @xmath60 ( i.e. terms for which @xmath1158 ) . in particular , @xmath1159 where @xmath1160 and @xmath1161 for @xmath1162 .
therefore , @xmath1163 which gives the second estimate of for @xmath1164 and @xmath1165 .
next , we observe that @xmath1166 is bilinear in @xmath1167 , involves up to two b - derivatives of each @xmath400 and @xmath1168 , and the coefficients depend on up to two b - derivatives of @xmath82 , thus @xmath1169 which gives the third estimate of for @xmath1170 , @xmath1165 . in summary , we obtain for integer @xmath1171 .
finally , we determine @xmath184 so that we have the tame estimate : given @xmath1172 , we can write @xmath1173 as in , with @xmath1160 and @xmath1174 ; hence , by theorem [ thmtamesolutionop ] , we obtain a solution operator @xmath1175 where @xmath1176 , provided @xmath1177 is small enough so that all dynamical and geometric hypotheses hold for @xmath1173 .
notice that the subprincipal term of @xmath1173 can differ from that of @xmath1133 by terms of the form @xmath1178 , @xmath311 , @xmath1179 , see ; however , since such terms eliminate constants , the simple rank @xmath67 resonance at @xmath18 with resonant state @xmath67 does not change ; and moreover such terms are _ small _ because of the factor @xmath71 , hence high energy estimates still hold in a ( possibly slightly smaller ) strip in the analytic continuation , see the remark below ( * ? ? ?
* theorem 1 ) .
since @xmath828 is independent of @xmath32 , we have for all @xmath1005 , in particular @xmath1180 .
now , implies that holds for @xmath1181 , @xmath1182 , so we need to control @xmath1183 derivatives of @xmath696 .
thus , we can apply nash - moser iteration , theorem [ thmnashmoser ] , to obtain a solution @xmath1162 of the pde , with the caveat that @xmath82 is a priori supported on a space slightly larger than @xmath21 .
however , local uniqueness for quasilinear hyperbolic equations , see e.g. @xcite , implies that @xmath82 in fact is supported in @xmath21 , and that @xmath82 is the unique solution of , finishing the proof of the first part .
[ rmkstrongerthandsresult ] in the asymptotically de sitter setting considered in @xcite , the above theorem extends ( * ? ? ?
* theorem 8.8 ) , at the cost of requiring the control of more derivatives , since we allow the dependence of the metric @xmath1186 on @xmath148 as well .
[ corquasilinearkds ] under the assumptions of theorem [ thmquasilinearwave ] , the quasilinear wave equation , resp .
, on a 4-dimensional asymptotically kerr - de sitter space with @xmath55 has a unique global smooth ( i.e. conormal , in the space @xmath1060 ) solution if the @xmath1187-norm , resp .
@xmath1188-norm , of the forcing term @xmath1189 is sufficiently small .
the only difference between wave and klein - gordon equations with mass @xmath140 ( which is to be distinguished from the black hole mass @xmath1 ) is that the resonance of the klein - gordon operator @xmath69 with largest imaginary part , which gives the leading order asymptotics , is no longer at @xmath18 for @xmath1190 .
thus , if we sort the resonances @xmath1191 of @xmath69 with multiplicity by decreasing imaginary part , assume @xmath1192 and moreover that the high energy estimates for the normal operator family of @xmath69 hold in @xmath133 , the only change in the statement of theorem [ thmexpansion ] for klein - gordon operators is that the conclusion now is @xmath1193 , where @xmath1194 , with @xmath1195 identified with @xmath1196 for a smooth cutoff @xmath995 near the boundary .
, we obtain @xmath1197 ; if @xmath145 , the statement of theorem [ thmexpansion ] is unchanged ; and if @xmath1198 and @xmath1199 are close enough together ( including the case that @xmath77 is a double resonance ) , one gets two terms in the expansion of @xmath82 . for brevity
, we only explain one scenario here .
see also the related discussion in @xcite .
] we thus obtain the following adapted version of theorem [ thmtamesolutionop ] : [ thmtamesolutionopkg ] in the notation of section [ subsecsolve ] , under the above assumptions and for @xmath1005 , define the space @xmath1200 then the operator @xmath1007 has a continuous inverse @xmath791 that satisfies the tame estimate @xmath1201 [ thmquasilinearkg ] under the above assumptions and the assumption @xmath1202 , let @xmath1127 and @xmath1128 , @xmath1129 for @xmath1130 ; define the map @xmath1203 by @xmath1132 and assume that @xmath1133 satisfies the assumptions of section [ subsecforward ] and of theorem [ thmtamesolutionopkg ] .
moreover , define @xmath1204 where @xmath1205 then there exists @xmath1136 such that for all forcing terms @xmath1137 satisfying @xmath1138 , the equation @xmath1206 has a unique solution @xmath1207 .
if more generally @xmath1141 with @xmath1142 and @xmath1143 , then there exists @xmath1136 such that for all forcing terms @xmath1137 satisfying @xmath1144 , the equation @xmath1208 has a unique solution @xmath1140 .
the proof proceeds as the proof of theorem [ thmquasilinearwave ] .
notice that we allow the nonlinear term @xmath60 to be more general , the point being that firstly , any at least quadratic expression in @xmath1209 with @xmath1210 gives an element of @xmath41 , and secondly , every element in @xmath1211 vanishes at the boundary , thus the normal operator family of the linearization of @xmath1212 at any @xmath1210 is equal to the normal operator family of @xmath78 , for which one has high energy estimates by assumption .
[ corquasilinearkgkds ] under the assumptions of theorem [ thmquasilinearkg ] and for @xmath5 and @xmath70 sufficiently small , the quasilinear klein - gordon equation , resp . , on a 4-dimensional asymptotically kerr - de sitter space with angular momentum @xmath5 has a unique global smooth ( i.e. conormal , in the space @xmath1213 ) solution if the @xmath1187-norm , resp .
@xmath1188-norm , of the forcing term @xmath1189 is sufficiently small .
finally , following the same arguments as used in the previous section , we indicate how to prove the general theorems stated in the introduction .
we continue to use , but need to generalize the setting considered in section [ subsecforward ] : namely , generalizing , we now allow @xmath89 to be any first order b - differential operator , and correspondingly need information on the skew - adjoint part of @xmath577 ; concretely , we define @xmath1214 at the ( generalized ) radial sets @xmath27 , using the same notation as in , by @xmath1215 moreover , at the trapped set @xmath1216 , we assume that @xmath1217 with @xmath171 the minimal normal expansion rate for the hamilton flow of the principal symbol of @xmath577 , and @xmath92 the mellin dual variable of @xmath9 after an identification of a collar neighborhood of @xmath25 in @xmath1 with @xmath1218 ; note that @xmath92 is elliptic on @xmath30 .
let @xmath1219 be the threshold weight for the first part of theorem [ thmtamehyptr ] , i.e. @xmath1220 with @xmath1221 as defined in .
then corollary [ corwavepropsing ] holds in the current , more general setting , provided we assume @xmath1222 and @xmath1223 .
likewise , we obtain the high energy estimates of theorem [ thmnhypsemimic ] under the assumption @xmath1224 . in order to generalize theorem [ thmexpansion ] , we first choose @xmath1225 such that @xmath1226 which holds for sufficiently small @xmath1227 in view of by the compactness of @xmath30 in @xmath106 .
we moreover assume that there are no ( nonzero ) resonances in @xmath1228 in the case of theorem [ thmintrogeneralkg ] ( theorem [ thmintrogeneralwave ] ) , and we assume further that @xmath1229 . then in the proof of theorem [ thmexpansion ] , ignoring the issue of threshold regularities at radial sets momentarily , we can use the contour shifting argument without loss of derivatives up to , but excluding , the weight @xmath1219 , corresponding to the contour of integration @xmath1230 . shifting the contour further down
, we can not use the non - smooth real principal type estimate at @xmath30 anymore and thus lose @xmath590 derivatives at each step ; the total number of additional steps needed to shift the contour down to @xmath1231 is easily seen to be at most @xmath1232 hence in order to have the final conclusion that @xmath82 has an expansion with remainder in @xmath41 , we need to assume that @xmath82 initially is known to have regularity @xmath1233 for any @xmath1234 , which in turn requires @xmath1235 and @xmath1236 for the first , lossless , part of the argument to work . taking the regularity requirements at the radial sets into account
, we further need to assume @xmath1237 . under these assumptions ,
the proof of theorem [ thmexpansion ] applies , mutatis mutandis , to our current situation , and we obtain a tame solution operator as in theorem [ thmtamesolutionop ] , which now loses @xmath1238 derivatives . thus , we can prove theorems [ thmintrogeneralkg ] and [ thmintrogeneralwave ] using the same arguments which we used in the proof of theorem [ thmquasilinearwave ] ; the ` loss of derivatives ' parameter @xmath184 now needs to satisfy the conditions @xmath1239 with the first condition being the actual loss of derivatives , the second one coming from @xmath1240 certainly being a high enough regularity for @xmath1241 to be @xmath1242 , which is required for the application of the non - smooth microlocal regularity results , and the last condition being the threshold regularity ( for the non - smooth estimates ) at the radial sets . | we extend the semilinear framework developed by the two authors in @xcite and the _ non - trapping _ quasilinear theory developed by the first author @xcite to solve quasilinear wave equations with normally hyperbolic trapping .
the most well - known example that fits into our general framework is wave - type equations on kerr - de sitter space .
the key advance is an adaptation of the nash - moser iteration to our framework . |
je thanks for hospitality during his stay at texas a&m university . this research is supported by the air force office of scientific research , darpa - quist , office of naval research , and the tamu telecommunication and informatics task force ( titf ) initiative .
k. s. johnson _
et al . _ ,
science * 280 * , 1583 ( 1998 ) ; v. westphal and s. w. hell , phys .
94 , 143903 ( 2005 ) .
et al . _ , ultramicroscopy * 13 * , 227 ( 1984 ) ; a. lewis _
et al . _ ,
nature biotechnology * 21 * , 1378 ( 2003 ) . s. w. hell , nature biotechnology , * 21 * , 1347 ( 2003 ) .
u. w. rathe and m. o. scully , lett .
phys . * 34 * , 297 ( 1995 ) ; a. n. boto _
et al . _ ,
lett . * 85 * , 2733 ( 2000 ) ; m. dangelo , m. v. chekhova and y. shih , _ ibid . _ * 87 * , 013602 ( 2001 ) .
m. o. scully and k. drhl , phys .
a * 25 * , 2208 ( 1982 ) ; m. o. scully and c. h. raymond ooi , j. opt .
b : quantum semiclass . opt .
* 6 * , s816 ( 2004 ) ; a. muthukrishnan , m. o. scully and m. s. zubairy , _ ibid .
_ s575 ( 2004 ) .
w. denk , j.h .
strickler and w.w .
webb , science * 248 * , 73 ( 1990 ) .
e. betzig , opt .
lett . * 20 * , 237 ( 1995 ) . c. hettich
_ et al . _ ,
science * 298 * , 385 ( 2002 ) .
r. h. dicke , phys . rev . * 93 * , 99 ( 1954 ) .
z. ficek and s. swain , _ quantum interference and coherence _
( springer , berlin , 2005 ) .
t. g. rudolph , z. ficek , and b. j. dalton , phys .
a * 52 * , 636 ( 1995 ) ; j. g. cordes and w. roberts , phys .
a * 29 * , 3437 ( 1984 ) .
a. m. herkommer , w. p. schleich , and m. s. zubairy , j. mod . opt . * 44 * , 2507 ( 1997 ) ; e. paspalakis and p. l. knight , phys .
a * 63 * , 065802 ( 2001 ) .
s. qamar , s .- y .
zhu , and m. s. zubairy , phys . rev .
a. * 61 * , 063806 ( 2000 ) ; s. qamar , s .- y .
zhu , and m. s. zubairy , opt .
commun . * 176 * , 409 ( 2000 ) ; f. ghafoor , s. qamar , and m. s. zubairy , phys .
a * 65 * , 043819 ( 2002 ) ; m. sahrai , h. tajalli , k. t. kapale , and m. s. zubairy , phys . rev .
a * 72 * , 013820 ( 2005 ) . | precision measurement of small separations between two atoms or molecules has been of interest since the early days of science . here , we discuss a scheme which yields spatial information on a system of two identical atoms placed in a standing wave laser field .
the information is extracted from the collective resonance fluorescence spectrum , relying entirely on far - field imaging techniques .
both the interatomic separation and the positions of the two particles can be measured with fractional - wavelength precision over a wide range of distances from about @xmath0 to @xmath1 .
the measurement of small distances is a fundamental problem since the early days of science .
it has become even more important due to recent interest in nanoscopic and mesoscopic phenomena @xcite .
starting from the invention of the optical microscope around 400 years ago , today s optical microscopy methodologies can basically be divided into lens - based and lensless imaging . in general ,
far - field imaging is lens - based and thus limited by criteria such as the rayleigh diffraction limit which states that the achievable resolution in the focus plane is limited to half of the wavelength of illuminating light .
further limitation arises from out - of - focus light , which affects the resolution in the direction perpendicular to the focal plane .
many methods have been suggested to break these limits @xcite .
lens - based techniques include confocal , non - linear femtosecond or stimulated emission depletion microscopy @xcite .
also non - classical features such as entanglement , quantum interferometry or multi - photon processes can be used to enhance resolution @xcite .
a particularly promising development is lensless near - field optics , which can achieve nanometer spatial resolution @xcite . roughly speaking
, the idea is to have light interactions close enough to the object to avoid diffraction .
this , however , typically restricts near - field optics to objects on a surface . in 1995 , betzig proposed a method to reach sub - wavelength resolution that is not limited to one spatial dimension by assuming non - identical , individually addressable objects @xcite .
subsequently , this was realized in a landmark experiment of hettich et al .
@xcite .
it combined near - field and far - field fluorescence spectroscopy techniques , using the fluorescence spectrum to label different molecules inside an inhomogeneous external electrical field .
they also noticed dipole - dipole interactions between adjacent objects @xcite and used it to correct the measurement result .
however , there is still great interest in achieving nanometer distance measurements by using optical illuminating far - field imaging only .
( color online ) two atoms in a standing wave field separated by a distance @xmath2 smaller than half of the wavelength @xmath3 of the driving field .
the distance of the two atoms is measured via the emitted resonance fluorescence.,width=302 ] in this communication , we propose a scheme to measure the distance between two adjacent two - level systems by driving them with a standing wave laser field and measuring the far field resonance fluorescence spectrum , which is motivated by the localization of single atom inside a standing wave field @xcite . in particular , we focus on distances smaller than the rayleigh limit @xmath1 .
our basic approach is that in a standing wave , the effective driving field strength depends on the position of the particles .
thus , each particle generates a sharp sideband peak in the spectrum , where the peak position directly relates to the subwavelength position of the particle .
as long as the two sideband peaks can be distinguished from each other , the position of each particle can be recovered . however , when the interatomic distance decreases , the two particles can no longer be considered independent . due to the increasing dipole - dipole interaction between the two particles ,
the fluorescence spectrum becomes complicated .
we find , however , that the dipole - dipole interaction energy can directly be extracted from the fluorescence spectrum by adjusting the parameters of the driving field
. since the dipole - dipole interaction energy is distance dependent , it yields the desired distance information .
we provide detailed measurement procedures and our estimates show that the scheme is applicable to inter - particle distances in a very wide range from @xmath1 to about @xmath4 .
our model system consists of two identical two - level atoms located at fixed points @xmath5 ( @xmath6 ) in a resonant standing wave laser field ( see fig .
[ fig - system ] ) .
the atomic transition frequency is @xmath7 .
the laser field has frequency @xmath8 , wavelength @xmath3 and wave vector @xmath9 .
we assume the two atoms to be arranged along @xmath10 .
the driving field rabi frequency of atom @xmath11 is @xmath12 , with @xmath13 .
we denote the raising ( lowering ) operator of the @xmath11th atom by @xmath14 ( @xmath15 ) . in the following
, we assume the transition dipole moments of the two atoms to be parallel and aligned perpendicular to the @xmath10 direction .
we also assume resonant driving , @xmath16 .
if the two atoms are far apart ( the distance between atom @xmath11 and @xmath17 @xmath18 ) , then they are independent , and the total master equation is given by the sum of the two single - particle master equations @xcite . if the two atoms come close , they dipole - dipole interact , causing a collective system dynamics .
this gives rise to a complex energy shift due to a virtual photon exchange between the two atoms .
the imaginary part of this shift corresponds to an incoherent coupling , whereas the real part shifts the energy of the collective states of the system .
the full collective master equation is given by @xcite @xmath19 - \sum_{i , j=1}^{2 } \gamma_{ij } \left ( [ s_i^+,s_j^-\rho ] - [ s_j^-,\rho s_i^+ ] \right ) \ , . \label{master}\end{aligned}\ ] ] the coherent evolution is governed by @xmath20 .
the free energy @xmath21 of the two atoms and the interaction with the driving laser field @xmath22 are the same as for two independent atoms .
the coherent energy shift of the collective states arises from the dipole - dipole interaction @xmath23 , which involves couplings of both atoms . for the considered geometry ,
the dipole - dipole interaction @xmath24 is given by @xmath25 the incoherent evolution first entails the independent spontaneous emission of the two atoms @xmath26 ( @xmath6 ) , as found for uncoupled atoms .
terms with @xmath27 ( @xmath28 ) are the incoherent dipole - dipole couplings , where @xmath29 for large distances , ( @xmath30 ) , we find @xmath31 and @xmath32 , where @xmath33 is the kronecker delta symbol . thus we recover the case of two independent atoms , as expected . for small distances ( @xmath34 ) ,
one finds maximum incoherent cross - coupling , and @xmath24 approaches the static dipole - dipole interaction , @xmath35 % [ 1 - 3\cos^2\alpha ] \
, , \qquad \gamma_{ij}\approx \gamma\ , .
\label{param - approx}\end{aligned}\ ] ] sample spectra for @xmath36 , @xmath37 , @xmath38 .
fixed distance @xmath39 ( solid lines ) and harmonic oscillation around @xmath39 ( dashed lines ) .
( a ) large separation case : @xmath40 . ( b ) intermediate separation , weak driving field : @xmath41 ( c ) as ( b ) , but strong driving field : @xmath42 ( d ) small separation : @xmath43.,width=302 ] our strategy is to identify the distance of the two atoms via the emitted resonance fluorescence .
we define @xmath44 as the unit vector in observation direction , and the observation angle @xmath45 as @xmath46 .
the total two - atom steady state resonance fluorescence spectrum @xmath47 up to a geometrical factor is given by @xcite @xmath48 where the subindex @xmath49 denotes the steady state .
in general , this resonance fluorescence spectrum is rather complicated @xcite .
the spectrum , however , simplifies considerably in limiting cases , where either the driving field rabi frequency or the dipole - dipole interaction dominates the dynamics .
this will be exploited in the following , where we present in detail a measurement procedure , which allows us to extract the distance between the two atoms and their positions relative to nodes of the standing wave field , both with fractional - wavelength precision .
the first step in the measurement sequence is to apply a standing wave laser field to the two atoms , which at an anti - node of the standing wave corresponds to a rabi frequency @xmath50 of a few @xmath51 .
depending on the relative separation of the atoms , different spectra can be observed . if the two atoms are well - separated ( about @xmath52 ) , then the dipole - dipole interaction is negligible . in this case
spectra as shown in fig .
[ sample - spectra](a ) are obtained .
the two sideband structures can be interpreted as arising from the ac - stark splitting due to @xmath53 and @xmath54 .
thus the sideband peak positions @xmath55 and @xmath56 can directly be related to @xmath53 and @xmath54 and therefore to the position of the two atoms relative to the standing wave field nodes .
consequently , we can obtain the distance @xmath39 .
within half a wavelength , however , in general two interatomic distances are possible for measured values of @xmath53 and @xmath54 ( see fig .
[ intermediate - peakpos](a , b ) ) @xcite .
an identification of the actual atomic positions is possible by changing the standing wave phase slightly , i.e. , shifting the positions of the ( anti- ) nodes . as shown in fig .
[ intermediate - peakpos](a , b ) , a combination of the possible positions for two different standing wave phases yields the actual separation . note that this complication is not present for nearby atoms , where the non - vanishing dipole - dipole energy allows to determine the distance directly . in fig .
[ sample - spectra](a ) , the distance of the two particles is @xmath57 .
from the spectrum accessible in experiments , the distance @xmath58 is obtained , if we allow for a total measurement uncertainty of about @xmath59 .
thus the actual and the measured distances match , and the @xmath59 uncertainty of the distance measurement corresponds to about @xmath60 .
( color online ) ( a , b ) obtaining the position of the two atoms via a phase shift of the standing wave field .
solid ( dashed ) lines show possible atom positions for given @xmath53 ( @xmath54 ) .
( a ) before , ( b ) after the phase shift . the only coinciding potential positions in ( a ) and ( b )
give the true atomic positions .
( c ) deviation @xmath61 of the doublet splitting @xmath62 from @xmath63 for the strong field , intermediate distance case .
@xmath64 , @xmath37 , and @xmath36 .
the positions of the atoms are @xmath38 ( solid ) , @xmath65 ( dashed ) , @xmath66 ( dotted ) .
, width=302 ] if the distance between the two atoms is intermediate ( about @xmath67 ) , then the initial weak field measurement in general yields a more complicated spectrum , see fig .
[ sample - spectra](b ) .
the reason is that then the dipole - dipole coupling and the driving field strength are comparable , and the two atoms are not independent .
in such a case , a quantitative interpretation of the spectrum is difficult .
however , increasing the rabi frequency @xmath50 leads to a spectrum as shown in fig .
[ sample - spectra](c ) .
the spectrum consists of a central peak , two inner sideband doublets , and two outer sideband doublets , each symmetrically placed around the driving field frequency @xmath8 .
the center positions of the inner and outer sideband doublets corresponds to the rabi frequencies @xmath53 and @xmath54 .
the sideband structures are split into doublets due to the dipole - dipole coupling of the two atoms . for large @xmath50 ,
the splitting approaches twice the energy @xmath24 , as shown in fig .
[ intermediate - peakpos](c ) .
thus the strong - field sideband doublet splitting directly yields @xmath24 and then the distance of the two atoms , via eqs .
( [ omega12]).for example , in fig . [ sample - spectra](c )
, the actual distance is @xmath68 . from the spectrum ,
a measurement would obtain @xmath69 , where again we have allowed for an uncertainty of about @xmath59 . from eq .
( [ omega12 ] ) , this yields a measured distance of @xmath70 , in good agreement with the actual value . on the other hand ,
comparing the center center positions of the inner and outer sideband doublets with @xmath50 , the positions of the individual atoms relative to standing wave field nodes can be obtained . for the setup in fig .
[ sample - spectra](c ) , we have @xmath71 , @xmath72 , @xmath73 . from the spectrum , using the above procedure , we obtain @xmath74 , @xmath75 , assuming a relative uncertainty of @xmath59 . from @xmath76 , this would yield a measurement result of @xmath77 , in good agreement with the actual position of the atoms . in the above two regimes
, the situation slightly complicates if both atoms are located near - symmetrically around a node or an anti - node . in this case , @xmath78 , such that the two sideband peaks ( or doublets ) overlap .
one way to resolve this is to adequately change the standing wave field phase . by this
, the symmetry can be lifted to give @xmath79 .
then the above procedure can be applied to yield the separation and positions .
if the two atoms are very close to each other ( distance @xmath80 ) , then the spectrum is dominated by the dipole - dipole interaction energy @xmath24 , which gives rise to sideband structure at each side of the fluorescence spectrum close to @xmath81 , and only weakly depends on the driving field . a typical spectrum for this parameter range is shown in fig .
[ sample - spectra](d ) .
as long as @xmath53 , @xmath54 , @xmath82 is satisfied , the sideband structures only have a small residual dependence on the rabi frequency .
thus , the sideband peak position @xmath83 can directly be identified with @xmath24 .
fig .
[ peakpos - small - distance](a ) shows the deviation of the sideband peak positions from @xmath24 versus the atomic separation distance for different rabi frequencies @xmath50 .
note that the effective rabi frequencies @xmath53 , @xmath54 also depend on the position of the first atom within the wavelength , with maximum values @xmath84 close to the anti - nodes . it can be seen that for weak @xmath85 , the experimentally accessible sideband peak position and @xmath24 coincide very well . with increasing rabi frequency , the deviation increases , until the driving field induces a splitting of the sideband peaks , indicated by the branching point in fig .
[ peakpos - small - distance](a ) . if the initial spectrum of the first measurement has insufficient signal - to - noise ratio , then the fluorescence intensity can be enhanced by increasing the driving field intensity .
note that due to the dependence of @xmath53 , @xmath54 on the position of the two atoms , different positions of the two atoms may require different laser field intensities .
it is also possible to extrapolate the result of several measurements to the driving field - free limit to increase the measurement accuracy .
deviation @xmath86 of the peak position @xmath83 from @xmath24 for closely - spaced atoms .
@xmath36 , @xmath37 , and ( a ) against the atomic separation .
@xmath38 , @xmath87 ( solid ) , @xmath88 ( dashed ) , @xmath89 ( dotted ) .
( b ) against the driving field rabi frequency .
@xmath90 , @xmath91 ( solid ) , @xmath92 ( dashed ) , @xmath93 ( dotted ) .
branches indicate splittings into two peaks.,width=302 ] via eqs .
( [ omega12 ] ) or ( [ param - approx ] ) , the measured @xmath24 can easily be used to obtain the interatomic separation .
the separation is measured with increasing accuracy in the region of large slope of @xmath24 vs @xmath39 . for maximum accuracy ,
eq .
( [ omega12 ] ) should be numerically solved for the separation . here
, we discuss the small separation limit eq .
( [ param - approx ] ) , and allow for a small uncertainty in @xmath24 ( @xmath94 ) .
we obtain @xmath95^{1/3 } \cdot[1 - \delta \omega_{12}/(3\omega_{12})]$ ] as the distance @xmath96 between the two atoms .
thus , the relative uncertainty of the final result is about 1/3 of the relative uncertainty of the measured @xmath24 .
consider , for example , the case shown in fig .
[ sample - spectra](d ) .
the actual distance is @xmath97 .
the measured dipole - dipole energy is @xmath98 , again with a relative measurement uncertainty of about @xmath59 . from eq .
( [ omega12 ] ) , the distance then evaluates to @xmath99 .
thus in this case , the uncertainty of the distance measurement would be about @xmath100 , i.e. , less than @xmath101 of the actual distance .
once the distance @xmath39 is known , the positions of the two atoms relative to nodes of the standing wave field can be obtained . for this
, we note from fig .
[ peakpos - small - distance](b ) that for otherwise fixed parameters the position of the branching point depends on the rabi frequencies @xmath53 , and thus on the position @xmath102 .
if in the experiment we increase @xmath50 up to the branching point , then the position of the atom pair relative to the field nodes can be deduced .
accurate analytic expressions for the position of the branching point , however , are involved , as the general expression of the fluorescence spectrum is complicated @xcite .
thus a numerical fit as shown in fig .
[ peakpos - small - distance](b ) should be used to evaluate @xmath102 .
finally , since for small distances the spectrum is almost independent to the driving field , the distance information can also be obtained using a travelling - wave field , which may be more convenient in practice .
the precise positioning of the atoms is limited by thermal or quantum position uncertainties @xcite .
we have simulated this effect by assuming a motional ground state harmonic oscillation with amplitude @xmath103 ( corresponding to a lamb - dicke parameter @xmath104 ) around the mean distance @xmath39 .
results averaged over this motion are shown with dashed lines in fig .
[ sample - spectra ] . for @xmath105 ,
the motion is negligible . with increasing ratio @xmath106 ,
the spectral peaks split up . in fig .
[ sample - spectra](d ) , two peaks emerge at the classical turning points of the distance oscillation . from these , the mean distance can again be obtained .
the possible separation measurement range is limited , as the dipole - dipole coupling @xmath24 increases with decreasing separation as @xmath107 . for our model to remain valid , however ,
@xmath108 should be fulfilled . from eq .
( [ param - approx ] ) , for @xmath109 hz , @xmath110 hz , we estimate @xmath111 as the theoretical resolution limit .
this limitation only applies to the distance of the two atoms itself ; the distance uncertainty in principle can be well below @xmath112 . note that these considerations neglect experimental uncertainties , and are subject to imperfections e.g. in the measurement of laser field parameters or the alignment of dipole moments or laser fields . in summary , we have discussed a microscopy scheme entirely based on optical far - field techniques .
it allows to measure the separation between and the position of two nearby atoms in a standing wave laser field with fractional - wavelength precision over the full range of distances from about @xmath0 up to the rayleigh limit @xmath1 . |
the inclusive @xmath0 cross section at @xmath1 and @xmath2 has been measured recently by both the atlas @xcite and cms collaborations @xcite . all measurements show a systematic tension when compared to next - to - leading order ( nlo ) qcd calculations .
the disagreement was first observed in the 7 tev data , and then it increased at 8 tev , where the tension with the nlo predictions reaches the @xmath3 level .
the small uncertainties quoted for the nlo calculations suggest that higher - order qcd corrections can not change this pattern .
this triggered a lot of interest and a number of models were suggested to explain this excess in terms of new light states , see e.g. refs.@xcite . however , before discussing any hint for new physics , one needs to fully control the uncertainty associated with the standard model prediction .
one issue that arises , is that refs .
@xcite quote the inclusive cross section , obtained by extrapolating the measured fiducial cross section through data driven monte carlo ( mc ) acceptances .
one reason for quoting the inclusive cross - section is that it is independent of the experimental setup , hence it is possible to make statements about whether two measurements by atlas and cms are in agreement , while fiducial cross - sections are different for the two experiments .
nevertheless , the extrapolation from the fiducial to the inclusive phase space relies on a monte carlo simulation , and thus obviously it depends on the generator used .
for instance , a generator that systematically underestimates the fiducial cross - section would lead to an overestimation of the resulting `` measured '' inclusive cross section .
hence , a comparison to theory should be made first in the fiducial region , for which experimental data is available , before extrapolating the result to the fully inclusive phase space .
a possible way to compare the measurements of both experiments avoiding a big extrapolation to the total inclusive phase space would be to extrapolate the atlas fiducial measurement from the cms one , and vice versa .
this extrapolation would still depend on the monte carlo generator used , but it would involve two cross sections that are typically of the same order of magnitude . in this note
we compare the measured fiducial cross sections to the nlo predictions and find that there is no sizable tension between the two ( at the 1-@xmath4 level ) .
we then study the effect of the extrapolation from the fiducial region to the inclusive one .
we find that the monte carlo acceptance computed with powheg overestimates the reduction due to the fiducial phase space cuts , leading to a larger total cross section when the extrapolation from the fiducial to the inclusive phase space is carried out .
we study the source of the reduction in the monte carlo prediction , and discuss the possible impact of higher - order corrections . at the moment , only atlas published the measured fiducial cross sections at 7 and 8 tev . since the larger disagreement is observed in the 8 tev data , we focus on the latter measurement for the three leptonic channels , @xmath5 , @xmath6 and @xmath7 .
we present here theory predictions for the fiducial cross section as defined by the atlas experiment at a centre - of - mass energy of 8 tev @xcite .
the relative fiducial cuts are summarized in table [ tab : fiducials ] .
our analysis equally applies to the 7 tev case , for which we find similar conclusions .
+ @xmath8 gev for the leading ( subleading ) lepton and charged leptons separated by @xmath9 + + muon pseudorapidity @xmath10 and electron pseudorapidity @xmath11 or @xmath12 + + no jets ( anti-@xmath13 @xcite , @xmath14 ) with @xmath15 gev and @xmath16 , separated from the electron by @xmath17 + + @xmath18 gev and @xmath19 gev for @xmath20 , @xmath21 , and @xmath22 , respectively + + @xmath23 gev and @xmath24 gev for @xmath20 , @xmath21 , and @xmath22 , respectively + + it is instructive to first compare the fiducial measurements to the next - to - leading order results from mcfm 6.6 @xcite , including the formally next - to - next - to - leading order ( nnlo ) contribution due to @xmath25 @xcite .
we work in the @xmath26 scheme with the electroweak parameters @xmath27gev , @xmath28gev , @xmath29gev , @xmath30gev , and @xmath31gev@xmath32 .
the branching ratio of @xmath33 into leptons is taken from the particle data book @xcite , namely @xmath34.-boson width @xmath35 within its error , as quoted by the pdg @xmath36 leads to a @xmath37 8.5% variation in the total cross section , since it scales as @xmath38 . however , the standard model calculation of the @xmath33 width has a much smaller uncertainty .
] we use the mstw2008nlo parton densities @xcite and central renormalization and factorization scales @xmath39 , where @xmath40 is the invariant mass of the leptonic system .
the scale uncertainty is obtained by varying both scales by a factor of two in either direction and keeping @xmath41 . with this setup
we get a total inclusive @xmath0 cross section of @xmath42 pb ( @xmath43 pb ) excluding ( including ) higgs mediation . for consistency , the quoted higgs mediated cross section has also been computed with mcfm . on the other hand , atlas finds @xmath44 pb . the measured fiducial cross sections at 8 tev are reported in the first column of table [ tab : fidatlas8vsnlo - mswt - m(ww ) ] , together with the total nlo predictions in the fourth column .
the latter is given by the sum of the cross - section without higgs mediation and the higgs mediated contribution , as reported in the second and third column , respectively .
the quoted theory uncertainty does not account either for the pdf error , that was found to be at @xmath45 level @xcite , or for any interference effect .
.atlas fiducial cross - sections in fb at 8 tev ( 1st column ) and two processes that contribute to @xmath0 production ( 2nd and 3rd column ) .
the last column contains the sum of the previous two contributions .
the theory predictions are obtained with mcfm with central renormalisation and factorisation scales set to the leptonic system invariant mass , and using the mstw2008nlo pdf set .
the ( formally nnlo ) @xmath46 channel is included .
following ref .
@xcite , all quoted numbers do not include electrons or muons coming from @xmath47 decays . [
tab : fidatlas8vsnlo - mswt - m(ww ) ] [ cols="^,^,^,^,^",options="header " , ]
despite the substantial tension with the nlo prediction reported by atlas for the inclusive @xmath0 cross section at 8 tev , our findings show that the measured fiducial cross sections are in good agreement ( at about the 1-@xmath4 level ) with the nlo result for all leptonic decay modes . the discrepancy observed by atlas at 8 tev
can be explained to a large extent by studying the monte carlo predictions used in the extrapolation to the inclusive cross section .
the prediction obtained with the powheg box reduces the nlo result by 9 - 11% . the monte carlo result for the jet - veto efficiency
is found to overestimate sudakov suppression effects with respect to the analytic resummation .
furthermore , at the les houches level , the leptons transverse momentum spectrum is slightly different than the nlo distribution .
this is a higher - order effect , and it leads to a systematic reduction in the fiducial cross section ( of about 3% in the present case ) at the parton - shower level with respect to nlo . as a result , the fiducial cross - section is reduced and the extrapolated ( " measured ) inclusive cross section tends to be overestimated .
the latter effect should be taken into account as a potential additional systematic uncertainty associated with the matching procedure of higher - order calculations to a parton shower . using the nnll+nnlo predictions for drell - yan and higgs production , we provide an estimate for the fiducial cross section , including these higher - order effects .
it suggests that the agreement found at nlo for the fiducial cross sections will remain good , or even improve once nnll+nnlo effects are included in the theory predictions .
a better assessment of higher order effects and related uncertainties is today possible by matching the recently computed nnlo corrections @xcite to the existing nnll resummation @xcite .
it would be also useful to encode the nnlo prediction in a fully exclusive nnlo + parton shower generator , on the line of what has been recently done for higgs- @xcite and drell - yan production @xcite .
region with their nnlo parton shower approach . ]
further studies are also required for the formally nnlo @xmath48-initiated channel .
at lowest order it amounts to 3 - 4% of the nlo cross section at the inclusive level ( i.e. of the same order of the higgs - mediated reaction ) , but it raises to 8 - 9% after fiducial cuts are taken into account since the jet veto is ineffective when applied to the tree level prediction . due to the large amount of initial state radiation ,
this channel will experience a sizable sudakov suppression , and has a potentially large k factor .
we estimate the nnll+nnlo effects in the jet - veto efficiency , but a more detailed study is necessary .
we would like to thank and p. nason and g. salam for valuable discussions , and a. banfi and k. hamilton for useful comments on the manuscript .
gz is supported by the erc grant 614577 .
pm is supported by the research executive agency ( rea ) of the european union under the grant agreement number pitn - ga-2010 - 264564 ( lhcphenonet ) , and by the swiss national science foundation ( snf ) under the grant pbzhp2 - 147297 .
this research was partly supported by the munich institute for astro- and particle physics ( miapp ) of the dfg cluster of excellence `` origin and structure of the universe '' ( pm and gz ) and by the mainz institute for theoretical physics ( mitp ) of the prisma excellence cluster ( gz ) .
we would like to thank the galileo galilei institute ( pm and gz ) and cern ( pm ) for hospitality while part of this work was carried out .
99 g. aad _ et al . _ [ atlas collaboration ] , phys . rev .
d * 87 * ( 2013 ) 11 , 112001 [ erratum - ibid .
d * 88 * ( 2013 ) 7 , 079906 ] [ arxiv:1210.2979 [ hep - ex ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 699 * ( 2011 ) 25 [ arxiv:1102.5429 [ hep - ex ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 721 * ( 2013 ) 190 [ arxiv:1301.4698 [ hep - ex ] ] .
b. feigl , h. rzehak and d. zeppenfeld , phys .
b * 717 * , 390 ( 2012 ) [ arxiv:1205.3468 [ hep - ph ] ] .
d. curtin , p. jaiswal and p. meade , phys .
d * 87 * , no . 3 , 031701 ( 2013 ) [ arxiv:1206.6888 [ hep - ph ] ] .
p. jaiswal , k. kopp and t. okui , phys .
d * 87 * , no .
11 , 115017 ( 2013 ) [ arxiv:1303.1181 [ hep - ph ] ] .
k. rolbiecki and k. sakurai , jhep * 1309 * ( 2013 ) 004 [ arxiv:1303.5696 [ hep - ph ] ] .
d. curtin , p. meade and p. j. tien , arxiv:1406.0848 [ hep - ph ] .
m. cacciari , g. p. salam and g. soyez , jhep * 0804 * ( 2008 ) 063 [ arxiv:0802.1189 [ hep - ph ] ] .
j. m. campbell and r. k. ellis , phys .
d * 60 * ( 1999 ) 113006 [ hep - ph/9905386 ] .
j. m. campbell , r. k. ellis and c. williams , jhep * 1110 * ( 2011 ) 005 [ arxiv:1107.5569 [ hep - ph ] ] .
j. beringer _
[ particle data group collaboration ] , phys .
d * 86 * ( 2012 ) 010001 .
a. d. martin , w. j. stirling , r. s. thorne and g. watt , eur .
j. c * 63 * ( 2009 ) 189 [ arxiv:0901.0002 [ hep - ph ] ] .
i. w. stewart and f. j. tackmann , phys . rev .
d * 85 * ( 2012 ) 034011 [ arxiv:1107.2117 [ hep - ph ] ] .
a. banfi , p. f. monni , g. p.
salam and g. zanderighi , phys .
* 109 * ( 2012 ) 202001 [ arxiv:1206.4998 [ hep - ph ] ] .
t. gehrmann , m. grazzini , s. kallweit , p. maierhfer , a. von manteuffel , s. pozzorini , d. rathlev and l. tancredi , arxiv:1408.5243 [ hep - ph ] .
p. nason , jhep * 0411 * ( 2004 ) 040 [ hep - ph/0409146 ] .
s. frixione , p. nason and c. oleari , jhep * 0711 * ( 2007 ) 070 [ arxiv:0709.2092 [ hep - ph ] ] .
s. alioli , p. nason , c. oleari and e. re , jhep * 1006 * ( 2010 ) 043 [ arxiv:1002.2581 [ hep - ph ] ] . t. melia , p. nason , r. rontsch and g. zanderighi , jhep * 1111 * ( 2011 ) 078 [ arxiv:1107.5051 [ hep - ph ] ] .
p. nason and g. zanderighi , eur .
j. c * 74 * ( 2014 ) 2702 [ arxiv:1311.1365 [ hep - ph ] ] .
t. sjostrand , s. mrenna and p. z. skands , jhep * 0605 * ( 2006 ) 026 [ hep - ph/0603175 ] .
z. skands , arxiv:0905.3418 [ hep - ph ] .
p. jaiswal and t. okui , arxiv:1407.4537 [ hep - ph ] .
p. nason and b. webber , ann .
nucl . part .
* 62 * ( 2012 ) 187 [ arxiv:1202.1251 [ hep - ph ] ] .
r. boughezal , x. liu , f. petriello , f. j. tackmann and j. r. walsh , phys .
d * 89 * , 074044 ( 2014 ) [ arxiv:1312.4535 [ hep - ph ] ] . | in this note we analyse the excess in the @xmath0 _ inclusive _ cross section recently measured at the lhc .
we point out that in fact for the atlas _ fiducial _ cross sections there is no excess in the measurements compared to the nlo qcd predictions .
we also argue that higher order effects to the fiducial cross section are small , and tend to cancel each other , hence the inclusion of nnlo and nnll corrections will not modify this agreement significantly .
we find that at 8 tev a substantial part of the disagreement with the nlo prediction for the total cross section observed by atlas is due to the extrapolation carried out with powheg . |
molecular outflows are ubiquitous in star - forming regions , with every protostar thought to undergo a period of outflow activity ( e.g. @xcite ) .
outflow observations provide unique information about the mass - loss history of the parent protostar and thereby the accretion process which determines its evolution .
recently there has been renewed interest in the suggestion that outflows may drive turbulence in molecular clouds @xcite , given that the total energy they contain may be a significant portion of a molecular cloud s turbulent energy and they can cause much disruption of their natal environment ( e.g. @xcite ; @xcite ; @xcite ) .
theoreticians are divided about whether an efficient mechanism exists to feed energy from a small quantity of fast - moving material in an outflowing jet into random isotropic motions on molecular cloud scales ( e.g. @xcite versus @xcite ) .
however , analyses of data from regions of molecular clouds with high densities of outflows suggest that outflows are not the dominant driving mechanism ( @xcite ; @xcite ) . if they were
, then cloud lifetimes should be short ( consistent with the rapid formation scenario proposed by @xcite ) or the star formation rate very high ( see @xcite ) . in addition , the detection of outflows from candidate star - forming cores , located through say ( sub)millimetre continuum surveys ( e.g. @xcite ) , offers a way to classify such sources as either starless or protostellar ( e.g. @xcite ; hereafter ) , which complements methods based on their spectral energy distributions ( seds , e.g. @xcite ) .
molecular outflows have diverse structures , reflecting both their driving sources and environments
. such diversity has lead to a wide range of plausible models ( see @xcite for a summary ) with no single model accounting for all the features of outflows observed to date .
young , collimated outflows point to jet - driven models ( e.g. @xcite ) , yet older , wider - angle outflows evoke wide - angle wind models ( e.g. @xcite ) .
some observational studies even suggest a two - component wind is necessary ( e.g. @xcite ) . in this paper we present a survey for molecular outflows towards the perseus molecular cloud ( hereafter simply perseus ) using our @xmath4co @xmath1 datasets ( @xcite ; hereafter ) and data presented here for the first time towards ngc1333 , the most active regions of star formation in perseus .
this outflow survey towards a single cloud is one of the largest to date , allowing us to look for statistical variations in outflow properties .
this work is the second in a series of papers surveying the kinematics of molecular gas in perseus .
presented the data acquisition and large - scale kinematic properties , whilst @xcite explores the detailed gas motions inside star - forming cores .
this study is arranged as follows : section [ sec : observations ] describes the observations we analyse , in particular the new data towards ngc1333 , and presents an overview of the survey before listing the naming conventions we apply to star - forming cores in our fields . in section [ sec : outflow_detection_rates ] , we look for an outflow detection towards the position of every submillimetre core identified with scuba @xcite and analyse the corresponding detection rates and core classifications .
we set out the methods we have chosen to analyse the detected outflows in section [ sec : outflowparams_method ] ( the outflows themselves are described in appendix [ appendix : outflows ] ) before exploring the evolution of these parameters in section [ sec : outflowevolution ] .
finally , we summarize our results in section [ sec : summary ] .
the @xmath4co @xmath1 data we analyse are towards the four largest clusters of star - forming cores in perseus ( see table [ table : regiondetails ] and fig . [
fig : overview ] for an overview of the data ) .
the datacubes of ic348/hh211 , l1448 and l1455 have been published previously where we refer the reader for details of the reduction .
briefly , the data were taken as ` basket - weaved ' scan maps using harp ( heterodyne array receiver project ; @xcite ) on the james clerk maxwell telescope ( jcmt ) over nine nights between december 2007 and january 2008 .
the data were taken in good to excellent observing conditions with a median system temperature of @xmath12k .
the final data products are sampled on a 3arcsec grid and are smoothed slightly spatially and spectrally to an equivalent full - width half maximum ( fwhm ) beam size of 16.8arcsec ( 0.020pc at 250pc , our assumed distance to perseus ) and velocity channel width of 1kms@xmath8 respectively .
smoothing to 1kms@xmath8 resolution provides additional sensitivity in the @xmath4co linewings .
lccccccccc no .
& driving @xmath13 & class @xmath14 & @xmath15 @xmath16 & lobes ?
@xmath17 & axis pa & & + & source & & ( kms@xmath8 ) & & e of n & r & b & r & b + + 1 & hrf41 & 0 & 7.11 & 2 & 37 & 260 & 145 & 27 & 26 + 2 & hrf42 & 0 & 7.52 & 2 & 56 & 169 & 33 & 8 & 14 + 3 & hrf43 & i & 8.24 & 2 & 143 & 112 & 115 & 31 & 32 + 4a@xmath18 & hrf44 & 0 & 7.62 & 2 & 111 & 103 & 131 & 22 & 28 + 4b@xmath18 & hrf44 & 0 & 7.62 & 2 & 18 & 120 & 460 & 32 & 20 + ... + @xmath13 driving source from @xcite .
+ @xmath14 source classification @xcite : 0=class 0 and i = class i. + @xmath16 source c@xmath3o @xmath1 velocity where available or the map median where not .
+ @xmath17 number of lobes : r for red- and b for blue - shifted only .
+ @xmath19 maximum co lobe length .
+ @xmath20 maximum co velocity ( relative to @xmath15 ) .
+ @xmath18 perpendicular outflows from the iras2a protobinary : a approximately east - west and b north - south . + lcccccc driving & @xmath21 & @xmath22 @xmath13 & @xmath23 @xmath14 & @xmath24 @xmath16 & @xmath25 @xmath17 & @xmath26 @xmath19 + source & ( @xmath27 m@xmath7 ) & ( @xmath28m@xmath7 km s@xmath8 ) & ( 10@xmath29 j ) & ( 10@xmath30 yr ) & ( 10@xmath31
m@xmath7 km s@xmath8 yr@xmath8 ) & ( 10@xmath32 l@xmath7 ) + + hrf41 & 7.05 & 5.64 & 12.05 & 0.93 & 6.07 & 10.57 + hrf42 & 1.13 & 0.58 & 0.73 & 1.11 & 0.53 & 0.53 + hrf43 & 36.30 & 30.79 & 69.32 & 0.44 & 70.29 & 128.91 + hrf44a & 9.37 & 5.65 & 9.85 & 0.57 & 9.93 & 14.11 + hrf44b & 42.16 & 26.05 & 41.00 & 1.36 & 19.22 & 24.64 + ... + | we present a census of molecular outflows across four active regions of star formation in the perseus molecular cloud ( ngc1333 , ic348/hh211 , l1448 and l1455 ) , totalling an area of over 1000arcmin@xmath0 .
this is one of the largest surveys of outflow evolution in a single molecular cloud published to date .
we analyse large - scale , sensitive co @xmath1 datasets from the james clerk maxwell telescope , including new data towards ngc1333 . where possible we make use of our complementary @xmath2co and c@xmath3o data to correct for the @xmath4co optical depth and measure ambient cloud properties . of the 65 submillimetre cores in our fields ,
we detect outflows towards 45 .
24 of these are marginal detections where the outflow s shape is unclear or could be confused with the other outflows .
we compare various parameters between the outflows from class 0 and i protostars , including their mass , momentum , energy and momentum flux .
class 0 outflows are longer , faster , more massive and have more energy than class i outflows .
the dynamical time - scales we derive from these outflows are uncorrelated to the age of the outflow driving source , computed from the protostar s bolometric temperature .
we confirm the results of @xcite , that outflows decrease in force as they age .
there is a decrease in momentum flux from the class 0 to i stage : @xmath5 compared to @xmath6m@xmath7kms@xmath8yr@xmath8 , suggesting a decline in the mass accretion rate assuming the same entrainment fraction for both classes of outflow . if @xmath9 is the flux expected in radiation from the central source , then @xmath10 and @xmath11 .
furthermore , we confirm there are additional sources of mass loss from protostars . if a core s mass is only lost from outflows at the current rate , cores would endure a few million years , much longer than current estimates for the duration of the protostellar stage . finally , we note that the total energy contained in outflows in ngc1333 , l1448 and l1455 is greater than the estimated turbulent energy in the respective regions , which may have implications for the regions evolution .
[ firstpage ] submillimetre stars : formation stars : mass - loss ism : jets and outflows
ism : individual : perseus . |
ultracold atoms can be prepared with almost complete control over their density , temperature and interactions .
they serve as model systems for exploring unsolved problems in many - body physics as well as for demonstrating well - known textbook physics , such as of ideal , non - interacting gases that do not exist in nature .
for example several phenomena related to pauli blocking of fermions were clearly demonstrated only after the advent of ultracold fermi gases @xcite .
here we use the tunability of atomic interactions near feshbach resonances to create a non - interacting fermi gas with two components .
this is a realization of an ideal , non - interacting free electron gas ( feg ) with spin up and down components , as assumed in the simple theory of metals .
we demonstrate how such an ideal fermi gas will respond to effective magnetic fields , which is described by pauli paramagnetism .
the paper is mainly pedagogical .
it explains how paramagnetism is observed in trapped atomic samples which have an inhomogeneous density due to the harmonic confinement potential .
furthermore , these atomic samples realize a canonical ensemble in the fixed atom numbers @xmath1 and @xmath2 of the two components , whereas in metals and solid - state physics , @xmath3 is the magnetization determined by the applied external magnetic field . in all previous studies of paramagnetism ,
the magnetization was weak since the applied magnetic field times the magnetic moment was much smaller than the fermi energy @xcite . with ultracold atoms , we can easily realize the strong field case where the chemical potential difference is larger than the fermi energy and therefore fully polarizes the gas .
besides its pedagogical purpose , this paper experimentally demonstrates pauli paramagnetism in a truly non - interacting system exactly described by basic theory , whereas measurements even in simple metals revealed major discrepancies due to interaction effects @xcite .
( a ) response of the feg to an external magnetic field @xmath4 .
electrons in the two spin states are filled up to the fermi energy @xmath5 .
an external magnetic field @xmath4 shifts the energy of the two states into opposite directions and produces the density difference , or magnetization .
( b ) density distributions @xmath6 and @xmath7 of a two - component trapped fermi gas at @xmath8 . as discussed in the text , the local density polarization and corresponding normalized effective magnetic field field varies across the atomic cloud .
this is illustrated in the diagrams next to the curve for @xmath6 , where the gaps between the two spin states represent the absence of spin - flip processes in the system mentioned in the text.,title="fig : " ] + from atomic density profiles .
( a ) simulated density profiles of a two - component fermi gas in a spherically symmetric harmonic trap at zero temperature . the total population imbalance @xmath9 .
the shaded area corresponds to the partially polarized region where @xmath10 is well - defined .
( b ) the local chemical potential difference @xmath11 with ( solid ) and without ( dashed ) normalization by the fermi energy @xmath12 $ ] .
for the fully polarized gas @xmath13=1 $ ] according to the definition in the text .
( c ) the density difference @xmath14 with ( solid ) and without ( dashed ) normalization by the total density @xmath15 .
( d ) normalized density difference @xmath16 vs. normalized chemical potential difference @xmath17 . for the fully polarized gas , @xmath18 .
a linear fit at low @xmath17 ( dashed line ) gives @xmath19 .
( e ) normalized magnetic susceptibility @xmath10 versus the normalized chemical potential difference @xmath17.,title="fig : " ] + pauli paramagnetism explains the magnetization of a free electron gas with two spin states , neglecting the contribution from orbital motion . in a feg , the applied field shifts the energies of the two spin states in opposite directions as in fig .
[ fig : paramagnetism](a ) . for a uniform feg at temperature @xmath20=0 ,
the total number @xmath21 of electrons in each spin state inside a volume @xmath22 is given by @xmath23 where @xmath24 is the density of states , @xmath5 the fermi energy , @xmath25 the mass of an electron , @xmath26 the bohr magneton and @xmath27 planck s constant .
the magnetic susceptibility @xmath28 is defined as @xmath29 , where @xmath30 . as a dimensionless quantity , we consider the normalized susceptibility @xmath31\right)},\ ] ] where the polarization @xmath32 is used as a more meaningful measure of magnetization .
n)^{2/3}/{2m}$ ] and the total density is @xmath33 . here
we experimetanlly simulate pauli paramagnetism using trapped fermionic alkali atoms .
the two lowest hyperfine states of @xmath0li , @xmath34 and @xmath35 , represent the two spin states of an electron .
this atomic system is different from free electron systems : ( 1 ) due to the slow dipolar relaxation @xcite , the population in each spin state is conserved and therefore , ( 2 ) an external magnetic field does not lead to any magnetization , it only shifts the energies of the two spin state ( only in the case of strongly magnetic dipolar gases such as cr the dipolar relaxation is fast , and therefore the spin populations follow an external magnetic field @xcite ) . to realize the effective magnetic field we introduce density imbalance while preparing the system @xcite as illustrated in fig .
[ fig : paramagnetism](b ) . in the grand - canonical description of a system with fixed atom number
, a lagrange multiplier ( called the chemical potential ) ensures the correct expectation value for the atom number . in our case ,
the atom numbers in both spin states are fixed . in a grand - canonical description , this leads to two lagrange multipliers ( the two chemical potentials for the two spin states ) , or to one lagrange multiplier for the total number , and one for the population difference .
it is the sum of the latter lagrange multiplier and the external magnetic field which constitutes the effective magnetic field in our ensemble .
the contribution from the external magnetic field is always canceled out by an additive component in this multiplier .
for example , a balanced gas always has zero effective magnetic field independent of the applied field .
therefore the effective magnetic field depends only on the population imbalance and not on the external magnetic field . for a uniform fermi gas at @xmath8 ,
the number in each spin state determines chemical potential , @xmath36 @xcite and satisfies @xmath37 where @xmath38 and @xmath39 .
@xmath40 is the same as in eq .
( [ eq : solid_state ] ) if @xmath25 is taken to be the mass of a @xmath0li atom . comparing with eq.([eq : solid_state ] ) shows that the chemical potential difference @xmath11 is the effective magnetic field that takes the place of the magnetic field in a feg . with this effective magnetic field ,
the magnetic susceptibility can be written as @xmath41 we show now how a single atomic density profile in a harmonic trap @xmath42 can be used to determine the normalized susceptibility as a function of magnetic field . in the local density approximation ,
the density at each point in the trap is that of the corresponding uniform gas with modified chemical potentials @xmath43 @xcite , where @xmath44 are global chemical potentials defined for the whole trapped cloud which constrains the total number of atoms in each spin state . this concept has been used to determine the equation of state of interacting fermi gases @xcite .
here we are interested in the normalized magnetic susceptibility which is the slope of the normalized density difference versus the normalized chemical potential difference @xcite .
we demonstrate this procedure first using simulated density profiles ( fig .
[ fig : simulation ] ) .
as @xmath45 varies across the cloud , the chemical potential difference @xmath46 remains constant , as shown in fig .
[ fig : simulation](b ) .
however , the total density @xmath47 changes across the cloud , and consequently the local fermi energy @xmath48 $ ] as well , with @xmath12 $ ] being the same as in eq .
( [ eq : solid_state ] ) if @xmath25 is taken to be the mass of @xmath0li .
the normalized magnetic susceptibility @xmath10 depends on the dimensionless effective magnetic field @xmath49 $ ] , which does vary across the cloud , from small values near the center to the saturated value of one at the edges . from fig .
[ fig : simulation](c ) , a single set of atomic density profiles contains a range of polarizations @xmath50 and normalized effective magnetic fields @xmath51 $ ] .
the slope of the @xmath16 vs. @xmath17 yields the normalized susceptibility @xmath52 the normalization of @xmath53 in eq .
( [ eq : susceptibility ] ) is chosen to ensure @xmath54 for a fully polarized gas .
the slope near the origin , corresponding to small values of @xmath55 , gives @xmath56 @xcite .
as shown in fig .
[ fig : simulation](e ) , the susceptibility decreases with increasing @xmath55 and drops to zero when the fermi gas reaches full polarization .
is the 2p@xmath57 excited state in @xmath0li .
two successive images are taken at two different imaging frequencies , ( a ) and ( b ) .
corresponding ccd images are shown on the right.,title="fig : " ] +
an ultracold one - component fermi gas of @xmath0li in the state @xmath58 is prepared by sympathetic cooling with bosonic @xmath59na atoms @xcite . @xmath0li
atoms are then loaded into an optical dipole trap followed by radio - frequency ( rf ) transfer to @xmath60 using an rf landau - zener sweep at 300 g. a superposition of the two lowest hyperfine ground states , @xmath61 and @xmath62 is prepared by a rf sweep at 300 g , and holding here for 500 ms results in an incoherent mixture . .
representative profiles of ( a ) 2d column densities for @xmath63 ( upper solid curve ) and @xmath64 ( lower solid curve ) . temperature and chemical potential
are obtained from the fit ( dotted curves ) as discussed in the text .
each elliptical average is plotted versus the axial position ( long axis of the ellipse ) .
( b ) 3d density difference obtained from a single double - shot image .
dotted curve is a theoretical curve based on the chemical potentials and the temperature obtained from the 2d fit in ( a ) .
( c ) the normalized 3d density difference is plotted as a function of the normalized chemical potential difference .
the solid line is the theoretical prediction .
( d ) the linear fit at low @xmath55 gives the low - field susceptibility @xmath53=0.95(1 ) .
, title="fig : " ] + further cooling is provided by lowering the optical trapping potential which leads to radial evaporation ( axial confinement is provided magnetically ) .
we increase the magnetic field in 200 ms to 528 g , where the scattering length is zero @xcite , leading to a non - interacting two - component fermi gas .
the oscillation frequencies of the final trapping potential are @xmath65 hz axially and @xmath66 hz radially .
the total population imbalance @xmath67 is about 33% .
the final temperature is about 0.22@xmath68 where @xmath69 $ ] with @xmath15 being the total peak density and @xmath70 the boltzmann constant . to obtain density profiles of each spin state of @xmath0li
, we use a double - shot phase contrast imaging technique that involves two images taken at two different imaging frequencies @xcite ( see fig . [
fig : imaging ] ) .
one frequency is tuned between the two resonant transition frequencies for @xmath63 and @xmath64 states to the 2p@xmath57 electronically excited state .
the resulting phase contrast picture is a measure of the density difference @xmath71 . to measure the density of each component , a second image is taken of the same cloud at a frequency detuned towards the @xmath63 transition by 10 mhz .
the two images can be taken in quick succession because of the non - destructive nature of phase contrast imaging . for dispersive imaging techniques such as phase contrast imaging
, it is crucial to carefully focus the imaging system , since the atomic cloud refracts the probe light ( in contrast to resonant absorption imaging ) .
one can check this by comparing in - trap profiles of @xmath63 atoms imaged at positive and negative detunings of 60 mhz and find the focal position which minimizes the difference @xcite .
figure [ fig : density_profile2 ] shows the experimental results .
the camera images provide two - dimensional ( 2d ) column densities @xmath72 which integrate the density distribution along the line of sight .
to reduce the noise for the following analysis we take advantage of the symmetry of the trap and perform quadrant averaging and elliptical averaging : first , the four quadrants around the center of the images are averaged .
second , the data is averaged along elliptical contours .
the aspect ratio of the ellipses is determined from in - trap images @xcite .
temperature and global chemical potential for each component are obtained by fitting 2d densities with theoretical non - interacting density profiles . since the temperature @xmath20 is small but non - negligible ,
data are fitted with finite - temperature fermi gas 2d distribution in a harmonic trap given by @xmath73,\end{aligned}\ ] ] where @xmath74 is the @xmath15th - order polylogarithm .
@xmath75 is the axis along the line of sight .
since the majority component has better signal - to - noise ratio , we determine the temperature and chemical potential first for this component , and keep the temperature as a fixed parameter in the fit of the minority profiles .
three - dimensional ( 3d ) densities [ fig .
[ fig : density_profile2](b ) ] are obtained by performing the inverse abel transformation of the column densities @xcite .
the linear fit of @xmath16 as a function of @xmath55 at small @xmath55 gives the susceptibility at finite temperature [ fig .
[ fig : density_profile2](d ) ] . for finite but low temperatures ,
the susceptibility is known to vary as @xcite @xmath76.\ ] ] @xmath68 changes from 0.24 to 0.35 within the fitting region in fig .
[ fig : density_profile2](d ) , affecting the susceptibility by about 3% . for @xmath68=0.29 ,
this theoretical calculation gives @xmath77 and the experimentally obtained value is 0.95(1 ) in fig .
[ fig : density_profile2](d ) .
the discrepancy between the experimental and theoretical value is most likely due to a residual dispersion effect leading to systematic uncertainty estimated to be 10 - 20% @xcite .
we did not attempt to quantify this effect since this paper is mainly pedagogical and a dispersion effect depending on the imbalance as well as two detunings would be time - consuming to quantify figure [ fig : density_profile2](c ) illustrates that we can easily approach the high field limit of a fully polarized gas @xmath78 . note that @xmath16 is smaller than 1 because at finite temperature the minority component extends beyond its thomas - fermi radius . in a metallic free electron gas , with a typical fermi temperatures of 10,000
k , the high field region would require magnetic fields of 10,000 tesla , 200 times stronger than the strongest continuous laboratory magnetic fields .
we have demonstrated how the _ ideal gas _ magnetic susceptibility can be measured in ultracold fermi gases . after considering finite - temperature effects , the experimental results are in quantitative agreement with prediction from pauli paramagnetism .
since the local spin polarization varies across the trapped atomic cloud , a single density profile gives the susceptibility for both low and high effective magnetic fields .
this method can be applied to the study of magnetic properties of strongly interacting fermi gases which are of current interest @xcite .
this work was supported by the nsf and the onr , by an afosr muri on ultracold molecules , and by aro grant no .
w911nf-07 - 1 - 0493 with funds from the darpa optical lattice emulator program .
lee acknowledges support from the samsung scholarship .
t. t. wang acknowledges support from nserc .
we are grateful to caleb a. christensen for experimental assistance . | we show how to use trapped ultracold atoms to measure the magnetic susceptibility of a two - component fermi gas . the method is illustrated for a non - interacting gas of @xmath0li , using the tunability of interactions around a wide feshbach resonances .
the susceptibility versus effective magnetic field is directly obtained from the inhomogeneous density profile of the trapped atomic cloud .
the wings of the cloud realize the high field limit where the polarization approaches 100% , which is not accessible for an electron gas . |
local dwarf galaxies serve as potential examples of young galaxies in formation . due in part to their typically low metal abundances ,
the characteristics of the interstellar medium ( ism ) differ markedly from those in larger , more metal - rich systems . in cases where spatially and temporally concentrated star formation occurs , `` dwarf starburst '' galaxies inject energies into the ism that can be a substantial fraction of their system binding energies ; these bursts overwhelm the shallow potential wells of the parent galaxy and result in spectacular galactic winds ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
starbursts in low - mass dwarfs represent the low - luminosity end of the starburst superwind spectrum ; outflows increase in strength up to the ultraluminous infrared galaxies @xcite . in extreme cases
, single star clusters may input sufficient kinetic energy to drive galactic winds in dwarf galaxies .
the responsible `` super - star clusters''(sscs ) have typical ages @xmath6 1 gyr and dynamical masses @xmath0 10@xmath710@xmath8 m@xmath4 ; the most massive may be globular cluster progenitors . while many massive starburst and interacting systems are able to attain the densities and pressures requisite for ssc formation ( i.e. , n@xmath9 10@xmath10 @xmath11 , p / k
@xmath12 10@xmath810@xmath13 @xmath11 k , radii @xmath6 5 pc ; see oconnell 2004 for a recent review ) , the formation rate and maximum luminosity of sscs appear to generally scale with the total star formation rate ; typically these clusters form a power - law mass distribution . hence ,
massive , interacting galaxies such as the `` antennae '' ( ngc4038/4039 ) are able to produce hundreds of massive young clusters @xcite .
dwarf galaxies , on the other hand , produce relatively few massive sscs , and their formation may follow a different track than those of massive clusters in larger galaxies ; the cluster luminosity functions appear to be discontinuous in dwarfs , with sscs being much more massive than other clusters ( though this may be a statistical effect ; see whitmore 2004 ) . at a distance of [email protected] mpc @xcite ,
ngc1705 is one of the best - studied dwarf starburst galaxies .
the optical emission is dominated by the `` super - star cluster '' ( ssc ) ngc1705 - 1 @xcite , a compact ( effective radius @xmath0 1.5 pc ; * ? ? ?
* ) , massive ( dynamical mass @xmath0 10@xmath7 m@xmath4 ; * ? ? ?
* ) cluster of age @xmath6 12 myr @xcite .
the age of the cluster and the amount of ionized gas relative to the uv luminosity are consistent with most of the o stars having died out , so the system may be more correctly deemed a `` post - starburst '' galaxy @xcite .
dominant sscs are comparatively rare in dwarf starburst galaxies , and ngc1705 is one of the few systems where the luminosity of a single cluster dominates the galaxy s uv light .
the recent star formation is driving a multi - phase outflow from the disk , seen in the neutral , ionized , coronal , and x - ray emitting gas phases @xcite .
the h i dynamics show normal solid - body rotation , though extra - planar gas may also be involved with the outflow @xcite .
ngc1705 has undergone strong recent star formation ; @xcite compare synthetic color - magnitude diagrams to optical and near - infrared photometry to estimate a rate of 0.3 m@xmath4yr@xmath15 ; combining with the total h i mass of 1.04@xmath210@xmath8 m@xmath4 ( using the integrated flux from meurer etal .
1998 scaled to our adopted distance ) implies a gas depletion timescale of only @xmath0 0.35 gyr .
given the galaxy s recent starburst , its sub - solar nebular metallicity [ z @xmath0 35% z@xmath4 ( using the solar oxygen abundance of asplund etal .
2004 ) ; lee & skillman 2004 ] , its gas - rich central disk ( n@xmath16 @xmath17 10@xmath18 @xmath19 ; meurer etal . 1998 ) , and the presence of a ssc , it presents an ideal example of a starbursting dwarf galaxy with an extreme star - forming ism .
we present _ spitzer _ imaging and spectroscopy of this system , obtained as part of the _ spitzer infrared nearby galaxies survey _ ( _ sings _ ; * ? ? ?
* ) , with the aim of characterizing the nature of warm dust and pah emission in the metal - poor , active ism .
this system has one of the highest current star formation rates of any dwarfs in the _ sings _ sample and is the only dwarf in the survey that contains a known ssc .
for a complete overview of the _ sings _ observational strategies , see @xcite .
ngc1705 was observed in standard irac imaging mode ( producing two [email protected] fields , offset by @xmath0 1 ) and in mips scan mapping mode ( covering @xmath00.5 , with foreground and background sources dominating the emission at @xmath121.5@xmath2d@xmath20 and beyond ) .
irs spectral mapping was performed using all four modules ; the resulting maps cover an area larger than the optical body in all modules .
irs and irac data were obtained in 2004 july ( 62.4 minutes in irs long - low and short - low modules , and 18.5 minutes in irac imaging ) ;
mips data were obtained in 2004 september ( 58.6 minutes in all three modules ) .
all data were processed by the _ sings _ pipelines .
the irac pipeline reduces basic calibrated data images ; flux levels are uncertain at the @xmath0 10% level due to systematics ( see discussion of aperture correction effects in [ s3 ] ) .
the mips instrument team data analysis tool @xcite was used to process the mips data , while the _ cubism _ tool ( j. d. smith etal . , in preparation ) was used for spectral extraction and analysis of the irs data .
systematic uncertainties limit the absolute flux calibration to @xmath0 10% in the mips 24 @xmath1 m band and to @xmath0 20% in the mips 70 and 160 @xmath1mbands ; the flux uncertainty of the irs data is @xmath0 20% .
the fwhms of the mips psfs are 6 , 18 , and 38 at 24 , 70 , and 160 @xmath1 m , respectively .
global fluxes were measured after foreground / background source removal and background subtraction , using matched 105 radius circular apertures , centered on the ssc , in all mips and irac wavebands ; this radius is slightly larger than the fields of view shown in figure [ figcap1 ] . no emission associated with ngc1705 is present outside the area shown in figure [ figcap1 ] . since our measured global flux densities agree within the errors with those published in @xcite , we adopt their values as our final global flux density measurements .
the 11 @xmath1 m pah map was created from the full spectral data cube by averaging the flux density over the wavelength region 11.115
11.860 @xmath1 m , and subtracting the average continuum in the flanking regions ( 10.805 11.115 @xmath1 m and 11.860 12.357 @xmath1 m ) .
the h@xmath5 image analyzed here was taken from the public archive of the palomar / las campanas imaging atlas of blue compact dwarf galaxies @xcite .
these images were obtained with the du pont 100-inch telescope of the las campanas observatory .
the total flux of the galaxy was matched to the value presented in @xcite ; note that the ssc has been removed from the h@xmath5 image used in this paper .
figure [ figcap1 ] shows images of ngc1705 at 8 different wavelengths .
( a ) shows the near - uv ( @xmath21 380 nm ) continuum , where the ssc is the brightest single uv source in the galaxy ( producing more than @xmath0 40% of the uv photons ; * ? ? ?
there is an underlying stellar continuum containing a sizeable fraction of intermediate - mass stars : in the planetary camera field of view shown in figure [ figcap1](a ) , @xcite find that @xmath0 50% of the detected stars have masses @xmath12 3 m@xmath4 ( corresponding to an age of @xmath6 500 myr ) , while @xmath0 10% have masses @xmath12 7 m@xmath4 ( ages @xmath6 60 myr ) .
emission is present throughout the optical body of the system ( see figure [ figcap1]b ; note that the ssc has been blanked in this image ) ; the highest - surface brightness emission is offset from the ssc by @xmath0 250 pc , coincident with the morphology of dust observed in the longer - wavelength _
spitzer _
bands ( see below ) . moving into the near- and mid - infrared ( ir ) , the irac 3.6 and 4.5 @xmath1 m images ( see figures [ figcap1]c , d ) are sensitive to the underlying red stellar population ( typically consisting of young red supergiants and an older stellar population ) . the 3.6 @xmath1 m band is also known to contain some pah emission bands @xcite as well as emission from h ii gas processes ( e.g. , * ? ? ?
the irac 5.8 and 8 @xmath1 m bands ( see figures [ figcap1]e , f ) are primarily sensitive to dust and pah emission , though a stellar component is present as well ( e.g. , * ? ? ? * ) . note that the ssc retains a high surface brightness in the near- and mid - ir , though by 24 @xmath1 m the off - nuclear dust emission regions contribute the bulk of the global flux density ( see figure [ figcap1]g ) .
lcccc[!ht ] @xmath5 ( j2000 ) & 04:54:14.59 & 04:54:12.42 & 04:54:13.48 & 04:54:13.5 + @xmath22 ( j2000 ) & @xmath2353:21:36.20 & @xmath2353:21:45.07 & @xmath2353:21:38.95 & @xmath2353:21:40 + h@xmath5 flux & [email protected] & [email protected] & [email protected] & [email protected] + irac 3.5@xmath1 m flux density & [email protected] & [email protected] & [email protected] & 28@xmath143 + irac 4.5@xmath1 m flux density & [email protected] & [email protected] & [email protected] & 19@xmath142 + irac 5.7@xmath1 m flux density & [email protected] & [email protected] & [email protected] & 12@xmath142 + irac 7.9@xmath1 m flux density & [email protected] & [email protected] & [email protected] & 22@xmath142 + mips 24@xmath1 m flux density & [email protected] & [email protected] & [email protected] & 52@xmath145 + mips 70@xmath1 m flux density & 126@xmath1463 & 109@xmath1455 & n / a & 1090@xmath14220 + mips 160@xmath1 m flux density & n / a & n / a & n / a & 1200@xmath14250 + [ t1 ] + there is a striking difference in the galaxy s morphology shortward and longward of @xmath0 5 @xmath1 m .
there is a negligible amount of off - nuclear dust emission in the 4.5 @xmath1 m band , while longer wavelengths show two strong mid- and far - ir emission peaks that are offset from the ssc and coincident with the h@xmath5 surface brightness maxima .
these dust emission regions , labeled as ngc1705 d1 and d2 ( see figure [ figcap1]f ) , are the highest - surface brightness regions in ngc1705 in the mips bands ( together contributing @xmath0 25% of the 24 @xmath1 m global flux density ; see below ) .
we discuss the nature of the off - nuclear emission in more detail in [ s3.1 ] .
we measure the flux densities of the three individual dust emission peaks ( ngc1705 d1 , d2 , and ngc1705 ssc-1 ) , at a resolution slightly larger than that of the mips 24 @xmath1 m band ( fwhm @xmath245.7 ) , in the irac , mips 24 @xmath1 m , and h@xmath5 images .
fluxes were extracted using non - overlapping apertures of 10 diameter , centered on the locations given in table [ t1 ] and shown in figure [ figcap1 ] . due to the distance and size of ngc1705 , we lose sufficient resolution by the 70 @xmath1 m band to distinguish the individual emission peaks .
no aperture corrections were applied to the measured flux densities ; aperture corrections for extended sources are estimated to be between @xmath0 425% ( e.g. , * ? ? ?
* ) . for point sources ,
the aperture corrections are estimated to be of order 57% in the irac bands for this aperture size , while at 24 @xmath1 m the correction may be as large as a factor of 2 ; this effect may be important for the discussion in [ s3.1 ] and [ s3.2 ] .
the values listed in table [ t1 ] demonstrate that the spectral energy distributions ( seds ) from both ngc1705 d1 and d2 are beginning to rise towards a far - ir peak characteristic of dust emission in normal galaxies ( see , e.g. , * ? ? ?
the sed of the ssc , however , appears to flatten out by the mips 24 @xmath1 m band , suggesting that it contributes only a small fraction of the total far - ir luminosity of the system ; sed model fits to each of the emission peaks is not feasible , given the lack of resolution at the longer mips wavelengths ( but see dale etal.2005 and [ s3.1 ] for the galaxy s global sed and model fit ) .
as shown in figures [ figcap1 ] and [ figcap2 ] , both the ir and h@xmath5 emission peaks ( ngc1705 d1 and d2 ) are displaced from the ssc by @xmath12 250 pc .
@xcite use uv spectroscopy to demonstrate that the ssc produces the bulk of the uv luminosity of ngc1705 , but too few ionizing photons to produce the h@xmath5emission seen hundreds of pc away at the locations of d1 and d2 .
thus , young , massive stars ( stellar masses in the range 10 - 30 m@xmath4 , with corresponding main sequence lifetimes @xmath25 530 myr , depending on stellar evolution models and adopted metallicity ; see tosi etal .
2001 and annibali etal.2003 ) appear to be the source of the ionizing photons leading to the observed off - nuclear h@xmath5 emission . to assess whether the stellar population creating this h@xmath5 emission also powers the observed dust emission at locations d1 and d2 , it is instructive to compare the star formation rates derived from h@xmath5 and from the ir . due to the distance of the system , the physical resolution at 70 and 160 @xmath1mprecludes
the use of a total ir - based star formation rate for the individual peaks d1 and d2 ( e.g. , * ? ? ?
we instead apply the monochromatic calibration between h@xmath5 and 24 @xmath1 m luminosity derived in m51 by @xcite : @xmath26 where l@xmath27 and l@xmath28 are the observed luminosities at 24 @xmath1 m and h@xmath5 , respectively ; the intrinsic h@xmath5/p@xmath5 ratio is assumed to be 8.734 ; and we apply the central wavelength of 24.0 @xmath1 m to convert the 24 @xmath1 m flux density to a luminosity measurement . performing this comparison , we find that the observed 24 @xmath1 m luminosities of d1 and d2 are factors of 3 and 5 lower than predicted from the h@xmath5 luminosities alone .
note that extinction at h@xmath5 , while minimal [ @xcite find only marginal a@xmath29 values of 0.0 and 0.47 mag for d1 and d2 , respectively ] , would make this discrepancy larger ; aperture corrections at 24 @xmath1 m ( which may be as large as a factor of two ) will bring these values into better agreement . a lower star formation rate derived from the ir compared to h@xmath5 has been found for other dwarf galaxies in the _ sings _ sample ( cannon etal.2005 ; walter etal . , in prep . ) .
this effect has been attributed , to first order , to the lower dust contents found in the metal - poor ism . in ngc1705 , h@xmath5 and 24 @xmath1
m star formation rates agree within a factor of a few , suggesting that a substantial component of the dust emission at d1 and d2 is powered by re - radiation of photons from the same sources that are responsible for the h@xmath5 emission .
the offset may be due to differences in ism physics between metal - rich and metal - poor galaxies . while the energetics are comparable , we can not rule out a contribution from embedded sources that do not contribute to the observed h@xmath5 emission . in this regard ,
it is worth noting that some low - metallicity galaxies harbor heavily - embedded star formation regions ( e.g. , sbs0335@xmath23052 ; houck etal.2004 , hunt etal .
2004 ) .
an estimate of the dust mass in each component is not possible , given our inability to resolve the two peaks at 160 @xmath1 m .
the total dust mass in ngc1705 can be very simplistically estimated by applying the following relation : @xmath30 where d is the distance to the galaxy , @xmath31 is the flux density , @xmath32 represents the absorption opacity of the dust at frequency @xmath33 ( given in li & draine 2001 ) , and @xmath34 is the planck function evaluated at temperature t , derived by fitting a blackbody modified by a @xmath35 emissivity function to the 70 and 160 @xmath1 m data .
as shown in figure [ figcap3 ] , such a single - temperature solution is unable to simultaneously fit the observed 24 , 70 and 160 @xmath1 m data ; the fit to the 70 and 160 @xmath1 m flux densities grossly underestimates the observed 24 @xmath1 m flux density , implying a two - component dust grain population ( e.g. , cool grains at @xmath0 28 k , warmer grains at @xmath0 55 k ) .
similar conclusions are reached based on _ iso _ observations of nearby galaxies @xcite .
the inability of a single blackbody function to fit the observed flux densities suggests that a more sophisticated model of the far - ir sed is needed .
the sed models of @xcite and @xcite provide a more robust treatment of the multiple grain populations that contribute to the far - ir emission in a galaxy ( see dale etal.2005 for an application of this method to the entire _ sings _ sample ) . using the fit parameters for the global sed of ngc1705 presented in @xcite and shown in figure [ figcap3 ] ,
the models of @xcite imply that the dust mass derived using single - temperature fitting ( here , to the observed 70 and 160 @xmath1 m data ) underestimates the true dust mass by a factor of @xmath0 9 .
this application of the @xcite models results in a global dust mass estimate of ( [email protected])@xmath210@xmath7 m@xmath4 .
a third estimate of the dust mass in the galaxy can be found using the models of li & draine ( 2001 , 2002 ) . here ,
radiation field strengths are varied via power - law distributions ; pah , silicate and graphite grains are illuminated and the resulting seds can be compared to the observations .
the fit of these models to our observations is also shown in figure [ figcap3 ] . the observed
sed is well - reproduced by @xmath0 2.3@xmath210@xmath36 m@xmath4 of gas and dust , assuming the latter is of `` lmc2-type '' ( see weingartner & draine 2001 ) .
the dust - to - gas ratio ( @xmath37 ) of lmc2-type dust is estimated to be @xmath0 0.003 , which implies a total dust mass of @xmath0 7@xmath38 m@xmath4 ( with an uncertainty of @xmath0 50% ) in ngc1705 . given the uncertainties , we consider the agreement of the dust mass estimates using the sed models of @xcite and @xcite and the models of @xcite to be satisfactory .
we adopt the mean value of these two estimates , ( [email protected])@xmath210@xmath7 m@xmath4 , as the global dust mass of ngc1705 .
comparing this total dust mass with the h i mass measured in a matching aperture ( 8.9@xmath210@xmath36 m@xmath4 ; meurer etal .
1998 ) implies a global @xmath37 of @xmath0 2.5@xmath210@xmath39 , uncertain by @xmath0 50% .
this value is a factor of @xmath0 24 lower than the galactic average values ( @xmath37@xmath40 @xmath0 0.0060.01 ; sodroski etal .
1997 , li 2005 ) , but larger than values found for the more metal - poor small magellanic cloud ( z @xmath0 23% z@xmath4 ; dufour 1984 ; average dust - to - gas ratios @xmath0 10 - 30 times @xmath41 @xmath37@xmath40 ; stanimirovic etal .
2000 ; bot etal.2004 , stanimirovic etal.2005 ) .
if there is an additional amount of cold dust in ngc1705 that radiates longward of 160 @xmath1 m ( similar to what is seen in some dwarf galaxies with very cold seds ; e.g. , tuffs etal .
2002 ) , then this value will be a lower limit .
since the irac 8 @xmath1 m band samples the 7.7 @xmath1 m pah emission complex as well as the underlying warm dust continuum ( see * ? ? ?
* ; * ? ? ?
* ) , it is instructive to quantify the amount of the diffuse emission at 8 @xmath1 m that is due to pah emission . using the short - low irs module ( @xmath42 6 - 14 @xmath1 m , r @xmath25 100 ) , a summation over the wavelength ranges surrounding the two brightest pah bands ( 7.5 - 9.1 @xmath1 m and 10.7 - 11.7 @xmath1 m ) yields a detection of the 11.3 @xmath1 m pah complex in ngc1705 d1 only , with a nominal rise in the continuum level surrounding the 7.7 @xmath1 m pah feature ( though of low s / n ; see the spectrum of this region presented in figure [ figcap4 ] ) .
there are no pah bands detected in ngc1705 d2 or towards the ssc . in the top panel of figure [ figcap4 ]
we also overlay a typical pah - dominated spectrum from observations of the star - forming spiral galaxy ngc7331 @xcite ; note the agreement of the relative strengths of the 11.3 and 7.7 @xmath1 m pah bands in ngc1705 d1 with those of the example pah spectrum .
figure [ figcap5 ] shows contours of a ( continuum subtracted ) irs spectral summation of the 11.3 @xmath1 m pah emission , overlaid on the _
hst_/wfpc2 f380w and h@xmath5 images .
pah emission is associated with high - surface brightness h@xmath5 emission in the eastern dust emission complex ( ngc1705 d1 ) , suggesting that local radiation density is an important parameter that affects the appearance of emission from these molecules in the ism ( but note the trend of decreasing pah emission strength with increasing radiation field density discussed in * ? ? ?
. however , radiation field strength alone seems to be insufficient for the emergence of pah emission : there is high surface brightness h@xmath5 emission associated with ngc1705 d2 , but no associated pah emission is seen ( see also the relative flux ratios in table [ t1 ] ) .
previous studies have detected the 7.7 and 11.3 @xmath1 m pah bands in the diffuse ism of the milky way , unassociated with regions of massive star formation and apparently excited by the local interstellar radiation field @xcite .
other factors also likely play important roles in the production and destruction of pah emission , including the local electron and gas densities @xcite and supernovae shocks @xcite . the extracted flux density in the 11.3 @xmath1 m pah band in ngc1705 d1 is found to be [email protected] mjy over a 0.745 @xmath1 m spectral region ( see discussion in [ s2 ] ) .
given the low s / n ratio of the associated 7.7 @xmath1 m pah feature in the irs spectrum , its flux density can be estimated by assuming a uniform surface brightness over the same aperture size ; from figure [ figcap4 ] , a peak value of @xmath0 3.5 mjysr@xmath15 is representative .
scaled to the 10 diameter circular aperture used for irac flux extraction , this corresponds to a flux density of @xmath0 2.6 mjy .
if the continuum intensity of @xmath0 1 mjysr@xmath15 seen between 9 and 10 @xmath1 m is representative of the background under the 7.7 @xmath1 m feature , then this value will decrease to @xmath0 1.8 mjy .
these values can be compared with the residual emission in the irac 8 @xmath1 m band after continuum removal ; the exemplary value of ( irac 8 @xmath1 m ) @xmath23 ( 0.232)@xmath43(irac 3.5 @xmath1 m ) used by @xcite yields a diffuse ( stellar subtracted ) 8 @xmath1 m flux density of [email protected] mjy ( though note that a single scaling factor is likely not applicable to all galaxies ) . depending on the underlying continuum levels , assumptions about the surface brightness distribution of the pah emission , and the scaling factor used to remove the stellar continuum
, it thus appears that most of the emission from region d1 in the irac 8 @xmath1 m band arises from pah emission .
this is also evident from inspection of figure [ figcap4 ] ; the 7.7 @xmath1 m pah complex is clearly detected in d1 , but the spectrum of d2 is flat in this region .
these arguments imply that spatially variable pah strengths compared to the underlying dust continuum make dissecting pah and warm dust emission a complex issue .
this is exemplified by region d2 , which shows high surface brightness h@xmath5 , starlight - subtracted 8 @xmath1 m and dust emission but no detectable pah features .
since the spectra of photodissociation regions ( pdrs ) and of h ii regions are quite different ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , the differences between d1 and d2 may have to do with how much of a pdr is left in one case vs the other .
irs spectra should be used to verify the presence and strength of pah emission that may constitute an unknown fraction of the detected diffuse emission at 8 @xmath1 m . while the spectra presented in figure [ figcap4 ] are of relatively low s / n , they offer interesting comparisons of the physical conditions in the three regions probed .
ngc1705d1 shows pronounced [ s iv ] @xmath44 10.51 @xmath1 m emission , while this appears to be absent in d2 and the ssc .
the [ ne ii ] 12.81 @xmath1 m line , which is blended with the 12.6 @xmath1 m pah feature , appears in the spectra of both d1 and d2 .
the relative strengths of [ s iv ] and [ ne ii ] in d1 compared to d2 argue for a stronger radiation field in the former ( see discussion in * ? ? ?
the low s /
n of the spectra precludes a detailed treatment of the ionization structure of the system .
the spectrum of the ssc shows a broad peak around 10 @xmath1 m that may be suggestive of emission from amorphous silicates , either ultra small grains undergoing single - photon heating , or larger grains being heated to steady temperatures of @xmath0 250 k ( e.g. , * ? ? ?
@xcite find a pronounced change in the strength of pah emission as a function of metallicity in a sample of 34 star - forming galaxies . above
the level of 12+log(o / h ) @xmath24 8.2 ( @xmath25 35% z@xmath4 ) , pah emission is detected via diffuse 8 @xmath1 m emission and confirmed with irs spectroscopy ; at lower metallicities , pah emission is weak or absent in their sample .
ngc1705 has a nebular metallicity equal to this transition point [ 12@xmath45log(o / h ) @xmath24 [email protected] , measured at various locations throughout the central disk ( including both d1 and d2 ) ; see lee & skillman 2004 ] ; the variable pah emission from different regions of the galaxy provides an interesting test case for this empirical relation .
@xcite compare the 8/24 @xmath1 m ratio ( r@xmath46 ) to the 4.5/8 @xmath1 m ratio ( after the 4.5 @xmath1 m image has had an estimate of the underlying stellar continuum removed by scaling and subtracting the irac 3.5 @xmath1 m image ; r@xmath47 ) .
the 8/24 @xmath1 m ratio clearly separates star forming galaxies according to both metallicity and the relative strengths of stellar and dust emission : metal - rich galaxies preferentially show high 8/24 @xmath1 m ratios and low ratios of stellar to dust emission at 8 @xmath1 m , while metal - poor galaxies show opposite trends .
these color separations are interpreted to be caused , to first order , by the decreased emission strengths of pah molecules in the low - metallicity ism .
computing the diagnostic color ratios as described in @xcite , we find ( r@xmath47 , r@xmath46 ) values of ( [email protected] , [email protected] ) , ( [email protected] , [email protected] ) , and ( [email protected] , [email protected] ) for ngc1705 d1 , d2 , and the total galaxy , respectively , using the values in table [ t1 ] .
note , however , that aperture corrections in the mips 24 @xmath1 m band can be significant for the 10 diameter apertures used here ( between @xmath0 2050% , depending on the extent of the source and the contamination from neighboring emission regions )
. taken at face value , each of the r@xmath46 ratios is consistent with the average values found in the pah - emitting galaxies of the @xcite study ( i.e. , r@xmath48 0.2 ) .
the r@xmath47 ratio of the pah emission region , d1 , falls exactly on the color boundary defined by @xcite ; the two other regions show higher r@xmath47 ratios , consistent with values found for systems with pah non - detections .
if aperture corrections at 24 @xmath1 m are significant , the r@xmath46 ratio will decrease for regions d1 and d2 ( but not for the global measurements ) , and the @xcite diagnostic color ratios correctly separate the regions d1 and d2 into pah - emitting and pah - deficient , respectively . in normal star - forming galaxies ,
there is a remarkably tight correlation between the total ir luminosity and the strength of radio continuum emission ( see , e.g. , de jong etal .
1985 , helou etal.1985 , condon 1992 , and references therein ) .
the tightness of the relation ( scatter of less than 0.3 dex over multiple orders of magnitude ) is especially remarkable , given that the physical processes that drive ir ( re - radiation of absorbed photons by dust grains ) and radio emission ( typically nonthermal synchrotron emission , though a thermal free - free component can also contribute ) are only indirectly associated with one another .
indeed , @xcite shows that the relation holds for ir- and radio - deficient galaxies , but that these systems require a complex interplay of star formation , dust content and magnetic fields to produce the apparently linear relation in this regime .
we explore the relation between the global radio and ir luminosities of ngc1705 .
the total ir ( tir ) flux ( derived by applying equation 4 from dale & helou 2002 ) and the radio continuum luminosity appear to be related via a constant value in star - forming galaxies [ commonly denoted by the `` q '' parameter , where q @xmath49 log(s@xmath50/s@xmath51 ) ; see , e.g. , helou etal.1985 , bell 2003 ] .
@xcite find a nondetection of ngc1705 in the radio continuum at 6 cm , using data that are sufficiently sensitive to detect the galactic w49 star formation complex at the assumed distance of 5.1 mpc ( rms @xmath24 0.05 mjybm@xmath15 ) .
using the global flux densities of ngc1705 in table [ t1 ] , we derive s@xmath50 @xmath25 8@xmath210@xmath52 wm@xmath53 ; assuming that `` q '' @xmath24 2.64 @xcite and that the radio emission is a mix of thermal and nonthermal processes ( i.e. , @xmath540.7 , where s@xmath55 @xmath56 ) for the extrapolation from 20 to 6 cm flux densities , this would imply an expected radio continuum flux density of @xmath0 2 mjy .
this is more than an order of magnitude higher than the johnson etal .
( 2003 ) upper limit . reducing the assumed value of `` q '' ( e.g. , q @xmath24 2.34
; yun etal .
2001 , though note that this value is derived using only information on the sed between 42122 @xmath1 m ) will increase the expected flux density by roughly 50% ; assuming a stronger nonthermal component in the extrapolation from 1.4 to 4.8 ghz ( e.g. , @xmath5 @xmath571.2 , indicative of pure synchrotron emission ) reduces the expected value by @xmath0 50% . applying a monochromatic `` q '' parameter ( e.g. , * ? ? ?
* ) yields similar estimates of the expected radio continuum flux density .
regardless of the extrapolation , it is clear that the radio continuum of ngc1705 is extremely weak compared to the observed far - ir flux densities .
the detection , using these _ spitzer _ data , of strong far - ir emission shows that there is significant dispersion amongst individual galaxies in the far - ir vs. radio correlation .
indeed , @xcite suggests that this correlation is dependent on numerous factors and upon galaxy type , and that both the ir and radio luminosities of dwarf galaxies are suggested to significantly underestimate the true star formation rate .
it is important to note , by comparison with the internal scatter in individual galaxies seen in the larger sample of @xcite , that the tightness of the global radio - ir correlation may depend on averaging over a significantly large number of emission regions . in dwarf galaxies , and
starbursting dwarf galaxies in particular , the ir and radio emission are typically dominated by one ( or at most a few ) luminous star formation regions , allowing significant deviations from the relations seen in more massive galaxies . in larger starburst galaxies ,
a deficiency of radio continuum emission compared to the far - ir luminosity has been attributed to a very young ( @xmath6 few myr ) , `` nascent '' starburst that has occured after an extended period ( @xmath12 100 myr ) of quiescence ( e.g. , * ? ? ?
* ) . in the case of ngc1705 , which is ir - bright but radio - faint
, the aforementioned galactic wind may play an important role in the lack of radio continuum emission .
the young age of the ssc ( @xmath6 12 myr ; * ? ? ?
* ) implies that thermal emission associated with the uv photons of massive stars should still be strong ( see timescale arguments in condon 1992 and cannon & skillman 2004 ) .
similarly , the `` post - starburst '' nature of the system @xcite and spatially resolved stellar photometry @xcite have shown that most of the o stars associated with the strong burst of star formation have exploded as sne , implying a strong synchrotron emission component .
if the burst has `` blown out '' of the disk and is venting hot gas into the halo ( as evidenced by uv absorption line spectroscopy and diffuse h@xmath5 and x - ray gas in the halo ; see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the energetic particles that give rise to radio continuum emission are likely escaping the disk ( see also the discussion in * ? ? ?
we have presented observations of the nearby dwarf starburst galaxy ngc1705 made with all three instruments on board the _ spitzer space telescope_. this is one of the few local dwarf galaxies known to host a ssc , and the only such system in the _ sings _ sample .
the galaxy underwent a strong starburst some 15 myr ago that produced the ssc and massive star formation in the field ; this starburst is driving a multiphase outflow from the ssc region and this has dramatic consequences for the multiwavelength properties of the galaxy .
the galaxy morphology changes significantly longward and shortward of @xmath0 5 @xmath1 m .
optical and short - wavelength irac imaging shows an underlying red stellar population , with the ssc dominating the luminosity .
longer - wavelength irac and mips imaging reveals warm dust emission arising from two off - nuclear regions that dominate the far - ir morphology of the system .
stellar and warm dust emission from the ssc is still prominent at 8 @xmath1 m , but the sed appears to flatten out toward longer wavelengths , suggesting that most of the warm dust emission arises from the off - nuclear regions .
there is high surface brightness h@xmath5 emission associated with these dust emission peaks .
the energy source for this off - nuclear nebular and dust emission , some @xmath0 250 pc from the ssc , is likely young massive stars in these regions ; some of these stars may be embedded , thus not contributing to the observed h@xmath5 luminosity .
the lack of extinction toward the ssc @xcite and the modest warm dust emission of the ssc in the mips bands suggests that dust near the ssc may have been removed by the multi - phase outflow or evacuated by the high uv flux from the ssc .
we derive a total dust mass of @xmath0 ( [email protected])@xmath210@xmath7 m@xmath4 .
this value represents the mean of the dust mass estimates from using the sed models of @xcite and @xcite and the models of @xcite .
a single - temperature blackbody function is unable to simultaneously fit the observed 24 , 70 and 160 @xmath1 m flux densities .
both sets of sed models predict a higher dust mass than a single - temperature blackbody fit to the 70 and 160 @xmath1 m data points . comparing to the h i mass implies a global dust - to - gas ratio @xmath0 24@xmath2 lower than the galactic average .
the strong diffuse emission in the irac 8 @xmath1 m band is dominated by warm dust emission ; irs spectroscopy of the ssc and both off - nuclear sources shows 11.3 @xmath1 m pah emission in the more luminous of the two dust emission peaks only ( no pah emission is associated with the ssc ) .
this serves as a prudent reminder that irac diffuse 8 @xmath1 m emission traces both dust continuum and pah emission , and that spectroscopy is essential for a robust separation of the two components .
the fact that variable pah emission strengths are seen in a galaxy with nebular metallicity equal to the empirical pah threshold found by @xcite highlights the complexity in interpreting the nature of diffuse 8 @xmath1 m emission . finally , we note that ngc1705 appears to deviate significantly from the canonical far - ir vs. radio correlation , in the sense that strong far - ir emission is present , but radio continuum emission remains undetected at meaningful sensitivity levels
. the galactic outflow may play an important role in affecting the radio continuum emission ; if significant numbers of energetic particles are escaping into the halo , the radio continuum emission will be significantly suppressed compared to the total ir luminosity .
dwarf starburst galaxies , typically dominated by one massive star formation region , appear to be susceptible to strong variations in this ratio ; the tightness of the radio - ir correlation may rely on a larger statistical sampling of star formation regions throughout an individual galaxy .
the _ spitzer space telescope _
legacy science program `` the spitzer nearby galaxies survey '' was made possible by nasa through contract 1224769 issued by jpl / caltech under nasa contract 1407 .
the authors thank henry lee for helpful discussions , gerhardt meurer for providing the h i image used in our analysis , and the anonymous referee for an exceptionally detailed report that helped to improve the manuscript . 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 , and nasa s astrophysics data system . | we present _ spitzer _ observations of the nearby dwarf starburst galaxy ngc1705 obtained as part of the _ spitzer infrared nearby galaxies survey_. the galaxy morphology is very different shortward and longward of @xmath0 5 @xmath1 m : optical and short - wavelength irac imaging shows an underlying red stellar population , with the central super star cluster ( ssc ) dominating the luminosity ; longer - wavelength irac and mips imaging reveals warm dust emission arising from two off - nuclear regions that are offset by @xmath0 250 pc from the ssc and that dominate the far - ir flux of the system .
these regions show little extinction at optical wavelengths .
the galaxy has a relatively low global dust mass ( @xmath0 2@xmath210@xmath3 m@xmath4 , implying a global dust - to - gas mass ratio @xmath0 24 times lower than the milky way average , roughly consistent with the metallicity decrease ) .
the off - nuclear dust emission appears to be powered by photons from the same stellar population responsible for the excitation of the observed h@xmath5 emission ; these photons are unassociated with the ssc ( though a contribution from embedded sources to the ir luminosity of the off - nuclear regions can not be ruled out ) .
low - resolution irs spectroscopy shows moderate - strength pah emission in the 11.3 @xmath1 m band in the more luminous eastern peak ; no pah emission is detected in the ssc or the western dust emission complex .
there is significant diffuse emission in the irac 8 @xmath1 m band after starlight has been removed by scaling shorter wavelength data ; the fact that irs spectroscopy shows spatially variable pah emission strengths compared to the local continuum within this diffuse gas suggests caution in the interpretation of irac diffuse 8 @xmath1 m emission as arising from pah carriers alone .
the nebular metallicity of ngc1705 falls at the transition level of @xmath0 0.35 z@xmath4 found by engelbracht and collaborators , below which pah emission is difficult to detect ; the fact that a system at this metallicity shows spatially variable pah emission demonstrates the complexity of interpreting diffuse 8 @xmath1memission in galaxies . ngc1705 deviates significantly from the canonical far - infrared vs. radio correlation , having significant far - infrared emission but no detected radio continuum . |
this paper deals with the minimal circle - fibered self - maps of two dimensional manifolds .
we recall that the pair @xmath0 consisting of a compact metric space @xmath1 and its continuous endomorphism @xmath2 is called a minimal dynamical system if @xmath1 does not have any non - empty compact subset @xmath3 satisfying @xmath4 .
the understanding of the structure of minimal systems has a clear importance for the discrete dynamics . during the last decades
, much progress has been made in studying minimal subsystems of @xmath5 in the case when @xmath6 is a low dimensional compact connected manifold , e.g. see [ 114 ] . in particular , auslander and katznelson have proved @xcite that the minimality of @xmath5 together with dim @xmath7 implies that @xmath8 and that @xmath9 is conjugate to an irrational rotation ( hence , @xmath9 is a homeomorphism ) . if dim @xmath10 then , due to the blokh - oversteegen - tymchatyn theorem @xcite , the minimal manifold @xmath6 must be either the 2-torus @xmath11 or the klein bottle @xmath12 .
it was also shown in @xcite that , in contrast with the minimal system @xmath13 , there exist minimal fiber - preserving systems @xmath14 which are not invertible .
the key dynamical and topological components of the proof in @xcite are , respectively , the rees example @xcite of a non - distal but point - distal torus homeomorphism and the roberts - steenrod characterisation @xcite of the monotone transformations of 2-dimensional manifolds . since the available constructions @xcite of the minimal homemorphisms of the klein bottle are technically quite involved , the similar question about the existence of minimal non - invertible self - maps of @xmath12 has been left open in @xcite .
in fact , more complicated topology of the klein bottle ( a skew product of two circles ) in comparison to the torus @xmath11 ( a direct product of two circles ) could potentially be an obstacle for the existence of minimal non - invertible self - maps of @xmath12 , cf .
* theorem c-11 and corollary 1 ) and @xcite .
nevertheless , the main result of this paper shows that [ main ] there exists a fiber - preserving transformation @xmath15 of the klein bottle , which is minimal and non - invertible .
theorem [ main ] is proved in the next section of our work .
the proof uses the hric - jger blow up technique proposed recently in @xcite . in the cited work
, the authors also sketched a new construction of a non - distal but point - distal torus homeomorphism .
we develop further their construction and adopt it to a more complicated topological situation . as a by - product , even without the use of the roberts - steenrod theory of monotone transformations of 2-dimensional manifolds , we are able to present a relatively short and explicit analytic construction of a fiber - preserving minimal non - invertible self - map of the two - torus , cf . @xcite .
set @xmath16 then @xmath17 .
we will fix the positive orientation of @xmath18 induced by the usual order on @xmath19 .
each ordered pair @xmath20 of points in @xmath18 defines two closed subarcs of @xmath18 whose endpoints are @xmath21 and @xmath22 .
the arc obtained by moving a point from @xmath21 to @xmath22 in the positive direction will be denoted by @xmath23 \subset { { \bbb t}}^1 $ ] .
hence , @xmath23 \cup [ y , x ] = { { \bbb t}}^1 $ ] so that @xmath24 $ ] and @xmath25 $ ] . by slightly abusing the notation , we also will write @xmath26 = { { \bbb t}}^1 , \ [ 0,0.5]= [ 1,0.5]$ ] and @xmath27 = \{0\}$ ] . next , consider the homeomorphism @xmath28 defined by @xmath29 @xmath30 let @xmath31 be an equivalence relation on @xmath32 in which each point @xmath20 is identified with all its images : @xmath33 and @xmath34 .
let @xmath35 denote the corresponding quotient map .
the quotient space @xmath12 is one of standard models of the klein bottle .
notice that a transformation @xmath36 of the torus induces a transformation of the klein bottle by the quotient map @xmath37 if and only if @xmath36 commutes with @xmath38 .
the desired minimal map @xmath39 will be constructed as a factor of a minimal and non - invertible transformation @xmath40 of the torus . on the other hand ,
the map @xmath40 will be constructed as a topological extension of the parry minimal homeomorphism @xmath41 of the form @xmath42 . here
@xmath43 is a rotation by an irrational angle @xmath44 and continuous function @xmath45 is such that @xmath46 ( i.e. @xmath47 commutes with @xmath38 ) .
moreover , the fourier coefficients of @xmath48 must satisfy several assumptions listed in @xcite ; in addition , we can choose them in such a way that @xmath49 and @xmath50 for all @xmath51 . in the sequel , we will use the notation @xmath52 for the composition @xmath53 take now some point @xmath54 , set @xmath55 and then choose the points @xmath56 and @xmath57 in such a way that @xmath58 for each @xmath59 .
then the @xmath47-orbits of @xmath60 do not intersect curves @xmath61 and @xmath62 .
let continuous @xmath63 have their graphs @xmath38-invariant and intersecting transversally at the points @xmath64 and @xmath65 , see figure 1 .
we will choose @xmath66 in such a way that they take zero values ( recall that @xmath67 is identified with @xmath68 ) in all points except for some neighbourhood @xmath69 of @xmath70 and @xmath71 of @xmath72 , respectively .
we define a fiber measure @xmath73 on the @xmath74-algebra @xmath75 of borelian subsets of @xmath76 as @xmath77}}{\phi(x)-\psi(x ) } , & \mbox{if } \
\phi(x)>\psi(x),\\ \frac{\lambda|_{[\phi(x),\psi(x)]}}{\psi(x)-\phi(x ) } , & \mbox{if } \ \psi(x)>\phi(x),\end{array}\right.\ ] ] where @xmath78 denotes a probabilistic dirac measure concentrated at @xmath22 ( i.e. @xmath79 ) and @xmath80}(a)$ ] denotes the lebesgue measure of intersection of a measured set @xmath81 and the arc @xmath82 $ ] .
consider also the measures @xmath83 and @xmath84 on @xmath75 defined by @xmath85 it is clear that @xmath86 @xmath87 .
in addition , @xmath88 inherits the symmetry properties of @xmath89 : for all @xmath90 and @xmath91 $ ] , we have @xmath92=1-\mu_{x+\frac12}[0,1-y].\ ] ] first we prove that , for all @xmath93 , it holds @xmath94=1-\mu_{x+\frac12}^0[0,1-y].\ ] ] for @xmath95\setminus \{x_1^ * , x_2^\ast\}$ ] ( i.e. for the fibers with the usual lebesgue or dirac measures ) , this relation is immediate .
now , let @xmath96 be such that @xmath97 and hence @xmath98
. then @xmath99=\left\{\begin{array}{cl}0 , & \mbox{if } \ 1-y<\psi\left(x+\frac12\right),\\ \frac{1-y-\psi\left(x+\frac12\right)}{\phi\left(x+\frac12\right)-\psi\left(x+\frac12\right ) } , & \mbox{if } \
\psi\left(x+\frac12\right)\leq1-y\leq\phi\left(x+\frac12\right),\\1 , & \mbox{if } \ 1-y>\phi\left(x+\frac12\right);\end{array}\right\}\\=\left\{\begin{array}{cl}0 , & \mbox{if } \
y>\psi(x),\\ \frac{\psi(x)-y}{\psi(x)-\phi(x ) } , & \mbox{if } \ \phi(x)\leq y\leq \psi(x),\\ 1 , & \mbox{if } \
y<\phi(x);\end{array}\right\}= \quad 1-\mu^0_x[0,y ] .
\end{gathered}\ ] ] similarly , ( [ mm ] ) holds when @xmath100 .
next , we claim that @xmath101=1-\mu_{x+\frac12}^0[1-a,1-b ] , \ \mbox{for } \ a , b \in [ 0,1],\ a < b , \ [ a , b ] \in { { \bbb t}}^1.\ ] ] this is trivial for the fibers over @xmath70 and @xmath72 . for
the rest of the fibers , ( [ 1-a ] ) follows from ( [ mm ] ) in view of @xmath102=\left\{\begin{array}{cl}\mu_x^0[0,b]-\mu_x^0[0,a ] , & \mbox{if } \
0\leq a\leq b\leq 1;\\ 1-\mu_x^0[0,a]+\mu_x^0[0,b ] , & \mbox{if } \
< a\leq 1.\end{array}\right.\ ] ] finally , we will show that the superscript @xmath68 in ( [ mm ] ) can be omitted .
indeed , set @xmath103 .
then @xmath104=\sum_{n \geq 0 } 2^{-n-2}\mu_{r^{n}(x+\frac12)}^0\sigma_{x+\frac12}^{n}[0,1-y]+(1-y)/2=\\ \sum_{n \geq 0 } 2^{-n-2}\mu^0_{x+\frac12+n\alpha}\left[\varrho_n\left(x+\frac12\right),1-y+\varrho_n\left(x+\frac12\right)\right]+(1-y)/2=\\ \sum_{n \geq 0 } 2^{-n-2}\mu_{x+n\alpha+\frac12}^0[-\varrho_n(x),1-y-\varrho_n(x)]+1/2-y/2=\\ \sum_{n \geq 0 } 2^{-n-2}(1-\mu_{x+n\alpha}^0[1+\varrho_n(x),y+\varrho_n(x)])+1/2-y/2=\end{gathered}\ ] ] @xmath105+y/2\right ) = 1-\mu_x[0,y].$ ] the next two results show that @xmath106 $ ] and @xmath107 $ ] have discontinuities only at the points from backward orbits of @xmath108 and @xmath109 : [ 22 ] let @xmath110 be sequences in @xmath76 which converge to @xmath111 and @xmath112 , respectively .
then @xmath113\leq\mu_{x_0}^0[y_0,z_0].\ ] ] moreover , if @xmath114 , then @xmath115\geq\mu_{x_0}^0[y_0,z_0).\ ] ] since the functions @xmath66 are continuous , this result is a straightforward consequence of the definition of @xmath84 .
in fact , for each @xmath116 , we have that @xmath117\to\mu_{x_0}^0[y_0,z_0]$ ] as @xmath118 . [ pi3 ]
let @xmath119 and @xmath120 converge to @xmath111 and @xmath121 , respectively
. then @xmath122\leq \limsup_{j\rightarrow\infty}\mu_{x_j}^n[0,y_j]\leq\mu_{x_0}^n[0,y_0 ] , \quad n \in { { \bbb n}}.\ ] ] furthermore , @xmath123\leq \limsup_{j\rightarrow\infty}\mu_{x_j}[0,y_j]\leq\mu_{x_0}[0,y_0].\ ] ] since @xmath124 are continuous functions and , by our assumption , @xmath125 for each @xmath126 , we have that @xmath127 @xmath128 \leq \liminf_{j\rightarrow\infty}\mu_{x_j}^n[0,y_j].\ ] ] the proof of the second inequality in ( [ pi ] ) is similar .
finally , by a direct calculation we get the following : @xmath129+y_j/2\right)\leq\sum_{n\geq0}2^{-n-2}\mu_{x_0}^n[0,y_0]+y_0/2=\mu_{x_0}[0,y_0],\ ] ] @xmath130+y_j/2\right)\geq\sum_{n\geq0}2^{-n-2}\mu_{x_0}^n[0,y_0)+y_0/2=\mu_{x_0}[0,y_0).\ ] ] hence , if @xmath131 $ ] is discontinuous at some point @xmath132 then @xmath133 .
following @xcite , we consider continuous fiber - preserving selfmap @xmath134 defined by @xmath135 , where @xmath136|\mu_x[0,y^\prime]\geq y\}.\ ] ] the existence of this minimum can be deduced , for example , from corollary [ pi3 ] .
it is clear that @xmath137 and @xmath138 because , for all small @xmath139 , @xmath140=1-\mu_{x+\frac12}[0,\epsilon]=1-\sum_{n \geq 0 } 2^{-n-2}\mu_{x+\frac12}^n[0,\epsilon]-\epsilon/2 < 1.\ ] ] obviously , @xmath141 is an increasing function of @xmath142 .
now , let @xmath143 [ respectively , @xmath144 denote the full [ respectively , backward ] orbit of a point @xmath90 .
as we have proved , @xmath107 $ ] is continuous at each point @xmath20 where @xmath145 .
therefore @xmath146= y \
\mbox{if } \ x \not\in \frak{o}^-\ ] ] that implies @xmath147 for all @xmath148 and @xmath149 .
[ lmmcontt ] the map @xmath150 is continuous and surjective .
moreover , @xmath151 commutes with @xmath38 and is invertible on the set @xmath152 .
it is clear that @xmath151 is continuous if and only if @xmath141 is a continuous function of @xmath153 .
so , take some @xmath154 and consider sequences @xmath155 and @xmath156 . from the definition of @xmath157 it holds @xmath158\geq y_j$ ] .
suppose that @xmath159 converges to some limit point @xmath160 .
then corollary [ pi3 ] yields @xmath161 suppose for a moment that @xmath162 and take @xmath163 such that @xmath164 for all large @xmath165
. then @xmath166 and therefore , due to corollary [ pi3 ] , we have for all large @xmath165 that @xmath167.\ ] ] as a consequence , @xmath168 for all large @xmath165 , a contradiction .
hence , @xmath169 and @xmath141 is continuous .
now , since @xmath141 depends continuously on @xmath170 and @xmath137 and @xmath138 , we obtain that @xmath171 .
in addition , if @xmath172 then @xmath141 is a strictly increasing function of @xmath22 and therefore @xmath151 is invertible on @xmath152 .
next , since @xmath173 we find that @xmath151 commutes with @xmath38 if and only if @xmath174 now , for @xmath175 it holds that @xmath176+\mu_{x+\frac12}[0,1-\tau_x(y)]=y+\mu_{x+\frac12}[0,1-\tau_x(y)],\ ] ] so that @xmath177=1-y=\mu_{x+\frac12}[0,\tau_{x+\frac12}(1-y)].\ ] ] thus @xmath151 commutes with @xmath38 on @xmath152 . finally , if @xmath178 we can find a sequence of points @xmath179 such that @xmath180 as @xmath118 .
but then @xmath181 this completes the proof .
we are ready to construct the non - invertible minimal map @xmath182 .
defining @xmath183 on the set @xmath184 , where @xmath185 , by @xmath186 we will extend it continuously on the whole torus .
for each @xmath187 , the map @xmath188 $ ] .
[ lmmpom ] we observe that @xmath189 and thus it suffices to establish the continuity of @xmath190 on the arcs @xmath191 $ ] and @xmath192 $ ] of @xmath18 separately .
for instance , consider the arc @xmath191 $ ] where @xmath193 so that @xmath194 $ ] .
we recall that , by our assumptions , none of the points @xmath195 , coincided with @xmath108 and @xmath109 and therefore the function @xmath196 $ ] is continuous at each point of the form @xmath197 . in consequence , the map @xmath198 $ ] is continuous as composition of two continuous applications : @xmath199,$ ] and @xmath200,$ ] @xmath201\}$ ] .
next , set @xmath202\times\mathbb{t}^1\right)$ ] and @xmath203\times\mathbb{t}^1\right)$ ] .
the map @xmath183 is uniformly continuous on @xmath204 .
it follows from ( [ dt ] ) that @xmath205 $ ] for each pair @xmath206 hence , @xmath207 ) , \ ( x , y ) \in\lambda.\ ] ] since @xmath208 is continuous , we need only to prove the uniform continuity of @xmath209 $ ] on @xmath204 .
this task can be simplified if we observe that , due to the weierstrass m - test and lemmas [ lmmcontt ] and [ lmmpom ] , the function @xmath210 + \sigma_x(\tau_x(y))/2 $ ] is continuous on @xmath32 while @xmath211}{2^{n+2 } } + \sigma_x(\tau_x(y))/2 = \sum_{n\geq 0 } \frac{\mu^{n+1}_{x}[0,\tau_x(y)]}{2^{n+2 } } + w(x , y).\ ] ] here we are using the relation @xmath212 = \mu_{r^{n}(r(x))}^0\circ\sigma_{r^{n-1}(r(x))}\circ\ldots\circ\sigma_{r(r(x))}\circ\sigma_{r(x)}[0,\sigma_x(\tau_x(y))]=\\ { } = \mu_{r^{n+1}(x)}^0\circ\sigma_{r^n(x)}\circ\ldots\circ\sigma_{r(x)}\circ\sigma_x[\sigma_x^{-1}(0),\tau_x(y)]=\\ = \mu_x^{n+1}[-r(x),\tau_x(y)]=\mu_x^{n+1}[-r(x),0]+\mu_x^{n+1}[0,\tau_x(y)].\end{gathered}\ ] ] recall also that @xmath213 , and that @xmath214 $ ] because of non - negativity of @xmath48 for @xmath215 $ ] . in consequence
, it suffices to establish the uniform continuity of the function @xmath216 $ ] on @xmath204 . in other words
, we have to prove that for each @xmath217 there exists @xmath218 such that for any @xmath219\setminus\frak{o}$ ] and @xmath220 satisfying @xmath221 it holds that @xmath222 here @xmath223 denotes the metric on @xmath18 naturally inherited from @xmath224 .
in particular , @xmath223 is shift - invariant .
observe also that we interpret @xmath225 as a point on @xmath18 .
set @xmath226-\mu_{x_2}^n[0,\tau_{x_2}(y_2)].$ ] since @xmath227 $ ] , we find that @xmath228 so let us take arbitrary @xmath217 , then there exists @xmath229 for which @xmath230 let @xmath231 denote open @xmath232 neighbourhood of the set @xmath233 .
due to corollary [ pi3 ] and lemma [ lmmcontt ] , the function is uniformly continuous for each @xmath234 .
we can choose @xmath235 to be small enough in order to have the closures of @xmath236 sets @xmath237 mutually disjoint .
obviously , each function from the finite set @xmath238 , @xmath239 , n = 0,1 , \dots , n_1\}$ ] , is uniformly continuous on @xmath240 .
take now @xmath241 small enough to assure that @xmath242 @xmath243 imply the inequality @xmath244 as well as \(i ) the existence of at most one integer @xmath245 $ ] such that @xmath246 ; \(ii ) @xmath247 once @xmath248 , @xmath249 .
a key observation is that in the case ( i ) the distance @xmath250 can not be large even when @xmath251 : @xmath252 thus , estimating separately the term @xmath253 ( whenever @xmath254 with properties described in ( i ) appears ) as @xmath255 , we obtain @xmath256 @xmath257 , which completes the proof .
the map @xmath258 commutes with @xmath38 , is uniformly continuous on @xmath259 and it admits a unique continuous extension @xmath183 on @xmath32 which also commutes with @xmath38 .
observe that all maps @xmath260 are bijective and @xmath261 .
first , we will prove that @xmath262 commutes with @xmath38 .
clearly , @xmath263 so that , for @xmath264 , @xmath265 but @xmath151 is injective on @xmath259 and therefore @xmath266 for each @xmath264 . hence , since @xmath267 are linear maps , @xmath268 is also uniformly continuous .
as the maps @xmath269 and @xmath270 are uniformly continuous and @xmath271 is dense in @xmath191\times\mathbb{t}^1 $ ] and @xmath272 is dense in @xmath273\times\mathbb{t}^1 $ ] , they can be uniquely continuously extended to the sets @xmath191\times\mathbb{t}^1 $ ] or @xmath273\times\mathbb{t}^1 $ ] , respectively .
since these maps coincide on the intersection @xmath274\times\mathbb{t}^1 ) \cap ( [ 1/2,0]\times\mathbb{t}^1)$ ] , they define a continuous self - map @xmath275 .
clearly , since @xmath258 commutes with @xmath38 , the set @xmath259 is dense in @xmath32 and the functions @xmath276 are continuous on @xmath277 , we obtain that @xmath266 for all @xmath278 .
similarly , @xmath279 on @xmath32 .
the map @xmath182 is minimal and non - invertible .
let @xmath280 be a non - empty compact and @xmath183-invariant set .
but then @xmath281 so , @xmath282 is also compact @xmath47-invariant set , which means @xmath283 because @xmath47 is minimal .
but since fibers are mapped to fibers and @xmath151 is bijective on @xmath259 , the set @xmath284 must contain whole @xmath259 .
but then @xmath285 and therefore @xmath183 is a minimal map . finally , we prove that @xmath183 is non - invertibile . for @xmath286 , consider the points @xmath287 > y_2=\mu_{x_1^\ast}[0,y_1^\ast]-\delta$ ] on the circle @xmath76 . for every @xmath288
, we have that @xmath289=\sum_{n=0}^\infty \frac{\mu_{r^n(x_1^\ast)}^0\sigma_x^n[0,y^\prime]}{2^{n+2}}+y^\prime/2=\sum_{n=1}^\infty \frac{\mu_{r^n(x_1^\ast)}^0\sigma_x^n[0,y^\prime]}{2^{n+2}}+y^\prime/2<\\ < \sum_{n=1}^\infty \frac{\mu_{r^n(x_1^\ast)}^0\sigma_x^n[0,y^*_1]}{2^{n+2}}+y^*_1/2 + 0.25-\delta= \sum_{n=0}^\infty \frac{\mu_{r^n(x_1^\ast)}^0\sigma_x^n[0,y^*_1]}{2^{n+2}}+y_1^\ast/2-\delta= y_2.\end{gathered}\ ] ] this yields immediately that @xmath290 and therefore @xmath291 since @xmath151 is invertible on the fiber over @xmath292 , we find that @xmath293 and therefore the non - degenerated interval @xmath294 $ ] is transformed by @xmath183 into a point .
so far we have obtained a non - invertible minimal self - map @xmath183 of the torus .
since @xmath183 commutes with @xmath38 , it induces a transformation @xmath295 of the klein bottle .
@xmath295 is a factor of @xmath183 by the fiber - preserving quotient map @xmath37 , so @xmath295 is a non - invertible minimal fiber - preserving transformation @xmath183 of the klein bottle .
the construction is completed . after changing the definition of @xmath84 ( where an appropriate weighted sum of _ all measures _
@xmath296 should be considered ) , we can similarly construct a minimal circle - fibered homeomorphism of the klein bottle having an asymptotic pair of points .
then the roberts - steenrod theory @xcite of monotone transformations of 2-dimensional manifolds can applied in order to obtain a different proof of theorem [ main ] , see @xcite for more detail .
the authors thank l-0.08cm39ubomr snoha and roman hric for valuable conversations .
the first author was supported by project sgs 2/2013 from the silesian university in opava .
support of this institution is gratefully acknowledged .
the second author was partially supported by fondecyt ( chile ) , project 1110309 .
this work was done during the second author sabbatical leave from the university of talca and his research stay at the mathematical institute of the silesian university in opava .
sergei trofimchuk acknowledges the hospitality of this mathematical institute . | applying the hric - jger blow up technique , we give an affirmative answer to the question about the existence of non - invertible minimal circle - fibered self - maps of the klein bottle .
in addition , we present a simpler construction of a non - invertible minimal self - map of two dimensional torus . |
charnonium is one of the simplest quark - antiquark composite particles .
charmonium physics has played an important role in the study of quantum chromodynamics ( qcd ) both perturbatively and nonperturbatively , since the first charmonium state @xmath2 was discovered in 1974 . during the past decade
, the study of charmonium has become more interesting because of the large difference between the predictions of the color - singlet model and the observations of @xmath2 and @xmath15 production at several experimental facilities e.g. at the fermilab tevatron @xcite .
the newly developed nonrelativistic qcd ( nrqcd ) factorization formalism @xcite allows the infrared safe calculation of inclusive heavy quarkonium production and decay rates . in the nrqcd production mechanism , a heavy quark - antiquark pair can be produced at short distances in a conventional color - singlet or a color - octet state , and then evolves into an observed quarkonium nonperturbatively . with this color - octet mechanism
, one may explain the tevatron data on the surplus production of @xmath16 and @xmath17 at large @xmath18 , though puzzles about their polarizations still remain ( for a recent review see @xcite and references therein ) . to further test the color octet mechanism ,
it is interesting to study the charmonium production in @xmath14 annihilation .
the @xmath2 inclusive production in @xmath14 annihilation has been investigated within the color - singlet model @xcite and the color - octet model @xcite .
the angular distribution and energy distribution of color - singlet @xmath2 production at @xmath19 have been discussed in @xcite . in @xcite
it is found that a clean signature of the color - octet mechanism may be observed in the angular distribution of @xmath2 production near the end point region . in @xcite contributions of various color - octet as well as color - singlet channels to the @xmath2 production cross sections are calculated in a wide range of @xmath14 collider energies .
moreover , the @xmath2 polarizations are predicted in @xcite .
recently , babar @xcite and belle @xcite have measured the direct @xmath2 production in continuum @xmath14 annihilations at @xmath19 .
the total cross section and the angular distribution seem to favor the nrqcd calculation over the color - singlet model @xcite , but some issues ( e.g. about the momentum distribution and polarization of @xmath2 ) still remain .
the situation has become even more complicated due to the very recent observation for the double @xmath20 production associated with @xmath2 by belle@xcite .
the measured exclusive cross section for @xmath21 process is an order of magnitude larger than the theoretical value@xcite , and the measured inclusive cross section for @xmath22 ( @xmath70.9 pb)@xcite is more than five times larger than nrqcd predictions which are only about 0.1 - 0.2 pb@xcite taking into account the differences in the values of the input parameters or methods . among other attempts to solve the @xmath23 inclusive production problem@xcite ,
@xmath5 annihilation into two photons was also studied@xcite , but the two photon contribution turned to be negligible at @xmath10gev , though it could prevail over one photon contribution at higher energies ( say , @xmath24gev)@xcite .
the double @xmath20 production associated with @xmath2 ( both exclusively and inclusively ) is very puzzling and needs a better understanding for both perturbative and nonperturbative qcd . on the other hand , experimentally , it is not clear whether the copious ( even dominant ) double @xmath20 production will also happen for charmonium states other than @xmath2 , such as @xmath1 and @xmath3(j=0 , 1 , 2 ) . among them
the @xmath8 and @xmath9 are more interesting since they have large branching fractions decaying into @xmath25 and might be easier to be detected .
in fact , in ref .
@xcite the inclusive production for @xmath8 and @xmath9 was searched for with the available integrated luminosity of about 30 fb@xmath26 at belle .
as more data are collected in the near future at b factories we hope that more accurate measurements for the p - wave and other s - wave charmonium states will be possible .
these measurements will be helpful to clarify the problems associated with @xmath2
double @xmath4 production . on the theoretical side ,
the calculations for inclusive s - wave and p - wave charmonia production via double @xmath4 are necessary in the framework of nrqcd , including both the color - singlet and color - octet contributions .
when we know the differences between nrqcd predictions and experimental data , we will have to further consider other mechanisms and methods in qcd to explain the differences .
this paper is organized as follows . in section 2
, we will calculate the complete @xmath27@xmath28 color - singlet inclusive production cross sections for @xmath1 and @xmath3(j=0,1,2 ) ( along with @xmath2 ) via double @xmath4 in @xmath5 annihilation through a virtual photon .
then we will compare the complete calculation with the calculation obtained in the charm quark fragmentation limit , and give their ratio as functions of the center - of - mass energies and determine the energy scales at which fragmentation approximations become reliable . in section 3 , we will further estimate the color - octet contributions to @xmath2 and @xmath3 inclusive cross sections via double charm .
finally , we summarize our results in section 4 .
the quarkonium can be described in term of fock states superposition within the nrqcd framework as follows @xmath29 > + o(v)q\bar{q}[^3p_j^{(8)}]g >
\nonumber\\&&+ o(v^2)q\bar{q}[^1s_0^{(8)}]g > + o(v^2)q\bar{q}[^3s_1^{(1,8)}]gg > + \ldots,\end{aligned}\ ] ] @xmath30 > + o(v)q\bar{q}[^3s_1^{(8)}]g > \nonumber\\&&+ o(v^2)q\bar{q}[^3p_j^{(1,8)}]gg > + \ldots,\end{aligned}\ ] ] where the superscript 1 or 8 labels the color configuration of the fock components . following the nonrelativistic qcd ( nrqcd ) factorization formalism , the scattering amplitude of the process @xmath31 in fig .
[ feynman ] and fig . [ fey2 ] is given by @xmath32 where @xmath33 is the intermediate @xmath4 pair produced at short distance , which subsequently evolves into a specific charmonium state at long distance , @xmath34 is the derivative of the amplitude with respect to the relative momentum between the quark and anti - quark in the bound state . for the case of color - singlet state
, the coefficient @xmath35 can be related to the origin of the radial wave function ( or its derivative ) of the bound state as @xmath36 the spin projection operator can be defined as@xcite @xmath37 we list the spin projection operators and their derivatives with respect to the relative momentum , which we will use in the calculations , as @xmath38 @xmath39 @xmath40.\ ] ] for p - wave states we need further relations to reduce the polarizations @xmath41 where m is the mass of the charmonium , which equals to @xmath42 in the nonrelativistic approximation .
the calculation of cross sections for @xmath43 charmonium @xmath44 is straightforward . using the definition in ref .
@xcite we get the differential cross section as follows @xmath45 where l = s for s - wave , l = p for p - wave and @xmath46 .
the functions @xmath47 and @xmath48 for different charmonium states are given in the appendix .
with eq .
( [ cross ] ) we can evaluate the inclusive cross sections for @xmath1 , @xmath2 and @xmath3 .
the input parameters used in the numerical calculations are@xcite @xmath49 @xmath50 now we give the numerical results at the belle and babar energy @xmath10 gev .
@xmath51 @xmath52 @xmath53 @xmath54 @xmath55 the @xmath2 production rate is in agreement with other references@xcite after taking into account the differences in the values of the input parameters . in the @xmath56 limit , where @xmath57 is defined as @xmath58
, the approximate fragmentation results will be equivalent to the complete calculations .
this is another check for the validity of the complete calculation . here
the fragmentation cross sections are written as @xmath59 where @xmath60 are the charm quark fragmentation functions into s - wave@xcite or p - wave@xcite charmonia .
the cross sections obtained in the complete calculation and in the fragmentation approximation as functions of the center - of - mass energies are plotted in fig .
[ fig1]-[fig5 ] .
all these cross sections are in units of @xmath61 , the cross section for @xmath62 annihilating into the @xmath4 quark pair , times @xmath63 .
one can find that the cross sections in complete calculations and fragmentation approximations ( all in units of the cross section for @xmath62 annihilating to the @xmath4 pair ) are proportional to the fragmentation probabilities for the charm quark fragmentating into charmonia when the @xmath64 limit is valid .
this is just what the fragmentation approach describes .
the results in these figures show that except for @xmath65 , the differences between fragmentation results and complete calculations are large at low energies . at the belle and babar energy @xmath6=10.6gev ,
the ratios of complete calculations to fragmentation calculations are @xmath66 for @xmath1 , @xmath2 , @xmath8 , and @xmath9 respectively . as the center - of - mass energy increases , the ratios of complete calculations to fragmentation results increase and can reach to @xmath67 when the center - of - mass energy is over 100 gev . moreover , the cross sections are rather sensitive to the input parameters . if we choose @xmath68 , and @xmath69 at @xmath10gev , the cross sections for @xmath70 , and @xmath3 ( j=0,1,2 ) become 77.0fb ,
192fb , 64.2fb , 18.3fb , 8.48fb respectively .
we next estimate the color - octet contribution to @xmath2 and @xmath3 production via double @xmath4 in @xmath14 annihilation .
the feynman diagrams are showed in fig [ feynman ] and fig [ fey2 ] . in fig .
[ feynman ] , the charmonium comes from the color - octet mediate states @xmath71 by emitting soft gluons at long distances . here
the color - octet contribution can be obtained from the corresponding color - singlet contribution divided by a factor of @xmath72 .
the matrix elements @xmath73 can be extracted from the tevatron data for @xmath2 and @xmath3 production ( see ref .
@xcite for detailed discussions ) .
accordingly , with some unavoidable uncertainties we set them as @xmath74 @xmath75 @xmath76 @xmath77 @xmath78 @xmath79 with these values of the matrix elements , the color - octet contributions to @xmath2 and @xmath3 in fig . [ feynman ] are about at least two orders of magnitude smaller than the color - singlet contributions , and therefore are negligible . in fig . [ fey2 ]
the color - octet contributions come from four different ( the upper two and the lower two ) diagrams . with their contributions ( including the interference terms ) , the differential cross section reads @xmath80 where @xmath81 takes the form @xmath82 + \nonumber\\ & & ( z-2)^2[3\delta^4 - 8\delta^2(3z-4)+32(2+z(z-2))]\nonumber\\ & & { \rm ln}[\frac{z \sqrt{4 + \delta^2 - 4 z}+2\sqrt{(1-z)(z^2-\delta^2)}}{z \sqrt{4 + \delta^2 - 4 z}-2\sqrt{(1-z)(z^2-\delta^2)}}]\}.\end{aligned}\ ] ] the numerical results can be obtained by using the parameters given above , and are @xmath83 @xmath84 the color - octet contribution to @xmath2 is only @xmath12 of the color - singlet cross section . for @xmath8 ,
the color - octet contribution is significant , which is about @xmath85 of the color - singlet cross section . with the approximation of heavy quark spin symmetry ,
the contributions of color octet @xmath86 to @xmath3 ( from color - octet @xmath87 mediate state to color - singlet @xmath88 final state by e1 transition ) satisfy the ratio @xmath89 for @xmath90 respectively .
their values are given by @xmath91 @xmath92 we show the angular distribution and energy distribution for @xmath8 in fig [ ang ] and fig [ z3 ] .
one can see that the color - octet contribution enhances the differential cross section significantly in the low energy ( small @xmath93 ) region .
in summary , we have calculated the complete @xmath94 inclusive production cross sections for @xmath1 , @xmath2 , and @xmath3(j=0 , 1 , 2 ) plus @xmath4 in @xmath5 annihilation through a virtual photon .
we consider both color - singlet and color - octet contributions , and give the analytical expressions for these cross sections .
the complete color - singlet calculations are compared with the approximate fragmentation calculations as functions of the center - of - mass energy @xmath6 .
we find that most of the fragmentation results substantially overestimate the cross sections ( e.g. by a factor of @xmath74 for @xmath8 and @xmath9 ) at the belle and babar energy @xmath10gev .
the fragmentation results become a good approximation only when @xmath6 is higher than about @xmath11gev .
we further calculated the color - octet contributions to these cross sections with analytical expressions .
we find that while the color - octet contribution to @xmath2 inclusive production via double charm is negligible ( only about @xmath12 ) , the color - octet contributions to @xmath8 and @xmath9 can be significant .
these results may serve as nrqcd predictions to compare with the experimental data observed or to be observed at belle and babar .
the authors thank z.z . song for useful discussions . this work was supported in part by the national natural science foundation of china , and the education ministry of china .
in this appendix , we give the functions of s and @xmath95 which are defined in eq .
( [ cross ] ) .
@xmath96 \nonumber \\ & & -3\delta^{2}(z-2)^{4}[8\delta^{6}(2+\delta^{2})-96\delta^{6}z-2\delta^{2}(192 - 48\delta^{2}+8\delta^{4}-\delta^{6})z^{2 } \nonumber \\ & & + 16\delta^{2}(8 + 6\delta^{2}-\delta^{4})z^{3}+\delta^{2}(192 + 40\delta^{2}-\delta^{4})z^{4 } + 8(32 - 4\delta^{2}+\delta^{4})z^{5 } \nonumber\\ & & -8(48+\delta^{2})z^{6}]\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] @xmath97 \nonumber \\ & & + 3\delta^{2}(z-2)^{4}[8\delta^{6}(6+\delta^{2})-32\delta^{6}z -2\delta^{2}(64 + 48\delta^{2}+16\delta^{4}-\delta^{6})z^{2 } \nonumber \\ & & + 16\delta^{2}(12-\delta^{2})(2+\delta^{2})z^{3}-(1024 + 320\delta^{2}-88\delta^{4}-\delta^{4})z^{4}\nonumber \\ & & + 8(96 - 28\delta^{2}-\delta^{4})z^{5}+8(16+\delta^{2})z^{6}]\nonumber\\ & & \times\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] @xmath98 \nonumber \\ & & -\delta^{2}(z-2)^{4}[8\delta^{4}(48 + 22\delta^{2}+3\delta^{4})-32\delta^{4}(24 + 5\delta^{2})z \nonumber \\ & & -2\delta^{2}(448 + 16\delta^{2}+8\delta^{4}-3\delta^{6})z^{2}+16\delta^{2}(56 - 10\delta^{2}-5\delta^{4})z^{3 } \nonumber \\ & & + \delta^{2}(1152 + 272\delta^{2}-3\delta^{4})z^{4}+8(32 - 92\delta^{2}+5\delta^{4})z^{5}-56(16+\delta^{2})z^{6 } \nonumber \\ & & + 512z^{7}]\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] @xmath99 \nonumber \\ & & + \delta^{2}(z-2)^{4}[8\delta^{4}(16 + 2\delta^{2}+3\delta^{4})-32\delta^{4}(8-\delta^{2})z\nonumber\\ & & -2\delta^{2}(320 - 272\delta^{2}+64\delta^{4}-3\delta^{6})z^{2}+16\delta^{2}(40 - 54\delta^{2}-5\delta^{4})z^{3}\nonumber \\ & & -(1024 - 720\delta^{4}-3\delta^{6})z^{4}+8(96 - 36\delta^{2}- 5\delta^{4})z^{5 } \nonumber \\ & & + 8(80 + 7\delta^{2})z^{6}-512z^{7}]\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] @xmath100 \nonumber \\ & & + 3\delta^{2}(z-2)^{4}[-192\delta^{10}+96\delta^{8}(26 + 3\delta^{2})z -64\delta^{6}(160 + 75\delta^{2}+3\delta^{4})z^{2}\nonumber\\ & & -16\delta^{4}(576 - 1664\delta^{2}-183\delta^{4}-2\delta^{6})z^{3}\nonumber\\ & & -4\delta^{2}(1536 - 4608\delta^{2}+4016\delta^{4}+152\delta^{6}+5\delta^{8})z^{4}\nonumber\\ & & + 2\delta^{2}(11264 - 23424\delta^{2}-160\delta^{4}+106\delta^{6}+3\delta^{8})z^{5}\nonumber\\ & & + 4(2048 - 4224\delta^{2}+9952\delta^{4}+248\delta^{6}-27\delta^{8})z^{6}\nonumber\\ & & -(20480 - 22528\delta^{2}+5312\delta^{4}-368\delta^{6}+3\delta^{8})z^{7}\nonumber\\ & & + 4(4096 - 6496\delta^{2}-600\delta^{4}+17\delta^{6})z^{8}-16(320 - 472\delta^{2}+7\delta^{4})z^{9}\nonumber \\ & & + 32(48+\delta^{2})z^{10}]\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] @xmath101 \nonumber \\ & & + 3\delta^{2}(z-2)^{4}[192\delta^{10}-288\delta^{8}(10+\delta^{2})z + 64\delta^{6}(224 + 59\delta^{2}+3\delta^{4})z^{2}\nonumber\\ & & + 16\delta^{4}(192 - 1248\delta^{2}-121\delta^{4}-2\delta^{6})z^{3}\nonumber\\ & & + 4\delta^{2}(512 - 12288\delta^{2}+2384\delta^{4}-56\delta^{6}+5\delta^{8})z^{4}\nonumber\\ & & + 2\delta^{2}(3072 + 35968\delta^{2}-160\delta^{4}+50\delta^{6}-3\delta^{8})z^{5}\nonumber\\ & & -4(2048 - 4992\delta^{2}+8224\delta^{4}-408\delta^{6}-23\delta^{8})z^{6}\nonumber\\ & & + ( 12288 - 51200\delta^{2}-3008\delta^{4}-1456\delta^{6}-3\delta^{8})z^{7}\nonumber\\ & & -4(2048 - 8992\delta^{2}-1224\delta^{4}-17\delta^{6})z^{8}+16(192 - 616\delta^{2}-7\delta^{4})z^{9}\nonumber\\ & & + 32(16+\delta^{2})z^{10}]\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] @xmath102\nonumber \\ & & -3\delta^{2}(z-2)^{4}[-192\delta^{8}(3+\delta^{2})+96\delta^{6}(48 + 28\delta^{2}+3\delta^{4})z\nonumber\\ & & -16\delta^{4}(768 + 808\delta^{2}+217\delta^{4}+12\delta^{6})z^{2}\nonumber\\ & & + 16\delta^{4}(1600 + 652\delta^{2}+105\delta^{4}+2\delta^{6})z^{3}\nonumber\\ & & + 4\delta^{2}(7168 - 4352\delta^{2}-360\delta^{4}-59\delta^{6}-5\delta^{8})z^{4}\nonumber\\ & & -2\delta^{2}(24576 - 3968\delta^{2}+1024\delta^{4}-64\delta^{6}-3\delta^{8})z^{5}\nonumber\\ & & + \delta^{2}(17408 - 7296\delta^{2}+136\delta^{4}-51\delta^{6})z^{6}\nonumber\\ & & -(8192 - 12800\delta^{2}-8576\delta^{4}-300\delta^{6}+3\delta^{8})z^{7}\nonumber\\ & & + 2(8192 - 6656\delta^{2}-1328\delta^{4}+17\delta^{6})z^{8}-128(80 - 10\delta^{2}+\delta^{4})z^{9}\nonumber\\ & & + 128(24 + 5\delta^{2})z^{10}]\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] @xmath103\nonumber \\ & & + 3\delta^{2}(z-2)^{4}[-192\delta^{8}(1+\delta^{2})+96\delta^{6}(16 + 12\delta^{2}+3\delta^{4})z\nonumber\\ & & -16\delta^{4}(256 + 344\delta^{2}+59\delta^{4}+12\delta^{6})z^{2 } + 16\delta^{4}(448 + 404\delta^{2}-9\delta^{4}+2\delta^{6})z^{3}\nonumber\\ & & + 4\delta^{2}(5120 + 1792\delta^{2}-856\delta^{4}+135\delta^{6}-5\delta^{8})z^{4}\nonumber\\ & & -2\delta^{2}(16384 + 6016\delta^{2}-1088\delta^{4}+24\delta^{6}-3\delta^{8})z^{5}\nonumber\\ & & + ( 32768 - 25600\delta^{2}-7040\delta^{4}-1864\delta^{6}-57\delta^{8})z^{6}\nonumber\\ & & -(57344 - 81408\delta^{2}-11904\delta^{4}-884\delta^{6}-3\delta^{8})z^{7}\nonumber\\ & & + 2(16384 - 27648\delta^{2}-2384\delta^{4}-17\delta^{6})z^{8}-128(48 - 118\delta^{2}-\delta^{4})z^{9}\nonumber\\ & & -128(8 + 5\delta^{2})z^{10}]\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] @xmath104\nonumber \\ & & -3\delta^{2}(z-2)^{4}[-192\delta^{6}(144 + 57\delta^{2}+5\delta^{4 } ) + 96\delta^{6}(1008 + 304\delta^{2}+15\delta^{4})z\nonumber\\ & & + 16\delta^{4}(4224 - 6392\delta^{2}-1731\delta^{4}-60\delta^{6})z^{2}\nonumber\\ & & -16\delta^{4}(13632 - 916\delta^{2}-705\delta^{4}-10\delta^{6})z^{3}\nonumber\\ & & -4\delta^{2}(15360 - 56448\delta^{2}-8648\delta^{4}+433\delta^{6}+25\delta^{8})z^{4}\nonumber\\ & & + 2\delta^{2}(96256 - 29568\delta^{2}-13280\delta^{4}-340\delta^{6}+15\delta^{8})z^{5}\nonumber\\ & & + ( 16384 - 193536\delta^{2}-28672\delta^{4}+14680\delta^{6}+399\delta^{8})z^{6}\nonumber\\ & & -5(8192 - 11776\delta^{2}-3712\delta^{4}+604\delta^{6}+3\delta^{8})z^{7}\nonumber\\ & & + 2(10240 + 9728\delta^{2}-2784\delta^{4}-79\delta^{6})z^{8}+512(4 - 47\delta^{2}+2\delta^{4})z^{9}\nonumber\\ & & + 256(12 + 19\delta^{2})z^{10}]\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] @xmath105\nonumber \\ & & -3\delta^{2}(z-2)^{4}[-192\delta^{6}(48 + 3\delta^{2}+5\delta^{4 } ) + 288\delta^{6}(112 + 8\delta^{2}+5\delta^{4})z\nonumber\\ & & + 16\delta^{4}(1920 - 3784\delta^{2}+71\delta^{4}-60\delta^{6})z^{2}\nonumber\\ & & -16\delta^{4}(6336 - 3660\delta^{2}+289\delta^{4}-10\delta^{6})z^{3}\nonumber\\ & & -4\delta^{2}(13312 - 43392\delta^{2}+9928\delta^{4}-805\delta^{6}+25\delta^{8})z^{4}\nonumber\\ & & + 2\delta^{2}(92160 - 76928\delta^{2}+6944\delta^{4}-676\delta^{6}+15\delta^{8})z^{5}\nonumber\\ & & + ( 16384 - 322560\delta^{2}+121856\delta^{4}+10920\delta^{6}+365\delta^{8})z^{6}\nonumber\\ & & -(24576 - 248320\delta^{2}+96896\delta^{4}+5332\delta^{6}-15\delta^{8})z^{7}\nonumber\\ & & + 2(14336 - 40960\delta^{2}+12768\delta^{4}+79\delta^{6})z^{8}-512(36 - 55\delta^{2}+2\delta^{4})z^{9}\nonumber\\ & & -256(4 + 19\delta^{2})z^{10}]\ln\frac{z\sqrt{4+\delta^{2}-4z}+2\sqrt{(1-z)(z^{2}-\delta^{2 } ) } } { z\sqrt{4+\delta^{2}-4z}-2\sqrt{(1-z)(z^{2}-\delta^{2})}}\}.\end{aligned}\ ] ] j.h .
khn , j. kaplan , and e.g.o .
sadiani , nucl . phys .
* b 157 * , 125 ( 1979 ) ; w.y .
keung , phys . rev . * d 23 * , 2072 ( 1981 ) ; j.h .
khn and h. schneider , phys . rev .
* d 24 * , 2996 ( 1981 ) ; j.h .
khn and h. schneider , z. phys .
* c 11*,263 ( 1981 ) ; l. clavelli , phys . rev .
* d26 * , 1610 ( 1982 ) .
k@xmath107hn , j. kaolan and e.g.o .
safiani , nucl . phys .
* b157 * , 125 ( 1979 ) ; b. guberina , j.h .
k@xmath108hn , r.d .
peccei and r. r@xmath107ckl , nucl . phys . *
b174 * , 317(1980);p .
cho and a.k .
leibovich phys .
* d53 * , 150(1996 ) . plotted against the center - of - mass energy .
dotted line illustrates the fragmentation calculation and solid line illustrates the complete calculation .
the cross sections are in units of @xmath61 times @xmath63.,width=529,height=302 ] | motivated by the recent observation of double charm quark pair production by the belle collaboration , we calculate the complete @xmath0 inclusive production cross sections for @xmath1 , @xmath2 , and @xmath3(j=0 , 1 , 2 ) plus @xmath4 in @xmath5 annihilation through a virtual photon .
we consider both color - singlet and color - octet contributions , and give the analytical expressions for these cross sections .
the complete color - singlet calculations are compared with the approximate fragmentation calculations as functions of the center - of - mass energy @xmath6 .
we find that most of the fragmentation results substantially overestimate the cross sections ( e.g. by a factor of @xmath74 for @xmath8 and @xmath9 ) at the belle and babar energy @xmath10gev .
the fragmentation results become a good approximation only when @xmath6 is higher than about @xmath11gev .
we further calculate the color - octet contributions to these cross sections with analytical expressions .
we find that while the color - octet contribution to @xmath2 inclusive production via double charm is negligible ( only about @xmath12 ) , the color - octet contributions to @xmath8 and @xmath9 can be significant .
pacs number(s ) : 12.40.nn , 13.85.ni , 14.40.gx inclusive charmonium production via double @xmath13 in @xmath14 annihilation + kui - yong liu department of physics , peking university , beijing 100871 , people s republic of china and department of physics , liaoning university , shenyang 110036 , people s republic of china + zhi - guo he department of physics , peking university , beijing 100871 , people s republic of china + kuang - ta chao china center of advanced science and technology ( world laboratory ) , beijing 100080 , people s republic of china and department of physics , peking university , beijing 100871 , people s republic of china |
( color online ) field @xmath4 and temperature @xmath1 dependence of the amplitude @xmath39 of the normal - state torque of cecoin@xmath0 single crystals . ( a ) @xmath4 dependence of @xmath39 at measured temperatures of 1.9 , 3 , 4.5 , 5 , 6 , 8 , and 10 k. ( b ) @xmath1 dependence of @xmath39 at measured magnetic field of 14 t. inset : angular @xmath26 dependent torque @xmath78 measured in the normal state at @xmath79 k and 7 t in cecoin@xmath0 single crystals .
the solid line is a fit of the data with @xmath80 .
( a ) angular @xmath26 dependence of the reversible torque @xmath49 measured at 1.9 k and 3 t on cecoin@xmath0 single crystals .
the solid line is a fit of the data with @xmath81 .
( b ) angular dependence of the irreversible torque @xmath50 measured at 1.9 k and 1 t on cecoin@xmath0 single crystals .
sharp peaks are present at @xmath69 , @xmath76 , @xmath77 and @xmath82 .
plot of the angular @xmath26 dependence of the upper critical field @xmath60 and lower critical field @xmath83 for magnetic field parallel to the @xmath29 plane , magnetization @xmath63 , free energy @xmath65 , reversible torque @xmath49 , and irreversible torque @xmath50 for @xmath8 wave symmetry .
field @xmath4 dependence of the amplitude @xmath52 of the reversible torque in the mixed state of cecoin@xmath0 single crystals measured at @xmath84 k. the solid line is a guide to the eye .
inset : temperature @xmath1 dependence of @xmath52 measured at @xmath85 t. the solid line is a guide to the eye . | in - plane torque measurements were performed on heavy fermion cecoin@xmath0 single crystals in the temperature @xmath1 range @xmath2 k @xmath3 k and applied magnetic field @xmath4 up to 14 t. the normal - state torque is given by @xmath5 .
the reversible part of the mixed - state torque , obtained after subtracting the corresponding normal state torque , shows also a four - fold symmetry .
in addition , sharp peaks are present in the irreversible torque at angles of @xmath64 , 3@xmath7/4 , 5@xmath7/4 , 7@xmath7/4 , etc . both the four - fold symmetry in the reversible torque and the sharp peaks in the irreversible torque of the mixed state imply @xmath8 symmetry of the superconducting order parameter .
the field and temperature dependences of the reversible mixed - state torque provide further evidence for @xmath8 wave symmetry .
the four - fold symmetry in the normal state has a different origin since it has different field and temperature dependences than the one in the mixed state .
the possible reasons of the normal state four - fold symmetry are discussed .
the superconductivity in the heavy fermion superconductor cecoin@xmath0 is unconventional as exemplified by the non - fermi liquid behavior,@xcite the giant nernst effect present in the normal state , @xcite its proximity to quantum critical points , @xcite the pauli limiting effect , @xcite and the possible multiband picture in the superconducting state.@xcite unconventional superconductivity is always a subject of great interest .
knowing the pairing symmetry , which is related to the ground state and gap energy , is essential to the understanding of the pairing mechanism and the origin of superconductivity .
for a conventional superconductor , which is described by the bcs theory , the pairing is phonon mediated and the pairing symmetry is @xmath9 wave . for an unconventional superconductor , the quasiparticle gap vanishes at certain points on the fermi surface .
for example , most of the experimental evidence on high temperature superconductors , such as yba@xmath10cu@xmath11o@xmath12 , indicates that the @xmath13 wave symmetry dominates in these materials .
it has been established that the superconducting order parameter of cecoin@xmath0 displays @xmath14-wave symmetry .
@xmath15 in and @xmath16co nuclear magnetic resonance measurements@xcite and torque measurements@xcite have revealed a suppressed spin susceptibility , which implies singlet spin pairing .
a @xmath17 term is present in the low temperature @xmath1 specific heat , consistent with the presence of nodes in the superconducting energy gap.@xcite nuclear quadrupole resonance and nuclear magnetic resonance measurements on cecoin@xmath0 have revealed that the nuclear spin lattice relaxation rate @xmath18 has no hebel - slichter coherence peak just below the superconducting transition temperature @xmath19 , and it has a @xmath20 dependence at very low temperatures , which indicates the existence of line nodes in the superconducting energy gap.@xcite nevertheless , the direction of the gap nodes relative to the brillouin zone axes , which determines the type of @xmath14-wave state , namely @xmath13 or @xmath8 , is still an open question and an extremely controversial issue .
for example , angular dependent thermal conductivity measurements in a magnetic field have revealed four - fold symmetry consistent with @xmath13 symmetry.@xcite neutron scattering experiments by eskildsen et al . revealed a square lattice oriented along [ 110 ] ,
also consistent with @xmath13 wave symmetry .
@xcite however , field - angle - dependent specific heat measurements have found the symmetry of the superconducting gap to be @xmath8.@xcite furthermore , theoretical calculations by ikeda et al .
strongly suggest @xmath8 wave symmetry when taking into account the available @xmath21 experimental data of the fflo states@xcite [ @xmath21 is the temperature at which the upper critical field @xmath22 changes from second order to first order ] .
in contrast , recent calculations by tanaka et al .
@xcite based on the fermi liquid theory and by vorontsov et al .
@xcite based on a unified microscopic approach , support the @xmath13 gap symmetry .
this extremely controversial issue needs to be resolved through an experimental technique that allows the direct measurement of the nodal positions .
the experiments which are phase sensitive usually include surface or boundary effects , while the experiments which detect bulk properties are not phase sensitive . in the study presented here we use torque measurements to clarify the gap symmetry .
torque is a bulk measurement so it provides information on the order parameter of the bulk , not only the surface .
( the order parameter of the surface might be different from that of the bulk ) .
it also _ directly _ probes the nodal positions on the fermi surface with high angular resolution since torque is the angular derivative of the free energy .
hence , such an experimental technique is ideal to determine the direction of the gap nodes relative to the brillouin zone axes and , in fact , it has already been successfully used to identify the nodal positions of untwinned yba@xmath10cu@xmath11o@xmath12 single crystals@xcite and tl@xmath10ba@xmath10cuo@xmath23 thin films.@xcite in addition , recent theoretical calculations by h. adachi have shown that low - field torque measurements can be used to detect the nodal positions of a @xmath14-wave superconductor with a small fermi - surface anisotropy,@xcite as is the case of cecoin@xmath0 .
in - plane torque measurements were performed on single crystals of cecoin@xmath0 both in the normal state and in the mixed state .
normal - state torque shows a four - fold symmetry .
the reversible part of the angular dependent mixed - state torque data , obtained after subtracting the corresponding normal state torque , also shows a four - fold symmetry with a positive coefficient .
the @xmath1 and @xmath4 dependences of the coefficients of the four - fold torques in the normal and mixed states are different . hence
, these two four - fold symmetries have clearly different origin .
sharp peaks in the irreversible torque data were observed at angles equal with @xmath64 , 3@xmath7/4 , 5@xmath7/4 , and 7@xmath7/4 , etc .
the symmetry of the free energy extracted from the four - fold symmetry of the reversible torque of the mixed state coupled with the position of the sharp peaks in the irreversible torque of the mixed state point unambiguously towards @xmath8 wave symmetry in cecoin@xmath0 .
further support for the @xmath8 wave symmetry is provided by the @xmath4 and @xmath1 dependence of the mixed - state reversible torque .
single crystals of cecoin@xmath0 were grown using the flux method .
high quality crystals with regular shape and shiny surfaces were chosen to carry out the torque measurements .
the single crystals were etched in concentrated hcl for several hours to remove the indium left on the surface during the growth process .
the crystals were then rinsed thoroughly in ethanol .
a piezoresistive torque magnetometer was used to measure the angular dependence of the in - plane torque of cecoin@xmath0 both in the normal state and mixed state .
the torque was measured over a large temperature range ( 1.8 k @xmath24 10 k ) and magnetic field @xmath4 range ( 1.5 t @xmath25 14 t ) by rotating the single crystal in fixed magnetic field .
the angle @xmath26 for the in - plane rotation was defined as the angle made by the field with the @xmath27-axis of the single crystal .
the contributions of the gravity and puck to the total torque signal were measured and subtracted from it as discussed elsewhere.@xcite the experiments were carried out in a physical property measurement system ( ppms ) . in such a system with a one axis rotator , it is very difficult to ensure an in - plane alignment of better than about @xmath28 degrees .
if misalignment exists , i.e. the magnetic field is not completely within the @xmath29 plane of the single crystal , there should be a @xmath30 term [ see eq .
( 6 ) of ref .
@xcite ] .
indeed , the angular - dependent in - plane torque signal has a @xmath30 term [ @xmath31 and @xmath26 of eq .
( 6 ) are @xmath26 and @xmath32 , respectively , in the present case ] in addition to the @xmath33 term .
in fact , the amplitude of the measured @xmath30 term gives a misalignment @xmath34 .
so , we attribute this @xmath30 term to the misalignment of the single crystal .
the torque data shown in this paper are after subtracting this @xmath30 term .
previously , we have shown that the @xmath35-axis rotation torque signal measured in the mixed state has a paramagnetic component which is comparable with the diamagnetic component.@xcite the former component is a result of the anisotropy of the susceptibilities along the @xmath27 and @xmath36 axes .
therefore , such a paramagnetic torque signal is absent in the present measurements in which the torque is measured while rotating the single crystal along the @xmath36 axis , since @xmath37 .
nevertheless , @xmath1 and @xmath4 dependent torque measurements in the normal state reveal that the normal - state torque signal is not negligible , is reversible , and it has a four - fould symmetry [ see inset to fig .
1(b ) ] . the solid line is a fit of the data with @xmath38 .
the field and temperature dependence of the amplitude @xmath39 , gives the @xmath4 and @xmath1 dependence of the normal - state torque .
the coefficient @xmath39 has an @xmath40 dependence up to 14 t for all measured temperatures from 1.9 k to 10 k [ see fig .
1(a ) ] . as the temperature increases , the slope of the plots in fig .
1(a ) decreases .
this is consistent with the temperature dependence of @xmath41 shown in fig .
1(b ) ; i.e. , @xmath41 increases , hence @xmath39 decreases , with increasing @xmath1 . a straight line fit of the data with @xmath42 gives @xmath43 k , which has a value close to the single ion kondo temperature @xmath44 [ reported to be between 1 and 2 k , ref.@xcite ) ] and @xmath45 t@xmath46n@xmath47m@xmath47 .
hence , @xmath48 .
therefore , in approaching the superconducting transition , @xmath39 decreases with decreasing @xmath4 and increases with decreasing @xmath1 .
we subtract the corresponding normal - state torque from the torque measured in the superconducting state .
the torque data measured in the mixed state of cecoin@xmath0 , obtained by subtracting the corresponding normal - state torque from the measured torque , have both reversible and irreversible components . the reversible torque @xmath49 is the average of the torque data measured in clockwise and anti - clockwise directions , while the irreversible torque @xmath50 is the average of the antisymmetric components of the torque data measured in clockwise and anti - clockwise directions .
figure 2(a ) shows the reversible part of the angular dependent in - plane torque data measured in the mixed state at 1.9 k and in a magnetic field of 3 t. clearly , there is a four - fold symmetry present in the torque data , although the data points are somewhat scattered .
the solid line is a fit of these mixed - state data with @xmath51 .
the coefficient @xmath52 is positive since the torque displays a maximum at @xmath53 .
figure 2(b ) shows the irreversible part of the mixed - state torque data @xmath54 measured at @xmath55 k and @xmath56 t for a single crystal with a mass of 2.6 mg .
the @xmath54 data have sharp peaks at @xmath64 , 3@xmath7/4 , 5@xmath7/4 , and 7@xmath7/4 , etc .
the nodal positions of cecoin@xmath0 can be obtained from the reversible and irreversible mixed - state torque data , as previously done in the study of yba@xmath10cu@xmath11o@xmath57.@xcite specifically , theoretical calculations predict that the in - plane upper critical field @xmath58 has a four - fold symmetry for a @xmath14-wave superconductor.@xcite in the case of @xmath8 wave symmetry , the angular variation of the upper critical field @xmath59 ; hence , it has maxima at @xmath64 , 3@xmath7/4 , 5@xmath7/4 , 7@xmath7/4 , etc .
figure 3 shows the angular dependence of the reversible and irreversible torque obtained by starting from this angular dependence of @xmath60 , as follows .
the lower critical field @xmath61 is out of phase with @xmath58 ( see fig .
3 ) since the thermodynamic critical field @xmath62 is independent of the magnetic field orientation . therefore , the magnetization @xmath63 , given by m@xmath64 , has the same angular dependence as @xmath60 ( see fig .
3 ) .
the easy axis of magnetization ( maximum magnetization ) should correspond to free energy @xmath65 minima .
this implies that , for the @xmath8 symmetry , @xmath65 has minima at @xmath64 , 3@xmath7/4 , 5@xmath7/4 , 7@xmath7/4 , etc .
( see fig .
3 ) . the torque is the angular derivative of the free energy @xmath65 ; i.e. , @xmath66 .
hence , the reversible torque data for a material with @xmath8 wave symmetry should display a four - fold symmetry with maxima at @xmath53 , @xmath67 , @xmath68 , etc ( see fig .
3 ) . also ,
the free energy minima act as intrinsic pinning centers for vortices , so the irreversible torque data for a material with @xmath8 wave symmetry should display peaks at the same angles at which the free energy has minima ; i.e. , at @xmath69 , 3@xmath7/4 , 5@xmath7/4 , 7@xmath7/4 , etc .
notice that the angular dependence of reversible and irreversible torque data of cecoin@xmath0 shown in figs .
2(a ) and 2(b ) is the same as the @xmath26 dependence of @xmath49 and @xmath50 , respectively , shown in fig .
3 , obtained from the theoretically predicted angular dependence of the upper critical fields for a material with @xmath8 symmetry .
therefore , the reversible along with the irreversible torque data in the mixed state unambiguously imply that the wave symmetry of cecoin@xmath0 is @xmath8 . to further understand the four - fold symmetry displayed by the present torque measurements in the mixed state
, we studied the field and temperature dependence of the amplitude @xmath52 .
figure 4 is a plot of the @xmath4 dependence of @xmath52 , which gives the @xmath4 dependence of the torque , obtained by fitting the angular dependent torque data measured in different magnetic fields .
note that @xmath70 increases with increasing @xmath4 , reaches a maximum , and then decreases with further increasing @xmath4 .
also note that the four - fold symmetry vanishes close to @xmath58 ( @xmath71 t ) .
this field dependence of the magnitude of @xmath52 is the same as the field dependence of the basal - plane reversible torque in the mixed state of a layered @xmath72 wave superconductor [ see fig .
8(a ) of ref . @xcite ] calculated by adachi et al.@xcite based on the quasiclassical version of the bcs - gorkov theory with a fermi surface which is isotropic within the basal plane .
the sign difference between the data of the present fig .
4 and fig .
8(a ) of ref .
@xcite , which is for a @xmath72 wave symmetry , further indicates that the present data reflect @xmath8 symmetry since the torque data have opposite signs for the @xmath8 and @xmath72 wave symmetries .
the inset to fig .
4 is a plot of the temperature dependence of the amplitude @xmath52 .
note that @xmath52 , hence the torque , decreases with increasing @xmath1 and vanishes towards @xmath19 .
the fact that both the @xmath1 and @xmath4 dependences of the reversible mixed - state torque vanish at the superconducting - normal state phase boundary further indicates that the observed four - fold symmetry is related with superconductivity ; hence , it reflects the gap symmetry .
we note that the behaviors of @xmath73 and @xmath74 are totally different ( compare figs . 1 and 4 ) .
so the four - fold symmetries present in normal and mixed states have different origin .
the origin of the four - fold symmetry in the normal state is not yet clear to us .
torque measurements on lacoin@xmath0 single crystals , which also have a tetragonal structure but are not superconducting and the @xmath75 electrons are absent , give some clues on the normal state four - fold symmetry of cecoin@xmath0 .
the angular dependent torque data for lacoin@xmath0 are shown in fig .
5 . clearly , the @xmath33 symmetry observed in cecoin@xmath0 is completely absent here .
the difference in the normal - state torque between cecoin@xmath0 and lacoin@xmath0 could be due to the presence of heavy electrons in the former compound and their absence in the latter one .
the crystalline electric field , which is important in heavy fermion systems , might be responsible for the four - fold symmetry in the normal - state torque .
also , the field induced order , possibly quadrupolar order , could be another reason for the four - fold symmetry in the normal - state torque of cecoin@xmath0 .
no doubt , the origin of this normal - state four - fold symmetry present in the torque data requires further study .
nevertheless , this is beyond the scope of this paper . in - plane angular dependent torque measurements
were performed on cecoin@xmath0 single crystals both in the normal and mixed states .
normal - state torque measurements show a four - fold symmetry .
the reversible torque in the mixed state , obtained after subtracting the corresponding normal state contribution , also shows a four - fold symmetry with maxima at @xmath53 , @xmath67 , @xmath68 , etc .
sharp peaks in the irreversible torque data were observed at @xmath69 , @xmath76 , @xmath77 , etc .
these latter peaks correspond to minima in the free energy of a @xmath8 wave symmetry .
the mixed state four - fold symmetry and the peak positions in the irreversible torque point unambiguously towards @xmath8 wave symmetry of the superconducting gap .
the field and temperature dependences of the amplitude of the normal - state torque is different from that of the mixed - state torque , which indicates that the normal - state torque has a different origin .
+ + * acknowledgments * the authors would like to thank professor k. machida for fruitful discussions .
this research was supported by the national science foundation under grant no .
dmr-0406471 at ksu and the us department of energy under grant no .
de - fg02 - 04er46105 at ucsd .
[ ] + 28 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , , , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , , , * * , ( ) . , , , , , , , , * * , ( ) . , , , , , , , , , * * , ( ) . , , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , , , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , , , , , , , ,
* * , ( ) . , , , , , * * , ( ) .
, , , , , , , * * , ( ) . , , , , , ( ) . , , , , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , , * * , ( ) . , , , , , , , * * , ( ) . , * * , ( ) . , , , , * * , ( ) . , * * , ( ) . , , , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , , , , , , * * , ( ) . ,
* * , ( ) . |
for conformally compact manifolds that are hyperbolic near infinity , we now have fairly good control over the growth of the resonance counting function .
upper and lower bounds have been obtained for various cases in @xcite . in this paper
we develop techniques which can provide a sharp constant for the upper bound , and apply these specifically to the case where the manifold is a compactly supported perturbation of the hyperbolic space @xmath1 .
the techniques are inspired by stefanov s recent proof of sharp upper bounds on the resonance counting function for perturbations of the euclidean laplacian @xcite .
let @xmath2 denote the positive laplacian on @xmath1 .
we can write the green s function associated to @xmath2 explicitly : if @xmath3 , then @xmath4 where @xmath5 is the gauss hypergeometric function and @xmath6 . from this expression
we quickly deduce that @xmath7 admits an analytic extension to @xmath8 if @xmath9 is even , and a meromorphic extension with poles at @xmath10 for @xmath11 if @xmath9 is odd . in the latter case
the multiplicities of the poles are given by @xmath12 let @xmath13 denote the resonance set for @xmath1 ( empty for @xmath9 even ) , with resonances repeated according to multiplicity .
the associated resonance counting function is defined by @xmath14 for @xmath9 odd , an asymptotic for @xmath15 is easily deduced by integrating ( [ m0.k ] ) . for later usage , we introduce the constant @xmath16 the resonance counting function asymptotics for @xmath1 are then summarized by @xmath17 as @xmath18 .
the main result of this paper concerns the resonance counting function @xmath19 for @xmath20 a compactly supported perturbation of @xmath2 . to describe the class of perturbations
precisely , let @xmath21 for some @xmath22 .
we assume that @xmath23 is a smooth riemannian manifold , possible with boundary , such that for some compact @xmath24 , we have @xmath25 in other words , @xmath23 agrees with @xmath1 near infinity .
note that @xmath26 is allowed to have a more complicated topology than @xmath1 , as illustrated in figure [ xg ] .
, with @xmath27 replacing the closed ball @xmath28 .
] let @xmath29 denote the laplacian on @xmath23 , and @xmath30 with @xmath31 .
we then define the perturbed operator @xmath32 where some self - adjoint boundary condition is imposed if @xmath26 has a boundary . since @xmath7 functions as a good parametrix for @xmath33 near infinity , it is straightforward to prove meromorphic continuation of @xmath34 .
we can thus define the resonance set @xmath35 , with resonances repeated according to multiplicity , and the associated counting function @xmath36 the arguments of cuevas - vodev @xcite and borthwick @xcite are easily extended to show that @xmath37 our goal in this paper is to refine this estimate by producing an explicit constant @xmath38 for this bound , which is sharp in the sense that @xmath39 holds in at least some cases . as in stefanov s work @xcite , such a result requires a slightly regularized version of the counting function .
the basis of our estimate is the following relative counting formula : @xmath40 with @xmath41 the relative scattering determinant for @xmath20 and @xmath42 the corresponding relative scattering phase .
this formula holds for a general class of background manifolds @xmath43 and for much more general perturbations ; see proposition [ relcount ] . from the relative counting formula ( [ rcf ] ) , the role that the asymptotic ( [ n0.asymp ] ) for @xmath15 will play is clear .
the contribution from the relative scattering phase @xmath42 is similarly easy to account for , because it satisfies a weyl - type asymptotic as @xmath18 , @xmath44 where @xmath45.\ ] ] it is for this result that we must require smoothness of @xmath46 and @xmath47 . in various asymptotically hyperbolic settings ,
the scattering phase asymptotic was established by guilop - zworski @xcite , guillarmou @xcite , and borthwick @xcite . by adapting of the arguments from @xcite
, we can extend the result to the class of perturbations considered here , for a general class of background manifolds @xmath48 ) .
once we have the scattering phase asymptotic , the final step in estimating the right - hand side of ( [ rcf ] ) is to study the integral of @xmath49 over a half - circle .
it is here that we specialize to @xmath1 as the background space . with a combination of singular value techniques and asymptotic analysis of legendre functions , we produce a bound @xmath50 with @xmath51_+}{x^{n+2}}\>dx\>d\theta,\ ] ] where @xmath52_+$ ] denotes the positive part and @xmath53 \\ & \qquad + \log \left| \frac{\cosh r - \sqrt{1 + \alpha^2 \sinh^2 r}}{\cosh r + \sqrt{1 + \alpha^2 \sinh^2 r } } \right| .
\end{split}\ ] ] the @xmath54-dependence of @xmath55 is approximated by @xmath56 for @xmath54 large , so that @xmath55 is roughly proportional to @xmath57 .
the estimate ( [ logt.est ] ) leads directly to our main result : [ main.thm ] for @xmath58 , a compactly supported perturbation of the laplacian @xmath2 on @xmath1 as described above , we have @xmath59 where @xmath60 .
to highlight the dependence of @xmath38 on @xmath20 , we note that the constant @xmath61 is dimensional , @xmath62 depends on @xmath54 and on @xmath63 only through its volume , and @xmath55 depends only on @xmath54 .
none of these components depends on @xmath47 .
the factor @xmath64 is included in the formula ( [ sharp.est ] ) so that an asymptotic result of the same form as ( [ sharp.est ] ) would be equivalent to @xmath39 , with the same constant .
note that the only missing ingredient needed to establish such an asymptotic result is a lower bound of the same form as ( [ logt.est ] ) . to demonstrate the sharpness of theorem [ main.thm ] , we consider explicitly the case of scattering by a spherical obstacle in @xmath1 , for which @xmath65 , and @xmath66 with dirichlet boundary conditions on @xmath67 .
figure [ dresplot ] shows a sample resonance set for a spherical obstacle in @xmath68 . in @xmath68 .
all points off the real axis have multiplicity two ; on the real axis the multiplicities are more complicated . ]
[ dir.thm ] if @xmath20 is the dirichlet laplacian on @xmath69 , then @xmath70 in @xmath68 . ]
figure [ nobst ] shows that resonance counting functions @xmath19 for spherical obstacles in @xmath68 with several values of @xmath54 .
these graphs are based on exact computation of the resonances .
the approximate values of the asymptotic constants for these cases are @xmath71 already at @xmath72 we can see that the behavior of @xmath19 is consistent with the predictions of theorem [ dir.thm ] .
the paper is organized as follows . in [ bbox.sec ]
we develop basic spectral results , such as meromorphic continuation of the resolvent , in a very general `` black box '' perturbation setting . in [ scatt.sec ]
we narrow the context somewhat , in order to establish a nice factorization formula for the relative scattering determinant , from which the relative counting formula ( [ rcf ] ) follows .
another application of the factorization is the poisson summation formula for resonances , which leads to ( [ sig.weyl ] ) . the process of estimating the scattering determinant begins in [ pkernel.sec ] , with a formula that expresses this determinant in terms of the poisson kernel on @xmath1 . in [ tau.est.sec ]
we exploit this relation to prove ( [ logt.est ] ) and complete the proof of theorem [ main.thm ] .
a few explicit spherically symmetric examples are considered in [ ex.sec ] , which contains the proof of theorem [ dir.thm ] .
finally , the asymptotic analysis of legendre functions that is needed for [ tau.est.sec ] and [ ex.sec ] is developed in the appendix .
* acknowledgment .
* i would like to thank to plamen stefanov for suggesting the extension of his results to the hyperbolic setting .
that suggestion occurred during a workshop at the banff international research station , and subsequently part of the work was carried out during a stay at the centre international des rencontres mathmatiques in luminy .
i am grateful to both of these institutions for their support , and also thank the emory icis for additional travel support for the luminy meeting .
in geometric scattering theory , the term `` black box '' refers to a general class of perturbations of the euclidean laplacian in @xmath73 introduced by sjstrand - zworski @xcite .
although in standard usage this terminology is specific to the euclidean setting , the same abstract formulation can be adapted to other settings . in this section
, we will discuss black box perturbations in an asymptotically hyperbolic context .
our goal is to set up the definition of resonances by demonstrating meromorphic continuation of the resolvent , and then to prove a global estimate of the counting function .
it makes sense to do this in a general setting , since only minor changes are required to adapt previously published arguments .
this section essentially amounts to a review of known results .
an asymptotically hyperbolic metric on @xmath74 admits , by definition , a compactification @xmath75 with boundary defining function @xmath76 such that @xmath77 is a smooth , compact riemannian manifold with boundary and @xmath78 on @xmath79 .
we will assume that @xmath43 is _ even _ in the sense introduced by guillarmou @xcite .
this means that the taylor series of @xmath80 at @xmath81 contains only even powers of @xmath76 . under this assumption
the resolvent @xmath82 admits a meromorphic to @xmath83 , with poles of finite rank @xcite . appropriating the terminology from the euclidean case
, we define a class of perturbations of @xmath84 as follows .
given a compact @xmath85 , we consider the hilbert space @xmath86 where @xmath87 is some abstract hilbert space filling in for @xmath88 .
on @xmath89 we consider a self - adjoint operator @xmath20 with domain @xmath90 , satisfying the following assumptions : 1 . @xmath91 .
if @xmath92 and @xmath93 vanishes near @xmath28 , then @xmath94 .
2 . for @xmath94 ,
@xmath95 3 . as a map @xmath96
, @xmath97 is compact . here
the notations @xmath98 and @xmath99 denote the orthogonal projections @xmath100 and @xmath101 , respectively
. we will refer to an operator @xmath20 defined as above as a _ black box _ perturbation of @xmath84 .
given that meromorphic continuation of the resolvent is already known for @xmath84 , it is relatively easy to extend this result to @xmath20 .
[ pcontinue ] let @xmath43 be an even asymptotically hyperbolic manifold and @xmath20 a black box perturbation of @xmath84 .
the resolvent @xmath102 admits for any @xmath103 a meromorphic continuation to @xmath104 as an operator @xmath105 , with poles of finite rank .
the resolvent @xmath106 serves as a suitable parametrix for @xmath34 near the boundary .
let @xmath107 be cutoff functions equal to 1 on @xmath28 , such that @xmath108 on the support of @xmath109 and @xmath110 on the support of @xmath111 .
let @xmath112 , and define @xmath113 as an operator on @xmath114 with dirichlet boundary conditions imposed on @xmath115 , so that @xmath116 is self - adjoint .
we can naturally regard @xmath117 as an operator on @xmath89 .
then for @xmath118 such that @xmath119 we set @xmath120 then @xmath121 where @xmath122 with @xmath123 ( p_1 - z_0)^{-1 } \chi , \\
l_2(s , z_0 ) & : = ( s(n - s ) - z_0 ) \chi_1 ( p_1 - z_0)^{-1 } \chi , \\
l_3(s ) & : = [ \delta_{g } , \chi_0 ] r_0(s ) ( 1-\chi ) . \end{split}\ ] ] our goal is to prove that @xmath124 is compact and then apply the analytic fredholm theorem .
consider first the error term @xmath125 , which we can write as @xmath126 { \mathbbm{1}}_{x - k } ( p_1 - z_0)^{-1 } \chi\ ] ] by definition , @xmath127 maps @xmath89 to @xmath128 and we have assumed that the latter is contained in @xmath129 . since @xmath130 $ ] is first order with smooth coefficients whose compact support is contained in @xmath131 , we see that @xmath132 $ ] is compact as a mapping @xmath133 .
hence @xmath134 is compact @xmath96 .
the black box assumption that @xmath135 is compact implies that @xmath136 is compact on @xmath114 . and the resolvent identity @xmath137\ ] ] then shows that @xmath138 is compact on @xmath89 .
finally , the error term @xmath139 has a smooth kernel contained in @xmath140 .
this implies that for @xmath141 , @xmath139 is a compact operator on @xmath142 for @xmath104 . after adding the pieces together ,
these arguments show that @xmath124 is compact on @xmath143 for @xmath144 . using the self - adjointness of @xmath116 and the standard resolvent estimate , @xmath145 we can insure that @xmath146 is small by choosing @xmath147 large . similarly
, we can make @xmath148 small by choosing @xmath149 in the first quadrant sufficiently far from the real axis and the line @xmath150 .
thus for some @xmath151 we have @xmath152 , implying that @xmath153 is invertible at this point .
the analytic fredholm theorem then applies to define @xmath154 meromorphically on @xmath142 for @xmath104 .
the claimed result follows from @xmath155 because @xmath156 maps @xmath105 for @xmath104 .
the fact that @xmath34 admits meromorphic continuation as a bounded operator on @xmath89 for @xmath157 ( the @xmath158 case ) implies , as an immediate corollary , that @xmath159 theorem [ pcontinue ] allows us to define resonances associated to @xmath20 as the poles of @xmath34 , with multiplicities given by @xmath160 then @xmath35 is defined to be the set of resonances of @xmath20 , repeated according to the multiplicities @xmath161 .
the corresponding counting function is @xmath162 the remaining goal of this section is to establish an order - of - growth estimate for @xmath19 .
this requires first of all that @xmath43 be hyperbolic near infinity , in the sense that sectional curvatures all equal @xmath163 outside some compact set .
( no resonance bounds are currently known in the asymptotically hyperbolic case without this extra condition . )
such asymptotically hyperbolic manifolds are even in particular .
we must also make some extra assumptions of @xmath20 : 1 .
the operator @xmath20 must be bounded below , so that the set ( [ sig.disc ] ) is actually finite .
2 . the singular values of the resolvent of the cutoff operator @xmath116 introduced in the proof of theorem [ pcontinue ] satisfy a growth estimate , @xmath164 for some @xmath165 independent of @xmath166 and @xmath167 . the natural way to satisfy
the growth estimate ( [ p1.sing ] ) is to assume that @xmath168 for some riemannian manifold @xmath23 , possibly with boundary , and that @xmath20 is an elliptic self - adjoint pseudodifferential operator of order @xmath169 .
then to establish ( [ p1.sing ] ) we can start by using the resolvent estimate ( [ std.res ] ) to estimate @xmath170 let @xmath171 denote the dirichlet laplacian on @xmath172 . since @xmath173 has order @xmath163 , the operator @xmath174 is zeroth order and thus bounded on @xmath175 .
then ( [ p1.sing ] ) follows from @xmath176 ( the fact that @xmath165 can be chosen independently of @xmath166 follows from the resolvent estimate ( [ std.res ] ) . )
[ upper.bound ] let @xmath74 be a conformally compact manifold , hyperbolic near infinity , and @xmath20 a black box perturbation of @xmath84 that satisfies the extra assumptions ( i ) and ( ii ) .
then @xmath37 this is a fairly minor generalization of the upper bound proved by cuevas - vodev @xcite and borthwick @xcite .
this is because for those arguments the interior metric enters only in the interior parametrix term , i.e. , the first term on the right in ( [ rparametrix ] ) .
the difficult part of the upper bound analysis involves the terms supported near infinity , and this part of the argument applies immediately to @xmath20 by the assumption that @xmath177 . to apply the argument from cuevas - vodev , we need to check some estimates on the interior error terms @xmath134 and @xmath178 . for the former , the fact that @xmath179 implies that @xmath134 is bounded as a map @xmath180 .
if @xmath181 denotes the dirichlet laplacian on @xmath182 , then we can estimate @xmath183 where we can use ( [ res.id ] ) and ( [ std.res ] ) to see that we may take @xmath165 to be independent of @xmath118 . for the @xmath138 term , we first of all note that ( [ std.res ] ) implies @xmath184 by the assumption ( [ p1.sing ] ) we can immediately estimate @xmath185 for the argument in @xcite one needs to set @xmath186 for each @xmath103 such that @xmath187 , so the precise dependence of these estimates on @xmath149 and @xmath118 is significant .
the estimates ( [ muk.l1 ] ) and ( [ muk.l2 ] ) correspond precisely to the interior estimates ( * ? ? ?
* eq.s ( 2.234 ) ) .
the proof of ( * ? ? ? * prop .
1.2 ) then gives a bound @xmath188\bigr\ } \le c_\varepsilon r^{n+1}\ ] ] to fill in the missing sector containing the negative real axis , we apply the argument from borthwick @xcite . here the interior parametrix enters only in the proof of ( * ? ? ?
* lemma 5.2 ) .
the required bound is that for some constant @xmath189 , @xmath190 for @xmath191 .
since @xmath20 is self - adjoint and bounded below by assumption , this follows from the standard resolvent estimate .
the proof of ( * ? ? ? * prop .
5.1 ) then shows that @xmath192\bigr\ } \le c_\varepsilon r^{n+1}.\ ] ] the combination of estimates in the two regions gives the global result .
colin guillarmou has noted a mistake in the original argument from @xcite , which propagated through the arguments in @xcite and @xcite .
the faulty claim is that one can choose a family of cutoffs @xmath193 such that @xmath194 in some neighborhood of @xmath195 and also so that , in local coordinates isometric to the unit half - disk in @xmath1 , @xmath196 factors as @xmath197 in the coordinates @xmath198 .
it is not possible to satisfy these assumptions simultaneously .
fortunately , this problem is relatively easy to fix .
there are two sets of cutoffs used in these proofs .
( all three proofs use the same construction . ) the inner cutoffs @xmath193 must form a partition of unity near the boundary , but are not actually required to factor in local coordinates . the essential requirement for the inner cutoffs
is that their derivatives satisfy quasi - analytic estimates ( see ( * ? ? ?
* eq . ( 2.6 ) ) for example ) , and this is easily obtained without reference to a factorization .
the local factorization assumption is crucial only for outer cutoffs @xmath199 ( with @xmath200 on the support of @xmath196 ) .
we may keep this assumption in place because the outer cutoffs do not form a partition of unity .
for this section we continue to assume , as in theorem [ upper.bound ] , a conformally compact background manifold @xmath43 that is hyperbolic near infinity .
the restriction @xmath201 defines a riemannian metric on @xmath79 , whose conformal class is independent of @xmath76 .
thus @xmath79 is commonly referred to as the `` conformal infinity '' of @xmath43 .
a black box perturbation @xmath20 shares the same conformal infinity , since @xmath20 agrees with @xmath84 outside a compact set .
the scattering matrices @xmath202 and @xmath203 , associated to @xmath20 and @xmath84 , respectively , are pseudodifferential operators on @xmath79 defined as in @xcite . away from the diagonal , we can realize the kernel of the scattering matrix as a boundary limit of the resolvent : @xmath204 where @xmath205 = @xmath20 or @xmath206 .
( this relationship can be extended to the diagonal if one is sufficiently careful - see @xcite . )
this connection allows us to see that @xmath202 and @xmath203 differ by a smoothing operator , as follows . by applying @xmath34 to ( [ pmik ] ) from the left
, we obtain the identity @xmath207 then taking boundary limits as in ( [ sc.reslim ] ) gives the kernel of @xmath202 on the left , while on the right we obtain the kernel of @xmath203 as the limit of @xmath156 , plus a smooth contribution from the @xmath124 term .
this implies that the relative scattering matrix @xmath208 is determinant class , and we define the relative scattering determinant @xmath209 let @xmath210 denote the hadamard product over the resonance set @xmath211 : @xmath212 where @xmath213 [ detsrel.factor ] assume that @xmath74 is conformally compact and hyperbolic near infinity , and @xmath20 is a black box perturbation of @xmath84 satisfying the extra assumptions ( i ) and ( ii ) from [ bbox.sec ] .
the relative scattering determinant admits a factorization @xmath214 where @xmath215 is a polynomial of degree at most @xmath216 . since structure near
infinity is unchanged from @xmath74 , the arguments of guillarmou @xcite relating the resolvent and scattering pole multiplicities apply to @xmath20 .
thus the proof of ( * ? ? ?
7.2 ) shows that ( [ detsrel.pp ] ) holds with with @xmath215 a polynomial of unknown degree .
to control the degree , we use the fact that @xmath20 is bounded from below to obtain @xmath217 for some @xmath218
. then the proof of ( * ? ? ?
* lemma 5.2 ) gives that @xmath219 for @xmath220 with dist@xmath221 .
the same estimate applies to @xmath222 . in the formula
@xmath223 the hadamard products have order @xmath216 .
thus the @xmath224 estimates imply that @xmath225 in the half - plane @xmath226 , for any @xmath227 .
since @xmath215 is already known to be polynomial , the degree of @xmath215 is at most @xmath216 .
one nice application of proposition [ detsrel.factor ] is a jensen - type formula connecting the resonance counting functions to a contour integral involving the relative scattering determinant . to state
this we introduce the relative scattering phase of @xmath20 , defined as @xmath228 with branches of the log chosen so that @xmath229 is continuous starting from @xmath230 . by the properties of the relative scattering matrix ,
@xmath229 is real and @xmath231 .
the following relative counting formula is the asymptotically hyperbolic analog of a formula developed by froese @xcite for schrdinger operators in the euclidean setting .
[ relcount ] assume that @xmath20 is a black box perturbation of @xmath74 as in proposition [ detsrel.factor ] . as @xmath232 , @xmath233 according to proposition [ detsrel.factor ] , for @xmath234
, @xmath41 has zeros when @xmath235 or @xmath236 and the latter case occurs only if @xmath237 lies in the discrete spectrum of @xmath84 .
likewise , poles of @xmath41 for @xmath238 occur when either @xmath239 or @xmath240 , the latter only if @xmath237 lies in the discrete spectrum of @xmath20 .
all of these are counted with multiplicity of course .
let @xmath241 denote the contour @xmath242 ) ) \cup [ { \tfrac{n}{2}}+ it,{\tfrac{n}{2}}-it]$ ] , as shown in figure [ econtour ] .
assuming @xmath243 is not the absolute value of a resonance in @xmath244 or @xmath13 , we have @xmath245 where @xmath246 is the counting function for the ( finite ) set @xmath247 ( the resonances coming from the discrete spectrum ) . .
] evaluating the contour integral yields @xmath248 now if we divide by @xmath243 and integrate , we obtain the claimed formula with remainder given by @xmath249 our second important application of proposition [ detsrel.factor ] is to establish the poisson formula , which will lead to weyl - type asymptotics for the relative scattering phase .
define the meromorphic function @xmath250 by @xmath251,\ ] ] for @xmath252 .
the connection between @xmath250 and the relative scattering determinant established by patterson - perry ( * ? ? ?
5.3 and lemma 6.7 ) depends only on the structure of model neighborhoods near infinity , and so carries over to our case without alteration .
this yields the following birman - krein type formula : [ birman.krein ] for @xmath252 we have the meromorphic identity , @xmath253 by the functional calculus , @xmath254 is essentially the fourier transform of the continuous part of the wave 0-trace ( see ( * ? ? ?
* lemma 8.1 ) for the precise statement ) . by propositions [ detsrel.factor ] and [ birman.krein ]
we can write @xmath255\ ] ] taking the fourier transform just as in the proof of ( * ? ? ?
1.2 ) then gives a relative poisson formula : [ rel.poisson ] assume that @xmath20 is a black box perturbation of @xmath74 as in proposition [ detsrel.factor ] .
the difference of regularized wave traces satisfies @xmath256 - { \operatorname{0-tr}}\left [ \cos \left(t \sqrt{\delta_{g_0 } - \tfrac{n^2}4}\,\right ) \right ] \\ & \qquad = \frac12 \sum_{\zeta\in { \mathcal{r}}_p } e^{(\zeta-\frac{n}2)|t| } - \frac12 \sum_{\zeta\in { \mathcal{r}}_0 } e^{(\zeta-\frac{n}2)|t| } , \end{split}\ ] ] in the sense of distributions on @xmath257 .
the desired asymptotics of the scattering phase correspond to the big singularity of the wave trace at @xmath258 .
this singularity is very much analogous to that worked out by duistermaat - guillemin @xcite in the compact case .
the following result was proven for riemann surfaces , possible with internal boundary , by guillop - zworski ( * ? ? ?
* lemma 6.2 ) and for higher dimensional asymptotically hyperbolic manifolds ( without boundary ) by joshi - s baretto @xcite .
let @xmath43 be a riemannian manifold that is conformally compact and hyperbolic outside some compact set @xmath85 ( a more restrictive class than asymptotically hyperbolic ) .
then we consider another riemannian manifold @xmath23 , possibly with boundary , with compact @xmath24 such that @xmath259 .
let @xmath29 denote the laplacian on @xmath23 .
we may also include a potential @xmath30 , supported in @xmath27 . given this setup
we define the operator @xmath32 acting on @xmath260 with some self - adjoint boundary condition imposed on the internal boundary @xmath67 .
clearly @xmath20 is a black box perturbation of @xmath84 , and it satisfies assumptions ( i ) and ( ii ) of [ bbox.sec ] by the remark preceding theorem [ upper.bound ] . [ wt.sing ] assume that @xmath261 as described above . if @xmath262 has support in a sufficiently small neighborhood of @xmath206 and @xmath263 in some smaller neighborhood of @xmath206 , then @xmath264{p - n^2/4}}\,\right ) \right]\>dt
\sim \sum_{k=0}^\infty a_k |\xi|^{n-2k},\ ] ] where @xmath265 by finite speed of propagation we can use cutoffs to split the wave trace into internal and external pieces : @xmath266{p - n^2/4}}\,\right)\right ] & = { \operatorname{tr}}\left[\cos \left(t \sqrt{\smash[b]{p_1 - n^2/4}}\,\right ) \chi\right ] \\ & \qquad + { \operatorname{0-tr}}\left[\cos \left(t \sqrt{\smash[b]{\delta_{g_0 } - n^2/4}}\,\right ) ( 1-\chi)\right ] .
\end{split}\ ] ] the small time behavior of the first term ( which is an actual trace ) is given by ivrii s result for compact manifolds with boundary @xcite .
for the exterior term we can apply @xcite . using proposition [ birman.krein ] and the fourier transform relationship between @xmath267 and the wave @xmath206-trace , we can extract from proposition [ wt.sing ] the asymptotic behavior of the relative scattering phase , defined as @xmath228 with branches of the log chosen so that @xmath229 is continuous starting from @xmath230 . by the properties of the relative scattering matrix , @xmath229 is real and @xmath231 .
[ scphase.cor ] as @xmath268 , @xmath269\>\xi^{n+1 } + o(\xi^n).\ ] ] the argument to derive corollary [ scphase.cor ] from theorem [ rel.poisson ] and proposition [ wt.sing ] requires almost no change from that given for @xmath270 by guillop - zworski ( * ? ? ?
1.5 ) , so we omit the details .
proposition [ detsrel.factor ] ( and in particular the bound on the order of @xmath41 ) supplies the additional information needed to extend their result to @xmath271 . the leading coefficient is initially given by a difference of @xmath206-volumes , and we use @xmath272 to reduce this to a difference of the volumes of @xmath27 and @xmath28
since the asymptotics of @xmath42 are given by corollary [ scphase.cor ] , application of the formula from proposition [ relcount ] requires only estimation of @xmath273 in the half - plane @xmath157 . to facilitate this estimation , we need a more explicit realization of @xmath41 as a fredholm determinant
. this realization will involve the poisson kernel for the background metric @xmath74 . for the moment
we assume only that @xmath43 is an even asymptotically hyperbolic metric .
the poisson kernel can be derived from the kernel of the resolvent @xmath7 by the limit @xmath274 for @xmath275 and @xmath276 .
this kernel defines the poisson operator @xmath277 for @xmath278 , where @xmath279 is the metric induced on @xmath79 by @xmath280 .
for @xmath281 we can solve @xmath282 by setting @xmath283 .
moreover , for @xmath284 with @xmath237 not in the discrete spectrum of @xmath84 , @xmath93 has a two - part asymptotic expansion as @xmath285 , @xmath286 where @xmath203 is the scattering matrix .
this expansion , for general choice of @xmath287 , uniquely determines the scattering matrix via meromorphic continuation . the same construction works for @xmath202 . in particular ,
if we manage to find a family of solutions of @xmath288 such that @xmath289 for @xmath290 and @xmath149 in some suitable region , then @xmath202 can be identified as the map @xmath291 .
[ tauq ] suppose that @xmath20 is a black box perturbation of @xmath43 with support in @xmath28 .
let @xmath292 be cutoff functions such that @xmath293 and @xmath294 .
the relative scattering matrix can be written as the fredholm determinant @xmath295 where @xmath296 r_p(s ) [ \delta_0,\chi_1]e_0(n - s).\ ] ] since all of the operators in question are meromorphic families , we can restrict @xmath149 to some convenient set like @xmath150 , @xmath297 to avoid poles in the proof . given @xmath290 ,
consider the ansatz @xmath298 as a solution of @xmath299 .
then @xmath300 implies that @xmath301e_0(s)f + ( p - s(n - s ) ) u ' = 0.\ ] ] after applying @xmath34 on the left , we see that @xmath299 may be solved by setting @xmath302e_0(s)f.\ ] ] using the assumption on supports of @xmath109 and @xmath303 , we can derive @xmath304 u ' + ( 1-\chi_2 ) ( p - s(n - s))u ' \\ & = - [ \delta_0 , \chi_2]u ' . \end{split}\ ] ] this is compactly supported , so that @xmath7 may be applied to give @xmath305 u ' \\ & = -r_0(s)[\delta_0 , \chi_2 ] r_p(s ) [ \delta_0,\chi_1]e_0(s)f . \end{split}\ ] ] from this we can deduce the asymptotic behavior of @xmath306 as @xmath285 , @xmath307 r_p(s ) [ \delta_0,\chi_1]e_0(s)f.\ ] ] using the definition ( [ up.def ] ) of @xmath93 and the known asymptotic ( [ e0.asym ] ) for @xmath308 , we thus derive the expansion @xmath309
r_p(s ) [ \delta_0,\chi_1]e_0(s)f .\ ] ] we can rewrite this as @xmath310 using the identity @xmath311 which follows immediately from ( [ e0.asym ] ) . from ( [ u.ffp ] ) we read off
that @xmath312 and the determinant follows . in order to use lemma [ tauq ] to estimate @xmath41 , we need explicit knowledge of the background poisson operator @xmath313 . at this point
we specialize to @xmath314 and work out formulas for @xmath315 . in the usual @xmath316 coordinates , @xmath317
, we can read off immediately from ( [ r0.def ] ) that @xmath318^s.\ ] ] however , our application requires that @xmath319 be written in geodesic polar coordinates and then decomposed into spherical harmonics . the easiest way to do
this is to rederive @xmath319 from scratch . in geodesic polar coordinates , @xmath320 and
the hyperbolic metric is given by @xmath321 where @xmath322 denotes the standard sphere metric on @xmath323 .
it is thus natural to adopt the boundary defining function @xmath324 so that @xmath279 , the metric induced on @xmath79 by @xmath80 , is also the standard sphere metric .
the laplacian on @xmath1 is @xmath325 the eigenfunctions of @xmath326 are spherical harmonics @xmath327 with @xmath328 here @xmath329 and @xmath330 with @xmath331 [ e0.coeff ] for @xmath1 the poisson kernel in geodesic polar coordinates admits an expansion @xmath332 with coefficients given by @xmath333 where @xmath334 is the legendre function . if we expand the poisson kernel with respect to the spherical harmonic basis as in ( [ e0.expand ] ) , then the equation @xmath335 implies the coefficient equations , @xmath336 after a standard change of variables this becomes the legendre equation . since the poisson kernel is smooth in the interior , we select the legendre solutions that are recessive for @xmath337 , namely @xmath338 for some constants @xmath339 . the constant @xmath339 may be identified from the asymptotic expansion ( [ e0.asym ] ) as @xmath340 .
for the coefficients this expansion implies that @xmath341_l \rho^s,\ ] ] with @xmath342_l(s)$ ] the matrix elements of the scattering matrix @xmath203 , which will be diagonal in the spherical harmonic basis . using ( [ rhor ] ) and the well - known asymptotics of the legendre @xmath20-function , the leading terms in our ansatz as @xmath343 are @xmath344 from which we deduce @xmath345 for future reference , note that we can also read off from this construction the ( well - known ) matrix elements of @xmath203 , @xmath346_l = 2^{n-2s } \frac{\gamma({\tfrac{n}{2}}-s)}{\gamma(s-{\tfrac{n}{2 } } ) } \frac{\gamma(l+s)}{\gamma(l+n - s)}.\ ] ]
in this section we will combine the formula for @xmath41 from lemma [ tauq ] with the explicit fourier coefficients of the poisson kernel given in proposition [ e0.coeff ] .
we can then use estimates of the legendre @xmath20-function developed in the appendix to produce an estimate for the @xmath273 term in the counting formula from proposition [ relcount ] . throughout this section ,
the background metric is restricted to @xmath314 .
we assume that @xmath20 is a black box perturbation of the hyperbolic laplacian @xmath2 . as in [ intro.sec ] , the support of the perturbation is assumed to lie within @xmath347 the main result of this section is the following : [ inttau.prop ] let @xmath20 be a black box perturbation of the hyperbolic laplacian @xmath2 on @xmath1 . for @xmath348 , we can estimate @xmath349 as @xmath350 , where @xmath55 was defined by ( [ brad.def ] ) . before proceeding with the proof
, we note that the combination of proposition [ relcount ] , corollary [ scphase.cor ] , and theorem [ inttau.prop ] , immediately yields the proof of theorem [ main.thm ] .
note also that the restriction to @xmath261 ( from the more general black box class ) is needed only for corollary [ scphase.cor ] .
the proof of theorem [ inttau.prop ] will be broken into several stages , starting with : [ tau.sum.lemma ] assuming @xmath20 and @xmath54 are defined as above , fix some small @xmath351 and @xmath352 and define @xmath353 . for @xmath354 with dist@xmath355 ,
the relative scattering determinant can be estimated by @xmath356 where @xmath357^{\frac12 } \left [ \int_{r_2}^{r_3 } |a_l(s ; r)|^2\>(\sinh r)^n\>dr\right]^{\frac12 } , \ ] ] with @xmath358 the coefficients from proposition [ e0.coeff ] , and @xmath165 depends only on @xmath359 , @xmath241 , and @xmath54 .
let @xmath109 and @xmath303 be smooth cutoffs as in lemma [ tauq ] , such that @xmath360 for @xmath361 and @xmath362 for @xmath363
. then we can rewrite the @xmath364 from lemma [ tauq ] as @xmath365 } [ \delta_0 , \chi_2 ] r_p(s ) [ \delta_0,\chi_1]{\mathbbm{1}}_{[r_1,r_2]}e_0(n - s),\ ] ] where @xmath366}$ ] denotes the characteristic function @xmath367}(r)$ ] , acting as a multiplication operator . by lemma [ tauq ] and the cyclicity of the trace we have @xmath368 r_p(s ) [ \delta_0,\chi_1]{\mathbbm{1}}_{[r_1,r_2]}e_0(n - s ) e_0(s)^t { \mathbbm{1}}_{[r_2,r_3]}\bigr).\ ] ] for @xmath354 , under the assumption dist@xmath369
, we can apply the spectral theorem and standard elliptic estimates to obtain @xmath370 r_p(s ) [ \delta_0,\chi_1 ] \bigr\vert \le c,\ ] ] where @xmath165 depends on @xmath359 , @xmath241 , and @xmath54 . under these restrictions , @xmath371
where @xmath372}e_0(n - s ) e_0(s)^t { \mathbbm{1}}_{[r_2,r_3]}\ ] ] using proposition [ e0.coeff ] , the eigenfunctions of @xmath373 can then be written down explicitly . if we define @xmath374 } \overline{a_l(s;r ) y_l^m(\omega)},\ ] ] then @xmath375 where @xmath376 is given by ( [ laml.def ] ) . to see that @xmath377 , counted with multiplicities , contains all of the nonzero eigenvalues of @xmath373 , suppose that @xmath378 and @xmath379 for all @xmath380 .
then by ( [ e0.expand ] ) we have @xmath381 } w = 0 $ ] , which implies that @xmath382 .
hence , after possible rearrangement , the sequences @xmath377 and @xmath383 correspond .
the claimed estimate follows from ( [ tau.sv ] ) .
lemma [ laml.def ] reduces our problem to the estimation of the @xmath376 s , which we take up next .
[ lambda.lemma ] assume that @xmath157 and @xmath384 , and set @xmath385 and @xmath386 .
assuming that @xmath387 in the definition ( [ laml.def ] ) of @xmath376 , we have the bound @xmath388 where @xmath389 was defined in ( [ hdef ] ) , with a constant @xmath165 that depends only on @xmath54 . from proposition [ e0.coeff ]
we obtain the explicit formula @xmath390^{\frac12 } \\ & \qquad \times\left [ \int_{r_2}^{r_3 } \bigl|p_{s-\frac{n+1}2}^{-l - \frac{n-1}2}(\cosh r)\bigr|^2\>\sinh r\>dr\right]^{\frac12 } , \end{split}\ ] ] where we have exploited the symmetry @xmath391 . by conjugation , if necessary , we can assume that @xmath392 $ ] . applying the estimate from corollary [ pq.bounds ] then yields @xmath393^{\frac12 } \left [ \int_{r_2}^{r_3 }
e^{2k { \operatorname{re}}(\phi(\alpha , r)-p(\alpha))}\>\sinh r\>dr\right]^{\frac12}. \end{split}\ ] ] by ( [ dzph ] ) , @xmath394 is increasing as a function of @xmath395 . hence @xmath396^{\frac12 } \le e^{k { \operatorname{re}}(\phi(\alpha , r_3)-p(\alpha ) ) } \cosh r_3,\quad j=1,2.\ ] ] the first factor on the right side of ( [ lam.bound1 ] ) can be estimated directly via stirling s formula for @xmath397 , @xmath398 + o(\log k ) , \end{split}\ ] ] as @xmath399 , uniformly for @xmath400 .
we can extend the same estimate to @xmath401 , using @xmath402 and our assumption that @xmath403 , which implies @xmath404 after we note that @xmath405,\ ] ] we obtain from ( [ lam.bound1 ] ) and ( [ gamma.bd1 ] ) the estimate @xmath406 now we can combine lemmas [ tau.sum.lemma ] and [ lambda.lemma ] to estimate @xmath41 . the strategy here is similar to stefanov s in ( * ? ? ?
[ logtau.prop ] for @xmath407 and @xmath408 we have @xmath409 uniformly for @xmath410 , with @xmath411_+}{x^{n+2}}\>dx,\ ] ] where @xmath52_+$ ] denotes the positive part . since @xmath412 the assumption that @xmath410 implies that @xmath413 remains a distance @xmath414 from @xmath415 for @xmath416 sufficiently large .
the hypothesis of lemma [ tau.sum.lemma ] is thus satisfied , yielding the estimate ( [ tau.sum ] ) with a @xmath165 that depends only on @xmath359 , @xmath241 , and @xmath54 . to apply lemma [ lambda.lemma ] to estimate the right - hand side of ( [ tau.sum ] )
, we need to distinguish the terms according to the sign of @xmath417 . for large @xmath416
the sum is dominated by terms with @xmath418 , which occurs for @xmath419 outside a certain neighborhood of the origin , as shown in figure [ hplot ] .
, shown for @xmath420 . ]
let @xmath421 be the implicit solution of the equation @xmath422 , so that @xmath423 precisely when @xmath424
. given some @xmath227 , we will subdivide the sum ( [ tau.sum ] ) by breaking at values where @xmath425 and @xmath426 , leaving us with three parts .
the dominant part of the sum will be @xmath427 ( recall that @xmath428 where @xmath429 . ) for @xmath419 in this range , assuming @xmath430 , we apply lemma [ lambda.lemma ] to obtain @xmath431 using this estimate together with the asymptotic @xmath432 we have @xmath433 ( the sum could be restricted to @xmath434 , but this would not improve the bound . ) we can estimate @xmath435 with a constant that depends only on @xmath436 .
thus @xmath437 with this estimate , the sums over lower order terms in ( [ sigp.est ] ) are easily controlled , and we obtain @xmath438 where @xmath165 depends only on @xmath359 , @xmath241 , and @xmath54 . because @xmath439 is an increasing function of @xmath440 , the right - hand side of ( [ sigplus.sum ] ) is easily estimated by the corresponding integral , @xmath441 making the substitution @xmath442 gives @xmath443_+}{x^{n+2}}\>dx + c a^n \log a \\ & = b(\theta , r_3)a^{n+1 } + c a^n \log a , \end{split}\ ] ] with @xmath165 depending only on @xmath359 , @xmath241 , and @xmath54 .
the middle term in ( [ tau.sum ] ) will be @xmath444 the number of terms in this sum is @xmath445 , and we can control them using ( [ log1lam ] ) , noting also that @xmath446 for @xmath447 in the given range .
using an integral estimate as we did for @xmath448 , we thus obtain @xmath449 where @xmath165 depends only on @xmath359 , @xmath241 , and @xmath54 .
the final portion of the sum is @xmath450 we use the fact that @xmath451 in this range to estimate @xmath452 this implies @xmath453 for some @xmath454 , where @xmath455 depends on @xmath456 as well as @xmath359 and the @xmath457 s . adding the three parts @xmath458 of ( [ tau.sum ] ) together
now yields @xmath459 + c({\varepsilon},\eta , r_0,\delta ) e^{-ca},\ ] ] where we have made the dependence of the constants explicit . since @xmath389 is a strictly increasing function of @xmath395 , we can absorb the @xmath460 term into the first term by assuming that @xmath456 is small relative to @xmath241 and replacing @xmath436 with @xmath461 . with this change
, we obtain from ( [ r3.delta ] ) the estimate @xmath462 the constant @xmath463 may well blow up as @xmath464 .
the best we can do here is to observe that ( [ r0.r4 ] ) implies @xmath465 \le b(\theta , r_0 + 4\eta ) - b(\theta , r_0),\ ] ] for any @xmath466 .
since @xmath467 is uniformly continuous on @xmath468\times [ r_0 , r_0 + 1]$ ] , we can now let @xmath464 to obtain the claimed @xmath469 estimate . for any @xmath470
, we can integrate the result from proposition [ logtau.prop ] over @xmath410 , which gives , @xmath471 the factorization given by proposition [ detsrel.factor ] , together with the minimum modulus theorem ( see e.g. ( * ? ? ?
8.71 ) ) , implies that for any @xmath227 , there exists a sequence @xmath472 such that @xmath473 . in sectors of the form
@xmath474 $ ] , where @xmath475 is analytic , we can apply a phragmn - lindelf argument , using ( [ tau.order.est ] ) , @xmath476 for @xmath150 , and the estimate from proposition [ logtau.prop ] for @xmath477 , to conclude that @xmath478 uniformly for @xmath474 $ ] .
thus , @xmath479
suppose that @xmath480 and we consider a black box perturbation @xmath261 , where both the metric @xmath46 and potential @xmath47 are spherically symmetric .
the symmetry assumption guarantees that the perturbed poisson kernel is `` diagonalized '' by spherical harmonics , in the sense that @xmath481 the coefficients @xmath358 will satisfy ( [ ml.def ] ) for @xmath482 and are thus expressible in terms of legendre functions . following the convention of olver , we use the legendre q - function in the form @xmath483 where @xmath484 is the standard definition .
this makes @xmath485 an entire function of either @xmath486 or @xmath487 , which is much more convenient for identifying resonances .
we can formulate the general solution of ( [ ml.def ] ) for @xmath488 as @xmath489,\ ] ] where @xmath490 in particular examples , @xmath339 and @xmath491 will be determined by matching @xmath492 and its first derivative to the corresponding solutions for @xmath493 .
the scattering matrix elements can be read off from the asymptotics of these solutions as @xmath343 , using @xmath494_l\bigr),\ ] ] in the same way that we found @xmath342_l$ ] in ( [ s0.coeff ] ) .
indeed , from the well - known asymptotic ( * ? ? ?
* eq . ( 12.09 ) ) @xmath495 we can see from ( [ al.general ] ) that the scattering matrix elements are given by @xmath496_l = - 2^{n-2s } \frac{\gamma({\tfrac{n}{2}}-s)}{\gamma(s-{\tfrac{n}{2 } } ) } \frac{a_l(s)}{b_l(s)}.\ ] ] consider the case where @xmath20 is the laplacian for a spherical obstacle of radius @xmath54 in @xmath1 . imposing the dirichlet boundary condition at @xmath497
gives coefficients @xmath498 in this case , from ( [ sp.coeff ] ) we see that @xmath499_l = 2^{n-2s } \frac{\gamma({\tfrac{n}{2}}-s)}{\gamma(s-{\tfrac{n}{2 } } ) } \frac{{\mathbf{q}}_{-\nu-1}^{k}(\cosh r_0)}{{\mathbf{q}}_{\nu}^{k}(\cosh r_0)}.\ ] ] with this observation we can give the : our goal is to show that @xmath500 in conjunction with proposition [ relcount ] and corollary [ scphase.cor ] , ( [ logtau.asym ] ) would imply that @xmath501 and this is equivalent to the stated asymptotic for @xmath19 .
using ( [ obst.s ] ) with ( [ s0.coeff ] ) gives the relative scattering matrix elements , @xmath502_l = \frac{\gamma(l+n - s)}{\gamma(l+s ) } \frac{{\mathbf{q}}_{-\nu-1}^{k}(\cosh r_0)}{{\mathbf{q}}_{\nu}^{k}(\cosh r_0)},\ ] ] with @xmath167 and @xmath487 defined as in ( [ k.nu ] ) . with the connection formula ( *
* eq . ( 12.12 ) ) , @xmath503 we can rewrite the coefficient in the form @xmath504_l = 1 - \cos ( \pi \nu)\ > \gamma(k-\nu ) \frac{p_{\nu}^{-k}(\cosh r_0)}{{\mathbf{q}}_{\nu}^{k}(\cosh r_0)}.\ ] ] now consider @xmath505_l\bigr|.\ ] ] defining @xmath419 by @xmath506 , we can use ( [ dir.ss ] ) to write this as @xmath507 where @xmath508 assuming that @xmath509 , corollary [ poverq ] gives the estimate @xmath510},\ ] ] with constants depending only on @xmath359 . applying
stirling s formula and avoiding the poles by assuming @xmath384 as in the proof of lemma [ lambda.lemma ] , we have @xmath511 + o(\log \alpha)\ ] ] since @xmath512 = h(\alpha , r_0),\ ] ] the full estimate is @xmath513 as in the proof of proposition [ logtau.prop ] we divide the sum ( [ lt.eta ] ) into three pieces @xmath514 , and @xmath515 , with breaks at @xmath516 for some @xmath227 .
the dominant piece is @xmath517 using the lower bound from ( [ etak.est ] ) , but otherwise arguing as in the proof of proposition [ logtau.prop ] , we have @xmath518 the estimates on @xmath519 and @xmath515 are identical to those in proposition [ logtau.prop ] : @xmath520 and @xmath521 hence we conclude that @xmath522 for @xmath523 and @xmath524 , with constants that depend only on @xmath54 and @xmath359 . integrating , over @xmath525 , and using proposition [ logtau.prop ] to control the errors from @xmath526 $ ] , we obtain the estimate @xmath527 valid for any @xmath528 .
( this @xmath529 combines the terms proportional to @xmath359 and @xmath456 from above . ) in combination with theorem [ inttau.prop ] , this proves ( [ logtau.asym ] ) . with some care , the explicit scattering matrix provided by ( [ sp.coeff ] ) can be used to compute resonances .
scattering poles and zeros are defined a renormalized scattering matrix @xmath530 , in which the infinite rank poles and zeros coming from the gamma functions are removed , @xmath531 the scattering multiplicity is then defined by @xmath532.\ ] ] for @xmath1 , the connection between scattering multiplicities and resonances is given by @xcite @xmath533 for @xmath534 , the term @xmath535 plays a role only if @xmath20 has discrete spectrum . in the examples that we will consider explicitly , @xmath270 and the discrete spectrum is empty . for these cases ,
the resonances are precisely the poles of the @xmath530 .
consider first the spherical obstacle of radius @xmath54 in @xmath68 , for which the scattering matrix is given by ( [ obst.s ] ) . from this expression
we can read off the resonance set @xmath536 figures [ dresplot ] and [ nobst ] were thus obtained through numerical computation of zeroes of the legendre q - function . as a second example , we consider scattering in @xmath68 by a radial step potential of the form @xmath537 in this case , with @xmath538 , the coefficient solutions for @xmath539 are legendre @xmath20 functions @xmath540 , with @xmath541 the corresponding resonance set is @xmath542\big|_{z = \cosh r_0 } = 0\bigr\}.\ ] ] where @xmath543 is the wronskian .
resonance counting functions for @xmath544 and @xmath545 , with @xmath546 , are shown in figure [ npot ] .
we should note that theorem [ main.thm ] does not apply to the step potential , because the lack of smoothness means that we can not derive scattering phase asymptotics through corollary [ scphase.cor ] .
in view of the scattering phase asymptotics proved by christiansen @xcite in the black box euclidean case , one might hope that the smoothness requirement in our case could be loosened .
however , the technique of robert used in @xcite does not seem to be applicable to the conformally compact hyperbolic case . in any case , it is interesting to compare the putative upper bound suggested by theorem [ main.thm ] to the empirical results based on ( [ r.pot ] ) . for both of the cases shown in figure [ npot ] , the constant from the theorem would be @xmath547 .
the numerical results thus suggest that @xmath19 satisfies an asymptotic with a constant significantly smaller than upper bound that theorem [ main.thm ] would predict . .
] our final example is a `` transparent '' spherical obstacle .
let @xmath548 where @xmath549 then @xmath550 = 0\bigr\},\ ] ] where @xmath551 figure [ ntrans ] shows resonance counting functions for @xmath552 and @xmath553 . once again , theorem [ main.thm ] does not apply because of the lack of smoothness .
however , in this case the predicted constants , @xmath554 at least roughly match the observed behavior , so that one might believe that the theorem would give a sharp result if extended to this case . with @xmath555 . ]
in this section we will estimate the growth of the legendre functions @xmath556 and @xmath557 as @xmath558 simultaneously .
we wish to extract the leading asymptotic behavior , with error bounds uniform in @xmath559 for @xmath560 .
the construction of these estimates leans heavily on techniques from olver @xcite . throughout this discussion
we identify @xmath561 and switch freely between the two variables .
let @xmath562 then the legendre equation reduces to @xmath563 with @xmath564 if @xmath565 then the equation ( [ wleg ] ) has turning points ( points where @xmath287 vanishes to first order ) when @xmath566 .
by conjugation , it suffices to assume @xmath567 and so we focus on the upper turning point . to obtain uniform estimates near this point
, we introduce the complex variable @xmath568 defined by integrating @xmath569 starting from @xmath570 on the left and from @xmath571 ( the turning point ) on the right . throughout this section
we assume principal branches for the logs and square roots , under the restriction that @xmath572 $ ] . integrating both sides of ( [ zeta.fdz ] )
yields @xmath573 where @xmath574 .
\end{split}\ ] ] the expression ( [ phi.def ] ) is well - defined by principal branches for @xmath575 $ ] , and we extend the definition to the positive real axis by continuity .
( at the apparent singularity at @xmath576 , this extension yields @xmath577 . )
the region of interest , namely @xmath572 $ ] and @xmath578 , corresponds to the sector @xmath579 $ ] , as illustrated in figure [ phiplot ] .
figure [ zetaplot ] show the corresponding picture for @xmath568 , and illustrates in particular how passing from @xmath580 to @xmath568 resolves the singularity at the turning point . for future reference
, we note that @xmath580 satisfies the equation @xmath583 implying in particular that @xmath584 .
the fact that @xmath585 is an increasing function of @xmath395 will be important later , and is not so evident from ( [ phi.def ] ) .
[ legpq.airy ] assuming that @xmath591 , @xmath592 $ ] and @xmath593 , we have @xmath594^{\frac14 } } \>e^{-k p(\alpha ) } \bigl [ { { \rm ai}}\bigl(k^{\frac23 } e^{\frac{2\pi i}3 } \zeta\bigr ) + h_1(k , \alpha , r)\bigr],\ ] ] and @xmath595^{\frac14 } } \
> e^{kq(\alpha ) } \bigl [ { { \rm ai}}\bigl(k^{\frac23 } \zeta\bigr ) + h_0(k,\alpha , r)\bigr],\ ] ] where @xmath568 is defined by ( [ zeta.def ] ) and ( [ phi.def ] ) , @xmath596 and @xmath597 are defined in ( [ pal.def ] ) and ( [ qal.def ] ) , respectively . the error terms satisfy @xmath598 with @xmath165 independent of both @xmath419 and @xmath395 .
if we set @xmath599 , then the equation ( [ wleg ] ) transforms to : @xmath600 a perturbed version of the airy equation , with the extra term given by @xmath601 following olver ( * ? ? ?
11.9.1 ) , we consider solutions of the form @xmath602 for @xmath603 , where the error terms satisfy @xmath604,\ ] ] let us focus first on the legendre @xmath20-function .
as @xmath605 , the airy function @xmath606 is exponentially decreasing for @xmath607 and exponentially increasing for @xmath608 $ ] . since @xmath609 is recessive at zero , and @xmath610 corresponds to @xmath611 , we choose the solution @xmath612 from ( [ w.airy ] ) .
the assumption that the solution is recessive as @xmath610 implies boundary conditions , @xmath613 which we must impose on ( [ hsig.eq ] ) . to identify the legendre @xmath20 function with a multiple of @xmath612 , we compare the well - known asymptotic @xmath614 to the behavior of the ansatz @xmath615^{\frac14 } } \bigl({{\rm ai}}(k^{\frac23}e^{\frac{2\pi i}3 } \zeta ) + h_1(k , \alpha , r)\bigr).\ ] ] away from the negative real axis , the airy function has the asymptotic behavior ( * ? ? ?
* eq . ( 4.4.03 ) ) @xmath616,\ ] ] uniformly for @xmath617 , with a constant that depends only on @xmath227 . to cover the negative real axis we have also (
* eq . ( 4.4.05 ) ) , @xmath618,\ ] ] uniformly for @xmath619 $ ] .
( these estimates agree where they overlap . ) as @xmath337 , we have @xmath620 , which is in the range covered by ( [ ai.asym1 ] ) . along with the asymptotic behavior of @xmath568 deduced from ( [ phi.asym1 ] ) , this yields @xmath621 thus for the legendre @xmath20-function we find @xmath622 which proves ( [ p.asym ] ) . to complete the analysis of the @xmath20 case , it remains to control the size of the error term @xmath623 .
the error bounds may be derived as in the proof of ( * ? ? ?
11.9.1 ) , starting from the differential equation ( [ hsig.eq ] ) satisfied by @xmath624 .
using the boundary condition ( [ h.bndry1 ] ) , we can apply variation of parameters to transform this to an integral equation , @xmath625\ > \frac{f(r')^\frac12 \sinh r'}{\zeta(r')^\frac12 } \>dr',\ ] ] where @xmath626 then , using the method of successive approximations as in ( * ? ? ?
6.10.2 ) , together with the bounds on the airy function and its derivatives developed in @xcite , we obtain the bound , @xmath627 where @xmath628 using ( [ fz.gz ] ) and ( [ psi.gfz ] ) , direct computation shows that @xmath629 + \frac{5}{16 } \frac{(1+\alpha^2 \sinh^2 r)^{\frac12}}{\zeta^\frac52\sinh^2 r}.\ ] ] because we require estimates that are uniform in both @xmath419 and @xmath395 , the analysis of ( [ psi.intg ] ) is somewhat complicated . for some small @xmath454 , we will break the estimation into 3 different zones as described below .
we use the notation @xmath630 to mean that the ratio @xmath631 is bounded above and below by positive constants that do not depend on @xmath419 or @xmath395 .
_ zone 1 : _ assume that @xmath632 and @xmath633 . the first term in the formula ( [ phi.def ] ) for @xmath580 dominates for large @xmath395 and the second term for small @xmath395 .
we can thus derive the bounds , @xmath634 in ( [ psi.intg ] ) gives @xmath635 it is then relatively straightforward to control the contribution of these terms to ( [ psi.def ] ) .
for @xmath633 , we obtain @xmath636 for some @xmath165 that depends on @xmath637 but not on @xmath419 . _ zone 2 : _ assume that @xmath632 and @xmath638 .
in this case we claim that @xmath639 ) then gives @xmath640 for @xmath638 , we obtain @xmath641 for some @xmath165 that depends on @xmath637 but not on @xmath419 .
_ zone 3 : _ assume that @xmath642 .
this puts us near the turning point .
it is convenient to use the @xmath561 variable here .
the turning point occurs at the point @xmath643 which lies near the path of integration for ( [ psi.def ] ) only when @xmath644 is close to @xmath645 .
note that @xmath646 so that the assumption @xmath642 translates to @xmath647 using the definition of @xmath580 as @xmath652 , we can write @xmath653 in the form @xmath654 then from ( [ pk.est ] ) we can derive estimates of the same form for @xmath653 , @xmath655 ) and ( [ qk.est ] ) , with the fact that @xmath656 and the formula for @xmath657 given in ( [ psi.gfz ] ) , we obtain @xmath658 for @xmath638 , the result is @xmath659 for @xmath660 the corresponding estimate is @xmath661 now we can combine the estimates of contributions to @xmath662 from all three zones , namely ( [ zone1 ] ) , ( [ zone2 ] ) , ( [ zone3a ] ) , and ( [ zone3b ] ) , to obtain @xmath663 $ ] and @xmath664 , with @xmath165 independent of both @xmath395 and @xmath419 . applying the resulting estimate of @xmath662 in ( [ h.bound ] ) then gives @xmath665 this completes the error analysis in the @xmath20 case .
we turn now to the legendre @xmath666-function and the proof of ( [ q.asym ] ) .
we want the solution to be recessive at @xmath667 , so we set @xmath668 in the ansatz ( [ w.airy ] ) and impose the condition @xmath669 using ( [ phi.asym2 ] ) and ( [ ai.asym1 ] ) we have @xmath670^{\frac14 } } { { \rm ai}}(k^{\frac23}\zeta ) \sim \frac{k^{-\frac16}(\frac{\alpha}2)^{-\frac12}}{2\pi^{\frac12 } } e^{-(k\alpha+\frac12)r},\ ] ] as @xmath340 . from the asymptotic ( [ q.asym ] ) we find @xmath671 hence , for the @xmath666-legendre function we have @xmath672 which proves ( [ q.asym ] ) .
to control @xmath673 we use the boundary condition ( [ h2.bc ] ) to transform the differential equation ( [ hsig.eq ] ) for @xmath673 into an integral equation , @xmath674\ > \frac{f(r')^\frac12 \sinh r'}{\zeta(r')^\frac12 } \>dr',\ ] ] where @xmath675 by conjugation , it suffices to assume that @xmath592 $ ] . using the asymptotics , ( [ ai.asym1 ] ) and ( [ ai.asym2 ] ) , and the first error estimate from ( [ hestimates ] )
, we have @xmath685 \bigr| \le c e^{k{\operatorname{re}}\phi}\ ] ] if we assume that @xmath686 , for some @xmath454 , then the estimate ( [ p.upper ] ) follows immediately from ( [ p.asym ] ) . on the other hand ,
if @xmath687 , then by the assumption that @xmath688 $ ] , we deduce that @xmath447 , @xmath689 , and the ratio @xmath690 $ ] are all @xmath691 . for @xmath692 , we can use ( [ aih.est ] ) to complete the estimate
if @xmath693 , then @xmath694 is bounded by ( [ hestimates ] ) and the fact that @xmath606 is regular at the origin .
( it is only because of this last case that the factor @xmath695 must be included in the final estimate . )
[ poverq ] assuming that @xmath680 , @xmath470 , and @xmath696 $ ] , we have uniform bounds for @xmath167 sufficiently large : @xmath697},\ ] ] meaning the the ratio of the two sides is bounded above and below by constants depending only on @xmath359 .
( the upper bound extends to @xmath698 , but the lower bound does not . ) by ( [ p.asym ] ) and ( [ q.asym ] ) we have @xmath699 } \>\frac{{{\rm
ai}}\bigl(k^{\frac23 } e^{\frac{2\pi i}3 } \zeta\bigr ) + h_1(k , \alpha , r ) } { { { \rm ai}}\bigl(k^{\frac23 } \zeta\bigr ) + h_2(k,\alpha , r)}\ ] ] for some @xmath454 , consider first the case where @xmath700 .
the assumption that @xmath701 is bounded away from @xmath645 implies that @xmath702 $ ] , so that we can apply ( [ ai.asym1 ] ) to estimate both of the airy functions in ( [ pqratio ] ) . by choosing @xmath637 sufficiently large
, we can assume that the factor @xmath703 appearing in ( [ ai.asym1 ] ) is bounded away from zero , since @xmath704 in our case . by the estimates ( [ hestimates ] ) and the assumption @xmath680 , by choosing @xmath167 sufficiently large we can assume that @xmath705 and @xmath706 are arbitrarily small relative to the airy function estimates . under these assumptions
we have @xmath707 the bound then follows immediately . | for certain compactly supported metric and/or potential perturbations of the laplacian on @xmath0 , we establish an upper bound on the resonance counting function with an explicit constant that depends only on the dimension , the radius of the unperturbed region in @xmath0 , and the volume of the metric perturbation . this constant
is shown to be sharp in the case of scattering by a spherical obstacle . |
we consider the specific system outlined in the figure [ f : picture ] . here
two electron spins are localized in adjacent quantum dots , coupled via tunneling .
one of the dots is capacitively coupled to a transmission line resonator . a modest ( 100 mt ) external magnetic field along an axis @xmath7 zeeman - splits the spin - aligned states , @xmath8 and @xmath9 , while the spin - anti - aligned states are used as a qubit degree of freedom : @xmath10 and @xmath11 . the notation @xmath12 labels the number of electrons in the left and right quantum dots .
in addition to the qubit states , an auxiliary singlet state with two electrons in one quantum dot , @xmath3 , is coupled via tunneling @xmath0 to the separated singlet , @xmath13 ( fig . [ f : picture]b ) .
the energy of the auxiliary state is determine largely by the bias @xmath14 due to the electric field ; control of @xmath14 allows to control the qubit s evolution .
the hamiltonian for this three state system is : @xmath15 we have introduced a static magnetic field gradient between the two dots @xmath16 , which mixes the singlet and triplet states .
this system has been studied in detail in refs . .
parameter ranges for experiments are @xmath17ev and @xmath18ev .
state preparation , measurement , 1-qubit gates , and local 2-qubit gates can be achieved by changing the bias @xmath14 between @xmath3 and @xmath13 @xcite .
interaction with the resonator is included naturally by writing the bias as a contribution from static fields , @xmath2 , set by voltages on the gates defining the double dot , and a contribution from the resonator itself : @xmath19 where @xmath20 is the capacitive coupling of the resonator to the dot , while @xmath21 is the total capacitance of the double dot .
the voltage due to the resonator of length @xmath22 , capacitance per unit length @xmath23 , and impedance @xmath24 is quantized as @xcite @xmath25 we now deduce an effective hamiltonian for the system when the splitting between eigenstates of eq .
[ e : hdd ] , @xmath26 , is comparable to the fundamental mode frequency of the resonator , @xmath27 ( i.e. , the cavity detuning @xmath28 is small ) . neglecting the higher energy modes , we write the hamiltonian for the resonator itself as @xmath29 , where @xmath30 is the lowest energy mode destruction operator , and the interaction with the dot is @xmath31 where @xmath32 is the vacuum rabi coupling between the double dot and the resonator . in practice , for @xmath33 ghz , @xmath34 mhz ( i.e. , @xmath35 ) is achievable @xcite ) .
a reduction of @xmath36 by lowering @xmath37 results in a comparable reduction of @xmath38 .
finally , as @xmath39 resonators are possible in the microwave domain @xcite , the quantization of the resonator voltage is appropriate . in practice ,
the need to work with a finite local external magnetic field ( 100 mt ) may limit @xmath40 @xcite .
we seek a set of parameters @xmath41 such that the system can be coupled to and controlled by the resonator . this `` operating point '' should maximize the coupling to the resonator and minimize the noise in the combined system . to motivate the optimal choice ,
let us first proceed with small tunnel coupling @xmath0 and cavity coupling @xmath38 . in this case
the eigenstates are @xmath42 , with eigenenergies @xmath43 . setting the energy difference @xmath44 ,
degenerate perturbation theory in the tunnel coupling @xmath0 reveals an avoided crossing at this balanced point between @xmath45 and @xmath3 with an energy gap @xmath46 around @xmath47 ( fig .
[ f : picture]b ) and mixing angle @xmath48 . working in the rotating frame with the rotating wave approximation , we find an effective hamiltonian for the combined system to be @xmath49 where , at @xmath50 , @xmath51 and @xmath52 we choose @xmath53 for two reasons .
first , at this point the gap @xmath26 is insensitive to first - order changes in @xmath2 and @xmath54 .
such zero - derivative points are optimal for controlling quantum bits with imperfect electronics and in the presence of low frequency noise @xcite . in our case
this suppresses fluctuations in control electronics ( @xmath2 ) and in the magnetic field gradient ( due to hyperfine interactions ) , which are the dominant dephasing terms .
second , the coupling coefficient @xmath55 is directly proportional to the energy splitting in the system , as would be expected of a `` bare '' coupling of comparable strength
. the cost of working at this point is increased sensitivity to fluctuations in @xmath0 ( it effects both @xmath55 and @xmath26 at first order ) and a higher probability of relaxation due to the enhanced charge admixture of the qubit states ( see the analysis of inelastic effects below ) .
the former could be mitigated : as no time - dependent control of @xmath0 is required , the gates that set @xmath0 can be heavily filtered to greatly reduce potential noise . for long term storage of the quantum information
, we can adiabatically map the states @xmath56 to the spin states @xmath57 via change of @xmath2 , going to a well separated regime .
as long as the change of @xmath2 is slow with respect to the splitting @xmath54 , the process is entirely adiabatic ; furthermore , we are only sensitive to charge fluctuations on a frequency scale faster than @xmath58 . the spin states , protected by occassional refocusing , are mostly insensitive to charge relaxation and dephasing @xcite .
thus , when long quantum memory times or insensitivity to charge - based relaxation is necessary , the system can be adiabatically mapped to such a separated regime ; when operation with the resonator is necessary , they can be mapped back to the balanced regime of @xmath59 .
we now proceed to generalize the above analysis to an arbitrary set of parameters @xmath60 , treating the interaction with the cavity as a perturbation .
we write the eigenenergies of @xmath61 as @xmath62 ( @xmath63 , and @xmath64 ) and the eigenvectors as @xmath65 where @xmath66 is a 3x3 unitary matrix . in the rotating wave approximation
, we may neglect all cavity interaction terms diagonal in the eigenstates .
furthermore , when the energy differences are non - degenerate ( @xmath67 ) , we may also only keep coupling to the transition with energy @xmath68 .
then , @xmath69 thus , @xmath70 . keeping @xmath0 fixed , we plot @xmath55 and the expected error in operations due to relaxation ( calculated below ) . of immediate interest
is the regime of @xmath71 , where @xmath72 and @xmath73 .
this suggests that maximal coupling to the cavity occurs when @xmath74 , and @xmath54 are all comparable . , in units of bare coupling @xmath38 , as a function of field gradient @xmath54 and bias @xmath2 .
( b ) calculated probability of error in swapping a double - dot state with the resonator state due to charge - based relaxation using @xmath75ev and an inverse temperature of 10 @xmath76ev .
[ f : coupling],width=288 ]
we now demonstrate the favorable noise properties of the balanced system .
three kinds of error are considered : relaxation of the charge system in a time @xmath77 , additional dephasing of the charge system in a time @xmath78 , and decay of the cavity photon at a rate @xmath79 .
we assume the additional dephasing arises from low frequency fluctuations @xcite , due to changes of the electrostatic gates and magnetic field gradient , e.g. , as noise on @xmath2 and @xmath54 . for charge - based relaxation , coupling to a phonon bath with spectral density @xmath80 will lead to decay @xcite .
using the spin - boson model in the perturbative regime yields an overall error rate from relaxation and incoherent excitation : @xmath81 we estimate @xmath82 at @xmath83ev/@xmath84 using measured charge relaxation times @xcite . for a double quantum dot system in gaas and small @xmath26 , @xmath85 @xcite .
this indicates @xmath86 for the low energy limit .
reducing the resonant frequency lowers relaxation rates . near the optimal point , with @xmath87 2 ghz , @xmath88s .
additional dephasing ( @xmath89 term ) arises from variations of the energy gap , @xmath90 with @xmath91 .
we assume a high frequency cutoff of the noise at @xmath92 .
the contribution to error at first order arises from @xmath93 .
when the pre - factor @xmath94 is non zero , there is first - order dephasing .
in the rotating frame , the off - diagonal density matrix element evolves as @xmath95\ .
\label{e : decay}\ ] ] for example , when @xmath96 , the decay is gaussian with a time constant @xmath97 , where the characteristic time @xmath98 .
bare dephasing times of @xmath99 ns were observed for qubit frequencies of @xmath100 ghz @xcite .
however , longer @xmath101 times might be achieved for qubits at @xmath102 ghz through better high- and low - frequency filtering of electronics noise .
we will take @xmath103 ns for the remainder of the paper ; this limit arises from the combined effect of charge - based terms and nuclear spin - related dephasing @xcite . at the zero derivative point ( @xmath104 ) second - order terms
must be considered @xcite . to the leading order @xcite , we find dephasing occurs over a timescale @xmath105 .
more detailed calculations indicate that the underlying physics of the bath becomes important near zero - derivative points @xcite .
thus the optimal point combines the maximal coupling @xmath55 to the resonator with an extended @xmath106 time ( fig .
[ f : dephasing ] ) . in this regime , with @xmath107 ghz , @xmath108 .
we note that away from the optimal point , even assuming maximal coupling , @xmath109 .
for example , in the strongly biased regime ( @xmath110 ) , @xmath111 .
we conclude that only near the optimal point may the strong coupling regime be achieved .
( black ) , @xmath112 ( blue ) , and @xmath113 ( red ) .
the region of @xmath72 is marked in grey .
[ f : dephasing],width=288 ]
we now discuss potential applications .
coherent control of the system with little sensitivity to low frequency noise is achieved by driving the resonator on resonance with a microwave pulse .
we can describe this scenario by making the cavity state a coherent state , @xmath114 , letting us replace @xmath30 with @xmath115 .
we would expect rabi oscillations between @xmath116 and @xmath117 with rabi frequency @xmath118 .
low frequency components of the driving field ( dc offsets ) are greatly reduced because of the gap induced by @xmath0 and the insensitivity of the gap energy @xmath26 on variations of @xmath14 .
another technique is a quantum non - demolition measurement .
when the microwave transition is detuned from the cavity by @xmath119 , the evolution yeilds a qubit - dependent phase of @xmath120 ( for @xmath56 states respectively ) for an injected microwave field over a time @xmath121 ( limited by qubit decay ) . assuming that the noise level of the detector exceeds the background by about a factor of ten , the fidelity of such a qnd measurement of the encoded spin qubits is given by @xmath122 @xcite .
this single qubit qnd measurement can be completed in a time @xmath123s .
in addition , a wide variety of cavity qed quantum control techniques may be accessible . as an example , we focus on the swap operation , where the qubit state is `` swapped '' with a photonic state of the resonator . when there are no photons in the cavity and @xmath124 , the state @xmath125 is stationary , while @xmath126 oscillates with rabi frequency @xmath127 to the state @xmath128 ( we use @xmath129 to indicate @xmath130 photons in the cavity mode ) . when the time spent oscillating is @xmath131 , a quantum state of the singlet triplet system is mapped to the existence or absence of a cavity photon
: @xmath132 this process can be controlled by rapidly changing @xmath133 to and from zero .
in essence , we can convert one quantum bit ( the double - dot system ) to another quantum bit ( the cavity photon ) , which can then be used for a variety of quantum information tasks , such as long - distance quantum gates , quantum communication , and quantum repeaters @xcite .
furthermore , it may allow for coupling to other qubit systems , such as atoms , molecules @xcite , or superconducting qubits @xcite .
as an example , we now detail the expected errors for the swap operation .
both @xmath77 and @xmath79 contribute to the decay of the system , which lead to population of the state @xmath134 , while dephasing , @xmath135 , leads to phase errors in the transformation . including only relaxation terms , yields a probability of error in `` swap '' of @xmath136 .
in addition , pure dephasing ( @xmath135 ) leads to an error with probability @xmath137 .
the optimal point has @xmath138 , @xmath139 , and @xmath140 from noise in @xmath2 and @xmath54 .
optimizing against both noise sources suggests that for a bare dephasing time of @xmath141 ns , a cavity frequency of @xmath142 1.5 ghz is optimal for swap operations , with a probability of error @xmath143 .
in summary , we have shown how spin states of double quantum dots can be coupled directly to microwave resonators . an optimal operating point with maximum coupling to the resonator and minimum coupling to some sources of charge and spin noise is found .
this suggests a variety of powerful quantum control techniques may become possible for such a system . | we describe a coherent control technique for coupling electron spin states associated with semiconductor double - dot molecule to a microwave stripline resonator on a chip . we identify a novel regime of operation in which strong interaction between a molecule and a resonator can be achieved with minimal decoherence , reaching the so - called strong coupling regime of cavity qed .
we describe potential applications of such a system , including low - noise coherent electrical control , fast qnd measurements of spin states , and long - range spin coupling
. controlling quantum behavior of realistic solid - state systems is an intriguing challenge in modern science and engineering .
while over the past several decades much progress has been made in manipulation of isolated atomic and optical systems , only recently have solid - state systems demonstrated controlled , coherent behavior at a single quantum level @xcite .
the complex environment of solid state systems make them significantly more challenging to isolate and operate coherently .
furthermore , robust quantum control techniques analogous to those used in amo physics still need to be developed .
recently , a novel approach to quantum manipulation of spin - based quantum systems in semiconductor quantum dots has been proposed and experimentally realized @xcite .
this approach combines spin and charge manipulation to take advantage of the long memory times associated with spin states and , at the same time , to enable efficient readout and coherent manipulation of coupled spin states using intrinsic interactions .
while this allows one to consider architectures based on pulsed quasi - static electrical control and static magnetic field @xcite , for many purposes it is desirable to develop fast and robust quantum control techniques based on microwave manipulation . in this letter
we describe a technique for electrical coupling of electron spin states associated with semiconductor double - dot molecule to microwave stripline resonator on a chip @xcite .
the essential idea is to use an effective electric dipole moment associated with exchange coupled spin states of a double - dot molecule to couple to the oscillating voltage associated with a stripline resonator .
taking into account the main decoherence mechanisms for both spin and charge degrees of freedom , we identify an optimal point of operation in which strong interaction between molecule and resonator can be achieved with minimal decoherence , thereby enabling the strong coupling regime of cavity qed .
finally , we describe potential applications including low - noise coherent electrical control , fast qnd measurements of spin states , and long - range coupling of pairs of spins . before proceeding we
note the early proposals for achieving the strong coupling regime of cavity qed with cooper pair qubits and single electrons in double - dot molecules @xcite .
strong coupling of superconducting qubits to stripline cavities have been recently implemented in pioneering experiments by schoelkopf and co - workers @xcite .
this system has enabled a range of beautiful demonstrations ranging from control and measurement of the qubit through the resonator to swap of the qubit with the state of the resonator . finally ,
ideas similar to the present work have been proposed recently , using vertical quantum dot systems @xcite . , capacitively coupled to a transmission line resonator .
( b ) energy level diagram showing the ( 0,2 ) singlet and the four ( 1,1 ) two - spin states .
( c ) low energy spectrum in units of @xmath0 with small gradient @xmath1 : for large , negative @xmath2 , the ground state is @xmath3 .
the @xmath4 triplet ( red ) and @xmath5 triplets ( dashed grey ) are far from resonance with the @xmath6 triplet .
note the optimal point ( gray box ) occurs at the left - most avoided crossing .
[ f : picture],width=288 ] |
recently , the development of numerical schemes for strong as well as weak approximation of stochastic differential equations ( sdes ) has focused amongst others on runge kutta type schemes @xcite .
this is due to the increasing complexity of stochastic taylor expansions and the desire to avoid derivatives in higher order approximation schemes . in section [ sec
: srk - methods ] , a class of stochastic runge kutta ( srk ) methods due to rler @xcite for the weak approximation of it sde systems with a multi dimensional wiener process is considered .
this class contains as a special case the second order srk scheme proposed by platen @xcite as well as the class of srk methods proposed by tocino and vigo - aguiar @xcite .
order conditions for coefficients of the srk methods have been calculated by applying the multi - colored rooted tree analysis due to rler @xcite .
in contrast to earlier work on this topic , the aim of the present paper is to analyze these order conditions with the objective to determine a full classification of the coefficients for this class of srk methods .
a full classification for order one srk schemes with @xmath0 stage and for order one srk schemes with deterministic order two for @xmath1 stages as well as for second order srk schemes with @xmath2 stages is calculated in section [ sec : parameter - families ] .
further , some optimal schemes are derived from this classification in section [ sec : optimal - schemes ] by taking into account additional higher order conditions .
their performance is studied by some numerical examples in section [ sec : numerical - example ] .
+ + we denote by @xmath3 the solution of the @xmath4-dimensional it sde defined by @xmath5 with an @xmath6-dimensional wiener process @xmath7 and @xmath8 $ ] .
we assume that the borel - measurable coefficients @xmath9 and @xmath10 satisfy a lipschitz and a linear growth condition such that the existence and uniqueness theorem @xcite applies . in the following ,
let @xmath11 denote the @xmath12th column of the diffusion matrix @xmath13 for @xmath14 .
+ + let a discretization @xmath15 with @xmath16 of the time interval @xmath8 $ ] with step sizes @xmath17 for @xmath18 be given .
further , define @xmath19 as the maximum step size .
let @xmath20 denote the space of all @xmath21 fulfilling a polynomial growth condition and let @xmath22 if @xmath23 and @xmath24 for all @xmath25 and @xmath26 @xcite . a time discrete approximation @xmath27 converges weakly with order @xmath28 to @xmath29 as @xmath30 at time @xmath31 if for each @xmath32 exist a constant @xmath33 and a finite @xmath34 such that @xmath35 holds for each @xmath360,\delta_0[\,$ ] .
we consider stochastic runge kutta methods as proposed in @xcite for the weak approximation of sde ( [ st - lg - sde - ito-1 ] ) .
therefore , the @xmath4-dimensional approximation process @xmath37 of an explicit @xmath38-stage srk method is defined by @xmath39 and @xmath40 for @xmath18 with stage values @xmath41 for @xmath42 and @xmath43 . here ,
@xmath44 and @xmath45 , @xmath46 for @xmath47 with @xmath48 for @xmath49 are the vectors and matrices of coefficients of the srk method .
we choose @xmath50 for @xmath51 with a vector @xmath52 @xcite . in the following , the product of column vectors is defined component - wise .
the coefficients of the srk method ( [ srk - method - ito - wm - allg01 ] ) are determined by the following butcher tableau : [ cols="^,^,^,^ " , ] ) and sde ( [ simu : dm - sde]).,title="fig:",width=257 ] ) and sde ( [ simu : dm - sde]).,title="fig:",width=257 ] ) and sde ( [ simu : dm - sde]).,title="fig:",width=257 ] ) and sde ( [ simu : dm - sde]).,title="fig:",width=257 ] + + due to the results in figure [ bild001 ] , we can see that for both test equations the so called optimal srk scheme rdi1wm attains much better orders of convergence than the well known order one em scheme .
the same holds for the optimal srk schemes rdi3wm and rdi4wm compared to the order two schemes exem and pl1wm .
clearly , the optimal srk schemes rdi1wm , rdi3wm and rdi4wm need some additional computational effort compared to the schemes em , exem and pl1wm , respectively .
therefore , we take the number of evaluations of the drift function @xmath53 and of each diffusion function @xmath54 , @xmath55 , as well as the number of random variables that have to be simulated as a measure for the computational effort . then we can compare the computational effort versus the errors of the analyzed schemes .
the results are presented in figure [ bild002 ] , and again , rdi1wm performs much better than the scheme em for both test equations .
further , rdi3wm yields similar results like rdi4wm which is for higher precisions slightly better than pl1wm and significantly better than the scheme exem for the test equation ( [ simu : nonlinear - sde2 ] ) . considering the multi - dimensional test equation ( [ simu : dm - sde ] ) ,
the scheme rdi3wm is again close to rdi4wm which performs for higher precisions slightly better than exem .
however both optimal srk schemes rdi3wm and rdi4wm are significantly better than the srk scheme pl1wm .
in the present work , a full classification of the coefficients for a class of explicit srk methods of order @xmath56 for @xmath0 and order @xmath57 for @xmath1 stages as well as for the orders @xmath58 and @xmath59 with @xmath2 stages is calculated .
based on this classification , coefficients for so called optimal srk schemes are determined by considering additional higher order conditions .
optimal coefficients for srk methods of order @xmath57 , @xmath58 and @xmath59 are calculated and similarly to the deterministic setting @xcite , better convergence results are expected for these schemes in general . finally , the srk schemes based on optimal coefficients are applied to some test equations . here , it turned out that the proposed optimal srk schemes attain higher orders of convergence than the well known schemes under consideration and they also perform very well if the computational effort is taken into account .
+ + for future research , it would be interesting to extend the presented classification to diagonal or fully implicit srk methods .
further , the given classification may be applied in order to determine coefficients for srk methods with optimal stability properties .
the authors are very grateful to the unknown referees for their fruitful comments and suggestions . | in the present paper , a class of stochastic runge kutta methods containing the second order stochastic runge
kutta scheme due to e. platen for the weak approximation of it stochastic differential equation systems with a multi dimensional wiener process is considered .
order one and order two conditions for the coefficients of explicit stochastic runge kutta methods are solved and the solution space of the possible coefficients is analyzed .
a full classification of the coefficients for such stochastic runge kutta schemes of order one and two with minimal stage numbers is calculated .
further , within the considered class of stochastic runge kutta schemes coefficients for optimal schemes in the sense that additionally some higher order conditions are fulfilled are presented . and stochastic runge kutta method , stochastic differential equation , classification , weak approximation , optimal scheme + msc 2000 : 65c30 , 60h35 , 65c20 , 68u20 |
the success of the dynamical triangulation model in two dimensions summarized by its equivalence to the liouville theory in the continuum limit motivated intensive studies of the model in higher dimensions through monte carlo simulation .
it turned out , however , that with an action obtained by naive discretization of the einstein - hilbert action , the phase transition observed is of first order both in three dimensions and in four dimensions @xcite . in lattice theories in general , it is often preferable ( _ e.g .
_ improved actions to reduce the finite lattice spacing effects ) or even necessary ( _ e.g . _ the wilson term to eliminate doublers ) to modify the lattice actions from that obtained by naive discretization of the action in the continuum .
we may therefore hope to obtain a second order phase transition in the present case by adding a local term to the naive einstein - hilbert action . _ _ _ _ the new term we consider is motivated by a possible contribution from the measure for the path integral over the metric and has been first studied in four dimensions @xcite .
recently this modified action has been studied in three dimensions by the monte carlo renormalization group and a possibility was suggested that the first order phase transition terminates at a finite coefficient of the new term @xcite .
is the minus of the @xmath0 in ref .
@xcite ] we clarify this possibility by high statistics monte carlo simulation using multi - canonical technique .
the naive discretization of the einstein - hilbert action in the dynamical triangulation model in @xmath1 dimensions gives @xmath2 where @xmath3 and @xmath4 are the number of vertices and @xmath5simplices respectively in the simplicial manifold .
we modify the action as @xmath6 \ , , \end{aligned}\ ] ] where @xmath7 is the new coupling constant , and @xmath8 is the number of @xmath5simplices that contain the node @xmath9 . the summation in ( [ eq : muterm ] ) is taken over all the nodes in the manifold .
this term @xcite can be considered as the contribution from the path integral measure @xmath10 , since @xmath8 can be interpreted as the `` local volume '' , which corresponds to @xmath11 in the continuum . in the following we consider the canonical ensemble with fixed @xmath12 given by the partition function : @xmath13 where the summation in ( [ eq : partfunc1 ] ) is taken over all possible triangulations of @xmath1-dimensional sphere that contain @xmath4 @xmath1-simplices .
the @xmath14 is the partition function for fixed @xmath15 , which can be given by @xmath16 where the summation is now taken over all possible triangulations that contain @xmath4 @xmath1-simplices and @xmath15 vertices .
the details of the numerical simulation will be reported elsewhere @xcite .
in order to investigate the phase diagram with the modified action , we measure the microcanonical entropy @xmath17 defined by @xmath18 from which we calculate the microcanonical inverse temperature @xmath19 by @xmath20 precise measurement of the microcanonical entropy near the first order phase transition point is possible by using the multi - canonical technique .
this was previously done for the unmodified action with @xmath21 in ref .
@xcite , and the critical exponents consistent with the first order phase transition have been successfully extracted by the finite - size scaling analysis . in fig .
[ fig : k ] , we show the microcanonical inverse temperature @xmath19 for several values of @xmath7 . note that the node distribution is given for each @xmath22 and @xmath0 by @xmath23 and the double peak structure in the node distribution , which is a signal of the first order phase transition , can be probed directly by looking at the microcanonical inverse temperature @xmath19 .
the double peaks appear if and only if ( i ) @xmath19 has both local maximum and local minimum and ( ii ) @xmath24 is set to a value between the local minimum and the local maximum .
one can see that the local minimum and the local maximum observed for @xmath21 @xcite merge at @xmath25 , beyond which the @xmath19 becomes a monotonous function of @xmath15 .
this suggests that the first order phase transition ceases to exist at @xmath26 , beyond which it becomes a crossover .
for each @xmath7 below @xmath26 , we define the critical coupling @xmath27 of the phase transition as the position of the peak of the node susceptibility @xmath28 . the end point of the phase transition line
is given by @xmath29 for @xmath30 , and @xmath31 for @xmath32 . in fig .
[ fig : freeenergy ] , we plot the free energy @xmath33 , which is nothing but the minus log of the node distribution , as a function of @xmath15 near @xmath34 for @xmath21 and @xmath35 .
the two local minima observed for @xmath36 merge into one at @xmath37 .
the free energy curve at @xmath38 has a wide flat bottom , which suggests the existence of a massless mode .
in order to examine a possible continuum limit at the end point , we measure the boundary area distribution @xmath39 , which is the number of boundaries with the area @xmath40 at the geodesic distance @xmath41 .
the corresponding quantity in two dimensions called the loop length distribution is calculated analytically and is found to possess a continuum limit @xcite . the scaling behavior
expected from this result has been correctly reproduced by numerical simulation @xcite . in fig .
[ fig : bad ] , we plot the boundary area distribution at the end point of the first order phase transition line as a function of @xmath42 .
one can see a reasonable scaling behavior .
the power of @xmath43 in the scaling variable @xmath42 implies that the fractal dimension is 3 .
we would like to remark that the scaling behavior we observe is much better than the one that has been claimed to exist with the unmodified action @xcite .
our results suggest that the first order phase transition becomes second order at the end point , where one can obtain a continuous theory .
a natural question to be asked is what the continuum theory is .
one may suspect that we can not construct any continuum theory in three - dimensional quantum gravity , since we have no physical degrees of freedom .
this argument is too naive , however .
an example of field theory which has no physical degrees of freedom that is still completely well - defined is two - dimensional yang - mills theory .
the fact that we have to fine - tune two parameters to obtain the continuum limit implies that there are two relevant operators around the fixed point .
a possible interpretation of the continuum theory is therefore the @xmath44 gravity .
the observation that the fractal dimension at the fixed point is approximately three is also suggestive of this interpretation . the fractal dimension we extracted , however ,
is still preliminary , and the value might increase as we increase the system size .
needless to say , a lot of work is yet to be done to understand the nature of the continuum theory . in 4d
, the situation is different from the 3d case in that singular vertices which are shared by huge number of @xmath45-simplices exist in the strong coupling phase @xcite .
we can add a local term in the action that eliminates these singular vertices , if one wishes , and then the situation becomes more or less the same @xcite .
we therefore expect that the first order phase transition can be changed into second order by modifying the action . | we investigate a new phase structure of the three - dimensional dynamical triangulation model with an additional local term by monte carlo simulation .
we find that the first order phase transition observed for the naive einstein - hilbert action terminates at a finite coefficient of the new term .
the phase transition turns into second order at the endpoint , beyond which it becomes a crossover . # |
99 m. fukugita and t. yanagida , phys . lett.*b174 * , 45(1986 ) . g. branco , t. morozumi , b. nobre and m. rebelo , nucl . phys .
* b 617 * , 475 ( 2001 ) . t. endoh , s. kaneko , s. k. kang , t. morozumi , and m. tanimoto , hep - ph 0209020 . | how is cp violation of low energy related to cp violation required from baryon number asymmetry ?
we give an example which shows a direct link between cp violation of neutrino oscillation and baryogenesis through leptogenesis .
when the sphaleron process is active , the sum of baryon number and lepton number is not a conserved quantity ; @xmath0 anomaly @xmath1 .
therefore , the evolution equation of baryon number and lepton number becomes a coupled equation .
the present baryon number can be written in terms of the `` initial '' lepton number and baryon number as ; b_now=(b - l)_ini.+ ( b+l)_ini .
( b - l)_ini .. fukugita and yanagida proposed `` baryogenesis without grand unification '' @xcite .
then , @xmath2 while the lepton number production is possible because their model includes the heavy majorana neutrinos and their decays lead to the lepton number asymmetry ; l_ini ~- .
the purpose of my talk is give a specific example which shows `` a direct link '' between the size and sign of baryon number and cp violation in neutrino oscillatio ; p ( _ _ e)-p(| _ |_e ) ~j = i m ( k_e1 k_e2^)(k_1 k_2^)^ , [ j ] where @xmath3 is mns matrix .
|l_l _ k _ l. @xmath4 is related to mns matrix . by taking the basis in which the mass matrix for the heavy majorana neutrinos and charged leptons are real diagonal
, the mns matrix can be obtained through the diagonalization of @xmath5 on the other hand the lepton number asymmetry in the same basis is given as ; _ 1 & = & = - , where @xmath6 ( tev ) and we take @xmath7 .
we consider `` the minimal seesaw '' which generates @xmath8 and @xmath9 simultaneously .
the minimal model is @xmath10 model @xmath11 3 light neutrinos @xmath12 and 2 heavy majorana neutrionos @xmath13 and @xmath14 .
= & & |l^i m_l_i l^i + |^i m_d_ij n_rj + |(n_rj)^c m_j n_rj , where @xmath15 .
dirac mass term @xmath16 matrix , m_d&= & ( cc m_11 & m_12 + m_21 & m_22 + m_31 & m_32 + ) = u_l ( cc 0 & 0 + m_2 & 0 + 0 & m_3 + ) v_r , where @xmath17 ( @xmath18 ) is a @xmath19 ( @xmath20 ) unitary matrix .
the important properties of the model are one light neutrino is exactly massless : @xmath21=0 $ ] and there are three cp violating phases since there are @xmath22 imaginary elements in @xmath23 . by writing @xmath18 as follows ; v_r&= & ( cc c_r & s_r + -s_r & c_r ) ( cc ( -i ) & 0 + 0 & ( i ) ) , we can show the leptogenesis is determined by a cp phase @xmath24 , _ 1 & & ~- im[(m_d^ m_d)_12 ^ 2 ] + & & ~-(m_3 ^ 2-m_2 ^ 2)^2 s_r^2 c_r^2 . on the other hand , cp violation in neutrino oscillation , @xmath25 in eq.([j ] ) , depends on all three cp violating phases because @xmath3 is determined by the diagonalization as @xmath26 and it is sensitive to all cp phases in @xmath23 .
we give an example for the model in which j is determined by leptogenesis phase @xmath24 .
suppose @xmath27 is a real orthogonal matrix @xcite as @xmath28 .
mns matrix @xmath3 has the following form ; k= o_23 o_12 ( ccc 1 & 0 & 0 + 0 & & + 0 & & + ) p , where @xmath29 is a diagonal phase matrix which is not relevant for @xmath25 .
@xmath30 and @xmath31 are determined through the diagonalization of @xmath32 .
therefore @xmath30 and @xmath31 do not depend on @xmath17 at all .
it depends on @xmath33 besides @xmath34 and @xmath35 .
those four quantities can be determined from the heavy majorana decay width , @xmath36 and @xmath37 and two light neutrino masses scales @xmath38 and @xmath39 .
taking @xmath40 , @xmath41 , we have ; & & k= ( ccc & & + - & & + & - & + ) p. it is easy to see @xmath42 .
we can show there is a correlation between @xmath31 and @xmath24 .
if @xmath24 vanishes , @xmath43 becomes a real symmetric matrix .
then , @xmath31 must vanish in the limit . on the other hand ,
if @xmath24 does not vanish , the sign of @xmath24 determines the sign of lepton number asymmetry which in turn determines the excess of matter ( anti - matter ) . in our model ,
the sign of @xmath25 reflects the sign of @xmath24 .
we found the following correlation holds .
b -b
l -l _
r -_r - j -j . in fig.1 and fig.2
, we show the correlation @xmath44 , where @xmath45 and @xmath46 , by identifying the two light neutrino masses as @xmath47 ev and @xmath48(lma ) .
we take @xmath49 .
the figures are obtained for fixed @xmath50 ( ev ) and varying @xmath51 . .
@xmath52 ( ev ) . ] .
@xmath52 ( ev ) . ]
* summary * + * leptogenesis phase ( @xmath24 ) certainly affects the neutrino oscillation cp violation through ( @xmath31 ) .
however , if we measure @xmath25 only , we can not distinguish the phase for the leptogenesis ( @xmath31 ) from the other phases in @xmath17 because only a certain combination of them appear .
the isolation must be done using some other quantities , double @xmath53 decay etc @xcite .
* we show the correlation between cp violating phase for leptogenesis and cp violating phase for neutrino oscillation for a specific choice of @xmath17 . |
dusty circumstellar disks have been the focus of intense observational interest in recent years , largely because they are thought to be the birthplaces of planetary systems .
these observational efforts have yielded many new insights on the structure and evolution of these disks . in spite of major developments in
spatially resolved observations of these disks , much of our knowledge of their structure is still derived from spatially _
un_resolved spectroscopy and spectral energy distributions ( seds ) .
the interpretation of this information ( as well as spatially resolved data ) requires the use of theoretical models , preferentially with as much realism and self - consistency as possible .
such disk models have been developed and improved over many years .
when they are in reasonable agreement with observations they can also serve as a background onto which other processes are modeled , such as chemistry , grain growth , and ultimately the formation of planets .
this chapter reviews the development of such self - consistent disk structure models , and discusses the current status of the field .
we restrict our review to models primarily aimed at a comparison with observations
. we will start with a concise resum of the formation and viscous evolution of disks ( section [ sec - viscevol ] ) .
this sets the radial disk structure as a function of time .
we then turn our attention to the vertical structure , under the simplifying assumption that the gas temperature equals the dust temperature everywhere ( section [ sec - diskstruct ] ) . while this assumption is valid in the main body of the disk , it breaks down in the disk surface layers .
the formation of stars and planetary systems starts with the gravitational collapse of a dense molecular cloud core .
since such a core will always have some angular momentum at the onset of collapse , most of the infalling matter will not fall directly onto the protostar , but form a disk around it while matter falls onto the disk , viscous stresses within the disk will transport angular momentum to its outer regions . as a consequence of this ,
most of the disk matter moves inward , adding matter to the protostar , while some disk matter moves outward , absorbing all the angular momentum ( _ lynden - bell and pringle _ , @xcite ) . during its formation and evolution
a disk will spread out to several 100 au or more ( _ nakamoto and nakagawa _
, @xcite , henceforth nn94 ; _ hueso and guillot _ , @xcite , henceforth hg05 ) .
this spreading is only stopped when processes such as photoevaporation ( this chapter ) , stellar encounters ( _ scally and clarke _ , @xcite ; _ pfalzner et al .
_ , @xcite ) or a binary companion ( _ artymowicz and lubow _ , @xcite ) truncate the disk from the outside . during the collapse phase , which lasts a few@xmath0 years
, the accretion rate within the disk is very high ( @xmath1 @xmath2yr ) , but quickly drops to @xmath3 @xmath4yr once the infall phase is over ( nn94 , hg05 ) .
the optical and ultraviolet excess observed from classical t tauri stars ( cttss ) and herbig ae / be stars ( haebes ) confirms that this on - going accretion indeed takes place ( _ calvet et al .
_ , @xcite , and references therein ) . in fig .
[ fig - hueso ] we show the evolution of various disk and star parameters
. evolution of various disk and star quantities as a function of time after the onset of collapse of the cloud core ( after _ hueso and guillot _ , @xcite ) .
solid lines : stellar mass ( upper ) and disk mass ( lower ) .
dotted line : accretion rate _ in the disk_. , width=302 ] an issue that is still a matter of debate is what constitutes the viscosity required for disk accretion , molecular viscosity is too small to account for the observed mass accretion rates .
turbulent and magnetic stresses , however , can constitute some kind of anomalous viscosity .
the magnetorotational instability ( mri ) , present in weakly magnetized disks , is the most accepted mechanism to drive turbulence in disks and transport angular momentum outwards ( _ balbus and hawley _
, @xcite ; _ stone and pringle _ , @xcite ; _ wardle _ , @xcite , and references therein ) .
there is a disk region ( 0.2 @xmath5 4 au , for typical ctts disk parameters according to _ dalessio et al .
_ , @xcite ) in which the ionization fraction is smaller than the minimum value required for the mri . neither thermal ionization ( requiring a temperature higher than 1000 k ) , cosmic ray ionization ( requiring a mass surface density smaller than @xmath6 100 g/@xmath7 ) ( _ jin _ , @xcite ; _ gammie _ , @xcite ) , nor x - rays ( _ glassgold et al .
_ , @xcite , @xcite ) are able to provide a sufficient number of free electrons to have mri operating near the midplane . _
gammie _ ( @xcite ) proposed a layered accretion disk model , in which a `` dead zone '' is encased between two actively accreting layers .
the precise extent of this dead zone is difficult to assess , because the number density of free electrons depends on detailed chemistry as well as the dust grain size distribution , since dust grains tend to capture free electrons ( _ sano et al .
_ , @xcite , and references therein ) .
if the disk dust is like the interstellar dust , the mri should be inhibited in large parts of the disk ( _ ilgner and nelson _ , @xcite ) , though this is still under debate ( e.g. _ semenov et al .
_ , @xcite ; ) .
there are also other ( non - magnetic ) mechanisms for anomalous viscosity , like the baroclinic instability ( _ klahr and bodenheimer _ , @xcite ) or the shear instability ( _ dubrulle et al .
_ , @xcite ) , which are still subject to some controversy ( see the recent review by _ gammie and johnson _ , @xcite ) .
angular momentum can also be transferred by global torques , such as through gravitational spiral waves ( _ tohline and hachisu _
, @xcite ; _ laughlin and bodenheimer _ , @xcite ; _ pickett et al . _ , @xcite and references therein ) or via global magnetic fields threading the disk ( _ stehle and spruit _ , @xcite ) ,
possibly with hydromagnetic winds launched along them ( _ blandford and payne _ , @xcite ; _ reyes - ruiz and stepinski _ , @xcite ) .
to avoid having to solve the problem of viscosity in detail , but still be able to produce sensible disk models , _ shakura and sunyaev _ ( @xcite ) introduced the `` @xmath8-prescription '' , based on dimensional arguments . in this recipe the vertically averaged viscosity @xmath9 at radius @xmath10 is written as @xmath11 , where @xmath12 is the pressure scale height of the disk and @xmath13 is the isothermal sound speed , both evaluated at the disk midplane where most of the mass is concentrated .
the parameter @xmath8 summarizes the uncertainties related to the sources of anomalous viscosity , and is often taken to be of the order of @xmath14 for sufficiently ionized disks . from conservation of angular momentum ,
the mass surface density @xmath15 of a _ steady _ disk ( i.e. with a constant mass accretion rate @xmath16 ) , for radii much larger than the disk inner radius @xmath17 , can be written as @xmath18 . with @xmath19 , where @xmath20 is the keplerian angular velocity
, we see that for @xmath21 , @xmath22 as shown by fu ori and ex lupi type outbursts ( _ gammie and johnson _
, @xcite , and references therein ) .
these outbursts can have various triggering mechanisms , such as thermal instability ( _ kawazoe and mineshige _ , @xcite ; _ bell and lin _ , @xcite ) ; close passage of a companion star ( _ bonnell and bastien _ @xcite ; _ clarke and syer _ , @xcite ) ; mass accumulation in the dead zone followed by gravitational instability ( _ gammie _ , @xcite ; _ armitage et al .
_ , @xcite ) .
disks are therefore quite time - varying , and constant @xmath8 steady disk models should be taken as zeroth - order estimates of the disk structure . given the challenges of understanding the disk viscosity from first principles , attempts have been made to find observational constraints on disk evolution ( _ ruden and pollack _
, @xcite ; _ cassen _ , @xcite ; _ hartmann et al . _ , @xcite ; _ stepinski _ , @xcite ) .
for example , _ hartmann et al . _ ( @xcite ) study a large sample of cttss and find a decline in mass accretion rate with time , roughly described as @xmath23 , which they compare to the analytic similarity solutions of _ lynden - bell and pringle _ ( @xcite ) for the expanding disk .
with the radial structure following from accretion physics , as described above , the next issue is the vertical structure of these disks .
many authors have modeled this with full time - dependent 2d/3d ( magneto / radiation- ) hydrodynamics ( e.g. , _ boss _ , @xcite , @xcite ; _ yorke and bodenheimer _ , @xcite ; _ fromang et al .
_ , @xcite ) .
while this approach is obviously , it suffers from large computational costs , and often requires strong simplifying assumptions in the radiative transfer to keep the problem tractable . for comparison to these models
are therefore less practical .
the main objective of the models described in this section is the determination of the density and temperature structure of the disk . for a given surface density @xmath24 , and a given temperature structure @xmath25
( where @xmath26 is the vertical coordinate measured upward from the midplane ) the vertical density distribution @xmath27 can be readily obtained by integrating the vertical equation of hydrostatics : @xmath28 where @xmath29 with @xmath30 .
since the main source of opacity is the dust , most models so far make the assumption that the gas temperature is equal to the dust temperature , the temperature of the disk is set by a balance between heating and cooling .
the disk cools by thermal emission from the dust grains at infrared wavelengths .
this radiation is what is observed as infrared dust continuum radiation from such disks .
line cooling is only a minor coolant , and only plays a role for @xmath31 when gas and dust are thermally decoupled .
dust grains can be heated in part by radiation from other grains in the disk .
the iterative absorption and re - emission of infrared radiation by dust grains in the disk causes the radiation to propagate the disk in a diffusive way .
viscous dissipation of gravitational energy in the disk due to accretion .
once the temperature structure is determined , the sed can be computed .
the observable thermal emission of a dusty disk model consists of three wavelength regions .
the main portion of the energy is emitted in a wavelength range depending on the minimum and maximum temperature of the dust in the disk .
we call this the `` energetic domain '' of the sed , which typically ranges from 1.5 @xmath32 m to about 100 @xmath32 m . at shorter wavelength
the sed turns over into the `` wien domain '' . at longer wavelengths the sed turns over into the `` rayleigh - jeans domain '' , a steep , nearly powerlaw profile with a slope depending on grain properties and disk optical depth ( ) .
, width=302 ] is a perfectly flat disk being irradiated by the star due to the star s non - negligible size ( _ adams and shu _ , @xcite ; _ friedjung _ , @xcite ) .
the stellar radiation impinges onto the flat disk under an irradiation angle @xmath33 ( with @xmath34 the stellar radius ) . neglecting viscous dissipation , the effective temperature of the disk
is set by a balance between the irradiated flux ( with @xmath35 the stellar luminosity ) and blackbody cooling , which yields @xmath36 .
the energetic domain of its sed therefore has a slope of @xmath37 with @xmath38 , follows from the that any disk with @xmath39 has an sed slope of @xmath40 .
observations of cttss , however , show sed slopes typically in the range @xmath41 to 1 ( _ kenyon and hartmann _ , @xcite ) , i.e. much less steep .
the seds of herbig ae / be stars show a similar picture , with a somewhat larger spread in @xmath42 .
_ meeus et al .
_ ( @xcite , henceforth m01 ) divide the seds of herbig ae / be stars into two groups : those with strong far - infrared flux ( called ` group i ' , having slope @xmath43 ... @xmath44 ) and those with weak far - infrared flux ( called ` group ii ' , having slope @xmath45 ...
@xmath46 ) .
all but the most extreme group ii sources have a slope that is clearly inconsistent with that of a flat disk .
it was recognized by _ kenyon and hartmann _ ( @xcite ) that a natural explanation for the strong far - infrared flux ( i.e. shallow sed slope ) of most sources is a flaring geometry of the disk . the flaring geometry a significant portion of the stellar radiation at large radii where the disk is cool , where @xmath47 is the height above the midplane where the disk becomes optically thick to the impinging stellar radiation .
a closer look at the physics of an irradiation - dominated disk ( be it flat or flared ) reveals that its surface temperature is generally higher than its interior temperature ( _ calvet et al .
_ , @xcite ; _ malbet and bertout _ , @xcite ; ) . because of the shallow incidence angle of the stellar radiation @xmath48 , the _ vertical _ optical depth of this warm surface layer is very low
as a consequence , the thermal radiation from these surface layers produces dust features in _
emission_. this is exactly what is seen in nearly all non - edge - on t tauri and herbig ae / be star spectra ( e.g. m01 ; _ kessler - silacci et al .
_ , @xcite ) , indicating that these disks are nearly always dominated by irradiation .
armed with the concepts of disk flaring and hot surface layers , a number of authors published detailed with direct applicability to observations .
the aforementioned cg97 model ( with refinements described in _
chiang et al .
_ , @xcite ) is a two - layer model for the interpretation of seds and dust emission features from _ non - accreting _ ( ` passive ' ) disks .
_ lachaume et al . _ ( @xcite ) extended it to include viscous dissipation .
the models by _
dalessio et al .
_ ( @xcite ) solve the complete 1 + 1d disk structure problem including irradiation and viscous dissipation ( using the @xmath8 prescription ) .
the main input parameters are a global ( constant ) mass accretion rate @xmath49 and @xmath8 .
the surface density profile @xmath24 is calculated self - consistently . , width=302 ] vertical temperature distribution of an irradiated @xmath8-disk at 1 au , for a fixed @xmath15 ( chosen to be that of a disk model with @xmath50yr for @xmath51 ) , but varying @xmath49 , computed using the models of _ dalessio et al . _ ( @xcite ) .
, width=302 ] models of describe the seds of cttss reasonably well .
however , _
dalessio et al .
_ ( @xcite ) argue that they tend to slightly overproduce far - infrared flux and have too thick dark lanes in images of edge - on disks .
they also show that the percentage of expected edge - on disks appears to be overpredicted .
they suggest that dust sedimentation could help to solve this problem . _
chiang et al . _ ( @xcite ) find similar results for a subset of their herbig ae / be star sample : the meeus group ii sources ( see also cg97 ) .
they fit these sources by dust settling through a reduction of the disk surface height .
self - consistent computations of dust sedimentation produce similar seds and confirm the dust settling idea ( _ miyake and nakagawa _
, @xcite ; _ dullemond and dominik _ , @xcite , henceforth dd04b ; _ dalessio et al . _ , @xcite ) .
the disk thickness and far - infrared flux can also be reduced by grain growth ( _ dalessio et al .
_ , @xcite ; _ dullemond and dominik _ , @xcite ) . from comparing infrared and ( sub-)millimeter spectra of the same sources ( _ acke et al .
_ , @xcite ) , it is clear that the ( sub-)millimeter usually require mm - sized grains in the outer regions of the disk , while infrared dust emission features clearly prove that the disk surface layers are dominated by grains no larger than a few microns ( see ) .
it appears that a bimodal size distribution can fit the observed spectra : the very inner part of the disk is dust - free due to dust sublimation ( see for a discussion of this region ) . the dusty part of the disk can therefore be expected to have a relatively abrupt inner edge at about @xmath52 au for a 50@xmath53 star ( scaling roughly with @xmath54 ) .
if the gas inward of this dust inner rim is optically thin , which mostly the case ( _ muzerolle et al .
_ , @xcite ) , then this dust inner rim is illuminated by the star at a @xmath5590 degree angle , and is hence expected to be much hotter than the rest of the disk behind it . .
this is a natural explanation , since dust sublimation occurs typically around 1500 k , and a 1500 k blackbody bump fits reasonably well to the near - infrared bumps in those sources . _
tuthill et al . _ ( @xcite ) independently discovered a bright half - moon ring around the herbig be star lkha-101 , which they attribute to a bright inner disk rim due to dust sublimation .
_ dullemond et al .
_ ( @xcite ; henceforth ddn01 ) extended the cg97 model to include such a puffed - up rim , and _ dominik et al .
_ ( @xcite ) showed that the meeus sample of herbig ae / be stars can be reasonably well fitted by this model .
however , for meeus group ii sources these fits required relatively small disks ( see , however , section [ subsec-2dtrans ] ) .
the initial rim models were rather simplified , treating it as a vertical blackbody ` wall ' ( ddn01 ) .
_ isella and natta _ ( @xcite ) improved this by noting that the weak dependence of the sublimation temperature on gas density is enough to strongly round off the rim .
rounded - off rims appear to be more consistent with observations than the vertical ones : .
there is still a worry , though , whether the rims can be high enough to fit sources with a strong near - infrared bump . with near - infrared interferometry the rim can be spatially resolved , and thus the models can be tested .
the measurements so far do not yet give images , but the measured ` visibilities ' can be compared to models .
in this way one can measure the radius of the rim ( e.g. , _ monnier et al .
_ , @xcite ; _ akeson et al .
_ , @xcite ) and its inclination ( e.g. , _ eisner et al .
_ , @xcite )
. moreover it can test whether indeed the near - infrared emission comes from the inner rim of the dust disk in the first place ( some doubts have been voiced by _
vinkovic et al .
_ , @xcite ) .
we refer to the for a more in - depth discussion of interferometric measurements of disks .
the inner rim model has so far been mainly applied to herbig ae / be stars because the rim appears so apparent in the ( nir ) .
but _ muzerolle et al . _
( @xcite ) showed that it also applies to t tauri stars . in that case , however , the luminosity from the magnetospheric accretion shock is required in addition to the stellar luminosity to power the inner rim emission .
, width=302 ] , width=302 ] the models described so far are all based on an approximate 1 + 1d ( or two - layer ) irradiation - angle description . in reality
the structure of these disks is 2-d , if axisymmetry can be assumed , and 3-d if it can not . over the last 10 years
many multi - dimensional dust continuum radiative transfer programs and algorithms were developed for this purpose ( e.g. , _ whitney et al .
_ , @xcite ; _ lucy et al .
_ , @xcite ; _ wolf et al .
_ , @xcite ; _ bjorkman and wood _ , @xcite ; _ nicolinni et al .
_ , @xcite ; _ steinacker et al .
_ , @xcite ) .
most applications of these codes assume a given density distribution and compute spectra and images .
there is a vast literature on such applications which we will not review here ( see ) .
but there is a trend to include the self - consistent vertical density structure into the models by iterating between radiative transfer and the vertical pressure balance equation ( _ nomura _ , @xcite ; _ dullemond _ , @xcite , henceforth d02 ; _ dullemond and dominik _ , @xcite , henceforth dd04a ; _ walker et al .
_ , @xcite ) .
although , there is an obvious interest in direct observations of the gas . these disks .
moreover , it is important to estimate how long disks remain gas - rich , and whether this is consistent with the formation time scale of gas giant planets ( _ hubickyj et al .
_ , @xcite ) .
unfortunately , gas lines often probe those regions of disks in which the gas temperature is difficult to compute .
the disk models we described above assume that the gas temperature in the disk is always equal to the local dust temperature .
while this is presumably true for most of the matter deep within optically thick disks , in the tenuous surface layers of these disks ( or throughout optically thin disks ) the densities become so low that the gas will thermally decouple from the dust .
the gas will acquire its own temperature , which is set by a balance between various heating- and cooling processes .
these processes depend strongly on the abundance of various atomic and molecular species , which , for their part , depend strongly on the temperature .
the gas temperature , density , chemistry , radiative transfer and radiation environment are therefore intimately intertwined and have to be studied as a .
this greatly complicates the modeling effort , and the first models which study this in detail have only recently been published . on stationary models , i.e. models that are in chemical , thermal and hydrostatic equilibrium .
for the tenuous regions of disks the chemical time scales are short enough that this is valid , in contrast to the longer chemical time scales deeper in the disk ( e.g. , _ aikawa and herbst _ , ; _ willacy et al .
_ , @xcite ) .
the models constructed so far either solve the gas temperature / chemistry for a _ fixed _ gas density structure , or include the gas density in the computation to obtain a self - consistent thermo - chemical - hydrostatic structure .
the physics and chemistry of the surface layers of protoplanetary disks strongly resembles that of photon dominated regions ( pdrs , _ tielens and hollenbach _
, @xcite ; _ yamashita et al . _
@xcite ) . in those the gas temperature
generally greatly exceeds the dust temperature . but
the dust - gas coupling gradually takes over the gas temperature balance as one gets deeper into the disk , typically beyond a vertical column depth of , the uppermost surface layer contains mostly atomic and ionized species , since the high uv irradiation effectively dissociates all molecules ( _ aikawa et al . _ , ) .
the photochemistry is driven by the stellar irradiation and/or in case of nearby o / b stars , by external illumination . in flaring disk models
, the stellar radiation penetrates the disk under an irradiation angle @xmath56 like the one described in the previous section .
this radiation gets diluted with increasing distance from the central star and attenuated by dust and gas along an _ inclined _ path into the disk .
the stellar radiation therefore penetrates less deep into the disk than external uv radiation .
the thermal balance of the gas in disks is solved by equating all relevant heating and cooling processes . for this gas thermal balance equation ,
a limited set of key atomic and molecular species is sufficient : e.g. , h@xmath57 , co , oh , h@xmath57o , . for most atoms and molecules , the statistical equilibrium equation has to include the pumping of the fine structure and rotational levels by the cosmic background radiation , from the optical depth of the line ( similar to the approach of _ tielens and hollenbach _ , @xcite for pdrs ) .
the optical depth used for this escape probability is the line optical depth in the _ vertical _ direction where the photons most readily escape . of the most critical ingredients of these models
is the uv and x - ray radiation field ( stellar and external ) , in the literature the far ultraviolet radiation field is often represented by a single parameter @xmath58 describing the integrated intensity between 912 and 2000 normalized to that of the typical interstellar radiation field @xcite .
however , several papers have shown the importance of a more detailed description of the radiation field for calculations of the chemistry and the gas heating / cooling balance ( _ spaans et al . _ , @xcite ; _ kamp & bertoldi _ , ; _ bergin et al . _ , @xcite ; _ kamp et al . _ , @xcite ; ) .
for instance , in t tauri stars the radiation field is dominated by strong ly@xmath8 emission , which has consequences for the photodissociation rate of molecules that can be dissociated by ly@xmath8 photons . the photoelectric heating process , on the other hand
, depends strongly on the overall shape of the radiation field , which is much steeper in the case of cool stars .
a similar problem appears in the x - ray spectra of cool m stars , which are dominated by line emission .
induced grain photoelectric heating is often a dominant heating process for the gas in the irradiated surface layers . its efficiency and thus the final gas temperature depends strongly on the grain charge , dust grain size and composition ( pahs , silicates , graphites , ices , etc . ) .
x - rays from the central star the uppermost surface layers , as this subsection focuses on of the optically thick disk at @xmath59 where of the dust surface area per hydrogen nucleus to the interstellar value , which is roughly @xmath60h .
@xmath61 contained in the surface layer ( @xmath62 ) is usually small , @xmath63 , , width=302 ] the detailed temperature structure of the surface layers of optically thick young disks was studied for the first time by , @xcite , and .
[ 100auslice ] shows the vertical structure in a disk model with 0.01 @xmath64 at 100 au around a 0.5 @xmath64 t tauri star .
fine structure line cooling of neutral oxygen .
molecules can shield themselves from the dissociating radiation .
as soon as the fraction of molecular hydrogen becomes larger than 1% , h@xmath57 line cooling .
molecular line emission cools the gas down to hundred k before the densities are high enough for gas and dust to thermally couple . to the cooling .
instead co , which has a rich rotational spectrum at low temperatures , becomes an important coolant . at larger
radii the from the central star drops as well as the . too low for the endothermic destruction of h@xmath57 by o atoms and hence the contains substantial fractions of molecular hydrogen
. detailed models of the gas temperature have shown that gas and dust are collisionally coupled at optical depth @xmath65 .
thus the basic assumption @xmath66 of the disk structure models presented in the previous section is justified .
the pure rotational lines of h@xmath57 such as j = 2 0 s(0 ) [ 28 @xmath32 m ] , j = 3 1 s(1 ) [ 17 @xmath32 m ] , j = 4 2 s(2 ) [ 12 @xmath32 m ] and j = 6 4 s(4 ) [ 8 @xmath32 m ] trace the warm gas ( 100 - 200 k ) in the disks . .
the detection of the mid - ir h@xmath57 lines at low spectral resolution ( e.g. , with spitzer ) is hindered by the low line - to - continuum ratio .
the mid - infrared line spectra of molecular hydrogen from a t tauri disk model ( @xmath67 @xmath64 ,
@xmath68 @xmath69 , @xmath70 k , @xmath71 @xmath64/yr ) with ( solid line ) and without ( dotted line ) uv excess .,width=302 ] , width=302 ] another tracer of the physics in the tenuous surface layers ( see also the chapter by ) .
it has been detected in a number of externally illuminated proplyds in the orion nebula @xcite as well as in t tauri and herbig ae / be stars .
@xcite explain by the photodissociation of the oh molecule , which leaves about 50% of the atomic oxygen formed in the upper @xmath72 level of the @xmath73 line .
need oh abundances higher than those predicted from disk models to fit the emission from the disks around herbig ae / be stars .
reveal the presence ( few 1000 k ) ; hence the [ oi ] line might arise partly from thermal excitation .
, the dust grains grow to centimeter sizes and the disks become optically thin .
in addition , as we shall discuss , the gas in the disk ultimately disappears , turning the disk into a debris disk .
it is therefore theoretically conceivable that there exists a transition period in which the disk has become optically thin in dust continuum , but still contains a detectable amount of gas . measuring the gas mass in such transition disks sets a timescale for the planet formation process .
the spitzer legacy science program ` formation and evolution of planetary systems ' ( feps ) has set upper limits on gas masses of @xmath74 @xmath75 around solar - type stars with ages greater than 10 myr ( _ meyer et al .
_ , @xcite ; _ hollenbach et al .
_ , @xcite ; _ pascucci et al .
_ , @xcite ) .
several groups have so far these transition phases of protoplanetary disks : @xcite modeled the disk structure and gas / dust emission from intermediate aged disks around low - mass stars , , , and modeled the gas chemistry and line emission from a - type stars such as @xmath76 pictoris and vega .
_ jonkheid et al . _
( @xcite ) studied the gas chemical structure and molecular emission in the disk around hd141569a .
these models are all based on the same physics as outlined above for the optically thick protoplanetary disks .
the disks are still in hydrostatic equilibrium , so that the disk structure in these low mass disks is similar to that in the more massive disks with the midplane simply removed .
however , some fundamental differences remain : the minimum grain size in these disks is typically a few microns , much larger than in the young protoplanetary disks ; in addition , the dust may have settled towards the midplane , and much of the solid mass may reside in larger particles ( @xmath77 cm ) than can be currently observed .
this reduces the grain opacity and the dust - to - gas mass ratio compared to the younger optically thick disks .
optically thin to stellar uv and @xmath78 kev x - ray photons . at columns greater than , the gas opacity becomes large enough to shield h@xmath57 and co , allowing significant molecular abundances . for disks extended to 100 au , very little mass ( very roughly @xmath79 @xmath75 ) is needed to provide this shielding .
detection can be significantly hampered by the low line - to - continuum ratio ( weak narrow line against the bright dust thermal background ) .
these lines generally originate from 110 au . , width=302 ]
have shown that beyond 40 au the dominant coolant for the latest tenuous stages of disk evolution is the [ cii ] 158 @xmath32 m line .
the fine structure lines of c , o and c@xmath80 trace only the surface of these tenuous disks : . since typical gas temperatures
are higher than in molecular clouds , co lines from the upper rotational levels ( j = 4 3 ) are predicted to be stronger than the lower j lines . @xcite
have recently detected the co j = 3 2 line in hd141569 and disk modeling by _ jonkheid et al .
_ ( @xcite ) shows that the profile excludes a significant contribution from gas inwards of @xmath81 au and estimate the total gas mass to @xmath82 .
the above section has shown that in the surface layers of the disk the gas temperature can become very high , greatly exceeding the dust temperature .
the warm surface gas can flow off the disk and escape the gravity of the star .
since the heating process responsible for these high temperatures is the _ radiation _ from the central star or a nearby o - star , this process is called `` photoevaporation '' . the viscous evolution ( i.e. accretion and spreading ) of the disk , discussed in section [ sec - viscevol ] ,
can be strongly affected by this photoevaporation process .
typically , it significantly shortens the ` lifetime ' of a disk compared to pure viscous evolution .
photoevaporation can also create inner holes or truncate the outer disk .
this has relevance to observations of such disks , such as the percentage of young stars with infrared excess versus their age ( _ haisch et al .
_ , @xcite ; _ carpenter et al .
_ , @xcite ) , or the inferred ` large inner holes ' of some disks ( e.g. , _ calvet et al .
_ , @xcite ; _ bouwman et al .
_ , @xcite ; _ forrest et al .
_ , @xcite ; _ dalessio et al .
_ , @xcite ) .
it has also far - reaching consequences for the formation of planets , as we will discuss below .
photoevaporation has already been discussed in earlier reviews ( _ hollenbach et al .
_ , @xcite ; _ hollenbach and adams _ , @xcite ; _ richling et al .
_ , @xcite ) . however , these reviews mainly focused on the heating by a nearby massive star ( such as the famous case of the proplyds in orion ) .
in contrast , in this section we will exclusively review recent results on photoevaporation by the central star , progress in this field since ppiv has been mostly theoretical , since observations of diagnostic gas spectral lines for the case of photoevaporation by the central , low mass star requires greater sensitivity , spectral resolution , and spatial resolution than currently available .
we will , however , discuss the implications for the observed ` inner holes ' and disk lifetimes .
photoevaporation results when stellar radiation heats the disk surface and resulting thermal pressure gradients drive an expanding hydrodynamical flow to space .
as shown in section [ sec - surface ] the main heating photons lie in the fuv , euv and x - ray energy regimes .
x - rays , however , were shown to be of lesser importance for photoevaporation ( _ alexander et al . _ , @xcite ) , and we will not consider them further .
there are two main sources of the strong euv and fuv excesses observed in young low mass stars : accretion luminosity and prolonged excess chromospheric activity .
recent work ( _ alexander et al . _ , @xcite ) has shown that euv photons do not penetrate accretion columns , so that accretion can not provide escaping euv photons to power photoevaporation . _
alexander et al .
_ ( @xcite ) present indirect observational evidence that an active chromosphere may persist in t tauri stars even without strong accretion , and that euv luminosities of @xmath83 photons / s may persist in low mass stars for extended ( @xmath84 yrs ) periods to illuminate their outer disks .
fuv photons may penetrate accretion columns and also are produced in active chromospheres .
they are measured in nearby , young , solar mass stars with little accretion and typically ( with great scatter ) have luminosity ratios @xmath85 or @xmath86 photons / s .
euv photons ionize the hydrogen in the very upper layers of the disk and heat it to a temperature of @xmath87 k , independent of radius .
fuv photons penetrate deeper into the disk and heat the gas to @xmath88 k , depending on the intensity of the fuv flux , the gas density and the chemistry ( as was discussed in section [ sec - surface ] ) . whether the euv or fuv heating is enough to drive an evaporative flow depends on how the resulting compares to the local escape speed from the gravitationally bound system .
a characteristic radius for thermal evaporation is the `` gravitational radius '' @xmath89 , where the sound speed equals the escape speed : @xmath90 early analytic models made the simple assumption that photoevaporation occurred for @xmath91 , and that the warm surface was gravitationally bound for @xmath92 .
however , a closer look at the gas dynamics shows that this division happens not at @xmath89 but at about @xmath93 @xmath94 ( _ liffman _
, @xcite ; _ adams et al . _ , @xcite ; _ font et al . _ , @xcite ) , and that this division is not entirely sharp .
in other words , photoevaporation happens _
mostly _ outside of the `` critical radius '' @xmath95 , though a weak evaporation occurs inside of @xmath96 .
since these are important new insights since ppiv , we devote a subsubsection on them below . with @xmath97
k the critical radius for euv - induced photoevaporation is @xmath98 @xmath99 au . however , there is no fixed @xmath100 because the fuv - heated gas has temperatures that depend on fuv flux and gas density , i.e. , on @xmath10 and @xmath26 .
therefore , @xmath100 depends on @xmath10 and @xmath26 , and may range from 3 - 150 au for solar mass stars .
the evaporative mass flux @xmath101 depends not only on the temperature of the photon - heated gas , but also on the vertical penetration depth of the fuv / euv photons . for euv photons this is roughly set by the strmgren condition that recombinations in the ionized layer equal the incident ionizing flux . neglecting dust attenuation
, this penetration column the penetration depth is an important quantity because it sets the density at the base of the photoevaporative flow : the deeper the penetration depth , the higher the density .
the flux of outflowing matter is the product of local density and sound speed within this heated layer .
this is why the complex surface structure models of section [ sec - surface ] are so important for fuv - driven photoevaporation .
for euv - driven photoevaporation , on the other hand , the situation is less complicated , since the temperature in the ionized skin of the disk is independent of @xmath10 and @xmath26 , as long as @xmath102 , where @xmath103 is the bottom of the ionized layer , i.e. the base of the flow .
for this simple case , the evaporative mass flux .
although fuv - heated layers have lower temperatures than the euv - heated skin they are at higher densities and may equally well initiate the flow and determine the mass flux as euv photons ( see _ johnstone et al .
_ , @xcite for a similar situation for externally illuminated disks ) .
one way to understand why the disk can evaporate at radii as small as @xmath104 is to consider the evaporative flow as a bernoulli flow along streamlines ( _ liffman _
, @xcite ; _ adams et al . _ , @xcite ) .
these streamlines initially rise nearly vertically out of the disk and then bend over to become asymptotically radially outward streamlines .
the gas at its base lies deep in the gravitational potential . as a simplification
let us now treat these streamlines as if they are entirely radial streamlines ( ignoring their vertical rise out of a disk ) .
then the standard atmospheric solution has a density that falls off from @xmath10 to roughly @xmath89 as exp@xmath105 .
the gas flows subsonically and accelerates , as it slowly expands outward , until it passes through a sonic point at @xmath106 ( @xmath107 is the classic parker wind solution for zero rotation ) . for @xmath108 ,
the mass flux is reduced considerably by the rapid fall - off of the density from @xmath10 to @xmath109 . for @xmath110 ,
the mass flux is roughly given by the density at @xmath109 times the sound speed times the dilution factor @xmath111 that accounts for mass conservation between @xmath10 and @xmath109 : @xmath112 .
assuming the same @xmath113 and @xmath13 at all @xmath10 , we see that @xmath114 and that @xmath115 .
this demonstrates that @xmath116 for this simplified case , and that even for @xmath117 evaporation is weak , but not zero . in fig .
[ 100auslice ] the base of the flow is marked with the large dot ( though that figure shows a static , non - evaporating model with only fuv heating ) . in that figure
, @xmath118 is the temperature such that the sound speed equals the escape speed ; @xmath119 is roughly where the photoevaporation flow originates ( i.e. , where @xmath120 ) .
although central star fuv models are not yet published , several central star euv models have appeared in the literature .
_ hollenbach et al .
_ ( @xcite ) first outlined the essential physics of euv - induced flows by the central star and presented an approximate analytic solution to the mass loss rate for a disk larger than @xmath89 .
the basic physics is the strmgren relation , @xmath121 , where @xmath122 is the hydrogen recombination coefficient and @xmath123 is the electron density in the ionized surface gas .
this sets the hydrogen nucleus ( proton ) number density at the base of the flow @xmath124 , and therefore an identical proportionality for the mass loss rate : @xmath125 in units of @xmath126 .
radiation hydrodynamical simulations ( _ yorke and welz _ , ; _ richling and yorke _ , ) find a similar power - law index for the dependence of the mass - loss rate on the euv photon rate of the central star .
this result applies for both high and low mass central stars , and is valid for a weak stellar wind .
the effect of a strong stellar wind is such that the ram pressure reduces the scale height of the atmosphere above the disk and the euv photons are allowed to penetrate more easily to larger radii .
this increases the mass - loss rate from the outer parts of the disk .
it is noteworthy that the diffuse euv field , caused by recombining electrons and protons in the disk s ionized atmosphere inside @xmath96 , controls the euv - induced mass - loss rates ( _ hollenbach et al .
_ , @xcite ) for disks .
this effect negates any potential for self - shadowing of the euv by the disk .
let us first assume a disk that does not viscously evolve : it just passively undergoes photoevaporation . for disks with size @xmath127
, the photoevaporation proceeds from outside in .
the mass flux rate at @xmath128 is much higher than inside of @xmath128 , because the gas at @xmath128 is least bound .
in addition , disk surface densities generally fall with @xmath10 ( see section [ sec - viscevol ] ) .
therefore , the disk shrinks as it photoevaporates , and most of the mass flux comes from the outer disk radius .
however , for disks with @xmath129 , two types of disk evolution may occur .
for euv photons , _ hollenbach et al .
_ ( @xcite ) showed that the mass flux @xmath130 beyond @xmath96 goes roughly as @xmath131 the timescale for complete evaporation at @xmath10 goes as @xmath132 .
as long as @xmath15 does not drop faster than @xmath131 , the disk will evaporate first at @xmath133 , and , once a gap forms there , will then steadily erode the disk from this gap outwards .
if , on the other hand , @xmath132 decreases with @xmath10 , then the disk shrinks from outside in as in the @xmath134 case .
the photoevaporation by the fuv from the central star has not yet been fully explored , but preliminary work by gh06 suggests that the mass flux @xmath130 in the outer disks around solar mass stars _
increases _ with @xmath10 . in this case
, the disk evaporates from outside in for most generally assumed surface density laws , which decrease with @xmath10 .
now let us consider a disk that is actively accreting onto the star ( see section [ sec - viscevol ] ) .
in general , if the photoevaporation drills a hole somewhere in the disk or ` eats ' its way from outside in , the forces of viscous spreading tend to move matter toward these photoevaporation regions , which can accelerate the dissipation of the disk .
if the disk has a steady accretion rate @xmath49 , then a gap forms once @xmath135 exceeds @xmath49 . since @xmath136 , the gap first forms at the minimum radius ( @xmath137 ) and then works its way outward .
_ clarke et al .
_ ( @xcite ) presented time - dependent computations of the evolution of disks around low mass stars with @xmath138 photons s@xmath139 .
their model combines euv photoevaporation with a viscous evolution code .
after @xmath140 to 10@xmath141 years of viscous evolution relatively unperturbed by photoevaporation , the viscous accretion inflow rates fall below the photoevaporation rates at @xmath89 . at this point ,
a gap opens up at @xmath89 and the inner disk rapidly ( on an inner disk viscous timescale of @xmath142 yr ) drains onto the central star or spreads to @xmath89 where it evaporates . in this fashion , an inner hole is rapidly produced extending to @xmath89 .
evolution of the surface density of a euv - photoevaporating disk ( figure adapted from _ alexander et al . , _
this simulation starts from a given disk structure of about 0.05 @xmath143 ( marked with ` start ' in the figure ) .
initially the disk accretes and viscously spreads ( solid lines ) . at @xmath144yr
the photoevaporation starts affecting the disk .
once the euv - photoevaporation has drilled a gap in the disk at @xmath78 au , the destruction of the disk goes very rapidly ( dashed lines).the inner disk quickly accretes onto the star , followed by a rapid erosion of the outer disk from inside out .
in this model the disk viscosity spreads to @xmath145 au ; however , fuv - photoevaporation ( not included ) will likely truncate the outer disk.,width=302 ] _ alexander et al .
_ ( @xcite , @xcite ) extended the work of _ clarke et al .
_ they show that once the inner hole is produced , the diffuse flux from the atmosphere of the inner disk is removed and the attenuation of the direct flux by this same atmosphere is also removed .
this enhances the euv photoevaporation rate by the direct euv flux from the star , and the effect magnifies as the inner hole grows as @xmath146 , again derivable from a simple strmgren criterion .
the conclusion is that the outer disk is very rapidly cleared once the inner hole forms . the rapid formation of a cleared out inner hole almost instantly changes the nature and appearance of the disk .
the above authors near- to mid - infrared color ( in magnitudes ) versus 850 @xmath32 m flux for photoevaporation / viscous evolution models .
the data are taken from _
hartmann et al . _ ( @xcite ) and _ andrews and williams _ ( @xcite ) : 850 @xmath32 m detections ( circles ) and upper limits ( triangles ) are plotted for both cttss ( filled symbols ) and wttss ( open symbols ) .
evolutionary tracks are shown for models with stellar masses 0.5 ( dashed ) , 1.0 ( solid ) , and 2.0 @xmath143(dotted ) , at a disk inclination of @xmath147 to the line of sight .
the thick tracks to the right and left show the 1 @xmath143 model at @xmath148 and @xmath149 , respectively .
crosses are added every 1 myr to show the temporal evolution .
initially the ( optically thin ) 850 @xmath32 m flux declines slowly at constant ( optically thick ) infrared color .
however , once the viscous accretion rate falls below the photoevaporation rate , the disk is rapidly cleared from the inside - out .
( figure adapted from _ alexander et al . _
@xcite . ) , width=264 ] _ matsuyama et al .
_ ( @xcite ) pointed out that if the euv luminosity is created by accretion onto the star , then , as the accretion rate diminishes , the euv luminosity drops and the timescale to create a gap greatly increases .
even worse , as discussed above , the euv photons are unlikely to escape the accretion column .
only if the euv luminosity remains high due to chromospheric activity does euv photoevaporation play an important role in the evolution of disks around isolated low mass stars . _
alexander et al .
_ ( @xcite ) argue this is the case .
_ ruden _ ( @xcite ) provides a detailed analytic analysis which describes the evolution of disks in the presence of viscous accretion and photoevaporation and compares his results favorably with these two groups . the processes which disperse the gas influence the formation of planets .
small dust particles follow the gas flow .
if the gas is dispersed before the dust can grow , all the dust will be lost in the gas dispersal and planetesimals and planets will not form . even if there is time for particles to coagulate and build sufficiently large rocky cores that can accrete gas ( _ pollack et al .
_ , @xcite ; _ hubickyj et al .
_ , @xcite ) , the formation of gas giant planets like jupiter and saturn will be suppressed if the gas is dispersed before the accretion can occur .
furthermore , gas dispersal helps trigger gravitational instabilities that may lead to planetesimal formation ( _ goldreich and ward _
, @xcite ; _ youdin and shu _ , @xcite ; _ throop et al . _ , @xcite ) , affects planet migration ( e.g. , _ ward _ , @xcite ) and influences the orbital parameters of planetesimals and planets ( _ kominami and ida _ , @xcite ) .
( @xcite ) showed that with @xmath150 photons s@xmath139 , the early sun could have photoevaporated the gas beyond saturn before the cores of neptune and uranus formed , leaving them gas poor . however , this model ignored photoevaporation inside of @xmath89 . the current work by _ adams et al .
_ ( @xcite ) would suggest rather rapid photoevaporation inside of 10 au , and make the timing of this scenario less plausible .
fuv photoevaporation ( either from external sources or from the central star ) may provide a better explanation .
preliminary results from gh06 suggest that the early sun did not produce enough fuv generally to rapidly remove the gas in the outer giant planet regions . _ adams et al . _ and _
hollenbach and adams _ , ( @xcite ) discuss the external illumination case , . a number of observations point to the truncation of kuiper belt objects ( kbos ) beyond about 50 au ( e.g. , _
allen , bernstein , and malhotra _
, @xcite ; _ trujillo and brown _ , @xcite ) .
_ adams _ _ et al . _ ( @xcite ) and _ hollenbach and adams _ ( @xcite , @xcite )
show that photoevaporation by a nearby massive star could cause truncation of kbos at about 100 au , but probably not 50 au .
the truncation is caused by the gas dispersal before the dust can coagulate to sizes which survive the gas dispersal , and which can then later form kbos .
models of fuv photoevaporation by the early sun are needed . in young disks
, dust settles toward the midplane under the influence of the stellar gravity and coagulates .
once coagulated dust has concentrated in the midplane , the roughly centimeter - sized particles can grow further by collisions or by local gravitational instability ( _ goldreich and ward _ , @xcite ; _ youdin and shu _ , @xcite ) .
a numerical model by _ throop and bally _ ( @xcite ) follows the evolution of gas and dust independently and considers the effects of vertical sedimentation and external photoevaporation .
the surface layer of the disk becomes dust - depleted which leads to dust - depleted evaporating flows . because of the combined effects of the dust settling and the gas evaporating , the dust - to - gas ratio in the disk midplane is so high that it meets the gravitational instability criteria of _ youdin and shu _ ( @xcite ) , indicating that kilometer - sized planetesimals could spontaneously form .
these results imply that photoevaporation may even trigger the formation of planetesimals .
presumably , photoevaporation by the central star may also produce this effect .
in this chapter we have given a brief outline of how disks form and viscously evolve , what their structure is , what their spectra look like in dust continuum and in gas lines , and how they might come to their end by photoevaporation and viscous dispersion .
the disk structure in dust and gas is summarized in fig .
[ fig - picto ] . evidently , due to the broadness of the topic we had to omit many important issues .
for instance the formation of disks is presumably much more chaotic than the simple picture we have discussed . in recent years there is a trend to outfit even the observation - oriented workhorse models with ever more detailed physics .
this is not a luxury , since the volume of observational data ( both spectral and spatial ) is increasing dramatically , as shown by various other chapters in this book .
for instance , with observational information about dust growth and sedimentation in disks , it will be necessary to include realistic dust evolution models into the disk models . additionally , with clear evidence for non - axial symmetry in many disks ( e.g. , _ fukagawa et al .
_ , @xcite ) modelers may be forced to abandon the assumption of axial symmetry .
the thermal - chemical study of the gas in the disk surface layers is a rather new field , and more major developements are expected in the next few years , both in theory and in the comparison to observations .
these new insights will also influence the models of fuv - photoevaporation , and thereby the expected disk lifetime . a crucial step to aim for in the future
is the unification of the various aspects discussed here .
they are all intimitely connected together and mutually influence each other .
such a unification opens up the perspective of connecting seemingly unrelated observations and thereby improving our understanding of the bigger picture .
hollenbach d. and adams f. ( 2005 ) in _ star formation in the interstellar medium _
( d. lin , d.johnstone , f. adams , d. neufeld , and e. ostriker , eds . ) , asp conference series , pp .
3 , provo : pub . astr . | = 11pt = 0.65 in = 0.65 in |
some thirteen years ago we made a first attempt to extract information about the prototype symbiotic star z and by applying techniques of time - series analysis to its optical light curve ( lc ) ( formiggini & leibowitz ( 1994 ) , hereafter paper i ) .
we have collected all historical and modern measurements of the brightness of the star known to us , mostly visual estimates by countless amateur astronomers all over the world . in this
we have been greatly helped by the rich and invaluable data archive of the american association of variable star observers ( aavso ) .
the lc analyzed in that work covered 98yr of the star history . since our first analysis
, data on additional 14yr of this star photometric behavior have been accumulated .
this fact , together with some experience and knowledge that we gained on the photometric variability of two other prototype symbiotics , bf cyg and yy her , ( leibowitz & formiggini 2006hereafter paper ii ; formiggini & leibowitz 2006hereafter paper iii ) , prompted us to make a second attempt to understand z and through its long - term lc . in this work we present the results of our reanalysis of the more extended lc , now covering 112yr . some of our conclusions reconfirm part of what we asserted in paper i. here we present also new conclusions about the system .
the basic parameters of the z and symbiotic star and the common knowledge and understanding of this system are summarized in many papers ( e.g. kenyon & webbink 1984 ; mikolajewska & kenyon 1996 ; sokoloski et al .
important developments in the study of this system in the thirteen years since the publication of paper i , that are related to our present analysis , may be summarized as follows : from polarimetric data , schmid & schild ( 1997 ) were able to determine the orbital plane inclination of the binary system _ i_=47@xmath0@xmath112@xmath0 .
the spectroscopic orbital period was found to be identical to the photometric one derived in paper i,(mikolajewska & kenyon 1996 ; fekel et al .
sokoloski & bildstein ( 1999 ) discovered coherent 28 minute oscillations , which were interpreted as reflecting the spin period of the magnetic white dwarf ( wd ) component of the system .
this episode was detected , however , only during a decline from one upward fluctuation in the brightness of the star ( within our h6 event see below ) that started in 1997 .
another attempt to observe rapid oscillations during the 2000 - 2002 fluctuation of that event yielded a null result ( sokoloski et al .
2006 ; gromadzki et al .
recently , sokoloski et al . ( 2006 ) made a thorough analysis of multi - wavelength observations of the 1996 - 97 and the 2000 - 02 fluctuations of event h6 .
they propose that the first one is probably a cataclysmic instability episode in an accretion disk around the wd of the system .
they suggest that the second , much more intense outburst , and presumably also many other similar fluctuations of z and , constitute a new type of a `` combination nova '' outburst .
this is an eruption and a large pour of energy from the system , the origin of which is disk instability , coinciding or preceding an enhanced nuclear burning on the surface of the wd .
for the construction of the visual lc of z and , we used the payne - gaposhkin ( 1946 ) data , covering the period from 1895 to 1944 and the visual magnitude estimates of aavso .
our procedure in bringing the two data sources to a satisfactory common scale and zero level is described in paper i section 2 .
the lc shown in fig .
1 here is the same one presented in fig . 1 of paper i , extended up to the month of september 2007 .
the points displayed in the figure after jd=2425000 are the magnitude estimate values in equal time bins of 10 d width . due to the scarcity of the earlier data ,
those points are presented as given in the original sources . in our analysis of the data
we have eliminated the first three points marked in the plot by open circles , on account of their large separation in time from the rest of the data .
our results , however , do not change significantly if we add these points to the lc .
the other twenty nine points marked by circles will be discussed in section 4.4 .
the curve underneath the observed data points displays the running mean of the lc above it .
the running window that we have used is 760 d wide .
the sine wave curve depicted in the figure will be explained in section 4.1 .
the two lower curves were shifted downward from their nominal y values to facilitate eye comparison .
basic characteristics of the optical lc of z and become qualitatively apparent by just looking at fig .
the temporal behavior of the star is marked by six outburst events , during most of which the mean brightness of the system rose by more than one magnitude .
this is clearly seen in the running mean curve in fig 1 .
this curve displays the mean over 760 d of the star magnitude , namely , the average magnitude of the star over one binary orbital cycle of the system ( see below ) .
superposed on these brightening episodes are large fluctuations that reach up to two magnitudes peak to peak amplitude .
the outburst episodes which we designate as h1 to h6 , are marked by the heavy lines along the time axis .
the individual events ( red dots in fig.1 ) are however quite different from each other , in their duration , in their structure , as well as in the amplitude of the fluctuations .
we refer to all of them as the high ( h ) states of the system . in between the h states
there are clear time intervals during which the brightness of the system is down and in a much quieter state .
these are the low ( l ) states of the system ( blue crosses in in fig.1 ) .
it can clearly be seen in fig .
1 that the brightness of the star at all l states is cyclically modulated with one and the same periodicity .
our first quantitative treatment of the lc was a computation of its power spectrum ( ps ) ( scargle , 1982 ) .
we also applied on it the aov technique of period search in time - series ( schwarzenberg - czerny 1989 ) .
both yield very similar results .
2(a ) presents the ps of the observed lc seen in fig . 1
, in the period range from 200d to the entire length of the lc , nearly @xmath240000 d. there is a high rise of the spectrum at the extreme low frequency region of it .
it indicates some variability on time scale of the length of the lc .
this can either be a real effect or an artifact due to some systematic error in our procedure of normalization of the different data sources in bringing them to a common magnitude level .
asides from this feature , the ps is dominated by two peaks , both are statistically significant .
the major one is at the frequency f1 , corresponding to the period _
p_1=7550@xmath1722 d. the uncertainty quoted here and throughout this work is p@xmath3/t , p being the corresponding period and t is the time length of the lc .
the second high peak in the ps near f=0.0015 d@xmath4 will be discussed in section 4.3 . the lowest curve in fig .
1 is a sine wave of the _ p_1 periodicity , fitted to the data by least squares .
the purpose of presenting it is only to show qualitatively how faithful is _
p_1 in representing the quasi cycle of the outburst occurrences .
we note that our present _
p_1 value and the value quoted in paper i for the characteristic repetition time of the outbursts of z and are within the corresponding error estimates .
the difference between the two values is due to the new outburst of the system of the present era , that was added in the last fourteen years to the older lc analyzed in paper i. in any event , as we shall discuss in section 5.2 , we do not claim , nor do we believe that the _ p_1 number represents a period of a strictly cyclic process .
the outbursts of z and take place quasi - periodically and _ p_1 is presently the best estimate for the average time interval between centres of successive outburst events . the inserts in fig .
2 ( a)(c ) are blowups of the high frequency end of the corresponding ps .
they , and the arrows in all frames of fig .
2 will be discussed in section 6.2.1 .
2(b ) is the ps of the combined lc of all low states of the system .
the unquestionable highly statistical significant peak is at the frequency f2 , corresponding to the period _ [email protected] d. fig .
3(a ) displays twice the l state lc folded onto the _ p_2 periodicity .
it is quite evident that the system has preserved the phase of the _ p_2 cycle , as well as the amplitude of this variation , throughout the entire 112yr of the observations , i.e. for more than 44 successive cycles .
this is in spite of the six time intervals of different lengths , those of the h state of the system , that are dispersed within these 112yr , during which the p2 periodicity all but disappeared from the lc of the star .
note the absence of any spectral feature in fig 2(c ) at the frequency of the high peak seen in fig 2(b ) ( see section 4.3 ) fig .
2(c ) is the ps of the h state lc . before combining the magnitudes of all the outbursts of the star into this single h lc
, we subtracted from each individual h lc a polynomial of first or second degree .
this process detrended each individual h lc from variability on the time scale of the h time interval itself .
when we subtract from each h lc only the corresponding mean magnitude we obtain similar results .
the highest peak in fig .
2(c ) , which is statistically highly significant , is at the frequency f3 , corresponding to the period _
[email protected] d. the cluster of lower peaks around it are either aliases of the high one or resulting from the highly non harmonic structure of the _ p_3 periodicity . for example , the second highest peak to the right of the f3 one , corresponds to the period of 606 d. this is the beat period of _ p_3 and _ p_1 , due to the gaps in the h lc , ( the time intervals of the l state ) , that break up the h lc with the 7550 d periodicity .
3(b ) is the lc of the h state , folded onto the _ p_3 periodicity .
the twenty nine points marked by red circles are the ones marked by circles in fig .
1 . we shall refer to them shortly .
with the exception of these points , all other 1533 points seen in the figure are evidence of the persistence of the _ p_3 period in the h state of the system .
the figure demonstrates that the period _
p_3 has the same phase in all its appearances , including the h6 event and the h1 event .
these two events are separated from each other by 55 cycles of this period and the @xmath2100 years that separate them contain the five time intervals of the l state , during which the _
p_3 period is absent from the lc .
there is no peak at the f3 frequency in the ps of the l state .
note the saw tooth structure of the mean profile of the _ p_3 photometric cycle .
4 displays in details the quality of the fit of the p3 periodicity to the individual fluctuations in each of the six outburst events .
each frame presents the observed detrended lc of one event .
the harmonic function is a sine wave fitted to the combined h lc of the star .
thus the amplitude and the phase of the wavefronts are the same in all six frames .
note that the scale of the x and y axis in fig .
4 ( e ) is different from all the rest .
4 ( e ) presents a time range that is wider than event h5 , as defined in fig .
1 . the outburst event itself is the segment delineated by the two vertical lines in the figure .
the dotted line is the run of harmonic wave of the _ p_2 periodicity that is fitted by least squares to the l light curve . in all other frames , i.e. all h events
except h5 , one can see in the figure that while the structure of the observed individual fluctuations varies considerably and is markedly different from the smooth sine wave structure , the phase seems to fit very well .
four points in h4 ( surrounded by green ellipse in fig .
4 ( d ) ) around jd 40000 are exceptions , showing a sharp rise in the brightness of the star that starts near a minimum of the _ p_3 periodicity .
we shall return to these four outliner points , as well as to fig .
4 ( e ) in the next section .
most of the 29 h state points that are out of phase with the 1533 measurements that exhibit the _ p_3=658 d oscillations , are those that constitute the two peaks of event h5 .
they are shown in details in fig .
4 ( e ) . the figure presents a wider segment of the lc . the h5 event as we identify it , is delineated by the two vertical lines in the figure .
the solid sinusoid line is the harmonic wave of the _ p_3 periodicity that is fitted by least squares to the entire h lc of the star , as in all other frames of fig .
the dotted line is the _ p_2=759 d harmonic wave fitted to the overall l lc .
one can see in fig .
4(e ) , that outside the two vertical lines the observed points follow very accurately the dotted line . in fact
most of the dotted line is obscured by the observed points that in the resolution of the figure fall right on top of it .
the lc in the section between the lines is composed of two well defined peaks and a sequence of lower points in between them , and a few that follow the second one .
these lower points too exhibit two oscillations , the maxima of which are systematically higher by 0.1 - 0.2 magnitude than the maxima of the 759 wave .
they also follow much better the solid line representing the 658 wave .
thus , the underlying rise in the brightness of the star of the h5 event , as compared with the brightness level in the l state , seems to follow the 658 periodicity , in phase with all other h events .
the two peaks of this event , as well as the four outlier points in event h4 , seem to be some sporadic flaring activity of the system that is probably related to the magnetic active phase of the star ( see section 5.3 ) but occur not in phase with the 658 d periodicity .
for the first fluctuation of the h5 event , the exceptional nature of it is apparent even in the photometric data .
4 ( e ) shows that this peak is different from all others .
its duration is only one half of the 658 d period which is the typical duration of nearly all other fluctuations observed in the h state of the system .
the two out of phase fluctuations of h5 will be discussed further in section 6.1 .
the optical brightness of z and in its low state is clearly modulated by the _ p_2=759 d periodicity .
this period coincides well with the spectroscopic periodicity found for this binary star ( mikolajewska & kenyon 1996 ; fekel et al . 2000 ) and it is therefore well established as the orbital period of the system ( paper i ) .
the profile of the optical photometric cycle is nearly sinusoidal , with minimum light obtained near spectroscopic conjunction , when the giant star is in front of the wd of the system ( sokoloski et al .
the cool component at this phase is obscuring part of the emission nebula , the excitation centre of which is the wd .
this binary configuration is also the phase of minimum light reflection from the hemisphere of the giant illuminated by the hot component ( formiggini & leibowitz 1990 ) .
six outbursts of z and have been recorded so far .
they are quite different from each other in duration , in the amount of excess optical radiation emitted in them , as well as in their structure .
the outbursts seem to take place , however , with a fairly constant time interval of some 7550 d , i.e. some twenty one earthly years , between the centres of successive ones . an outburst is generally manifested photometrically in the optical region by one or more magnitude increase in the average brightness of the system , and by intense periodic oscillations with a period _ p_3=658.4 d of strongly varying amplitude .
very similar features characterize the outbursts of two other symbiotic stars , bf cyg ( paper ii ) and yy her ( paper iii ) .
in these two systems the triggering of the outbursts has been ascribed to magnetic activity in the outer layers of the giant component of these two binary systems .
a solar - like magnetic dynamo mechanism has been suggested as the driver of the phenomenon and as the clock that maintains the quasi periodic cyclic nature of it .
the similarities in the phenomenology suggest also a similarity in mechanism .
we propose that in z and too , each of the six recorded outbursts resulted from intense , solar - like magnetic activity that takes place quasi - cyclically in the giant star of this system .
the quasi - periodic nature of the magnetic cycle in the giant of z and is not dissimilar to the timing of the magnetic activity of our own sun .
the solar cycle is also far from being strictly periodic , with intervals between successive maxima varying between nine and thirteen years ( babcock 1961 ; fligge , solanki & beer 1999 ; mursula & ulich 1998 ) .
similarly , vast variations in the format of different maxima of the activity cycle as observed in z and , characterize also the activity cycles of our sun ( solanki et al .
2002 ) . during outbursts
, the brightness of the star fluctuates violently with a distinct periodicity _
p_3=658.4 d. while the structure of these oscillations is far from being stable , the star seems to preserve the periodicity and the phase of their occurrences . with the two exceptions discussed in section 6.1 , the clock that controls the _
p_3 cycle appears to be stable throughout the @xmath2100 years period of the observations , i.e. for more than 55 cycles of the 658 d periodicity .
the persistence of the _ p_3 periodicity and phasing , implies that the clock controlling them must be a rather stable one .
in addition to the binary period , another clock in the system that is able to withstand the violent activity of the system with the tenfold or more variation in its output power , could be the spin of the giant star .
we suggest that the period of the observed outburst oscillations and its stable phase are the combined effect of these two coherent cycles of the system , according to the following qualitative model .
the system consists of a hot white dwarf ( wd ) star and an m giant , in binary orbit around each other with the period _
p_2=759 d. sokoloski et al .
( 2006 ) showed quite convincingly that the violent activity of the 2000 - 2002 z and fluctuation , and presumably also of the other violent fluctuations of the system , consist of a nuclear runaway process ( nra ) on the surface of the wd , combined with instability episodes in an accretion disk around the wd that seems to precede it or to coincide with it .
we suggest that the cause and the trigger of these two processes that constitute the `` combination nova '' events is the enhanced activity of the giant star , regulated semi - periodically by the giant inner magnetic dynamo .
one of the consequences of this activity is an intensification of the giant stellar wind , as is the case for the sun . in the case of our sun ,
detailed measurements of solar wind various parameters in the last few years , especially by _ ulysses _ , _ soho _ and other space probes , reveal this phenomenon as a multi - components , complex process .
however , a global structure , in mass density as well as in velocity , that is regulated to a very large extent by a dipole magnetic field that can be ascribed to the sun , is clearly apparent . during maximum activity seasons of the solar cycle , the wind pattern is changing , when both density and velocity taking different spatial configuration , also governed , however , by a global magnetic dipole field . for recent review of solar wind dynamics over the solar cycle
see schwenn ( 2006 ) .
notice in particular fig . 5 of that paper .
during maximum activity seasons of the solar cycle , for example , the wind velocity from the magnetic pole regions is strongly enhanced ( marsch 2006 ; ofman 2006 ) .
it is very much likely that for the giant of z and as well , during active seasons , stellar wind from the poles of a dipole magnetic field of the star is more enhanced , in velocity , in mass density or in both , than the wind from the magnetic equatorial regions .
the same qualitative model would apply if the magnetic field near the poles is inhibiting mass loss rather than intensifying it , and the wind from the magnetic equator is the more enhanced component .
we hypothesize that the axis of the dipole is inclined with respect to the spin axis of the star and rotates with it .
this is the case in the sun , although the inclination angle there is quite small ( babcock 1961 ; bilenko 2002 ) .
however , the phenomenon of magnetic oblique rotators with large inclination angles is not uncommon among stars ( kurtz 1990 ) .
we should note , however , that the existence of the hypothesized magnetic dipole field of the giant star would not be easily detected directly . in the solar case , the dipole field that is associated and presumably controls the global structure of the wind , reaches a maximum intensity of no more than 50 gauss .
the local fields in spots and in other features of surface and coronal activities are orders of magnitude stronger . for a distant observer ,
as we are with respect to the giant in z and , they mask any observable diagnostic of the global field , such as measurable polarization or the zeeman effect .
the polarization observed by schmid and schild ( 1997 ) is due to raman scattering and the geometry involved is on a much wider scale of the z and nebula that far exceeds the binary dimensions .
we further note , however , that the essential requirement here is not the dipole magnetic field .
what is needed for our interpretation is the existence on the surface of the giant of one , or two antipodal active regions , fixed in the outer atmosphere and rotating with the star .
f@xmath5rnandez - castro et al . (
1988 ) showed that even during quiescence states of z and , some nuclear burning is taking place on the surface of the wd .
this implies that even at quiescence , the giant is loosing mass at a large rate .
this would indicate in turn , that the giant star is not far from filling its gravitational roche lobe of the binary system ( sokoloski et al .
the gravitational pull of the wd creates a tidal wave in the outer layers of the giant .
the high tides are fixed in the rotating frame of the system . if the giant rotation is not locked to the orbital cycle , the tidal wave is traveling in the atmosphere of the spinning giant .
for an observer on the surface of the giant , the period of one cycle , namely , the synodic period of the wave , is the beat of the binary orbital period of the system and the sidereal spin period of the star ( see next section ) . during phases of maximum activity of the giant , whenever the bulge of the tidal wave is sweeping across the neighborhood of either the north or the south magnetic poles of the star ( or
, one might say , when the magnetic pole is sweeping through the tidal bulge ) a specially intense mass loss rate is ensued .
this results in enhanced cataclysmic dynamical processes in the disk around the wd .
it also increases the rate of supply of fuel for the nuclear runaway burning on the surface of the wd , and thus greatly intensifying it . due to the strong dependence of the rate of nuclear burning on the surface of a wd on temperature and density , even small enhancements in the accretion rate onto the disk and eventually to the wd surface
are enough to excite vast variations in the energy output rate of the system ( sokoloski et al .
this enhancement may occur once every synodic period of the traveling tidal wave , if the magnetic activity near one pole is significantly more enhanced than near the other .
it may occur twice every synodic cycle if the two antipodal regions on the surface of the giant are more similar in their surface activity .
this qualitative picture explains well the phase preservation in the _
p_3 periodicity in all outburst events of the system ( with the exception of the flares mentioned above ) .
it is the coherence of the binary cycle on one hand , and that of the spin period of the giant on the other , that ensure it .
we observe that during outbursts , the period of the activity cycle is _
p_3=658 d. this is , as explained , the synodic period of the tidal wave in the giant atmosphere .
if there are two antipodal centres of activity on the surface of the rotating giant , the synodic period is in fact 1317 d. for a binary orbital period of _
p_2=759 d , this would be the case if the sidereal period of the giant spin is 482 or 1790 d. if the intense wind at maximum activity season is emanating mainly from just one pole , the giant sidereal spin period is 352 d. in this case it could also be 5000 d if the giant rotates in the retrograde direction .
according to our suggested scenario ( section 5.3.1 ) , during seasons of maximum magnetic activity of the giant , whenever a magnetic pole is sweeping through the tides bulge , a large flux of matter is flowing from the giant onto the roche lobe of the wd . thus , during the giant activity season , what is being supplied in large quantities to the disk around the wd is not only mass but also a great deal of angular momentum . the disk instability , which according to the combination nova scenario ( sokoloski et al .
2006 ) precedes the nra on the surface of the wd , may be a response of the disk to the sudden increase in the influx of angular momentum , no less than to the increase in the rate of mass supply .
this may be an additional reason for the prompt response of the disk to the changing rate of mass accretion .
in contrast , the two out of phase eruptions of h5 may be understood as instability events that erupted in the disk spontaneously . in this event too , the supply of matter to the disk was modulated by the p3 periodicity , in phase with the oscillations in all other h events , as evident in the low amplitude variation of the low lying points of the h5 event , as described in section 4.4
. however , the amount of material supplied to the disk in this event , and hence also of angular momentum , was obviously smaller than in all other h events . due to the smallness of these two parameters
, they disturbed the disk less sharply than in the common , more powerful maximum activity seasons of the giant . by saying that in h5 the disk instabilities occurred spontaneously we mean that the conditions for instability developed in the disk more slowly .
the timing of their onset was regulated by the internal physics of the disk itself , as is the case in the common disk instability events in dwarf novae .
this is why they are not in phase with all other p3 oscillations .
the small amplitude brightness variation during h5 that is in phase with the p3 periodicity possibly reflects the brightness response of the disk to the increase varying supply of material from the giant .
it may also be the result of modest variations in the nuclear burning rate on the surface of the wd , due to variations in the rate of materials falling onto the wd surface either directly or after spiraling through the accretion disk . in this section we report on two other findings in the lc of the star that are very much in line with our suggested scenario .
the two additional features that we identify in the ps are , however , clearly below an appreciable statistical significance .
we therefore do not consider them as proofs of our model .
we do believe , however , that it is worthwhile mentioning them here for what they are , since if their reality is better confirmed by future observations , they do provide considerable support for our interpretation .
the inserts in the three frames of fig .
2 are blow - ups in the y direction of the high frequency ends of the corresponding psa . the arrows in frames ( b ) and ( c ) designated `` 1 '' , are pointing at peaks corresponding to the periods [email protected] d in the l ps ( fig . 2 ( b ) ) , and [email protected] d in the h ps ( fig . 2 ( c ) ) .
well within the error in the peak position these number are precisely half the 482 value , one of the candidates for the sidereal spin period of the giant .
a detection of a signal with 241 d periodicity is expected if there is a concentration of bright or dark spots around the magnetic poles of the rotating giant .
2 ( a ) shows , however , that the 241 d peak is absent from the ps of the entire lc , namely , the curve of the combined sequence of the low and the high states .
this means that if the _ p_=241 d is a real feature of z and , there is a phase shift in this periodicity between the low and the high states of the system .
if the 241 d periodicity is indeed half the spin period of the giant this would indicate that the hot or dark spots on the surface of the giant that are responsible for this light variations , change their location on the surface of the star between the two states of the system .
for example they could be concentrated more around the poles in one state , moving towards the magnetic equator in the other .
we note also that in the ps of the high state ( fig .
2 ( c ) ) there is a weak signal also at _
[email protected] d ( peak `` 2 '' ) . due to the low statistical significance of these signals
we refrain from using them for making a choice between the possible values that we suggest for the sidereal spin period of the giant .
mikolajewska et al .
( 2002 ) report on preliminary results of ir observations that indicate the presence of an ellipsoidal effect in the lc of z and .
if the star is nearly filling its roche lobe , a significant distortion of its outer layers from spherical symmetry is indeed expected . such a distortion
would be reflected in the lc of the system if the contribution of the giant photospheric light to the total optical luminosity of the system in the observed spectral region is not entirely negligible .
we do find in the h lc a weak signal of a period which is half the orbital period , as expected for the ellipsoidal effect .
arrow `` 3 '' in fig .
2 ( c ) is pointing at a peak in the ps corresponding to the period [email protected] d. the peak is below any statistical significance and it is undetectable in fig . 2 ( b ) , the ps of the l lc .
we therefore do not consider it as a positive detection of the ellipsoidal effect as stated above , in spite of the near perfect coincidence of its frequency with twice the orbital frequency .
we nonetheless wanted to find out where an assumption that it is a true sign of the ellipsoidal shape of the giant could lead us . since the giant spin is not locked to the orbital revolution , the tidal bulge is noncollinear with the centres of the two stars . if the spin period is 482 or 352 d , the orbital angular velocity is slower than that of the spin and the bulge is leading in front this line ( lecar , wheeler & mckee , 1976 ) .
by least squares we computed the sine wave of half the orbital period that best - fit the h lc .
we compare this lc to the sine wave of the orbital period itself that is best fitted to the l state of the system .
we found that one minimum of the `` ellipsoidal '' lc is preceding the orbital minimum by 130 d. the other `` ellipsoidal '' minimum is obviously preceding the maximum of the orbital cycle by the same number of days . in units of angle
this means that the direction of the high tides protrusion is leading the binary centres line by @xmath260@xmath0 .
if the reality of the _ p_2/2 signal could be confirmed , this result would be an important source of quantitative information about dissipative processes in the outer layers of the giant , those that are responsible for this noncollinearity .
we have performed in the past time - series analysis similar to the one presented in this work , on the historical lcs of two other prototype symbiotics , bf cyg and yy her ( paper ii and iii ) .
the three symbiotic stars are very similar in the basic astronomical parameters that characterize the corresponding binary systems .
table 1 presents some of these parameters for the three stars .
thus z and seems to become a third symbiotic system in which evidence for the operation of a solar - like magnetic dynamo cycle has been discovered .
the stars harboring the process in the symbiotic systems are giants , whereas the sun is a main sequence star .
the intensity of the surface activity of the symbiotic giants is probably also different from that of the sun , being of a more grandeur scale in the giants .
note , however , that the dramatic effects of the activity cycles of these stars are mainly due to the fact that they are members in interacting binary systems .
the activity cycle on the giant is mainly a trigger for the vast outpour of energy , which is generated in the extreme conditions on and around the wd component of the symbiotic system .
the quasi - periodicity of the phenomenon in all four stars is however of the same order of 15 - 20 yr .
the rotation rate of the giant of z and is much more distant from synchronization with the orbital cycle than in the other two symbiotics .
thus the synodic period of the tidal wave in its outer atmosphere is much shorter than the dynamo magnetic cycle . in bf cyg and yy
her , the near synchronization makes the spin - orbit beat period to be of the same order as that of the magnetic dynamo cycle .
this is why in z and we do not find in the lc any trace for an effect of the tides on the stellar magnetic cycle itself .
this is also the case in the sun with its spin period of merely @xmath227 d. the tides in the sun , excited primarily by jupiter , are obviously also of a much smaller scale than in the symbiotics . in any event , in both z and and in the sun , over one activity cycle , the displacement vector due to the radial motion induced by tides averages to zero , since one activity cycle includes many tidal oscillations . in bf cyg and yy
her , on the other hand , the radial oscillations of the outer envelope of the giant that are due to tides have the same characteristic time as that of the magnetic cycle itself . the role of tides in the creation and maintenance of the dynamo may therefore be comparable to that of differential rotation , convection and meridional flow that are currently believed to be at the dynamical roots of the magnetic dynamo phenomenon ( paper iii and references therein ) .
this is why in these two stars the synodic period of the tide wave is manifested in the timing of the magnetic dynamo cycle of their corresponding giant components .
the giant of z and is thus joining the class of stars that harbor magnetic dynamo process , as have been suggested already in the past ( soker 2002 , soker & kastner 2003 ) . in z and , as well as in bf cyg and yy
her , the observational effects of this phenomenon are gaining very large magnification , due to the binary interaction with the nearby wd companion .
this strengthens the proposition that symbiotic stars may turn out to be important laboratories for studying the stellar magnetic dynamo cycle , an astronomical process that affects earthlings directly .
new analysis of an updated optical lc of z and covering the last 112yr of the star history reestablishes the well known 759.1 d photometric periodicity of the star , known to be the binary orbital period of the system .
the analysis confirms the semi - periodic nature of the outburst phenomenon of this symbiotic star .
the characteristic time interval between the centres of successive outbursts is 7550 d. it is suggested that the quasi - periodicity of the outburst events is regulated by a solar - type magnetic dynamo process in the giant of the system .
enhanced mass loss during maximum phases of this activity cycle is the trigger and provides the fuel for the intense energy outpour of energy during outbursts , emanating mainly from the vicinity of the wd component .
in addition to an increase by one magnitude or more in the mean brightness of the system , outbursts are also characterized by violent periodic oscillations . our analysis reveals a persistent periodicity of _
p_3=658 d of these oscillations that preserves its phase throughout the entire 112yr of the observations . the oscillations result from periodic modulation of the mass accretion rate onto the wd component of the system , from the bulge of a tidal wave in the atmosphere of the giant which is fixed in the binary rotation frame . the synodic period of the tidal wave that circulates in the atmosphere of the giant is the measured _ p_3 value , or twice this number .
this is the beat period of the giant spin and the orbital periods of the system .
the sidereal spin period of the giant is either 482 or 1790 d. it could also be 352 d or possibly 5000 d , if the giant is spinning in the retrograde direction .
@lrrrcrc@ + + & z and & ref .
& bf cyg & ref .
& yy her & ref .
+ + giant sp . type & m4 iii & 1 & m5 iii & 1 & m4 iii & 1 + luminosity l @xmath6 & 620 - 1600 & 2 & 5200 & 2 & 1100 & 2 + binary period & 759.0 & 5 & 757.3 & 3 & 593.2 & 4 + giant spin period & 482@xmath7 & 6 & 798.8 & 3 & 551.4 & 4 + tidal wave period & 1317 & 6 & 14580 & 3 & -7825 & 4 + solar - type period & 7550 & 6 & 5375 & 3 & 4650 & 4 + + + * ) this is a slightly preferred number among four possible values .
\1 ) mrset & schmid ( 1999 ) , 2 ) mrset et al .
( 1991 ) , 3 ) leibowitz & formiggini ( 2006 ) , + 4 ) formiggini & leibowitz ( 2006 ) , 5 ) fekel et al .
( 2000 ) , 6 ) this paper .
we acknowledge with thanks the variable star observations from the aavso international database contributed by observers worldwide and used in this research . this research is supported by isf - israel science foundation of the israeli academy of sciences .
babcock h.w . , 1961 ,
apj , 133 , 572 bilenko i.a . , 2002 , a&a , 396 , 657 fekel f.c . ,
hinkle k.h . ,
joyce r.r . ,
skrutskie m.f . , 2000 , aj , 120 , 3255 f@xmath5rnandez - castro t. , cassatella a. , gim@xmath5nes a. , viotti r. , 1988 , apj , 324 , 1016 fligge m. , solanki s.k . , beer j. , 1999,a&a , 346 , 313 formiggini l. , leibowitz e.m . , 1990 ,
a&a , 227 , 121 formiggini l. , leibowitz e.m . , 1994 , a&a , 292 , 534 ( paper i ) formiggini l. , leibowitz e.m . , 2006 , mnras , 372 , 1325 ( paper iii ) gromadzki m. , mikolajewski m. , tomov t. , bellas - velidis i. , dapergolas a. , galan c. , acta astron , 56,97 kenyon s.j . , webbink r.f . , 1984 , apj , 279 , 252 kurtz d.w . , 1990 , ara&a 28 , 607 lecar m. , wheeler j.c . , mckee c.f . , 1976 ,
apj 205 , 556 leibowitz e.m . , formiggini l. , 2006 , mnras , 366 , 675 ( paper ii ) marsch e. , 2006 , advances in space research , 38,921 mikolajewska j , kenyon s.j .
1996 , aj , 112 , 1659 mikolajewska j. , kolotilov e.a .
, shenavrin v.i . ,
yudin b.f . , 2002 , in g@xmath8nsicke k , beuermann k. ,
reinsch k. eds , asp conf .
vol 261 , the physics of cataclismic variables and related objects , astron .
, san francisco , p.645 mrset u. , nussbaumer h. , schmid h. m. , vogel m. , 1991 , a&as , 248 , 458 mrset u. , schmid , h. m. , 1999 , a&as , 137 , 473 mursula , k , ulich th . , 1998 , geophysical research letters , 25 , 1837 ofman , l. , 2006 , advances in space research 38,64 payne - gaposhkin c. , 1946 , apj , 104,362 scargle j.d . , 1982 , apj .
, 263 , 835 schmid h.m . , schild h. , 1997 , a&a , 327 , 219 schwarzenberg - czerny a. , 1998 , mnras , 241 , 153 schwenn r. 2006 , space science review , 124 , 51 soker n. , 2002 , mnras , 337 , 1038 soker n. , kastner j. h. , 2003 , apj , 592 , 498 2002 , mnras sokoloski j. l. , bildstein , l. , 1999 , apj,517 , 919 sokoloski j. l. et al . , 2006 , apj , 636 , 1002 solanki s.k . , krivova n.a .
, schssler m. , fligge m. , 2002 , a&a , 396,1029 | we have reanalyzed the long - term optical light curve ( lc ) of the symbiotic star z andromedae , covering 112yr of mostly visual observations . two strictly periodic and one quasi - periodic cycles can be identified in this lc . a p1=7550 d quasi periodicity characterizes the repetition time of the outburst episodes of this symbiotic star .
six such events have been recorded so far . during quiescence states of the system ,
i.e. in time intervals between outbursts , the lc is clearly modulated by a stable coherent period of p2=759.1 d. this is the well known orbital period of the z and binary system that have been measured also spectroscopically .
a third coherent period of p3=658.4 d is modulating the intense fluctuations in the optical brightness of the system during outbursts .
we attribute the trigger of the outbursts phenomenon and the clock that drives it , to a solar type magnetic dynamo cycle that operates in the convection and the outer layers of the giant star of the system .
we suggest that the intense surface activity of the giant star during maximum phases of its magnetic cycle is especially enhanced in one or two antipode regions , fixed in the atmosphere of the star and rotating with it .
such spots could be active regions around the north and the south poles of a general magnetic dipole field of the star .
the p3 periodicity is half the beat of the binary orbital period of the system and the spin period of the giant .
the latter is then either 482 or 1790 d. if only one pole is active on the surface of the giant , p3 is the beat period itself , and the spin period is 352 d. it could also be 5000 d if the giant is rotating in retrograde direction .
we briefly compare these findings in the lc of z and to similar modulations that were identified in the lc of two other prototype symbiotics , bf cyg and yy her .
[ firstpage ] binaries : symbiotic stars : individual : z and stars : magnetic fields stars : oscillations . |
in the last 35 years v@xmath1o@xmath2 has been the subject of wide experimental and theoretical investigations , after its identification as the prototype of mott - hubbard systems . @xcite during the seventies an intensive set of measurements@xcite was performed in order to clarify its crystal and magnetic structure as well as its transport properties .
two metal - insulator transitions were found : one from a paramagnetic metallic ( pm ) to an antiferromagnetic insulating ( afi ) phase , below @xmath3 k , associated with a corundum - to - monoclinic crystal distortion ; the other from the pm to a paramagnetic insulating ( pi ) phase at higher temperatures ( above @xmath4 k ) . the pi phase can also be obtained by means of a small ( @xmath5 ) cr - doping : in this case cr ions act as a negative pressure , thus enlarging the average cation - cation distance and inducing a mott - insulator transition . as reviewed , eg , in ref . [ ] , all transitions are due to the interplay between band formation and electron coulomb correlation . in 1978 castellani ,
natoli and ranninger@xcite gave a realistic description of the complex magnetic properties and phase diagram of v@xmath1o@xmath2 , focusing on the peculiar magnetic structure observed in the afi phase , which breaks the high - temperature trigonal symmetry and can not be explained in terms of a single - band hubbard model .
they realized that the introduction of the extra degree of freedom represented by the degeneracy of t@xmath6-vanadium orbitals was necessary to describe correctly the spin structure , whose driving mechanism was the ordered pattern of t@xmath6-orbitals throughout the whole crystal .
this model was found to be in need of correction only more than 20 years later , when two experiments of non - resonant magnetic x - ray scattering@xcite and linear dichroism@xcite independently demonstrated that the average spin value on each vanadium ion is @xmath7 , contrary to the previously supposed @xmath8.@xcite these latter measurements stimulated a renewed interest in the system , that led to a new picture for its ground - state , where the two nearest v - ions are linked together in a stable molecule with spin @xmath9.@xcite when considered together with the available sets of x - ray diffraction data,@xcite and polarized neutron experiments,@xcite this model predicts that the afi phase is not magnetoelectric , as further confirmed by two experiments.@xcite yet , this seems not to be in keeping with the detection of non - reciprocal effects in linear dichroism , in ref .
, whose existence requires a magnetoelectric ( me ) phase . indeed , on the basis of diffraction data,@xcite the magnetic space group ( msg ) for the afi phase of v@xmath1o@xmath2 turns out to be @xmath10 , which is not magnetoelectric , because it explicitly contains both time - reversal ( @xmath11 ) and inversion ( @xmath12 ) symmetries .
however , a necessary condition to reveal a non - reciprocal signal is that none of them be a symmetry operator.@xcite thus , as better detailed in ref . [ ] , the interpretation of ref .
[ ] is in contradiction with the usually accepted msg . in this work we present a new experiment of vanadium k
edge absorption and linear dichroism , with the aim of clarifying the physical picture and solve the controversy about the ground - state symmetry of v@xmath1o@xmath2 . in order to better compare with the results of ref .
[ ] , we used the same single crystal , which is 2.8@xmath0 cr - doped .
the theoretical framework at the basis of this experiment is the one introduced in ref .
[ ] , where the possibility of a structural origin for the observed dichroism was suggested and a numerical prediction of the effect was carried out . as shown in the next two sections , the present measurement confirms that prediction and calls for a final experiment to check the effect of a me annealing procedure@xcite when cooling the system down to its afi phase .
in the monoclinic low - temperature phase pure v@xmath1o@xmath2 , as well as the 2.8@xmath0 chromium - doped sample used in the experiment , crystallizes in the body - centered classical space group @xmath13,@xcite with two groups of four v ions in the unit cell , which are translationally equivalent ( together with their oxygen environment ) apart from the opposite direction of magnetic moments.@xcite x - ray absorption spectra at v k edge ( e@xmath14=5465 ev ) were collected at the id26 beamline of the european synchrotron radiation facility in fluorescence mode .
the x - ray beam energies were selected by a double crystal si(220 ) , fixed exit monochromator with an energy resolution of about 0.3 ev at v k edge .
the id26 beamline worked in the so - called gap - scan mode , that is tuning the maximum of the undulator emission to the actual monochromator rocking curve .
two mirrors provided an harmonic free , small ( diameter @xmath15 ) , intense x - ray spot size on the sample , linearly polarized in the horizontal plane . the fluorescence intensity emitted from the sample
was collected in the total - fluorescence mode by measuring the current from two photodiodes , mounted parallel to the polarization of the incoming beam in order to minimize the elastic contribution to the spectrum .
the 2.8@xmath0 chromium - doped v@xmath1o@xmath2 single crystal ( about @xmath16 mm@xmath17 ) was mounted in a closed cycle he refrigerator , properly oriented with respect to the horizontal linear polarization of the x - ray beam and cooled down to 80 k. absorption spectra were collected in quick - exafs mode ( 20 seconds per scan ) in the energy range 5300 - 6300 ev ( about 4000 points per scan ) .
our geometrical setup allowed to have very narrow bragg peaks . in order to remove them ,
several ( @xmath1860 ) spectra were collected at different angles by rocking the sample ( @xmath19 2 deg . ) around the vertical axis . with this procedure bragg peaks
were shifted and could be easily distinguished from the true spectral features .
averaging up all the scans then provided exceptional quality , low noise , xas spectra suitable for accurate quantitative studies .
all data were collected for three orthogonal directions of polarization , ie , along the corundum c - axis , and in the hexagonal plane , along the monoclinic b@xmath20-axis and orthogonal to it , as shown in fig .
[ fig1 ] . here
we analyze the data collected in the near - edge region .
at v k edge , the x - ray absorption cross - section for the @xmath21 vanadium ion in the unit cell is given by : the operator @xmath23 is the usual matter - radiation interaction operator expanded up to the quadrupolar term , with photon polarisation @xmath24 and wave vector @xmath25 .
@xmath26 ( @xmath27 ) is the ground ( excited ) state of the crystal and @xmath28 ( @xmath29 ) its energy .
the sum is extended over all the excited states of the system , @xmath30 is the energy of the incoming photon and @xmath31 is the fine - structure constant .
the total cross - section is obtained by summing up over all 8 ions in the unit cell : @xmath32 .
the corresponding sum - rule for both structural and non - reciprocal linear dichroism was already derived in ref .
[ ] in the framework of multiple scattering theory .
here we just remind that on the basis of the usually accepted @xmath10 msg,@xcite linear dichroism in the hexagonal plane is expected to be at a maximum when one of the two orthogonal polarizations is parallel to the monoclinic b@xmath20-axis , thus orthogonal to the glide plane , and is zero when both polarizations are at @xmath33 degrees with the latter .
figure [ fig1 ] shows the experimental data in the xanes region for the three polarization directions : when the x - ray polarization is along the c - axis ( c ) , dichroism with respect to the other two directions is apparent ; less manifest , but still clear is the dichroism within the hexagonal plane , with the x - ray polarizations respectively parallel to the monoclinic b@xmath20-axis ( b ) and perpendicular to it ( a ) .
such dichroism is explictly shown in fig .
[ fig3](a ) . in fig .
[ fig2 ] , instead , we have plotted a numerical simulation of the `` in - plane '' xanes , as calculated in the framework of the multiple scattering ( ms ) theory within the muffin - tin approximation , and superimposed the experimental curve ( b ) for a better comparison . as in ref .
[ ] , we chose a cluster containing 135 atoms , _
i.e. _ , 54 vanadium and 81 oxygen atoms in the correct ratio 2 to 3 , having a radius of 6.9 , enough to get convergence with the cluster size .
the spectrum calculated with the real part of the hedin - lundqvist ( hl ) potential is convoluted with a lorentzian function having an energy dependent damping @xmath34 , derived from the universal mean free path curve by the relation @xmath35 , where @xmath36 is the photoelectron kinetic energy ( in rydbergs ) and @xmath37 its wave vector .
the agreement between experimental measurements and numerical simulations is quite good , even though the energy distance between the peak at 5467 ev and the highest peak is underestimated in the theory ( 18 ev vs. 20.5 ev ) .
this is probably a consequence of the potential adopted to describe the electron - atom scattering in the ms equations , whose hl exchange - correlation part@xcite is known to be too attractive near the edge.@xcite this compression of the energy scale in the theoretical signal is highlighted in fig .
[ fig3 ] , where experimental and theoretical `` in - plane '' linear dichroisms are compared . in spite of it , all other main features of the experiment are well described by our calculation , as illustrated in fig .
[ fig3](c ) , where the theoretical energy scale is expanded by a factor of 1.14 in order to better guide the eye : in this case all main peaks and valleys are correctly reproduced at almost all energies .
the same happens for xanes in fig .
[ fig2 ] , apart from the experimental signal around 5470 ev , to be probably related to the upper hubbard band,@xcite which is missed by our single - particle approach .
even more important , no extra scale factors are needed to describe the relative intensity of linear dichroism and xanes spectra , as both experiment and theory show a ratio of about 3@xmath0 .
finally , the calculated intensity is found to be of purely dipolar ( e1-e1 ) origin , due to the structural monoclinic distortion , and not of magnetic origin . all this is to be compared with the absence of structural linear dichroism measured in ref .
[ ] , as derived from their fig .
2 , where the average signal in the two magnetic field configurations is shown .
the apparent contradiction can be rationalized only in two ways , as suggested in ref .
[ ] : either by assuming a transition to another monoclinic domain , subsequent to the me annealing procedure , or by an accidental geometrical setup such that the two orthogonal polarizations were at 45@xmath38 from the b@xmath20-monoclinic axis . in this latter case
there is no structural dichroism , as shown in ref .
[ ] . in the present experiment , when the system was heated up to its pi phase and then cooled back to the afi phase , without any me annealing procedure , we found a transition to the same monoclinic domain .
even though a proper explanation of this phenomenon is still lacking , it is interesting to note that all @xmath39 cr - doped v@xmath1o@xmath2 single crystals used in resonant x - ray scattering experiments@xcite had the tendency to form one big monodomain , instead of a statistical mixture of the three possible . under the assumption of polarization at 45@xmath38 from the b@xmath20-axis , the appearence of a non - reciprocal signal in the afi phase might find its origin in the me annealing procedure , that might have driven the system to a nearby excited me state that is known to exist from ref .
[ ] . however , we would like to stress again that the relative intensity of the linear dichroism at the main peak is about 3@xmath0 of xanes in all three cases , i.e. , the present experiment , our theoretical calculation , and ref .
[ ] , thus strongly pointing towards a common origin of the three , given also the similarity in shape . in this respect , the possibility that the signal of ref .
[ ] be of magnetic e1-e2 origin is in contradiction with the resonant x - ray scattering results of paolasini it et al.@xcite : there it is shown that the @xmath40 reflection , of e1-e1 nature , is 100 times more intense than the @xmath41 , of e1-e2 magnetic origin , which is moreover restricted to the pre - edge region ( @xmath42 ev ) and not extended for more than 40 ev , like the linear dichroism of fig .
[ fig3 ] . the factor 10 in the amplitude ratio of e1-e1 and e1-e2 signals for a single @xmath43-ion site
is also confirmed by our ab - initio calculations , as well as by the practical absence of states with @xmath44 symmetry in the energy range 5474 - 5505 ev . in conclusion , the e1-e1 origin of this linear dichroism experiment
is well supported by a robust theoretical framework and calls for a further , definitive , investigation of the ground state properties of v@xmath1o@xmath2 in its low - temperature phase subsequent to a magnetoelectric annealing .
mcwhan , t.m .
rice , and j.p .
remeika , phys .
* 23 * , 1384 ( 1969 ) d.b .
mc whan and j.p .
remeika , phys .
b * 2 * 3734 ( 1970 ) ; d.b .
mcwhan , j.p .
remeika , s.d .
bader , b.b .
triplett , and n.e .
philips , phys .
b * 7 * , 3079 ( 1973 ) ; h. kuwamoto , j.m honig , and j. appel , phys . rev .
b * 22 * , 2626 ( 1980 ) and references therein . p.d .
dernier and m. marezio , phys .
* 2 * , 3771 ( 1970 ) r.b .
moon , phys .
lett . * 25 * , 527 ( 1970 ) j. spaek , j. sol
. state chem .
* 88 * , 70 ( 1990 ) c. castellani , c.r .
natoli and j. ranninger , phys .
b * 18 * , 4945 ( 1978 ) ; ibid .
, 4967 ( 1978 ) ; ibid . , 5001 ( 1978 ) l. paolasini , c. vettier , f. de bergevin , f. yakhou , d. mannix , a. stunault , w. neubeck , m. altarelli , m. fabrizio , p.a .
metcalf and j.m .
honig , phys .
lett . * 82 * , 4719 ( 1999 ) j .- h . park , l.h .
tjeng , a. tanaka , j.w .
allen , c.t .
chen , p. metcalf , j.m .
honig , f.m.f .
de groot , g.a .
sawatzky , _ phys .
_ b * 61 * , 11506 ( 2000 ) f. mila , r. shiina , f .-
c . zhang , a. joshi , m. ma , v. anisimov , t.m .
rice , phys .
lett . * 85 * , 1714 ( 2000 ) s. di matteo , n.b .
perkins , c.r .
natoli , phys .
b * 65 * , 054413 ( 2002 ) a. tanaka , jour . phys . soc
* 71 * , 1091 ( 2002 ) d.n .
astrov , sov .
jetp , * 11 * , 708 ( 1960 ) s. di matteo , a.g.m .
jansen , phys .
b * 66 * , 100402(r ) ( 2002 ) j. goulon , a. rogalev , c. goulon - ginet , g. benayoun , l. paolasini , c. brouder , c. malgrange , p.a .
metcalf , phys .
lett . * 85 * , 4385 ( 2000 ) s.
di matteo , y. joly , c.r .
natoli , phys .
b * 67 * , 195105 ( 2003 ) wei bao , c. broholm , g. aeppli , s.a .
carter , p. dai , t.f .
rosenbaum , j.m .
honig , p. metcalf , s.f .
trevino , phys .
b * 58 * , 12727 ( 1998 ) the me annealing procedure consists in the application of external electric and magnetic fields when cooling the sample through a phase transition .
l. hedin , b.i .
lundqvist , j. phys .
c , * 4 * , 2064 ( 1971 ) c.r .
natoli , m. benfatto , s. della longa , k. hatada , j. synchrotr .
10 * , 26 ( 2003 ) and references therein l. paolasini , s. di matteo , c. vettier , f. de bergevin , a. sollier , w. neubeck , f. yakhou , p.a .
metcalf , and j.m .
honig , j. electr .
spectr . & rel . phen . *
120 * , 1 ( 2001 ) s.y .
ezhov , v.i .
anisimov , d.i .
khonskii and g.a .
sawatzky , phys .
. lett . * 83 * , 4136 ( 1999 ) a. bombardi , f. de bergevin , s. di matteo , l. paolasini , p.a .
metcalf , j.m .
honig , physica b * 345 * , 40 ( 2004 ) | we performed near - edge x - ray absorption spectroscopy ( xanes ) at v k edge in the antiferromagnetic insulating ( afi ) phase of a 2.8 @xmath0 cr - doped v@xmath1o@xmath2 single crystal .
linear dichroism of several percent is measured in the hexagonal plane and found to be in good agreement with ab - initio calculations based on multiple scattering theory .
this experiment definitively proves the structural origin of the signal and therefore solves a controversy raised by previous interpretations of the same dichroism as non - reciprocal .
it also calls for a further investigation of the role of the magnetoelectric annealing procedure in cooling to the afi phase . |
amphiphilic molecules in aqueous solution form fluid bilayers with vanishing surface tension .
this causes them to undergo strong shape fluctuations , governed by the canham - helfrich curvature energy @xcite @xmath3 where @xmath4 is the surface element , @xmath5 corresponds to the doubled mean curvature of the surface at each point , and @xmath6 is the bending rigidity .
thermal undulations renormalize @xmath6 as follows @xcite : @xmath7 , \ ] ] where @xmath8 is an ultraviolet wavevector cutoff set by the inverse width of the molecules in the membrane , and @xmath1 is an infrared cutoff determined by its finite size . in practice
, membranes occur in the form of spherical vesicles , and @xmath9 is determined by their surface area . at finite temperatures , the model is only defined for finite planar surfaces . for @xmath1 larger than the de gennes - taupin persistence length @xmath10 @xcite , the renormalized bending rigidity
@xmath11 vanishes . beyond the persistence length
, the normal vectors of the surface are uncorrelated , and the membrane is crumpled .
the renormalization group flow extracted from the perturbative result ( [ eq : alpha1 ] ) as well as nonperturbative calculations in the large-@xmath12 limit @xcite , where @xmath12 is the dimension of the embedding space , exclude the possibility of a phase transition in the canham - helfrich model , even for tensionless membranes .
recently , the model has been extended by a kinetic term to include quantum fluctuations @xcite .
a one - loop renormalization group analysis showed that quantum effects stiffen the membrane .
the ground state corresponds to a flat configuration , where the normal vectors of the surface are strongly correlated .
the flat phase exists up to a critical temperature @xmath13 above @xmath14 , thermal fluctuations overcome quantum effects , and the membrane is always crumpled . in this paper
, we analyze the behavior of the quantum membrane exactly for very large dimension @xmath12 of the embedding space at all temperatures .
since the model is exactly solvable in this limit , we can calculate all of its relevant properties , in particular its order parameter and phase diagram .
the surface describing the membrane is parametrized by a vector field @xmath15 in the @xmath12-dimensional embedding space , where @xmath16 is a two - dimensional parameter space . in this parametrization
, the hamiltonian ( [ helfmodel ] ) reads @xmath17 where @xmath18 is the metric induced by the embedding , and @xmath19 $ ] .
the symbol @xmath20 denotes the derivative with respect to the parameters @xmath21 , and @xmath22 is the laplace - beltrami operator . as in ref .
@xcite , we add to the hamiltonian ( [ helfmodel2 ] ) a kinetic term to account for quantum fluctuations : @xmath23 where @xmath24 is now time - dependent , @xmath25 is the bare mass density , and the dot indicates a time derivative .
the euclidean action describing the quantum membrane is thus @xmath26 , \label{euclaction}\ ] ] and the partition function @xmath27 can be represented as a functional integral over all possible surface configurations @xmath28 : @xmath29/\hbar).\ ] ]
for large @xmath12 it is useful to consider @xmath31 as an independent field @xcite , and impose relation ( [ gmunu ] ) with help of a lagrange multiplier @xmath32 .
we consider the case where the classical action will have an extremum around an almost flat configuration . in the @xmath12-dimensional generalization of the monge parametrization of an almost flat surface
, the metric tensor becomes @xmath33 the partition function for the membrane can then be written as @xmath34 with the euclidean action @xmath35 - \frac{c_0}{4 } \lambda_{aa}^2 \right\}.\ ] ] we have included a bare surface tension @xmath36 to absorb infinities arising in the process of renormalization .
the renormalized , physical surface tension @xmath37 will be set equal to zero at the end of our calculations .
we have further added a term proportional to @xmath38 , with the proportionality constant @xmath39 being the in - plane compressibility of the membrane .
this is also necessary to absorb infinities , and the renormalized compressibility @xmath40 will be set equal to zero at the end to describe an incompressible planar fluid .
note that the functional integration over the lagrange multipliers @xmath41 in ( [ partfunc3 ] ) has to be performed along the imaginary axis for convergence .
the functional integral over all possible surface configurations @xmath42 in ( [ partfunc3 ] ) is gaussian , and can be immediately carried out , yielding an effective action @xmath43 with @xmath44 \label{s0tilde},\ ] ] and @xmath45 .
\label{s1}\ ] ] for large @xmath12 , the partition function ( [ partfunc3 ] ) is dominated by the saddle point of the effective action ( [ effectaction ] ) with respect to the metric @xmath31 and the lagrange multiplier @xmath46 , and we are left with a mean - field theory in these fields . for very large membranes ,
translational invariance allows us to assume that this saddle point is symmetric and homogeneous @xcite , such that @xmath47 with constant @xmath48 and @xmath49 .
there , the functional trace in ( [ s1 ] ) becomes an integral @xmath50 @xmath51 over the @xmath52-dimensional phase space , after replacing @xmath53 and @xmath54 .
at zero temperature , the phase space integral in ( [ s1 ] ) yields @xmath55\right\},\ ] ] where the ultraviolet divergences of the integral have been regularized by a wavevector cutoff @xmath8 .
the first term in ( [ s1t0 ] ) is a constant and renormalizes the surface tension to @xmath56 the quadratically divergent term renormalizes the bending rigidity in the second term of ( [ s0tilde ] ) .
the logarithmically divergent term proportional to @xmath57 modifies the in - plane compressibility to @xmath58 where @xmath59 is a renormalization scale .
we now set @xmath37 and @xmath40 equal to zero , to describe a tensionless incompressible membrane .
the renormalized effective action is then @xmath60\right\},\ ] ] where we have defined the critical bending rigidity @xmath61 and the constants @xmath62 , @xmath63 . from the second - derivative matrix of @xmath64 with respect to @xmath65 and @xmath66 we find that the stability of the saddle point is guaranteed only for @xmath67 .
note that the integration over @xmath41 in ( [ partfunc3 ] ) along the imaginary axis requires a maximum of ( [ seffld ] ) with respect to @xmath68 for stability .
the extremization of ( [ seffld ] ) with respect to @xmath65 leads to two different solutions for the saddle point , namely @xmath69 & = & \frac{1}{a } \left ( \frac{1}{\alpha^{1/2 } } - \frac{1}{\alpha_{\rm c}^{1/2 } } \right).\end{aligned}\ ] ] these describe two different phases existing at zero temperature .
for @xmath70 , @xmath71 is the only possible solution for the saddle point .
this solution corresponds to the flat phase , as we shall verify below . for @xmath72 , on the other hand , there exists a solution of eq.([lambdasp2 ] ) for nonzero @xmath66 .
this solution corresponds to the crumpled phase .
the behavior of the effective action ( [ seffld ] ) is shown in fig .
[ fig : freezero ] . as @xmath73 approaches the critical point from below ,
i. e. the membrane softens , @xmath66 becomes nonzero , and the surface crumples . to determine the saddle point solution for @xmath65 we extremize ( [ seffld ] ) with respect to @xmath66 . in the flat phase , where @xmath74 , we obtain @xmath75 showing that the total area of the membrane increases as @xmath73 approaches @xmath76 from below , with a crumpling transition at @xmath76 . in the crumpled phase
, @xmath65 is given by @xmath77 with nonzero @xmath66 .
as @xmath73 approaches @xmath76 from above , @xmath66 tends to zero , and @xmath65 goes again to infinity .
the positivity of @xmath65 and the stability of the saddle point imply that there is an upper bound for the bending rigidity , given by @xmath78 below which an incompressible membrane becomes unstable .
the behavior of @xmath66 in the two phases is shown in fig.[fig : lambda ] .
the behavior of @xmath79 is shown in fig .
[ fig : rho ] .
at finite temperature , the phase space integral in ( [ s1 ] ) involves a sum over the matsubara frequencies
@xcite : @xmath80 for small @xmath49 , a series expansion leads to @xmath81 \nonumber \\ & & + \frac{\hbar d}{2}\sqrt{\pi } \sum_{m=3}^{\infty } \frac{(-1)^{m+1}\lambda_0^m}{m \ , 2^{2 m } \pi^m } \left(\frac{\hbar}{k_{\rm b } t}\right)^{m-2 } \left ( \frac{\nu_0}{\alpha_0}\right)^{\frac{m-1}{2 } } \frac{\gamma(\frac{m-1}{2})}{\gamma(\frac{m}{2 } ) } \zeta(m-1 ) \bigg\}\end{aligned}\ ] ] as in the zero - temperature discussion , we absorb the logarithmic divergence by renormalizing the in plane compressibility via eq.([compressibility ] ) , setting @xmath40 equal to zero for incompressible membranes .
the surface tension receives now a temperature dependent renormalization @xmath82 and @xmath36 is chosen to make @xmath83 for tensionless membranes at all temperatures .
extremization of the renormalized combined effective action ( [ s0tilde ] ) and ( [ s1 ft ] ) with respect to @xmath65 leads again to two possible solutions for the saddle point , namely @xmath84 or @xmath85 , with @xmath86 + \lambda_{\rm t } \left [ 1 - \gamma + \ln \left(\frac{l^2}{8 \pi } \frac{k_{\rm b } t}{\hbar } \sqrt{\frac{\alpha}{\nu } } \right)\right ] \nonumber \\ & & + 32 \pi^{3/2 } \sum_{m=3}^{\infty } \frac { ( -1)^{m+1 } \lambda_{\rm t}^{m-1}}{m \ , 2^{2 m } \pi^m } \left(\frac{\hbar}{k_{\rm b } t}\right)^{m-2 } \left ( \frac{\nu}{\alpha}\right)^{\frac{m-2}{2 } } \frac{\gamma(\frac{m-1}{2})}{\gamma(\frac{m}{2 } ) } \zeta(m-1 ) \nonumber \\ & & = \frac{1}{a } \left [ 1 - \alpha^{1/2 } \left ( 1 - \sqrt{\frac{\alpha_{\rm c}}{\alpha_{\rm t } } } \right)\right ] \left ( \frac{1}{\alpha^{1/2 } } - \frac{1}{\alpha_{\rm t}^{1/2}}\right)\end{aligned}\ ] ] where @xmath87\ ] ] is the critical bending rigidity at finite temperature .
alternatively , we find the critical temperature at a fixed bending rigidity @xmath11 : @xmath88 in qualitative agreement with the perturbative critical temperature in eq .
( [ eq : tc_pert ] ) . for @xmath89 , eq .
( [ lambdat ] ) has no solution for @xmath90 . in this case , the membrane is in the flat phase , the only available solution for the saddle point being @xmath71 . for @xmath91 , @xmath90 is nonzero , and the membrane is crumpled .
let us now examine the saddle point solutions for @xmath65 . in the crumpled phase where @xmath85 is nonzero , we may expand the effective action into a high - temperature series .
extremization with respect to @xmath92 leads to @xmath93 \left [ 1 - \alpha^{1/2 } \left ( 1 - \sqrt{\frac{\alpha_{\rm c}}{\alpha_{\rm t } } } \right)\right ] - a \sqrt{\alpha } \lambda_{\rm t } \nonumber \\ & - & \hbar d \alpha \frac{\sqrt{\pi}}{2 } \sum_{m=3}^{\infty } \frac{(-1)^{m+1}\lambda_{\rm t}^{m-1}}{2^{2 m } \pi^m } \left ( 1 - \frac{2}{m } \right ) \left(\frac{\hbar}{k_{\rm b } t}\right)^{m-2 } \left ( \frac{\nu}{\alpha}\right)^{\frac{m-1}{2 } } \frac{\gamma(\frac{m-1}{2})}{\gamma(\frac{m}{2 } ) } \zeta(m-1 ) .\end{aligned}\ ] ] the positivity of @xmath65 and the stability of the saddle point again define an upper bound for the inverse bending rigidity , given by @xmath94 with @xmath95 for temperatures lower than @xmath96 the effective action becomes unstable if the rigidity is lower than @xmath97 . above @xmath96 ,
the membrane is stable at all rigidities . in the flat phase ,
the situation is more delicate . for @xmath71 ,
@xmath65 can be calculated exactly , and we obtain @xmath98,\ ] ] with an infrared regulator @xmath1 equal to the inverse lateral size of the membrane . for low temperatures , ( [ rhoftl0 ] ) may be approximated by @xmath99 \left [ 1 - \alpha^{1/2 } \left ( 1 - \sqrt{\frac{\alpha_{\rm c}}{\alpha_{\rm t } } } \right)\right].\ ] ] at high temperatures , however , the positivity of @xmath65 is not guaranteed . for fixed , but high temperatures , and for fixed membrane lateral size @xmath1
, there is a characteristic value of the inverse bending rigidity defined by @xmath100 above which @xmath65 changes sign , and ( [ rhoftl0 ] ) is no longer applicable . interestingly , for all @xmath1 and at all temperatures @xmath101 , the critical bending rigidity @xmath102 is larger than @xmath103 , so that the crumpling transition still occurs .
the behavior of @xmath65 is depicted in fig .
[ fig : rho ] below .
note that eq .
( [ tcr ] ) reflects the existence of a persistence length . at fixed temperature , for @xmath104 ,
the membrane is flat at scales smaller than @xmath105 and crumpled at larger scales .
this agrees with the de gennes - taupin persistence length @xmath106 . as the projected area @xmath9 of the membrane approaches infinity , the root @xmath107 of the branch @xmath108 ( see fig . [
fig : rho ] ) goes to zero , and the branch @xmath109 becomes unphysical .
the phase diagram of the quantum membrane is plotted in fig.[fig : phase ] .
as the lateral size @xmath1 of the membrane goes to infinity , the inverse critical bending rigidity @xmath107 goes to zero , and the crumpling transition is washed out . in the limit of infinite area , the ratio @xmath110 .
the membrane is crumpled at large scales , and flat at scales smaller than the persistence length .
its behavior can thus be described by the classical canham - helfrich model alone .
let us finally characterize the two phases in terms of the correlation functions between the normal vectors to the surface of the membrane . for the solution @xmath71 , this correlation function coincides with the one we found perturbatively for the zero temperature case , namely @xmath111 this solution corresponds to the flat , low temperature phase , where the normal vectors are strongly correlated .
such behavior can incidentally also be obtained in the large-@xmath12 limit at high temperatures by adding curvature terms of higher orders to the canham - helfrich hamiltonian ( [ helfmodel ] ) to stabilize a negative bending rigidity @xmath6 @xcite . for nonzero @xmath112
the correlation function behaves as @xmath113 in this case , the normals to the membranes are uncorrelated beyond a length scale @xmath114 .
the exponential decay of the correlation function shows that this solution corresponds to the crumpled phase .
the length scale @xmath114 may also be identified as the persistence length @xmath106 @xcite .
we have analyzed the temperature behavior of a membrane subject to thermal and quantum fluctuations in the limit of large embedding space dimension .
we found that at zero temperature there is a crumpling transition at some critical stiffness @xmath115 . for membranes of finite lateral size , quantum fluctuations are still relevant and a crumpling transition occurs also at nonzero temperature in an auxiliary mean field approximation as the lateral size of the membrane goes to infinity , the transition disappears , and the membrane is always crumpled in spite of quantum fluctuations , this being a consequence of the infrared divergences in the flat phase . | the phase diagram of an incompressible fluid membrane subject to quantum and thermal fluctuations is calculated exactly in a large number of dimensions of configuration space . at zero temperature
, a crumpling transition is found at a critical bending rigidity @xmath0 . for membranes of fixed lateral size
, a crumpling transition occurs at nonzero temperatures in an auxiliary mean field approximation .
as the lateral size @xmath1 of the membrane becomes large , the flat regime shrinks with @xmath2 . |
casimir effect is an interesting quantum phenomenon that has attracted the attention of physicists for more than 60 years @xcite . in recent years
, the finite temperature correction to the casimir effect has attracted increasing interest both theoretically and experimentally @xcite .
it has been known that in the high temperature regime , the vacuum fluctuations of quantum fields can give rise to casimir effect that is very different from the effect at zero temperature . since the pioneering work by ambjrn and wolfram @xcite , where casimir energies for scalar field and electromagnetic field in a @xmath6-dimensional rectangular cavity were calculated , casimir effect in higher dimensional spacetime has gradually received more attention , especially in recent years , when some long standing problems in physics have found solutions by embedding our physical @xmath7-dimensional spacetime in higher dimensional spacetime .
calculations of casimir energy in spherically symmetric configurations have been of interest to physicists for more than 45 years now . in @xcite ,
boyer calculated the casimir force acting on a perfectly conducting three dimensional spherical shell and found that it is repulsive .
this result was later confirmed by several groups of researchers @xcite using different methods .
the casimir energy of a massive scalar field confined in the interior and exterior of a three dimensional spherical shell was computed in @xcite .
generalization to higher dimensions was carried out in @xcite . later in @xcite , a systematic approach for scalar field ,
spinor field and electromagnetic field in @xmath6-dimensional spherical cavity was given .
in @xcite , we reconsider the electromagnetic casimir effect of a @xmath6-dimensional spherical shell .
most of these works only considered the zero temperature casimir effect .
so far finite temperature casimir effect on spherical shells has only been considered for the case when @xmath8 @xcite . in this work ,
we consider the finite temperature casimir effect for higher dimensional spherical shells .
we treat scalar field with dirichlet and neumann boundary conditions , and electromagnetic field with perfectly conducting and infinitely permeable boundary conditions . for scalar field ,
the results for zero temperature casimir energy of a @xmath6-dimensional spherical shell can be found in @xcite , whereas the corresponding results for electromagnetic field were obtained in @xcite .
the current work can be considered as an extension of these to the finite temperature region .
of particular interest to us is the high temperature asymptotic behavior of the casimir free energy . in this paper
, we use units where @xmath9 .
the casimir free energy of a quantum field in a @xmath6-dimensional bounded region @xmath11 is defined as @xmath12where @xmath13 are the eigenfrequencies of the field .
the zero temperature casimir energy @xmath14 is generically divergent , but the thermal correction @xmath15is always finite . using zeta regularization ,
the regularized casimir free energy is defined as ( see e.g. @xcite ) : @xmath16 ^ 2\zeta_{t}(m;0)\right),\ ] ] where @xmath17 is a normalization constant , and @xmath18 ^ 2\right)^{-s}\end{aligned}\]]is the corresponding thermal zeta function .
let @xmath19be the associated heat kernel .
it is well known that the heat kernel has an asymptotic expansion of the following form as @xmath20 ( see e.g. @xcite ) : @xmath21the coefficients @xmath22 are related to the zeta function @xmath23 by @xmath24 it has been shown that in the high temperature region ( see e.g. @xcite ) , the regularized casimir energy behaves as @xmath25 here @xmath26 is the riemann zeta function .
on the other hand , it is well known that ( see e.g. @xcite ) : @xmath27 hence , the regularized zeta function is well defined if and only if @xmath28 . in this article , we are going to consider the casimir free energy of a spherical shell in @xmath0-dimensional minkowski spacetime with radius @xmath29 .
for regularization purpose , we need to enclose this spherical shell in a larger spherical shell of radius @xmath30 .
let @xmath31 be the annular region @xmath32 between the two spherical shells .
the regularized casimir free energy of the spherical shell with radius @xmath29 is defined as @xmath33where @xmath34 is the ball of radius @xmath35 .
namely , we take the sum of the energy inside the spherical shell and the energy outside the shell , and subtract away the energy when the shell is absent .
let @xmath36 and @xmath37 denote @xmath38 by @xmath39 .
as in @xcite , one finds that @xmath40 can be written as a sum of three terms : @xmath41 where the term @xmath42 only depends on @xmath43 , and @xmath44 is the interacting term which is regular for all @xmath45 and goes to zero in the limit @xmath46 .
hence , @xmath47 and @xmath48@xmath49 can be considered as contribution from the interior region , and @xmath42 can be considered as contribution from the exterior region . in the same way
, one can show that the thermal zeta function can be written as @xmath50and the regularized casimir free energy is given by @xmath51 ^ 2\zeta_{t}(a;0)\right).\ ] ] as discussed above , the regularized casimir free energy is free of ambiguities if and only if @xmath52 . in the high temperature limit , the casimir free energy has an asymptotic expansion @xmath53as was discussed in @xcite , we need to renormalize this casimir free energy by subtracting away all the terms of order @xmath54 and above , so that in the high temperature limit , the renormalized casimir free energy behaves like @xmath55 in other words , @xmath56 from this , we see that the heat kernel coefficients @xmath57 , @xmath58 , are important for the renormalization of the casimir free energy .
the vanishing of the coefficient @xmath59 will render the casimir free energy well - defined .
after renormalization , in the limit @xmath60 , @xmath61the leading term is of order @xmath62 with coefficient @xmath63 .
the next - to - leading order term is of order @xmath64 with coefficient @xmath65 finally , there are terms of order @xmath66 and @xmath67 if and only if @xmath52 , i.e. , when the casimir free energy is not well - defined . in this work ,
our main goal is to compute the heat kernel coefficients @xmath68 , @xmath2 , the derivative of the thermal zeta function @xmath69 at @xmath70 and the derivative of the zeta function @xmath71 at @xmath70 . the spectrum of a scalar field in a spherical symmetric cavity subject to dirichlet ( d ) or neumann ( n ) boundary conditions is well known ( see e.g. @xcite ) .
the spectrum of an electromagnetic field in a spherical symmetric cavity with perfectly conducting ( pc ) or infinitely permeable ( ip ) boundary conditions was studied in @xcite . from the spectrum
, one can construct the corresponding zeta functions . for scalar field with dirichlet or neumann boundary conditions ,
@xmath72where @xmath73 @xmath74 \left[-ck_{\nu}(az)-azk_{\nu}'(az)\right ] \bigr\}. \end{split}\]]here @xmath75 when @xmath76 , @xmath77 when @xmath78 .
the part that involves the modified bessel function @xmath79 comes from the interior region , and the part that involves the modified bessel function @xmath80 comes from the exterior region . notice that @xmath81 only happens for neumann boundary conditions , and it only affects the term with @xmath82 for which @xmath83 . in this case , @xmath84 \left[-ck_{\nu}(az)-azk_{\nu}'(az)\right]=i_{\nu+1}(az)k_{\nu+1}(az ) .
\end{aligned}\]]hence , @xmath85where we have used the fact that @xmath86 . for electromagnetic field with perfectly conducting or infinitely permeable boundary conditions , one can divide the spectrum into contributions from transverse electric ( te ) modes and from transverse magnetic ( tm ) modes , and the corresponding zeta function can also be written as a sum of the te contribution and the tm contribution , namely @xmath87 for perfectly conducting boundary conditions , @xmath88 for infinitely permeable boundary conditions , @xmath89here @xmath90 the corresponding thermal zeta functions can be written down in the same way .
for example , for scalar field with dirichlet boundary conditions , @xmath91 namely , one replace the @xmath92 with @xmath93 and sum over @xmath94 from @xmath95 to @xmath96 . from this , we see that the basic constituents of the zeta functions and the thermal zeta functions are @xmath97 and @xmath98 .
the analytic continuations of the zeta functions @xmath99 and @xmath100 have been discussed in a number of works .
see e.g. @xcite . here
we apply the same method to perform the analytic continuations of the zeta functions @xmath101 and @xmath102 and the thermal zeta functions .
first consider the functions @xmath103 and @xmath104 .
debye uniform asymptotic expansions of the modified bessel function states that @xcite : as @xmath105 , @xmath106where @xmath107and @xmath108 and @xmath109 are defined recursively by @xmath110it follows that @xmath111 let @xmath112 then @xmath113@xmath114 and @xmath115 can be computed recursively by @xmath116one can prove by induction that they are polynomials in @xmath117 of the form @xmath118 from the debye asymptotic expansions , we find that @xmath119 and @xmath120\left[-ck_{\nu}(az)-azk_{\nu}'(az)\right]\bigr\ } \\\sim & \frac{z}{1+z^2}-2\sum_{i=1}^n \frac{zt(z)^3m_{2i , c}'(t(z))}{\nu^{2i}}\\&+\frac{d}{dz}\left(\ln \bigl\ { \left[ci_{\nu}(\nu z)+\nu zi_{\nu}'(\nu z)\right]\left[-ck_{\nu}(az)-azk_{\nu}'(az)\right]\bigr\}-\ln\frac { \nu(1+z^2)^{1/2}}{2 } -2\sum_{i=1}^n \frac{m_{2i , c}(t(z))}{\nu^{2i}}\right ) , \end{split}\]]where @xmath121 is any positive integer . using this ,
we can write the zeta functions @xmath122 as @xmath123where @xmath124
^ 2-m^2\right)^{-s } \frac{z}{1+z^2}\\ = & \frac{1}{2}a^{2s}\left(\nu^2+a^2m^2\right)^{-s } ; \end{split}\ ] ] for @xmath125 , @xmath126 ^ 2-m^2\right)^{-s}\left\{\frac{zt(z)^3d_{2i}'(t(z))}{\nu^{2i}}\right\}\\ = & -2a^{2s}\sum_{k=0}^{2i}d_{2i , k}\frac{\gamma\left(s+i+k\right)}{\gamma(s)\gamma\left(i+k\right)}\nu^{2k}(\nu^2+a^2m^2)^{-s - i - k},\\ a_{(\text{r},c),i}^{\text{int},\nu}(s ) = & - \frac{\sin ( \pi s)}{\pi}\int_{\frac{a m}{\nu}}^{\infty}dz\left(\left[\frac{\nu z}{a}\right]^2-m^2\right)^{-s}\left\{\frac{zt(z)^3m_{2i , c}'(t(z))}{\nu^{2i}}\right\}\\ = & -2a^{2s}\sum_{k=0}^{2i}m_{2i , k}(c)\frac{\gamma\left(s+i+k\right)}{\gamma(s)\gamma\left(i+k\right)}\nu^{2k}(\nu^2+a^2m^2)^{-s - i - k};\end{split}\ ] ] and @xmath127 ^ 2-m^2\right)^{-s } \frac{d}{dz}\left(\ln \left(i_{\nu}(\nu z)k_{\nu}(\nu z)\right)-\ln\frac{1 } { 2 \nu ( 1+z^2)^{1/2 } } -2\sum_{i=1}^n \frac{d_{2i}(t(z))}{\nu^{2i}}\right),\end{split}\]]@xmath128 ^ 2-m^2\right)^{-s } \\ & \times\frac{d}{dz}\left(\ln \left\ { \left[ci_{\nu}(\nu z)+\nu zi_{\nu}'(\nu z)\right]\left[-ck_{\nu}(\nu z)-\nu zk_{\nu}'(\nu z)\right]\right\}-\ln\frac { \nu(1+z^2)^{1/2}}{2 }
-2\sum_{i=1}^n \frac{m_{2i , c}(t(z))}{\nu^{2i}}\right ) .
\end{split}\ ] ] now notice that @xmath129 and @xmath130 can be expanded as @xmath131 let @xmath132be the hurwitz zeta function . substituting , and into , we find that the zeta functions , , and can be written as @xmath133here @xmath134 by taking @xmath135 , one can guarantee that @xmath136 does not contain any pole on the half - plane @xmath137 .
hence , this gives the analytic continuation of the zeta functions to @xmath137 .
next , define @xmath138 @xmath139 ^ 2+[2\pi p at]^2\right)^{s}}.\end{aligned}\]]we find in the same way that the thermal zeta functions can be written as @xmath140here @xmath141 by taking @xmath135 , one can guarantee that @xmath142 does not contain any pole on the half - plane @xmath143 .
hence , this gives the analytic continuation of the thermal zeta functions to @xmath143 .
using the fact that @xmath144 has only one pole at @xmath145 with residue @xmath146 , we read immediately from that , for dirichlet boundary conditions , @xmath147}y_{d;d - n+2i-1}\sum_{k=0}^{2i } d_{2i , k } \frac{\gamma\left(\tfrac{d - n}{2}+i+k\right ) } { \gamma\left(i+k\right)}\right),\\ \hat{c}_{d}=&-\frac{1}{2}\sum_{j=0}^{d-2}y_{d;j}\zeta_h\left(-j;\tfrac{d-2}{2}\right)-\sum_{i=1}^{\left[\frac{d-1}{2}\right]}y_{d;2i-1}\sum_{k=0}^{2i } d_{2i , k},\\ \hat{c}_{d+1}=&-\frac{1}{a}\sum_{i=1}^{\left[\tfrac{d}{2}\right]}y_{d;2i-2}\sum_{k=0}^{2i}d_{2i , k}\frac{\gamma\left(i+k-\tfrac{1}{2}\right)}{\gamma(i+k ) } ; \end{split}\ ] ] .[t1]heat kernel coefficients for dirichlet boundary condition [ cols="^,^,^,^,^,^,^",options="header " , ] similarly , @xmath148 ^
2+[2\pi p at]^2\right)^{s}}\\ = & \sum_{p=-\infty}^{\infty}\frac{1}{\gamma(s)}\int_0^{\infty } dt\ , t^{s-1}\exp\left\{-t\left(\left[\tfrac{d}{2}\right]^2+[2\pi p at]^2\right)\right\}\\ = & \sum_{p=-\infty}^{\infty}\frac{1}{\gamma(s)}\frac{\sqrt{\pi}}{2\pi at}\int_0^{\infty } dt\ , t^{s-1 - 1/2}\exp\left\{-t\left[\tfrac{d}{2}\right]^2-\frac{p^2}{4ta^2t^2}\right\}\\ = & \frac{\gamma\left(s-1/2\right)}{\gamma(s ) } \frac{1}{2\sqrt{\pi } at}\left(\frac{2}{d}\right)^{2s-1}+\frac{2}{\gamma(s ) } \frac{1 } { \sqrt{\pi } at}\sum_{p=1}^{\infty } \left(\frac{p}{dat}\right)^{s-1/2}k_{s-1/2}\left(\frac{dp}{2at}\right)\\ = & x_{\text{singular}}(s)+x_{\text{regular}}(s),\end{aligned}\ ] ] where @xmath149 has poles at @xmath150 and @xmath151 is a regular function . substituting these into
, we can compute @xmath152 . using and ,
one can then write down the expressions for the renormalized casimir free energy . we find that for dirichlet boundary conditions : @xmath153}\sum_{\substack{j=0\\j\neq 2i-2}}^{d-2}\sum_{k=0}^{2i}y_{d;j}d_{2i , k}\frac{\gamma\left(i+k-\tfrac{1}{2}\right ) } { \gamma(i+k)}\zeta_h\left(2i - j-1;\tfrac{d-2}{2}\right)\\&+\frac{1}{4\sqrt{\pi}a}\sum_{i=1}^{\left[\tfrac{d}{2}\right ] } \sum_{k=0}^{2i}y_{d;2i-2}d_{2i , k}\frac{\gamma\left(i+k-\frac{1}{2}\right)}{\gamma(i+k)}\left(\psi\left(i+k-\tfrac{1}{2}\right)-\psi(1)-2\psi\left(\tfrac{d-2}{2}\right)\right)\\ & + \frac{2}{\sqrt{\pi}a } \sum_{j=0}^{d-2}y_{d;j}\sum_{i=1}^{\left[\tfrac{d}{2}\right]}\sum_{k=0}^{2i}\frac{d_{2i , k}}{\gamma(i+k)}\sum_{n=0}^{\infty}\sum_{p=1}^{\infty } \left(n+\tfrac{d-2}{2}\right)^{k+j - i+1/2}\left(\frac { p}{2at}\right)^{i+k-1/2}k_{i+k-1/2}\left(\frac { p}{at}\left(n+\tfrac{d-2}{2}\right)\right)\\ & -\frac{t}{2}\sum_{l=0}^{\infty}\sum_{p=-\infty}^{\infty}b_d(l)b_{\text{d},\left[\frac{d}{2}\right]}^ { l+\frac{d-2}{2}\prime}(a;0;2\pi p t)-\frac{\hat{c}_{d+1}}{2\sqrt{\pi}}\ln(a\mu)+\frac{1}{\sqrt{\pi}}\sum_{n=0}^{d-1}2^{d - n } \gamma\left(\frac{d - n+1}{2}\right)\zeta_r(d - n+1 ) \hat{c}_{n } t^{d - n+1}.\end{aligned}\ ] ] = 0.43 = 0.43 = 0.43 = 0.43 for neumann boundary conditions : @xmath154}\sum_{\substack{j=0\\j\neq 2i-2}}^{d-2}\sum_{k=0}^{2i}y_{d;j}m_{2i , k}\left(\tfrac{2-d}{2}\right)\frac{\gamma\left(i+k-\tfrac{1}{2}\right ) } { \gamma(i+k)}\zeta_h\left(2i - j-1;\tfrac{d}{2}\right)\\ & + \frac{1}{4\sqrt{\pi}a}\sum_{i=1}^{\left[\tfrac{d}{2}\right ] } \sum_{k=0}^{2i}y_{d;2i-2}m_{2i , k}\left(\tfrac{2-d}{2}\right)\frac{\gamma\left(i+k-\frac{1}{2}\right)}{\gamma(i+k)}\left(\psi\left(i+k-\tfrac{1}{2}\right)-\psi(1)-2\psi\left(\tfrac{d}{2}\right)\right)\\ & + \frac{2}{\sqrt{\pi}a}\sum_{j=0}^{d-2}y_{d;j}\sum_{i=1}^{\left[\tfrac{d}{2}\right]}\sum_{k=0}^{2i}\frac{m_{2i , k}\left(\tfrac{2-d}{2}\right)}{\gamma(i+k)}\sum_{n=0}^{\infty}\sum_{p=1}^{\infty } \left(n+\tfrac{d}{2}\right)^{k+j - i+1/2}\left(\frac { p}{2at}\right)^{i+k-1/2}k_{i+k-1/2}\left(\frac { p}{at}\left(n+\tfrac{d}{2}\right)\right)\\ & -\frac{t}{2}\sum_{l=1}^{\infty}\sum_{p=-\infty}^{\infty}b_d(l)b_{\left(\text{r},\tfrac{2-d}{2}\right),\left[\frac{d}{2}\right]}^ { l+\frac{d-2}{2}\prime}(a;0;2\pi p t ) -\frac{d}{8 a } + \frac{1 } { 2\sqrt{\pi } a } \sum_{p=1}^{\infty } \sqrt{\frac{dat}{p}}k_{1/2}\left(\frac{dp}{2at}\right)-\frac{t}{2}\sum_{p=-\infty}^{\infty } b_{d,1}^{\tfrac{d}{2}\prime}(a;0;2\pi pt)\\&+\frac{1}{\sqrt{\pi}a}\sum_{k=0}^2d_{2,k}\left(\tfrac{d}{2}\right)^{2k}\left(\frac{\gamma\left(k+1/2\right)}{\gamma(k+1 ) } \frac{1}{2 } \left(\frac{2}{d}\right)^{2k+1}+\frac{2}{\gamma(k+1 ) } \sum_{p=1}^{\infty } \left(\frac{p}{dat}\right)^{k+1/2}k_{k+1/2}\left(\frac{dp}{2at}\right)\right)\\&-\frac{\hat{c}_{d+1}}{2\sqrt{\pi}}\ln(a\mu)+\frac{1}{\sqrt{\pi}}\sum_{n=0}^{d-1}2^{d - n } \gamma\left(\frac{d - n+1}{2}\right)\zeta_r(d - n+1 ) \hat{c}_{n } t^{d - n+1}.\end{aligned}\ ] ] = 0.43 = 0.43 = 0.43 = 0.43 for perfectly conducting boundary conditions : @xmath155}\sum_{\substack{j=0\\j\neq 2i-2}}^{d-2}\sum_{k=0}^{2i}\left(x_{d;j}d_{2i , k}+y_{d;j}m_{2i , k}\left(\tfrac{d-2}{2}\right)\right)\frac{\gamma\left(i+k-\tfrac{1}{2}\right ) } { \gamma(i+k)}\zeta_h\left(2i - j-1;\tfrac{d}{2}\right)\\&+\frac{1}{4\sqrt{\pi}a}\sum_{i=1}^{\left[\tfrac{d}{2}\right ] } \sum_{k=0}^{2i}\left(x_{d;2i-2}d_{2i , k}+y_{d;2i-2}m_{2i , k}\left(\tfrac{d-2}{2}\right)\right ) \frac{\gamma\left(i+k-\frac{1}{2}\right)}{\gamma(i+k)}\left(\psi\left(i+k-\tfrac{1}{2}\right)-\psi(1)-2\psi\left(\tfrac{d}{2}\right)\right)\\ & + \frac{2}{\sqrt{\pi}a } \sum_{j=0}^{d-2 } \sum_{i=1}^{\left[\tfrac{d}{2}\right]}\sum_{k=0}^{2i}\frac{\left(x_{d;j}d_{2i , k}+y_{d;j}m_{2i , k}\left(\tfrac{d-2}{2}\right)\right ) } { \gamma(i+k)}\sum_{n=0}^{\infty}\sum_{p=1}^{\infty } \left(n+\tfrac{d}{2}\right)^{k+j - i+1/2}\left(\frac { p}{2at}\right)^{i+k-1/2}k_{i+k-1/2}\left(\frac { p}{at}\left(n+\tfrac{d}{2}\right)\right)\\ & -\frac{t}{2}\sum_{l=1}^{\infty}\sum_{p=-\infty}^{\infty}\left ( h_d(l)b _ { \text{d},\left[\frac{d}{2}\right]}^ { l+\frac{d-2}{2}\prime}(a;0;2\pi p t)+b_d(l)b_{\left(\text{r},\tfrac{d-2}{2}\right),\left[\frac{d}{2}\right]}^ { l+\frac{d-2}{2}\prime}(a;0;2\pi p
t)\right)\\&-\frac{\hat{c}_{d+1}}{2\sqrt{\pi}}\ln(a\mu)+\frac{1}{\sqrt{\pi}}\sum_{n=0}^{d-1}2^{d - n } \gamma\left(\frac{d - n+1}{2}\right)\zeta_r(d - n+1 ) \hat{c}_{n } t^{d - n+1}.\end{aligned}\ ] ] = 0.43 = 0.43 = 0.43 = 0.43 finally for infinitely permeable boundary condition : @xmath156}\sum_{\substack{j=0\\j\neq 2i-2}}^{d-2}\sum_{k=0}^{2i}\left(x_{d;j}m_{2i , k}\left(\tfrac{4-d}{2}\right)+y_{d;j}d_{2i , k}\right)\frac{\gamma\left(i+k-\tfrac{1}{2}\right ) } { \gamma(i+k)}\zeta_h\left(2i - j-1;\tfrac{d}{2}\right)\\&+\frac{1}{4\sqrt{\pi}a}\sum_{i=1}^{\left[\tfrac{d}{2}\right ] } \sum_{k=0}^{2i}\left(x_{d;2i-2}m_{2i , k}\left(\tfrac{4-d}{2}\right)+y_{d;2i-2}d_{2i , k } \right)\frac{\gamma\left(i+k-\frac{1}{2}\right)}{\gamma(i+k)}\left(\psi\left(i+k-\tfrac{1}{2}\right)-\psi(1)-2\psi\left(\tfrac{d}{2}\right)\right)\\ & + \frac{2}{\sqrt{\pi}a } \sum_{j=0}^{d-2 } \sum_{i=1}^{\left[\tfrac{d}{2}\right]}\sum_{k=0}^{2i}\frac{\left(x_{d;j}m_{2i , k}\left(\tfrac{4-d}{2}\right)+y_{d;j}d_{2i , k } \right ) } { \gamma(i+k)}\sum_{n=0}^{\infty}\sum_{p=1}^{\infty } \left(n+\tfrac{d}{2}\right)^{k+j - i+1/2}\left(\frac { p}{2at}\right)^{i+k-1/2}k_{i+k-1/2}\left(\frac { p}{at}\left(n+\tfrac{d}{2}\right)\right)\\ & -\frac{t}{2}\sum_{l=1}^{\infty}\sum_{p=-\infty}^{\infty}\left ( h_d(l)b _ { \left(\text{r},\tfrac{4-d}{2}\right),\left[\frac{d}{2}\right]}^ { l+\frac{d-2}{2}\prime}(a;0;2\pi p t)+b_d(l)b_{\text{d},\left[\frac{d}{2}\right]}^ { l+\frac{d-2}{2}\prime}(a;0;2\pi p
t)\right)\\&-\frac{\hat{c}_{d+1}}{2\sqrt{\pi}}\ln(a\mu)+\frac{1}{\sqrt{\pi}}\sum_{n=0}^{d-1}2^{d - n } \gamma\left(\frac{d - n+1}{2}\right)\zeta_r(d - n+1 ) \hat{c}_{n } t^{d - n+1}.\end{aligned}\ ] ] = 0.43 = 0.43 = 0.43 = 0.43 from these expressions , we see that when @xmath157 , the renormalized casimir free energy depend on the normalization constant @xmath17 through the term @xmath158from , we find that when @xmath159 , @xmath160where @xmath161\right ) + \frac { \psi(1)+\ln ( 4\pi ) + \ln ( at)}{\sqrt{2\pi}}\hat{c}_{d+1}.\ ] ] using the results of sections [ hkc ] and [ zetaprime ] , we listed @xmath162 in tables [ t9 ] and [ t10 ] .
for perfectly conducting spherical shell in @xmath8 dimensions , our result agree with that obtained in @xcite . in figures
[ f1 ] , [ f2 ] , [ f3 ] , [ f4 ] , [ f5 ] , [ f6 ] , [ f7 ] and [ f8 ] , we plot @xmath163 when @xmath164 , for dirichlet , neumann , perfectly conducting and infinitely permeable boundary conditions , as a function of @xmath165 for @xmath165 between 0.001 and 10 , and compare each the these to the corresponding @xmath166 . from the graphs , it is observed that when @xmath167 , @xmath162 gives a good approximation to @xmath168 .
we have discussed the regularization and renormalization of the casimir free energy of a spherical shell in @xmath0-dimensional minkowski spacetime .
we consider scalar field with dirichlet and neumann boundary conditions , and electromagnetic field with perfectly conducting and infinitely permeable boundary conditions .
the casimir free energy is regularized using zeta functional method .
the high temperature asymptotic expansion of the regularized casimir free energy contains terms that are proportional to @xmath169 , which have to be subtracted away in the renormalization process . here
@xmath68 are the heat kernel coefficients .
we have derived formulas for these heat kernel coefficients .
when @xmath170 , they are computed explicitly . for @xmath58 , it is found that @xmath68 is zero if @xmath171 is even .
as is well known , the regularized and renormalized casimir energy is unambiguously defined ( not depending on a normalization constant ) if and only if @xmath52 , which is the case when @xmath6 is odd . in the high temperature limit , the leading term of the renormalized casimir free energy behaves as @xmath172where @xmath3 is the corresponding zeta function , and @xmath173 . we have derived the general expression for @xmath174 , and computed its explicit value when @xmath175 .
this might be of independent interest since it is closely related to a functional determinant .
finally , we derive explicit expressions for the renormalized casimir free energy .
when @xmath164 , it is computed numerically and the result is shown graphically .
this work is supported by the ministry of higher education of malaysia under frgs grant frgs/1/2013/st02/unim/02/2 .
i have benefited from discussions with a. flachi and k. kirsten when this work is done .
consider @xmath176}^n\frac{c^{2i}}{i}\zeta_h\left(2i - j;\tfrac{d}{2}\right)+y_d(c),\end{aligned}\ ] ] where @xmath177}\frac{c^{2i}}{i}\frac{1}{\nu^{2i}}\right).\end{aligned}\ ] ] it follows that @xmath178taking derivative with of @xmath179 with respect to @xmath180 , @xmath181 } c^{2i-1 } \frac{1}{\nu^{2i}}\right).\end{aligned}\ ] ] now , @xmath182 } c^{2i-1 } \frac{1}{\nu^{2i}}\right)\\ = & \nu^{j-1}\left(\frac{c}{\nu - c}+\frac{c}{\nu+c}-2\sum_{i=1}^{\left[\frac{j+1}{2}\right ] } c^{2i-1 } \frac{1}{\nu^{2i-1}}\right)\\ = & \nu^{j-1}c\left(\frac{1}{\nu - c}+\frac{1}{\nu+c}-2\sum_{i=1}^{\left[\frac{j+1}{2}\right ] } c^{2i-2 } \frac{1}{\nu^{2i-1}}\right)\\ = & \nu^{j-2}c\left(\frac{c}{\nu - c}-\frac{c}{\nu+c}-2\sum_{i=2}^{\left[\frac{j+1}{2}\right ] } c^{2i-2 } \frac{1}{\nu^{2i-2}}\right)\\ = & \nu^{j-2}c^2\left(\frac{1}{\nu - c}-\frac{1}{\nu+c}-2\sum_{i=1}^{\left[\frac{j-1}{2}\right ] } c^{2i-1 } \frac{1}{\nu^{2i}}\right)\\ = & \hspace{2cm}\vdots\\\end{aligned}\ ] ] if @xmath183 is even , @xmath184 can be reduced to @xmath185if @xmath183 is odd , @xmath184 can be reduced to @xmath186 hence , if @xmath6 is even , @xmath187whereas if @xmath6 is odd , @xmath188here @xmath189 .
hence , if @xmath6 is even , @xmath190- j\int_0^c du\,u^{j-1 } \left(\ln\gamma\left(\tfrac{d}{2}+u\right)+\ln\gamma\left(\tfrac{d}{2}-u\right ) \right)\right\};\end{aligned}\ ] ] whereas if @xmath6 is odd , @xmath191 therefore , @xmath192\right.\\&-\left .
j\int_0^cdu\,u^{j-1 } \left((-1)^j\ln\gamma\left(\tfrac{d}{2}+u\right)+\ln\gamma\left(\tfrac{d}{2}-u\right ) \right)\right\}.\end{aligned}\ ] ] 10 m. bordag , g. l. klimchitskaya , u. mohideen and v. m. mostepanenko , _ advances in the casimir effect _ , oxford university press , oxford , 2009 . g. l. klimchitskaya , u. mohideen and v. m. mostepanenko , rev . mod . phys .
* 81 * , 1827 ( 2009 ) .
k. a. milton , l. l. deraad , jr . and j. schwinger , ann
. phys . * 115 * ( 1978 ) , 388 .
p. candelas , ann . phys .
* 143 * ( 1982 ) , 241 . v.
v. nesterenko and i. g. pirozhenko , phys .
d * 57 * ( 1998 ) , 1284 .
m. e. bowers and c. r. hagen , phys .
d * 59 * ( 1998 ) , 025007 .
s. leseduarte and a. romeo , ann . phys .
* 250 * ( 1996 ) , 448 .
s. leseduarte and a. romeo , eur .
* 34 * ( 1996 ) , 79 .
m. bordag , e. elizalde , k. kirsten and s. leseduarte , phys .
d * 56 * ( 1997 ) , 4896 . c. m. bender and k. a. milton , phys .
d * 50 * ( 1994 ) , 6547 .
a. romeo , phys .
d * 52 * ( 1995 ) , 7308 .
a. romeo , phys .
d * 53 * ( 1996 ) , 3392 .
k. kirsten , _ spectral functions in mathematics and physics _ , chapman & hall/ crc , boca raton , fl , 2002 .
e. elizalde , s. d. odintsov , a. romeo , a. a. bytsenko , and s. zerbini , _ zeta regularization techniques with applications _ , world scientific publishing co. inc . ,
river edge , nj , 1994 .
m. bordag , v. v. nesterenko and i. g. pirozhenko , phys .
d * 65 * , 045011 ( 2002 ) .
l. p. teo , arxiv:1304:7557 ( 2013 ) .
m. bordag , e. elizalde and k. kirsten , j. math . phys . * 37 * , 895 ( 1996 )
. m. bordag , e. elizalde , k. kirsten , and s. leseduarte , phys . rev .
d * 56 * , 4896 ( 1997 ) .
e. elizalde , m. lygren , and d.v .
vassilevich , commun .
phys . * 183 * , 645 ( 1997 ) .
j. s. dowker and k. kirsten , commun .
* 7 * , 641 ( 1999 ) .
m. bordag , k. kirsten , and s. dowker , commun . math
. phys . * 182 * , 371 ( 1996 ) . | we consider the finite temperature casimir free energy acting on a spherical shell in @xmath0-dimensional minkowski spacetime due to the vacuum fluctuations of scalar and electromagnetic fields .
dirichlet , neumann , perfectly conducting and infinitely permeable boundary conditions are considered .
the casimir free energy is regularized using zeta functional regularization technique . to renormalize the casimir free energy , we compute the heat kernel coefficients @xmath1 , @xmath2 , from the zeta function @xmath3 .
after renormalization , the high temperature leading term of the casimir free energy is @xmath4 .
explicit expressions for the renormalized casimir free energy and @xmath5 are derived .
the dependence of the renormalized casimir free energy on temperature is shown graphically . |
on a complete probability space @xmath0 , we consider the following stochastic process @xmath1 in @xmath2 defined by @xmath3 where @xmath4 is a standard brownian motion , @xmath5 is a poisson process with intensity @xmath6 independent of @xmath7 , and we denote by @xmath8 the compensated poisson process @xmath9 .
the parameters @xmath10 are unknown and @xmath11 and @xmath12 are closed intervals of @xmath13 and @xmath14 , where @xmath15 .
let @xmath16 denote the natural filtration generated by @xmath7 and @xmath17 .
we denote by @xmath18 the probability law induced by the @xmath19-adapted cdlg stochastic process @xmath20 starting at @xmath21 , and by @xmath22 the expectation with respect to @xmath18 .
let @xmath23 and @xmath24 denote the convergence in @xmath18-probability and in @xmath18-law , respectively . for @xmath10 , we consider an equidistant discrete observation of the process @xmath20 which is denoted by @xmath25 , where @xmath26 for @xmath27 , and @xmath28 .
we assume that the high - frequency observation condition holds .
that is , @xmath29 let @xmath30 denote the density of the random vector @xmath31 under the parameter @xmath32 . for @xmath33 ,
set @xmath34 .
the aim of this paper is to prove the following lan property .
[ theorem ] assume condition .
then , the lan property holds for all @xmath10 with rate of convergence @xmath35 and asymptotic fisher information matrix @xmath36 .
that is , for all @xmath37 , as @xmath38 , @xmath39 where @xmath40 is a centered @xmath41-valued gaussian vector with covariance matrix @xmath42 theorem [ theorem ] extends in the linear case and in the presence of jumps the results of gobet in @xcite and @xcite for multidimensional continuous elliptic diffusions .
the main idea of these papers is to use the malliavin calculus in order to obtain an expression for the derivative of the log - likelihood function in terms of a conditional expectation .
some extensions of gobet s work with the presence of jumps are given for e.g. in @xcite , @xcite , and @xcite .
however , in the present note , we estimate the coefficients and jump intensity parameters at the same time . the main motivation for this paper is to show some of the important properties and arguments in order to prove the lamn property in the non - linear case
whose proof is non - trivial . in particular , we present four important lemmas of independent interest which will be key elements in dealing with the non - linear case . the key argument consists in conditioning on the number of jumps within the conditional expectation which expresses the transition density and outside it .
when these two conditionings relate to different jumps one may use a large deviation principle in the estimate .
when they are equal one uses the complementary set .
within all these arguments the gaussian type upper and lower bounds of the density conditioned on the jumps is again strongly used .
this idea seems to have many other uses in the set - up of stochastic differential equations driven by a brownian motion and a jump process .
we remark here that a plain it - taylor expansion would not solve the problem as higher moments of the poisson process do not become smaller as the expansion order increases .
in this section we introduce the preliminary results needed for the proof of theorem [ theorem ] . in order to deal with the likelihood ratio in theorem [ theorem ]
, we will use the following decomposition @xmath43 for each of the above terms we will use a mean value theorem and then analyze each term .
we start as in gobet @xcite applying the integration by parts formula of the malliavin calculus on each interval @xmath44 $ ] to obtain the following expressions for the derivatives of the log - likelihood function w.r.t . @xmath45 and @xmath46 .
moreover , using girsanov s theorem , we obtain the following expression for the log - likelihood function w.r.t .
@xmath47 . for any @xmath48
, we denote by @xmath49 the transition density of @xmath50 conditioned on @xmath51 .
[ prop1 ] for all @xmath52 , and @xmath53 , @xmath54,\\ & \dfrac{\partial_{\sigma}p^{\theta,\sigma,\lambda}}{p^{\theta,\sigma,\lambda}}\left(\delta_n , x_{t_k},x_{t_{k+1}}\right)=\dfrac{1}{\delta_n}{\mathrm{e}}_{x_{t_k}}^{\theta,\sigma,\lambda}\left[\left(b_{t_{k+1}}-b_{t_{k}}\right)^2\big\vert x_{t_{k+1}}^{\theta,\sigma,\lambda}=x_{t_{k+1}}\right ] -\frac{1}{\sigma } , \\ &
\dfrac{\partial_{\lambda}p^{\theta,\sigma,\lambda}}{p^{\theta,\sigma,\lambda}}(\delta_n , x_{t_k},x_{t_{k+1}})={\mathrm{e}}_{x_{t_k}}^{\theta,\sigma,\lambda } \left[-\dfrac{b_{t_{k+1}}-b_{t_{k}}}{\sigma}+\dfrac{\widetilde{n}_{t_{k+1}}^{\lambda}-\widetilde{n}_{t_k}^{\lambda}}{\lambda}\bigg\vert x_{t_{k+1}}^{\theta,\sigma,\lambda}=x_{t_{k+1}}\right ] .
\end{split}\ ] ] we next present the four lemmas mentionned in the introduction . consider the events @xmath55 , for all @xmath56 and @xmath53 .
[ lemma4 ] for all @xmath57 , and @xmath56 , @xmath58 for all @xmath59 and @xmath53 , we introduce the random variable @xmath60 . \end{split}\ ] ] we remark that heuristically the indicator functions @xmath61 and @xmath62 outside and inside the conditional expectation correspond to restrictions on @xmath63 and @xmath64 , respectively .
[ lemma5 ] for all @xmath65 and @xmath53 , @xmath66 we next fix @xmath67 , and analyze @xmath68 in two separate cases as follows @xmath69 furthermore , we write @xmath70 and @xmath71 , where @xmath72 and @xmath73 contain the terms @xmath74 , and @xmath75 and @xmath76 contain the terms @xmath77 in ( [ eq44 ] ) . [ lemma6 ] assume that @xmath78 and @xmath79 , for some constant @xmath80 .
then for all @xmath81 , and for @xmath82 large enough , @xmath83 for all @xmath84 and @xmath53 , set @xmath85\right],\\ m_{2,p}^{\bar{\theta},\bar{\sigma},\bar{\lambda}}:&=\sum_{j=0}^{\infty}{\mathrm{e}}_{x_{t_k}}^{\theta,\sigma,\lambda}\left[{\bf 1}_{j_j}{\mathrm{e}}_{x_{t_k}}^{\bar{\theta},\bar{\sigma},\bar{\lambda}}\left[{\bf 1}_{j_j^c}\left(n_{t_{k+1}}-n_{t_{k}}\right)^p\big\vert x_{t_{k+1}}^{\bar{\theta},\bar{\sigma},\bar{\lambda}}=x_{t_{k+1}}\right]\right ] .
\end{split}\ ] ] [ lemma7 ] assume that @xmath78 and @xmath79 , for some constant @xmath80 .
then , for any @xmath86 , @xmath84 , and for @xmath82 large enough , there exist constants @xmath87 such that for all @xmath67 , and @xmath88 , @xmath89 we next recall a convergence in probability result , and a central limit theorem for triangular arrays of random variables . for each @xmath90 , let @xmath91 and @xmath92 be two sequences of random variables defined on the filtered probability space @xmath93 , and assume that they are @xmath94-measurable .
[ zero ] ( * ? ? ?
* lemma 9 ) assume that @xmath95 \overset{{\mathrm{p}}}{\longrightarrow } 0 $ ] , and @xmath96\overset{{\mathrm{p}}}{\longrightarrow } 0 $ ] , as @xmath97 .
then @xmath98 as @xmath97 . [ clt ] ( * ? ?
* lemma 4.3 ) assume that there exist real numbers @xmath99 and @xmath100 such that as @xmath97 , @xmath101 \overset{{\mathrm{p}}}{\longrightarrow } m , \quad \sum_{k=0}^{n-1}\left({\mathrm{e}}\left[\zeta_{k , n}^2\vert \mathcal{f}_{t_k } \right]-\left({\mathrm{e}}\left[\zeta_{k , n}\vert \mathcal{f}_{t_k}\right]\right)^2\right)\overset{{\mathrm{p}}}{\longrightarrow } v , \text { and } \sum_{k=0}^{n-1}{\mathrm{e}}\left[\zeta_{k , n}^4\vert \mathcal{f}_{t_k}\right ] \overset{{\mathrm{p}}}{\longrightarrow } 0 . \end{split}\ ] ] then as @xmath97 , @xmath102 , where @xmath103 is a centered gaussian variable with variance @xmath104 .
for @xmath105 $ ] , set @xmath106 .
applying the markov property and proposition [ prop1 ] to each term in ( [ eq : dec ] ) , we obtain that @xmath107d\ell , \end{split}\ ] ] @xmath108 -\frac{1}{\sigma(\ell ) } \right)d\ell , \end{split}\ ] ] and @xmath109d\ell .
\end{split}\ ] ] now using equation , we obtain the following expansion of the log - likelihood ratio @xmath110 where @xmath111d\ell\right ) , \\ \eta_{k , n}&:=\dfrac{v}{\sqrt{n}}\int_0 ^ 1\dfrac{1}{\delta_n}\left(\dfrac{\sigma^2}{\sigma(\ell)^3}\left(b_{t_{k+1}}-b_{t_{k}}\right)^2-\dfrac{\delta_n}{\sigma(\ell)}\right)d\ell,\\ m_{k , n}&:=\dfrac{v}{\sqrt{n}}\int_0 ^ 1\dfrac{1}{\delta_n}\dfrac{1}{\sigma(\ell)^3}\bigg\{\left(\theta\delta_n+\widetilde{n}_{t_{k+1}}^{\lambda}-\widetilde{n}_{t_{k}}^{\lambda}\right)^2 + 2\sigma\left(b_{t_{k+1}}-b_{t_{k}}\right)\left(\theta\delta_n+\widetilde{n}_{t_{k+1}}^{\lambda}-\widetilde{n}_{t_{k}}^{\lambda}\right)\\ & \qquad\qquad-{\mathrm{e}}_{x_{t_k}}^{\theta_n,\sigma(\ell),\lambda_n } \bigg[\left(\theta_n\delta_n+\widetilde{n}_{t_{k+1}}^{\lambda_n}-\widetilde{n}_{t_{k}}^{\lambda_n}\right)^2\\ & \qquad\qquad+2\sigma(\ell)\left(b_{t_{k+1}}-b_{t_{k}}\right)\left(\theta_n\delta_n+\widetilde{n}_{t_{k+1}}^{\lambda_n}-\widetilde{n}_{t_{k}}^{\lambda_n}\right ) \bigg\vert x_{t_{k+1}}^{\theta_n,\sigma(\ell),\lambda_n}=x_{t_{k+1}}\bigg]\bigg\}d\ell , \\ \beta_{k , n}&:=-\dfrac{w}{\sqrt{n\delta_n}}\dfrac{1}{\sigma^2}\left(\sigma\left(b_{t_{k+1}}-b_{t_{k}}\right)+\dfrac{w\delta_n}{2\sqrt{n\delta_n}}-\dfrac{u\delta_n}{\sqrt{n\delta_n}}\right)\\ & \qquad+\dfrac{w}{\sqrt{n\delta_n}}\int_0 ^ 1{\mathrm{e}}_{x_{t_k}}^{\theta_n,\sigma,\lambda(\ell)}\left[\dfrac{\widetilde{n}_{t_{k+1}}^{\lambda(\ell)}-\widetilde{n}_{t_k}^{\lambda(\ell)}}{\lambda(\ell)}\bigg\vert x_{t_{k+1}}^{\theta_n,\sigma,\lambda(\ell)}=x_{t_{k+1}}\right]d\ell,\\ r_{k , n}&:=\dfrac{w}{\sqrt{n\delta_n}}\dfrac{1}{\sigma^2}\int_0 ^ 1\left(\widetilde{n}_{t_{k+1}}^{\lambda(\ell)}-\widetilde{n}_{t_k}^{\lambda(\ell ) } -{\mathrm{e}}_{x_{t_k}}^{\theta_n,\sigma,\lambda(\ell)}\left [ \widetilde{n}_{t_{k+1}}^{\lambda(\ell)}-\widetilde{n}_{t_k}^{\lambda(\ell ) } \bigg\vert x^{\theta_n,\sigma,\lambda(\ell)}_{t_{k+1}}=x_{t_{k+1}}\right]\right)d\ell . \end{split}\ ] ] we next show that the random variables @xmath112 are the terms that contribute to the limit in theorem [ theorem ] , and @xmath113 and @xmath114 are the negligible contributions . indeed , using girsanov s theorem and lemma [ lemma7 ]
, we can show that the conditions of lemma [ zero ] under @xmath18 hold for each term @xmath113 and @xmath114 .
that is , [ lemma1 ] assume condition .
then , as @xmath38 , @xmath115 \overset{{\mathrm{p}}_x^{\theta,\sigma,\lambda}}{\longrightarrow}-\dfrac{u^2}{2\sigma^2}-\dfrac{v^2}{2}\dfrac{2}{\sigma^2}-\dfrac{w^2}{2\sigma^2}\left(1+\dfrac{\sigma^2}{\lambda}\right)+\dfrac{uw}{\sigma^2}\\ & \sum_{k=0}^{n-1}\left({\mathrm{e}}^{\theta,\sigma,\lambda}\left[\xi_{k , n}^2+\eta_{k , n}^2+\beta_{k , n}^2\vert \mathcal{f}_{t_k}\right]-{\mathrm{e}}^{\theta,\sigma,\lambda}\left[\xi_{k , n}\vert\mathcal{f}_{t_k}\right]^2-{\mathrm{e}}^{\theta,\sigma,\lambda}\left[\eta_{k , n}\vert \mathcal{f}_{t_k}\right]^2-{\mathrm{e}}^{\theta,\sigma,\lambda}\left[\beta_{k , n}\vert \mathcal{f}_{t_k}\right]^2\right ) \\ & \qquad \qquad\qquad \qquad\overset{{\mathrm{p}}_x^{\theta,\sigma,\lambda}}{\longrightarrow}\dfrac{u^2}{\sigma^2}+2\dfrac{v^2}{\sigma^2}+\dfrac{w^2}{\sigma^2}\left(1+\dfrac{\sigma^2}{\lambda}\right)\\ & \sum_{k=0}^{n-1 } { \mathrm{e}}^{\theta,\sigma,\lambda}\left[\xi_{k , n}^4+\eta_{k , n}^4+\beta_{k , n}^4\vert \mathcal{f}_{t_k}\right]\overset{{\mathrm{p}}_x^{\theta,\sigma,\lambda}}{\longrightarrow}0 \\ & \sum_{k=0}^{n-1}\left({\mathrm{e}}^{\theta,\sigma,\lambda}\left[\xi_{k , n}\eta_{k , n}\vert \mathcal{f}_{t_k}\right]-{\mathrm{e}}^{\theta,\sigma,\lambda}\left[\xi_{k , n}\vert \mathcal{f}_{t_k}\right]{\mathrm{e}}^{\theta,\sigma,\lambda}\left[\eta_{k , n}\vert \mathcal{f}_{t_k}\right]\right)\overset{{\mathrm{p}}_x^{\theta,\sigma,\lambda}}{\longrightarrow}0\\ & \sum_{k=0}^{n-1}\left({\mathrm{e}}^{\theta,\sigma,\lambda}\left[\xi_{k , n}\beta_{k , n}\vert \mathcal{f}_{t_k}\right]-{\mathrm{e}}^{\theta,\sigma,\lambda}\left[\xi_{k , n}\vert \mathcal{f}_{t_k}\right]{\mathrm{e}}^{\theta,\sigma,\lambda}\left[\beta_{k , n}\vert \mathcal{f}_{t_k}\right]\right)\overset{{\mathrm{p}}_x^{\theta,\sigma,\lambda}}{\longrightarrow}-\dfrac{uw}{\sigma^2}\\ & \sum_{k=0}^{n-1}\left({\mathrm{e}}^{\theta,\sigma,\lambda}\left[\eta_{k , n}\beta_{k , n}\vert \mathcal{f}_{t_k}\right]-{\mathrm{e}}^{\theta,\sigma,\lambda}\left[\eta_{k , n}\vert \mathcal{f}_{t_k}\right]{\mathrm{e}}^{\theta,\sigma,\lambda}\left[\beta_{k , n}\vert \mathcal{f}_{t_k}\right]\right)\overset{{\mathrm{p}}_x^{\theta,\sigma,\lambda}}{\longrightarrow}0 .
\end{aligned}\ ] ] finally , lemma [ clt ] applied to @xmath116 concludes the proof of theorem [ theorem ] .
second author acknowledges support from the european union programme fp7-people-2012-cig under grant agreement 333938 .
third author acknowledges support from the lia cnrs formath vietnam and the program arcus mae / idf vietnam . | in this paper , we consider a linear model with jumps driven by a brownian motion and a compensated poisson process , whose drift and diffusion coefficients as well as its intensity are unknown parameters .
supposing that the process is observed discretely at high frequency we derive the local asymptotic normality ( lan ) property . in order to obtain this result , malliavin calculus and girsanov s theorem
are applied in order to write the log - likelihood ratio in terms of sums of conditional expectations , for which a central limit theorem for triangular arrays can be applied .
0.5 * rsum * 0.5*la proprit lan pour un modle linaire avec sauts .
* dans cet article , nous considrons un modle linaire avec sauts dirig par un mouvement brownien et un processus de poisson compens do nt les coefficients et lintensit dpendent de paramtres inconnus .
supposant que le processus est observ haute frquence , nous obtenons la proprit de normalit asymptotique locale .
pour cela , le calcul de malliavin et le thorme de girsanov sont appliqus afin dcrire le logarithme du rapport de vraisemblances comme une somme desprances conditionnelles , pour laquelle un thorme centrale limite pour des suites triangulaires peut tre appliqu .
, , |
since the mid eighties , the field of quarkonium physics has faced a number of puzzles challenging our understanding of qcd at the interplay between its short- and long - distance domains .
the puzzles related to the quarkonium production at the tevatron have been attributed to the colour - octet mechanism ( com ) , _ i.e. _ the non - perturbative transition of heavy quark - antiquark pairs in colour - octet state into quarkonia ( see @xcite for reviews ) . since a few years , we know that @xmath6 and @xmath7 corrections to the colour - singlet mechanism ( csm ) @xcite are essential to try to explain the @xmath8 dependence of the @xmath0 and @xmath9 cross sections observed in high - energy hadron collisions @xcite .
polarisation predictions are also dramatically affected by qcd corrections , both in the inclusive case and in the production of quarkonia with a prompt photon @xcite .
as far as the @xmath8-integrated yield is concerned , colour - singlet @xmath10 configurations have been shown to be sufficient - wave decay at nlo , earlier regulated by an ad - hoc binding energy @xcite , which can however be rigorously cured @xcite in the more general framework of nrqcd . ] to account for the experimental data @xcite . in this letter
, we reassess the importance of the leading-@xmath11 contribution to @xmath12 _ i.e. _ from the colour - singlet transitions . in the previous analyses of @xmath13
@xcite , these have been disregarded since formally appearing at higher orders in @xmath14 or @xmath15 .
in fact , they are not negligible at all .
quotes such as _
`` @xmath16 offers a clean test of the colour - octet contributions '' _ from @xcite and _ `` if the @xmath13 production is really detected , it would be a solid basis for testing the color - octet mechanism of the nrqcd '' _ from @xcite are overstated if not misleading .
this observable is not cleaner than the inclusive production for instance .
we have identified two classes of important colour - singlet contributions .
the first comes from the strange - quark gluon fusion which produces a @xmath17 pair where the charm quark fragments into a @xmath0 ( see fig .
[ diagram - a ] ) .
this is reminiscent of the leading-@xmath8 contribution to @xmath18 for instance @xcite . in the past , @xmath17 has indeed been identified as a probe of the strange quark pdf @xcite , as @xmath17 , could in principle be discriminated experimentally owing to the presence of an additional charmed hadron in the final state . ] .
the other class is simply a contribution _ la _ vector - meson dominance .
the @xmath19 quarkonium bound - state is simply produced by an off - shell photon ( or @xmath20 ) emitted by the quark which also radiates the @xmath1 boson ( see fig . [ diagram - b ] ) .
the latter contribution is clearly enhanced in @xmath21 collisions at the tevatron owing to the presence of valence antiquarks in the antiproton , whereas the former contribution is getting larger at lhc energies in @xmath22 collisions thanks to the enhancement of the gluon pdf at lower @xmath23 . in any case , these csm processes are not at all negligible compared to the leading colour - octet contributions ( see fig .
[ diagram - c ] ) .
interestingly , both these born contributions possess a leading-@xmath8 contribution ( @xmath24 ) . not only are they significant at low @xmath8 , but they remain large at large @xmath8 .
this is at variance with the inclusive case where the born contributions are not leading power in @xmath8 . in section 2 ,
we briefly discuss how we have evaluated the cross sections of the different contributions . in section 3
, we present and discuss our results .
section 4 gathers a detailed discussion of the relative size of the csm contribution via an off - shell _ photon _ w.r.t that of the com via an off - shell _ gluon_. we finally conclude in section 5 .
in the csm @xcite , the amplitude for the production of a @xmath19 quarkonium @xmath25 of a given momentum @xmath26 and of polarisation @xmath27 accompanied by other partons , noted @xmath28 , and a @xmath1 boson is written as the product of the amplitude to create the corresponding heavy - quark pair , a spin projector @xmath29 and @xmath30 , the radial wave function at the origin in the configuration space .
precisely , one has @xmath31 where one defines @xmath32 , @xmath33 , and where @xmath34,@xmath35 are the heavy - quark spin components and @xmath36 is the projector onto a colour - singlet state .
@xmath29 has a simple expression in the non - relativistic limit : @xmath37 where @xmath38 is the quarkonium polarisation vector .
once one sums over the heavy - quark spin components , one obtains traces which can be evaluated in a standard way .
in particular , for lo evaluations without loops one can simply use the framework described in @xcite based on the tree - level matrix element generator madonia @xcite .
another possibility would be to use helac - onia @xcite . for the cross - section evaluation
, we have used the parameters @xmath39 gev@xmath40 and br@xmath41 . neglecting relativistic corrections ,
one has in the csm , @xmath42 .
we have taken @xmath43 gev and @xmath44 .
the uncertainty bands for the resulting predictions are obtained from the _ combined _ variations of the heavy - quark mass within the range @xmath45 gev , with the factorisation @xmath46 and the renormalisation @xmath47 scales chosen among the couples @xmath48 . for the colour - octet contributions ,
the only relevant parameter is the nrqcd long distance matrix elements ( ldme ) @xmath49 } ) \rangle $ ] .
we have set it to @xmath50 gev@xmath51 , _
i.e. _ the value obtained in the recent global nlo analysis of butenschoen and kniehl @xcite .
this value is also of the order of what was obtained in another recent nlo nrqcd fit @xcite .
there are of course drawbacks in using ldme obtained from nlo fits .
first , nlo corrections to the hard part of color - octet processes to inclusive production show a @xmath52 factor higher than one which leads to a reduction of the co ldme compared to those extracted from a lo fit . yet ,
a comparison to the nlo results of @xcite for co channels indicate that our evaluation is reasonable
. moreover , various co contributions can interfere and fits can yield negative values .
for instance , a recent nlo fit has obtained such a negative result for this ldme @xcite .
it would not make much sense to use such a value in lo computations since the cross section would then be negative .
it is therefore important to recall that our choice is also close to the lo analysis of @xcite and from analyses which partially took into account qcd corrections @xcite .
our leading - order results for the differential cross sections in @xmath8 are shown in fig .
[ fig : dsigdpt ] for the tevatron @xmath53 , and for the lhc at @xmath54 tev @xmath55 and @xmath56 tev @xmath57 .
+ + at the tevatron , the com contribution ( orange band ) is significantly larger than that of the csm via @xmath58 fusion ( dark green band ) .
however , it is of similar size as the csm contribution via @xmath59 ( light blue band ) .
note that the light - blue band actually also contains other electroweak contributions appearing at the same order , _
i.e. _ via @xmath60 , but the yield is strongly dominated by processes via @xmath59 . at lhc energies , the three contributions are of the same order .
the total csm cross section is thus about twice as large as the com one , probably a bit more at 14 tev and at large @xmath8 ( see fig . [
fig : dsigdptc ] ) .
such results clearly demonstrate that , contrary to earlier claims in the literature @xcite , the yield for the production of @xmath0 in association with a @xmath1 boson _ can not _ actually be used as a clean probe of the com .
this remains true over the whole range in @xmath8 .
the born csm contributions considered here are indeed leading @xmath8 at variance to the inclusive case where leading-@xmath8 contributions @xcite only appear at higher orders in @xmath15 .
+ in addition to the @xmath8 dependence , we present in fig .
[ fig : dsigdy ] our csm results for the differential cross sections in @xmath61 for the lhc at @xmath54 tev @xmath53 and @xmath56 tev @xmath55 .
one observes that the csm yields via @xmath59 and via @xmath58 fusion are of the same order at the lhc energies , with an increasing proportion of @xmath58 fusion as the energy increases .
as we have seen above , the contributions from the csm via an off - shell photon and from the com via an off - shell gluon namely via a @xmath62}_1 $ ] state are similar , with nearly exactly the same @xmath8 dependence . we have found it instructive to investigate this . to this end , we have evaluated the cross section for a @xmath63 annihilation into a @xmath19 quarkonium in both channels .
apart from the radiation of the @xmath1 , this is the same process as discussed above .
the partonic cross section for the singlet contribution via an off - shell photon , @xmath64 , is : @xmath65}_{via \
\gamma^\star}=\frac{(4\pi\alpha)^2e_q^2e_q^2}{m^3_\mathcal qs}\,\delta\left(x_1x_2-m^2_\mathcal q / s\right ) @xmath66 the heavy quark charge and @xmath67 the light quark charge . for the octet contribution via a @xmath68}$ ] state
, one can follow petrelli _ et al . _ @xcite and obtain @xmath69}_{via \
g^\star}=\frac{(4\pi\alpha_s)^2 \pi}{27m^3_\mathcal qs}\,\delta\left(x_1x_2-m^2_\mathcal q / s\right ) \langle\mathcal o_{{{\cal q } } } ( \phantom{\!\!\!s}^3s^{[8]}_1)\rangle.\ ] ] where @xmath70}_1)\rangle$ ] is to be fitted to reproduce the @xmath8 spectrum of the data at the tevatron and at the lhc and , in principle also , the yield polarisation . in the singlet case
, one can connect the wave function at the origin to a similar matrix element : @xmath71})\rangle=2n_c(2j+1)\frac{|r(0)|^2}{4\pi}$ ] .
one can now make the ratio of the singlet to octet contribution : @xmath72}_{via \ \gamma^\star}}{\hat\sigma^{[8]}_{via \
g^\star}}=\frac{6\alpha^2e_q^2e_q^2\langle \mathcal o_{{{\cal q } } } ( \phantom{\!\!\!s}^3s^{[1]}_1)\rangle}{\alpha^2_s\langle\mathcal o_{{{\cal q } } } ( \phantom{\!\!\!s}^3s^{[8]}_1)\rangle}.\ ] ] the difference in the colour structure gives the relative factor @xmath73 between the octet and the singlet contributions . in the @xmath0 case , we have @xmath74}_1)\rangle=1.45 $ ] gev@xmath40 , @xmath75}_1)\rangle=2.2\times 10^{-3}$ ] gev@xmath40 as we used above and @xmath76 .
the ratio is then about two thirds for @xmath77 fusion . in the case of @xmath9 production ( @xmath78}_1)\rangle\simeq 10 $ ] gev@xmath40 , @xmath79})\rangle=0.4\div 3\ , \times\ , 10^{-2}$ ] gev@xmath40 at lo @xcite and @xmath80 ) , the ratio is similar to that of @xmath0 . along the same lines ,
if the quark line emits a @xmath1 boson , one expects the same ratio up to factors involving the quark ( @xmath81 and @xmath82 ) electric charges and 2 for the @xmath9 , since the natural scale of the process would then be @xmath83 rather than @xmath84 ; the strong coupling would then be smaller and the electroweak one larger . ] .
this convincingly explains the similarity between the orange ( com ) and light blue ( csm via @xmath59 ) bands on fig .
[ fig : dsigdpt ] c ) for instance .
both observations can surely be extended to nlo in @xmath15 and then by using , in a coherent manner , the nrqcd ldme values recently fit in _
e.g. _ @xcite . as we mentioned earlier ,
the ldme values extracted from these recent fits unfortunately depend much on the data sets which were used . from the simple computations done above , it can not therefore be excluded at all that csm contributions via an off - shell photon would in fact be larger than that from com contributions via an off - shell gluons in specific processes , as the ones studied here , where quark
antiquark annihilation is dominant .
let us emphasise that the partonic process @xmath85 is expected to take over @xmath86 and @xmath87 fusions in inclusive @xmath0 and @xmath9 production at energies close to threshold .
further quantitative statements however require a dedicated survey within the csm in particular in what concerns the feed - down from @xmath88
beside the discussion of the contributions from co and cs channels , there are additional important aspects to keep in mind when comparing the theoretical predictions for @xmath89 to data .
first , experimental analyses of @xmath1 production usually proceed by looking at the @xmath1 leptonic - decay products , in particular @xmath90 .
such events can be tagged by the presence of a missing mass carried by the undetected neutrino .
this however also means that one can not strictly enforce that the invariant mass of the @xmath90 pair equals that of the @xmath1 .
an unexpected consequence of this in the present study is that the ( rare ) 3-body decay , @xmath91 has previously been considered in @xcite . ] @xmath92 can not be disentangled from genuine @xmath93 events .
in fact , its contribution is not negligible with the typical cuts used at the lhc .
we have indeed found that with the cuts used by atlas @xcite ( @xmath94 gev , @xmath95 gev , @xmath96 , @xmath97}>40 $ ] gev ) , the process @xmath98 contributes nearly equally to that of @xmath99 , where @xmath100 .
second , atlas has evaluated @xcite , using the pocket formula @xmath101 , that a significant double - parton - scattering ( dps ) contribution as high as 40 % is to be expected provided that this formula makes sense and that one can use the effective cross section @xmath102 as extracted from the @xmath103 jets analysis @xcite . in principle , the dps signal is reducible since the @xmath0 and the @xmath1 should completely be uncorrelated in @xmath104 and @xmath8 . in practice , since one expects only a handful of events per fb@xmath105 , it will be very complicated to subtract with a good accuracy the sps signal by looking at the @xmath106 or @xmath107 distributions .
third , as for most quarkonium - production observables , feed - down from excited - quarkonium states can be important and proceeds from different partonic reactions .
we have indeed computed that the cross section for @xmath108 times the branching @xmath109 is about @xmath110 times larger than the direct cross section for @xmath89 . in short , the feed - down from @xmath88 is expected to be larger than in the inclusive case and can not be disregarded .
this simply comes from the possibility of a fragmentation contribution at @xmath111 .
summing the direct contribution to the feed - down from @xmath88 and @xmath112 , we find a total cross - section of @xmath113 fb at @xmath114 tev , comparable to the cross - section @xmath115 fb for dps - subtracted prompt @xmath13 recently obtained by the atlas collaboration @xcite .
we have shown that the lo csm contributions to direct @xmath12 are not negligible compared to the contribution arising from co transitions which were previously thought to be dominant .
these csm contributions arise from two sub - processes : a ) the fusion of a gluon and a strange quark which turns into a charm quark by the emission of the @xmath1 , the charm quark subsequently fragments into a @xmath116 pair ; b ) the annihilation of a quark @xmath81 and an antiquark @xmath117 into an off - shell photon , @xmath59 , and a @xmath1 , the @xmath59 subsequently fluctuates into a @xmath0 .
the former process appears at @xmath118 and the latter at @xmath5 compared to @xmath119 for the com process which is however suppressed in the @xmath120 expansion of nrqcd .
we have also noted that , for any @xmath19 quarkonium - production process involving quark
antiquark annihilation , the csm process via an off - shell photon numerically competes with the com one via an off - shell gluon through a @xmath68}$ ] octet . finally , owing to the uncertainties on the co ldme , the small rate for this process at the lhc and the possibility for large dps contributions , our conclusion is that the study of direct @xmath89 yields can not serve as a clean probe of the colour - octet mechanism , as previously stated in the literature .
we thank b. gong , d. price and j.x .
wang for useful discussions .
this work is supported in part by the france - china particle physics laboratory and by the p2i network .
j. p. lansberg , int .
j. mod .
a * 21 * , 3857 ( 2006 ) .
n. brambilla _ et al .
j. c * 71 * , 1534 ( 2011 ) .
z. conesa del valle , _ et al . _ ,
( ps ) * 214 * , 3 ( 2011 ) .
chang , nucl .
b * 172 * , 425 ( 1980 ) ; r. baier and r. rckl , phys .
b * 102 * , 364 ( 1981 ) ; r. baier and r. rckl , z. phys .
c * 19 * , 251 ( 1983 ) .
j. campbell , f. maltoni and f. tramontano , phys .
lett . * 98 * , 252002 ( 2007 ) .
p. artoisenet , j. p. lansberg and f. maltoni , phys .
b * 653 * , 60 ( 2007 ) .
b. gong and j. x. wang , phys .
* 100 * , 232001 ( 2008 ) .
b. gong and j. x. wang , phys .
d * 78 * , 074011 ( 2008 ) .
r. li and j. x. wang , phys .
b * 672 * , 51 ( 2009 ) .
j. p. lansberg , phys .
b * 679 * , 340 ( 2009 ) .
s. j. brodsky and j. p. lansberg , phys .
d * 81 * , 051502 ( 2010 ) .
j. p. lansberg , pos ichep * 2010 * , 206 ( 2010 ) .
r. barbieri , r. gatto and e. remiddi , phys .
b * 61 * ( 1976 ) 465 .
h. -s .
shao , comput .
commun .
* 184 * ( 2013 ) 2562 m. butenschoen and b. a. kniehl , phys .
d * 84 * ( 2011 ) 051501 k. -t .
chao , y. -q .
ma , h. -s .
shao , k. wang and y. -j .
zhang , phys .
* 108 * ( 2012 ) 242004 b. gong , l. -p .
wan , j. -x .
wang and h. -f .
zhang , phys .
lett * 110 * , 042002 ( 2013 ) e. braaten , s. fleming and a. k. leibovich , phys .
d * 63 * ( 2001 ) 094006 c. -f .
qiao , l. -p .
sun , d. -s .
yang and r. -l .
zhu , eur .
j. c * 71 * ( 2011 ) 1766 [ atlas collaboration ] , atlas - conf-2013 - 042 . | we show that the colour - singlet contributions to the hadroproduction of @xmath0 in association with a @xmath1 boson are sizable , if not dominant over the colour - octet contributions .
they are of two kinds , @xmath2 at @xmath3 and @xmath4 at order @xmath5 .
these have not been considered in the literature until now .
our conclusion is that the hadroproduction of a @xmath0 in association with a @xmath1 boson can not be claimed as a clean probe of the colour - octet mechanism .
the rate are small even at the lhc and it will be very delicate to disentangle the colour - octet contributions from the sizable colour - singlet ones and from the possibly large double - parton - scattering contributions . during this analysis ,
we have also noted that , for reactions such as the production of a @xmath0 by light quark
antiquark fusion , the colour - singlet contribution via an off - shell photon is of the order of the expectation from the colour - octet contribution via an off - shell gluon .
this is relevant for inclusive production at low energies close to the threshold .
such an observation also likely extends to other processes naturally involving light - quark annihilation .
quarkonium production , vector boson production 12.38.bx , 14.40.gx , 13.85.ni |
to model two @xmath0yb atoms of mass @xmath113 on a single site , we first consider the two - body problem in a 3d harmonic trap , @xmath114 . in this case , the center - of - mass and relative coordinates can be decoupled , and in the relative basis , we have hamiltonian @xmath115 , with @xmath116 and interaction part @xmath117 here , @xmath118 labels the relative harmonic oscillator states with angular momentum @xmath119 ( these are the only ones that are affected by the short - range interactions ) , @xmath120 is the real - space harmonic oscillator wavefunction at @xmath121 , and the non - interacting energy @xmath122 .
the triplet and singlet configurations are defined as @xmath123 , with corresponding interaction strengths @xmath124 , which do not depend directly on the nuclear spin . for simplicity ,
we focus on a zero - range interaction here , but finite - range corrections are easily incorporated into the formalism provided the range of the interactions is much smaller than the harmonic oscillator length @xmath125 . to determine the two - body eigenstates , we consider the general wavefunction in the singlet - triplet basis @xmath126 and then insert this into the schrdinger equation as follows : @xmath127 by summing over @xmath118 and replacing the bare interactions @xmath124 with the scattering lengths @xmath128 , we arrive at the matrix equation for the regular part of the two - channel wave function @xmath129 : @xmath130 \boldsymbol{\psi}_{\rm reg } = 0\end{aligned}\ ] ] where @xmath131 and the quantity @xmath132 contains the pair propagator in a harmonic potential @xcite @xmath133 the interaction energies are thus obtained from the condition @xmath134 . for finite - range corrections ,
we need to replace @xmath135 by an energy - dependent matrix that depends on the scattering phase shifts , as in eq . of the main text : @xmath136 where the collision energy is @xmath137 .
we are interested in the leading order terms of the low - energy expansion : @xmath138 where @xmath139 denotes the effective range .
formally , such an effective range may be included by using a two - channel model for each of the singlet and triplet interactions .
for the trapped two - body problem , the low - energy expansion of @xmath140 is thus equivalent to making the replacement @xmath141 in eq . .
the interaction energies of the two - body problem in a 3d harmonic trap are dependent on the trap energy spectrum for the single - channel problem @xcite . in order to take into account the anharmonic character of the lattice potential
, we expand this potential to the 6th order and correct the harmonic energy spectrum using second - order perturbation theory .
this result matches very well with the diagonalization of the full hamiltonian performed by deuretzbacher et .
al . @xcite .
using the determined scattering lengths and the corresponding effective ranges of the interactions , one can calculate the scattering amplitude resonance position in the bulk @xmath142 g and the zero crossing @xmath143 g .
these values correspond to zero energy scattering and the error bars were obtained allowing 10% of uncertainty in the determination of the scattering lengths .
in the cross - dimensional thermalization measurement , the atomic cloud is transferred into a crossed dipole trap operating at the magic wavelength with trap frequencies @xmath144hz .
simultaneously , the magnetic field is changed in a two - step ramp : first it reaches a value 100 g above the value of interest in 100ms and then is finally ramped with a fixed speed of 2g / ms .
directional heating is produced in a controlled way by flashing the lattice potential along the vertical axis .
twenty pulses are applied over a time of three trap oscillation periods , in order to ensure that the cloud is thermalized along the excitation direction . after a variable hold time
, the trap is switched off and the atoms in the ground state are imaged after 12ms of free flight . in order to calibrate the magnetic field in our experiments
, a measurement of the zeeman shift of the @xmath145 @xmath146 @xmath147 transitions between stretched states is performed .
a subsequent measurement of the zeeman shift of the @xmath145 @xmath146 @xmath148 transitions showed very good agreement with previously published values @xcite .
atoms and ( b ) a mixture of @xmath149 atoms .
inset of ( a ) : trap geometry of thermalization experiment , with dipole trap beams in light red , heating lattice beams in dark red and the imaging beam in blue .
the images are oriented in the @xmath150-plane . ] as described in the main text , a cross - dimensional thermalization experiment is performed to determine the magnetic - field dependence of the elastic scattering cross - section of a @xmath149 mixture .
the aspect ratio of the cloud is measured by performing a gaussian fit to the density distribution obtained by absorption imaging after 12ms of time of flight . as an example , the time evolution of the cloud sizes for two magnetic field values is plotted in fig .
[ figsup : aspectratio ] . the magnetic - field independent thermalization of the @xmath151 mixture
is shown for comparison in [ figsup : aspectratio](a ) .
for the cross - thermalization rate measurement , two different values of atomic density are used . for magnetic fields below 200 g , close to the feshbach resonance ,
the density is lower in order to compensate for the large scattering cross - section . at higher magnetic fields ,
the higher value of atomic density is used .
the two data sets are merged together using as reference the corresponding magnetic - field independent thermalization rates of the @xmath151 mixture .
n. poli , z. w. barber , n. d. lemke , c. w. oates , l. s. ma , j. e. stalnaker , t. m. fortier , s. a. diddams , l. hollberg , j. c. bergquist , a. brusch , s. jefferts , t. heavner , and t. parker , phys .
a * 77 * , 050501 ( 2008 ) . | we report on the experimental observation of a novel interorbital feshbach resonance in ultracold @xmath0yb atoms .
this opens up the possibility of tuning the interactions between the @xmath1 and @xmath2 metastable state , both possessing zero total electronic angular momentum .
the resonance is observed at experimentally accessible magnetic field strengths and occurs universally for all hyperfine state combinations .
we characterize the resonance in the bulk via interorbital cross thermalization as well as in a three - dimensional lattice using high - resolution clock - line spectroscopy .
our measurements are well described by a generalized two - channel model of the orbital - exchange interactions .
magnetic feshbach resonances have become an indispensable tool in the study of ultracold quantum gases , enabling the tuning of interaction strengths over a wide parameter range @xcite .
this tunability has given rise to an impressive set of experimental results , including the realization of the bec bcs crossover in degenerate fermi gases @xcite as well as the discovery of novel few - body phenomena such as efimov trimers @xcite . whereas the majority of alkali atomic species feature feshbach resonances , this is not the case for alkaline - earth - like atom states without electronic angular momentum .
in this letter , we report on the observation of a recently predicted novel type of feshbach resonance @xcite , enabling the tuning of interorbital interactions based on the zeeman shift of different nuclear spin states of the atoms . in a feshbach resonance , a bound state of the energetically inaccessible molecular potential ( closed channel ) couples to the scattering continuum of the energetically allowed entrance channel .
this coupling dramatically affects the atomic scattering properties whenever the bound state energy is close to the open channel energy . in magnetic feshbach resonances
, an external magnetic field can tune these states into and out of resonance , as they have different magnetic moments such as in the prototypical example of singlet and triplet scattering channels of two alkali atoms @xcite . .
a finite magnetic field induces a deformation of the scattering potential for large atomic separation ( yellow solid lines ) , proportional to the differential magnetic moment .
the inset illustrates the origin of the differential magnetic moment due to different land g - factors of the @xmath1 ( bright symbols ) and @xmath2 ( dark symbols ) states . ] in contrast , alkaline - earth - type atoms with two valence electrons such as ytterbium and strontium possess no electronic angular momentum in the atomic ground state , either for the electronic spin singlet @xmath1 ( denoted @xmath3 ) or for the triplet @xmath2 ( @xmath4 ) .
this and the associated suppression of hyperfine coupling make these atoms superb candidates for atomic clocks @xcite and for the investigation of new exotic many - body states with extended su(@xmath5 ) symmetry @xcite .
however , this also implies that no magnetic feshbach resonances are expected within the two orbital states . instead
, optical feshbach resonances @xcite , based on the coupling to a bound molecular state via a one- or two - photon process have been investigated @xcite .
because of their limited tunability and lifetime , such optical feshbach resonances have been difficult to exploit in experiments . until now
, magnetic feshbach resonances in ytterbium have only been observed when atoms are specifically prepared in the @xmath6p@xmath7 state possessing electronic magnetic moment and hyperfine coupling @xcite .
however , recent measurements @xcite of the scattering properties of a @xmath1 @xmath2 atom pair in two different nuclear spin states suggest the existence of a shallow molecular bound state , leading to the prediction of a magnetically accessible scattering resonance @xcite . the scattering resonance is described by interaction channels possessing both orbital and nuclear degrees of freedom in @xmath0yb ( @xmath8 ) . as in @xcite
, the interaction can be expressed in the basis consisting of symmetric and antisymmetric superpositions @xmath9 of the electronic orbitals ( @xmath4,@xmath3 ) and the two nuclear spin states ( @xmath10,@xmath11 ) with @xmath12 .
associated to @xmath13 are the orbital singlet scattering length @xmath14 @xcite and the very large triplet scattering length @xmath15@xmath16 @xcite , with @xmath16 denoting the bohr radius .
the two orbitals exhibit different land g - factors , due to a weak hyperfine coupling of the @xmath2 with the @xmath17 state , giving rise to a differential magnetic moment @xmath18 , where @xmath19 is the bohr magneton .
the presence of a magnetic field will therefore mix the singlet and triplet states , introducing a coupling between the orbital and spin degrees of freedom . in this case , the non - interacting system has the eigenbasis @xmath20 and @xmath21 . two atoms entering the potential in the open channel eigenstate @xmath22 couple to @xmath23 through the orbital - mixing interaction term . at short distances
the molecular interaction potentials dominate and are independent of the magnetic field due to the symmetric nature of the @xmath13 states .
nevertheless , the entrance channel @xmath22 can be brought into resonance with bound states of the interorbital molecular potentials by shifting the entrance energy using the differential zeeman shift @xmath24 as illustrated in fig .
[ fig : schematics ] .
the resonance occurs when @xmath25 comes close to the binding energy of the bound state in the closed channel @xmath26 , with @xmath27 . for one- and two - orbital mixtures .
circles and diamonds mark the thermalization rate of the @xmath28 and the @xmath29 mixture , respectively with @xmath30 ( @xmath31 ) .
square symbols show the values for a @xmath28 mixture with @xmath32 ( @xmath33 ) rescaled to @xmath34 .
error bars denote the @xmath35 uncertainty of the fit to the cloud aspect ratio .
the inset shows a conversion of the thermalization rate to the scattering length @xmath36 , based on the reference provided by the ground state @xmath37 @xcite ( dashed line ) . in the inset ,
the offset due to the residual ground state thermalization rate @xmath38 has been subtracted ( see main text ) . ] to determine the magnetic field dependence of the scattering length , we perform cross - dimensional thermalization measurements in a 3d harmonic trap @xcite .
a mixture of @xmath39 and @xmath40 is prepared in an out - of - equilibrium state and rethermalizes with the mean rate of elastic collisions , which can be derived from enskog s equation @xcite . in the limit of low - energy scattering , the collision rate ( and therefore the thermalization rate )
is proportional to the elastic scattering cross section @xmath41 .
our experiments begin by creating a degenerate fermi gas via evaporative cooling in a crossed dipole trap @xcite .
the desired spin mixture is prepared by optical pumping , resulting in a two - component @xmath42 gas with typical temperatures @xmath43 and a total atom number of @xmath44 per spin state , where @xmath45 is the fermi temperature . in order to populate the @xmath2 state ,
the atoms are first loaded into a one - dimensional , state - independent , i.e. , magic - wavelength @xmath46 nm lattice @xcite in the lamb - dicke regime .
this lattice is coaligned with the 579 nm excitation beam used to apply a @xmath47-pulse on the @xmath48 optical clock transition at an applied magnetic field of @xmath49 g .
next , the atoms are released adiabatically into a magic - wavelength dipole trap ( see supplementary information @xcite ) .
the sample is then heated along the @xmath50-direction by repeated bragg pulses using a standing wave along this axis , and the subsequent thermalization of excitation into the orthogonal directions is observed by measuring the cloud aspect ratio @xmath51 after 13ms of time of flight .
we determine the thermalization rate @xmath52 by a single exponential fit to the time dependence @xmath53 @xcite .
we observe a change in the thermalization rate over 2 orders of magnitude as a function of the magnetic field as shown in fig .
[ fig : crosstherm ] . in contrast , for a spin mixture in the ground - state orbital @xmath3 , we find the thermalization rate @xmath54 to be independent of the magnetic field .
the resulting @xmath55-field dependence of @xmath52 exhibits the characteristic shape of a feshbach resonance with a peak position @xmath56 g , and a zero crossing at @xmath57 g , both determined by quadratic fits within @xmath58 g ( @xmath59 g ) regions around the resonance ( zero - crossing ) position , respectively .
the excitation process has an efficiency of approximately 90% and the thermalization rate in the zero - crossing regime of @xmath36 is therefore bounded from below by the thermalization of the @xmath60 atoms with the residual @xmath61 atoms .
a unique property of this new type of feshbach resonance is that it involves only two channels , which additionally are fully su(@xmath5 ) symmetric in the absence of a magnetic field , making the coupling universal with respect to the choice of @xmath62 states .
we demonstrate this by repeating the measurement with another spin mixture with a different value of the differential magnetic moment @xmath63 .
when @xmath64 is taken into account by rescaling the magnetic field axis correspondingly , the data sets collapse onto a single curve without further adjustments ( see fig . [
fig : crosstherm ] ) , demonstrating the universal behavior with respect to different @xmath62-state combinations .
mixture held at a magnetic field @xmath55 near the orbital feshbach resonance .
( a ) number of atoms in the state @xmath65 with @xmath66 ( circles ) and the residual @xmath67 fraction with @xmath68 ( diamonds ) .
all data points represent averages of at least eight individual measurements and points less than 2.5 g apart are binned to reduce visual clutter .
error bars indicating the standard error of the mean are smaller than the marker size .
( b ) relative population of @xmath65 , i.e. , number of atoms in @xmath65 normalized to @xmath69 , where @xmath70 is the ground - state atom number without losses , averaged for fields @xmath71 g , and @xmath72 is the residual atom number in @xmath67 .
the inset shows the time trace of the total atom number decay close to the feshbach resonance . ] in order to characterize the loss channel of the resonance , we use the same preparation protocol as for the cross thermalization , omitting the heating procedure . after a fixed hold time of 150ms , the remaining ground - state atoms are imaged spin selectively using an optical stern - gerlach technique @xcite .
as shown in fig .
[ fig : losses]a , when reducing the holding field we observe a magnetic field - dependent loss of the @xmath65 population starting close to the position of the resonance . however , the minimum of the loss feature does not occur until the field approaches @xmath73 .
below 50 g , we observe a significant repopulation of the @xmath67 state , since orbital exchange becomes energetically favorable at low fields @xcite .
accounting for this exchange , the loss feature still remains strongly shifted towards lower magnetic fields compared to the resonance position in the elastic channel ( fig .
[ fig : losses]b ) .
this is in fact similar to observations in other mixtures of fermions @xcite , but the shift is possibly enhanced due to the intrinsically large size of the shallow bound state . at the resonance position , we extract a lifetime of @xmath74ms at a temperature @xmath75 and an initial peak density of @xmath76 atoms/@xmath77 . assuming three - body decay for the @xmath78@xmath65 mixture , we obtain a loss rate coefficient of . the observed loss is comparable to previously reported values in alkali fermi gases , e.g. , @xmath79k mixtures @xcite .
this realization of a stable strongly interacting fermi gas appears favorable for exploring novel superfluidity phenomena with this feshbach resonance in @xmath0yb @xcite . in a third experiment
, we directly probe the two - particle interaction on individual sites of a deep , 3d magic - wavelength lattice , as described in @xcite .
this offers the advantage of determining the interaction shift of the atom pair with high resolution using clock - line spectroscopy .
furthermore , in such a setting the two - particle problem can be separated from possible many - body effects and allows for precise determination of the interaction energies .
the energies of the two - particle states that can be excited on the open channel transition at varying magnetic fields are shown in fig .
[ fig : onsite ] .
the energy is given relative to that of two non - interacting spatially separated .
the data points correspond to resonance positions obtained in an @xmath80 spin mixture for an isotropic lattice depth of @xmath81 , where @xmath82 is the recoil energy of the lattice light .
the lowest observed branch ( circles ) has negative interaction energy for @xmath83 g . for low fields
, this state corresponds to the molecular branch of @xmath84 . extrapolating the measurements to @xmath73
, we find an on - site bound state energy of @xmath85khz .
@xmath3 atom pair interaction energy on a lattice site .
the values are given relative to the energy of two non interacting atoms .
the assignment of the spectroscopy resonances ( points ) is based on the observed transition strengths , which decrease for higher lying energy states .
solid lines are solutions of the two - particle problem .
light blue bands indicate the range of variations of the theoretical model spanned by varying @xmath86 and @xmath87 by 10% .
( inset ) schematic diagram of the excitation geometry . ]
the next higher - lying energy branch was used in @xcite to determine the value of @xmath88 .
we verify with spin - sensitive measurements that this state evolves from the antisymmetric superposition state @xmath89 to the ( band - excited ) open - channel state @xmath22 with increasing magnetic fields . using spin - exchange oscillations at low fields , an indirect measurement of @xmath90
was also performed @xcite . to model the on - site @xmath91-@xmath92 interaction ,
we generalize the two - channel description of ref .
@xcite to the case of isotropic confinement and energy - dependent collisions . in the absence of a magnetic field ,
the interactions in the singlet and triplet channels decouple .
consequently , the energy shift @xmath93 is determined by the condition @xmath94+\pi(e)=0 $ ] , which depends on the singlet and triplet scattering phase shifts , @xmath95 , and the renormalised pair propagator in the lattice site , @xmath96 @xcite . however , in general we need to take into account the coupling introduced by the magnetic field , and this converts the problem into a matrix equation .
thus , the condition for the interaction energy becomes @xmath97 = 0 , \label{eq : energycondition}\ ] ] with @xmath98 and @xmath99 .
for details of the derivation , see @xcite .
note that our model goes beyond the usual solution of the two - body problem in a harmonic trap @xcite and accounts for how both open and closed channels are modified under confinement .
we apply a low - energy expansion of the phase shifts up to and including effective range .
the magnetic field dependence of all measured interaction shifts is very well reproduced by our model ( solid lines in fig .
[ fig : onsite ] ) . a best fit of the model with @xmath86 as a free parameter yields @xmath100 .
the corresponding effective range @xmath101 is calculated analytically @xcite , assuming a long - range van der waals potential with @xmath102a.u .
@xcite .
the fit uncertainty on @xmath86 is @xmath103 , but we expect that the uncertainty from model simplifications such as the lattice and scattering phase expansions are comparable or larger than this . to indicate the variability , we plot a range corresponding to @xmath104% variation of both scattering lengths as shaded areas in fig .
[ fig : onsite ] .
the value for @xmath105 obtained applying the model to the @xmath73 data from @xcite is consistent with the one reported there , and has been used in this work with effective range @xmath106 .
the spectroscopy results also enable us to derive a feshbach resonance position @xmath107 g and the zero - crossing @xmath108 g for the bulk at @xmath109nk .
both values are in good agreement with the results obtained from the independent cross - thermalization measurements . in conclusion ,
we have observed a new type of scattering resonance between different atomic orbitals of @xmath0yb arising from strong interorbital spin - exchange interactions . because of the su(@xmath5)-symmetric nature of the exchange interaction @xcite , the resonance occurs universally for any spin combination .
we have precisely characterized this feshbach resonance using a new model of the on - site atom pair interaction energy shift that incorporates the effect of confinement on both open and closed channels . even in the degenerate ,
strongly interacting quantum gas on resonance , we observe a long lifetime , making our system promising for observing two - orbital fermi gases with exotic order parameters @xcite .
furthermore , the tunability of the e - g interaction strength suggests novel avenues for the experimental implementation of two - orbital many - body lattice models @xcite .
we acknowledge helpful input from frank deuretzbacher and valuable discussions with hui zhai .
this work was supported by the erc through the synergy grant uquam and by the european union s horizon 2020 funding ( d.r.f . ) .
the work of m.m.p . , j.l . , and i.b .
was performed in part at the aspen center for physics , which is supported by national science foundation grant no .
phy-1066293 .
35ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/revmodphys.82.1225 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.120403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.040403 [ * * , ( ) ] link:\doibase 10.1126/science.1100818 [ * * , ( ) ] link:\doibase 10.1007/978 - 3 - 642 - 21978 - 8 [ _ _ ] , edited by , , vol .
( , , ) http://dx.doi.org/10.1038/nature04626 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.135301 [ * * , ( ) ] \doibase doi:10.1016/s0370 - 1573(99)00025 - 3 [ * * , ( ) ] \doibase doi:10.1016/j.physrep.2004.03.003 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.78.1311 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.87.637 [ * * , ( ) ] link:\doibase 10.1038/nphys1535 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.77.2913 [ * * , ( ) ]
link:\doibase 10.1103/physrevlett.101.203201 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.123201 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.173201 [ * * , ( ) ] link:\doibase 10.1038/nphys3061 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.120402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.70.414 [ * * , ( ) ] link:\doibase 10.1103/physreva.71.043408 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.103002 [ * * , ( ) ] @noop link:\doibase 10.1103/physreva.77.050501 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.123201 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.012719 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.190401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.91.240402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.91.020402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.083201 [ * * , ( ) ] http://stacks.iop.org/1367-2630/13/i=5/a=055003 [ * * , ( ) ] \doibase 10.1023/a:1018705520999 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.032726 [ * * , ( ) ] link:\doibase 10.1103/physreva.58.4222 [ * * , ( ) ] link:\doibase 10.1103/physreva.59.1998 [ * * , ( ) ] link:\doibase 10.1103/physreva.89.012711 [ * * , ( ) ] @noop ( ) , * supplemental material * + * * m. hfer@xmath110 , l.riegger@xmath110 , f. scazza@xmath110 , c. hofrichter@xmath110 , d. r. fernandes@xmath110 , + m. m. parish@xmath111 , j. levinsen@xmath111 , i. bloch@xmath110 & s. flling@xmath112 |
the present work is part of a comprehensive investigation of the local interstellar medium ( ism ) towards the southern coalsack and the chamaeleon - musca dark clouds .
a comparison of the colour excess @xmath9 vs. distance diagrams for the chamaeleon - musca complex @xcite and the southern coalsack @xcite shows remarkable similarities .
the jump of the colour excess to higher values occurs approximately at the same distance , and the observed minimum value of this rise is almost the same : @xmath10 .
although the clouds are apart by more than 15 these facts suggested that they might be dense condensations embedded in an extended interstellar structure . in order to investigate the possible physical association between these two clouds ,
a new photometric investigation was carried out .
the campaign resulted on very precise @xmath11 data for 1017 stars covering these clouds and the connecting area : @xmath12 and @xmath13 @xcite .
analysis of the colour excess @xmath14 distance diagrams for these stars indicated that exists a foreground region with very little reddening ( @xmath9 @xmath5 0006 ) , that is bound , at the distance range of 150 @xmath15 30 pc , by a transition region where the colour excesses suddenly increase to a mininum value of @xmath9 @xmath5 0050 for the most diffuse lines of sight and @xmath9 @xmath5 0100 for the densest ones . beyond this transition region
the diagrams suggest the existence of a second low column density volume , as the complete range of colour excess from @xmath5 0050 to @xmath16 , shown by the stars in the very narrow distance slot centered on 150 pc , remains unchanged for at least another 350 pc ( * ? ? ?
* hereafter paper i ) .
the onset of this minimum colour excess for the whole surveyed area suggests that the obscuring material is distributed in an extended interstellar sheet - like structure .
the existence of such an absorbing feature , at a distance identical to the molecular clouds , also suggests that coalsack , chamaeleon and musca can be dense condensations embedded on the diffuse medium composing the sheet - like feature .
it has been also noted in paper i that the jump in @xmath9 , caused by the dust feature , seems to have a dependence on the galactic latitude that can be represented by @xmath17 = [ 0\fm050;0\degr ] \rightarrow [ 0\fm100;-8\degr ] \rightarrow [ 0\fm150;-15\degr]$ ]
. the quoted dependence might indicate , if we postulate that the dust sheet is roughly perpendicular to the galactic plane , that the sheet does not have the same optical thickness , because @xmath18 _ vs. _ @xmath19 does not follow a simple @xmath20 law . on the other hand , it might also be the result of approaching the tangential point of a warped sheet - like structure with the same column density , curved away from the sun .
as pointed out in paper i , when viewed in connection to the other data on the local ism , the existence of these two low - reddening volumes has led to the idea that the dust sheet could be part of a large scale structure , probably related to the interface of the local and loop i bubbles .
such an interface of dense matter may have resulted from the compression of the ism by the action of energetic events of the sco - cen ob association from the far side ( @xcite ) and either by a supernova explosion near the sun or by whatever created the local low density region from the near side ( e.g. @xcite ) .
if the dust sheet is part of a small bubble , it could be associated to the lcc - shell proposed by @xcite .
note however that @xcite find that gas is moving out from sco - cen in the fourth galactic quadrant , but towards the association in the first quadrant .
recently , observational evidences of an annular volume of dense neutral matter , that supposedly would have been formed during the collision of the two bubbles , was found on the rosat all - sky survey data by @xcite .
remarkably , as noticed by @xcite , the molecular gas from @xmath21 to @xmath22 and @xmath23 seemingly define a large complex of clouds around 100 - 150 pc from the sun , comprising , e.g. , @xmath24 oph , r cra , g317 - 4 , musca , coalsack , chamaeleon and the lupus clouds . in this case
a schematic drawing of the interaction zone between the two bubbles and the dark clouds towards the sco - cen association would be like fig .
the position and sizes of the dark clouds were obtained from @xcite and the ring - like contours from @xcite .
the dashed square delineates the region surveyed in this work . to investigate the kinematics of the interstellar gas components towards the coalsack , chamaeleon and musca dark clouds we have obtained high - resolution ( @xmath0 ) spectra of the interstellar absorption lines . from the radial velocities , column densities , velocity dispersions and the known distance of the stars we aim to study
the physical association proposed for these clouds .
in order to investigate the interstellar gas components towards the coalsack , chamaeleon and musca dark clouds 60 b - type stars covering the area defined by the galactic coordinates : @xmath2 and @xmath25 have been selected from the photometric @xmath11 sample by @xcite .
all targets are located within 500 pc from the sun and have well determined colour excesses @xmath9 and photometric distances .
sao257142 , sao258697 and ppm377063 were observed to complement the data at more negative latitudes .
the basic information on the 65 stars is listed in table 1 .
for each one the visual magnitude @xmath26 , the galactic longitude and latitude , the colour excess and the distance are given in successive columns . based on the @xmath11 data the mean errors in the colour excess and distance values are 0018 and 15 - 30% , respectively .
the positions of the stars within the surveyed area are given in fig .
[ obpos ] .
the clouds contours defined by the thick lines are the lowest opacity level of the photographic catalogue by @xcite and the thin one is the outer 2 k km s@xmath27 velocity integrated co emission contour for the southern coalsack @xcite . velocity integrated co emission contour for the coalsack .
the ( @xmath28 ) , ( @xmath29 ) and ( @xmath30 ) signs indicate the stars with line - of - sight _ inside _ the clouds contours , in their _ outskirts _ and towards the _ sheet - like _ structure , respectively .
the horizontal and vertical dashed lines delineate eight sub - areas , identified by the roman numbers , that are used to discuss the details of the velocity structure . ]
the data sample consists of three general divisions based upon the star s location within the area : _ inside _ the clouds contours , in their _ outskirts _ and towards the _ sheet - like _ structure .
they are , respectively , indicated by the ( @xmath28 ) , ( @xmath29 ) and ( @xmath30 ) signs .
the horizontal and vertical dashed lines delineate eight sub - areas , identified by the roman numbers , that will be used to discuss the details of the velocity structure .
note that 5 stars are within 100 pc , while 7 other are between 100 and 150 pc and another 7 between 150 and 200 pc from the sun . data . ]
rcrrcrclcrrcl & @xmath26 & @xmath31 ( ) & @xmath32 ( ) & @xmath9 & @xmath33(pc ) & & & @xmath26 & @xmath31 ( ) & @xmath32 ( ) & @xmath9 & @xmath33(pc ) + & & + 239370 & 8.754 & 294.5 & 3.1 & 0.132 & 427 & & 239779 & 7.743 & 298.1 & 3.6 & 0.107 & @xmath34468 + 239702 & 7.160 & 297.6 & 2.6 & 0.079 & 379 & & 240265 & 7.183 & 302.5 & 3.2 & 0.106 & @xmath34489 + 240621 & 8.621 & 305.3 & 3.0 & 0.274 & 454 & & 240368 & 4.617 & 303.3 & 3.7 & 0.014 & @xmath34136 + 240645 & 4.624 & 305.5 & 2.8 & 0.022 & 65 & & 251460 & 7.167 & 295.0 & @xmath43.7 & 0.199 & @xmath34312 + 251530 & 8.217 & 294.7 & @xmath40.4 & 0.105 & 285 & & 251554 & 8.215 & 296.1 & @xmath44.2 & 0.272 & @xmath34291 + & & & & & & & & & & & & + 251545 & 8.018 & 295.3 & @xmath41.9 & 0.115 & 417 & & 251625 & 7.332 & 296.9 & @xmath43.2 & 0.149 & @xmath34347 + 251582 & 6.880 & 296.0 & @xmath42.7 & 0.164 & 291 & & 251717 & 4.724 & 297.6 & @xmath40.7 & 0.072 & @xmath34174 + 251837 & 7.519 & 298.9 & 1.1 & 0.090 & 309 & & 251802 & 7.111 & 299.2 & @xmath43.9 & 0.103 & @xmath34422 + 251841 & 4.053 & 299.3 & @xmath41.3 & 0.006 & 99 & & 251894 & 7.897 & 300.6 & @xmath46.7 & 0.083 & @xmath34348 + 251979 & 8.378 & 301.3 & 1.2 & 0.139 & 483 & & 251903 & 4.868 & 300.1 & @xmath40.3 & 0.015 & @xmath34124 + & & & & & & & & & & & & + 251988 & 8.838 & 301.9 & @xmath46.8 & 0.139 & 402 & & 251928 & 6.920 & 300.2 & 0.8 & 0.096 & @xmath34313 + 252114 & 8.092 & 303.7 & @xmath40.7 & 0.226 & 498 & & 251987 & 6.266 & 301.7 & @xmath43.6 & 0.021 & @xmath34117 + 252245 & 6.674 & 305.3 & @xmath45.7 & 0.114 & 123 & & 252262 & 7.258 & 305.7 & @xmath44.6 & 0.050 & @xmath34172 + 252284 & 4.529 & 306.7 & 1.6 & 0.018 & 96 & & 252294 & 7.810 & 306.0 & @xmath44.6 & 0.108 & @xmath34441 + 256763 & 7.978 & 296.2 & @xmath415.8 & 0.465 & 229 & & 252321 & 6.198 & 306.1 & @xmath46.9 & 0.108 & @xmath34196 + & & & & & & & & & & & & + 256798 & 7.675 & 296.6 & @xmath414.5 & 0.220 & 198 & & 256789 & 7.970 & 296.2 & @xmath414.1 & 0.161 & @xmath34227 + 256849 & 6.883 & 297.3 & @xmath411.8 & 0.286 & 234 & & 256900 & 8.113 & 299.3 & @xmath49.9 & 0.194 & @xmath34211 + 256924 & 4.240 & 301.3 & @xmath416.5 & 0.007 & 86 & & 256980 & 8.440 & 302.6 & @xmath48.7 & 0.159 & @xmath34460 + 256955 & 3.841 & 301.4 & @xmath49.3 & 0.003 & 97 & & 256953 & 7.777 & 301.9 & @xmath415.3 & 0.160 & @xmath34284 + + 239327 & 8.684 & 294.5 & 1.6 & 0.068 & 500 & & 256828 & 6.399 & 295.9 & @xmath410.3 & 0.157 & @xmath34196 + 239423 & 8.239 & 294.8 & 4.0 & 0.036 & 411 & & 256834 & 5.562 & 296.3 & @xmath410.5 & 0.192 & @xmath34263 + 239531 & 7.455 & 296.2 & 2.5 & 0.056 & 459 & & 256843 & 6.079 & 296.6 & @xmath410.6 & 0.237 & @xmath34135 + 240041 & 7.758 & 300.5 & 2.8 & 0.123 & 231 & & 257002 & 9.163 & 303.7 & @xmath410.6 & 0.148 & @xmath34368 + 251629 & 8.121 & 296.2 & @xmath40.0 & 0.105 & 433 & & 257032 & 7.628 & 305.5 & @xmath47.8 & 0.093 & @xmath34284 + & & & & & & & & & & & & + 251699 & 5.896 & 298.5 & @xmath46.7 & 0.006 & 99 & & 257033 & 6.045 & 305.3 & @xmath49.4 & 0.145 & @xmath34140 + 251774 & 9.160 & 298.1 & 0.2 & 0.107 & 384 & & 257047 & 6.618 & 305.3 & @xmath411.9 & 0.020 & @xmath34175 + 251874 & 8.124 & 300.2 & @xmath45.0 & 0.157 & 322 & & 257068 & 8.747 & 306.3 & @xmath410.2 & 0.126 & @xmath34276 + 251942 & 8.072 & 300.5 & 1.8 & 0.094 & 454 & & 257069 & 6.308 & 305.9 & @xmath413.0 & 0.069 & @xmath34149 + 252136 & 9.000 & 304.1 & @xmath43.2 & 0.078 & 353 & & 257142 & 5.000 & 307.0 & @xmath418.0 & & @xmath34169@xmath35 + & & & & & & & & & & & & + 256745 & 5.964 & 297.1 & @xmath418.3 & 0.076 & 255 & & 258697 & 6.800 & 306.0 & @xmath420.3 & & @xmath34270@xmath35 + 256788 & 6.759 & 294.7 & @xmath411.1 & 0.143 & 190 & & 377063@xmath36 & 5.100 & 295.0 & @xmath421.0 & & @xmath34165@xmath35 + 256800 & 5.585 & 294.5 & @xmath49.7 & 0.106 & 243 & & & & & & & + + high resolution ( @xmath37 60000 ) spectra of the d@xmath38 and d@xmath39 absorption lines superposed on the continua spectra of the selected stars have been obtained with the 1.4 m coud auxiliary telescope ( cat ) and the coud echelle spectrometer ( ces ) at the european southern observatory ( eso ) , in la silla ( chile ) .
the observations were carried out during 5 nights in april 1996 and 6 nights in april 1997 , using the remote control facilities at the eso s headquarters in garching bei mnchen ( germany ) .
the ces optics was optimized to give a reciprocal dispersion of 1.8 / mm at the central wavelength of 5890 , resulting in a spectral coverage of nearly 60 .
the entrance slit width of 459 @xmath40 m , used for all the observations , yielded an actual instrumental resolution ( fwhm ) @xmath41 m or a velocity resolution @xmath42 = 5.3 km s@xmath27 ( fwhm ) .
this corresponds to an actual resolving power of @xmath37 56000 , as determined by measurements of the intrinsic width of the thorium lines .
the ccd was a loral / lesser 2688 x 512 , with 16 pre - scan pixels .
it has a pixel size of 15 @xmath40 m @xmath43 15 @xmath40 m and is relatively free of blemishes . since the optical train selects only a single order , it was always possible to position the spectrum on a clean portion of the ccd .
the dark current and the readout noise are very low , 1.8 @xmath44/pixel / hour at 164 k and 8.3 @xmath44 , respectively . the combined efficiency of the cat / ces using the long camera is about 8% in the used central wavelength .
additional information regarding the cat / ces can be found in @xcite .
multiple sets of bias , dark and flat - field exposures were obtained each night for calibration purposes .
exposures of a thorium lamp were taken before and after each object exposure to ensure a better wavelength calibration .
the ces , however , showed to be a very stable instrument .
exposure times ranged from 2 minutes to 2.5 hours , depending on the visual magnitude and spectral type . in all cases
signal - to - noise ratios ( s / n ) greater than 150 in the raw spectra were achieved .
at least two exposures were obtained at each setup position and the integration times were limited to a maximum of 30 minutes to minimize the effects of cosmic ray events in the individual spectra .
although _ hipparcos _ satellite was launched in 1989 august , the catalogue only became available in 1997 @xcite .
for this reason , the observing sample was selected based on our photometric distance determination . however , the paralactic distances may now serve as a basis for comparison with our photometric distances .
a search in the _ hipparcos _ catalogue provided us 51 stars in common with our sample one of them , sao239370 , appears with negative value for its trigonometric parallax , and was excluded from the comparison .
[ hipparcos ] shows a plot containing the distances based on the _ hipparcos _ parallax vs. photometric distances .
the denoted errors for the photometric distances were estimated for each star based on the individual uncertainties of their photometric measurements @xcite , and ranges from 11 to 49 per cent . in general , the agreemment is very good testifying the high quality of our photometric distance determination .
there are few stars showing rather large discrepancy between their estimated photometric and parallatic distances , usually those beyond 200pc and large uncertainties in their trigonometric parallax .
however , as will be seen further , the main interstellar structures discussed in this work are located nearer than that , and consequently these uncertainties in the distance estimation do not affect substantially our conclusions .
the initial processing of the ccd frames employed iraf routines to subtract the bias , divide by a normalized flat - field , and remove cosmic rays from the rows occupied by the stellar spectrum and the background regions .
the one - dimensional spectra were then extracted using iraf routines .
it is worth reiterating that usually the dome flat - fields provided the best correction for the strong vignetting present in the spectra borders , besides that the affected parts were trimmed off the images .
the wavelength calibration was established with a set of 20 to 30 lines , identified using the thorium line wavelengths of @xcite .
a second - order chebyshev function was fitted to the line wavelength as a function of the pixel number , with rms scatter of the residuals about the fit smaller than 0.002 ( 0.05 km s@xmath27 ) in all cases .
the dispersion solution was applied to all science exposures with no rebinning of the data , and the final spectra normalized to unity continuum with cubic splines .
the individual normalized spectra were then co - added on a pixel - by - pixel basis , weighted by the inverse of the rms deviation of the continuum fit .
no noticeable degradation in spectral resolution resulted from the co - addittion procedure , since the individual spectra have been obtained within an hour or two of each other , and the velocity changes in the observer rest frame , over this time scale , are very small compared to the instrumental resolution .
numerous absorption lines arising from the earth s atmosphere occur in this region of the spectrum , and two lines in particular ( at 5889.637 and 5890.09 ) are very close to the rest wavelength of the d@xmath45 line .
@xcite discusses the origin and identification of these atmospheric ( telluric ) lines in the wavelength region . to reduce telluric line contamination
the stellar spectra were divided out by a purely atmospheric absorption template spectrum obtained from a fast rotating , unreddened early - type star with no interstellar sodium , and taken at similar air mass . for both observing
runs spectra of @xmath46leo and @xmath47lup were taken each night .
the equivalent width of the lines for these stars are estimated to be @xmath6 1 m and 1.4 m , respectively ( @xcite ) .
the stars sao251050 and sao224833 ( @xmath48lup ) , initially chosen as candidates for telluric line correction , showed the presence of interstellar lines .
since most of the telluric lines arise from water molecules , their strengths change throughout the night as the water content of the earth s atmosphere changes .
hence , the template spectra were properly scaled to match the strengths of the atmospheric lines in the object spectra .
the success of the atmospheric correction procedure is apparent from the fact that all the telluric lines , which are clear of the interstellar lines , have been satisfactorily removed ( see fig .
[ spectra ] ) .
finally the observed wavelengths were brought to the velocity scale and converted to the local standard of rest ( lsr ) velocity frame , assuming a solar motion of 20 towards @xmath46 = 18@xmath49 and @xmath50 = 30@xmath51 @xcite .
to access the zero point of the velocities the wavelengths of some of the numerous telluric lines gathered in the observed spectral range were measured , providing an external accuracy of the order of 4 m and showing that the lines match quite well , seemingly with no systematic offset .
.3 cm .3 cm .3 cm .3 cm .3 cm .3 cm .3 cm .3 cm .3 cm .3 cm .3 cm .3 cm .3 cm .3 cm
all observed absorption lines show appreciable velocity structure , and thus profile fitting provides the best method for determining accurate interstellar column densities as well as to discern and determine the properties of the individual interstellar clouds contributing to the line profile .
the observed profiles of the d@xmath45 and d@xmath52 absorption - lines are shown in fig .
[ spectra ] .
the dots represent the object spectra and the solid line gives the adopted multicomponent fit , as described below . the line - fitting program ( fits6p )
has been kindly provided to us by dr .
further details of the employed method can be found in @xcite and @xcite .
the zero point of the velocities in the upper and lower panels are with respect to the rest wavelength of the d@xmath45(5889.9510 ) and d@xmath52(5895.9242 ) lines , respectively . in each panel
the tick marks and numbers above the profiles identifies the different components used to fit the d@xmath45 and d@xmath52 lines .
voigt line profiles parameterized by their velocity dispersion parameter b , cloud component velocity and cloud column density @xmath53 , were convolved with a gaussian instrumental profile of 5.3 km s@xmath27 width and fitted to the observed data .
inspection of the calibration lamp spectra supports our assumption of a gaussian instrumental profile since no asymmetric tails are observed in the thorium lines .
both the d@xmath45 and d@xmath52 lines were fitted simultaneously , with the fewest components that seemed reasonable .
inflections in line shape , as a result of additional components in the wings of the stronger lines , often indicated whether another component should be added to the model .
the hyperfine structure splitting of @xmath54 1 of the individual d lines were not included in the adjustments since the resolution of the data is sufficiently broad to be affected by the splitting . in most cases
the assigned initial values of b were from 2.0 to 4.0 to all components , but larger values are sometimes required , e.g. sao239531 . also the three free parameters ( @xmath53 , , b ) for each component were allowed to vary in an iterative nonlinear least - squares fit to the observed profile .
some parameters occasionally were held fixed at `` reasonable '' values in order to facilitate convergence of the fits , but are thought to be as well determined as those allowed to vary .
the average of the d@xmath45 and d@xmath52 best fit values of these parameters are listed in table 2 .
a total of 125 entries are given .
each successive column gives the star s name , component number , lsr radial velocity , heliocentric radial velocity @xmath55 , logarithm of the column density , velocity dispersion b and the equivalent widths @xmath56 and @xmath57 , respectively .
the distribution of the absolute mean errors for , and b are shown in figs .
[ param]a - c .
for narrow , isolated lines of moderate strength such errors are typically 1%-10% for and few hundredths of a for b and . for stronger ,
wider and more blended lines the uncertainties in are typically 5%-20% , and 0.1 - 0.4 in b and .
for the strongest saturated lines , uncertainties in the actual component structure are likely to be the dominant error source .
the theoretical uncertainty in the placement of the continuum level is approximately equal to the rms scatter divided by the square root of the number of points used to define it @xcite . the statistical error on line flux values
are given by @xmath58 = @xmath59 , where @xmath60 is the number of observed data points in the absorption profile , @xmath61 is the resolution element in / pixel , and s / n is the signal - to - noise ratio of the exposure @xcite . in most cases
the continua are very well defined and statistical errors probably dominate the equivalent width errors . typically for our observations , @xmath62 , @xmath63 = 0.027 / pix and s / n @xmath64 150 , so that @xmath65 1 m . the star @xmath48lup has been previously observed at similar resolution ( 4 ) by @xcite . the parameters for ( , , b ) are ( @xmath66 , @xmath67 , @xmath68 ) , and those obtained in this work are ( @xmath69 , @xmath70 , @xmath71 ) .
the remarkable aggreement between the values is rather encouraging .
additional absorption features occur in some of the spectra of low @xmath72 stars shown in fig .
[ spectra ] .
the stellar lines are expected to be very weak throghout the class of b stars ( being @xmath6 2 m for stars earlier than b7 ) , but the cases where stellar lines have been identified were removed from the multicomponent fit .
lcrrccrrclcrrccrr star & c & & & & b & & & & star & c & & & & b & & & & & & km s@xmath27 & & & & & & & km s@xmath27 & 239327&1 & 0.5 & 10.0 & 12.3 & 3.2 & 162.9 & 122.4&&251928&1 & 3.5 & 11.4 & 11.5 & 3.5 & 48.5 & 26.2 & 2 & @xmath47.4 & 2.0 & 11.5 & 8.0 & 60.6 & 34.5 & & & 2 & @xmath43.3 & 4.6 & 12.6 & 2.1 & 139.5 & 118.9239370&1 & 0.6 & 9.8 & 13.0 & 2.5 & 188.2 & 168.9 & & & 3&@xmath411.5 & @xmath43.5 & 11.4 & 4.8 & 41.7 & 21.9 & 2 & @xmath43.2 & 6.0 & 11.9 & 7.1 & 137.1 & 76.7&&251942&1 & 1.2 & 8.9 & 12.6 & 2.4 & 157.9 & 134.6239423&1 & @xmath42.2 & 6.8 & 11.7 & 5.0 & 88.5 & 48.8 & & & 2 & @xmath46.2 & 1.5 & 11.7 & 2.7 & 68.3 & 39.8 & 2 & @xmath49.4 & @xmath40.4 & 11.2 & 1.4 & 27.0 & 14.8&&251979&1 & @xmath42.1 & 5.5 & 12.7 & 3.4 & 215.4 & 188.5239531&1 & @xmath41.3 & 7.5 & 11.8 & 2.7 & 89.5 & 52.3&&251987&1 & @xmath42.0 & 6.1 & 11.4 & 5.4 & 43.0 & 21.7 & 2 & @xmath45.9 & 3.0 & 11.8 & 9.5 & 108.5 & 56.9 & & & 2&@xmath417.0 & @xmath48.9 & 10.2 & 0.3 & 3.3 & 1.2239702&1 & @xmath40.5 & 7.9 & 12.6 & 2.4 & 154.5 & 132.0&&251988&1 & 1.4 & 9.9 & 13.9 & 1.4 & 150.1 & 138.3 & 2&@xmath411.8 & @xmath43.4 & 11.4 & 5.9 & 50.5 & 26.5 & & & 2 & @xmath44.5 & 3.9 & 11.5 & 3.3 & 56.2 & 30.5239779&1 & 1.0 & 9.2 & 12.5 & 2.2 & 133.9 & 111.0&&252114&1 & @xmath42.4 & 4.8 & 12.8 & 3.1 & 205.3 & 176.4 & 2 & @xmath47.7 & 0.5 & 11.6 & 3.2 & 65.8 & 37.1&&252136&1 & 7.7 & 15.0 & 10.9 & 0.2 & 9.3 & 6.2240041&1 & @xmath41.9 & 5.7 & 12.9 & 2.4 & 176.4 & 156.5 & & & 2 & @xmath43.1 & 4.3 & 12.3 & 3.1 & 152.2 & 115.4 & 2 & @xmath46.0 & 1.6 & 11.6 & 6.2 & 67.4 & 35.8 & & & 3&@xmath412.3 & @xmath45.0 & 11.1 & 5.8 & 23.8 & 11.5240265&1 & 0.5 & 7.5 & 12.5 & 1.9 & 121.4 & 105.2&&252245&1 & @xmath40.7 & 6.6 & 12.3 & 3.2 & 162.6 & 125.1 & 2 & @xmath46.3 & 0.7 & 11.0 & 1.5 & 19.3 & 10.4&&252262&1 & @xmath40.8 & 6.2 & 12.0 & 1.9 & 88.7 & 62.7 & 3&@xmath413.9 & @xmath47.0 & 11.2 & 6.0 & 29.1 & 14.9 & & & 2 & @xmath49.9 & @xmath42.8 & 10.7 & 4.7 & 10.3 & 5.2240368&1 & @xmath40.4 & 6.3 & 10.7 & 1.9 & 10.1 & 5.0&&252284&1 & @xmath44.9 & 1.0 & 11.1 & 4.7 & 26.1 & 12.3 & 2 & @xmath49.7 & @xmath43.1 & 10.9 & 4.3 & 16.1 & 8.4&&252294&1 & @xmath40.6 & 6.3 & 12.5 & 1.6 & 101.2 & 95.1240621&1 & 0.8 & 6.9 & 12.6 & 2.9 & 178.5 & 133.9 & & & 2 & @xmath45.3 & 1.6 & 11.9 & 4.0 & 113.8 & 65.9 & 2&@xmath411.8 & @xmath45.6 & 12.7 & 5.9 & 312.8 & 269.5&&252321&1 & @xmath40.9 & 6.3 & 12.8 & 2.7 & 182.2 & 156.9 & 3&@xmath423.3&@xmath417.2 & 12.6 & 5.3 & 274.4 & 225.2&&256745&1 & 3.1 & 14.0 & 12.4 & 1.6 & 103.5 & 88.4 & 4&@xmath432.2&@xmath426.0 & 12.3 & 2.3 & 134.3 & 102.4 & & & 2 & @xmath44.9 & 6.0 & 10.9 & 5.0 & 16.8 & 8.5 & 5&@xmath438.8&@xmath432.7 & 15.3 & 0.5 & 277.4 & 223.2&&256763&1 & 2.3 & 13.3 & 12.4 & 3.9 & 186.5 & 148.7 & 6&@xmath449.6&@xmath443.5 & 12.5 & 2.6 & 160.7 & 131.7&&256788&1 & 3.5 & 14.4 & 12.9 & 2.6 & 179.0 & 165.9 & 7&@xmath458.1&@xmath452.0 & 12.6 & 3.0 & 188.9 & 149.6 & & & 2 & @xmath44.9 & 6.0 & 11.2 & 2.7 & 29.8 & 15.1240645&1 & @xmath43.8 & 2.3 & 11.2 & 5.1 & 28.0 & 13.4&&256789&1 & 11.2 & 22.0 & 10.9 & 2.1 & 16.5 & 8.6251460&2 & @xmath42.9 & 7.1 & 13.0 & 1.8 & 144.6 & 130.6 & & & 2 & 3.4 & 14.2 & 12.7 & 2.0 & 138.7 & 122.0 & 3&@xmath410.4 & @xmath40.4 & 11.3 & 8.2 & 38.8 & 19.9 & & & 3 & @xmath43.2 & 7.7 & 11.2 & 6.5 & 27.3 & 13.9 & 1 & 3.3 & 13.3 & 11.2 & 8.6 & 27.6 & 14.0&&256798&1 & 3.8 & 14.5 & 12.8 & 2.0 & 141.3 & 131.1251530&1 & @xmath40.3 & 9.4 & 13.0 & 2.0 & 152.4 & 145.2&&256800&1 & 1.7 & 12.5 & 12.4 & 3.9 & 189.1 & 153.8 & 2 & @xmath45.6 & 4.1 & 11.7 & 7.3 & 86.4 & 47.8&&256828&1 & 0.8 & 11.3 & 12.7 & 3.0 & 191.6 & 171.6251545&1 & @xmath42.0 & 7.7 & 14.9 & 1.1 & 214.4 & 167.8&&256834&1 & 3.4 & 13.8 & 12.9 & 2.4 & 168.2 & 157.9 & 2&@xmath413.2 & @xmath43.5 & 10.1 & 0.2 & 2.8 & 0.8 & & & 2 & @xmath43.5 & 7.0 & 12.1 & 1.5 & 80.6 & 66.8251554&1 & 36.7 & 46.5 & 11.0 & 6.9 & 17.7 & 9.0&&256843&1 & 0.8 & 11.2 & 14.6 & 2.1 & 258.6 & 235.9 & 2 & @xmath41.6 & 8.2 & 12.9 & 2.7 & 190.6 & 174.7&&256849&1 & 1.6 & 11.8 & 14.2 & 1.6 & 182.4 & 156.5 & 3&@xmath411.8 & @xmath42.0 & 10.9 & 4.3 & 16.1 & 8.2&&256900&1 & 1.4 & 11.0 & 12.6 & 2.2 & 144.2 & 120.6251582&1 & @xmath42.6 & 7.0 & 12.7 & 2.3 & 161.1 & 140.5 & & & 2 & @xmath47.8 & 1.7 & 10.9 & 2.8 & 13.4 & 7.3251625&1 & @xmath41.8 & 7.6 & 12.9 & 2.6 & 188.4 & 166.8&&256924&1 & @xmath47.7 & 1.9 & 10.5 & 2.4 & 6.1 & 2.6 & 2&@xmath414.0 & @xmath44.5 & 11.2 & 4.8 & 29.9 & 15.4&&256953&1 & 2.1 & 11.4 & 13.1 & 1.8 & 145.3 & 131.9251629&1 & 11.8 & 21.0 & 10.9 & 0.3 & 9.1 & 4.9&&256955&1 & @xmath41.6 & 7.3 & 11.0 & 5.7 & 20.4 & 10.0 & 2 & @xmath41.8 & 7.4 & 12.5 & 3.7 & 213.2 & 162.9&&256980&1 & 9.8 & 18.2 & 11.1 & 6.5 & 27.6 & 9.0 & 3 & @xmath48.8 & 0.4 & 12.6 & 1.2 & 83.4 & 86.7 & & & 2 & 1.4 & 9.9 & 12.4 & 2.7 & 147.1 & 133.0 & 4&@xmath418.1 & @xmath48.9 & 11.6 & 3.2 & 59.1 & 34.0 & & & 3 & @xmath48.2 & 0.2 & 11.7 & 3.5 & 77.8 & 36.9251699&1 & @xmath42.4 & 7.0 & 11.2 & 5.4 & 29.2 & 15.0&&257002&1 & 17.8 & 26.1 & 11.4 & 0.8 & 31.6 & 16.9251717&1 & @xmath42.2 & 6.7 & 12.5 & 2.3 & 145.4 & 121.1 & & & 2 & 10.4 & 18.7 & 11.4 & 0.6 & 24.3 & 18.3251774&1 & @xmath42.4 & 6.2 & 12.3 & 3.4 & 162.5 & 122.4 & & & 3 & @xmath40.6 & 7.7 & 12.5 & 4.4 & 223.1 & 169.4 & 2&@xmath411.2 & @xmath42.6 & 11.1 & 4.0 & 20.1 & 14.4 & & & 4&@xmath414.2 & @xmath45.9 & 12.5 & 3.1 & 170.1 & 150.4251802&1 & @xmath45.4 & 4.2 & 12.7 & 3.5 & 212.2 & 178.6&&257032&1 & 2.4 & 9.9 & 12.6 & 3.3 & 188.3 & 162.5 & 2&@xmath416.1 & @xmath46.5 & 12.2 & 2.9 & 128.7 & 98.6&&257033&1 & 34.3 & 42.0 & 10.7 & 3.5 & 9.8 & 5.0 & 3&@xmath427.2&@xmath417.6 & 11.2 & 2.1 & 26.0 & 15.0 & & & 2 & 0.2 & 7.9 & 13.1 & 1.9 & 155.7 & 141.1 & 4&@xmath439.4&@xmath429.8 & 11.6 & 2.1 & 56.5 & 35.8 & & & 3&@xmath411.4 & @xmath43.7 & 10.4 & 0.3 & 4.2 & 2.3 & 5&@xmath451.4&@xmath441.8 & 10.4 & 0.2 & 2.6 & 3.1 & & & 4&@xmath431.3&@xmath423.6 & 10.8 & 5.1 & 12.4 & 6.3251837&1 & @xmath41.7 & 6.5 & 12.4 & 2.7 & 149.8 & 117.7&&257047&1 & 1.0 & 9.0 & 11.4 & 2.9 & 46.3 & 25.0 & 2&@xmath410.2 & @xmath42.0 & 11.5 & 4.3 & 56.7 & 28.5 & & & 2 & @xmath48.7 & @xmath40.6 & 10.6 & 5.2 & 9.0 & 2.8251841&1 & 2.4 & 10.9 & 11.0 & 3.1 & 18.2 & 8.1&&257068&1 & @xmath40.1 & 7.4 & 12.6 & 3.6 & 211.5 & 183.2 & 2 & @xmath46.0 & 2.5 & 11.1 & 3.4 & 26.0 & 12.8&&257069&1 & 1.0 & 9.0 & 12.3 & 1.9 & 116.3 & 90.6251874&1 & @xmath41.6 & 7.1 & 12.6 & 5.3 & 281.9 & 232.4&&257142&1 & 1.6 & 9.7 & 12.0 & 2.3 & 100.8 & 72.0251894&1 & 4.4 & 13.2 & 12.3 & 1.8 & 103.7 & 86.6&&258697&1 & @xmath40.3 & 8.4 & 12.3 & 2.7 & 132.7 & 114.5 & 2 & @xmath42.0 & 6.8 & 12.5 & 1.4 & 98.7 & 74.7&&377063&1 & 4.2 & 15.8 & 11.5 & 2.8 & 50.3 & 28.3 & 3 & @xmath49.3 & @xmath40.5 & 10.9 & 2.6 & 14.1 & 7.6 & & 251903&1 & @xmath41.0 & 7.2 & 11.1 & 5.5 & 23.3 & 10.2 & & & 2&@xmath411.6 & @xmath43.5 & 10.9 & 1.2 & 14.3 & 7.2 & & histograms of the individual distributions of , and b are given in figs .
[ param]d to [ param]f . inspection of fig .
[ param]d shows a clear division of the column densities around @xmath5 12.0 2 .
the average column density of the sub - samples is 11.2 2 and 12.7 2 , respectively , both with small dispersion of @xmath15 0.6 2 .
figure [ param]e shows that the distribution of the lsr velocities , although centered around 0 , is skewed to the negative side .
the velocity dispersion parameter b shows a concentration around 2.5 and barely suggests another around 5.5 .
these trends are somewhat confirmed by figs .
[ param]g - i , that shows the plots of b @xmath14 , @xmath14 and @xmath14 b , respectively . for the sake of clarity
few points of very high negative and positive lsr velocities have been excluded from these plots .
as one can see the lower column densities span the whole velocity dispersion range , and are concentrated mostly on the negative velocity range . on the other hand ,
the higher column densities are concentrated around b @xmath5 3 , and have velocities spread by @xmath15 3 around zero in the lsr frame . the apparent lack of points with @xmath6 10.5 2 for all b may be due to our detection limits
. the lack of points with b @xmath6 1.5 for @xmath64 11.5 2 may be due to our inability to resolve closely blended components . to investigate the distribution and kinematics of the interstellar gas components along the line - of - sight plots of , and b as a function of the stellar distance are given in fig .
[ dis_nbv ] .
the _ left _ column shows those stars with line - of - sight _ inside _ the clouds contours ; the _ middle _ column those in their _ outskirts _ ; and the _ right _ column those towards the _
sheet_. the plotting simbols indicate a further division of the components according to the line profile characteristics . the line profiles with a _ single weak _ feature are represented by the open triangles , while those profiles with a _
single strong _ feature are represented by the solid triangles .
the profiles with _ two weak _ features are represented by the plusses and asterisks , which indicate the features with velocity similar to the single weak and single strong components , respectively .
the profiles with _ two or more _ features are represented by the squares and circles .
the open and solid squares indicate those features similar to the single weak and to the single strong components .
the open circles indicate the other components , i.e. , the third , fourth and so on .
the structure of the column density @xmath14 distance diagrams for the stars _ inside _ the clouds contours , in their _ outskirts _ and towards the _ sheet _ is fairly similar . _
single weak _
components are seen for the stars closer than 60 - 80 pc from sun ( open triangles and plusses ) , and for most of the more distant stars along the distance range ( open squares ) . on the other hand ,
the _ single strong _
component is seen only for stars more distant than 120 - 150 pc from the sun ( solid triangles , asterisks and solid squares ) .
the open circles tend to appear at distances greater than 300350 pc , but without clear correlation with the column densities or the radial velocities .
the velocities of the components identified along the line of sight also have a similar structure for the stars _ inside _ the clouds contours , in their _ outskirts _ and towards the _ sheet_. the _ single weak _ component ( open triangles , open squares and plusses ) has velocity more negative than @xmath45 , while the _ single strong _ component ( solid triangle , solid square and asterisks ) is centered at 0 in the lsr frame .
the fact that the plusses and asterisks show kinematic structure consistent with those of the _ single weak _ and the _ single strong _ components , respectively , suggests that the material have some inhomogeneities .
the open circles appear in the whole range of radial velocities .
the velocity dispersions of the components also have a similar structure for the stars _ inside _ the clouds contours , in their _ outskirts _ and towards the _ sheet_. the _ single weak _ component has b @xmath5 5.5 . on the other side the _ single strong _ component has b @xmath5 2.5 .
the b value is a measure of both thermal and turbulent velocities in an interstellar cloud .
we have b @xmath73 , where @xmath74 is the rms turbulent velocity , @xmath75 is the boltzmann s constant , @xmath76 is the kinetic temperature and @xmath77 is the mass of the atom .
there is no independent way to determine the degree of turbulence within the material , but some qualitative considerations can be done .
the typical kinetic temperature of interstellar diffuse clouds is generally taken to be about 80 115 k , from observations of rotational excitation of h@xmath45 @xcite . for this corresponds to 0.24 @xmath6 b @xmath6 0.29 . in this case
the upper limits to @xmath74 for the _ weak _ and _ strong _ components would be 1.8 and 3.9 . assuming a sound speed for diffuse interstellar clouds about 0.7 @xcite the observed components are either significantly hotter than a standard diffuse cloud or are subject to supersonic turbulent motions . to look for a possible association of the interstellar gas components with the observed dust features in this direction , , and b as a function of the @xmath9 colour excess have been plotted in fig .
[ eby_nbv ] .
the columns and symbols have the same meaning as for fig . [ dis_nbv ] .
it can be seen that the _ strong _ component is only picked up by the reddened stars , i.e. , those with @xmath78 , whereas the _ weak _ component is picked up by both reddened and unreddened stars , suggestting that dust and gas apparently have the same distribution along the line of sight .
the average column density of the _ weak _ component ( @xmath54 11.2 2 ) , would correspond to @xmath79 @xmath80 2 .
it is worth noticing at this point that , in fig .
6a of paper i , around 60 - 70 pc the dispersion of the @xmath9 increases , with the lower limit becoming slightly shifted to positive values .
this indicates that the spectroscopy has provided a finer tuning of the existing low column density components , more difficult to disentagle based only on the colour excess values . on the other hand
, it shows that the observed increase in the dispersion of @xmath9 is real , and not due to observational errors , attesting the high quality of the photometric data and method .
there appears to be a common envelope of minimum and maximum column density , centered in a narrow distance slot around 120 - 150 pc , that remain unchanged up to 350 - 400 pc , suggestting that a second volume of low density may have been reached .
the complete range of column densities from @xmath54 10.7 to 11.7 , is also shown by the stars centered around 60 pc , suggesting a sheet - like structure .
interestingly , the minimum and maximum column densities of the _ strong _ component ( = 12.3 and 13.2 2 ) correspond to @xmath81 and 0300 , respectively , if it is assumed that the standard relationship between and log @xmath79 @xcite and @xmath9 and log @xmath79 @xcite are valid .
this may suggest that the high column densities may be the gas related to the dust sheet - like structure , reported in paper i. since the coalsack , musca and chamaeleon are also located around 150 @xmath15 30 pc they seem also to be part of the gas composing the sheet . assuming that the observed stars are within different dark clouds the column densities , radial velocities and velocity dispersions , shown in figs .
[ spectra ] to [ dis_nbv ] , could be either the effect of a common structure containing the embedded molecular clouds or just an effect of different superposing structures . in order to investigate how
the parameters ( , , b ) are distributed over the studied area we have used the eight sub - areas delineated by the horizontal and vertical dashed lines and identified by the roman numbers in fig .
[ obpos ] .
figures [ lgal_nbv]a and b show the distribution of ( _ left _ ) , ( _ middle _ ) and b ( _ right _ ) against the galactic longitude and latitude , respectively , for the eight sub - areas .
the symbols have the same meaning as for fig . [ dis_nbv ] , and
the vertical and horizontal dashed lines indicate the division between the sub - areas .
the _ strong _ and _ weak _ components are present and well separated everywhere .
the gas components form two extended interstellar features , that cover the whole connecting area .
note , however , that the radial velocity of the _ strong _ component seems to increase from @xmath82 near the galactic plane to + 3 near the southern parts .
this trend would be expected from an expanding bubble with the center around @xmath83 , @xmath84 .
this fact is quite interesting given the fact that high thermal pressure has been measured for the hot gas phase for a nearby direction ( @xmath85 , @xmath86 ) from a combination of rosat 0.25 kev and optical data ( knude et al .
in preparation ) .
the analysis of the identified components indicates that the interstellar gas is distributed in two extended sheet - like structures permeatting the whole area , one at @xmath1 60 pc and another around 150 pc from the sun .
the nearby feature is approaching the sun with average radial velocity of @xmath47 , has low average column density ( @xmath5 11.2 2 ) and is either hotter or more turbulent than the _ strong _ component , with velocity dispersion b @xmath5 5 .
the more distant feature has column densities between 12.3 @xmath6 @xmath6 13.2 , average velocity dispersion b @xmath5 2.5 and seems associated to the dust sheet observed towards the coalsack , musca and chamaeleon direction .
its velocity is centered around 0 , but there is a trend for increasing from @xmath43 near @xmath7 to 3 near @xmath8 .
there seems to be a common envelope of minimum and maximum column density , centered in a narrow distance slot around 120 150 pc , that remains unchanged up to 350 pc , suggestting that a volume of low density may have been reached .
after that , a third group of high negative and positive velocity components seemingly arise around 300 - 350 pc from the sun , however , without indications of an extended structure .
details of the overall distribution of the ism ( @xmath1 500 pc ) along the studied direction have been extensively discussed in paper i. the kinematics of the nearby gas has been reviewed by @xcite .
details of the velocity structure in the region ( 360@xmath87 295 and 0@xmath88 30 ) , encompassing most of the sco - cen association , have been discussed by @xcite . in the immediate solar neighbourhood
the observational data suggest that the sun is located within but close to the edge of the local interstellar cloud ( lic ) .
first identified from its kinematics by @xcite , the lic was modelled by @xcite as roughly spherical with dimensions of 58pc , warm ( t @xmath89k ) , low density ( @xmath90 0.1 @xmath91 ) and mostly neutral ( @xmath92 ) .
its largest column density is about @xmath93@xmath94 towards @xmath95 , @xmath96 . moving at a heliocentric velocity of @xmath97 towards @xmath98 , @xmath99 ( cf .
@xcite ) , the lic flows through the solar system and resonantly scatters solar hi ( lyman @xmath46 ) and hei ( @xmath100584 ) radiation .
this flow , is referred to as the local interstellar wind ( lisw ) .
the lic and other warm clouds with similar temperatures but a wide range of metal depletions , are embedded in an irregularly shaped region , whose radius ranges from 30 to 300 pc , and is deficient in dense neutral hydrogen compared to the galactic average ( e.g. @xcite ) . usually called local bubble , this cavity is thought to contain hot ( t @xmath5 10@xmath101 k ) , low - density ( @xmath102 @xmath6 0.025 @xmath91 ) gas and as recently shown by the rosat and euve shadowing experiments , coexists with the neutral atomic and molecular gas within its interior ( e.g. @xcite ) .
it is , however , worth noticing that @xcite has recently shown that heliospheric x - ray emission can account for about 25%-50% of the observed soft x - ray background intensities , which may put in doubt if the local bubble really contains the million degree hot gas .
optical interstellar absorption line studies within 100pc of the sun show evidence of a flow of material analogous to the lisw . assuming that the sun is embedded within a single coherently moving interstellar wind , from a direction ( @xmath103,@xmath104 ) with velocity @xmath105
, @xcite obtained the following vector ( @xmath103 , @xmath104 , @xmath105 ) = ( 345 , @xmath410 , @xmath415 ) . in a similar analysis @xcite obtained ( 354 , + 3 , @xmath412 ) and @xcite obtained ( 301 , + 59 , @xmath49 ) .
a possible description for this flow assumes that the lism is material which belongs to the surface of an expanding shell centered on the sco - cen association .
@xcite has obtained an expansion velocity of 7 9 assuming the center at ( @xmath106 , @xmath107 ) , the distance to the sun of 140 pc and the radius of 110 pc , as suggested by @xcite .
another possible explanation for the existence of the neutral clouds and their predominant flow away from the sco - cen association is a consequence of the interaction between the local and loopi bubbles ( see sect.[lbli ] ) and subsequent local fragmentation of the interaction zone @xcite .
since the closest star in our sample is at 65 pc the lic can not be addressed properly , neither the distance of the lower column density component . however , according to ( * ? ? ?
* and references therein ) the column densities of the nearby feature observed in this work should not be much closer than 50 pc from the sun . for the latitude / longitude region observed in our work the velocities predicted by the interstellar wind flow vector and the expanding shell models give essentially the same results ( * ? ? ?
3 ) . in any case
the velocity of the low column density component is consistent with the previous results , and indicate an outflow from the sco - cen association .
a more recent discussion by @xcite also using line data and _ hipparcos _ distances proposes components at three distinct heliocentric velocities in the puppis - vela region .
their component a with the velocity in the range from + 6 to + 9 covers part of the region presently discussed .
@xmath94 , while open circles represent components for which @xmath108@xmath94 .
note the well defined velocity distribution for the _ strong _ component and that it appears for stars beyond @xmath109pc . ] from the distances of table 1 and heliocentric velocities from table 2 we construct fig .
[ vel_dist ] of v@xmath110 vs. distance and notice that @xcite s component a is seen in two stars closer than 100pc and in fact seems present in all stars all the way to 450pc , this implies that component a covers all of our surveyed region , indeed a sheet nearer than about 100pc .
two stars marginally have absorption features at component b s lower velocity limit ( 12 ) , these stars are at 100pc and 130pc respectively .
otherwise component b is seen in most stars beyond 180pc ; the upper distance limit is very sharply defined by component b s upper limit at 15 .
no stars beyond @xmath5350pc shows any indication of component b. component c ( 21 23 ) is virtually absent from our region and is only observed for galactic longitude less than 275 in @xcite s survey , far from our region .
four stars closer than 100pc show absorption in a narrow interval from + 1 to + 4 , this interval is hardly represented between 100 and 400pc from where it is again present in almost any star .
a fifth group of absorption has velocities between @xmath42 and @xmath44 for three stars between 120 and 140pc , lines in this interval disappear for larger distances but reappear between 300 and 400pc .
locally , that is for distances less than 150pc , we confirm three interstellar kinematic groups with absorption lines in the intervals : [ @xmath45,0 ] , [ 0,5 ] and [ 5,10 ] the second of which represents the volume within 100pc whereas the other two both are present between 120 and 150pc .
the local ism thus seems more complicated than proposed in @xcite investigation .
one could speculate whether the interval [ 0,5 ] is representing the local undisturbed material and [ 5,10 ] and [ @xmath45,0 ] the expansion of the local and the sco cen bubbles respectively ?
[ vel_dist ] also indicates that the distribution of heliocentric velocities for the diffuse components changes character between 300 and 350pc .
closer than 300pc there is a set of velocities exceeding 10 , this set disappears beyond 350pc . on the other hand a new set with heliocentric velocities smaller than 0 appears beyond 300pc .
could this kinematic transition at 300350pc be connected to the expansive movement of the backside of the lower centaurus crux feature being countered by an opposite motion ? the expansion has ceased at about 350pc and the counter motion halting the expansion is first noticed at about 300pc . a soft x - ray shadow has recently been discovered on the edges of the loop i bubble , cast by a warped annular volume of dense neutral matter that supposedly formed during its collision with the local bubble @xcite .
the shadow counterpart , a huge hi ring was also identified on the @xcite data .
@xcite suggested that the steep increase of the column density caused by the annular feature , from less than 10@xmath111 to @xmath64 7 @xmath112 , occurs at a distance of @xmath5 70 pc from the sun .
this distance is thought to be supported by results from optical and uv spectral analysis of stars near the center of loop i ( @xmath113 and @xmath114 ) by @xcite , who found the presence of a neutral gas wall of @xmath115 2 at a distance of 40 @xmath15 25 pc .
see also the determination of the radial extent of the local hot bubble recently estimated by @xcite based on the _ rosat _ and _ iras _ all - sky survey data .
however , for the hydrogen column density of the annular feature , the colour excess and the column densities obtained in this work suggest that the interaction zone between the two bubbles is located around 120 - 150 pc from the sun .
this result is corroborated by @xcite an improved version of the 3d mapping of the local bubble cavity has been recently obtained by @xcite where it is found that for the region investigated here the 2050m isocontours at @xmath116 occurs at @xmath117pc , but at @xmath118 the distance is close to 70pc . the hydrogen column densities and velocities observed by @xcite are more consistent with our lower column density component at @xmath1 60 pc . in the same strip of latitude ,
the wall of neutral gas has a column density @xmath119 2 , as suggested by @xcite . moreover ,
the hydrogen column densities towards the interaction ring , used by ( * ? ? ?
their fig .
4 ) suggest a value around 3 @xmath43 10@xmath120 2 at 70 pc , also more consistent with our nearby feature .
since @xcite have observed only unreddened directions our results cast some doubt on the distance to the interaction zone being located around 70 pc .
it seems that the annular region is twisted and folded , with different directions having different distances .
in addition to the velocities indicative of the flow of material in the last section , some of the more distant stars also have components at nearly 0 in the lsr @xcite . in the longitude range @xmath121 there are a number of low velocity ( @xmath122 @xmath123 ) components with generally strong lines , consistent with the proximity of the @xmath24 oph and lupus clouds @xcite . in the longitude range @xmath124
there are also several components , in the positive latitudes , with lsr velocities which are small but clearly negative ( @xmath122 @xmath125 ) .
as mentioned in sect.[kinem ] , the feature around 120 150 pc observed in our work , has column densities suggestting that the gas is associated to the dust sheet observed towards the coalsack , musca and chamaeleon direction .
it is certainly not appropriate to make generalizations exclusively from our data , but the fact that the velocities observed here are also in the same range ( @xmath126 @xmath127 ) , suggests that the dust and gas feature around 120 150 pc , seems to be part of an extended large scale feature of similar kinematic properties , supposedly identified with the interface of the local and loop i bubbles .
the investigation of the interstellar gas components towards the southern coalsack , chamaeleon and musca dark have produced the following results : * the interstellar gas is distributed in two extended sheet - like structures permeatting the whole searched area , one at @xmath1 60 pc and another around 120 150 pc from the sun .
* the nearby feature is approaching to the sun with average radial velocity of @xmath128 , has low average column density @xmath5 11.2 2 and has velocity dispersion b @xmath5 5 .
the more distant feature has column densities between 12.3 @xmath6 @xmath6 13.2 , average velocity dispersion b @xmath5 2.5 and seems associated to the dust sheet observed towards the coalsack , musca and chamaeleon direction .
its velocity is centered around 0 , but there is a trend for increasing from @xmath43 near @xmath7 to 3 near @xmath8 . * in aggreement with several independent investigations the nearby low column density feature indicates a general outflow from the sco - cen association , while the dust and gas feature around 120 150 pc seem to be part of an extended large scale feature of similar kinematic properties , supposedly identified with the interaction zone of the local and loop i bubbles .
assuming that the interface and the ring - like volume of dense neutral matter have similar properties , our results suggest that the interaction zone between the bubbles is twisted and folded .
this paper is based on observations collected at the european southern observatory ( eso , la silla , chile ) .
the staff at eso , both in la silla and garching are thanked for the assistence during the observing runs .
welty is thanked for the line - fitting program and dr .
egger is thanked for the ring - like feature contours .
the brazilian agencies cnpq and fapemig , and the danish nbifafg are acknowledged for supporting this research .
w. corradi wishes to express his gratitude to the nbifafg members for the invaluable help during the development of this work in denmark .
bochkarev n.g . , 1987 , ap&ss , 138 , 229 bowyer s. , lieu r. , sidher s.d . , et al . , 1995 , nat , 375 , 212 breitschwerdt d. , freyberg m.j . , egger r. , 2000 , a&a , 361 , 303 bruhweiler f.c . , 1996 in : bowyer s. , malina r.f .
iau colloq .
152 , astrophysics in the extreme ultraviolet .
kluwer , dordrecht , p. 261
centurion m. , vladilo g. , 1991 , apj , 372 , 494 cha a.n .
, sahu m.s .
, warren moos h , blaauw , a. , 2000 , apjs , 129 , 281 corradi w.j.b . ,
franco g.a.p . , 1995 , a&as , 112 , 95 corradi w.j.b . , franco g.a.p . , knude j. , 1997 , a&a , 326 , 1215 ( paper i ) cox d.p . , reynolds r.j . , 1987 , ara&a , 25 , 303 cowie l.l .
, songaila a , 1986 , araa , 24 , 499 cravens t.e . , 2000 , apj , 532 , 153 crawford i.a .
, 1991 , a&a , 247 , 183 crutcher r.m . , 1982 , a&a , 254 , 82 dame , t.m . , ungerechts , h. , cohen , r.s . , de geus , e.j . , , 1987 , apj , 322 , 706 de geus e.j . , 1992 ,
a&a , 262 , 258 dickey j.m . ,
lockman f.j . , 1990 , ara&a , 28 , 215 dodorico s. , ghigo m. , ponz d. , 1987 in : an atlas of the thorium argonium spectrum for the eso echelle spectrograph in the @xmath129 region , european southern observatory egger r.j . , aschenbach b. , 1995 , a&a , 294 , l25 esa , 1997 , the hipparcos and ticho catalogues , esa sp-1200 .
esa publications division , noordwijk feitzinger j.v .
, stwe j.a
. , 1984 , a&as , 58 , 365 ( erratum : 63 , 203 ) ferlet r. , vidal - madjar a. , gry c. , 1985 , apj , 298 , 838 franco g.a.p . , 1989 , a&a , 215 , 119 franco g.a.p . , 1991 , a&a , 251 , 581 frisch p.c . , york d.g . , 1983 , apj , 271 , l59 frisch p.c . , york d.g
, 1986 in : smoluchowski r. , bahcall j.n .
, matthews m.s .
the galaxy and the solar system .
university of arizona press , tucson , p. 83 gehrels n. , chen w. , 1993 , nat , 361 , 706 gnova r. , beckman j.e . , bowyer s. , spicer t. , 1997 , apj , 484 , 761 hartquist , t.w .
, 1994 , ap&ss , 216 , 185 hobbs l.m . , 1978 , apj , 222 , 491 howarth i.d . ,
phillips a.p . , 1986 ,
mnras , 222 , 809 iwan d. , 1980 , apj , 169 , 25 kaper l. , pasquini l. , b. , 1996 , in :
cat / ces operating manual " kerp j. , herbstmeier u. , mebold u. , 1993 , a&a , 268 , l21 knude j. , 1978 , in : reiz a. , andersen j. ( eds . ) astronomical papers dedicated to b. strmgren .
copenhagen univ .
, copenhagen , p. 273 lallement r. , 1998 , lectures notes in physics , 506 , 19 lallement r. , bertin p. , 1992
, a&a , 266 , 479 lallement r. , ferlet r. , lagrange a.m. , lemoine m. , vidal - madjar a. , 1995 , a&a 304 , 461 lallement r. , vidal - madjar a. , ferlet r. , 1986 , a&a , 168 , 225 lallement r. , welsh b.y . , vergely j.l . , crifo f. , sfeir d. , 2003 , a&a in press mihalas , d. , binney j. , 1981 , in galactic astronomy " 2nd edition ( san francisco , freeman ) nyman l .- . , bronfman l. , thaddeus p. , 1989
, a&a , 216 , 185 paresce f. , 1984 , aj , 89 , 1022 redfield s. , linsky j.l . , 2000 , apj , 534 , 825 sembach k.r . , danks a.c . ,
savage b.d . , 1993 ,
a&as , 100 , 107 snowden s.l . , cox d.p .
, mccammon d. , sanders w.t . , 1990 ,
apj , 354 , 211 snowden s.l . ,
mebold u. , hirth w. , herbstmeier u. , schmidt j.h.m.m . , 1991 , sci , 252 , 1529 snowden s.l . , freyberg m.j . ,
plucinsky p.p . ,
et al . , 1995 , apj , 454 , 643 snowden s.l .
, egger r. , finkbeiner d.p .
, freyberg m.j . ,
plucinsky p.p . , 1998 , apj , 493 , 715 savage b.d .
, bohlin r.c . , drake j.f . , et al . , 1977 ,
apj , 216 , 291 spitzer , jr .
, l. : 1978 , `` physical processes in the interstellar medium '' ( new york : john wiley & sons ) sfeir d.m . , lallement r. , crifo f. , welsh b.y . , 1999 , a&a , 346 , 785 tinbergen j. , 1982 , a&a , 105 , 53 wang q.d .
, yu k.c . , 1995 ,
aj , 109 , 698 warwick r.s . ,
barber c.r . ,
hodgkin s.t .
, pye j.p . , 1993 , mnras , 262 , 289 weaver h. , 1979 in : burton w.b . ( ed . )
84 , large - scale characteristics of the galaxy .
reidel , dordrecht , p. 295
welsh b.y . ,
vedder p.w .
, vallerga j.v . , 1990 , apj , 358 , 473 welsh b.y .
, vedder p.w .
, vallerga j.v . , craig n. , 1991 , apj , 381 , 462 welsh b.y . , craig n. , vedder p.w . , vallerga j.v . , 1994 ,
apj , 437 , 638 welty d.e .
, hobbs l.m . , kulkarni v.p . , 1994 ,
apj , 436 , 152 | the results of a spectroscopic programme aiming to investigate the kinematics of the local interstellar medium components towards the southern coalsack and chamaeleon - musca dark clouds are presented .
the analysis is based upon high - resolution ( @xmath0 ) spectra of the insterstellar nai d absorption lines towards 63 b - type stars ( @xmath1 500 pc ) selected to cover these clouds and the connecting area defined by the galactic coordinates : @xmath2 and @xmath3 .
the radial velocities , column densities , velocity dispersions , colour excess and photometric distances to the stars are used to understand the kinematics and distribution of the interstellar cloud components .
the analysis indicates that the interstellar gas is distributed in two extended sheet - like structures permeating the whole area , one at @xmath1 60 pc and another around 120 - 150 pc from the sun .
the nearby feature is approaching to the local standard of rest with average radial velocity of @xmath47 , has low average column density @xmath5 11.2 2 and velocity dispersion b @xmath5 5 .
the more distant feature has column densities between 12.3 @xmath6 @xmath6 13.2 , average velocity dispersion b @xmath5 2.5 and seems associated to the dust sheet observed towards the coalsack , musca and chamaeleon direction .
its velocity is centered around 0 , but there is a trend for increasing from @xmath43 near @xmath7 to 3 near @xmath8 .
the nearby low column density feature indicates a general outflow from the sco - cen association , in aggreement with several independent lines of data in the general searched direction .
the dust and gas feature around 120 150 pc seem to be part of an extended large scale feature of similar kinematic properties , supposedly identified with the interaction zone of the local and loop i bubbles .
assuming that the interface and the ring - like volume of dense neutral matter that would have been formed during the collision of the two bubbles have similar properties , our results rather suggest that the interaction zone between the bubbles is twisted and folded .
[ firstpage ] stars : distances - ism : clouds - ism : individual objects : southern coalsack - chamaeleon - musca - loop i - local bubble |
we make a more detailed study , both analytical and numerical , of some _ exact relativistic _
solutions we recently proposed to describe static , axially symmetric massive core - shell systems ( vieira and letelier 1996a , 1997 , hereafter vl1 and vl2 , respectively ) .
we add in this way to the efforts of modeling many situations of interest in astrophysics involving massive shell - like structures around centers , as for example black holes or neutron stars surrounded by massive shell and ring remnants .
a nice illustration of this possibility is offered by the famous supernova 1987a plus its physical rings , see e. g. panagia et al .
( 1996 ) , meyer ( 1997 ) and chevalier ( 1997 ) . at least in a seminal sense
, the model could also describe galaxies since there are many of them exhibiting massive rings ( sackett and sparke 1990 , arnaboldi et al .
1993 , reshetnikov and sotnikova 1997 ) and shells ( malin and carter 1983 , quinn 1984 , dupraz and combes 1987 , barnes and hernquist 1992 ) , while many others possibly have galactic nuclei hiding black holes ( see kormendy and richstone 1995 for a review ) .
specifically , we implement here a monopolar core ( a black hole in the relativistic case ) plus an exterior shell of dipoles , quadrupoles and octopoles .
obviously , these multipoles are shell - like legendre expansions i. e. their corresponding terms increase with the distance in the intermediate vacuum between the core and the shell . beyond its own applicability
, the model also points to the possibility of a realistic description of any axially symmetric relativistic core - shell configuration whose approximation as a static system should be valid at some useful time scale , via a systematic multipolar legendre expansion .
in fact , a further step in this program was lastly started in letelier and vieira ( 1997 ) by giving stationarity ( rotation ) to the relativistic case of a monopolar core plus a purely dipolar shell .
this is an important improvement since real celestial objects do rotate .
the present analysis will be extended to rotating core - shell models in a forthcoming contribution . additionally , we consider the newtonian counterpart of the relativistic model above , which was only sketched in the previous works , and show that they could describe core - shell systems `` per se '' interesting in astronomy .
so , we can have massive circumstellar dust shells around certain types of stars as byproducts of their death , see e. g. barlow et al .
( 1994 ) for an example of ( irregular ) circumstelar shells around luminous blue variables and a review by groenewegen et al .
( 1998 ) about dust shells around carbon mira variables .
another potential application for core - shell models is to the more speculative possibility of hollow galactic halos of dark matter made of neutrinos recently considered by ralston and smith ( 1991 ) , madsen ( 1991 ) and barnes ( 1993 ) .
moreover , we will see that the study of the relativistic core - shell model in parallel to its newtonian counterpart is useful also to clarify the physical content of the former one .
after poincar ( 1957 ) and the kam theory ( after kolmogorov 1954 , arnold 1963a , 1963b and moser 1967 ) it became well established that non - integrability and hence chaos is a general rather than exceptional manifestation in the context of dynamical systems ( modeling or not physical situations ) ( see berry 1977 ) . given this ubiquitous fact , an important issue in astronomical modeling is to study in which extent in phase space chaoticity rises in models that are relevant to describe real systems and what are its consequences .
for example , binney ( 1982a ) discusses the difficulties of constructing stationary self - consistent models when a significant fraction of orbits are irregular since they may not obey neither vlasov s nor jean s equations .
then , it is remarkable that a wide class of fully integrable potentials , the so called stckel potentials ( lynden - bell 1962 , de zeeuw 1985 and de zeeuw et al .
1986 ) , are feasible starting points to describe , by themselves or by adding perturbations , real disc , elliptical or even triaxial galaxies . for more realistic tridimensional models
there are evidences for rounder and smoother mass distributions generating only relatively small fractions of chaotic orbits ( schwarzschild 1979 , binney and tremaine 1987 and evans et al .
1997 ) , while flattening and/or sharpening of the mass distributions , for example through increasing triaxiality and/or putting cusps and central masses to mimic black holes , tends to increase the chaoticity and force us to take it into account ( gerhard and binney 1985 , schwarzschild 1993 , merrit and fridman 1996 , norman et al .
1996 , merrit 1997 and valluri and merrit 1998 ) . on the other hand ,
the emergence of chaos in two - dimensional models has more loose correlations with morphological aspects , see for example binney and spergel ( 1982 ) , richstone ( 1982 ) , binney ( 1982b ) , gerhard ( 1985 ) and sridhar and touma ( 1997 ) .
we address in this work also a numerical study about the chaoticity of orbits trapped in the bound gravitational zones between the core and the external massive shell . beyond its own applicability
as seen above , we take advantage of models that are generic within the class of axially symmetric core - shell distributions and moreover offer exact relativistic and newtonian counterparts to be compared , in order to achieve some understanding about chaoticity related to two aspects : firstly , to the role of reflection symmetries shared by almost all models in astronomy and astrophysics , and secondly to the consequences of treating exact central black holes as just newtonian central masses . in the first part of this article ( section [ analitico ] ) we present analytical results concerning some properties like unicity and analyticity of the relativistic solutions themselves , in connection with both the linearized core - shell solutions and the corresponding newtonian models taken as limiting cases . in section [ sol ]
we enlarge the relativistic solution presented in vl1 for a monopolar core plus a shell of quadrupoles and octopoles in order to include a dipolar shell component .
this complete solution was only outlined in vl2 . in this later reference
it was emphasized that the case of a core plus purely dipolar shell is physically non - trivial already in the newtonian gravity , moreover , it was shown that the newtonian dipolar case is integrable , in contrast with its chaotic relativistic counterpart , which allowed us to characterize the chaotic behavior in this case as an intrinsic general relativistic effect .
here , we take advantage of this striking difference of the dipolar case in order to reinforce the numerical conclusions presented in section [ caos ] concerning the non - trivial dynamical role played by the presence / absence of the _ discrete _ reflection symmetry about the equatorial plane on the chaoticity of the orbits . in section [ reg ]
we proceed to a full elimination of axial ( conical ) singularities ( hereafter css ) from the relativistic model , which was only partially accomplished in vl1 .
in fact , css are not globally removable from neither static nor stationary many - body relativistic solutions since they are self - consistently demanded by einstein s equations to strut the otherwise unstable configuration against its own gravity ( see robertson and noonan 1968 and letelier and oliveira 1998 for a recent review ) .
so , in the present case the most we can do is really move all them outside the shell , thus obtaining a true vacuum in the intermediate space between the core and the shell .
we also discuss in this section the conditions for the kruskal - type analyticity of the solutions at the horizon of the central black hole and use both aspects to find the conditions under which the unicity of the models is assured . in section [ rw ]
we present the linearization of the relativistic model in the multipole strengths via the so called regge - wheeler ( rw ) formalism .
in particular , we see that our model exemplifies the fact that css survive to the linearization process , their presence in the intermediate vacuum being in fact an obstruction to the application of the rw formalism .
this does reinforce the interpretation of css as a kind of singular matter distribution necessary to the dynamical consistency of static and stationary relativistic models . however being intrinsecally relativistic manifestations , an adiabatic treatment of css in the newtonian limit allows for estimates concerning the emission rate of gravitational waves by two coalescing rotating black holes ( araujo et al .
1998 ) . in section
[ new ] we discuss newtonian core - shell models , on their own as well as limiting cases of the corresponding relativistic ones .
we put the hnon - heiles - like structure of the relativistic models in the appropriate astronomical context and stress the physical content of the various terms present in them , in particular we see that the apparently naive constant relativistic solution does in fact hide a relativistic homoeoid ( see chandrasekhar 1987 for newtonian homoeoids ) .
the numerical part of this article is presented in section [ caos ] and deals with a study of the chaoticity of orbits trapped in the intermediate vacuum between the core and the shell .
specifically , we shall explore in this section the fact that , due to its generality , axially symmetric core - shell expansions are particularly suitable to study the dynamical role played by a discrete symmetry of the model namely the reflection symmetry about an equatorial plane , on the chaoticity of orbits .
the reflection symmetry is present in the model if and only if the shell is of even - type , i. e. _ iff _ @xmath1-poles , @xmath2 , occur in the shell expansion , and it is broken if we add any odd , @xmath3-poles to the expansion .
the hypothesis of reflection symmetry is a widely spread assumption in astronomical modeling and its twofold justification lies firstly in that too many real celestial objects to be modeled , for example stars themselves , star clusters and galaxies seem in fact very symmetric with respect to a middle plane .
another reason for this symmetry assumption is that the resulting model gets strongly simplified in both analytical and numerical aspects . of course
, we do not expect the reflection symmetry be realized in nature exactly , so it is relevant to search for possible detectable dynamical effects arising from deviations of that symmetry .
a strong additional motivation for this numerical study is to compare from the point of view of the chaotic behavior what happens when the presence of true central black holes are simplified by reducing them to newtonian central masses .
this is a commom practice in current modeling , as for example in gerhard and binney ( 1985 ) , sridhar and touma ( 1997 ) and valluri and merrit ( 1998 ) .
surprisingly enough , the numerical findings show marked effects on the chaoticity of the orbits in the intermediate vacuum between the core and the shell linked to the presence / lack of the reflection symmetry in the relativistic as well as newtonian models .
we also find strong quantitative differences in the chaoticity manifested in both relativistic and newtonian cases . finally , we present the conclusions with some discussion and prospects in section [ discussion ] .
we deal here with static , axially symmetric models , for which the weyl coordinates @xmath4 are the starting point : @xmath5 -e^{-2\nu}\rho^2d\phi^2,\label{hh1}\ ] ] where @xmath6 and @xmath7 are functions of @xmath8 and @xmath9 only . except where units are explicitly required ( particularly in section [ new ] ) , we use non - dimensional variables : @xmath10 ( idem for @xmath8 and @xmath9 ) , and @xmath11 where @xmath12 is the light velocity in vacuum and @xmath13 is some convenient unit of length .
einstein s equations in absence of matter ( vacuum ) reduce in this case to the usual laplace equation for @xmath6 @xmath14 and the quadrature @xmath15\;d\rho+ 2\rho\nu_{,\rho}\nu_{,z}\;dz\label{hh3}\ ] ] for @xmath7 . before proceeding ,
we mention that the weyl coordinates should be considered somewhat deceiving if we insist in naively transferring their image contents about mass configurations to the newtonian common sense .
for example , the spherical shape of the horizon of a black hole is compressed into a bar in weyl s coordinates , the interior portion of the black hole space - time being wholly discarded from the portrait .
apart from the well known fact that the relativistic context is not cast in a straight relation with the newtonian one in strong regimes , they are mathematically sound as will be clear later when we will compare both versions of the core - shell models .
we pass to the prolate spheroidal coordinates @xmath16 , which have a direct link with the `` spherical '' ones @xmath17 ( @xmath18 ) we will use later , @xmath19 { \mbox{,\hspace{0.4cm}$u \geq 1,$}}\nonumber\\ & & v = cos\theta=\frac{1}{2}[\sqrt{\rho^2+(z+1)^2}- \sqrt{\rho^2+(z-1)^2 } ] { \mbox{,\hspace{0.4cm}$-1\leq v\leq 1$}},\label{hh4}\end{aligned}\ ] ] or , in terms of @xmath20 , @xmath21 eqs .
( [ hh2 ] ) and ( [ hh3 ] ) are written in terms of @xmath22 as @xmath23_{,u}+[(1-v^2)\nu_{,v}]_{,v}=0,\label{hh5b}\ ] ] @xmath24,\nonumber\\ & & \gamma_{,v}=\frac{u^2 - 1}{u^2-v^2}[v(u^2 - 1)(\nu_{,u})^2- v(1-v^2)(\nu_{,v})^2 + 2u(1-v^2)\nu_{,u}\nu_{,v}].\label{hh5a}\end{aligned}\ ] ] in the prolate spheroidal coordinates @xmath25 , @xmath26 , laplace equation ( [ hh5b ] ) can be separated and solved in terms of standard legendre polynomials @xmath27 ( see for example moon and spencer 1988 , and , for relativistic applications , reina and treves 1976 ) : @xmath28[c_{\ell } q_{\ell}(v)+d_{\ell}p_{\ell}(v)].\label{hh6}\ ] ] the particular solution picked out from the general one above is determined by the matter distribution whose model is wanted .
we are interested here in monopolar core plus shell - type models , so we are guided by the newtonian case ( to be detailed in section [ new ] ) to the specific solution of the form ( up to third order ) @xmath29 @xmath30 is an integration constant and @xmath31 describes the monopolar core , which either reduces to a black hole if we put @xmath32 and identify @xmath33 with the schwarzschild radius of the core ( in which case @xmath17 above are schwarzschild s coordinates ) , or can be switched off by simply putting @xmath34 .
the remaining terms correspond to the multipoles originated from the exterior shell of matter : dipole @xmath35 , quadrupole @xmath36 and octopole @xmath37 ( note the opposite sign convention for dipoles in this definition with respect to that appearing in vl2 ) .
the nontrivial character of shell dipoles in both newtonian and general relativistic theories of gravity has been anticipated in vl2 while shells made of quadrupoles plus octopoles were considered in vl1 .
if non - monopolar cores should be considered , we should add to ( [ hh7 ] ) terms of the form @xmath38 , @xmath39 .
much more sophisticated multipolar relativistic treatments for core , shell and core - shell models are considered by thorne ( 1980 ) , zhang ( 1986 ) and suen ( 1986a , 1986b ) , respectively , in terms of higher order systematic expansions of the metric in de donder coordinates .
our approach is more modest at this stage and inspired only on the zeroth order , newtonian limit of the full relativistic situation .
the odd multipoles of @xmath26 in ( [ hh7 ] ) ( @xmath35 and @xmath37 in the present case ) break the reflection symmetry about the plane @xmath40 , and we shall show in section [ reg ] that we need of both @xmath35 and @xmath37 being simultaneously either present or absent if we want to rule out conical singularities of the intermediate vacuum between the core and the shell . we rewrite ( [ hh7 ] ) as @xmath41 after integrating ( [ hh5a ] ) we obtain the corresponding solution for @xmath7 : @xmath42,\\ \gamma_q=-4\kappa{\cal{q}}u(1-v^2)\\ { \mbox{\hspace{1.0cm}}}+(1/2){\cal{q}}^2[u^4(1-v^2)(1 - 9v^2 ) -2u^2(1-v^2)(1 - 5v^2)-v^2(2-v^2)],\\ \gamma_o=-(2/5)\kappa { \cal{o}}v[5(3u^2 - 1)(1-v^2)-4]\\ { \mbox{\hspace{1.0cm}}}+(3/100){\cal{o}}^2[-25u^6(1-v^2 ) ( 5v^2 + 2v-1)(5v^2 - 2v-1)\\ { \mbox{\hspace{1.0cm}}}+15u^4(1-v^2)(65v^4 - 40v^2 + 3 ) -3u^2(1-v^2)(25v^2 - 3)(5v^2 - 3)\\ { \mbox{\hspace{1.0cm}}}-v^2(25v^4 - 45v^2 +
27)],\\ \gamma_{dq}=-4{\cal{d}}{\cal{q}}uv(u^2 - 1)(1-v^2),\\ \gamma_{do}=(3/10){\cal{d}}{\cal{o}}[u^2(5u^2 - 6)(1-v^2 ) ( 1 - 5v^2)+v^2(5v^2 - 6)],\\ \gamma_{qo}=-(6/5){\cal{q}}{\cal{o}}uv(u^2 - 1)(1-v^2 ) [ ( 5u^2 - 1)(3v^2 - 1)+2(1-v^2)].\label{hh9 } \end{array}\ ] ] this is the complete solution , which was outlined in vl2 .
the additional integration constant @xmath43 will be fixed in the next section in connection with the elimination of conical singularities from the intermediate vacuum , while the integration constant @xmath30 will prove to be necessary in assuring analyticity to the full core - shell solution at the horizon in the absence of conical singularities .
the terms proportional to @xmath44 in @xmath45 , @xmath46 and @xmath47 represent nonlinear interactions between the black hole and the external shell
. we shall be interested in two classes of singularities of the weyl spacetime .
the first one are the strong singularities that are located in the points wherein the scalar polynomial invariants of the curvature tensor blow up . for a static axially symmetric spacetime solution to the vacuum field equations we have only two non - vanishing invariants .
they are ( carminati and mclenaghan 1991 ) @xmath48 and @xmath49 where @xmath50 is the weyl trace - free tensor , which for vacuum solutions coincides with the riemann curvature tensor .
after some algebraic manipulations , they reduce to @xmath51,\\ w_2=3\kappa ^3\rho^{-2}(\nu _ { , \rho}/\rho-\sigma ) [ \sigma ( 3\rho^3\nu _ { , \rho}^2\nu _ { , z}^2+\rho ( 1 - 3\rho\nu _ { , \rho } ) \sigma + 2\nu _ { , \rho } ) \\ { \mbox{\hspace{1.0cm}}}-\nu _ { , z}^2 ( \gamma _ { , \rho}+3\nu _ { , \rho } ( 1 - 2\rho\nu _ { , \rho } ) ) + \rho^3 ( \nu _ { , z}^6+\nu _ { , \rho}^6 ) \\ { \mbox{\hspace{1.0cm}}}+r\nu _ { , \rho z } ( 2\nu _ { , z } ( 3\nu _ { , \rho } ( 1-\rho\nu _ { , \rho } ) + \rho\nu _ { , z}^2 ) + \nu _ { \rho z } ) \\ { \mbox{\hspace{1.0 cm } } } + \nu _ { , \rho\rho } ( 3 \gamma _ { , \rho } ( 1 - 2\rho\nu _ { , \rho } ) + 4r^2\nu _ { , \rho}^3-\rho\nu _ { , zz } ) ] \label{hh9a } \end{array}\ ] ] where @xmath52 an @xmath53 . in the present case , these scalars are singular on the position of the attraction center and will be also singular in some directions of the spatial infinity ( the specific directions will depend on the signs of @xmath54 , @xmath55 , etc . ) , in other words they are singular in the position of the sources of the gravitational field .
there is another class of singularities that do not show up in a simple way in the curvature , the so called conical singularities ( css ) ( sokolov and starobinskii 1977 ) .
these singularities are in some sense like the distributions related to a low dimensional newtonian potential , e.g. , for an infinite massive wire the potential is @xmath56 , we have that the laplace equation , the analog to the curvature , gives us @xmath57 for @xmath58 while for @xmath59 we need to use some global property , for example gauss theorem , to get @xmath60 . to be more precise ,
let us consider a conical surface , @xmath61 , embedded in the usual euclidean three dimensional space @xmath62 , so that we have on the cone @xmath63 with @xmath64 and @xmath65 .
note that the coordinate range of @xmath66 as well as @xmath67 are the usual ones .
the ratio between the arc of the circumference and its radius is @xmath68 in this case .
if we compute the riemann tensor for this last metric we find that @xmath69 for @xmath70 . by using the gauss - bonnet theorem in this case
, we can pick up the curvature singularity as being of the form @xmath71 .
it is well known that css arise when we consider weyl s solutions near the symmetry axis ( robertson and noonan 1968 and letelier and oliveira 1998a ) .
they are interpreted as a geometric consequence of the presence of some kind of `` strut '' , necessary to consistently prevent any static non - spherically symmetric model from collapsing due to self - gravity . in this sense , the uniqueness of the core - type , one - body solution for the ( static ) spherically symmetric einstein s equations ( namely , the schwarzschild one ) should relate to the fact that it is impossible to attach struts in a perfectly round metric .
for the weyl solutions , we can always consider a small disk centered in the symmetry axis ( @xmath72 ) and impose that the ratio between the circumference and its radius equals @xmath73 .
this is the condition to eliminate css from the intermediate vacuum ( then attaching them to the infinity ) , which amounts to impose on the function @xmath7 the well known conditions of elementary flatness @xmath74 these two conditions fix the constant @xmath43 and impose an additional constraint on the _ odd _ shell multipoles ( @xmath35 and @xmath37 here ) in the presence of the black hole : @xmath75 @xmath76=0.\label{hh11a}\ ] ] the condition ( [ hh11 ] ) fixing @xmath43 is necessary in all cases to rule out the conical singularities . from ( [ hh11a ] ) we see that the former is also a sufficient condition in two cases : i ) the shell of dust is left alone by switching off the core ( @xmath77 ) , and ii ) in the presence of the black hole ( @xmath78 ) , the shell is made only of even - type multipoles ( @xmath36 here ) .
we also see that it is possible to eliminate conical singularities for a single or pure shell component only if it is of the even type .
then , it follows that the core - shell dipole solution presented in vl1 , as well as those with @xmath79 presented in vl2 , all have conical singularities since they do satisfy ( [ hh11 ] ) but _ not _ ( [ hh11a ] ) .
if we include the necessary condition ( [ hh11 ] ) in ( [ hh9 ] ) , we have the following for @xmath7 : @xmath80,\\ \gamma_o=-(2/5)\kappa { \cal{o}}v[5(3u^2 - 1)(1-v^2)-4]\\ { \mbox{\hspace{1.0cm}}}-(3/100){\cal{o}}^2(u^2 - 1)(1-v^2 ) [ ( 25u^4 - 20u^2 + 1)(25v^4 - 14v^2 + 1)\\ { \mbox{\hspace{1.0cm}}}+30u^2v^2(5v^2 - 1)+6(1-v^2)],\\ \gamma_{dq}=-4{\cal{d}}{\cal{q}}uv(u^2 - 1)(1-v^2),\\ \gamma_{do}=(3/10){\cal{d}}{\cal{o}}(u^2 - 1)(5u^2 - 1 ) ( 1-v^2)(1 - 5v^2),\\ \gamma_{qo}=-(6/5){\cal{q}}{\cal{o}}uv(u^2 - 1)(1-v^2 ) [ ( 5u^2 - 1)(3v^2 - 1)+2(1-v^2)].\label{hh12 } \end{array}\ ] ] here , it is explicitly seen that only the first terms in @xmath45 and @xmath47 , respectively , do not vanish identically at @xmath81 .
they cancel one another at those points only if we use the additional condition ( [ hh11a ] ) in the equation above , which amounts to put there either @xmath34 or @xmath82 plus @xmath32 . in the later case , @xmath7
lastly becomes @xmath83,\\ \gamma_{do}=-2{\cal{o}}v(3u^2 - 1)(1-v^2)-(1/4 ) { \cal{o}}^2(u^2 - 1)(1-v^2)\times \\ { \mbox{\hspace{1.3cm}}}(75u^4v^4 - 42u^4v^2 - 42u^2v^4 + 18u^2v^2 + 3u^4 + 3v^4 + 1),\\ \gamma_{qdo}=-2{\cal{q}}{\cal{o}}uv(u^2 - 1)(1-v^2)(3u^2 - 1 ) ( 3v^2 - 1),\label{hh12a } \end{array}\ ] ] remembering that now @xmath35 is present only through @xmath37 .
it remains an arbitrariness in the full solution , namely the constant @xmath30 . to show that it plays a non - trivial role in assuring analyticity to the solution at the black hole horizon , we start by writing the solution in the schwarzschild coordinates @xmath17 : @xmath84-e^{-p}r^2 sin^2\theta d\phi^2,\label{hh13}\ ] ] with @xmath85 and @xmath86 given respectively by ( [ hh8 ] ) and ( [ hh9 ] ) with @xmath78 . at first sight
we should expect that the singularity of this metric at the horizon remains only a coordinate defect since all metric functions here do differ from the corresponding schwarzschild ones by well behaved exponentials . to see what really happens there
, we go to kruskal coordinates @xmath87 , defined as @xmath88 in order to eliminate the usual horizon divergence coming from the factor @xmath89 .
the new , kruskal components @xmath90 are obtained from the old , schwarzschild ones @xmath91 via @xmath92 the line element ( [ hh13 ] ) reads in kruskal s coordinates as @xmath93\\ { \mbox{\hspace{1.0cm}}}-e^{q - p}r^2d\theta^2-e^{-p } r^2 sin^2\theta d\phi^2,\label{hh16 } \end{array}\ ] ] where @xmath94 ( and hence @xmath95 ) is implicitly , analytically given in terms of @xmath96 by the first relation in ( [ hh14 ] ) , @xmath97 , and @xmath98 and @xmath99 are defined by @xmath100 an inspection of ( [ hh16 ] ) and ( [ hh17 ] ) clearly shows that this metric is analytic at the horizon ( @xmath101 or @xmath102 ) if and only if @xmath99 is analytic there . if we define @xmath103 , from which follows that @xmath104 , this amounts just to impose the finiteness of @xmath105 .
in fact , we see that @xmath106 , with @xmath107 a constant depending only upon @xmath30 , @xmath43 , @xmath35 , @xmath36 and @xmath37 , and @xmath108 being a function only of @xmath109 , so that the limit @xmath110\}\label{hh18}\ ] ] is finite ( and equals to @xmath111 ) if and only if the condition @xmath112 holds identically .
this additional constraint on @xmath30 and @xmath43 reads @xmath113=0,\label{hh19 } \end{array}\ ] ] which fixes @xmath30 .
we see that the analyticity at the horizon and the non - existence of conical singularities are independent conditions , however if the necessary condition for the absence of conical singularities ( [ hh11 ] ) also holds , ( [ hh19 ] ) reduces to the following first order relation @xmath114 in any case both conditions taken jointly make the solution to be unique .
it is worth to emphasize that assuring analyticity to the metric at the horizon has nothing to do with removing neither strong nor conical singularities from curvature invariants , since the vacuum between the core and the shell are free from strong singularities while the conical ones are removable from there through conditions ( [ hh11 ] ) and ( [ hh11a ] ) above .
secondly , however kruskal coordinates are historically linked to the question of geodesic completness , this issue is not necessary to our purposes .
in fact , we use here the analyticity of the kruskal coordinates near the horizon of the central black hole only to assure that the shell is a _ truly _
controllable perturbation of the black hole in that region if expanded at any truncation order in the shell parameters , in the same sense considered by vishveshwara ( 1970 ) for linear perturbations .
relations ( [ hh19 ] ) and ( [ hh19a ] ) show that analyticity near the horizon is a nontrivial property of the relativistic shells , having to be forced into the solution to assure its analytic behavior there .
we consider now the linearization of the solution with respect to the shell parameters , i. e. , we consider the shell as a small perturbation ( e. g. formed by dust ) of the central black hole geometry . due to the spherical symmetry of the schwarzschild background
, it is assumed that _ any _ linear perturbation could be expanded in spherical harmonics , an approach due firstly to regge and wheeler ( 1957 ) to study the stability of black holes under small perturbations .
we shall see that this is not true if conical singularities are present , in spite of the perturbation we are dealing with being a perfectly linear one .
we firstly summarize the regge - wheeler approach .
the metric is expanded as @xmath115 , where @xmath116 is the schwarzschild metric , @xmath117 is some small parameter and @xmath118 is the general first order perturbation of the metric .
we put @xmath119 in the einstein vacuum equations @xmath120 , where @xmath121 is the ricci tensor , and obtain @xmath122 , which leads to the well known regge - wheeler ( rw ) differential equations @xmath123 for the perturbation @xmath124 . before trying to solve these equations , it is possible to expand @xmath125 in tensor spherical harmonics in the schwarzschild coordinates @xmath17 ( see mathews ( 1962 ) , zerilli ( 1970 ) and thorne(1980 ) ) .
@xmath124 falls in one of the two following general classes of perturbations , depending on its parity under rotations about the origin performed on the 2-dimensional manifold @xmath126const . , @xmath127const .
: one class ( the even - type one ) has parity @xmath128 and its general ( symmetric ) form is @xmath129 where @xmath130 is the standard @xmath131-mode spherical harmonic , @xmath132 , @xmath133 , @xmath134 and @xmath135 are the corresponding functions of the non - angular coordinates and we define the partial differential operators @xmath136 , @xmath137 , and @xmath138 .
the _ odd _ class has parity @xmath139 and its general ( symmetric ) form is @xmath140 where @xmath141 and @xmath142 , and @xmath143
. the superposition of perturbations are valid in the linearized theory , so in the matrices above we can assume the einstein summation convention in the indices @xmath131 ( truncated at the convenience ) for the case of a more general multi - mode perturbation .
we can simplify these matrices further by exploring the freedom to make arbitrary , first order in @xmath117 , coordinate transformations around @xmath144 .
in particular , the matrices achieve its most simple or canonical forms in the so called regge - wheeler infinitesimal gauge ( regge and wheeler 1957 ) .
nonetheless , we do not need to go to that gauge here . if we linearize ( [ hh13 ] ) in the shell strengths we obtain for the perturbation @xmath124 @xmath145 with @xmath146 and @xmath147 given respectively by ( [ hh8 ] ) and ( [ hh9 ] ) after dropping the second order terms in the shell strengths appearing in those equations ( remember that @xmath95 and @xmath97 ) . the linear perturbation ( [ hh30 ] ) is a compulsory solution of the linearized equations @xmath148 , since it is the first order term of the expansion of an exact solution of the full equations .
having in mind the axial symmetry of our static diagonal perturbation , an inspection of it shows that it is a superposition of even - type modes only .
then , it should be fitted by the matrix ( [ hh28 ] ) with the terms @xmath149 and @xmath150 retained .
we find that this fitting is impossible in general .
in fact , the linearized solution ( [ hh30 ] ) is a regge - wheeler perturbation only if it satisfy the following additional constraints ( here @xmath32 ) : @xmath151 on the other hand , the rw formalism lies on a sound mathematical basis , namely the multipole expansion theory ( for a review with emphasis in general relativity , see thorne 1980 ) , so the reason for this drawback must have a physical origin .
in fact , as it is shown in sokolov and starobinskii ( 1977 ) , to conical singularities there correspond a curvature ( and hence a ricci ) tensor proportional to a dirac delta term centered at the symmetry axis . from einstein
s equations , this means that there correspond to them a certain energy - momentum tensor with this same structure ( it does not matter here how exotic they should be ) and hence we does not have a true vacuum between the core and the shell , as it is supposed from the starting in the rw formalism . by comparing ( [ hh31 ] ) with the conditions ( [ hh11 ] ) plus ( [ hh11a ] ) with @xmath32 for schwarzschild s metric , we see that the former amounts to rule out the conical singularities from the linear approximation .
in other words , conical singularities _ survive _ to the linearization and their presence is an obstruction to the application of the rw formalism .
so , what is some surprising in all this is that conical singularities strutting the model against its own gravity persist after the linearization , contrary to the loosely accepted idea that they should lie in the very non - linear realm of general relativity .
we now consider the newtonian limit of the model ( [ hh13 ] ) in the schwarzschild coordinates @xmath17 .
we assume that there exist a region @xmath152 in the vacuum between the core s horizon and the shell where the conditions of weak gravitational field and slow motion of test particles occur .
then , eintein s equations reduce in @xmath152 to laplace s equation for the newtonian potential @xmath153 , which relates to the metric @xmath154 only through the temporal component as @xmath155 .
the remaining components of the metric are irrelevant to this approximation .
now , we consistently assume that @xmath152 is far away from the horizon , where the schwarzschild coordinates approximate to the usual time plus the euclidean spherical ones ( we maintain the notation @xmath17 in the later approximation ) .
next , we pass to cylindrical coordinates @xmath156 via @xmath157 , @xmath158 and expand the component @xmath159 of ( [ hh13 ] ) to the first order in @xmath30 , @xmath54 , @xmath55 and @xmath160 , to obtain after some manipulation : @xmath161[1-\frac{1}{r}]\\ { \mbox{\hspace{3.5cm}}}+{\cal{q}}(2z^2-r^2)[1- \frac{2}{r}][(1-\frac{1}{r})^2- \frac{1}{3r^2}]\\ { \mbox{\hspace{3.5cm}}}+{\cal{o}}(2z^3 - 3zr^2 ) [ 1-\frac{2}{r}][1-\frac{1}{r}][1- \frac{2}{r}+\frac{2}{5r^2}].\label{hh21 } \end{array}\ ] ] the equation above was presented in vl1 ( without the dipole ) in a rather obscure ( yet correct ) form for the sake of obtaining the newtonian limit . by assumption ,
the non - dimensional constants satisfy @xmath162 ; moreover , @xmath94 satisfies @xmath163 in the region @xmath152 , so that each square bracket appearing in ( [ hh21 ] ) reduces to unity in this approximation .
the only place where @xmath94 itself survives is in the term @xmath164 of ( [ hh21 ] ) , just that due to the monopolar core with mass @xmath165 . the final step is to rewrite the surviving terms with the unit of length @xmath166 appearing explicitly ( remember that until now @xmath94 stood for @xmath167 , etc . ) .
the result for @xmath153 is @xmath168 for comparison , let us briefly recall _ ab initio _ the proper newtonian formulation : let the coordinate origin stay at the center of mass of the monopolar core ( with mass @xmath165 ) , @xmath9 be the symmetry axis of the core shell system and @xmath169 be the region between the smallest and the largest spheres centered at the origin that isolate the inner vacuum from the core and the shell . we have to solve laplace s equation in @xmath169 for the axially symmetric newtonian potential . by using the standard legendre expansion ,
we arrive at the following gravitational potential @xmath170 felt by test particles evolving in @xmath169 : @xmath171,\label{hh23}\ ] ] where @xmath172 , and @xmath173 are the usual cartesian coordinates .
@xmath174 are respectively the constant , dipole , quadrupole and octopole shell strengths given by the following volume integrals over the shell with mass distribution @xmath175 ( @xmath176 standing for the legendre polynomial of order @xmath177 ) : @xmath178{c } \int\!\int\!\int \\ [ -0.2 cm ] { \mbox{\tiny { shell } } } \end{array } \rho ( r,\theta)\frac{p_n(cos\theta)}{r^{n+1}}dv.\label{hh24}\ ] ] obviously , the regions @xmath152 and @xmath169 above must have a non - empty intersection if the newtonian approximation to the full relativistic case is valid , so we assume this and compare both expressions ( [ hh22 ] ) and ( [ hh23 ] ) for the potential in @xmath179 , thus obtaining @xmath180 where @xmath181 is the mass of the shell .
the question is : for which newtonian shells are these constraints on the constants @xmath182 attainable ? to exemplify ,
let the shell be an homogeneous ring of mass @xmath181 , radius @xmath183 , and centered on the @xmath9-axis at @xmath184 , whose density in cylindrical coordinates is @xmath185 .
the integrals @xmath182 in this case are @xmath186 in the limit @xmath187 the ring reduces to a particle with mass @xmath181 placed at @xmath184 , while if @xmath188 it reduces to a ring placed at the equatorial plane with vanishing odd multipoles . in the case of @xmath189 , and given the characteristic core lenght @xmath166 ( a relativistic parameter we know , but that can be crudely anticipated already in the newtonian frame by assuming light can be trapped by gravity like ordinary matter ) , we see that all constraints in ( [ hh25 ] ) are satisfied for , say , @xmath190 ( for the limiting case of a point particle at @xmath184 , we let @xmath187 and this condition become @xmath191 ) . although intuitively expected , this example enlightens and realizes the criteria to overlap both theories , that is to say , to assure the validity of the assumption @xmath192 , and should be compared to the treatment made for example in perry and bohun ( 1992 ) for weyl s solutions with usual , decreasing core - type multipoles .
we discuss now the physical role played by the constant @xmath30 in the model . obviously , @xmath193 , @xmath194 in ( [ hh1 ] ) is a solution of eqs .
( [ hh2 ] ) and ( [ hh3 ] ) for any values of the constants @xmath30 , @xmath43 . by the preceding discussion
, its newtonian limit leads to the constant potential @xmath195 . what is the physical meaning of a constant newtonian potential ? it does describe either trivially the empty space or rather the interior of a very special class of matter distributions , the so called _ homoeoids _ ( see chandrasekhar 1987 and binney and tremaine 1987 ) . the value of the constant potential inside an homoeoid is _ not _ arbitrary , having its value fixed by continuity requirements of the potential through the mass distribution of the homoeoid .
homoeoids are thus gravitationally undetectable from inside .
this shows that the constant solution above is a struted , relativistic version of a newtonian homoeoid .
if in addition we remove the conical singularity from inside the relativistic homoeoid by letting @xmath196 ( remember that conical singularities are not globally removable in static relativistic shell solutions since they are indispensable to strut the shell against its own gravity , so @xmath196 really moves them outside the shell ) , we see that a simple rescaling of the time and the radial coordinate reduces the metric inside the homoeoid to that of minkowski . on the other hand , the condition ( [ hh19 ] ) with @xmath197 shows that if we add a black hole inside an homoeoid then the metric does not extend analytically through the black hole horizon _ unless _ we maintain the `` strut '' in place with strength @xmath198 .
this is a non - intuitive aspect of the ( relativistic ) composition of a black hole with the rather simple homoeoidal shell .
an easy point that is worth to emphasize concerning the interpretation of multipolar expansions of core and shell types is the following : core multipoles measure deviations from _ sphericity _ of central mass distributions , while shell multipoles measure how much shells deviate from _ homoeoids _ , the later being a rather large class of distributions in which homogeneous spherical shells are very particular members .
we close this section with a brief discussion about the hnon - heiles structure of the present newtonian limit .
it was already pointed out in vl1 that the potential of the shell alone pertains to the hnon - heiles family of potentials , yet it does not suffice from its own to confine orbits in the intermediate vacuum ( see the opposite signs of @xmath199 and @xmath9 in the quadrupole term of ( [ hh22 ] ) ) .
though in practice we need to consider the full potential of a given galactic model viz .
its effective potential @xmath200 .
the hamiltonian @xmath99 of a test particle with mass @xmath201 is given in this case by @xmath202 where @xmath203 is the conserved angular momentum of the particle associated to the axial symmetry of the galaxy ( binney and tremaine 1987 ) .
it is also usual to assume that galaxies have further a reflection symmetry about an equatorial plane , in which case the motion restricted to that plane is integrable , in particular having a central stable circular orbit at some fixed radius , say , @xmath204 . to study a nonplanar stelar orbit as a small deviation from the planar one
, we perform a series expansion of @xmath200 in the variables @xmath205 , thus approximating the full motion by a bi - dimensional harmonic oscillator perturbed by higher order terms .
this is the very astronomical origin ( after truncating the series and idealizing the numeric coefficients ) of the cubic hnon - heiles polynomial , now a paradigm of non - integrable potential .
some history about this potential in astronomy can be traced , for example , from contopoulos ( 1960 ) , barbanis ( 1962 ) , van de hulst ( 1962 ) , ollongren ( 1962 ) and hnon and heiles ( 1964 ) . by adding the other terms to the terms originated from the shell expansion to form @xmath200 , we are able to confine test motions in the intermediate vacuum around the central stable orbit .
if the galaxy does not have reflection symmetry around a middle plane ( @xmath206 in the newtonian case and @xmath207 in the relativistic one ) , planar central stable orbits are not possible , there remaining only stable orbits of distorted , nonplanar type .
we will return in section [ caos ] to the discussion of the implications for orbit regularity of the largely assumed hypothesis of reflection symmetry about middle plones in galaxy modeling .
we study the chaoticity of orbits in the intermediate vacuum of newtonian static axially symmetric core - shell models and then compare them to the geodesics of the corresponding relativistic cases .
it is amazing that this class of models has relativistic and newtonian counterparts to be compared , there existing moreover sound observational motivations for both .
specifically , we analyse in both cases the role played by a _ discrete _ symmetry , namely the reflection symmetry around the equatorial plane , in regularizing orbits against chaoticity .
the hypothesis of reflection symmetry is a widely spread assumption in astronomical modeling and its twofold justification lies firstly in that too many real objects to be modeled seem in fact nearly symmetric with respect to a plane . on the other hand ,
the models are simplified in both analytical and numerical aspects by that assumption . of course
, we do not expect that symmetry be realized in nature exactly so it is relevant to search for possible orbital effects of its breaking .
it is important to realize that all cases treated here preserve the axial symmetry the next continuous symmetry after the missing spherical one , which allows us to isolate the dynamical effects of only breaking or preserving the reflection symmetry itself .
another advantage of the present study is that it is model independent ( of course within the class of models we are dealing with ) as we are focusing on general multipolar shell expansions instead on specific mass distributions .
all is made non - dimensional again in this section by formally taking @xmath208 , etc .
, @xmath209 , etc . , and @xmath210 in equations ( [ hh22 ] [hh23 ] ) plus ( [ hh27 ] ) , such that the newtonian system we consider comes from the hamiltonian ( @xmath211 ) @xmath212 on the other hand , the geodesic system comes from the lagrangean @xmath213 given by @xmath214 where the metric tensor @xmath215 and the coordinates @xmath216 are obtained from the non - dimensional invariant interval ( [ hh1 ] ) , and the dot stands for the derivative @xmath217 .
the proper definition of time - like geodesics out of the metric interval furnishes the first constant of motion @xmath218 .
the euler - lagrange formulation of the geodesic equations is @xmath219 from which the two additional constants of motion associated to the static nature and axiality of the relativistic system are read out , namely the relativistic energy @xmath220 and angular momentum @xmath221 defined by @xmath222 the remaining two equations in ( [ hh34 ] ) describe the dynamics for the variables @xmath223 .
we note that these @xmath223 are weyl s coordinates . from ( [ hh5 ] )
we see that whenever the schwarzschild coordinate @xmath94 satisfies @xmath163 it approximates the usual radial spherical one ( denoted here by the same letter @xmath94 ) and the weyl @xmath223 approximate the usual cylindrical ones @xmath224 .
we remember also that in the newtonian limit @xmath225 ( low velocities ) , @xmath226 , so the relativistic energy @xmath220 and the newtonian energy @xmath227 of an orbiting particle with mass @xmath201 are related through @xmath228 , and the corresponding angular momenta through @xmath229 .
we put @xmath230 and @xmath32 in this section .
we impose the conditions ( [ hh11 ] ) and ( [ hh11a ] ) for the absence of conical singularities on all running relativistic situations presented bellow , except of course for the purely dipolar shell .
all newtonian as well as relativistic poincar sections ( surfaces of section ) shown here are made at the plane @xmath40 ( with the appropriate coordinate interpretation in each case ) . in fig . 1 we present typical effective newtonian potential wells for the bounded orbits we are considering ( @xmath231 stands for @xmath200 and
the positive parts of the potential surfaces surrounding the wells are cut ) .
1(a ) and ( c ) show cases of oblate and prolate potentials , respectively , both possessing reflection symmetry about the equatorial plane given by @xmath40 .
oblate cases ( a ) have positive quadrupole strengths ( @xmath232 ) while prolate cases ( c ) have @xmath233 ( respectively @xmath234 and @xmath235 , see eqs .
( [ hh22 ] ) and ( [ hh23 ] ) and , for the specific case of a ring , eq .
( [ hh26 ] ) ) .
the presence of the reflection symmetry implies the vanishing of all odd shell multipoles ( @xmath54 , @xmath160 and @xmath236 , @xmath237 here ) .
we see that the oblate case has only one unstable equilibrium point ( in the equatorial plane ) while the prolate case has two of them ( symmetrically placed outside the equatorial plane ) , which is easy to understand in terms of the exterior oblate / prolate shell mass distributions .
figs . 1 ( b ) and ( d )
show the typical deformation of the previous oblate and prolate cases , respectively , when we break the symmetry of reflection about @xmath40 by introducing non - vanishing odd shell multipoles ( @xmath54 , @xmath160 ) .
on the other hand , the relativistic regime is felt in the situation we are concerned with mainly through the existence of one more unstable equatorial equilibrium point in addition to those already present in the newtonian potential .
this intrinsically relativistic additional unstable point is associated to the presence of the black hole at the center and marks the point above which the orbits fall into the black hole . in the remaining ,
the relativistic potentials are qualitatively similar to those displayed in fig . 1 for the newtonian case .
we have made a wider exploration of the shell parameters , energy and angular momentum than shown here , with the same conclusions in large .
the parameter values we actually chose to show are a compromise between the need to make consistent comparisons and the sharpening of the effects we found .
the case of a purely dipolar shell has been anticipated in vl2 , where we stressed that the monopolar core plus shell dipole is non - trivial already in the newtonian context .
moreover , we show that the newtonian case is integrable whereas the relativistic one is chaotic , which justifies the characterization of chaos in the purely dipolar case as an intrinsic general relativistic effect . here , we explore the fact that the newtonian dipolar shell breaks the reflection symmetry of the model without breaking the integrability of the motion itself . thus core - shell models provide us with two very distinct integrable situations from the point of view of the reflection symmetry : core plus dipolar shells ( d - cases ) , which do not have that symmetry , and purely monopolar cores ( kleperian cases ) which do have it . a typical poincar section of an integrable d - case is shown in fig .
firstly , we perturb both integrable configurations above with a reflection symmetry preserving , _ oblate _ quadrupolar term ( @xmath232 ) .
surfaces of section for core plus oblate quadrupolar shells ( oblate q - cases ) are shown in figs . 2(b , c , f ) while sections for core plus oblate dipolar - quadrupolar shells ( oblate dq - cases ) are shown in figs .
2(d , e ) .
these figures show that the breaking of the integrability of the d - cases ( without reflection symmetry ) by a quadrupole is much stronger than that of the keplerian case ( with reflection symmetry ) by the same quadrupole .
moreover : fig .
2(c ) shows that chaoticity in the oblate q - cases is in fact present only in a residual , `` microscopic '' level , fig .
2(f ) illustrates the robustness of the strong regularity of the oblate q - cases against the varying of the quadrupole strength , while fig .
2(e ) shows that the strong chaoticity in the dq - cases is very dependent on the energy , in contrast with the robustness of the almost regularity of the oblate q - cases also with the energy .
3 shows the effects of introducing a reflection symmetry breaking , octopolar perturbation on the two integrable configurations ( d - cases and keplerian cases ) , and also on the almost regular oblate q - cases .
3(a ) and ( f ) show the effects of a purely octopolar shell on the keplerian case ( o - case ) at two different octopole strengths .
we see that these effects are very much stronger in comparison with the almost - regular oblate q - cases .
3(b - e ) show the effects of the various combinations of shell multipoles and confirm the dominance of the chaotic effects associated to the odd multipoles over those related to the purely oblate q - cases .
in particular , the enhancement of chaoticity associated to the breakdown of the reflection symmetry is reinforced by the oblate qo - cases presented in fig .
3(c ) in comparison with the oblate q - cases , and figs . 3(d ) and ( e ) confirm the strong dependence with the energy of the chaoticity associated to the lack of reflection symmetry .
4 shows surfaces of section for _ prolate _ quadrupolar cases ( @xmath233 ) . figs .
4(a - c ) show the very much higher chaoticity of the prolate q - cases in comparison to the previous quasi - regular oblate q - cases .
this is to be expected in view of the prolate q - cases having two unstable equilibrium points ( and hence suffering the simultaneous influence of the two corresponding instability regions ) instead of only one occurring in the oblate q - cases .
these figures also show that the central primary stable orbit , deeply inside the accessible region , is strikingly the first one to bifurcate and drive the chaotic behavior with the rising of the energy .
the strong dependence with the energy of the whole set of bifurcations is also a distinctive aspect of these figures . in fig .
4(d ) we show the effect of applying the perturbing prolate quadrupolar term on the integrable , reflection symmetry missing d - cases ( prolate dq - cases ) .
we find that the absence of mirror symmetry in the prolate dq - cases causes the opposite to that occurring in the oblate dq - cases , namely the missing mirror symmetry does greatly enhance the orbit regularity in the prolate cases .
4(e ) does confirm that the breakdown of mirror symmetry associated to the addition of an octopolar term to the prolate q - cases ( prolate qo - cases ) strongly reduces the chaotic manifestation of the prolate q - cases themselves .
in particular , figs .
4(d - f ) show evidences that the lack of symmetry introduces a robust family of regular asymmetric orbits around a stable primary one in the bounded region .
this regularizing effect due to the lack of mirror symmetry in the prolate dqo - cases is easily understandable in terms of the progressive lack of influence of one among the two unstable equilibrium points of the potential on the bounded region accessible to the particle if odd shell multipoles are put in scene ( cf
. figs .
1(c ) and ( d ) ) .
we present in fig . 5 the findings for the corresponding _ relativistic _ core - shell models .
we choose for these figures the same non - dimensional multipole shell strengths and angular momenta of the corresponding newtonian figures and , except for the purely dipolar shell , we always eliminate the conical singularities ( or `` struts '' ) from the intermediate vacuum by fixing @xmath43 and @xmath35 in terms of @xmath36 and @xmath37 through the conditions ( [ hh11 ] ) and ( [ hh11a ] ) .
also , we always put @xmath238 here .
the relativistic energy @xmath220 is chosen in each case to be close that given by the approximate relation @xmath239 where @xmath240 is the newtonian energy .
5(a ) shows a typical surface of section for a black hole plus purely dipolar shells ( rd - cases ) , its strong chaoticity having to be compared with both the almost regularity of the _ oblate _ relativistic quadrupolar case ( rq - case ) shown in fig .
5(b ) ( its very small chaotic zones are visible if we zoom the cross - type regions between the islands ) and the corresponding integrable newtonian d - case shown in fig .
[ in particular , the smallness of the chaotic zones of the oblate rq - case shown in vl1 occasioned the misleading statement made by us therein of these cases not exhibiting chaos , which was promptly corrected in vieira and letelier 1996b . ] in fact , fig .
5(b ) does not represent the full chaotic behavior possible to oblate rq - cases . as we saw above about the relativistic counterparts of the newtonian potentials , the relativistic oblate cases have one more unstable scape point ( in fact , an infall point into the black hole ) inermost at the equator in addition to that one provided in both cases by the oblate shell itself .
hence , if we vary the energy and the angular momentum so that both unstable regions are equally accessible by the bound nearly equatorial geodesics , it is possible to enhance strongly the chaotic behavior of fig .
5(b ) to encompass even a significant portion of the more external region of the surface of section ( that associated to the nearly equatorial orbits ) .
we will see more about this enhancement for the relativistic cases below in fig .
5(c ) shows a surface of section for relativistic dipolar - octopolar cases ( rdo - cases ) .
remember that both odd components are simultaneously needed in view of the off - strut constraint ( [ hh11a ] ) .
5(d ) is the same for full oblate rdqo - cases
5(a - d ) show that the chaotic zones are greatly enlarged whenever odd shell multipoles are present , in contrast with the almost - regular oblate rq - case shown . figs .
5(e ) shows a surface of section for relativistic _
cases ( prolate rq - cases ) .
it is to be compared to the corresponding prolate newtonian one ( see fig .
4(b ) ) , in particular we note the evident fingerprints of the bifurcation series starting from the originally stable central primary orbit already present in the prolate q - cases .
we also see that the prolate rq - cases share with the prolate q - cases the same relatively wild chaotic behavior as compared to the respective oblate ( r)q - cases . like in the fig .
5(b ) , the dynamics would be more chaotic in fig .
5(e ) if the angular momentum and the energy ( unaltered here for the sake of full comparison ) were ajusted to allow all the three unstable points of the relativistic potential ( instead only the two ones related to the prolate character of the shell , see fig .
1 and the comments therein about the additional black hole unstable point ) have stronger combined influence on the motion .
so , the remaining differences between the relativistic figure 5(e ) and the newtonian one 4(b ) are due to the following : the central region ( in the surface of section ) is the only one able to become chaotic in the newtonian case since its orbits are just those ( non - planar ones ) that can reach the unstable regions near the two unstable scape points symmetrically placed outside the equatorial plane , while the more external ( in the surface of section ) quasi - equatorial orbits are regular since they are far from those two points and under the strong influence of the equatorial stable orbit in the boundary of the surface of section .
when we change the newtonian central mass by a true black hole then a third unstable scape point ( in fact a rolling down point into the black hole ) is placed innermost in the equatorial plane .
associated to it there is one more unstable region which can be reached , this time by the quasi - equatorial orbits . the specific combination of parameters of fig .
5(e ) is such that the central , non - planar orbits can feel only moderately the influence of the two scape points linked to the presence of the prolate shell , while the quasi - equatorial orbits are feeling strongly the nearby influence of the unstable point associated to the black hole . as we said above
, other combinations of parameters will allow a much stronger spread and eventually the overlaping of both types of chaotic regions . finally , fig .
5(f ) shows the restoration of a robust set of regular asymmetric orbits around a primary stable one when we break the reflection symmetry around the equatorial plane by switching on the odd multipoles onto the prolate rq - cases , much the same as occured in the prolate newtonian q - cases .
in the first part of this article we presented an unifying discussion concerning with the properties and , more important , the physical content of some relativistic , static , axially symmetric core - shell models on their own as well as in connection with the corresponding linearized and newtonian models taken as limiting cases . the analytical and observational motivations for these relativistic and newtonian models were also shown .
this was acomplished in a reasonably self - contained manner in the introduction and section 2 and needs no additional comments . in the second , numerical part of this work we explored the fact that the models i ) are generic within the class they pertain and ii ) have exact relativistic and newtonian counterparts to be compared , to study the chaoticity of bound orbits in the vacuum between the core and the shell .
specificaly , we firstly tested the relevance of the presence / absence of the reflection symmetry around the equatorial plane for the chaotic behavior of the orbits .
we find consistent evidences for a non - trivial role played by the reflection symmetry on the chaoticity of the dynamics in both the relativistic and newtonian cases .
we summarize these findings as follows : the _ breakdown _ of the reflection symmetry about the equatorial plane in both newtonian and relativistic core - shell models does i ) _ enhance _ in a significant way the chaoticity of orbits in reflection symmetric _ oblate _ shell models and ii ) _ inhibit _ significantly also the occurrence of chaos in reflection symmetric _ prolate _ shell models .
in particular , the lack of the reflection symmetry provides the phase space in the prolate case with a robust family of regular orbits around a stable periodic orbit that is otherwise missing at higher energies .
the other point we addressed about the chaotic behavior of orbits was for the consequences of substituting true central black holes by newtonian central masses .
we find that the relative extents of the chaotic regions in the relativistic cases are significantly larger than in the corresponding newtonian ones .
although not surprising in thesis , the strong differences between both regimes are in order in view of the procedure found in the literature of simulating the presence of a black hole at the core of a galaxy with a naive @xmath0 newtonian term ( see for example gerhard and binney 1985 , sridhar and touma 1997 and valluri and merrit 1998 ) .
this approximation is certainly valid far from the black hole but not in its proximity as is the case here .
these findings stress i ) the non - trivial role of the reflection simmetry in both relativistic and newtonian regimes , in contrast with its universal acceptance in astronomical modeling , ii ) the strong qualitative and quantitative differences between relativistic and newtonian regimes , in particular when dealing with orbits in the vicinity of black holes , and iii ) the intrincate interplay between both aspects when they are simultaneously present . the true dynamical aspect related to the role of reflection symmetries on the regularity of the orbits refers of course to the parity of the constants of motion ( the famous third one in the present axially symmetric case ) with respect to the coordinates , as well as their power to regularize orbits in phase space . translated to these terms
, the findings above are saying that the additional constant of motion in question is by far more powerful whenever i ) it is an _ even _ function of the coordinates for _ oblate _ cases , or ii ) it is an _
odd _ function of the coordinates for _ prolate _ cases .
since the issue of finding global or even approximate constants of motion given the dynamics is hard if not feasible in most cases , the former procedure of simply checking the presence / absence of reflection symmetries is of much more practical interest .
additional study is needed to see if and how our findings are extendible to more realistic configurations , with and without black holes / central masses .
the obvious improvement is to fill the intermediate vacuum with some reasonable mass distribution . while this is far from obvious in general relativity , it is easy to do in the newtonian context .
for example , we are considering just superpose our shell multipoles to some relevant potentials of celestial mechanics such as plummer - kuzmin , ferrers and others and even triaxial potentials .
they have one or more planes of symmetry , with or without axial symmetry , and so we can perturb them with shell terms to verify similar effects to those we found . here
we considered only tube orbits ( @xmath241 ) since we are interested in the effects of the existence of planes of symmetry on _ tridimensional _ motion ( somewhat similar effects to the found here are seen in gerhard 1985 for the restricted case of planar motions with respect to the existence of one or more lines of symmetry ) .
the model improvements above will allow us to study the effects of breaking the reflection symmetry also on tridimensional box orbits .
meantime , the fully relativistic program is in progress .
we have recently succeeded in giving rotation to a black hole ( i. e. in converting it into a kerr black hole ) plus a dipolar shell term ( letelier and vieira 1997 ) .
there are increasing evidences for the existence of black holes , particularly inside active galactic nuclei ( kormendy and richstone 1995 ) , which motivate us to consider also rotating core - shell models in the same lines followed here . | we make a detailed analysis of the exact relativistic core - shell models recently proposed to describe a black hole or neutron star surrounded by an axially symmetric , hollow halo of matter , and in a seminal sense also galaxies since there are massive shell - like structures as for example rings and shells surrounding
many of them and also evidences for many galactic nuclei hiding black holes .
we discuss the unicity of the models in relation to their analyticity at the black hole horizon and to the full elimination of axial ( conical ) singularities .
we also consider newtonian and linearized core - shell models , on their own to account for dust shells and rings around galaxies and supernova and star remnants around their centers , and also as limiting cases of the corresponding relativistic models to gain physical insight . secondly , these models are generic enough to numerically study the role played by the presence / lack of _ discrete _ reflection symmetries about planes , i. e. the presence / lack of equatorial planes , in the chaotic behavior of the orbits .
this is to be contrasted with the almost universal acceptance of reflection symmetries as default assumptions in galactic modeling .
we also compare the related effects if we change a true central black hole by a newtonian central mass .
our main numerical findings are : + 1- the _ breakdown _ of the reflection symmetry about the equatorial plane in both newtonian and relativistic core - shell models does i ) _ enhance _ in a significant way the chaoticity of orbits in reflection symmetric _ oblate _ shell models and ii ) _ inhibit _ significantly also the occurrence of chaos in reflection symmetric _ prolate _ shell models . in particular , in the prolate case
the lack of the reflection symmetry provides the phase space with a robust family of regular orbits that is otherwise not found at higher energies .
+ 2- the relative extents of the chaotic regions in the relativistic cases ( i. e. with a true central black hole ) are significantly larger than in the corresponding newtonian ones ( which have just a @xmath0 central potential ) . + |
although being very successful , the standard model is unable to resolve some puzzles . among them is a problem of fermion flavor .
the origin of hierarchies between charged fermion masses and ckm mixing angles is unexplained .
is there any underlying theory which might generate these hierarchies in a natural way ? moreover , in order to explain the neutrino data @xcite some extension of the standard model , generating neutrino masses and mixings , is necessary .
the number of fermion generations is a mystery .
do we have only three chiral families of quarks and leptons ? is any selection rule dictating the number of fermion generations ? motivated by these questions , in this paper we address these issues within the framework of supersymmetric ( susy ) @xmath0 grand unified theory ( gut ) .
latter s motivation is to have unified description of electro - weak and strong interactions @xcite , while susy provides natural understanding of gauge hierarchy problem as well as the successful gauge coupling unification @xcite .
for understanding hierarchies between fermion masses and mixings , we apply abelian flavor symmetry @xcite @xmath1 which by requirement is non - anomalous .
the @xmath1 , in combination with susy @xmath0 gut , due to anomaly constraint allows only three chiral families of matter @xmath3-plets and few @xmath0 singlet states .
we use some of these singlet states as right handed neutrinos , in order to build realist neutrino sector .
anomaly constraints fix @xmath1 charge assignment in such a way that texture zero quark and lepton yukawa matrices are generated .
together with natural understanding of hierarchies , model predicts inverted neutrino mass hierarchical scenario blending well with recent daya bay observation @xcite .
consider susy @xmath0 gut augmented with anomaly free @xmath1 flavor symmetry .
setup with anomaly free @xmath1 , will alow to gauge @xmath1 and remain within conventional 4-dimensional field theoretical framework ( without need of discussing @xmath5 s of a stringy origin @xcite ) . in a recent work @xcite the way of finding anomaly free @xmath1 flavor symmetry within susy
@xmath0 gut was suggested .
the finding was realized by embedding of @xmath4 in a single non - abelian group with anomaly free field content . in this way
, the @xmath1 charge assignment can be fixed . amongst several assignments , found in @xcite ,
there is one which also dictates the number of generations to be three .
this , as will be shown below , leads to very economical and attractive scenario for fermion masses and mixings . before showing this
, we briefly discuss the way of finding of such @xmath1 .
the states non - trivial under @xmath0 group , we introduce , will be just those of minimal susy @xmath0 .
these are scalar superfields @xmath6 and three families of matter @xmath3 supermultiplets .
we assume that @xmath7 is not charged under @xmath1 and thus does not contribute to the anomalies .
therefore , upon finding anomaly free @xmath1 charge assignment we will deal with three @xmath8-plets , one @xmath9-plet ( which is @xmath10 ) and four @xmath11-plets ( @xmath12three matter @xmath11-plets plus @xmath13 ) . as already mentioned , we search @xmath1 charge assignment by embedding of @xmath4 in non - abelian group with anomaly free matter @xcite .
let us consider @xmath14 group with chiral supermultiplets @xmath15 .
this simple set is anomaly free @xcite . here
@xmath16 is three index antisymmetric representation and @xmath17 is an anti - fundamental of @xmath14 .
decomposition of these states via the chain @xmath18 looks @xmath19 outside of parenthesis indicate @xmath20 and @xmath21 charges respectively .
note that @xmath20 and @xmath21 are coming from @xmath22 and @xmath14 respectively , with corresponding generators @xmath23 and @xmath24 .
the normalization factors @xmath25 and @xmath26 are omitted in eq .
( [ su7 ] ) .
now , in ( [ su7 ] ) , without change of charge assignments , we replace the pair of @xmath27-plets by the pair @xmath3 . with this replacement all anomalies ( at the level of @xmath0 and @xmath5 s ) will remain intact , i.e. will still vanish .
thus , we will have the following anomaly free content ( 10_-3 + 10_3)_3+(10_2+|5_-4)_-4 + 2 , which involves three families ( ! ) of matter @xmath3 supermultiplets plus four @xmath0 singlets . some of these singlets will be applied as right handed neutrinos ( rhn ) .
worth noting that , in difference from @xmath28 gut , the @xmath0 does not involve ( require ) the rhn states . in the @xmath0 scenario , we have just built up , the rhns are required for anomaly cancellation . by abelian symmetries
@xmath20 and @xmath21 , with charges given in eq . ( [ su7-content ] ) , we can build superposition @xmath29 .
this superposition is automatically anomaly free for arbitrary @xmath30 and @xmath31 , because the orthogonal generators @xmath32 and @xmath33 originate from single @xmath14 .
thus , using ( [ su7-content ] ) we can write the anomaly free set 10_-3|a+3|b+10_3|a+3|b+10_2|a-4|b+|5_-4|a-4|b+ 2(|5_-|a-|b+1_5|a-|b+1_6|b ) , where subscripts denote charges . as it turns out
, for building realistic phenomenology it is useful to add to this superposition another anomaly free @xmath5 , which can be found by similar procedure .
for instance , consider @xmath34-plet of @xmath35 group with a chain @xmath36 of decomposition @xcite : 27=16_1 + 10_-2 + 1_4=(10+|5 + 1)_1+(5+|5)_-2 + 1_4 . here
subscripts denote @xmath37 charges . in this content
, we can replace @xmath38 with @xmath39 and at the same time add @xmath40 .
moreover , we can add two @xmath0 singlets with @xmath37 charges @xmath41 and @xmath42 respectively . with this replacement and additions ,
the anomalies @xmath43 , @xmath44 , etc , will be unchanged .
this content allows to build superposition of three abelian groups @xmath20 , @xmath21 and @xmath37 : @xmath45 .
thus , the field content and @xmath46 charge assignment will look : @xmath47 + 1_5|a-|b+|c+1_5|a-|b+4|c+1_6|b+k|c+1_6|b - k|c , with 30|a(3 + 2p)=|c(2k^2 + 10p^2 - 27 ) .
relations between @xmath48 and @xmath49 ( imposed for @xmath50 ) given in eq .
( [ d - prime ] ) insures that all anomalies vanish .
clearly , with rational selection of @xmath51 and @xmath49 the value of @xmath52 also will be rational .
the set given in eq .
( [ d - prime ] ) is one simple selection among several options and opens up many possibilities for model building with realistic phenomenology .
we will identify these charges with the charges of @xmath1 flavor symmetry .
to the field content of eq .
( [ d - prime ] ) we add the pair @xmath53 .
this is needed to have , besides the matter fields , the higgs supermultiplets @xmath54 .
thus , in total we have three @xmath8-plets , one @xmath9-plet and four @xmath11-plets plus @xmath0 singlets . the @xmath1 charge assignment ( @xmath55 for @xmath56 and for remaining @xmath11-plets given in eq .
( [ d - prime ] ) ) is not unique .
we can exchange @xmath9-plet s @xmath1 charge with one of the @xmath11-plets charge . with this , all anomalies will still vanish .
in addition , out of the four @xmath11-plets , any of them can be identified with the higgs superfield @xmath13 . as it turns out , for the charges of the pair @xmath57 , we will have @xmath58 possible options for @xmath59 : ( q_h , q_|h)^(l)&= & \{(q , -q ) , ( q , -4|a-4|b-2|c ) , ( q , -|a-|b+|c ) , ( q , -|a-|b-2|c ) , . + & & ( -4|a-4|b-2|c , -|a-|b+|c ) , ( -4|a-4|b-2|c , -|a-|b-2|c ) , ( -4|a-4|b-2|c , q ) , + & & ( -|a-|b+|c , -4|a-4|b-2|c),(-|a-|b+|c , -|a-|b-2|c),(-|a-|b+|c , q ) , + & & .
( -|a-|b-2|c,-4|a-4|b-2|c),(-|a-|b-2|c,-|a-|b+|c),(-|a-|b-2|c , q ) } , with @xmath60 . note that we have left out possibilities obtained from those given in ( [ d - qhhbar ] ) by the substitution @xmath61 .
these @xmath58 options open up various possibilities for the model building @xcite .
below we present one of them , which we found to have nice and attractive properties with interesting implications for fermion masses and mixings . in order to break @xmath1 gauge symmetry
, we introduce @xmath0 singlet pair of flavon superfields @xmath62 with @xmath1 charges q(x)=-1 , q(|x)=1 . without loss of generality , we have normalized flavons charges modulo to one .
the scalar components of @xmath63 and @xmath64 acquire vevs and @xmath65 , is given in ref .
] = , = | , where @xmath66 gev is reduced planck scale , which will be treated as natural cut off for all higher dimensional non - renormalizable operators . in our approach , top quark ( and possibly bottom quark and tau lepton , in case of large @xmath67 ) will get mass at renormalizable level .
yukawa couplings of light families emerge after @xmath1 flavor symmetry breaking .
thus , the hierarchies between yukawa couplings and ckm mixing angles will be expressed by powers of small parameters @xmath68 . in the charge assignment
we need to fix the values of @xmath69 and @xmath70 .
if their ratios remain arbitrary there will be more than one extra @xmath5 symmetry , and that we have to avoid .
together with fixing @xmath71 , the values of @xmath72 and @xmath41 ( in ( [ d - prime ] ) and ( [ d - qhhbar ] ) ) should be selected in such a way as to have phenomenologically viable quark and lepton yukawa textures .
it turns out , that one selection leading to attractive yukawa sector , is the following : \ { |a , |b , |c}=\{- , , } , p = q = k=0 . in this case , in eq .
( [ d - qhhbar ] ) we pick up @xmath73 , which fixes charges of @xmath74 as @xmath75 .
the charges of matter @xmath3-plets ( and also @xmath0 singlets ) are also fixed and we will make the following identification : @xmath76 , where @xmath77 labels the flavor .
the model s field content and corresponding @xmath1 charges are given in table [ t : tab1 ] .
@xmath78 with this assignment , @xmath79 and @xmath80-type yukawa couplings are given by ccc & ccc 10_1 & 10_2 & 10_3 + c 10_1 + 10_2 + 10_3 & ( ccc ^4 & & ^2 + & |^2 & | + ^2 & | & 1 ) h , ccc & ccc |5_1 & |5_2 & |5_3 + c 10_1 + 10_2 + 10_3 & ( ccc ^4 & & ^2 + & |^2 & | + ^2 & | & 1 ) |h , [ ude - d - u2 ] where in front of each entry , dimensionless couplings(@xmath81 ) are assumed . as we will see shortly ,
good fit is achieved for @xmath82 . on the other hand , with these values , the matrix elements @xmath83
are so suppressed , that they are irrelevant and we can set them equal to zero .
thus , for all practical purposes , we can investigate the yukawa matrices : ccc & ccc & & + c + + & y_u , d , e ( ccc 0 & & 0 + & |^2 & | + 0 & | & 1 ) , [ texture-0 ] with zero textures . using the basis @xmath84 and @xmath85 , without loss of generality we can parameterize up and down yukawa matrices at gut scale to have forms : y_u ( ccc 0 & c|^2 & 0 + c|^2 & a_u|^2e^i_u & | + 0 & | & 1 ) _ t^0 , [ upy - du2 ] y_d ( ccc e^i & 0 & 0 + 0 & e^i & 0 + 0 & 0 & 1 ) ( ccc 0 & |^2 & 0 + k|^2 & a_d|^2e^i_d & b| + 0 & b| & 1 ) _ b^0 .
[ downy - du2 ] we have made field phase redefinitions in such a way that , in this basis , ckm matrix remains unity and in @xmath86 only one phase @xmath87 appears .
the phases @xmath88 and @xmath89 will not contribute to the quark masses , but will be important for the ckm matrix elements . from eqs .
( [ upy - du2 ] ) and ( [ downy - du2 ] ) , in a fairly good approximation we obtain the following relations valid at gut scale : these relations help to find a good fit . with proper selection of input parameters
@xmath90 , @xmath91 , @xmath92 , @xmath93 , @xmath94 we can get desirable values for fermion mass hierarchies and ckm mixing angles at gut scale .
then , using rg we can calculate these ratios at low scales : @xmath95 @xmath96 .
v_|_m_z=. v_|_m_g , if ( ) = ( ud , us , cd , cs , tb ) , where rg factors _ t = _m_t^m_g_t^2d , _ b = _m_h^m_g_b^2d , _
= _m_h^m_g_^2d are given in 1-loop approximation .
we will consider two cases with low / moderate and large values of the mssm parameter @xmath97 .
we take experimental value @xmath99 gev , determining top yukawa coupling at weak scale , and with @xmath98 we find @xmath100 . for this case ,
good fit is obtained for the following values of input parameters : @xmath101 @xmath102 _
u=0 , _
d=-0.065 , = -2.696 , =-0.97 .
these at gut scale give @xmath103 @xmath104 where @xmath105 .
performing renormalization ( using ( [ rg - mg - mt ] ) and @xcite ) , at low scales we get ( with input @xmath99 gev , @xmath106 gev ) : @xmath107 = m_z : |v_us|=0.2243 , |v_cb|=0.042 , |v_ub|=0.00349 , = 0.118 , = 0.34 .
these values of masses and ckm matrix elements are in good agreement with experiments @xcite , @xcite .
in this case we have @xmath109 .
input parameters are selected as : @xmath110 @xmath111 _
u=-0.055 , _
d=-0.0593 , = -2.73 , =-0.98 , giving at gut scale @xmath112 @xmath113 the renormalization procedure ( using ( [ rg - mg - mt ] ) and @xcite ) gives at low scales ( with input @xmath99 gev , @xmath106 gev ) : @xmath114 = m_z : |v_us|=0.2259 , |v_cb|=0.0409 , |v_ub|=0.00351 , = 0.135 , = 0.345 .
these agree well with experiments .
now let us discuss the charged lepton sector .
relevant yukawa couplings originate from @xmath115-type interactions of eq .
( [ ude - d - u2 ] ) ( while in practice @xmath116 has the structure of eq .
( [ texture-0 ] ) ) . without breaking the @xmath0 symmetry in these interactions
, one would get the asymptotic relation @xmath117 , which is unacceptable and is a well known problem for minimal susy @xmath0 gut .
however , by some specific extension , care can be exercised to solve this problem @xcite . without specifying origin of @xmath0 breaking in this sector
, we assume that it happens ( i.e. @xmath0 symmetry breaking ) in the sector of light families .
thus , in analogy of @xmath118 ( see eq . ( [ downy - du2 ] ) ) , in a basis @xmath119 , we parameterize @xmath116 to have the following form y_e ( ccc 0 & k_e_e |^2 & 0 + _
e |^2 & k_22a_d|^2e^i_e & b| + 0 & b| & 1 ) _
^0 , with _
^0=_b^0 .
[ ye - du2 ] in @xmath116 only one complex phase @xmath120 appears .
remaining phases are rotated away by proper phase redefinitions of the @xmath121 and @xmath122 states .
with @xmath123 we can avoid the relation @xmath124 , while keeping @xmath125 ( at the gut scale ) .
good fit can be obtained with @xmath126 = 55 , _
e=3.11 , k_e=0.07656 , k_22=3.3677 , _
e=-0.06 , and remaining parameters given in eqs .
( [ inp - tan5 ] ) and ( [ inp - tan55 ] ) respectively . with these
we obtain @xmath127 = m_g , for = 55 , = 2.06510 ^ -4 , = 0.0436 .
these lead to m_e(m_e)= 0.511 mev , m_(m_)=105.66 mev , m_(m_)=1.777 gev , in agreement with experiments .
the mixing angles originating from the charged lepton sector , for @xmath98 , are @xmath128 . while for @xmath108 we got @xmath129 .
note that while @xmath130 is large ( but not sufficiently ) , the @xmath131 and @xmath132 are too small .
this means that neutrino sector should be responsible for generating proper values of the lepton mixing angles .
to build the realistic neutrino sector , we apply the singlet states @xmath133 ( with @xmath1 charges given in table [ t : tab1 ] ) as right handed neutrinos .
their dirac type couplings ( to @xmath134 states ) and the mass matrix respectively are given by : ccc & ccc _ 1 & * 1*_2 & * 1*_3 + m_d c |5_1 + & ( ccc & & | + |^2 & |^2 & |^4 + ) h , ccc & ccc _ 1 & * 1*_2 & * 1*_3 + m_r c * 1*_1 + * 1*_2 + * 1*_3 & ( ccc ^2 & ^2 & 0 + ^2 & ^2 & 1 + 0 & 1 & |^2 ) m _ * , [ nu - coupl ] where @xmath135 is some mass scale and in the entries of these matrices the dimensionless couplings are omitted .
integration of heavy @xmath136 states leads to @xmath137 mass matrix for the light neutrinos : ccc & ccc & & + c + + & m_=m_dm_r^-1m_d^t ( ccc ^2 & |^2 & | + ) |m , [ nu - matrix1 ] with @xmath138 and @xmath139 are some dimensionless couplings .
note that , @xmath140 s @xmath141 block s determinant is zero .
it is convenient to work in a basis where charged lepton mass matrix is diagonal , i.e. rotate whole lepton doublets by unitary matrix which diagonalizes the matrix @xmath142 . in this basis ,
the weak leptonic current is diagonal and the neutrino mass matrix can be denoted by @xmath143 .
the convenience of this basis is that the diagonalizing matrix @xmath144 : u^t|m_u = m_^diag will coincide with the lepton mixing matrix .
the latter , in a standard parametrization , has the form : @xmath145 with @xmath146 and @xmath147 .
the phase matrices @xmath148 are given by : p_1=diage^-i_1 , e^-i_2 , e^-i_3 , p_2=diag1 , e^-i_1/2 , e^-i_2/2 , where @xmath149 are some phases .
to get some feeling about the results , obtained from the neutrino mass matrix , let us first ignore @xmath131 and @xmath132 mixings . since these angles are small , the picture qualitatively will remain unchanged .
( effects of these mixing angles are discussed in detail in an appendix ) . with @xmath150 , in a good approximation @xmath143 can be written as : ccc & ccc & & + c + + & ccc e & c & d + c & b^2 & ab + d & ab & a^2 ) .
[ nu - matrix ] note that @xmath141 block s determinant of the matrix ( [ nu - matrix ] ) is also zero : @xmath151 .
this , using ( [ ulept ] ) , leads to the following interesting relation ^2_13|s_12 ^ 2e^i_1+c_12 ^ 2| . using recent results from the neutrino experiments @xcite
, we can easily verify that the relation of eq .
( [ nu - pred1 ] ) is incompatible with normal hierarchical neutrino masses .
this conclusion remains robust taking into account the effects of @xmath152 and @xmath153 rotations coming from the charged lepton sector . with these , instead of eq .
( [ nu - pred1 ] ) we have the exact expression ^2_13=| @xmath154 is some phase .
derivation of ( [ nu - pred1-exact ] ) and forms of @xmath155 are given in appendix ( see eqs .
( [ i1])-([j ] ) ) .
one can investigate for what values of @xmath156 , desirable values of @xmath157 are obtained . in fig .
[ fig1 ] , region ( i ) corresponds to the values of @xmath156 as a function of @xmath158 , which give @xmath159 . on the other hand , region ( ii )
shows values of @xmath156 obtained within considered scenario ( for @xmath160 and @xmath161 ) .
we see that points of region ( ii ) are well below from points of region ( i ) . while fig .
[ fig1 ] corresponds to the best fit values of the neutrino oscillation parameters @xcite , the conclusion is same by taking them within @xmath162 error bars .
this demonstrates that within considered model , the normal hierarchical neutrino mass scenario can not be realized .
( 11.1,4.25 ) * ( ii ) * ( 10.8,7.5)*(i ) * [ fig1 ] on the other hand , inverted hierarchy in neutrino masses is possible within considered @xmath4 model . ) and @xcite ( within @xmath35 gut ) with inverted hierarchical neutrino masses .
] this is demonstrated in fig .
green dashed region includes points captured by two border bold curves ( obtained via eq . (
[ nu - pred1-exact ] ) within our model ) and two horizontal lines ( corresponding to the experimental values of @xmath157 within @xmath163 ) .
this figure corresponds to the best fit values of @xmath164 and @xmath165 , while free phases ( see eqs .
( [ * ] ) , ( [ j ] ) ) are varied within full ranges .
dashed lines correspond to the case with @xmath166 .
thus , inclusion of the charged lepton sector somewhat extents the allowed region .
all this demonstrates that inverted hierarchical scenario is easily realized .
[ fig2 ] shows that , the allowed region for @xmath158 is fixed as : 0.0008 evm_30.0044 ev , and using ( [ inv - masses ] ) we get : @xmath167 m_20.0493 ev1+^2 ^1/2 .
( 9.9,6 ) * ( b ) * ( 6.8,9.9)*(a ) * [ fig2 ] these imply @xmath168 ev , satisfying the current bound @xcite obtained from cosmology .
moreover , for neutrino less double @xmath169-decay parameter @xmath170 we obtain : m_|c_12 ^ 2m_1+s_12 ^ 2m_2e^-i_1 | , leading to : 0.011 evm_0.05 ev .
future experiments will be able to test viability of this scenario @xcite . in summary
, we have presented supersymmetric @xmath0 gut supplemented with non - anomalous @xmath1 flavor symmetry .
anomaly cancellation condition restricted the field content ( dictated three families of @xmath2 matter ) , as well as @xmath1 charge assignment .
texture zero yukawa matrices gave natural understanding of hierarchies between charged fermion mass and mixings .
model automatically involves @xmath0 singlet states utilized as right handed neutrinos .
inverted hierarchical neutrino mass scenario is predicted within considered model .
other phenomenological issues , such as doublet - triplet splitting , proton decay etc . , left beyond the scope of this paper , will be addressed elsewhere within more general class of models @xcite supplemented by anomaly free @xmath1 symmetry @xcite .
i am grateful to c.h . albright and k.s .
babu for useful discussions and comments .
i thank the center for theoretical underground physics and related areas ( cetup * 2012 ) for its hospitality and for partial support in lead , south dakota , where part of this work was done .
the partial support from shota rustaveli national science foundation ( contract # 03/79 ) is kindly acknowledged .
in this appendix we work out the details of contributions from the charged lepton sector to the neutrino sector . in particular , as was pointed out in sect .
[ nu - sect ] , we study impact of @xmath131 and @xmath132 mixing angles . with this study , we prove that within considered susy @xmath4 scenario , only inverted hierarchical neutrino mass scenario is realized
. charged lepton mass terms @xmath172 get diagonalized by transformations @xmath173 , where @xmath174 are unitary matrices such that l_e^m_er_e = m_e^diag .
let us rotate the neutrino states @xmath175 by the same unitary transformation as @xmath176-states : @xmath177 . with this ,
the weak current remains diagonal : @xmath178 . on the other hand , the neutrino mass couplings @xmath179
become @xmath180 with |m_=l_e^m_l_e^ * . upon transformation @xmath181
, the neutrino couplings can be diagonalized , i.e. u^t|m_u = m_^diag , and finally the weak current will be @xmath182 .
thus , the matrix @xmath144 in ( [ mnu - diag ] ) coincides with the lepton mixing matrix . from eqs .
( [ barmnu ] ) and ( [ mnu - diag ] ) we obtain l_e^m_l_e^*=u^*m_^diagu^ .
unitary matrix @xmath183 can be written as l_e = p_1^ll_23l_13l_12p_2^l , where @xmath184 are some diagonal phase matrices , @xmath185 , @xmath186 and @xmath187 correspond to the rotation angles @xmath131 , @xmath130 and @xmath132 respectively . without
loss of generality , the unitary matrices @xmath185 and @xmath186 can be taken to be real orthogonal matrices . since within our scenario @xmath131 and @xmath132 are small , we can write l_13l_12p_l(1+)p_l , where @xmath188 is real : = ( ) , with s_12^e_12^e , s_13^e_13^e .
[ ga ] without restricting any generality , we can take @xmath189 and using ( [ le ] ) , ( [ l13l13-ga ] ) in ( [ lmnu - umnudiag ] ) , we obtain ( 1+^t)m_^*(1+)=um_^diagu^t , with m_=p_l^*l_23^p_1^l*m_p_1^*l_23^*p_l^ * . because of smallness of @xmath190 and @xmath191 , further we use approximation and keep first powers of these angles ( and thus first powers of the matrix @xmath192 ) . with this , from ( [ f ] ) we get m_^*+^tm_^*+m_^*=t , with tum_^diagu^t .
note , that since @xmath141 block s determinant of @xmath193 is zero , similar applies to the @xmath141 block of the matrix @xmath194 ( see eq .
( [ f1 ] ) ) .
thus , @xmath195 can be parameterized as m_^ * = ( ) .
[ hatm - param ] with this , using matrix relation in eq .
( [ g ] ) , we derive @xmath196 @xmath197 ^2 + 2ds_13^e = t_33 , ab+cs_13^e+ds_12^e = t_23=t_32 . by iteration
( keeping @xmath198 and @xmath199 ) we obtain from ( [ h ] ) : @xmath200 ab = t_23-s_12^et_13-s_13^et_12 .
these three expressions , by eliminating @xmath201 and @xmath202 , give the following relation : t_23 ^ 2-t_22t_33=2s_12^e(t_23t_13-t_12t_33)+2s_13^e(t_23t_12-t_22t_13 ) .
substituting @xmath203 elements ( see eq . ( [ g ] ) ) in ( [ h11 ] )
we obtain m_1m_2(u_21u_32-u_22u_31)^2 + m_1m_3(u_21u_33-u_23u_31)^2+m_2m_3(u_22u_33-u_23u_32)^2 = -2s_12^ek_1 - 2s_13^ek_2 with @xmath204 k_2=m_1m_22_13c_23e^i(+_2)- ( m_1e^i_1-m_2)m_32_12s_23 e^-i(_1+_2+_1 + 2_2+_3 ) . using the form of @xmath144 of eq .
( [ ulept ] ) in left hand side of ( [ i ] ) , after some simplifications we obtain -^2_13e^i(2+_2)=s_12 ^ 2e^i_1+c_12 ^ 2 + s_12^ek_1+s_13^ek_2e^i(_1+_2 + 2_2 + 2_3 ) . introducing notations @xmath205 = _ 1+_2 + 2_2 + 2_2+args_12^ek_1+s_13^ek_2-argm_1s_12 ^ 2e^i_1+m_2c_12 ^ 2 , from ( [ * ] ) we get eq .
( [ nu - pred1-exact ] ) - the expression for @xmath206 .
let us first see if normal hierarchical neutrino masses are possible . in this case , @xmath207 and observed mass squire differences are @xmath208 and @xmath209 .
thus , two masses , say @xmath210 and @xmath211 , can be expressed as m_1= , m_2= . while @xmath208 , @xmath209 are both measured , the @xmath158 is unknown yet and we will treat it as a free parameter . taking into account ( [ i ] ) , we can easily verify that @xmath212 is excluded .
thus , @xmath213 ev . on the other hand , we can also have an upper bound for @xmath158 , set from the cosmological bound on a sum of three neutrino masses @xmath214 ev @xcite .
this , taking into account ( [ norm - masses ] ) , gives @xmath215 ev .
therefore , we will vary @xmath158 in a range 0.05 evm_30.34 ev . with help of ( [ nu - pred1-exact ] ) and using the best fit values of quantities @xmath216 , @xmath165 and neutrino mixing angles @xcite ,
we can see what values of @xmath156 are needed . in fig .
[ fig1 ] , dashed region ( i ) represents such values of @xmath156 ( versus @xmath158 ) . for fixed value of @xmath158 ,
the multiple values of @xmath156 are obtained because of free phases appearing in ( [ nu - pred1-exact ] ) . on the other hand , with ( [ j ] ) and ( [ i1 ] ) within our model with @xmath217 , @xmath218
we can calculate @xmath156 for different @xmath158 and remaining phases .
region ( ii ) of fig .
[ fig1 ] corresponds to this .
we see that regions ( i ) and ( ii ) do not overlap and therefore conclude that normal hierarchical neutrino mass scenario is not realized within considered model .
this conclusion remains robust even varying the values of @xmath216 , @xmath165 , @xmath219 within @xmath162 error bars . as turns out
, the inverted hierarchical neutrino masses blend well with relation ( [ nu - pred1-exact ] ) . in this case ,
@xmath208 and @xmath220 and thus : m_1= , m_2= .
if we set @xmath166 in ( [ nu - pred1-exact ] ) , we can easily see that for certain values of @xmath158 and @xmath221 all observable can be obtained within experimentally preferred ranges .
inclusion of @xmath156 do not change this positive result , but just offers slightly different choices of @xmath158 and various phases . for
illustration see fig .
[ fig2 ] , with corresponding discussion starting in a paragraph right before eq .
( [ predictions ] ) .
g. l. fogli , e. lisi , a. marrone , d. montanino , a. palazzo and a. m. rotunno , phys .
d * 86 * , 013012 ( 2012 ) .
f. p. an _ et al . _
[ daya - bay collaboration ] , phys .
lett . * 108 * , 171803 ( 2012 ) .
t. schwetz , m. tortola and j. w. f. valle , new j. phys . * 13 * , 109401 ( 2011 ) .
s. dimopoulos , s. raby and f. wilczek , phys .
d * 24 * , 1681 ( 1981 ) ; + w. j. marciano and g. senjanovic , phys . rev .
d * 25 * , 3092 ( 1982 ) ; + c. giunti , c. w. kim and u. w. lee , mod .
lett . a * 6
* , 1745 ( 1991 ) ; + p. langacker and m. -x .
luo , phys .
d * 44 * , 817 ( 1991 ) ; + u. amaldi , w. de boer and h. furstenau , phys .
b * 260 * , 447 ( 1991 ) .
m. dine , n. seiberg and e. witten , nucl .
b * 289 * , 589 ( 1987 ) ; + j. j. atick , l. j. dixon and a. sen , nucl . phys .
b * 292 * , 109 ( 1987 ) ; + m. dine , i. ichinose and n. seiberg , nucl .
b * 293 * , 253 ( 1987 ) .
e. dudas , s. pokorski and c. a. savoy , phys .
b * 356 * , 45 ( 1995 ) ; + m. -c .
chen , d. r. t. jones , a. rajaraman and h. -b .
yu , phys .
d * 78 * , 015019 ( 2008 ) ; + k. s. babu , m. frank and s. k. rai , phys .
lett . * 107 * , 061802 ( 2011 ) . for anomaly calculus within various groups
see : + j. banks and h. georgi , phys .
d * 14 * , 1159 ( 1976 ) ; + s. okubo , phys .
d * 16 * , 3528 ( 1977 ) .
g. -b .
zhao , s. saito , w. j. percival , a. j. ross , f. montesano , m. viel , d. p. schneider and d. j. ernst _ et al .
_ , arxiv:1211.3741 [ astro-ph.co ] ; + m. c. gonzalez - garcia , m. maltoni and j. salvado , jhep * 1008 * , 117 ( 2010 ) ; + m. moresco , l. verde , l. pozzetti , r. jimenez and a. cimatti , jcap * 1207 * , 053 ( 2012 ) . | supersymmetric @xmath0 gut augmented with anomaly free @xmath1 flavor symmetry is presented .
very economical field content and @xmath1 charge assignment are obtained by specific construction . in particular , three families of @xmath2 chiral matter , along the @xmath0 singlet states ( some of which serve as right handed neutrinos ) are obtained .
appealing texture zero yukawa matrices provide natural understanding of hierarchies between charged fermion masses and mixings .
the model predicts inverted hierarchical neutrino mass scenario with interesting implications .
cetup*12 - 019 + march 6 , 2013 + * * three family @xmath0 gut and + inverted neutrino mass hierarchy zurab tavartkiladze _ center for elementary particle physics , itp , ilia state university , 0162 tbilisi , georgia _ |
as the strongest spectral feature of the most abundant element in the universe , the ly@xmath0 transition of hydrogen ( @xmath7 ; 1216 ) has become one of the most widely utilized tools for detecting galaxies in the early universe ( e.g. , @xcite ) .
ly@xmath0 is versatile in that it can probe the extremes of the high - redshift galaxy census , such as faint , low - mass populations ( e.g. , @xcite ) and the high energy systems at the top - end of the mass function ( e.g. , @xcite ) .
ly@xmath0 photons are produced as a result of excitation or ionization of neutral hydrogen and can arise in a variety of astrophysical scenarios relating to high - redshift galaxies .
these include cooling radiation as a result of gravitational collapse ( e.g. , @xcite ) or extremely rapid gas motions ( i.e. , shocks ; @xcite ) , photo - ionization by stars ( e.g. , @xcite ) or an active galactic nucleus ( agn ; e.g. , @xcite ) , or fluorescence of neutral gas in the intergalactic medium ( igm ; e.g. , @xcite ) . however , the interpretation of the emergent ly@xmath0 spectra is non - trivial due to the resonant nature of the transition in neutral hydrogen , in which the line is optically thick for even small neutral fractions @xcite .
this property makes the escape of ly@xmath0 photons from a galaxy a complex radiative process that is highly dependent on the properties of the interstellar medium ( ism ) and surrounding circumgalactic medium , such as gas kinematics , covering fraction , geometry , and dust content .
the result of ly@xmath0 radiative transfer through the ism is a significant modulation of the intrinsic ly@xmath0 emission line profile , both spatially and spectrally .
the emergent emission line profile can be further modified by transfer through the surrounding igm , which can absorb a significant amount of radiation in the vicinity and blueward of ly@xmath0 ( e.g. , @xcite ) .
because of the amount of physics that can be encoded within the emergent ly@xmath0 line profile , a great deal of theoretical work has gone into understanding ly@xmath0 radiative transfer to aid in the interpretation of observed ly@xmath0 spectra of galaxies . in particular , the use of monte carlo numerical techniques has become standard practice ( e.g. , @xcite ) , and has provided useful predictions of the emergent ly@xmath0 line profile for comparison with observations for a variety of different physical scenarios .
the implementation of such radiative transfer codes are currently being advanced by post - processing more realistic models of galaxies drawn from hydrodynamic and cosmological simulations ( e.g. , @xcite ) .
non - active galaxies that are selected by virtue of their ly@xmath0 emission with equivalent width @xmath8 ( i.e. , ly@xmath0 emitters ; laes ) are typically low - mass ( @xmath9 ) with young stellar populations ( @xmath10 myr ) and are forming stars at rates of @xmath11 yr@xmath4 ( e.g. , @xcite ) .
in addition , some laes appear to have non - negligible dust content @xcite , low gas - phase metallicity @xcite , and compact , irregular spatial morphologies ( e.g. , @xcite ) .
at least some of these properties appear to not evolve significantly with redshift ( e.g. , @xcite ) . a significant fraction of ultraviolet ( uv ) continuum - selected galaxies ( i.e. , lyman break galaxies ; lbgs ) also show ly@xmath0 in emission ( e.g. , @xcite ) .
lbgs typically have lower ( some show ly@xmath0 in absorption ) , more evolved stellar populations , and are more massive than the typical ly@xmath0-selected galaxy ( e.g. , @xcite ) . in a star - forming galaxy , ly@xmath0
photons are produced from recombinations in hydrogen gas that was ionized by the uv radiation of massive , main sequence stars in regions . unlike uv continuum or photons from other optically thin transitions that hail from the same region ( e.g. , h@xmath0 ) , ly@xmath0 photons become resonantly trapped in the first parcels of neutral hydrogen they encounter . as compared to the optically thin photons , the path length of ly@xmath0 to escape the galaxy
is increased due to the scattering , making it especially sensitive to dust absorption .
the path length of ly@xmath0 photons through neutral hydrogen can be further altered by local velocity fields due to thermal motion , turbulence , and bulk motions ( e.g. , due to supernovae or stellar driven winds , or by galaxy - galaxy interactions ) , thereby shifting ly@xmath0 photons in and out of resonance and affecting the emergent line profile ( e.g. , @xcite ) .
this potentially makes the emergent ly@xmath0 spectrum a very powerful probe of gas kinematics ( e.g. , @xcite ) , especially when the spectrum can be spatially resolved ( e.g. , @xcite ) .
evidence for large - scale outflows is wide - spread in high - redshift star forming galaxies ( e.g. , @xcite ) . in lbgs , for example , the galaxy s redshift differs when measured from the interstellar ( is ) absorption lines versus the ly@xmath0 emission line ( by @xmath12 km s@xmath4 ) , which implies that one or both features are not at rest with respect to the stars in the galaxy @xcite .
systemic redshifts and gravitationally induced motion are expected to be well represented by nebular emission lines such as h@xmath0 , which is optically thin and whose strength depends upon the uv radiation field in the same vicinity as the origin of the ly@xmath0 photons @xcite . in lbgs ,
the is absorption lines are blueshifted with respect to the h@xmath0-based systemic redshift ( by @xmath13 km s@xmath4 ) while the peak of the ly@xmath0 emission line is typically redshifted ( by @xmath14 km s@xmath4 ; @xcite ) .
although the interpretation of the redshifted ly@xmath0 emission is non - trivial due to the transition s resonant scattering nature , the blueshifted is absorption is a strong indicator of a galactic outflow and is caused by the absorption of stellar light in swept - up material that is approaching along the observer s line of sight .
cccccl r.a . &
( j2000 ) & 10:00:14.18 & 10:00:28.33 & 10:00:27.23 & a11 + decl . & ( j2000 ) & + 02:14:26.11 & + 02:17:58.44 & + 02:17:31.50 & + @xmath15 & ( - ) & @xmath16 & @xmath17 & @xmath18 & a11 + @xmath19 & ( - ) & @xmath20 & @xmath21 & @xmath22 & f11 , s13 + @xmath23 & ( km s@xmath4 ) & @xmath24 & @xmath25 & @xmath26 & a11,f11 , s13 + @xmath27 & ( @xmath28 ergs s@xmath4 @xmath29 ) & @xmath30 & @xmath31 & @xmath32 & a11 + @xmath33 & ( @xmath34 ergs s@xmath4 ) & @xmath35 & @xmath36 & @xmath37 & + ew@xmath38 & ( ) & @xmath39 & @xmath40 & @xmath41 & b11 + @xmath42 & ( km s@xmath4 ) & @xmath43 & @xmath44 & @xmath45 & f11 , s13 + @xmath46 & ( mag ) & @xmath47 & @xmath48 & @xmath49 & cosmos + @xmath50 & ( mag ) & @xmath51 & @xmath52 & @xmath53 & b11 + sfr(h@xmath0 ) & ( yr@xmath4 ) & @xmath54 & @xmath55 & @xmath56 & f11 , s13 + @xmath57 & ( - ) & @xmath58 & @xmath59 & @xmath60 & f11 , s13 + stellar mass & ( @xmath61 ) & @xmath62 & @xmath63 & @xmath64 & f11 , s13 + recently , observations of rest - frame optical nebular emission lines ( e.g. , h@xmath65 , [ o iii ] , or h@xmath0 , depending on the redshift of the galaxy ) for @xmath66 laes observed in the near - infrared ( nir ) have become more common .
these data provide not only standard rest - frame optical emission line diagnostics for these high - redshift galaxies ( e.g. , @xcite ; song et al .
2013 , in preparation ) , but also a measure of their systemic redshifts @xcite .
thus , ly@xmath0 velocity offsets are now be measured for laes much in the same way as for lbgs ( e.g. , @xcite ) , which provides an additional constraint when comparing ly@xmath0 radiative transfer models to the observed ly@xmath0 spectra of laes .
however , velocity offsets of ly@xmath0 alone can not necessarily constrain the gas kinematics and may not correlate directly with bulk gas velocities .
useful kinematic information can be encoded within the offset ly@xmath0 line profile in the form of asymmetries and other finer - scale spectral features , such as multiple emission peaks .
this information can potentially be lost when observed at low spectral resolution , as has been used in multiple studies that probe ly@xmath0 velocity offsets ( e.g. , @xcite ; song et al . 2013 , in preparation ) .
multiple peaks are a natural consequence of ly@xmath0 resonant line transfer ( e.g. , @xcite ) , and the frequency of multiple - peaked ly@xmath0 spectra in lae and lbg samples has recently been studied by @xcite and @xcite , respectively .
both studies find that multiple - peaked ly@xmath0 line profiles comprise @xmath6750% of their respective samples , which should be regarded as lower limits due to the limited @xmath68 and spectral resolution of the ly@xmath0 data .
combined with the recent observational evidence that shows that laes have smaller ly@xmath0 velocity offsets as compared to lbgs @xcite , the spectral substructure that studies such as @xcite and @xcite have uncovered motivates observations of ly@xmath0 at higher spectral resolution . such observations will probe the ism and circumgalactic medium through which the ly@xmath0 photons traverse , and provide more stringent constraints on existing ly@xmath0 radiative transfer models . in this paper , we present new higher resolution ( 120 km s@xmath4 fwhm ) optical spectra of the ly@xmath0 emission of three galaxies that were initially discovered in the hetdex pilot survey ( hps ; @xcite ) .
we discuss these data in @xmath69[sec : data ] . in @xmath69[sec :
results ] , we present basic observational results from our optical spectroscopy and compare the three ly@xmath0 emission line profiles with those of other ly@xmath0 and uv continuum - selected samples . in @xmath69[sec : models ] , we present a quantitative comparison of the observed ly@xmath0 line profiles with the predictions of existing ly@xmath0 radiative transfer models with a spherical expanding shell gas geometry . in @xmath69[sec : discussion ] , we discuss our findings and finally present our conclusions in @xmath69[sec : conclusions ] . throughout this paper ,
we assume a flat @xmath70cdm cosmology ( @xmath71 km s@xmath4 mpc@xmath4 , @xmath72 , @xmath73 ; @xcite ) . for the atomic transitions discussed , we use the following vacuum wavelengths in quoted from the atomic line list v2.04 : ly@xmath0 ( 1215.670 ) ; [ o iii ] @xmath74 ( 5008.240 ) ; h@xmath0 ( 6564.610 ) .
all magnitudes are reported in the ab system @xcite .
the hobby - eberly telescope dark energy experiment ( hetdex ; @xcite ) will use 0.8 million laes as tracers of the galaxy power spectrum and measure the evolution of dark energy from @xmath75 with high precision .
this @xmath76420 deg@xmath77 ( @xmath769 gpc@xmath78 ) blind spectroscopic survey will be conducted with the hobby - eberly telescope and a revolutionary new multiplexed instrument called the visible integral - field replicable unit spectrograph ( virus ; @xcite ) . as a test - bed for hetdex
, a single prototype virus unit integral field spectrograph ( the mitchell spectrograph ; formerly known as virus - p ; @xcite ) has resided at the mcdonald observatory s harlan j. smith 2.7 m telescope since 2007 and has conducted a @xmath76100 night , @xmath76169 arcmin@xmath77 ( @xmath7610@xmath79 mpc@xmath78 ) pilot survey for hetdex in which 99 star - forming ( i.e. , non - agn ) laes were discovered at redshifts @xmath80 .
the survey design , catalog , and initial science results can be found in @xcite and @xcite . as described in @xmath69[sec : introduction ] , observations of the rest - frame optical nebular emission lines are especially important for constraining models of ly@xmath0 radiative transfer and gas kinematics in star - forming galaxies . given the aforementioned focus of this paper , we require prior knowledge of the galaxy systemic redshift as measured from the rest - frame optical nebular emission lines . for two hps laes ( catalog i d 194 and 256 )
, these data have been published in @xcite .
additionally , the only existing optical spectra previous to this work for these galaxies are from the hps in which the mitchell spectrograph observed with 5.5 resolution ( @xmath76400 km s@xmath4 fwhm at 4100 ) , yielding unresolved ly@xmath0 emission lines @xcite . given the existing data available for these two galaxies , we have chosen to conduct follow - up observations with higher resolution optical spectroscopy .
for convenience , some relevant properties of hps194 and hps256 are tabulated from the current hps publications in table [ table : galdata ] .
the only preselection for the galaxies in this follow - up study is that they have a ly@xmath0 flux
@xmath81 erg s@xmath4 @xmath29 to ease the detection of the rest - frame optical lines ( note that 86 of the 99 non - agn hps laes meet such a selection criteria ) .
additionally , note that the galaxies included in this study are bright enough in the continuum to be considered lbgs in most surveys ( e.g. , @xcite ) .
we obtained follow - up optical spectra of the ly@xmath0 emission of hps194 and hps256 using the mitchell spectrograph @xcite on the harlan j. smith 2.7 m telescope at the mcdonald observatory on the nights of ut 2010 february 13 and 14 , respectively . in an attempt to spectrally resolve the ly@xmath0 feature of these galaxies , we replaced the 831 line mm@xmath4 grating that was used to carry out the hps with a 2400 line mm@xmath4 grating .
this grating yields a spectral resolution of @xmath82 ( 120 km s@xmath4 fwhm at 4100 ) over the wavelength range @xmath83 ( ) @xmath84 ( @xmath85 ) . [ cols="^ " , ] many studies have used the results of ly@xmath0 radiative transfer through expanding shells to help explain the observed velocity offsets of the ly@xmath0 emission line from @xmath19 for star - forming galaxies and further to constrain the velocity of their large - scale outflows ( e.g. , @xcite ; song et al .
2013 , in preparation ) .
the analysis in @xmath69[subsec : modelcompare ] of the spectrally resolved , multiple - peaked ly@xmath0 emission of the three laes we present shows that it is difficult to constrain some of the physical parameters of each galaxy s outflowing gas .
this is especially true since even the best - fitting expanding shell radiative transfer models show major discrepancies when compared to the resolved ly@xmath0 spectra .
given these discrepancies , it is of note that and @xmath87 appear to be much better constrained than . for each galaxy ,
the best - fitting model has @xmath88 @xmath29 and a high value of @xmath87 as a result of their strikingly similar values of @xmath89 and @xmath90 .
in addition , as can be seen in figure [ fig : exmodelspec ] , there is little change in the position of the dominant redward ly@xmath0 peak over the range @xmath91 ( km s@xmath4 ) @xmath92 for an expanding shell model with low neutral column density ( i.e. , @xmath93 @xmath29 ; see @xmath69[subsec : gridtrends ] ) . especially since this strong redward peak dominates the observed ly@xmath0 line profile at low spectral resolution and/or @xmath68
, it appears that has the least influence on @xmath23 of the three physical parameters that we model for a galaxy with low neutral gas column density .
several of the aforementioned studies ( @xcite ; song et al .
2013 , in preparation ) do in fact utilize a lower spectral resolution that is similar to the hps when observing ly@xmath0 such that objects with line profiles similar to those we present here are unresolved ( see figure [ fig : vspec ] ) . in cases for objects with multiple - peaked ly@xmath0 line profiles with @xmath94 less than the instrumental resolution , much of the information that is encoded in the multiple - peaked profile is lost and the only measurable quantity from the line profile becomes @xmath15 ( or @xmath23 when a measure of @xmath19 exists ) . we have also previously noted the instrumental effect of observing an intrinsically asymmetric emission line at low resolution in which the peak flux can can be biased in the direction of the extended emission tail ( see @xmath69[subsubsec : lyaquantitative ] ) .
as a result of the loss of information , the measured @xmath23 for an unresolved line becomes a convoluted function of many parameters , including the line of sight velocity field , column density , and the temperature of the neutral gas in addition to parameters that were not explored in our models , such as dust content .
even when ignoring the possible effects of the igm on the ly@xmath0 line profile ( e.g. , @xcite ) , the matter is further complicated when considering ly@xmath0 radiative transfer through more realistic models of the gas distribution and velocity fields for galaxies ( see the previous subsection ) .
these models , which do not have spherically symmetric gas distributions , indicate that viewing angle should also be an important parameter in determining the degree of flux asymmetry between the multiple peaks , which can affect the value of the measured @xmath23 . as a result of all of these factors , attempts to equate the velocity offsets from systemic that are observed in ly@xmath0 spectra to physical outflow speeds need to be treated with extreme caution .
we also stress the importance of observing ly@xmath0 with high spectral resolution in order to extract the maximum amount of information from the line profile .
we do point out , however , that while the _ magnitude _ of the inflow or outflow speed is not yet clearly a measurable quantity from the ly@xmath0 line profile even when spectrally resolved , the detection and sign of a non - zero @xmath23 does at least indicate presence of an outflow ( neglecting igm absorption ) or potentially inflow ( for blueshifted ly@xmath0 emission ) along the line of sight . in figure
[ fig : cosmosthumbs ] , we show candels _ hst _ acs / wfc f606w and wfc3/ir f160w images of the three laes , which probe the rest - frame uv ( @xmath761800 ; this wavelength traces recent star formation through the continuum emission of young , massive stars ) and rest - frame optical ( @xmath764700 ) , respectively . as expected from near - uv _ hst _ morphological studies of laes at similar redshift ( e.g. , @xcite ) , the galaxies are compact with half - light radii @xmath95 1.6 , 1.1 , and 0.7 kpc for hps194 , hps256 , and hps251 , respectively , at @xmath762400 @xcite . at the spatial resolution of the _ hst
_ , two of the galaxies ( hps194 and hps251 ) are shown to have a companion continuum source at @xmath96 distance : _ hps194 : _ this galaxy consists of a compact source to the north - west with a possible tidal tail reaching towards the north ( this is visible as small clumps in the rest - frame uv image and is a very pronounced continuous feature in the rest - frame optical image ) .
a more diffuse companion is located @xmath97 to the south - west in projection ( @xmath98 kpc if the sources are at the same redshift ) .
as seen in figure [ fig : pointingthumbs ] , the nirspec slit used by @xcite is aligned along the two sources .
however , we were unable to detect h@xmath0 emission from the fainter source .
as will be presented by blanc et al .
2013 ( in preparation ) , we have also obtained deeper nir spectra of hps194 with the fire instrument @xcite at the 6.5 m magellan baade telescope . a preliminary reduction of the data yields no detection of the diffuse companion s rest - frame optical emission lines .
although we can not confirm from the available data if the compact and more diffuse sources lie at the same redshift , the disturbed , asymmetric morphology and possible tidal feature emanating from the north - west compact source is suggestive of an ongoing merger . _
hps251 : _ this galaxy has a fainter , compact companion located @xmath99 ( @xmath100 kpc ) to the south - south - west in projection . like hps194 , the nirspec slit used by song et al .
( 2013 , in preparation ) is aligned along the two sources .
unfortunately , two of the positions of the _ abba _ dither pattern positioned the dimmer source on the bottom edge of the slit , yielding insufficient @xmath68 in the final 2-d spectrum at the position of the companion .
however , recent follow - up observations using fire have confirmed detections of [ o iii ] @xmath101 and [ o iii ] @xmath74 for both sources at the same redshift ( these results will be presented by blanc et al .
2013 , in preparation ) . at the projected distance of @xmath100 kpc ,
the two components of hps251 are clearly interacting and will eventually merge .
while hps256 does not have a companion within @xmath102 , it does have several nearby continuum sources within @xmath103 to the south - east in projection ( @xmath104 kpc if the sources are at the same redshift ; see figure [ fig : pointingthumbs ] ) .
however , these sources are outside of the mitchell spectrograph s fiber and the nirspec slit used by @xcite was not aligned to include them .
@xcite explored the ly@xmath0 properties of close lbg pairs and found that all showed ly@xmath0 in emission when the projected separation was @xmath105 kpc .
their work supports the picture of galaxy - galaxy interactions triggering star - formation and clearing gas and dust sufficiently for the ly@xmath0 photons produced in the starburst to escape the galaxies with high .
@xcite also find that while the 1-d spectra of the close lbg pairs are often double - peaked , they are resolved spatially into two distinct offset ly@xmath0 lines and corresponding continua in 2-d spectra . for hps194 and hps251
, this may be an alternative explanation for the multiple - peaked ly@xmath0 emission and why the expanding shell models poorly represent the data .
however , our fiber - based ly@xmath0 spectra are spatially unresolved on the sky , so we can not investigate potential spatial offsets of the spectrally resolved ly@xmath0 emission components for these galaxies with the current data .
as seen in figure [ fig : pointingthumbs ] , the potential companion sources for hps256 and the confirmed companion to hps251 are located outside of the mitchell spectrograph s fiber , even when considering the rms pointing uncertainty .
yet , we still observe the multiple - peaked ly@xmath0 line profile morphology .
this makes it more likely that the multiple - peaked and asymmetric nature of the ly@xmath0 line for these galaxies is a consequence of radiative transfer effects in a non - static medium rather than being the result of the integrated ly@xmath0 emission from multiple sources at similar redshift .
since all three galaxies have at least one nearby continuum source within a projected distance of @xmath106 kpc and one of our galaxies has a _ confirmed _ companion within @xmath100 kpc , a possible link between galaxy - galaxy interactions and luminous ly@xmath0 emission should not be ignored .
this has recently been explored in the context of labs by @xcite who post - process hydrodynamical simulations of gas - rich binary major mergers with a 3-d radiative transfer code .
these authors find that such mergers produce copious ly@xmath0 emission ( with ly@xmath0 luminosity @xmath107 erg s@xmath4 ) that is extended over 20 - 50 kpc at @xmath108 as a result of shocked gas and the starburst induced by the gravitational interaction .
these properties are similar to typical @xmath108 labs ( e.g. , @xcite ) . due to the observational and selection methods utilized in the hps @xcite , the @xmath33 observed for our three laes ( @xmath109 erg s@xmath4 ; see table [ table : galdata ] ) is of the same order as that predicted by the @xcite simulations .
additionally , the _ total _ velocity widths of the ly@xmath0 emission for hps194 , hps256 , and hps251 are 667 , 612 , and 884 km s@xmath4 , respectively @xcite , which are each comparable to the median value of 780 km s@xmath4 for the labs of @xcite .
however , since the ly@xmath0 spectra we present in this work supply no spatial information , we can not currently assess the extended nature of the three laes individually .
the hps data is also of limited use in this regard due to the coarse spatial resolution and limited depth of the survey ( an upper limit of @xmath110 fwhm , which corresponds to the spatial resolution limit of the hps , can be placed on each lae s size ; @xcite ) .
future deep , spatially resolved ly@xmath0 spectra will be useful in determining if our laes are significantly extended , which would provide further evidence for the interaction scenario based on the simulations by @xcite .
recently , @xcite used measurements of the sfr of bright laes ( hps194 and hps256 were among the galaxies in their sample ) along with measurements of their dynamical mass ( which were assumed to be the upper limit of their gas mass ) and sizes to compare such galaxies to existing star formation scaling relations .
@xcite established that there are two distinct sequences of star formation in the kennicutt - schmidt @xmath111 vs. gas mass density @xmath112 plane : a temporally extended sequence for `` normal '' star - forming disk galaxies , and a more rapid sequence for starburst galaxies that was determined from measurements of sub - millimeter and ultraluminous infrared galaxies ( smgs and ulirgs , respectively ) .
the latter scaling relation is offset above the former by 1.1 dex in the @xmath111 vs. @xmath112 plane .
this bimodality between the star formation sequences of starburst and normal star - forming galaxies is largely due to the use of a bimodal conversion factor @xmath113 used to convert the observed carbon monoxide line luminosity @xmath114 to gas mass @xmath115 ( @xmath116 for normal star - forming galaxies and 0.8 for starburst galaxies ) .
the measurements by @xcite suggest that laes lie above the scaling relation for normal star - forming galaxies and are consistent with the @xcite starburst sequence at the @xmath117 level .
thus , laes appear to form stars more rapidly than a typical star - forming galaxy at a given @xmath112 .
note that the bimodality of @xmath118 has recently been challenged by the simulations of @xcite .
these authors show that while @xmath118 is systematically different for varying local galaxy conditions ( e.g. , metallicity and surface density ) , it varies smoothly among them .
the result is a continuous , unimodal star formation scaling relation where the highest @xmath111 disk galaxies overlap with the inferred mergers ( i.e. , starburst galaxies ) .
the measurements for laes by @xcite show that laes occupy a region in the @xmath111 vs. @xmath112 plane bounded by @xmath119 and @xmath120 .
data points within this bounded area are consistent with the scatter in the continuous star formation scaling relation of @xcite and lie in a region of elevated sfr density occupied by both high-@xmath121 disk galaxies and inferred mergers ( i.e. , low-@xmath121 ulirgs and high-@xmath121 smgs ) .
thus , the star formation observed for luminous laes is consistent with ( but not necessarily suggestive of ) that expected from an interacting or merging system . in this scenario , interactions could be responsible for dispersing gas and dust and allowing ly@xmath0 photons to escape through low column density ( i.e. , low optical depth ) `` windows '' in the overall neutral gas distribution . combined with a low dust optical depth ,
the gravitationally induced burst of star formation could result in the large we observe in addition to driving a strong large - scale outflow .
ly@xmath0 radiative transfer through the dispersed ( i.e. , lower ) outflowing gas could give rise to the asymmetric , multiple - peaked ly@xmath0 line profiles with small ly@xmath0 velocity offsets , as compared to the ly@xmath0 spectra of lbgs that typically have large velocity offsets and smaller .
interaction - induced inflows of gas ( resulting in star formation or shocks ) along the line of sight could also help to enhance the ly@xmath0 flux blueward of the line center .
additionally , multiple blueward ly@xmath0 peaks ( such as that observed for hps256 and hps251 ) could also be produced from shocked and/or fluorescing gas that is infalling along different sight lines .
the large blueshift ( @xmath6 km s@xmath4 ) of these ly@xmath0 peaks , however , likely indicates that such emitting material is not yet bound , unless radiative transfer effects couple favorably with the dynamics of the system to produce such large velocity offsets relative to the systemic velocity of the lae s regions . to verify an interaction - based scenario for luminous laes , a larger sample of galaxies with confirmed redshifts for nearby companion continuum sources and spatially resolved ly@xmath0 spectra are needed .
the irregular nature of these galaxies continuum morphologies and the confirmed and potential interactions with nearby companions may suggest that significant deviations away from ly@xmath0 point sources and spherical gas distributions and velocity fields are necessary for properly modeling ly@xmath0 radiative transfer through neutral gas on the galactic and circumgalactic scale .
this conjecture is especially intriguing given that the two galaxies that are most poorly represented by the spherical expanding shell model ly@xmath0 spectra ( i.e. , hps194 and hps251 ) are the same two galaxies with the strongest evidence of an interaction .
in this paper , we have obtained follow - up optical spectra of three laes drawn from the hetdex pilot survey @xcite with sufficient spectral resolution to resolve the ly@xmath0 emission line . with no preselection other than @xmath122 erg s@xmath4 @xmath29 ( which corresponds to @xmath123 erg s@xmath4 at @xmath124 )
, we find that all three galaxies at 120 km s@xmath4 fwhm spectral resolution display multiple ly@xmath0 emission peaks . using the nir spectra of these galaxies rest - frame optical emission lines from @xcite and song et al .
( 2013 , in preparation ) , we have determined the velocity structure of the ly@xmath0 emission relative to the systemic redshift of each galaxy .
our main results are as follows : \1 ) the prominent double - peaks of the ly@xmath0 emission line for each lae straddles and is asymmetric about the velocity zero - point .
the strongest emission peak is redshifted by 176 km s@xmath4 on average relative to the systemic velocity and its velocity offset and basic shape are strikingly similar among the three galaxies in our sample .
we observe larger variations from galaxy to galaxy in the emission blueward of the systemic velocity , including two of the three galaxies that display two separate weak blueshifted emission peaks .
the most blueshifted of these weak peaks for these two galaxies is offset from the systemic redshift by @xmath6 km s@xmath4 .
however , the velocity offset of the redward peak and the peak - to - peak velocity separation between the two most prominent peaks that straddle the velocity zero - point are @xmath125 smaller for each lae than the same measurements made on average for ly@xmath0 emitting lbgs with similar ly@xmath0 line profile morphologies .
this is true even when taking into account the spectral resolution differences for each sample .
\2 ) we have compared our ly@xmath0 spectra with the predicted line profiles of a grid of ly@xmath0 radiative transfer models in the popular and relatively idealized spherical expanding shell geometry to model large - scale galactic outflows .
in contrast to the findings of works such as @xcite for single - peaked asymmetric ly@xmath0 line profiles , we observe several key discrepancies between the best - fitting models and the data . visually , the redshifted ly@xmath0 emission component is acceptably reproduced by models with low column density of neutral gas ( @xmath88 @xmath29 ) . however , the blueshifted emission component ubiquitously has too little flux as compared to the data and has an incorrect velocity offset relative to the systemic velocity for two of the three galaxies ( hps194 and hps251 ) .
additionally , ly@xmath0 radiative transfer through an expanding shell can not produce the highly blueshifted emission peak that is observed for hps256 and hps251 .
\3 ) based on the above analysis , we caution against equating an observed velocity offset of ly@xmath0 directly with an outflow or inflow velocity , especially at low spectral resolutions where the ly@xmath0 line profile is unresolved and the intrinsic asymmetry of the emission line can bias the velocity offset measurement .
the measured velocity offset , especially for unresolved ly@xmath0 spectra , is a complex function of many parameters describing the neutral gas , especially the column density due to its large dynamic range and ability to significantly change the width and asymmetry of the emission components .
additionally , the ly@xmath0 line profile shape and the resulting velocity offset can also be highly dependent on the viewing angle for more realistic , non - spherical gas geometries .
\4 ) for luminous laes with @xmath123 erg s@xmath4 , like those in the hps lae sample , galaxy - galaxy interactions may play a significant role in producing and aiding the escape of copious ly@xmath0 photons and in shaping the emergent line profile by inducing star formation and clearing gas and dust .
such gravitational interactions may cause deviations away from spherical neutral gas geometries and velocity fields , such as those modeled in our expanding shell radiative transfer simulations .
the effects of a non - spherical outflow in addition to simultaneous inflow of neutral gas , as well as smaller - scale ly@xmath0 radiative transfer effects within the ism that are not captured in the expanding shell model could all contribute to the discrepancies we observe between the ly@xmath0 spectra of these galaxies and the predictions of the models .
in forthcoming work , we will extend this study to obtain spectrally resolved ly@xmath0 line profiles of a larger sample of @xmath126 high laes using multi - object techniques . of particular importance in the forthcoming work will be the ability to obtain at least one dimension of spatial information .
the combination of a larger sample and @xmath1271-d spectra will allow us to not only investigate the frequencies of various ly@xmath0 line profile morphologies among this unique sample , but also investigate the spatial distribution and possible extended nature of the ly@xmath0 emission .
the latter may provide important constraints on the emission and escape mechanisms of ly@xmath0 photons from these systems .
+ t.s.c . acknowledges the support of a national science foundation graduate research fellowship during this work .
this research is partially supported by the national science foundation under grant ast-0926815 and the texas norman hackerman advanced research program under grant 003658 - 0295 - 2007 .
the construction of the mitchell spectrograph was possible thanks to the generous support of the cynthia & george mitchell foundation .
we thank mcdonald observatory and its staff for graciously supporting the observations .
we would additionally like to thank the following individuals for useful scientific discussions : v. bromm , m. fumagalli , j. greene , d. jaffe , j. jardel , e. komatsu , k. kulas , e. mclinden , c. scarlata , c. steidel , v. tilvi , and s. tuttle .
finally , we acknowledge the anonymous referee for useful comments that have improved this paper . + | we present new results on the spectrally resolved ly@xmath0 emission of three ly@xmath0 emitting field galaxies at @xmath1 with high ly@xmath0 equivalent width ( @xmath2 ) and ly@xmath0 luminosity ( @xmath3 erg s@xmath4 ) . at 120 km s@xmath4 ( fwhm )
spectral resolution , the prominent double - peaked ly@xmath0 profile straddles the systemic velocity , where the velocity zero - point is determined from spectroscopy of the galaxies rest - frame optical nebular emission lines .
the average velocity offset from systemic of the stronger redshifted emission component for our sample is 176 km s@xmath4 while the average total separation between the redshifted and main blueshifted emission components is 380 km s@xmath4 .
these measurements are a factor of @xmath5 smaller than for uv continuum - selected galaxies that show ly@xmath0 in emission with lower ly@xmath0 equivalent width .
we compare our ly@xmath0 spectra to the predicted line profiles of a spherical `` expanding shell '' ly@xmath0 radiative transfer grid that models large - scale galaxy outflows .
specifically blueward of the systemic velocity where two galaxies show a weak , highly blueshifted ( by @xmath6 km s@xmath4 ) tertiary emission peak , the model line profiles are a relatively poor representation of the observed spectra . since the neutral gas column density has a dominant influence over the shape of the ly@xmath0 line profile , we caution against equating the observed ly@xmath0 velocity offset with a physical outflow velocity , especially at lower spectral resolution where the unresolved ly@xmath0 velocity offset is a convoluted function of several degenerate parameters . referring to rest - frame ultraviolet and optical _ hubble space telescope _ imaging , we find that galaxy - galaxy interactions may play an important role in inducing a starburst that results in copious ly@xmath0 emission , as well as perturbing the gas distribution and velocity field which have strong influence over the ly@xmath0 emission line profile . |
the study of difference equations ( des ) is an interesting and useful branch of discrete dynamical systems due to their variety of behaviors and ability to modelling phenomena of applied sciences ( see @xcite and the references therein ) .
the standard framework for this study is to consider iteration functions and sets of initial conditions in such a way that the values of the iterates belong to the domain of definition of the iteration function and therefore the solutions are always well defined .
for example , in rational difference equations ( rdes ) a common hypothesis is to consider positive coefficients and initial conditions also positive , see @xcite .
+ such kind of restrictions are also motivated by the use of de as applied models , where negative initial conditions and/or parameters are usually meaningless , @xcite . + but there is a recent interest to extend the known results to a new framework where initial conditions can be taken to be arbitrary real numbers and no restrictions are imposed to iteration functions .
it is in this setting where appears the forbidden set of a de , the set of initial conditions for which after a finite number of iterates we reach a value outside the domain of definition of the iteration function .
indeed , the central problem of the theory of des is reformulated in the following way : + _ given a de , to determine the good ( @xmath0 ) and forbidden ( @xmath1 ) sets of initial conditions . for points in the good set , to describe the dynamical properties of the solutions generated by them : boundedness , stability , periodicity , asymptotic behavior , etc . _
+ in this paper we are interested in the first part of the former problem : how can the forbidden set of a given de of order @xmath2 be determined ? in the previous literature to describe such sets , when it is achieved , is usually interpreted as to be able to write a general term of a sequence of hypersurfaces in @xmath3 . but those cases are precisely the corresponding to de where it is also possible to give a general term for the solutions .
unfortunately there are a little number of des with explicitly defined solutions .
hence we claim that new qualitative perspectives must be assumed to deal with the above problem .
therefore , the goals of this paper are the following : to organize several techniques used in the literature for the explicit determination of the forbidden set , revealing their resemblance in some cases and giving some hints about how they can be generalized .
thus we get a long list of des with known forbidden set that can be used as a frame to deal with the more ambitious problem of describe the forbidden set of a general de .
we review and introduce some methods to work also in that case . and finally we propose some future directions of research . + the paper is organized as follows : after some preliminaries , we review the riccati de , which is one of the few examples of de where the former explicit description is possible . as far as we know ,
almost all the literature where the forbidden set is described using a general term includes some kind of semiconjugacy with a riccati de .
des obtained via a change of variables or topological semiconjugacy are the topic of the rest of the section . in the following
we will discuss how algebraic invariants can be used to transform a given equation into a riccati or linear one depending upon a parameter , and therefore determining its forbidden set .
after we will deal with an example , found in @xcite , of description where the elements of the forbidden set are given recurrently but in explicit form .
we introduce a symbolic description of complex and real points of @xmath1 in section [ sec : symbolic ] , whereupon in section [ sec : qualitative ] we study some additional ways to deal with the forbidden set without an explicit formula .
we finalize with a list of open problems and conjectures .
+ to avoid an overly exposition , we have omitted some topics as the study of systems of difference equations ( sdes ) , ( see @xcite ) , the use of normal forms ( see @xcite ) , the systematic study of forbidden sets in the globally periodic case and the importance of forbidden sets in lorenz maps .
in this paper we deal with difference equations ( des ) and systems of difference equations ( sdes ) .
general definitions of these concepts are the following .
+ let @xmath4 be a nonempty set and @xmath5 a nonempty subset of @xmath4 .
let @xmath6 be a map .
a de of order @xmath7 is the formal expression @xmath8 ( [ eq : deorder1 ] ) represents a set of finite or infinite sequences of @xmath9 , as given @xmath10 , @xmath11 is constructed by recurrence using ( [ eq : deorder1 ] ) , that is , @xmath12 and if @xmath13 then @xmath14 .
when this process can be repeated indefinitely then @xmath15 and @xmath11 is named the solution of ( [ eq : deorder1 ] ) generated by the initial condition @xmath16 .
+ we call @xmath17 the _ forbidden set ( fs ) _ of ( [ eq : deorder1 ] ) , and @xmath18 _ the good set ( gs ) _ of ( [ eq : deorder1 ] ) . +
a de of order @xmath2 is @xmath19 where now @xmath20 is the iteration map , and the fs is defined as @xmath21 in a similar way , we define a sde of order @xmath7 as @xmath22 provided @xmath23 , @xmath24
. + system ( [ eq : sdeorder1 ] ) can be expressed as de of order @xmath7 of type ( [ eq : deorder1 ] ) using the vectorial notation @xmath25 and considering the map @xmath26 whose components are @xmath27 . + finally , a sde of order @xmath2 is defined as a set of equations @xmath28 using maps @xmath29 for @xmath24 . + in vectorial form
the set of equations ( [ eq : sdeorderk ] ) can be rewritten as in ( [ eq : deorderk ] ) @xmath30 with @xmath31 .
+ the former definitions depend on how is given the domain of definition @xmath9 of the iteration map . to remark that point ,
let s consider @xmath32 and the real de @xmath33 which forbidden set is @xmath34 when @xmath9 is defined as @xmath35 meanwhile @xmath36 is empty when @xmath9 is the natural domain of the constant function . +
to avoid degenerated or trival cases , some further restrictions must be imposed to @xmath9 .
natural and common restrictions consist usually in regarding which is the domain of definition of the iteration map .
therefore , in equation @xmath37 we say that @xmath38 as every solution with @xmath39 is 2-periodic and @xmath40 is not defined when @xmath41 .
it is implicitly assumed that @xmath42 .
+ we remark that ( [ eq:1overx ] ) is also a de over @xmath43 , the projective line , and then @xmath44 being every solution 2-periodic if @xmath45 using the rules @xmath46 , @xmath47 .
+ analogously , the de @xmath48 has @xmath44 in @xmath49 , @xmath50 when ( [ eq:1overx^2 + 1 ] ) is taken over the complex field and @xmath36 is again empty in the projective plane @xmath51 . + in practical applications an undefined zero division means that the denominator belongs to certain neighbourhood of zero .
therefore in de as @xmath52 where @xmath53 , @xmath54 are real polynomials , the forbidden set problem could be studied using @xmath55 for a machine - value @xmath56 .
+ in applied models where only positive values of the variables have practical meaning , @xmath9 could be @xmath57 .
+ as a final example of different ways to consider the @xmath9 set , let s recall the de associated to a lorenz map .
let @xmath58 be the unit interval @xmath59 $ ] .
let @xmath60 .
we say that @xmath61 is an expansive lorenz map if the following conditions are satisfied a. @xmath62 is increasing and continuous in each interval @xmath63 and @xmath64 $ ] b. @xmath65 and @xmath66 c. @xmath62 is topologically expansive , i.e. , there exists @xmath56 such that for any two solutions @xmath67 and @xmath68 of ( [ eq : deorder1 ] ) not containing the point @xmath69 , there is some @xmath70 with @xmath71 , and if some of the former solutions contains @xmath69 the inequality remains valid taking @xmath72 or @xmath73 . condition ( iii ) is equivalent to say that the preimages of the point @xmath69 are dense in @xmath74 , or , in the forbidden set notation , to say that @xmath75 .
+ it is obvious that @xmath62 can be arbitrarily defined in @xmath69 ( without bilateral continuity ) and in that case @xmath44 . in the standard definition
it is assumed that @xmath76 .
+ lorenz maps are an important tool in the study of the lorenz differential equations and the lorenz attractor , and also in the computation of topological entropy in real discontinuous maps .
see @xcite and the references therein .
+ in the following we deal mostly with des where iteration functions are quotients of polynomials , known as rational difference equations ( rdes ) .
let s briefly recall some well known results about riccati des .
let @xmath77 such that @xmath78 . a riccati de of order @xmath7 is @xmath79
some special cases occur for particular values of the parameters . if @xmath80 , @xmath81 or @xmath82 the equation is linear , constant or globally of period @xmath83 respectively ( see @xcite ) .
if none of those conditions stand , an affine change of variables transforms ( [ eq : riccatiorder1 ] ) into @xmath84 @xmath85 being the riccati number .
finally @xmath86 leads to the linear de @xmath87 .
the closed form solution of the linear de ables to write a closed form solution for ( [ eq : riccatiorder1_sf ] ) .
moreover , there is a correspondence between the solutions of the linear de containing the zero element and the finite or forbidden set generated solutions of ( [ eq : riccatiorder1_sf ] ) .
this is the idea behind the characterization of the fs of ( [ eq : riccatiorder1 ] ) that can be found , for example , in @xcite . from a topological point of view
, that characterization shows that @xmath36 is a convergent sequence or a dense set in @xmath49 or a finite set .
moreover , in the last case equation ( [ eq : riccatiorder1_sf ] ) is globally periodic ( see @xcite ) .
+ analogously , for the second order riccati de @xmath88 we can transform it into a linear one using the change of variables @xmath89 .
the forbidden set is then described as ( see @xcite ) @xmath90 where coefficients @xmath91 depend on the roots of the characteristic equation . roughly speaking @xmath36 is a countable union of plane hyperbolas convergent to a limit curve in some cases , dense in open sets in other cases and even a finite collection when the de is globally periodic .
+ our first proposal ( open problem [ op : qualitativedescriptionriccati ] ) is to clarify which kind of topological objects can be obtained in the riccati de of second order and to generalize to the riccati de of order @xmath2 @xmath92 we claim also ( conjecture [ conj : riccatigloballyperiodic ] ) that de ( [ eq : riccatiorderk ] ) is globally periodic if and only if @xmath36 is a finite collection of hypersurfaces in @xmath3 , generalizating what happens in the case of order 1 ( see @xcite ) . to study the higher order riccati de see also @xcite . + in the former discussion , it was important to determine the set of initial conditions whose solutions in the linear equation included the zero element .
this is called the zero set of the linear equation , @xmath93 .
for example , in the riccati de of order @xmath7 , @xmath93 is defined as the set of points @xmath94 such that the solution of the linear de includes the zero element . but @xmath95 and @xmath96 depend on the initial condition @xmath97 of ( [ eq : riccatiorder1_sf ] ) via the change of variables @xmath86 .
we can write the dependence as @xmath98 , @xmath99 .
therefore , @xmath36 can be expressed as @xmath100 generalizing to the riccati de of order @xmath2 ( [ eq : riccatiorderk ] ) , we get @xmath101 where @xmath102 are the terms of the associated linear equation of order @xmath103 .
+ note that the semiconjugacy formula @xmath86 has zero as a pole .
the forbidden can be regarded as the transformation of certain special set of the linear equation corresponding to the singularities of the semiconjugacy .
+ there is an increasing number of works in the recent literature that use the closed form solution of the riccati equation to describe the forbidden set of other rational difference equations .
let s review and complete some of them .
+ the order @xmath104 de @xmath105 can be transformed into the linear form @xmath106 using the semiconjugacy @xmath107 .
the forbidden set is a sequence of planes in @xmath108 .
the case @xmath109 and @xmath110 in ( [ eq : abozeid_1 ] ) was developed in @xcite , while @xmath111 , @xmath109 and @xmath112 is in @xcite .
see also @xcite for @xmath113 and @xmath114 . + a similar change of variables , @xmath115 , gives the former linear equation when it is applied to the following de of order @xmath116 , see @xcite .
@xmath117 we propose a generalization of this problem remarking that the following de @xmath118 becomes @xmath106 with @xmath119 ( open problem [ op : abozeid_generalized ] ) .
+ in @xcite , the change @xmath120 gives the riccati equation @xmath121 when we apply it to @xmath122 the authors explicitly describe @xmath36 as a sequence of plane hyperbolas .
+ equation ( [ eq : bajoliz2011 ] ) admits several generalizations .
if we use the change of variables @xmath123 applied to the riccati de @xmath121 , we obtain @xmath124 a rde of order @xmath103 whose forbidden set can be described using that of the riccati de . in this case
we have a rde of third degree .
a more interesting family of equations , where the degree remains equal to two , is @xmath125 where @xmath126 is a natural number .
remark that for @xmath127 we get equation ( [ eq : bajoliz2011 ] ) . using the change of variables @xmath128 , ( [ eq : bajoliz2011_generalized2 ] ) leads to @xmath129 that it is not a riccati difference equation , but can still be reduced to a linear form .
indeed , if with an affine change we transform it into @xmath130 we can introduce @xmath131 to get @xmath132 a linear de of order @xmath133 ( see open problem [ op : bajoliz2011_generalized ] ) . + in @xcite , it is shown that @xmath1 is a family of straight lines in @xmath134 in the case of equation @xmath135 here we get the riccati form @xmath136 via the change of variables @xmath137 .
+ given @xmath138 a possible generalization is @xmath139 since @xmath140 reduce ( [ eq : mcgrath2006_generalized ] ) to @xmath141 that can be linearized to an equation of order @xmath133 as we do for ( [ eq : bajoliz2011_generalized2 ] ) ( open problem [ op : mcgrath2006_generalized ] ) .
+ the difference equation @xmath142 of reference @xcite becomes @xmath143 with the change @xmath144 .
generalizing , in @xcite the rde of order @xmath103 @xmath145 is reduced to the riccati form @xmath146 when we use the multiplicative change of variables @xmath147 .
+ let s generalize the fs characterization of @xcite to the case where @xmath148 are arbitrary complex numbers .
if one of them is zero , then equation ( [ eq : shojaei2011 ] ) becomes trivial or globally periodic and @xmath36 is empty or the set of points with at least one zero component .
+ if @xmath149 , let @xmath150 .
therefore ( [ eq : riccati_shojaei ] ) transforms into @xmath151 the former equation admits the following closed form solution @xmath152 and from here it is easy to give the forbidden set expression . if @xmath153 , then @xmath154 and @xmath155 .
if @xmath69 is another root of the unity , then the equation is globally periodic and the forbidden set of the riccati equation is finite : @xmath156 supposing @xmath157 to be the smallest positive integer such that @xmath158 .
+ when we apply these results to ( [ eq : shojaei2011 ] ) , we get that @xmath159 that is , a countable union of generalized hyperbolas when @xmath69 is not a root of the unity ( @xmath15 ) or a finite union if @xmath160 . +
there are a number of rdes for which the closed form solution and fs is given in @xcite .
those equations can be grouped in three categories : * those obtained from the multiplicative de @xmath161 when we do change of variables as @xmath162 @xmath163 we deal then with rde of order @xmath83 and degree @xmath83 .
it is not difficult to obtain the closed form of ( [ eq : rhouma_multiplicative ] ) from which the fs expression is constructed .
* rde of order @xmath83 and degree @xmath83 resulting of the introduction of the variable @xmath164 in the general linear de @xmath165 * and another family of rde of order @xmath83 and degree @xmath83 , given in this case by the change @xmath166 applied to the riccati de of order @xmath7 @xmath167 in each of the former cases we propose some generalizations . given a mbius transformation @xmath168 ,
it must be possible to explicitly determine the fs of the rde constructed with the change @xmath169 applied to equation ( [ eq : rhouma_multiplicative ] ) or applied to equation ( [ eq : rhouma_linear ] ) .
these are the claims of open problems [ op : rhouma_multiplicative ] and [ op : rhouma_linear ] respectively . in the third case
, we have that the closed form solution of ( [ eq : rhouma_riccati ] ) and the change of variables able to write @xmath170 moreover , the relationship @xmath166 can be interpreted as a nonautonomous linear de when solved for @xmath171 , that is @xmath172 therefore a general expression for @xmath173 and for @xmath36 can be computed , and of course , the same idea must work for every change of variables @xmath174 where the explicit solution of the nonautonomous @xmath173 equation is known , and for every equation @xmath175 whose closed form is also known ( open problem [ op : rhouma_generalized ] ) .
an interesting modification of the former ideas is the use of invariants to describe the fs of some rdes . consider the following example from @xcite @xmath176 where @xmath177 .
this equation has the following invariant @xmath178 that is , for every solution @xmath179 of ( [ eq : palladino_1 ] ) there exist @xmath180 such that ( [ eq : palladino_1_invariant ] ) holds for every @xmath181 .
+ the presence of an invariant ables to write an alternative form of the de . indeed , solving ( [ eq : palladino_1_invariant ] ) for @xmath182 and changing the indices , we get the following riccati de @xmath183 where recall that constant @xmath184 depend on the values of @xmath185 and @xmath16 .
this is the basic idea to describe the fs of ( [ eq : palladino_1 ] ) .
+ note that formula ( [ eq : palladino_1_invariant ] ) implies the identity @xmath186 from where ( [ eq : palladino_1 ] ) is deduced .
+ a generalization of the former remark is the following .
let @xmath187 and @xmath188 be two mbius transformations , @xmath189 , @xmath190 , and consider the invariant @xmath191 therefore @xmath192 and from here we get @xmath193 a rde of order @xmath83 whose fs could be described with the former methodology as the invariant ( [ eq : palladino_generalized_invariant ] ) implies also that @xmath194 which is a riccati de of order @xmath7 ( open problem [ op : palladino_generalized ] ) .
for example , given @xmath195 and @xmath196 non zero complex numbers , the invariant @xmath197 leads to de @xmath198 and the invariant @xmath199 to de @xmath200 both of them not included in reference @xcite .
+ also , given @xmath201 , the invariant @xmath202 produces the de @xmath203 that can be studied in the same way because the @xmath184 dependent de @xmath204 is reduced to a linear equation of order @xmath103 using the same changes as in equation ( [ eq : bajoliz2011_generalized2 ] ) . +
another example of use of invariants is in @xcite .
de @xmath205 verifyes that @xmath206 is constant over each solution of ( [ eq : aghajanishouli ] ) .
therefore , from the equality @xmath207 the following linear relation is deduced @xmath208 and from here it is easy to give the closed and forbidden set of ( [ eq : aghajanishouli ] ) . in particular , the fs is @xmath209 a possible generalization is to consider invariants of the form @xmath210 as every solution will be associated to a linear recurrence ( open problem [ op : aghajanishouli ] ) .
+ finally , a general question is to determine which relationship exists between the poles and zeros of an algebraic invariant , from one side , and the elements of its associated de , from other .
one of the oldest examples in the literature concerning the forbidden set problem is in @xcite .
let @xmath211 be a real number , and consider the following rde of order @xmath83 @xmath212 map @xmath213 is the unfolding of ( [ eq : camouzisdevault ] ) as we have that @xmath214 , @xmath215 .
let @xmath216 be its inverse map , and @xmath217 .
therefore @xmath218 this is an obvious characterization of the fs as the set of inverse orbits of poles of the iteration map that can be improved as follows .
let s consider the subset @xmath219 and let @xmath220 be the sequence of functions @xmath221 defined inductively using the operator @xmath222 in the following way @xmath223 therefore we have [ th : theforbiddenset_cdv ] let @xmath224 be the subset of the forbidden set of ( [ eq : camouzisdevault ] ) defined by @xmath225 .
then @xmath226 in general , given a rde we can describe its fs as a set of implicitly defined hypersurfaces . in the second order case we get a set of implicitly defined curves .
for example , in the case of pielou s equation @xmath227 we can consider the unfolding map @xmath228 such that @xmath214 . by iterating map
@xmath229 we get @xmath230 @xmath231 @xmath232 @xmath233 from where the following first forbidden curves are deduced ( see figure [ fg : pielou ] ) : @xmath234 @xmath235{pielour1.eps } & \includegraphics[width=0.45\textwidth]{pielour2.eps}\\ \includegraphics[width=0.45\textwidth]{pielourm1.eps } & \includegraphics[width=0.45\textwidth]{pielourm2.eps } \end{array}$ ] remark that those curves are given in an implicit form .
we propose to find a way to construct the explicit form as in theorem [ th : theforbiddenset_cdv ] ( open problem [ op : explicitcurves ] ) .
let s consider the following rdes of order @xmath7 and degree @xmath83 @xmath236
@xmath237 where @xmath238 is a real or complex parameter . these equations are studied in @xcite in connection with their relationship with li - yorke chaos in rdes . + it is possible to describe the forbidden set of the particular case @xmath239 using a symbolic notation .
let @xmath240 be the inverse branches of the iteration function in ( [ eq : inverseparabola_a1 ] ) .
clearly , @xmath36 is the forward multiorbit of the poles of @xmath241 . moreover the unique preimage of pole @xmath242 in ( [ eq : inverseparabola_a1 ] ) is @xmath243 .
the preimages of pole @xmath7 are @xmath244 , where @xmath245 or @xmath246 for @xmath247 , and @xmath244 is the abbreviation of @xmath248 .
those preimages are always well defined in @xmath249 , that is , @xmath250 for every word , and the former representation is unique , that is , given two words such that @xmath251 then @xmath252 and @xmath253 for @xmath247 .
+ therefore @xmath254 let s see that words @xmath244 are distint from zero and are unique , as the remaining assertions are straightforward to demonstrate .
+ when @xmath255 , @xmath256 . for @xmath257 , if @xmath258 , then @xmath259 and therefore @xmath260 , what is impossible as @xmath242 does not belong to the image of @xmath261 or @xmath262 .
+ now , suppose that @xmath251 .
there is no lack of generality if we assume that @xmath263 , because @xmath261 and @xmath262 are injective maps .
let @xmath264 and @xmath265 . the former equality is rewritten as @xmath266 .
but note that @xmath267 from where we deduce that @xmath268 , what is again impossible . to avoid the contradictions both words @xmath244 and @xmath269 must be equal . note that this is a set in the complex field ( see figure [ fg : forbiddensetinverseparabola_c ] ) .
if we center in the real field , note that applying @xmath262 one time we stay over the reals , and applying it two consecutive times we get @xmath270 .
therefore the forbidden set of ( [ eq : inverseparabola_a1 ] ) over @xmath49 is @xmath271 numerically , @xmath272 is the projection of figure [ fg : forbiddensetinverseparabola_c ] over the real line .
.,title="fig:",scaledwidth=60.0% ] + in open problem [ op : symbolicdescription ] we propose to find the symbolic description of @xmath273 and @xmath272 for the families of rdes ( [ eq : inverseparabola ] ) and ( [ eq : inverselogistics ] ) .
as we have seen in the former sections , there are some works describing the fss of des . but
it should be remarked that the most part of des do nt have a closed form and therefore there is probably no explicit expression for their fss .
+ we claim that new perspectives should be taken in order to solve the problem of the fs for a generic de . specifically , given a de
we propose three general questions about its fs 1 . which are its topological and metrical properties ?
( openness , boundedness , fractal dimension , ... ) 2 .
what can be said about perturbations of its parameters ?
+ for example , given a family of des @xmath274 depending on parameter @xmath238 , and supposing that some feature of the fs is known when @xmath275 , what can be said about the fs of the de when @xmath276 ? 3 . are there some relationships between the properties of the fs and the dynamical behavior of the de ? there are , as far as we know , very few papers with a qualitative approach .
let s briefly review them .
+ in @xcite the problem of closedness of the fs is discussed in connection with the global periodicity question .
let @xmath277 be the real or complex field .
let @xmath278 and consider the @xmath2-dimensional de or dynamical system @xmath279 let @xmath280 be the domain of @xmath229 .
a set @xmath281 has full lebesgue measure in @xmath282 if the measure of @xmath283 is zero .
the map @xmath229 is almost a local diffeomorphism if there exists an open set @xmath284 of full lebesgue measure in @xmath282 such that @xmath229 is continuously differentiable on @xmath285 ( or analytic on @xmath285 when @xmath286 ) and @xmath287 a.e . @xmath288 .
let @xmath289 be an almost local diffeomorphism such that the set @xmath290 has zero lebesgue measure .
suppose also that @xmath282 is open and that @xmath229 is a continuous globally periodic map .
then the good set of ( [ eq : rubiomassegu_map ] ) is closed and has full measure in @xmath282 .
[ teo : rubiomassegu_maintheorem ] there are some consequences of this result .
the natural domain of a rational map @xmath291 is the set of points in @xmath292 such that any denominator in @xmath229 vanish .
the first result is for the rational case .
let @xmath282 be the natural domain of a rational map @xmath289 .
if @xmath229 is globally periodic and @xmath293 is not identically zero , then the good set of ( [ eq : rubiomassegu_map ] ) is open and has full lebesgue measure in @xmath292 . in the case of rde
, we say that a function @xmath294 depends effectively on its last variable if @xmath295 for some @xmath296 .
therefore for rde
@xmath297 we get let @xmath298 be a rational function , and @xmath282 its natural domain . if @xmath62 depends effectively on its last variable and ( [ eq : rubiomassegu_de ] ) is globally periodic , then the good set is open and has full lebesgue measure in @xmath292 .
a more concrete application is given in the case of the riccati de of order @xmath7 de ( [ eq : riccatiorder1 ] ) such that none of the quantities @xmath299 , @xmath300 and @xmath301 is zero , has an open good set if , and only if , the equation is globally periodic .
it is a natural question to wonder if this result applies also to any rde . again in @xcite is given the following not globally periodic example @xmath302 whose fs is @xmath303 and therefore closed .
numerical experiments show also that @xmath36 is not closed when the equation is regarded in @xmath249 ( see figure [ fg : fs_rubiomassegu ] , open problem [ op : rubiomassegu ] and conjecture [ conj : rubiomassegu ] ) .
+ elements of the fs of ( [ eq : rubiomassegu_ce ] ) .
green points are the unique elements of @xmath36 in @xmath49 .
fixed points are marked in yellow , and they seem to be in the clousure of @xmath36.,title="fig:",scaledwidth=40.0% ] + a second example of qualitative study of de is in @xcite . there
the author stands sufficient conditions for a de to have a good set of full lebesgue measure , generalizing a previous remark of @xcite . in the nonautonomous case ,
the main result is let @xmath298 be a function locally in fact of order @xmath2 and such that the set @xmath304 has lebesgue measure zero .
then the good set of equation ( [ eq : rubiomassegu_de ] ) has full measure in @xmath282 . here
a function @xmath62 is locally in fact of order @xmath2 if there exists an open set @xmath305 of full lebesgue measure in @xmath282 and such that @xmath295 a. e. @xmath288 .
+ in @xcite it is remarked the importance of determining the topological and geometrical properties of the good set in terms of the iteration function and its domain , and two related open problems are proposed ( see open problems [ op : rubiomassegu_equilibriumintheinterior ] and [ op : rubiomassegu_opengoodset ] ) .
+ also in @xcite there is a generalization of these considerations , studying the problem of avoiding forbidden sequences .
+ in @xcite the following de are considered : @xmath306 @xmath307 where @xmath308 and @xmath309 and @xmath310 are odd rationals with at least one of them positive .
an odd rational is a fraction of the form @xmath311 , @xmath312
. + again the objective is to determine somehow the forbidden set .
the author uses the term _ crash set _ referring to @xmath36 , emphasizing that the discontinuity is a pole .
+ none of the forbidden or crash sets of the former equations is explictly computed , but it is described qualitatively . in the case of ( [ eq : sedaghat_qualitative_1 ] ) , the iteration function belongs to the family of monotonic maps with pole , defined as the class of functions @xmath313 verifying 1 .
@xmath62 is continuous on its domain @xmath314 2 .
@xmath62 is injective on @xmath314 3 .
@xmath315 4 .
@xmath316 note that @xmath62 must be an increasing or a decreasing map , and that @xmath36 consists in the collection of preimages of the pole @xmath243 , qualitatively described in the next theorem .
@xmath53 stands for the set of fixed points of @xmath62 .
let @xmath62 be a monotonic map with pole .
a. if @xmath62 is a decreasing map , let @xmath317 be the fixed point in @xmath318 .
then the sequence @xmath319 is contained in the interval @xmath320 $ ] , and converges either to @xmath317 or to a @xmath83-cycle .
b. suppose that @xmath62 is an increasing map , @xmath321 , and @xmath322 .
then @xmath36 is contained in @xmath323 and converges monotonically to @xmath317 .
[ th : sedaghat_monotonicmapwithpole ] remark that there are some cases not included in this result .
when @xmath62 is increasing and @xmath324 , cobweb diagrams show that @xmath36 is a subset of @xmath325 $ ] and converges monotonically to @xmath317 . whereas for @xmath326
, @xmath36 could be finite or could have a more complicated structure .
for example in @xcite is wondered if in the last case @xmath36 can be dense outside a compact set .
( see open problem [ op : sedaghat_densefs ] and figure [ fg : cobweb ] ) .
+ @xmath235{cobweb31.eps } & \includegraphics[width=0.45\textwidth]{cobweb53.eps}\\ \includegraphics[width=0.45\textwidth]{cobweb75.eps } & \includegraphics[width=0.45\textwidth]{cobweb97.eps } \end{array}$ ] theorem [ th : sedaghat_monotonicmapwithpole ] applies not only to equation ( [ eq : sedaghat_qualitative_1 ] ) , but also to equations as : + @xmath327 and , in general to des of the form @xmath328 where @xmath308 and @xmath329 is a bijection such that @xmath330 .
+ the fs of ( [ eq : sedaghat_qualitative_2 ] ) is given in @xcite in terms of the values of @xmath238 and @xmath309 .
we propose to make a similar study for the family of de @xmath331 where @xmath309 is an odd rational and @xmath332 is the mbius transformation @xmath333 with @xmath334 ( open problem [ op : sedaghat_qualitative_1_generalized ] ) .
+ in the case of equation ( [ eq : sedaghat_qualitative_2 ] ) , in @xcite there is a description of the fs curves in an explicit form using a recurrent algorithm , in a similar way as it was made in section [ sec : fs_explicitform ] .
+ additional examples of estimative approaches where the fs is graphically represented can be found in @xcite for the following des @xmath335 @xmath336 in a similar way , we can estimate the fs of @xmath337 for differents values of parameters @xmath9 and @xmath338 .
see figure [ fg : devaultgalminasjanowskiladas ] and open problem [ op : devaultgalminasjanowskiladas ] .
@xmath235{fsdevaultetala1b5.eps } & \includegraphics[width=0.45\textwidth]{fscurvesdevaultetala1b5.eps}\\ \includegraphics[width=0.45\textwidth]{fsdevaultetala1b1.eps } & \includegraphics[width=0.45\textwidth]{fscurvesdevaultetala1b1.eps}\\ \includegraphics[width=0.45\textwidth]{fsdevaultetalam1bm1.eps } & \includegraphics[width=0.45\textwidth]{fscurvesdevaultetalam1bm1.eps } \end{array}$ ]
clarify which kind of topological objects can be obtained in the second order riccati de and generalize to the riccati de of order @xmath2 ( [ eq : riccatiorderk ] ) .
[ op : qualitativedescriptionriccati ] de ( [ eq : riccatiorderk ] ) is globally periodic if and only if @xmath36 is a finite collection of hypersurfaces in @xmath3 .
[ conj : riccatigloballyperiodic ] describe the fs of rde ( [ eq : abozeid_generalized ] ) .
[ op : abozeid_generalized ] describe the fs of rdes ( [ eq : bajoliz2011_generalized1 ] ) and ( [ eq : bajoliz2011_generalized2 ] ) .
[ op : bajoliz2011_generalized ] describe the fs of rdes ( [ eq : mcgrath2006_generalized ] ) .
[ op : mcgrath2006_generalized ] let @xmath339 .
describe the fs of the rdes obtained by applying the change @xmath169 to de ( [ op : rhouma_multiplicative ] ) .
[ op : rhouma_multiplicative ] let @xmath339 .
describe the fs of the rdes obtained by applying the change @xmath169 to de ( [ op : rhouma_linear ] ) .
[ op : rhouma_linear ] let @xmath175 a de of order @xmath103 such that its closed form is known .
let @xmath174 a change of variables that can be rewriten as a nonautonomous de @xmath340 with coefficients depending of @xmath341 and such that its closed form solution is also known .
determine the forbidden set of the de obtained by applying the change of variables to @xmath175 .
[ op : rhouma_generalized ] describe the fss of des ( [ eq : palladino_generalized_equation ] ) and ( [ eq : palladino_generalized_equation_order_k ] ) having algebraic invariants .
[ op : palladino_generalized ] let @xmath2 and @xmath342 be distint natural numbers greater than @xmath243 . determine the closed form solution and the forbidden set of des admiting an invariant of the form ( [ eq : aghajanishouli_invariant ] ) .
[ op : aghajanishouli ] for every rde of order @xmath83 , determine a region @xmath343 such that the forbidden curves in @xmath9 can be given in explicit form , and determine them .
[ op : explicitcurves ] find the symbolic description of @xmath273 and @xmath272 for the families of rdes ( [ eq : inverseparabola ] ) and ( [ eq : inverselogistics ] ) .
[ op : symbolicdescription ] let @xmath344 an almost local diffeomorphism such that @xmath345 has lebesgue measure zero .
suppose that equation ( [ eq : rubiomassegu_map ] ) is globally periodic .
obtain sufficient conditions in order that @xmath36 is closed .
[ op : rubiomassegu ] let @xmath346 be a rational function and @xmath282 its natural domain .
assume that @xmath62 depends effectively on its last variable .
then the good set of ( [ eq : rubiomassegu_de ] ) is open if , and only if , the equation is globally periodic .
[ conj : rubiomassegu ] to obtain necessary conditions in order to have that an equilibrium point of an autonomous difference equation belongs to the interior of the good set .
[ op : rubiomassegu_equilibriumintheinterior ] to obtain necessary conditions in order to have that the good set of a difference equation is open .
[ op : rubiomassegu_opengoodset ] let @xmath347 be a monotonic map with pole . complete the results of theorem [ th : sedaghat_monotonicmapwithpole ] , in the sense of describing qualitatively the fs of de @xmath348 when @xmath62 is an increasing map with no fixed point .
in particular , answer to the following questions : 1 . in which cases
@xmath36 is finite ? 2 .
can be @xmath36 dense outside a compact set ?
( see @xcite ) .
[ op : sedaghat_densefs ] determine the fs of de ( [ eq : devaultgalminasjanowskiladas ] ) .
[ op : devaultgalminasjanowskiladas ] determine the fs of de ( [ eq : sedaghat_qualitative_1_generalized ] ) by applying theorem [ th : sedaghat_monotonicmapwithpole ] in terms of @xmath238 , @xmath349 , @xmath69 and @xmath309 .
[ op : sedaghat_qualitative_1_generalized ] as we have seen throughout the paper , there are a lot of des with closed form solution and explicit forbidden set .
there are many ways to generalize them and therefore the literature on this subject is growing with more and more works devoted to that kind of de .
although those works have the value to enrich the set of known examples , may be it is time to put some order in the field . we propose to elaborate a database of des and sdes including , among others , the following topics : the forbidden and good sets , the asymptotic behavior , the boundedness of solutions , the relationships with other de or sde and the applied models related with them .
this paper has been partially supported by grant mtm2008 - 03679 from ministerio de ciencia e innovacn ( spain ) , project 08667/pi-08 fundacin sneca de la comunidad autnoma de la regin de murcia ( spain ) .
+ we thank the anonymous referees for their suggestions and comments . | in recent literature there are an increasing number of papers where the forbidden sets of difference equations are computed .
we review and complete different attempts to describe the forbidden set and propose new perspectives for further research and a list of open problems in this field .
universidad de murcia , departamento de matemticas + campus de espinardo , 30100 , murcia ( spain ) + [email protected] , [email protected] |
nonlinear dynamical systems are defined globally on manifolds . consider a system @xmath0 where @xmath1 .
the phase - space portrait of this system is defined as all the solution curves in @xmath2 ( see , e.g. , [ perko , 1991 ] . ) geometrically , these curves determine the motion of all the points in the space under this specific dynamical system .
moreover , the global manifold on which a system is stratified can be reached by studying the phase - space portrait .
for example , given a spherical pendulum ( see fig .
[ spherical pendulum ] for illustration ) , it has two degrees of freedom , and the _ lagrangian _ for this system is @xmath3 the _ euler - lagrange _ equations give @xmath4 so the system is given by two equations of motion , i.e. , @xmath5 in phase - space coordinates , if set @xmath6 , @xmath7 , @xmath8 , @xmath9 , we then have @xmath10 this is a 4-dimensional system .
now assume @xmath11 , where _ k _ is a constant , consequently @xmath12 and @xmath13 , the system will then become @xmath14 which stands for a single pendulum that sits on a 2-dimensional plane .
it is known that this system is defined on a _
klein bottle _ , ( see [ banks & song , 2006 ] and fig .
[ simple pendulum ] for an illustration . )
+ if set @xmath15 , i.e. , the system has a fixed angular velocity in @xmath16 , the system then becomes @xmath17 which is essentially a 3-dimensional hyperplane given by @xmath18 within the 4-dimensional space . moreover , since the vector field is periodic in both @xmath19 and @xmath20 with period @xmath21 , it is naturally defined within the cube @xmath22 as shown in fig .
[ special spherical pendulum](a ) . , title="fig:",width=288 ] + note that the phase - space portrait is a 2-dim single pendulum that sits on different slices defined by @xmath23 , and because we know that @xmath24 and @xmath25 are physically the same respectively , we can identify them by pairing the opposite sides via translation . in order to define the system on a compact manifold , we compress the infinite cube to a finite one , as shown in fig .
[ special spherical pendulum](b ) , and since the dynamics at the two ends are pointing the opposite directions , the identification will result in a self - intersection in the 3-dimensional euclidean space , fig . [ solid klein bottle ] shows an embedding in @xmath26 . + thus , we obtain the 3-manifold on which this special spherical pendulum is defined .
we call it 3-dimensional _ solid klein bottle_. in [ banks & song , 2006 ] , we showed that a dynamical system on a two - dimensional surface is given by a generalized automorphic function _
f_. in this paper , we will extend the previous result and propose to show how to generalise explicit differential equations that naturally have global behaviour on 3-manifolds .
again we will use the theory of automorphic functions to achieve it .
we shall now give a brief resum of 3-manifolds which will be needed in the following sections .
note that all the results are well - known , for example in [ ratcliffe , 1994 ] .
a @xmath27-manifold _ m _ without boundary is a @xmath27-dimensional hausdorff space that is locally homeomorphic to @xmath28,i.e .
, for every point @xmath29 @xmath30 there exists a homeomorphism that maps a neighbourhood _ a _ of _ x _ onto the @xmath27-dimensional euclidean space ; while if _ m _ has a boundary , then the homeomorphism maps _ a _ onto the upper - half @xmath27-dimensional euclidean space @xmath31 .
equivalently , a 3-manifold _ m _ is called a geometric 3-space .
assume that @xmath32 is a group which acts on a 3-dimensional geometric space _
x _ , then the orbit space of the action @xmath32 on @xmath33 is the set of @xmath32orbits , @xmath34 with the metric topology being the quotient topology , and the quotient map given by @xmath35 moreover , if @xmath32 is a discrete group of isometries of _ x _ , then @xmath32 is discontinuous and called a 3-dimensional _ fuchsian _ group .
in fact , it defines a fundamental region _
f _ of _ x _ which , together with its congruent counterparts , generates a tessellation of
_ x_. for a discrete group @xmath32 of isometries of a geometric space _
x _ , a subset _ f _ of _ x _ is a fundamental region if and only if 1 . the set _
f _ is open in _ x _ ; 2 .
the members of @xmath36 are mutually disjoint ; 3 .
@xmath37 for example , let @xmath38 be the translation of @xmath28 by @xmath39 for @xmath40 , then @xmath41 defines a discrete subgroup @xmath42 of @xmath43
. a fundamental region for @xmath42 will be the open unit cube in @xmath28 , as shown in fig .
[ cube ] , in fact , @xmath42 generates a tessellation of @xmath28 . by @xmath42
generated via translation @xmath39,width=192 ] if @xmath32 acts _ freely _ on _ x _ , the orbit space @xmath44 is then a 3-manifold which can also be called an _ x - space - form_. also , by assuming _
g _ is a group of similarities of a 3-dimensional geometric space _
x _ and _ m _ is a 3-manifold , we have an @xmath45atlas for _ m _ is a group of maps @xmath46 such that : 1 .
the set @xmath47 is an open connected subset of _ m _ for each _
@xmath48 maps @xmath47 homeomorphically onto an open subset of _ x _ for each _ i_. 3 .
@xmath49 4 .
if @xmath47 and @xmath50 overlap , then the map @xmath51 agrees in a neighbourhood of each point of its domain with an element in _
g_. note that @xmath52 consists of the _ charts _ of the _ ( x , g)atlas_. an _ ( x , g)structure _ is then defined as the maximal _ ( x , g)atlas _ for _ m_. hence a 3-manifold _ m _ with an _ ( x , g)structure _ is called an _ ( x , g)manifold_. it is well - known ( e.g. , [ ratcliffe , 1994 ] ) that the orbit space @xmath44 , together with the induced _
@xmath53atlas _ , is an _
@xmath53-manifold_. furthermore , we can obtain this 3-manifold by gluing one fundamental region _
f _ along the corresponding sides .
let @xmath54 be a family of fundamental regions in a geometric space _
x _ and @xmath32 be a group of isometries of _
x_. we can then construct the @xmath53manifold by applying the _ @xmath32-side - pairing_. a _ @xmath32-side - pairing _ for @xmath54 is a subset of @xmath32 , @xmath55 such that for each side _ s _ in @xmath54 , 1 .
there exists a side @xmath56 in @xmath54 that satisfies @xmath57 , 2 .
@xmath58 3 .
if _ s _ is a side of _ f _ in @xmath54 and @xmath59 is a side of @xmath60 , then @xmath61 the elements of @xmath32 are called the side - pairing transformations of @xmath54 , which generates an equivalence relation on the set @xmath62 , i.e. , the cycles of @xmath32 .
moreover , @xmath56 is uniquely determined by @xmath63 .
so if the @xmath32-side - pairing is proper , i.e. , each cycle of @xmath32 is finite and has solid angle sum @xmath64 , then by choosing two fundamental regions in @xmath54 , say @xmath65 and @xmath66 , the elements in @xmath32 will associate each side in @xmath65 with a unique one in @xmath66 , identifying the corresponding sides together will eventually generate a 3-manifold with an _
@xmath53-structure _ attached .
for instance , as in the previous example , after pairing the opposite sides of the unit cube by translations @xmath32 , we effectively end up with a 3-manifold _ m _ which is known as the _ cubical euclidean 3-torus _ ( see fig .
[ 3-torus ] for illustration ) .
in this section we shall first give a brief resum of automorphic functions .
more details can be found in , for example , [ ford , 1929 ; ratcliffe , 1994 ] .
to denote the points in @xmath26 , we use the following coordinates : @xmath67 also , we can think @xmath26 as a subset of hamilton s quaternions @xmath68 , so a point _ p _
@xmath69 can be expressed as a quaternion whose fourth term equals to zero , i.e. , @xmath70 where @xmath71 and @xmath72 , then a mbius transformation of @xmath73 is a finite composition of reflections of @xmath73 in spheres , where @xmath73 is the one - point compactification of @xmath26 , i.e. , @xmath74 it is exactly the linear fractional transformations of the form @xmath75 where @xmath76 and @xmath77 .
a _ mbius _ transformation is a conformal map of the extended 3-space , ( i.e. , riemann 3-manifold ) , denoted by @xmath78 .
moreover , ( [ mobius transformation ] ) can be represented in terms of a matrix @xmath79 in fact there exists a group homeomorphism : @xmath80 given by @xmath81 which becomes an isomorphism on the projective special linear group @xmath82 ( i.e. , those elements of @xmath83 of positive determinant modulo the scalar matrices ) .
it is known that 3-dimensional hyperbolic space ( or 3-dimensional hyperbolic manifold ) is the unique 3-dimensional simply connected riemannian manifold with constant sectional curvature @xmath84 ( see , e.g. , [ elstrodt , 1998 ] ) . also , since _ mbius _
transformations are defined on riemannian manifold , they can be used to generate a discrete group of discontinuous isometries , @xmath32 , of the upper half - space @xmath85 , where @xmath86 note that @xmath85 is a model for hyperbolic space , so we can use @xmath32 to tessellate @xmath85 and obtain a 3-manifold , _ m _ , by @xmath32-side - pairing either the fundamental region or a finite collection of discrete regions congruent to the fundamental region . obviously _
is with @xmath87-structure .
we shall continue using hamilton s quaternion @xmath68 , and the notation for points _ p _ in @xmath85 will be the same as that in @xmath26 , only with @xmath88 .
furthermore , since we restrict attention to the upper half - space , the automorphism group becomes @xmath89 ( linear fractional transformation with complex coefficients ) . if _
t _ is a map of the form ( [ mobius transformation ] ) , where @xmath90 , we have @xmath91 for an element @xmath92 , @xmath93 is classified as follows : 1 . if @xmath94 & @xmath95 , _ t _ is _ parabolic _ ; 2 . if @xmath96 & @xmath95 , _ t _ is _ hyperbolic _ ; 3 .
if @xmath97 & @xmath95 , _ t _ is _ elliptic _ ; to define explicit expressions for dynamical systems on _ m _ , we first need to find the so - called automorphic functions that are invariant under the elements of @xmath32 . by definition , an automorphic function _ a _ for the _ fuchsian _ group
@xmath32 is a meromorphic function generated on @xmath85 such that @xmath98 for all @xmath99 and @xmath100 @xmath101
. it would be nice if the dynamics on the 3-manifold _ m _ can be defined as @xmath102 where _ a _ is an automorphic function .
however , since we are dealing with vector fields , the solutions generated by ( [ automorphic function ] ) in @xmath85 are not @xmath32-invariant in the sense that dynamics at the boundary of the fundamental region wo nt match up when applying the @xmath32-side - pairing . in order to obtain systems
@xmath103 with @xmath32-invariant trajectories , we require the following invariance of the vector field _ f _ : the system @xmath104 will have @xmath32-invariant trajectories for any given discrete group @xmath32 of isometries of hyperbolic @xmath27-space _ x _ , if @xmath105 * proof . * to make the dynamics match up after the side - pairing , we require the `` ends '' of infinitesimal vectors in the direction of @xmath106 to map appropriately under @xmath32 ( see fig . [ matching vector fields ] for illustration ) . ,
width=192 ] hence we require @xmath107 for sufficiently small @xmath108 . thus @xmath109 @xmath110 so the lemma is proved .
@xmath111 to work out the relation between @xmath106 and @xmath112 explicitly , we have , from ( [ mobius transformation ] ) , @xmath113 @xmath114 @xmath115 therefore , for such a map @xmath116 , the invariance of the dynamical system _ f _ given by ( [ dynamical system ] ) can be written in the form @xmath117 note that ( [ modified dynamical system ] ) differs from the scalar invariance @xmath118 which is given by any automorphic function .
so we shall obtain vector fields _ f _ that satisfies ( [ modified dynamical system ] ) by modifying the _ poincar _ theta series ( see [ ford , 1929 ] ) which can be used to generate automorphic functions for those _ fuchsian _ groups with infinite elements .
let _ h _ be a rational function , which has no poles at the limit points of the isometry group @xmath32 , the theta series is given by @xmath119 where @xmath120 , @xmath121 are the elements of @xmath32 , and @xmath122 it is easy to verify that @xmath123 for each _ i _ , and by definition , two distinct theta series @xmath124 and @xmath125 with the same choice on _ m _ , we can have @xmath126 moreover , @xmath127 for each _ i _ , i.e. , _ f _ is an automorphic function . from ( [ modified dynamical system ] ) , we know that in the case of dynamical systems , some modification must be made to the theta series so that they can provide the invariance of the vector fields .
therefore instead of @xmath124 , we define @xmath128 while keep @xmath129 as usual .
the function @xmath130 satisfies @xmath131 for each _ i _ and so defines a @xmath32-invariant dynamical system if @xmath132 . * proof . *
since @xmath125 is the normal theta series , we have @xmath133 for each _ j _ , while for @xmath134 , we have @xmath135 since @xmath136 and @xmath137 hence @xmath138 therefore the result follows . @xmath139 an * automorphic vector field * on @xmath85 is a meromorphic , hypercomplex valued function _
f _ , such that it satisfies ( [ modified dynamical system ] ) for each isometry _ t _ in the _ fuchsian _ group @xmath32 . from the discussion above , we know that such functions _ f _ , generate dynamics situated on hyperbolic 3-manifolds , which is written in the form @xmath140 the trajectories are @xmath32-invariant on any fundamental region , we can then either `` wrap up '' one of them or choose a finite number and apply the @xmath32-side - pairing , both of which will give rise to systems sit on the resulting hyperbolic 3-manifold explicitly .
+ * example . *
it is known that the upper half - space @xmath85 can be tessellated by hyperbolic ideal tetrahedron . fig .
[ tessellation ] shows one particular representation . by hyperbolic tetrahedra , title="fig:",width=192 ]
+ let the @xmath32-side - pairing be either translations or simple expansions and contractions . according to fig.[tessellation ]
we then have the _ fuchsian _ group generated by the transformations @xmath141 choosing @xmath142 we can obtain a dynamical system by using the modified automorphic functions .
note that in this example , @xmath143 and @xmath144 do nt define poles within the phase - space , however , the system will have poles introduced by the modified theta series .
in fact , the whole _ z_plane will be covered with equilibria due to the fact that it contains only cusp points .
[ identification ] shows one possible construction of a hyperbolic 3-manifold by translation .
moreover , fig .
[ dynamics ] illustrates the solution trajectories of the system ( computed in maple ) , and the vector fields match up perfectly at the boundaries .
+ , where _ f _ is generated by @xmath143 and @xmath144.,width=153,height=153 ]
we now propose another way of generating dynamical systems on 3-manifolds . instead of using the upper half - space model , we shall now investigate hyperbolic 3-manifolds under the conformal ball model .
the same argument applies here , i.e. , given a group @xmath32 of isometries of @xmath33 and a proper @xmath32-side - pairing , we can form a 3-manifold @xmath145 with an @xmath53-structure by gluing a finite number of disjoint convex polyhedra .
moreover , if we take into consideration of the dynamical systems naturally situated on those solid fundamental polyhedra , the @xmath32-side - pairing will then yield a new system defined on the resulting manifold @xmath145 if and only if the trajectories match up according to the gluing pattern . again , as an example , we consider a regular ideal tetrahedron in @xmath146 , which has the shape in fig .
[ ideal tetrahedron ] .
+ let @xmath147 and @xmath148 be two disjoint regular ideal tetrahedrons in @xmath146 , illustrated in fig .
[ two tetrahedrons ] . for simplification ,
we regard them as regular tetrahedrons in the euclidean space .
+ because a _ mbius
_ transformation of the unit ball @xmath146 leaves it invariant , the permutation of the four vertices will determine the gluing pattern accordingly . if we label the sides and edges of @xmath147 and @xmath148 as in fig .
[ two tetrahedrons ] , there must exist an isometry of @xmath146 such that the sides of @xmath148 , namely , @xmath149 , @xmath150 , @xmath151 , @xmath152 , are mapped onto those of @xmath147 , i.e. , @xmath153 , @xmath154 , @xmath155 , @xmath156 , and exactly in this order .
it can be proved that this side - pairing is proper , hence implies that the resulting space will be a hyperbolic 3-manifold , say @xmath145 , which is known as the figure - eight knot complement .
now by assuming the existence of systems on these solid regular tetrahedrons , a new dynamical system can then be constructed on the resulting manifold via the side - pairing if and only if the trajectories match up on the corresponding boundaries of the polyhedra components . as an example , fig .
[ two systems ] illustrates this matching up by applying the side - pairing that we mentioned above .
note that the explicit dynamics in ( a ) and ( c ) are obtained by repeating ( b ) and ( d ) on all sides and edges of @xmath147 and @xmath148 , respectively . +
the classical reeb foliation of the sphere and the torus are well - known ( see [ moerdijk & mrcun , candel & conlon , 2000 ] ) .
these are obtained first from a heegaard splitting of the sphere @xmath157 where @xmath33 is a solid torus and each copy of @xmath33 carries the foliation shown below in fig .
[ reeb foliation ] .
each leaf apart from the bounding torus is a plane immersed into the solid torus . in this paper
we shall show that an infinite set of dynamical systems exists on the 3-sphere which are formed by taking a genus @xmath158 ( for any @xmath159 ) heegaard splitting of @xmath160 and finding a generalized reeb foliation on the solid genus @xmath158 bounded 3-manifolds .
each leaf ( apart from the bounding genus @xmath158 surfaces and a singular leaf ) will be an unbounded surface of infinite genus .
of course , it is well - known that every compact three - manifold has a ( nonsingular ) foliation ( see [ candel & conlon , 2000 ] ) , essentially proved by dehn surgery on embedded tori , each of which carries a reeb component . however , this is an existence result and it is difficult to use to define explicit dynamical systems on three - manifolds .
we begin by describing a simple system on the torus which can be mapped onto each leaf of the reeb foliation to give a system on @xmath161 with an infinite number of equilibria . the basic system on the torus will consist of a source , a sink and two saddles as shown in fig .
[ torus system ] .
( note that the converse of the _ poincar@xmath162 _ index theorem is not true , so it is not possible to have just a source and a saddle on the torus , although their total index would be @xmath163 . )
consider a single noncompact leaf in the reeb foliation consisting of a ` rolled up ' plane as in fig .
[ a single leaf ] , ] the plane @xmath164 cuts the leaf @xmath165 into an infinite number of cylinders plus a disk . mapping the dynamics of fig .
[ torus system ] onto each cylinder and adding a source at the origin of the disk gives the system on the plane shown in fig .
[ source ] .
we shall organize the dynamics on the leaf so that the sources lie ` below ' the point @xmath29 on the torus when the leaf is folded up .
note that the size of the shaded region in fig .
[ source ] depends on the leaf and shrinks to zero with origin ` below ' @xmath29 as in fig .
[ shaded region ] .
we shall now show that there is a ( singular ) foliation of a 3-manifold of genus @xmath158 containing a compact leaf consisting of the bounding genus @xmath158 surface , an uncountable number of unbounded leaves of infinite genus and a set of one - dimensional singular leaves .
consider first the genus @xmath166 case .
consider the orientable @xmath27-manifold with boundary consisting of the closed surface of genus @xmath166 .
there is a singular foliation of this manifold defined by a dynamical system with a singular one - dimensional invariant submanifold , an infinite number of noncompact invariant submanifolds of infinite genus and a single leaf consisting of the boundary .
[ reeb - like system ] * proof .
* we obtain the foliation by modifying the reeb foliation and its associated dynamical system introduced above .
hence consider two systems of the form in fig .
[ source ] , where one has the arrows reversed ( i.e. we reverse time in the corresponding dynamical system ) .
we then form the connected sum of the bounding tori by removing a disk around the source ( or sink ) at the point @xmath29
. then we ` plumb ' each leaf in a similar way ( again removing the source or sink ) .
this will require one singular line joining the origins of the leaves which occur just ` below ' @xmath29 .
see fig .
[ gluing pattern ] for illustration .
the leaves clearly have the form stated in the lemma .
@xmath139 + _ remarks . _
the nonsingular leaves ( apart from the genus @xmath166 boundary surface ) are embeddings of the surfaces shown in fig . [ typical leaf ] .
note that we must have at least one singular fibre in order to introduce such a foliation on a higher genus surface .
for we have any foliation of codimension @xmath167 of a compact orientable manifold @xmath145 of dimension @xmath27 with finite fundamental groups and genus @xmath168 , which is transversally oriented , must have a singular leaf .
* by novikov s theorem ( see [ moerdijk & mrcun , 2003 ] ) , any codimension @xmath167 transversely orientable foliation of @xmath145 has a compact leaf and if @xmath145 is orientable , this compact leaf is a torus containing a reeb component .
thus , if @xmath145 contains a compact leaf of genus @xmath168 , it is not a torus and hence there must exist a singular leaf .
@xmath139 _ remarks .
_ we can find a similar singular foliation of a genus @xmath166 3-manifold by adding a handle between the stable and unstable points on the torus in fig .
[ torus system ] .
this gives a typical leaf shown in fig .
[ adding handles ] , rather than the one in fig .
[ typical leaf ] .
we now define systems on 3-manifolds by gluing two systems of the form above situated on solid genus-@xmath158 surfaces by the use of a heegaard diagram .
we first recall the general theory of heegaard splittings of 3-manifolds ( see , e.g. [ hempel , 1976 ] ) .
a _ heegaard splitting _ of genus @xmath158 @xmath169 of a 3-manifold @xmath145 is a pair of solid cubes with @xmath158 handles @xmath170 , @xmath171 such that @xmath145 is obtained from @xmath170 and @xmath171 by gluing @xmath172 to @xmath173 . using a simplicial decomposition of @xmath145 and a dual complex , it can be seen that any 3-manifold has a heegaard splitting .
let @xmath174 be pairwise disjoint properly embedded @xmath166 cells in @xmath171 which cut @xmath171 into a 3-cell .
then @xmath175 cut @xmath176 into a 2-sphere with @xmath177 holes .
we call @xmath178 a _ heegaard diagram _ of @xmath169
. we can get back to @xmath145 from a heegaard diagram in the following way : 1 .
attach a copy of @xmath179 to @xmath170 ( @xmath180 is the 2-ball , @xmath181 $ ] ) for each @xmath182 by identifying @xmath183 with a neighbourhood of @xmath184 in @xmath172 .
the resulting manifold @xmath185 has a 2-sphere boundary .
2 . attach a copy of @xmath146 ( @xmath1863ball ) to @xmath185 via @xmath187 to @xmath188 .
this gives @xmath145 .
we can now state for any @xmath27-manifold @xmath145 , and any @xmath189 , there is a reeb - like dynamical system on @xmath145 given by gluing two systems of the form given in lemma @xmath190 .
* proof . *
let @xmath169 be a heegaard splitting of @xmath145 of genus @xmath158 and let @xmath191 be dynamical systems defined on @xmath192 , respectively , of the form given in lemma [ reeb - like system ] .
let @xmath193 be the homeomorphism defined in ( i ) , ( ii ) above . by using c - homeomorphisms of the type in [ lickorish , 1962 ]
, we can assume that @xmath194 is smooth .
now let @xmath195 be a solid genus-@xmath158 handle - body contained within @xmath171 ( as in fig .
[ solid handle - body ] ) so that @xmath196 and @xmath197 is a solid genus-@xmath158 handle - body properly contained in @xmath171 .
we can extend @xmath194 to a smooth map @xmath198 by the homotopy handle - body contained within @xmath171,height=192 ] @xmath199 let @xmath200 be the vector field corresponding to @xmath201 on @xmath202 .
then we ` twist ' the dynamics on @xmath171 by @xmath194 , i.e. , @xmath203 and extend this to @xmath171 in an obvious way using ( [ homotopy ] ) .
then the dynamics on @xmath173 match those on @xmath172 according to the heegaard diagram and the result is proved .
in this paper , we have considered a variety of methods for generating systems on 3-manifolds . we have shown how to construct dynamical systems explicitly on hyperbolic 3-manifolds .
this is achieved by using a modified theta series to obtain the ` generalized ' automorphic functions which ` uniformize ' the vector fields on the manifold .
here we concentrated on using the upper half - space model for the hyperbolic space , while it is also possible to use the disk model .
also we gave an example of how to generate such systems .
also we consider constructing dynamical systems with the help of reeb foliation .
this is achieved by defining a system on each leave and then using the connected sum method to link them together . | in this paper , we consider different ways of generating dynamical systems on 3-manifolds .
we first derive explicit differential equations for dynamical systems defined on generic hyperbolic 3-manifolds by using automorphic function theory to uniformize the upper half - space model .
it is achieved via the modification of the standard _ poincar _
theta series to generate systems invariant within each individual fundamental region such that the solution trajectories match up on the appropriate sides after the identifications which generate a hyperbolic 3-manifold .
then we consider the gluing pattern in the conformal ball model . at the end
we shall study the construction of dynamical systems by using the reeb foliation . + * * key words:**automorphic functions , hyperbolic manifolds , upper half - space model , _ poincar _ theta series , conformal ball model , reeb foliation , heegaard splittings .
[ section ] [ section ] [ section ] [ section ] |
one exciting aspect of dense stellar systems is the simultaneous importance of three principal areas of stellar astrophysics : dynamics , evolution , and hydrodynamics .
many simulation codes focus on one of these areas and have often been lifelong works in progress .
the first attempts at unifying these treatments into a coherent model to describe clusters have begun only recently . in june 2002 ,
specialists in stellar dynamics , stellar evolution , hydrodynamics , cluster observation , visualization , and computer science gathered at the american museum of natural history in new york city to begin discussing a framework for modelling dense stellar systems , without having to modify existing stellar codes extensively .
the workshop - style meeting , organized by p. hut and m. shara , became known as modest-1 @xcite .
the second such meeting , modest-2 , was organized by s. portegies zwart and p. hut and held in december 2002 at the anton pannekoek institute in amsterdam @xcite . from modest-2 ,
a set of eight `` working groups '' were established , each focusing on a different aspect of the modest endeavour . attempting to integrate stellar dynamics , evolution , and hydrodynamics codes into one fully functional package
will be challenging , largely because each area treats stellar properties that evolve on different time - scales .
however , by combining these areas , we will be able to better model the origins , dynamics , evolution , and death of globular clusters , galactic nuclei , and other dense stellar systems . in this paper ,
our focus is on modelling hydrodynamic interactions between stars .
the goal is to develop a software module for quickly generating collision product models , ultimately for any type of stellar collision , that could be incorporated into simulations of dense star clusters .
@xcite presented an appropriate formulation for treating parabolic single - single star collisions between low - mass main - sequence stars . here
we extend that study to include situations in which one of the parent stars is itself a thermally _ un_relaxed collision product .
such scenarios can occur during binary - single or binary - binary interactions , when the time between collisions is much less than the thermal relaxation time - scale .
stellar collisions and mergers can strongly affect the overall energy budget of a cluster and even alter the timing of important dynamical phases such as core collapse .
furthermore , hydrodynamic interactions are believed to produce a number of non - canonical objects , including blue stragglers , low - mass x - ray binaries , recycled pulsars , double neutron star systems , cataclysmic variables , and contact binaries .
such stars and systems are among the most challenging to model , but they are also among the most interesting observational markers .
blue stragglers , for example , exist on an extension of the main - sequence , but beyond the turnoff point .
blue stragglers are therefore appropriately named , as they are more blue than the remaining ordinary main - sequence stars , and , compared to other stars of similar mass , are straggling behind in their evolution .
this aberration from the common path of stellar evolution is believed to be due to mass transfer or merger in a binary system , or from the direct collision of two or more main - sequence stars ( for a review , see * ? ? ?
predicting the numbers , distributions , and other observable characteristics of stellar exotica will be essential for detailed comparisons with observations .
stellar dynamics codes determine the motions of stars .
the primary approaches to evolving clusters or galactic nuclei dynamically are direct n - body integrations , solving the fokker - planck equation ( e.g. , * ? ? ?
* ) , monte carlo approaches ( e.g. , * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and gaseous models ( e.g. , * ? ? ?
for a review of the ongoing nbody effort for accurate n - body simulations , see @xcite ; for a general review of cluster dynamics , see and @xcite .
the most important quantities that a stellar evolution software module can provide to a dynamics module are the stellar masses , as well as the stellar radii if collisions are included .
at least in principle , these results could come from a live ( i.e. , concurrent with the cluster dynamics ) stellar evolution calculation , from fitting formulae , or from interpolation among prior calculations .
due to the large number of stars , it would be wasteful to expend a considerable amount of time for a live computation of each star s evolution . for the ordinary stars
whose evolution is not perturbed by an event such as a collision , it is much more efficient and entirely appropriate to interpolate among , or to use analytic fitting formulae based upon , previously calculated evolutionary tracks . the parameter space associated with non - canonical stars , however , is too enormous to be adequately covered by interpolation or fitting formulae , and it will ultimately be necessary to invoke a full stellar evolution calculation in parallel with the stellar dynamics for such stars . although live stellar evolution calculations have not yet been combined with stellar dynamics codes , some parametrized codes , such as seba @xcite , sse @xcite and bse @xcite , have been successfully integrated .
when physical collisions between stars are modelled in a cluster calculation or a scattering experiment , it is usually done using a method known as `` sticky particles , '' in which a collision product is given a mass equal to the combined mass of the two parent stars and a velocity determined by momentum conservation . in a collision between two main - sequence stars , for example , the result would be modelled as a rejuvenated , thermally relaxed main - sequence star .
this simple method is a reasonable first approximation for many situations .
however , there are important characteristics of collision products that are neglected , including their rapid rotation , peculiar composition profiles , and enhanced sizes due to shock heating .
if the thermal relaxation time - scale of the collision product is much less than the time until its next collision , then it is appropriate to assume the product becomes instantaneously thermally relaxed , as is done in the classic simulations of @xcite .
this approximation becomes questionable when the collision has been mediated by a binary , as there is then at least one star in the immediate vicinity of the collision product and the likelihood of a subsequent collision will depend sensitively on the product s thermally _ un_relaxed size .
binaries are subject to enhanced collision rates for two primary reasons : ( 1 ) their collision cross section depends on the semi - major axis of the orbit , as opposed to the radius of a single star , and ( 2 ) due to mass segregation , binaries tend to be found in the core of the cluster , the densest and most active region . in clusters with a binary fraction exceeding about 20 per cent , binary - binary collisions are expected to occur more frequently than single - single and binary - single collisions combined @xcite .
it is probably not uncommon for binary fractions to be this large : the inner core of ngc 6752 , for example , is thought to have a binary fraction in the range 1538 per cent @xcite .
binary populations can lead to complex and chaotic resonant interactions .
these interactions tend to exchange energy between the binaries and the other stars in the cluster , and therefore are critical in determining its dynamics and observable characteristics @xcite .
a star intruding on a binary could , depending on parameters such as the separation of the binary and the velocity of the incoming star , escape to infinity , destroy the binary , form a new binary with a star from the original , or form a triple .
the outcomes of three body encounters can be categorized using a nomenclature based upon typical atomic processes ( see * ? ? ?
* who introduced the use of terms like ionization and exchange , to describe resultant scenarios ) .
see @xcite for an informative narration of the intimate interactions , usually involving binaries , that stars regularly undergo during cluster evolution .
the starlab computing environment is a very useful tool for modelling and analyzing all types of stellar phenomena .
the general technique for using starlab to determine the cross sections or branching ratios for the various outcomes of binary interactions is presented by @xcite .
they also highlight an example set of cases in which a @xmath1 star intrudes upon a binary with a @xmath2 primary and a @xmath3 secondary .
their assumed mass radius relation , @xmath4 , is appropriate for thermally relaxed main sequence stars and is applied both to the parent stars and any collision products .
they find that for binaries with semimajor axes of 0.2 , 0.1 , 0.05 , and 0.02 au , triple - star mergers comprise about 1 , 2 , 5 , and 15 per cent , respectively , of all merger events .
as the results of the present paper will help show , we expect that accounting for the enhanced , thermally unrelaxed size of the first collision product will greatly increase these percentages as well as the range of semimajor axes in which triple - star mergers are significant .
simulations of moderately dense galactic nuclei initially containing solar - mass main - sequence stars demonstrate that runaway mergers can readily produce stars with masses @xmath5 .
these massive stars then undergo further mergers to produce seed black holes with masses as large as @xmath6 @xcite
. this process may be responsible for massive black holes at the centres of most galaxies , including our own . for star clusters , recent n - body simulations
reveal that runaway mergers can lead to the creation of central black holes within a few million years ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . with the help of monte carlo simulations , @xcite
show that the runaway process will occur in a typical cluster with a relaxation timescale less than about 30 myr .
observational evidence for a possible intermediate - mass black hole in m15 has been recently reported by @xcite , although the data is more reasonably modelled with a large concentration of stellar - mass compact objects @xcite . mass loss and expansion due to shock heating when two stars collide are examples of hydrodynamical processes that can ultimately affect the future evolution of the cluster . mostly using the smoothed particle hydrodynamics ( sph , see [ sph ] ) method , numerous scenarios of stellar collisions and mergers have been simulated in recent years , including collisions between two main - sequence stars @xcite , collisions between a giant star and compact object @xcite , and common envelope systems @xcite .
the first published sph calculations of three - body encounters were done by @xcite , who performed over 100 very low resolution simulations and implemented a mass - radius relation appropriate for white dwarfs . other three- and four - body interaction simulations include binary - binary encounters among @xmath7 polytropes @xcite as well as neutron star
main - sequence binary encounters with a neutron star , main - sequence or white dwarf intruder @xcite .
see @xcite for more information concerning the use of sph in stellar collisions , and see @xcite for a qualitative overview of the progress in stellar collision research .
if the structure and composition profiles of colliding stars were available ( perhaps from a live stellar evolution calculation ) during a cluster simulation , then the sticky particle method could be replaced by a more detailed hydrodynamics module .
sph calculations could then , at least in principle , be run on demand within this cluster simulation in order to determine the orbital trajectory of the product(s ) , as well as their structure and chemical composition distributions .
however , at least @xmath8 sph fluid particles may be necessary to allow an accurate treatment of the subsequent evolution of collision products @xcite .
the trouble , therefore , is that the integration of just a single interaction could consume hours , days or even weeks of computing time ( depending on the initial conditions , desired resolution , and available computational resources ) .
although the use of equal - mass particles , or the more accurate sph equations of motion derived by @xcite , or both , could decrease the total number of particles required , it is still currently impractical to implement a full hydrodynamics calculation for every close stellar encounter during a cluster simulation .
one approach for incorporating strong hydrodynamic interactions and mergers into a grand simulation of a cluster , already successfully implemented by @xcite in the context of galactic nuclei , is to interpolate between the results of a set of previously completed sph simulations .
the sph database of freitag & benz treats all types of hyperbolic collisions between main - sequence stars ( mergers , fly - bys and cases of complete destruction ) , while also varying the parent star masses as well as the eccentricity and periastron separation of their initial orbit .
the tremendous amount of parameter space surveyed precludes having high enough resolution to determine the detailed structure and composition profiles of the collision products for all cases ; however , critical quantities such as mass loss and final orbital elements can be determined accurately .
a second possibility is to forgo hydrodynamics simulations and instead model collision products by physically motivated algorithms and fitting formulae that sort the fluid from the parent stars @xcite .
one advantage of such an approach is that it can handle cases in which one or both of the parent stars is itself a former collision product ( with chemical and structural profiles that are substantially different than that of a standard isolated star of similar mass and type ) . in this paper , we use both sph calculations and a much faster fluid sorting algorithm to study scenarios in which a newly formed collision product collides with a third parent star . by varying the order and orbital parameters of the collision , we investigate how factors such as shock heating affect the chemical composition and structure profiles of the collision product .
section [ procedure ] presents our procedures and numerical methods , both for our sph calculations ( [ sph ] ) and our fluid sorting algorithm ( [ sorting ] ) .
sph results are presented in
[ sphresults ] , and then compared to the results of our fluid sorting algorithm in [ mmasresults ] . in
[ discussion ] we discuss our findings and possible directions for future work .
one means by which we generate collision product models is with the parallel sph code used in @xcite .
the original serial version of this code was developed by @xcite , specifically for the study of stellar interactions such as collisions and binary mergers ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. introduced by @xcite and @xcite , sph is a hydrodynamics method that uses a smoothing kernel to calculate local weighted averages of thermodynamic quantities directly from lagrangian fluid particle positions ( for a review , see * ? ? ?
. each sph particle can be thought of as a parcel of gas that traces the flow of the fluid , with the kernel providing each particle s spatial extent and the means by which it interacts with neighbouring particles .
the sph code solves the equations of motion of a large number of particles moving under the influence of both hydrodynamic and self - gravitational forces .
all of the scenarios we investigate using sph involve a @xmath2 parent star , represented with 12800 equal - mass sph particles , and two @xmath1 parent stars , each represented with 9600 equal - mass particles .
each of our sph particles therefore has a mass of @xmath9 . for comparison ,
this particle mass is between the masses of the central particles used in the @xmath10 and @xmath11 calculations of @xcite , who used unequal - mass particles to study in detail the outer layers of the fluid in a collision between two @xmath12main - sequence stars . for our purposes
, the use of equal - mass particles is more appropriate , as it allows for higher resolution in the stellar cores and does not waste computational resources on the ejecta .
local densities and hydrodynamic forces at each particle position are calculated by appropriate summations over @xmath13 nearest neighbours .
the size of each particle s smoothing kernel determines the local numerical resolution and is adjusted during each time step to keep @xmath13 close to a predetermined value , 48 for the present calculations .
neighbour lists for each particle are recomputed at every iteration using a linked - list , grid - based parallel algorithm @xcite .
the hydrodynamic forces acting on each particle include an artificial viscosity contribution that accounts for shocks . as in @xcite , we adopt the artificial viscosity form proposed by @xcite , with @xmath14 , @xmath15 , and @xmath16 .
this form treats shocks well and has the tremendous advantage that it introduces only relatively small amounts of spurious shock heating and numerical viscosity in shear layers @xcite .
a number of physical quantities are associated with each sph particle , including its mass , position , velocity , and entropic variable @xmath17 .
here we adopt a monatomic ideal gas equation of state , appropriate for the stars in our mass range .
that is , @xmath18 , where the adiabatic index @xmath19 with @xmath20 and @xmath21 being the pressure and density , respectively .
the entropic variable is closely related ( but not equal ) to specific entropy : both of these quantities are conserved in the absence , and increase in the presence , of shocks .
our code uses an fft - based convolution method to calculate self - gravity .
the fluid density is placed on a zero - padded , 3d grid by a cloud - in - cell method , and then convolved with a kernel function to obtain the gravitational potential at each point on the grid .
gravitational forces are calculated from the potential by finite differencing , and then interpolated for each particle using the same cloud - in - cell assignment scheme . for each collision simulation in this paper ,
the number of grid cells is @xmath22 .
the ejecta leaving the grid interact with the enclosed mass simply as if it were a monopole .
following the same approach as in @xcite , we begin by using a stellar model from the yale rotational evolution code ( yrec ) to help generate sph models of the parent stars .
we focus on collisions involving 0.8 and @xmath1 main - sequence stars , with a primordial helium abundance @xmath23 and metallicity @xmath24 . using yrec ,
these stars were evolved with no rotation to an age of 15 gyr , the amount of time needed for the @xmath0 star to reach turnoff .
the total helium mass fractions for the @xmath25 and @xmath2 parent stars are 0.286 and 0.395 , and their radii are @xmath26 and @xmath27 , respectively .
see figs . 1 and 2 of @xcite for thermodynamic and composition profiles of the parent stars presented as a function of enclosed mass , as determined by yrec .
to generate our sph models , we use a monte carlo approach to distribute particles according to the desired density distribution , determining values of @xmath17 for each sph particle from its position . to minimize numerical noise , an artificial drag force
is implemented , with artificial viscosity turned off , to relax each sph parent model to the equilibrium configuration used to initiate the collision calculations .
fourteen different chemical abundance profiles are available from the yrec parent models to set the composition of the sph particles .
the abundances of an sph particle are assigned according to the amount of mass enclosed by an isodensity surface passing through that particle in the relaxed configuration .
[ refpar2 ] plots fractional chemical abundances ( by mass ) versus @xmath28 in each parent star in its relaxed sph configuration .
note that the dense core of the turnoff star is at a smaller @xmath17 , and its diffuse outer layers are at a larger @xmath17 , than all of the fluid in the @xmath1 star , which has direct consequences for the hydrodynamics of collisions involving these stars . also note that lithium and beryllium exist only in the outermost layers of the parent stars . , c@xmath29 , c@xmath30 , n@xmath31 , o@xmath32 , li@xmath33 , li@xmath34 , and be@xmath35 versus @xmath28 , where the entropic variable @xmath36 is in cgs units , for our @xmath1 ( solid curve ) and @xmath2 ( dotted curve ) parent stars , as determined by yrec . , width=317 ]
we focus on triple - star collisions , modelling each collision separately and in succession .
we do not consider fly - bys or grazing collisions in our sph calculations : all of our collisions lead to mergers .
we neglect any direct or indirect effects , including tidal forces , that the third star may have on the dynamics of the first collision .
we assume that the second collision occurs before the first collision product thermally relaxes : a reasonable approximation since contraction to the main - sequence occurs on a thermal time - scale , lasting at least @xmath37 yr for non - rotating products and as long as @xmath38 yr for rapidly rotating products @xcite , much longer than the typical time between collisions in some binary - single or binary - binary interactions ( but see [ future ] ) . the orbital trajectory in all our collisions is taken to be parabolic .
this is clearly not appropriate for galactic nuclei , where collisions are typically hyperbolic .
however , in globular clusters , the velocity dispersion is only @xmath39 km s@xmath40 , much less than the 600 km s@xmath40 escape speed from the surface of our @xmath2 turnoff star , and hence all single - single star collisions are essentially parabolic . for collisions involving binaries ( even including some hard binaries ) , the escape speed can still be large compared to the effective relative velocity at infinity .
for example , consider two @xmath2 turnoff stars in a circular orbit of radius 0.05 au in a globular cluster .
this is a hard binary , as each star moves with a velocity of about 60 km s@xmath40 with respect to the center of mass of the binary , a speed significantly larger than the cluster velocity dispersion .
yet , the effective relative velocity at infinity for a collision between one of the binary components and an intruder would typically not be much more than the orbital speed , and therefore still significantly less than the escape speed from our turnoff star .
we therefore expect collisions between a slow intruder and a binary component to be close to parabolic not only for all soft binaries , but also for some ( moderately ) hard binaries for the first single - single star collision , the stars are initially non - rotating and separated by 5 @xmath41 , where @xmath42 is the radius of our turnoff star .
the initial velocities are calculated by approximating the stars as point masses on an orbit with zero orbital energy and a periastron separation @xmath43 .
a cartesian coordinate system is chosen such that these hypothetical point masses of mass @xmath44 and @xmath45 would reach periastron at positions @xmath46 , @xmath47 , where @xmath48 and @xmath49 refers to the more massive star .
the orbital plane is chosen to be @xmath50 . with these choices , the centre of mass resides at the origin .
for the first collisions , the gravity grid maintains a fixed spatial extent from -4@xmath41 to + 4@xmath41 along each dimension . around the @xmath51-axis and
then by an angle @xmath52 around the @xmath53-axis .
the above figure shows how a vector @xmath54 initially pointing along the @xmath53-axis is transformed into the new vector @xmath55 by these rotations .
, width=317 ] for the second collision , we want to control the relative orientation of the first collision product s rotation axis and the orbital plane ( or , equivalently , the direction of approach of the third parent star ) . to do so , we begin with the final state of the first collision and make two rotations to its particle positions and velocities , through the angles @xmath56 and @xmath52 .
more specifically , the first rotation is clockwise through an angle @xmath56 about the @xmath51 axis , while the second rotation is clockwise through an angle @xmath52 about the @xmath53 axis ( see fig .
[ rotate ] ) .
finally , the particle positions and velocities are uniformly shifted parallel to the @xmath51-@xmath57 plane , and the third star is introduced such that the system s centre of mass will remain at the origin and the periastron positions ( in the two body point mass approximation ) will occur on the @xmath51 axis . in order to allow the bulk of the fluid to remain within the gravity grid
, the grid is extended up to a full width of @xmath58 in the @xmath51 and @xmath57 directions for some of the second collisions .
we use the same iterative procedure as @xcite to determine the bound and unbound mass .
sph structure and composition profiles presented in this paper result from averaging in 100 equally sized bins in the bound mass .
unfortunately , it is extremely difficult to use sph simulations to specify the equilibrium structure of the outermost few per cent of mass in any collision product .
some sph particles , although gravitationally bound , are ejected so far from the system s centre of mass that it would take many dynamical time - scales for them to rain back onto the central product and settle into equilibrium .
our requirement for stopping an sph calculation is that the entropic variable @xmath17 , when averaged over isodensity surfaces , increases outward over at least the inner 95 per cent of the bound mass in the first collision product , and at least 92 per cent in the second collision product .
many calculations are run longer in order to confirm that no rapid changes are still occurring in the structure and chemical composition distributions .
the results of parabolic collisions between low - mass main - sequence stars can be well explained by simple physical arguments . to a good approximation , the fluid from the parent stars sorts itself such that fluid with the lowest values of @xmath17 sinks to the core of the collision product while the larger @xmath17 fluid forms its outer layers .
therefore , the interior structure and the chemical composition profiles of the collision product can be predicted accurately using simple algorithms , instead of hydrodynamic simulations .
based on these ideas , @xcite have recently created a publicly available software package , dubbed make me a star ( mmas ) .
this package produces collision product models close to those of an sph code in considerably less time , while still accounting for shock heating , mass loss , and fluid mixing .
sorting the shocked fluid according to its entropic variable @xmath17 gives the @xmath17 profile of the collision product as a function of the mass @xmath59 enclosed inside an isodensity surface . in the case of the non - rotating products formed in head - on ( @xmath60 ) collisions ,
knowledge of the @xmath61 profile is sufficient to determine the pressure @xmath62 , density @xmath63 , and radius @xmath64 profiles .
using the @xmath17 profile determined by sorting , mmas numerically integrates the equation of hydrostatic equilibrium with @xmath65 to determine the @xmath21 and @xmath20 profiles , which are related through @xmath66 .
the outer boundary condition is that @xmath67 when @xmath68 , where @xmath69 is the desired ( gravitationally bound ) mass of the collision product .
the virial theorem provides a check of the resulting profiles .
this approach allows for the quick generation of collision product models , without hydrodynamic simulations , and has already been tested with single - single star collisions . presented in this paper
are the results from mmas for triple - star collisions ( see [ mmasresults ] ) .
our procedure is simple .
we call the mmas routine twice , using the output model from the first collision as one of the input parent models in the second . these mmas calculations therefore account for the differences in shock heating that arise from changing the order , or the periastron separations , or both , of the collisions . in addition to investigating all of the scenarios considered with the sph code
, we also use mmas to examine more completely how the sizes of products vary with the periastron separations of the collisions .
furthermore , we include a @xmath3 parent star of radius @xmath70 , whose structure is determined by yrec under the same conditions described in [ sph ] . for an off - axis collision , knowledge of the specific angular momentum distribution in the collision product
is necessary to determine its structure fully , which by itself is a challenging problem .
although mmas outputs an approximate specific angular momentum profile of the first collision product , we use only its entropic variable @xmath61 profile to help initiate the second collision .
that is , for one of the parent stars in the second collision , we always give as input to mmas the structure of a non - rotating star with the desired @xmath61 profile , a simplification that both eases and quickens computations .
the validity of this approximation is supported by the sph calculations presented in [ spin ] .
@xcite implemented version 1.2 of mmas , while the results of this paper use version 1.6 . besides cosmetic changes , the primary enhancement is that the structure of the collision product is integrated with a fehlberg fourth - fifth order runge - kutta method .
in addition , we have fine - tuned the fitted parameter @xmath71 from its previous value of -1.1 to the new value -1.0 , which has the effect of distributing shocks slightly more uniformly throughout the fluid .
table [ singlesingle ] summarizes five single - single star [ cols="^,^,^,^,^,^,^,^,^ " , ] from table [ singlesingle ] ( referred to as the first collision product ) is collided with a third parent star .
the table shows : the case number ; the name of the single - single collision that yielded the first collision product ; the mass @xmath72 of the third ( @xmath73 ) parent star ; the periastron separation @xmath74 of the second collision ; the rotation angles @xmath56 and @xmath52 ( see fig . [ rotate ] ) ; the time @xmath75 when the calculation was terminated ; the ratio of kinetic to gravitational binding energy @xmath76 in the centre - of - mass frame of the final sph collision product model ; the average radius of the isodensity surface enclosing 90 per cent of the bound mass , as calculated by sph , @xmath77 , and by mmas , @xmath78 ; and the mass of the product as calculated both by sph , @xmath79 , and by mmas , @xmath69 .
all twenty cases presented in table [ second ] involve two @xmath80 stars and a single @xmath2 star .
if mass loss were neglected completely , the mass of the final collision product would therefore simply be @xmath81 .
the sph calculated masses range from about 1.76 to @xmath82 , with the largest mass loss occurring for cases with successive head - on collisions . the cases in table [ second ] group naturally together in a variety of ways . cases 1 and 2
each involve two head - on collisions .
cases 5 and 6 differ only in the orientation of the first collision product s spin axis , and an identical statement can be made for cases 7 through 10 , as well as for cases 14 through 19 .
cases 2 , 11 , and 7 differ only in the periastron separation of the first collision , as do cases 12 , 13 and 14 . also
, many of the cases differ only in the periastron separation for the second collision : for example , cases 7 , 5 , and 14 , as well as cases 10 , 6 , 17 , and 20 . even without running an sph or mmas calculation
, one can generate a `` zeroth order '' collision product model , valid for all twenty cases , simply by sorting the fluid of the three parent stars by their @xmath17 values , with @xmath17 increasing from the core to the surface . in those regions in which more than one parent star contributes ,
chemical abundances can be determined by an appropriate weighted average : the fraction of fluid with entropic variable in some range @xmath83 that originated from any one parent star is just equal to the fluid mass in that same range from that star divided by the total fluid mass in that range from all three stars .
[ zeroth_order ] shows the composition profiles resulting from for a `` zeroth order '' collision product model generated by sorting the fluid from one @xmath84 and two @xmath1 stars according to their @xmath17 values , without accounting for mass loss , shock heating or fluid mixing . here
@xmath59 is the mass enclosed within a surface of constant density and @xmath85 is the total mass of the collision product , @xmath86 in this model .
, width=317 ] this exercise for the merger of two @xmath1 stars and a @xmath87 star , with shock heating , mass loss , and fluid mixing all completely neglected . the innermost 6 per cent and outermost 9 per cent of this collision product model consist of fluid that originated entirely in the @xmath2 parent star , and the profiles there consequently mimic those of the innermost and outermost regions , respectively , of that parent .
the profiles in the intermediate region , where all three parent profiles contribute , can be understood by looking back at fig .
[ refpar2 ] and remembering that the smaller @xmath17 fluid is placed deeper in the product .
for example , the c@xmath29 near @xmath88 originated in the @xmath1 star , while the higher @xmath17 , c@xmath29-rich fluid peaked near @xmath89 originated mostly in the @xmath2 star .
although some of the composition profiles turn out to agree reasonably with our more precise sph and mmas calculations , the structure of the star produced by this simple merging procedure does not .
in particular , because shock heating is neglected , energy is not conserved during the merger , and the resulting radius for the product is a considerable underestimate . the primary usefulness of the model presented in fig .
[ zeroth_order ] is that it serves as a reference to help us evaluate in what ways the shock heating , mass loss , and fluid mixing treated by our sph and mmas calculations affect the composition and structure profiles of the collision product .
cases 1 and 2 each involve two consecutive head - on collisions between the same three parent stars .
the only variation between these two cases is the order in which the collisions occur : in case 1 , the @xmath2 parent is involved in the first collision , while in case 2 it is involved in the second .
[ chem12 ] shows the resulting chemical composition profiles of the collision products . because the innermost few per cent of the final collision products consist of low-@xmath17 fluid that originated in the centre of the @xmath0 parent ( see fig .
[ frac12 ] ) , the composition profiles are nearly identical there . more generally , the resulting profile of each chemical species is at least qualitatively similar throughout the products .
the differences in the composition profiles of fig .
[ chem12 ] are arguably most pronounced for c@xmath29 . in the parent stars
, this element exists in appreciable amounts only in a relatively thin shell , and , as shown in fig .
[ refpar2 ] , this shell is at a higher value of @xmath17 in the @xmath2 star than in the @xmath1 star .
exactly where the c@xmath29-rich fluid is ultimately deposited depends on the details of the shock heating , and hence the order in which the stars collide .
in particular , the final c@xmath29 profile in the case 1 product has two distinct peaks at enclosed mass fractions of @xmath90 0.1 and 0.5 ( as in our zeroth order model , see fig [ zeroth_order ] ) , whereas the case 2 profile has a single extended peak centred near @xmath91 = 0.2 .
[ f12c13 ] displays , as a function of enclosed mass fraction @xmath91 within the product , how each parent contributes to the overall c@xmath29 profile .
the inner peak in the case 1 profile is due mostly to c@xmath29 that originated in the @xmath12 parent stars , while the outer peak is due mostly to the higher-@xmath17 , c@xmath29-rich fluid from the @xmath0 parent . in case 2 ,
the @xmath0 star is involved in only the second collision .
it therefore experiences less shock heating than in case 1 and more of its fluid is able to penetrate to the core of the final collision product .
consequently , much of the c@xmath29 from the @xmath1 stars is displaced out to larger enclosed mass fractions , while the c@xmath29 from the @xmath2 star is shifted inward .
the net result is the single extended peak that includes c@xmath29 from all three parent stars . in the final collision product.,width=317 ] to the mass of the collision product from each parent star as a function of enclosed mass fraction @xmath91 within the product for case 1 ( top ) and case 2 ( bottom ) , as determined by sph calculations .
each of these cases involves head - on ( @xmath92 ) collisions among one @xmath84 and two @xmath93 stars ; however in case 1 the @xmath2 star is part of the initial collision , whereas in the case 2 scenario it is part of the second collision .
different line types are used for each parent star : @xmath49 ( solid curve ) , @xmath94 ( dotted curve ) , and @xmath73 ( dashed curve ) .
parents with an index @xmath49 or 2 are involved in the first collision , while @xmath73 refers to the third parent star from the second collision . the contribution profile from the @xmath2 parent
is labelled in each case .
, width=317 ] versus the enclosed mass fraction @xmath91 in the final collision products of case 1 ( dashed curve ) and case 2 ( dotted curve ) , as determined by sph calculations .
the top pane shows the total c@xmath29 abundance .
the second pane shows the contribution from the @xmath2 parent , and the bottom two frames show the contribution from the @xmath1 parents , so that the curves in the bottom three panes add up to give the overall profile shown in the top pane .
, width=317 ] the detailed differences between the composition profiles of the other elements in fig .
[ chem12 ] can be understood by considering the @xmath17 and composition profiles of the parent stars , along with the distribution and amount of shock heating .
for example , the core of the @xmath87 parent star is rich in he@xmath95 and n@xmath31 , but depleted of c@xmath30 ( see fig .
[ refpar2 ] ) . the lower shock heating to the @xmath2 star in case 2 allows more of its core to sink to the centre of the collision product .
consequently , the he@xmath95 and n@xmath31 levels are enhanced as compared to case 1 , while the c@xmath30 levels are diminished , for final enclosed mass fractions @xmath91 in the range from 0.05 to 0.2 .
the mass that is ejected during the collisions comes preferentially from the outer layers of the parent stars , exactly where elements such as li@xmath33 , li@xmath34 , and be@xmath35 exist .
the surface abundances ( by mass ) in the final case 1 collision product for these three elements are approximately @xmath96 , @xmath97 , and @xmath98 , which is , respectively , about 30 , 6 , and 3 times less than at the surface of the @xmath0 parent star . in case 2 ,
the surface layers are comparably depleted in these elements : the corresponding abundances are @xmath98 , @xmath99 , and @xmath98 .
the bottom three panes in fig .
[ chem12 ] show that the distribution of li@xmath33 , li@xmath34 , and be@xmath35 does differ somewhat between cases 1 and 2 .
because the @xmath2 star suffers less shock heating in case 2 than in case 1 , it loses less of its mass as ejecta and , consequently , can contribute more li@xmath33 , li@xmath34 , and be@xmath35 to the outermost layers of the final collision product .
furthermore , when the fluid containing li@xmath34 and be@xmath35 from the @xmath1 stars is involved in both collisions ( case 2 ) , it is shocked more and ultimately either ejected or deposited in the outer @xmath10010 per cent of the product .
however in case 1 , the one @xmath1 star that is involved in only a single collision can deposit its li@xmath34 and be@xmath35 of comparatively low-@xmath17 deeper in the product , resulting in flattened profiles extending further into the interior .
[ frac12 ] shows , as a function of @xmath91 , the fractional contribution to the final product s mass from each of the three parent stars for cases 1 and 2 , as determined by sph calculations . in each case , the innermost few per cent of the final collision product consists of low-@xmath17 fluid that originated in the centre of the @xmath0 parent .
because the first collision in cases 1 and 2 is head - on ( @xmath60 ) , fluid from the first two parent stars is _ not _ distributed axisymmetrically in the first collision product ( the composition distribution is therefore not axisymmetric , even though the structure of the first product is ) . in case 2 , the @xmath2 star strikes the first collision product on the side with fluid from the first ( @xmath49 ) @xmath1 parent .
fluid from the first @xmath1 parent is therefore heated more than fluid from the second , and the former is buoyed out to larger enclosed mass fractions in the final product . in off - axis collisions , rotation induces shear mixing , so that if two identical stars are involved in the first collision , they contribute essentially equally within the final product : @xmath101 .
the profiles of fig .
[ chem12 ] and fig .
[ f12c13 ] demonstrate that the order in which the stars collide can influence shock heating enough to affect , at least slightly , the chemical composition distribution within the final collision product . while the difference in resulting chemical composition profiles is small , fig .
[ slog12 ] shows that the difference in the structure of the collision product would be completely negligible for most purposes .
although changing the order of these head - on collisions affects how the shock heating is distributed ( and hence where any particular fluid element settles ) , it does not greatly affect the overall amount of heating that occurs nor the amount of mass that is ejected .
at least for low mach number collisions ( as with parabolic collisions ) that are nearly head - on ( so that the merger occurs quickly ) , shock heating can be thought of as a mild perturbation ; consequently , the final @xmath17 profile , and hence the structure of the final product , is not sensitive to the collision order in such cases .
, the natural logarithm of the average entropic variable @xmath17 , and the base 10 logarithm of the average density @xmath21 , are all plotted as a function of the average distance @xmath102 from the centre of the collision product to an isodensity surface .
units are cgs.,width=317 ] we now investigate how the direction of approach of the third star toward the first collision product affects the final collision product .
one might wonder , for example , whether an impact in the first collision product s equatorial plane ( @xmath103 or 180@xmath104 ) would yield a qualitatively different result than if the impact had instead occurred on the rotation axis .
cases 5 and 6 , cases 7 to 10 , and cases 14 to 19 can all be used to examine such effects , as the cases within each set differ only in the angles @xmath56 and @xmath52 , by which the first product is rotated ( see fig . [ rotate ] ) .
we find that while the spin of the final product is of course sensitive to such variations ( e.g. , see the @xmath76 column of table [ second ] ) , the composition profiles are nearly unaffected . fig .
[ chem14151719 ] shows the chemical abundance star , with a different orientation of the first collision product s rotation axis in each case .
[ chem14151719 ] , width=317 ] profiles of the collision product resulting in four cases in which the angle of approach of the third star is varied ( cases 14 , 15 , 17 , and 19 ) .
each of these cases involves an off - axis collision between a case k collision product and a @xmath2 star . in case 14 , the first collision product s spin vector is parallel to the orbital angular momentum of the second collision . in the other cases , the case k collision product
is tilted various ways according to the values of @xmath56 and @xmath52 listed in table [ second ] . in case 19 , for example , the case k collision product is flipped over 180@xmath104 so that it rotates in an opposite direction to that of the case 14 rotation . in case 14
, the fluid of the first product is rotating with the third star s motion as it impacts ( @xmath105 ) . consequently , the merger process is relatively gentle . for larger @xmath56 ,
the relative impact velocity is larger and the merger is somewhat more violent .
cases 14 , 17 , 15 , and 19 have @xmath56 values of 0 , 45 , 90 , and 180@xmath104 , respectively ; as @xmath56 increases , slightly less fluid from the @xmath0 star can sink down into the core of the final collision product , and the c@xmath29 profile rises at a slightly smaller enclosed mass fraction @xmath91 ( see fig . [ chem14151719 ] ) . nevertheless , as shown in fig .
[ f14151719 ] , the contribution of each from the third parent star as a function of enclosed mass fraction @xmath91 within the collision product of cases 14 ( solid curve ) , 15 ( dotted curve ) , 17 ( long dashed curve ) , and 19 ( short dashed curve ) , as determined by sph calculations . in these scenarios ,
the first two parent stars are both @xmath1 ; the third parent star is @xmath2 and approaches from a different angle @xmath56 relative to the rotation of the first collision product in each case .
the fractional contribution from each of the first two parent stars is essentially equal and can therefore be determined easily from the @xmath106 curve : @xmath107 .
[ f14151719 ] , width=317 ] parent star to the product varies very little from case to case .
consequently , the chemical profiles in the collision products also vary little as @xmath56 is changed . indeed , the he@xmath95 , c@xmath30 , n@xmath31 , and o@xmath32 profiles in fig .
[ chem14151719 ] all look remarkably similar to the corresponding profiles in fig .
[ zeroth_order ] for our simple , zeroth order model . however , the c@xmath29 profile has a single broad peak , for the same reasons as in case 2 .
furthermore , because of mass loss , the beryllium and lithium surface abundances are found to be greatly less than our zeroth order model ( that neglects mass loss ) would indicate .
the structure of the final collision product ( see fig .
[ slog14151719 ] ) can be affected by the direction of approach .
[ slog14151719 ] , width=317 ] for two primary reasons .
firstly , having larger relative velocity at impact leads to larger shock heating .
notice , for example , how the case 19 product has the largest @xmath17 values in fig . [ slog14151719 ] .
secondly , having less angular momentum in the system leads to a more compact product .
for example , the case 19 product has the largest enclosed mass fraction for almost any average radius @xmath102 , despite the additional shock heating undergone in this case .
furthermore , by comparing the final masses @xmath79 listed in table [ second ] for the products of cases 5 and 6 , of cases 7 through 10 , and of cases 14 through 19 , one can see that the amount of mass ejected is only very weakly dependent upon the direction of approach of the third parent star , varying by about @xmath108 or less within each of these sets of cases .
we now investigate the effects that the periastron separation of the first collision has on the final collision product .
cases 12 , 13 , and 14 all involve off - axis collisions with first collision products that are created from the same @xmath1 parent stars but with different periastron separations ( cases j , jk , and k , respectively ) , and hence different inherited angular momenta .
[ frac121314 ] plots the fractional contribution of the third parent star within the merger product . of the first collision : cases 12 ( solid curve ) , 13 ( dotted curve ) , and 14 ( dashed curve ) . in all three cases , the first two parent stars are @xmath1 , while the third parent is @xmath2 .
the @xmath2 parent penetrates into the product the least in case 12 , because of the relatively small amount of shock heating suffered by the first product during the first collision in this case .
[ frac121314],width=317 ] in all cases , the low-@xmath17 core of the @xmath0 star is able to sink to the core of the final product . however , as the periastron separation of the first collision is increased , the two @xmath1 parent stars experience more shock heating , and the @xmath2 parent is able to have more fluid penetrate down to the depths near @xmath109 .
[ chem121314 ] presents chemical composition profiles of the first collision : cases 12 ( solid curve ) , 13 ( dotted curve ) , and 14 ( dashed curve ) .
[ chem121314],width=317 ] of the final collision product for these three cases .
these profiles demonstrate that the angular momentum of the first collision product has only a small effect on the final collision product . as expected from fig .
[ frac121314 ] , the profiles of the three cases are essentially identical in the innermost 5 per cent of the bound mass , because only the core of the @xmath110 parent contributes there .
the abundance profiles of each chemical species are at least qualitatively , and usually quantitatively , similar throughout the products .
the variations that do exist can be understood in terms of the different shock heating during the first collisions . because the amount of shock heating increases with periastron separation , the case j , jk , and k products have increasingly larger values of @xmath17 at almost any enclosed mass fraction ( this trend is not immediately evident in fig . [ slog121314 ] only because @xmath28 is being plotted versus radius and not enclosed mass ) .
the fluid from the @xmath0 star is therefore able to penetrate the case j product the least , the case jk product a little more , and the case k product even more still ( see fig . [ frac121314 ] ) .
consequently the rise in c@xmath30 and c@xmath29 abundance is pushed out to increasingly larger enclosed mass fractions @xmath91 in fig .
[ chem121314 ] as one considers cases 12 , 13 , and 14 , in that order . in case 12 and arguably case 13 , the cases with the lesser amounts of shock heating , traces of two separate peaks are evident in the c@xmath29 profile . as in our zeroth order model
( see fig .
[ zeroth_order ] ) , the inner peak is due mostly to c@xmath29 from the @xmath12 stars while the outer peak is mostly due to c@xmath29 from the @xmath0 star . of the first collision : cases 12 ( solid curve ) , 13 ( dotted curve ) , and 14 ( dashed curve ) .
the particular quantities plotted are as in fig .
[ slog12 ] .
[ slog121314],width=317 ] fig .
[ slog121314 ] shows that the structure of the bulk of the fluid in the final collision product is not significantly affected by the periastron separation of the first collision , and hence the spin of the first collision product .
there is , nevertheless , a visible trend for the enclosed mass fraction at a given average radius to decrease for products with more spin .
for example , the isodensity surface with an average radius of @xmath111 encloses about 94 per cent of the case 12 product , about 92 per cent for the case 13 product , and only about 90 per cent of the case 14 product .
such a trend is expected , simply because of expansion due to rotational support .
cases 10 , 17 , and 20 can be used to investigate the effects that the periastron separation of the second collision has on the profiles of the final product .
cases 10 , 17 , and 20 involve collisions between a case k product and a @xmath0 star , with periastron separations for the second collision of @xmath112 , 0.505 , and @xmath113 , respectively , corresponding to a number of passages or interactions @xmath114 , 2 , and 3 , again respectively .
see the discussion of fig . 6 in @xcite for the details of how @xmath115 is determined .
[ frac101720 ] reveals the way in which the @xmath2 parent of the second collision : cases 10 ( solid curve ) , 17 ( dotted curve ) , and 20 ( dashed curve ) . in all three cases ,
the first two parent stars are @xmath1 , while the third parent is @xmath2 .
the first collision product results from case k. [ frac101720],width=317 ] contributes to the final product in each of these three cases . as usual , the low-@xmath17 core of the @xmath0 star sinks to the core of the collision product . as the periastron separation of the second collision is increased , the resulting collision products tend toward larger mass - averaged values of @xmath17 .
fluid from the the @xmath0 star therefore can penetrate the case 10 product the most , the case 17 product a little less , and the case 20 product even less still .
[ chem101720 ] shows that the composition profiles in these three cases are essentially identical in the innermost few per cent of the collision products .
indeed , the abundance profiles are again at least qualitatively , and usually quantitatively , similar throughout the entire product . as before , slight variations do result from having different distributions of shock heating .
in particular , the rise in c@xmath30 and c@xmath29 abundance is drawn in to smaller enclosed mass fractions @xmath91 in fig .
[ chem101720 ] as one examines cases 10 , 17 , and 20 , in that order .
[ slog101720 ] reveals the differences in the final product structure for these three cases .
the top pane shows that the mass distribution of the final product is affected by the periastron separation of the second collision in a way that is simple to understand : increasing the second periastron separation increases both shock heating and rotation , and so a given radius encloses less mass . of the second collision :
cases 10 ( solid curve ) , 17 ( dotted curve ) , and 20 ( dashed curve ) .
[ chem101720],width=317 ] of the second collision : cases 10 ( solid curve ) , 17 ( dotted curve ) , and 20 ( dashed curve ) . [ slog101720],width=317 ] in [ direction ] we found that the direction of approach of the third star only weakly affects the profiles and mass of the final product .
we therefore do not account for the angles @xmath56 and @xmath52 of the second collision when applying our fluid sorting package mmas . as a result , the product model that mmas generates is identical within each of the following sets : cases 5 and 6 , cases 7 through 10 , and cases 14 through 19 .
in all twenty cases presented in table [ second ] , the final product mass given by mmas agree with those from sph to within 1.5 per cent .
[ chem12mmas ] compares the chemical composition profiles of final collision products , as determined by both mmas and sph models , for two scenarios ( cases 1 and 2 ) in which each collision is head - on ( @xmath92 ) .
these cases involve the same three parent stars ; however the order in which the stars collide is varied .
the mmas abundance profiles maintain the same qualitative shape as those of the sph data for almost all of the elements .
one possibly important difference is that the mmas package slightly over - mixes the core in case 1 , and , consequently , the central helium abundance is not quite as high as in the sph calculation .
another noteworthy difference is that the li@xmath33 profile , especially in case 1 , is not well represented near the surface . because li@xmath33 exists in an even thinner shell at the surface of the @xmath0 star than does li@xmath34 and be@xmath35 , its abundance profile in the product is particularly sensitive to the mass loss distribution during the collisions .
note that mmas does correctly predict that most li@xmath33 is ejected during the collisions .
furthermore , the abundance profiles generated by mmas much more closely resemble the sph results than our zeroth order model does ( see fig .
[ zeroth_order ] ) , indicating that mmas is capturing the important effects of mass loss and shock heating . in scenarios such as cases 1 and 2 for which the final product is non - rotating ,
it is straightforward to obtain the enclosed mass @xmath59 and density @xmath21 profiles from the @xmath17 profile by integrating the equation of hydrostatic equilibrium ( see [ sorting ] ) .
[ slog12mmas ] shows the resulting structure of the final collision products .
the kink in the @xmath17 profile a little inside @xmath116 marks the boundary within which fluid from only the @xmath0 star contributes , and mmas reproduces this feature quite well .
the central density of the sph model is slightly less than that of the mmas model , mostly due to how density is calculated as a smoothed average in sph . despite this difference ,
the overall structure of the collision product is extremely well reproduced by mmas .
[ chem4mmas ] compares the chemical composition profiles for sph and mmas data for case 4 , a situation in which both collisions are off - axis .
the most noticeable discrepancies are that mmas again slightly over - mixes the core and underestimates the surface li@xmath33 abundance .
nevertheless , the chemical abundance profiles produced by the mmas package and the sph code are extremely similar . for example
, mmas correctly reproduces the c@xmath29 abundance , with three peaks each corresponding to a different parent star .
the inner peak is due to the low-@xmath17 fluid from the @xmath12 star involved in only the second collision , the middle peak represents fluid from the other @xmath12 star , and the outer peak represents high-@xmath17 fluid from the @xmath0 parent ( see fig .
[ compare4c13 ] ) . in the final collision product for both the mmas ( solid curve ) and the sph ( dotted curve ) data of the case 4 collision product .
[ chem4mmas ] , width=317 ] versus the enclosed mass fraction @xmath91 in the final case 4 collision product , as determined both by mmas ( solid curve ) and by sph ( dotted curve ) .
the top pane shows the total c@xmath29 abundance , while the bottom three panes show the contributions from each individual parent star : c@xmath117 is the contribution from the @xmath2 parent , c@xmath118 is the contribution from the @xmath1 parent in the first collision , and c@xmath119 is the contribution from the @xmath1 parent in the second collision .
[ compare4c13 ] , width=317 ] note that this feature is reproducible by mmas only because it accounts for shock heating in each collision ( compare to our zeroth order model of fig .
[ zeroth_order ] , in which there are only two peaks in the c@xmath29 profile ) . in many of the mmas models , small kinks , or discontinuities ,
are evident in some of the abundance profiles : such features mark locations outside of which an additional parent star either starts or stops contributing .
for example , in the case 5 and 6 collision product , a kink exists in the c@xmath30 and c@xmath29 profiles near @xmath120 ( see fig .
[ chem56mmas ] ) . as is evident from fig .
[ compare56 ] , fluid inside of the in the final collision product for both the mmas ( solid curve ) and the sph data of the case 5 ( dotted curve ) and case 6 ( dashed curve ) collision products .
note that our mmas results do not distinguish between cases 5 and 6 , as they differ only in the orientation of the first product s spin .
[ chem56mmas ] , width=317 ] of each parent star @xmath121 versus the enclosed mass fraction @xmath91 in the case 5 and 6 collision products , as determined both by mmas ( solid curve ) and by sph .
the dotted curve corresponds to the case 5 sph results , while the dashed curve gives the case 6 sph results .
the same mmas model is valid for both cases , as they differ only in the direction of the first product s rotation axis .
the @xmath49 and 2 parents are @xmath122 , while the @xmath73 parent is @xmath2 .
[ compare56 ] , width=317 ] @xmath123 shell originated solely in the @xmath0 parent star . in the range @xmath124 , all three parent stars contribute .
the smoothing that is inherent to the sph scheme makes it difficult to resolve such features with our hydrodynamics code .
it is possible that similarly abrupt changes in abundance could occur in nature within real collision products . by comparing the sph data within fig .
[ chem56mmas ] , as well as within fig .
[ compare56 ] , we also see an example of the trend discussed in [ direction ] .
namely , the direction of the first collision product s rotation axis ( or equivalently the direction of approach of the third parent ) has little effect on the final collision product .
indeed , when using mmas , our approach is to neglect completely the rotation of the first collision product , which is why the same mmas model applies to both cases 5 and 6 .
[ slog1819mmas ] plots the entropic variable @xmath17 versus the enclosed mass fraction for the final collision products of cases 18 , 19 , and 20 , as determined both by sph and mmas . cases 18 and 19 differ only in the direction of approach of the third star , and we again see that this variation has little effect on the sph results .
the kink in all of the profiles slightly inside @xmath125 marks the boundary within which fluid from only the @xmath0 star contributes .
mmas again reproduces this feature quite well .
mmas does underestimate the shock heating to the core and hence the central value of @xmath17 , although some of this discrepancy is due to the spurious heating evident in longer sph simulations @xcite .
nevertheless , it is likely that this difference between the mmas and sph models would last only as a transient during the thermal relaxation in a stellar evolution calculation .
it is also worth noting that , while the sph calculations need to be terminated before all of the bound fluid can settle into equilibrium ( see [ sph ] ) , the mmas @xmath17 profile does steadily increase outward throughout the entire product .
unfortunately it is a difficult task to determine the overall size of a collision product , either with sph simulations or with a package like mmas .
whenever there is any mass loss in an sph simulation , there will also be sph particles that are nearly unbound and , in practice , still moving away from the product when the simulation is terminated .
these particles would ultimately form the outermost layers of the collision product , but it would take an utterly unfeasible amount of time to wait for them to come back and settle into equilibrium .
the entropic variable @xmath17 profile produced by mmas seems quite reasonable , both because it increases all the way out to the surface and because the sph results tend to approach its form as more of the fluid settles into equilibrium .
however , there are no simulation data to compare against for the very outermost layers of a product and so the exact form of the profile there is difficult to validate .
not surprisingly , the radius of the collision product is rather sensitive to the @xmath17 profile .
for example , simply by changing the parameter @xmath71 from -1.0 to the still very reasonable value of -1.1 , which tends to distribute slightly more shock heating to the outer layers ( see * ? ? ?
* ) , the radii of our mmas final collision product models for case 1 and for case 2 increase by about a factor of 2 .
despite such uncertainties , it is still interesting to get a crude estimate of the sizes of the collision products immediately from mmas . in making these estimates ,
we do not account for the expansion due to rotation , but instead simply integrate the equation of hydrostatic equilibrium for a non - rotating star with the same @xmath17 profile , using the outer boundary condition that the pressure vanishes .
the radii calculated therefore represent the sizes that the products would have if some mechanism were to brake their rotation without disturbing their @xmath17 profiles .
[ rvr ] plots the radii at various enclosed mass fractions for products generated in single - single star collisions involving 0.4 , 0.6 and @xmath0 parent stars , as determined by mmas .
these radii are plotted against the normalized periastron separation @xmath126 , which we allow to exceed unity slightly to account for bulges in the parent stars .
the general trend is that as the periastron separation increases , the collisions are more long - lived , there is more shock heating , and the radii of the collision products increase . because the fluid in the deep interior of the product is largely shielded from shocks , the @xmath17 profile there , and hence the radius @xmath102 profile , are not too strongly dependent on the periastron separation of the collision . as a result , the radii versus periastron separation curves of fig .
[ rvr ] become closer to horizontal as one looks to smaller enclosed mass fraction . for the cases examined in fig .
[ rvr ] , the full ( 100 per cent enclosed mass ) radius of the collision product is always at least about twice the sum of the radii of the parent stars , and often even much larger than this .
for example , if two @xmath0 stars suffer a grazing ( @xmath127 ) collision , the collision product then has a full radius of about @xmath128 , about 20 times larger than the sum @xmath129 of the parent star radii .
we therefore expect that the collisional cross - section of these first products will be significantly enhanced over that of their thermally relaxed counterparts .
[ bw ] is similar to fig .
[ rvr ] , but for triple - star collisions .
we use different line types to represent various normalized periastron separations @xmath126 for the first collision , and along the horizontal axis we vary the normalized periastron of the second collision . the curves
give the radii at three different enclosed mass fractions .
for each of the six @xmath43 values in a frame of fig .
[ bw ] , we performed a nested loop over 45 equally spaced values of the normalized periastron separation for the second collision , from 0 to 1.1 . therefore , mmas treated 270 different triple - star collisions ( in a few minutes on a pentium iv workstation ) for each of the four plots .
note that there is a general trend for the radius of the collision product to increase as the first periastron separation increases , as expected ; this effect is mild for the 50 per cent enclosed mass radius , and rather dramatic for the full radius . even more significant is the second periastron separation , with grazing second collisions resulting in products that are substantially larger than those from head - on collisions : the shock heating suffered by the already diffuse outer layers of the first collision product is severe when multiple pericentre passages occur before merger .
once @xmath74 grows large enough for the third star s initial impact to be outside of the first product s core , so that more than one pericentre passage would result before merger , then the shock heating is no longer as sensitive to @xmath74 and the full radius surfaces in fig .
[ bw ] tend to plateau .
how strongly the full radius varies with the @xmath74 therefore depends on the mass distribution within the first product . for first products with a more uniform density , such as in the product of two @xmath130 stars ,
the final product size increases more gradually and consistently with @xmath74 . as the mass of any one of the three parent stars is increased , the trend is for the radius of the collision product to increase as well .
for example , fig .
[ bw](a ) shows that for collisions in which two @xmath12 stars collide and then a @xmath0 collides with the first product , the final collision product radius does not exceed a few times @xmath131 . if one of the @xmath12 stars is substituted with a @xmath0 star , then the final radius can be as large as about @xmath132 [ see fig .
[ bw](b ) ] .
this extreme size is due to the phenomenally diffuse outer layers of the product : the average density of such a star is only @xmath133 g @xmath134 .
the noise visible on some of the full radius curves is due to approaching the limiting numerical precision during the structure integration in these diffuse regions . from fig .
[ bw ] we see that the radius that encloses 95 per cent of the total mass , while still large , is often orders of magnitude smaller than the full radius of the final product .
because of the low densities involved , the full radius calculated is rather sensitive to the details of the shock heating during the collision . changing the mmas parameter @xmath71 from -1.0 to -1.1 ,
for example , can increase the full radius by a factor of a few , although the radius enclosing 50 per cent of the total mass does not change by more than a few per cent .
nevertheless , any reasonable form and amount of shock heating yields products that are significantly larger than a thermally relaxed star with the same mass and composition .
colliding the same three parent stars in a different order does not drastically affect the mass of the final product , although it does significantly affect its size .
consider , for example , frames ( c ) and ( d ) of fig .
if two @xmath130 stars collide and then the resulting product collides with a @xmath0 star , the final product typically has a radius of order @xmath135 , but if the @xmath0 star is switched into the first collision instead , the final radius is usually in the range from @xmath136 .
the primary reason for this difference is that a collision between the 0.4 and @xmath0 stars yields a product with especially diffuse outer layers , and , as a result , is subject to a larger number of passages and hence more shock heating during a second collision .
we have used sph and the software package mmas to study triple - star collisions .
although such collisions span a tremendous amount of parameter space , our modest number of sph calculations do provide some valuable insights .
for the ( parabolic ) encounters that we consider , we find that the order in which stars collide ( see [ order ] ) , the angle of approach of the third star ( [ direction ] ) , and the periastron separation of the collisions ( [ spin ] ) have only a slight effect on the chemical composition distribution within the final collision product .
the order and orbital parameters of the collisions can , however , significantly affect the size and structure of the product .
the results of [ comparison ] help establish that the simple fluid sorting algorithm of mmas reproduce the important features of our sph models , even when one of the parent stars is itself a collision product . the mmas package can therefore be considered an adequate , if not an accurate , substitute for a hydrodynamics code in many situations . this realization will help simplify the process of generating collision product models in cluster simulations , because a full hydrodynamics calculation will not necessarily need to be run for each collision . indeed ,
we hope the mmas package will be used to help account for stellar collisions in dynamics simulations of globular clusters . toward this end
, mmas is already being incorporated into two software packages , triptych and tripletych , that respectively treat encounters between two stars and among three stars ( see * ? ? ?
these packages are controlled through a web interface and treat the orbital trajectories , possible merger(s ) , and evolution of the merger product and therefore incorporate three main branches of stellar astrophysics : dynamics , hydrodynamics , and evolution .
the product size estimates of [ sizes ] are admittedly crude .
for example , partial ionization and radiation pressure are neglected .
although the exact size of a collision product is difficult to determine , our calculations indicate that the first and final collision products are always significantly larger than their thermally relaxed counterparts would be .
indeed , according to our mmas calculations in [ sizes ] , the final collision product can have a radius up to @xmath137 , easily exceeding the size of a typical red giant .
furthermore , these calculations have assumed that some mechanism has braked the often rapid rotation of the products , so any rotation that does remain will only further enhance the size of the products .
the extended sizes of the products will increase the multi - star collision rate over that calculated in previous treatments of binary - single and binary - binary encounters .
all of the scenarios we consider with sph in this paper involve one @xmath84 and two @xmath12 stars . without shock heating ,
the low-@xmath17 fluid of the @xmath0 star would sink to the core of the final collision product , while its high-@xmath17 portions of the @xmath0 would settle in the outer layers .
the intermediate layers of the product would consist of fluid with the same @xmath17 range from all the parents . simply sorting the fluid in this way , without running a hydrodynamics calculation ,
can therefore provide a zeroth - order model of the collision product that captures some of its important qualitative features ( see fig .
[ zeroth_order ] ) .
however , non - uniform shock heating during the collisions somewhat alters the relative values of the entropic variable @xmath17 in the fluid , resulting in a slightly different sorting pattern .
because the amount and distribution of shock heating are dependent on the details of a collision , the sorting of the fluid varies with , for example , the order in which the stars collide .
shock heating can have larger consequences on the chemical abundance profiles of elements , such as c@xmath29 , that exist in substantial amounts only in a small shell in the initial parent stars .
however , the chemical abundance profiles of most elements , particularly helium , are always qualitatively the same , regardless of how the three stars are merged . because the abundance and distribution of helium ( and hence hydrogen ) is one of the most important factors in determining the collision product s subsequent course of stellar evolution , we believe that the order and geometry of the collisions will not significantly affect the stellar evolution of the product .
indeed , @xcite have recently presented a set of stellar evolution calculations for a collision product for which the starting yrec models were generated from sph calculations of different resolutions .
the variations in the initial helium profiles of their models are roughly comparable to those in our helium profiles resulting from colliding three parent stars in different ways .
although @xcite do find detailed differences in the evolution , especially during the `` pre - main - sequence '' contraction , the evolutionary tracks and time - scales are quite similar .
we therefore feel that , for low - velocity collisions , the hydrodynamical details of how three stars are merged will not significantly affect the stellar evolution of the collision product the major caveat here being that the geometry of the collisions can of course affect the rotation of the product , which in turn can greatly affect its evolution @xcite .
surface abundances of lithium and beryllium are particularly interesting to monitor , as these elements can be used as observational indicators of mixing and perhaps collisional history . as in the single - single star collisions presented by @xcite , we find that the triple - star mergers presented here yield collision products that are severely depleted of lithium and somewhat of beryllium at the surface
. even in the relatively gentle ( parabolic ) cases that we have considered , the collisions are energetic enough to expel most of the lithium and beryllium from the outer layers of the parents .
there are many scenarios to explore when dealing with collisions in environments as chaotic as dense stellar systems .
different orbital geometries besides the parabolic trajectories treated in this paper still need to be considered in more detail .
large stellar velocities in galactic nuclei lead to hyperbolic collisions . in globular clusters , perturbations to a binary can lead to an elliptical collision , while an encounter with a very hard binary can lead to significantly hyperbolic collisions .
future studies may want to include a more detailed look at the hydrodynamics during grazing encounters , which could be done efficiently with the help of grape ( short for gravity pipe ) special purpose hardware for calculating the self - gravity of the system .
furthermore , encounters involving more than three stars , such as in binary - binary interactions , may also warrant further examination : for example , the final collision product generated in a triple - star merger is typically so extended that it could immediately start suffering roche lobe overflow if left in orbit around a fourth star . collisions among a larger variety of stellar types and masses , reflective of the diverse populations of clusters , will also need to be explored . we have been concentrating on low - mass main - sequence stars , but collisions between high - mass main - sequence stars in young compact star clusters , or giants located in the dense cores of globular clusters , for example , are frequent
. a logical first step would be to examine high - mass main - sequence stars in a runaway merger scenario .
it would therefore be very useful to develop a generalization of the fluid sorting method that includes radiation pressure in the equation of state . due to shock heating during the collision , the product is much larger than a thermally equilibrated main - sequence star of the same mass .
how much of an effect this increased radius has on the effective cross section for merger is subject to many variables , including the structure of the product s outer layers and the velocity of approaching stars . in environments such as active galactic nuclei , where relative velocities tend to be high ,
the low - density outer layers of a newly formed collision product could likely get stripped by passing stars .
however , in globular clusters , where stellar velocities tend to be small , collisions with even low - density envelopes may lead to significantly increased rates of merger .
it would be useful to develop a robust collision module that could quickly predict whether any given collision trajectory will lead to a merger , and , if not , describe how the stars are affected by the interaction .
one simple approximation often implemented in cluster simulations is that a collision product instantaneously achieves its thermally relaxed radius , a good approximation when the time between collisions is much longer than the thermal time - scale . arguing instead that the global thermal time - scale of the first product can be much larger than the time between collisions in interactions involving binaries
, we make a different approximation in this paper , namely that the first product s radius ( and more generally its structure ) does not substantially evolve between collisions .
future scattering experiments could model thermally relaxing stars and study more carefully the timescale between collisions mediated by binaries .
the thermal time - scale in the outer layers of a collision product can be orders of magnitude less than its global thermal time - scale ( see table 1 of * ? ? ?
* ) , so that it may actually be necessary to follow the thermal contraction and stellar orbits simultaneously .
indeed , in the extremely low density layers of a collision product , it is even possible for the thermal time - scale to be comparable to the ( hydro)dynamical time - scale , so that the product could undergo significant thermal contraction even before it reaches hydrodynamical equilibrium .
it would be helpful if future stellar evolution calculations of collision products included a detailed description of the products size and structure throughout the thermal relaxation stage .
how quickly the outer layers of the thermally expanded product change with time will substantially affect its likelihood of subsequent collisions .
initial conditions for such stellar evolution calculations could be provided by the publicly available mmas package .
the primary hurdle for incorporating collisions into realistic stellar dynamics simulations is currently the stellar evolution of the collision products .
such stars are highly non - canonical , typically with very peculiar structural and composition profiles , and present a challenging set of initial conditions for stellar evolution codes . to make matters even more intricate ,
rotation , which is typically rapid after merging , will affect the structural properties and chemical compositions of the stars as they evolve ( e.g. , * ? ? ?
this rapid rotation also has the effect of ejecting mass as the product thermally contracts .
studying this emitted mass will be worthwhile , as it may likely carry away angular momentum and at least partially brake rotating collision products .
we would like to thank fred rasio for helpful comments and the use of his sph code , josh faber for having parallelized this code , randall perrine for assistance in preliminary sph calculations , alison sills for providing yrec models for the parent stars , and the referee marc freitag for valuable comments that helped improve this paper .
we are also grateful to the participants of the first two modest workshops for useful discussions , especially jarrod hurley , piet hut , steve mcmillan , onno pols , simon portegies zwart , and peter teuben .
this work was supported by nsf grants ast-0071165 , mri-0079466 , and ast-0205991 .
this work was also supported by the national computational science alliance under grant ast980014n and utilized the ncsa sgi / cray origin2000 parallel supercomputer .
rasio f. a. , freitag m. , grkan m. a. , 2003 , to appear in carnegie observatories astrophysics series , vol . 1 : coevolution of black holes and galaxies , " ed
. l. c. ho , cambridge : cambridge univ .
press , astro - ph/0304038 | in dense stellar clusters , binary - single and binary - binary encounters can ultimately lead to collisions involving two or more stars during a resonant interaction .
a comprehensive survey of multi - star collisions would need to explore an enormous amount of parameter space , but here we focus on a number of representative cases involving low - mass ( 0.4 , 0.6 , and @xmath0 ) main - sequence stars . using both smoothed particle hydrodynamics ( sph ) calculations and a much faster fluid sorting software package ( mmas ) , we study scenarios in which a newly formed product from an initial collision collides with a third parent star . by varying the order in which the parent stars collide , as well as the orbital parameters of the collision trajectories , we investigate how factors such as shock heating affect the chemical composition and structure profiles of the collision product .
our simulations and models indicate that the distribution of most chemical elements within the final product is not significantly affected by the order in which the stars collide , the direction of approach of the third parent star , or the periastron separations of the collisions .
although the exact surface abundances of beryllium and lithium in the product do depend on the details of the dynamics , these elements are always severely depleted due to mass loss during the collisions .
we find that the sizes of the products , and hence their collisional cross sections for subsequent encounters , can be sensitive to the order and geometry of the collisions .
for the cases that we consider , the radius of the product formed in the first ( single - single star ) collision ranges anywhere from roughly 2 to 30 times the sum of the radii of its parent stars .
the size of the final product formed in our triple - star collisions is more difficult to determine , but it can easily be as large or larger than a typical red giant . although the vast majority of the volume in such a product contains diffuse gas that could be readily stripped in subsequent interactions , we nevertheless expect the collisional cross section of a newly formed product to be greatly enhanced over that of a thermally relaxed star of the same mass
. our results also help establish that the algorithms of mmas can quickly reproduce the important features of our sph models for these collisions , even when one of the parent stars is itself a former product .
[ firstpage ] stars : chemically peculiar globular clusters : general galaxies : star clusters hydrodynamics
blue stragglers stars : interiors |
minimal surfaces @xmath7 of general type with @xmath0 fall into two classes according to the behavior of their albanese map @xmath8 .
indeed , since @xmath9 , either @xmath10 , where @xmath11 is a smooth curve of genus @xmath12 , or @xmath13 is surjective and @xmath7 is of _ albanese general type_. the surfaces which belong to the former case satisfy @xmath14 and are nowadays completely classified , see @xcite and @xcite .
those belonging to the latter case present a much richer and subtler geometry , and their full description is still missing ; we refer the reader to the introduction of @xcite and the references given there for a recent account on this topic .
so far , the only known example of a surface of general type with @xmath0 and @xmath15 was the one given in @xcite ; in that case , the albanese map is a generically finite quadruple cover of an abelian surface with a polarization of type @xmath16 .
as the title suggests , in this paper we investigate surfaces with the above invariants and whose albanese map is a generically finite _
double _ cover .
the results that we obtain are quite satisfactory , indeed we are not only able to show the existence of such _ new _ surfaces , but we also provide their complete classification , together with a detailed description of their moduli space . before stating our results , let us introduce some notation .
let @xmath17 be a @xmath18-polarized abelian surface and let us denote by @xmath19 { \longrightarrow}\widehat{a}[2]$ ] the restriction of the canonical homomorphism @xmath20 to the subgroup of @xmath12-division points .
then @xmath21 consists of four line bundles @xmath22 .
let us denote by @xmath23 the set @xmath24 .
our first result is [ theo.a ] given an abelian surface @xmath25 with a symmetric polarization @xmath26 of type @xmath18 , not of product type , for any @xmath27 there exists a curve @xmath28 whose unique non - negligible singularity is an ordinary quadruple point at the origin @xmath29 .
let @xmath30 be a square root of @xmath31 , and if @xmath32 assume moreover @xmath33 .
then the minimal desingularization @xmath7 of the double cover of @xmath25 branched over @xmath34 and defined by @xmath35 is a minimal surface of general type with @xmath0 , @xmath36 and albanese map of degree @xmath12 .
conversely , every minimal surface of general type with @xmath0 , @xmath36 and albanese map of degree @xmath12 can be constructed in this way . when @xmath37 we obtain instead a minimal surface with @xmath38 , see proposition [ prop.invariants ] and remark [ rem.pg=3 ] .
we use the following terminology : * if @xmath32 we say that @xmath7 is a _ surface of type _ @xmath39 .
furthermore , if @xmath40 we say that @xmath7 is of type @xmath41 , whereas if @xmath42 we say that @xmath7 is of type @xmath43 * if @xmath44 we say that @xmath7 is a _ surface of type _
@xmath45 . since @xmath9 ,
the results in @xcite imply that the degree of the albanese map is a topological invariant , see proposition [ prop.degree.alb ] .
therefore we may consider the moduli space @xmath46 of minimal surfaces of general type with @xmath0 , @xmath36 and albanese map of degree @xmath12 .
let @xmath2 , @xmath3 , @xmath4 be the algebraic subsets whose points parameterize isomorphism classes of surfaces of type @xmath41 , @xmath47 , @xmath45 , respectively . therefore @xmath46 can be written as the disjoint union @xmath48 our second result is [ theo.b ] the following holds : * @xmath2 , @xmath3 , @xmath4 are the connected components of @xmath49 * these are also _ irreducible _ components of the moduli space of minimal surfaces of general type@xmath50 * @xmath51 are generically smooth , of dimension @xmath52 , respectively@xmath50 * the general surface in @xmath2 and @xmath3 has ample canonical class@xmath50 all surfaces in @xmath4 have ample canonical class . this work is organized as follows .
in section [ sec.prel ] we fix notation and terminology and we prove some technical results on abelian surfaces with @xmath53-polarization which are needed in the sequel of the paper . in the section [ sec.d=2 ]
we give the proof of theorem a. moreover we provide a description of the canonical system @xmath54 in each of the three cases @xmath41 , @xmath47 , @xmath45 .
it turns out that if @xmath7 is either of type @xmath41 or of type @xmath45 then the general curve in @xmath54 is irreducible , whereas if @xmath7 is of type @xmath47 then @xmath55 , where @xmath56 is a base - point free pencil of curves of genus @xmath6 .
finally , section 3 is devoted to the proof of theorem b. such a proof involves the calculation of the monodromy action of the paramodular group @xmath57 on the set @xmath58 $ ] , with @xmath59 this is probably well - known to the experts but , at least to our knowledge , it is nowhere explicitly written , so we dedicated to it an appendix .
* acknowledgments * m. penegini was partially supported by the dfg forschergruppe 790 _ classification of algebraic surfaces and compact complex manifolds_. f. polizzi was partially supported by progetto miur di rilevante interesse nazionale _ geometria delle variet@xmath60 algebriche e loro spazi di moduli_. both authors are grateful to f. catanese for stimulating discussions related to this work and to c. ciliberto for pointing out the existence of surfaces of type @xmath45 and for useful remarks .
moreover , it is a pleasure to thank j. s. ellenberg and l. moret - bailly for suggesting the computation in the appendix ( see the ` mathoverflow ` thread http://mathoverflow.net/questions/49970/2-torsion-line-bundles-on-abelian-varieties ) .
both authors wish to thank the referee for many detailed comments and suggestions that considerably improved the presentation of these results .
* notation and conventions .
* we work over the field @xmath61 of complex numbers .
if @xmath25 is an abelian variety and @xmath62 is its dual , we denote by @xmath63 and @xmath64 the zero point of @xmath25 and @xmath65 , respectively .
moreover , @xmath66 $ ] and @xmath58 $ ] stand for the subgroups of @xmath12-division points . if @xmath67 is a line bundle on @xmath25 we denote by @xmath68 the morphism @xmath69 given by @xmath70 .
if @xmath71 is non - degenerate then @xmath72 is an isogeny , and we denote by @xmath73 its kernel .
a coherent sheaf @xmath74 on @xmath25 is called a _ it - sheaf of index i _
if @xmath75 if @xmath74 is an it - sheaf of index @xmath76 and @xmath77 it the normalized poincar bundle on @xmath78 , the coherent sheaf @xmath79 is a vector bundle of rank @xmath80 , called the _ fourier - mukai transform _ of @xmath74 . by `` _
_ surface _ _ '' we mean a projective , non - singular surface @xmath7 , and for such a surface @xmath81 denotes the canonical class , @xmath82 is the _ geometric genus _ , @xmath83 is the _ irregularity _ and @xmath84 is the _ euler - poincar characteristic_. if @xmath85 , we denote by @xmath86 the albanese map of @xmath7 . throughout the paper ,
we denote cartier divisors on a variety by capital letters and the corresponding line bundles by italic letters , so we write for instance @xmath87 .
if @xmath88 is any complete linear system of curves on a surface , its base locus is denoted by @xmath89 . if @xmath90 is any scheme , by first - order deformation " of @xmath90 we mean a deformation over @xmath91/(\epsilon^2)$ ] , whereas by small deformation " we mean a deformation over a disk @xmath92 . in section [ sec.prel ]
we use the following special case of eagon - northcott complex .
let us consider a short exact sequence of sheaves on @xmath7 of the form @xmath93 where @xmath26 , @xmath46 are line bundles , @xmath94 is a rank @xmath12 vector bundle and @xmath95 is a point .
then the symmetric powers @xmath96 and @xmath97 fit into short exact sequences @xmath98
in this section we prove some technical facts about abelian surfaces with polarization of type @xmath99 which are needed in the sequel of the paper .
the crucial results are proposition [ prop.2l-1 ] , proposition [ prop.2l-2 ] , corollary [ cor.sing.curv ] and corollary [ cor.restriction ] . for the statements
whose proof is omitted we refer the reader to @xcite , @xcite , ( * ? ? ?
* chapter 10 ) , @xcite and @xcite .
let @xmath25 be an abelian surface and @xmath100 an ample divisor on @xmath25 with @xmath101 .
then @xmath100 defines a positive definite line bundle @xmath102 on @xmath25 , whose first chern class is a polarization of type @xmath99 . by abuse of notation
we consider the line bundle @xmath26 itself as a polarization .
moreover we have @xmath103 so the linear system @xmath88 is a pencil . _
@xcite _ [ prop.barth.class ] let @xmath104 be a @xmath99-polarized abelian surface and let @xmath105 .
then we are in one of the following cases : * @xmath106 is a smooth , connected curve of genus @xmath107 * @xmath106 is an irreducible curve of geometric genus @xmath12 , with an ordinary double point@xmath50 * @xmath108 , where @xmath109 and @xmath110 are elliptic curves and @xmath111 * @xmath112 , with @xmath109 , @xmath113 , @xmath114 elliptic curves such that @xmath115 , @xmath116 , @xmath117 .
moreover , in case @xmath118 the surface @xmath25 admits an isogeny onto a product of two elliptic curves , and the polarization of @xmath25 is the pull - back of the principal product polarization , whereas in case @xmath119 the surface @xmath25 itself is a product @xmath120 and @xmath121 .
let us denote by @xmath122 the matrix @xmath123 , and by @xmath124 the moduli space of @xmath125-polarized abelian surfaces ; then there exists a zariski dense set @xmath126 such that , given any @xmath127 , all divisors in @xmath88 are irreducible , i.e. , of type @xmath128 or @xmath129 , see ( * ? ? ?
* section 3 ) .
if @xmath127 , we say that @xmath26 is a _ general
_ @xmath18-polarization .
if @xmath88 contains some divisor of type @xmath118 , we say that @xmath26 is a _ special _ @xmath130-polarization .
finally , if the divisors in @xmath88 are of type @xmath119 , we say that @xmath26 is a _ product _ @xmath130-polarization .
in the rest of this section we assume that @xmath26 is not a product polarization
. then @xmath88 has four distinct base points @xmath131 , which form an orbit for the action of @xmath132 on @xmath25 .
moreover all curves in @xmath88 are smooth at each of these base points , see ( * ? ? ?
* section 1 ) .
let us denote by @xmath133 the involution @xmath134 on @xmath25 .
then we say that a divisor @xmath11 on @xmath25 is _ symmetric _ if @xmath135 .
analogously , we say that a vector bundle @xmath74 on @xmath25 is symmetric if @xmath136 .
since @xmath67 is ample , @xcite implies that , up to translations , we may suppose that @xmath67 is symmetric and that the base locus of @xmath88 coincides with @xmath73 .
moreover : * for all sections @xmath137 we have @xmath138 .
in particular , all divisors in @xmath88 are symmetric ; * we may assume that @xmath139 and that @xmath140 are @xmath12-division points , satisfying @xmath141 .
there exist exactly three @xmath12-torsion line bundles @xmath142 , @xmath143 , @xmath144 on @xmath25 , with @xmath145 , such that the linear system @xmath146 contains an irreducible curve which is singular at @xmath63 .
more precisely , one shows that @xmath147 and that the unique curve @xmath148 which is singular at @xmath63 actually has an ordinary double point there . denoting by @xmath19 { \longrightarrow}\widehat{a}[2]$ ] the homomorphism induced by @xmath20 on the subgroups of @xmath12-division points ,
both @xmath149 and @xmath150 are isomorphic to @xmath151 .
indeed , we have @xmath152 let @xmath153 be the set @xmath154 .
[ prop.q.nodal ] let @xmath155 and @xmath156 .
then @xmath157 since @xmath158 , translating by @xmath95 we see that @xmath159 if and only if @xmath160 now the claim follows because this holds precisely when @xmath161 . for any @xmath162 ,
let us consider the linear system @xmath163 consisting of the curves in @xmath164 having a point of multiplicity at least @xmath5 at @xmath63 .
we first analyze the case @xmath32 .
[ prop.2l-1 ] we have @xmath165 , that is the linear system @xmath166 is a pencil .
moreover , if @xmath167 then we are in one of the following cases : * @xmath11 is an irreducible curve of geometric genus @xmath6 , with an ordinary quadruple point @xmath168this corresponds to the general case@xmath169 * @xmath11 is an irreducible curve of geometric genus @xmath12 , with an ordinary quadruple point and an ordinary double point@xmath50 * @xmath170 , where @xmath171 is an irreducible curve of geometric genus @xmath12 with an ordinary double point@xmath50 * @xmath67 is a special @xmath125-polarization and @xmath170 , where @xmath171 is the union of two elliptic curves intersecting transversally in two points .
since the three curves @xmath172 belong to @xmath173 and each @xmath174 is irreducible , by bertini theorem it follows that the general element @xmath175 is irreducible and smooth outside @xmath63 . on the other hand we have @xmath176 ,
so @xmath11 has an _ ordinary _ quadruple point at @xmath63 . blowing up this point , the strict transform of @xmath11 has self - intersection @xmath177 ,
so @xmath178 is a pencil .
assume first that @xmath26 is a general polarization .
we have shown that the general curve in @xmath173 belongs to case @xmath128 . in order to complete the proof , observe that @xmath178 contains the following distinguished elements : * three reduced , irreducible curves @xmath179
, @xmath180 , @xmath181 such that @xmath182 has an ordinary quadruple point at @xmath63 , an ordinary double point at @xmath183 and no other singularities ( see ( * ? ? ?
* corollary 4.7.6 ) ) .
these curves are as in case @xmath129 ; * three non - reduced elements , namely @xmath184 , @xmath185 , @xmath186 .
these curves are as in case @xmath118 . moreover ,
all other elements of @xmath173 are smooth outside @xmath63 ; this can be seen by blowing - up @xmath63 and applying the zeuthen - segre formula to the fibration induced by the strict transform of the pencil , see ( * ? ? ?
* section 1.2 ) .
finally , assume that @xmath26 is a special polarization .
then there is just one more possibility , namely @xmath170 , where @xmath171 is the translate of a reducible curve @xmath187 by a suitable @xmath12-division point .
this yields case @xmath119 .
let us consider now the case where @xmath31 is non - trivial . in the sequel
, @xmath188 always denotes a permutation of @xmath189 .
[ prop.2l-2 ] let @xmath162 be non - trivial .
then @xmath190 is empty , unless @xmath191 .
more precisely , for all @xmath192 we have @xmath193 , so that @xmath194 consists of a unique element , namely the curve @xmath195 .
assume that there exists an effective curve @xmath196 .
blowing up the point @xmath63 , the strict transform @xmath197 of @xmath11 is numerically equivalent to the strict transform of a general element of the pencil @xmath178 . since @xmath198 is non - trivial , by the description of the non - reduced elements of @xmath199 given in proposition [ prop.2l-1 ] we must have @xmath200 , so @xmath201 . summing up , propositions [ prop.2l-1 ] and [ prop.2l-2 ] imply [ cor.2l-3 ] let @xmath162 .
then @xmath202 is empty , unless @xmath203 . in this case @xmath204 [ cor.sing.curv ]
let @xmath104 be a @xmath125-polarized abelian surface , and let @xmath11 be a _ reduced _ divisor numerically equivalent to @xmath205 which has a quadruple point at @xmath206 .
then @xmath11 belongs to one of the following cases , all of which occur : * @xmath11 is irreducible , with an ordinary quadruple point at @xmath95 and no other singularities@xmath50 * @xmath11 is irreducible , with an ordinary quadruple point at @xmath95 , an ordinary double point and no other singularities@xmath50 * @xmath207 , where @xmath208 is irreducible and numerically equivalent to @xmath100 , with an ordinary double point at @xmath95 and no other singularities . since @xmath209 , the singularity of @xmath11 at @xmath95 is again an ordinary quadruple point . there exists a rank @xmath12 , indecomposable vector bundle @xmath94 on @xmath25 , such that @xmath210 @xmath211 moreover @xmath94 is symmetric and it is isomorphic to the unique locally free extension of the form @xmath212 let @xmath213 be the @xmath99-polarization on @xmath65 which is dual to @xmath26 .
then @xmath214 is a it - sheaf of index @xmath12 and its fourier - mukai transform @xmath215 is a rank @xmath12 vector bundle on @xmath25 , which satisfies by ( * ? ? ?
* theorem 14.2.2 ) and ( * ? ? ?
* corollary 2.8 ) .
in addition , ( * ? ? ? * proposition 14.4.3 ) implies @xmath216 . finally , hirzebruch - riemann - roch implies @xmath217 and by ( * ? ? ?
* proposition 2.2 ) and ( * ? ? ?
* proposition 2.4 ) , since @xmath26 is not a product polarization , we infer that @xmath94 is symmetric and isomorphic to the unique locally free extension .
[ prop.coh.s2.f ] let @xmath162 .
the following holds : * if @xmath218 , then @xmath219 * if @xmath220 , then @xmath221 tensoring with @xmath198 we obtain @xmath222 , that is @xmath223 has a non - trivial section . by (
* proposition 5 p. 33 ) there exists an effective divisor @xmath11 and a zero - dimensional subscheme @xmath224 such that @xmath223 fits into a short exact sequence @xmath225 where @xmath226 . then @xmath227 and @xmath228 now there are three possibilities : * @xmath11 is an elliptic curve ; * @xmath11 is a principal polarization ; * @xmath229 .
we want to show that @xmath230 and @xmath231 can not occur . in case
@xmath230 we have @xmath232 , then by we obtain @xmath233 and @xmath234 .
thus ( * ? ? ?
* lemma 10.4.6 ) implies that @xmath26 is a product polarization , contradiction . in case @xmath231 , the index theorem yields @xmath235 , so using we deduce @xmath236 , @xmath234
. tensoring by @xmath237 and setting @xmath238 , we obtain @xmath239 since @xmath240 is also principal polarization , by applying the same argument used in ( * ? ? ?
* proof of proposition 2.2 ) we conclude again that @xmath67 must be a product polarization .
therefore the only possibility is @xmath241 , namely @xmath229 .
it follows @xmath242 , that is @xmath243 consists of a unique point @xmath206 and becomes @xmath244 moreover , since @xmath223 is locally free , we have @xmath245 , see ( * ? ? ?
* example 1.7 ) or ( * ? ? ?
* theorem 12 p. 39 ) . applying the eagon - northcott complex to and tensoring with @xmath246
we get @xmath247 hence @xmath248 on the other hand , since @xmath249 , proposition [ prop.q.nodal ] yields @xmath250 using serre duality , the isomorphism @xmath251 and , since @xmath252 if and only if @xmath253 we obtain @xmath254 for all @xmath162 .
moreover hirzebruch - riemann - roch gives @xmath255 , hence we get @xmath256 this completes the proof .
[ prop.coh.s3.f ] for any @xmath162 , we have @xmath257 by hirzebruch - riemann - roch we obtain @xmath258 , so it suffices to show that @xmath259 for @xmath260 .
the sheaf @xmath261 satisfies it of index @xmath177 and @xmath262 , so its fourier - mukai transform @xmath263 is a line bundle on @xmath264 , which satisfies it of index @xmath12 by ( * ? ? ?
* theorem 14.2.2 ) and has @xmath265 by ( * ? ? ?
* corollary 2.8 ) .
this means that @xmath266 is a @xmath99 polarization . since @xmath74 is a symmetric vector bundle , by using ( * ? ? ?
* corollary 2.4 ) we obtain @xmath267 that is the rank @xmath12 vector bundle @xmath223 is the fourier - mukai transform of @xmath268 .
therefore , taking the isogeny @xmath269 and using ( * ? ? ?
* proposition 3.11 ) , we can write @xmath270 on the other hand , @xmath271 is a finite map so we have @xmath272 since @xmath273 by using we deduce @xmath274 the right - hand side vanishes for @xmath275 , so we are done .
let @xmath276 be the blow - up of @xmath25 at @xmath63 and let @xmath277 be the exceptional divisor .
since @xmath278 , by abusing notation we will often identify degree @xmath177 line bundles on @xmath279 with degree @xmath177 line bundles on @xmath25 , and we will simply write @xmath198 instead of @xmath280 . the strict transform of the pencil @xmath281 gives the base - point free pencil @xmath282 in @xmath279 , whose general element is a smooth curve of genus @xmath6 .
[ prop.restriction ] let @xmath283 be a smooth curve and let @xmath162 .
then @xmath284 if and only if @xmath285 .
if @xmath32 the result is clear , so we assume that @xmath286 is non - trivial . since @xmath287 ,
by using the short exact sequence @xmath288 and serre duality , we obtain @xmath289 in order to compute the last cohomology group , we will exploit the vector bundle @xmath74 .
in fact , applying the eagon - northcott complex to and tensoring with @xmath290 , we get @xmath291 by using , proposition [ prop.coh.s2.f ] and proposition [ prop.coh.s3.f ] , we obtain @xmath292 since @xmath34 is a smooth curve of genus @xmath6 , by using and riemann - roch we deduce @xmath293 this completes the proof .
[ cor.restriction ] let @xmath294 and let @xmath34 be a smooth curve in the pencil @xmath295 .
then @xmath296 if and only if @xmath297 . for all @xmath298 the effective curve @xmath299 does not intersect @xmath34 , so @xmath300 .
hence proposition [ prop.restriction ] yields @xmath301 and the claim follows .
in the sequel , @xmath7 will be a smooth minimal surface of albanese general type with @xmath0 and @xmath302 will be its albanese map , which we suppose of degree @xmath12 .
let @xmath303 be the branch locus of @xmath13 and let @xmath304 \ar[dr]_{\alpha } & x \ar[d]^{f } \\ & a}\ ] ] be the stein factorization of @xmath13 .
then @xmath305 is a finite double cover and , since @xmath7 is smooth , it follows that @xmath90 is normal , see ( * ? ? ?
* chapter i , theorem 8.2 ) .
in particular @xmath90 has at most isolated singularities , hence the curve @xmath306 is reduced .
[ prop.branch ] assume that @xmath36 and that the albanese map @xmath307 is a generically finite double cover .
then there exists a polarization @xmath308 of type @xmath99 on @xmath25 such that @xmath306 is a curve in @xmath309 whose unique non - negligible singularity is an ordinary quadruple point @xmath95 .
@xmath306 is linearly equivalent to @xmath310 for some divisor @xmath311 in @xmath25 .
there is a
canonical resolution " diagram @xmath312 where @xmath313 is smooth and @xmath276 is composed of a series of blow - ups , see ( * ? ? ?
* chapter iii , section 7 ) .
let @xmath314 be the centers of these blow - ups , and let @xmath315 be the inverse image of @xmath316 on @xmath279 ( with right multiplicities such that @xmath317 .
then the branch locus @xmath318 of @xmath319 is smooth and can be written as @xmath320 where the @xmath321 are even positive integers , say @xmath322 .
let us recall a couple of definitions : * a _ negligible singularity _ of @xmath306 is a point @xmath323 such that @xmath324 , and @xmath325 for any point @xmath316 infinitely near to @xmath326 * a @xmath327$]- _ singularity _ of @xmath306 is a pair @xmath328 such that @xmath316 belongs to the first infinitesimal neighbourhood of @xmath323 and @xmath329 .
for example , a double point and an ordinary triple point are negligible singularities , whereas a @xmath330$]-point is not . by using the formulae in @xcite we obtain @xmath331 which imply @xmath332 if @xmath333 for all @xmath76 , then all the @xmath316 are negligible singularities and gives @xmath334 , a contradiction
. then we can assume @xmath335 , @xmath336 .
therefore yields @xmath337 , that is @xmath338 is a polarization of type @xmath99 on @xmath25 .
now we have two possibilities : * @xmath339 is not infinitely near to @xmath340 ; then @xmath341 contains an ordinary quadruple point @xmath95 and ( possibly ) some negligible singularities ; * @xmath339 is infinitely near to @xmath340 ; then latexmath:[$d_a
\in and ( possibly ) some negligible singularities . but in case @xmath231 the surface @xmath313 contains a @xmath343-curve , hence @xmath344 , a contradiction . therefore @xmath306 must be a curve of type @xmath230 .
the existence of such a curve was proven in corollary [ cor.sing.curv ] , so we are done .
[ rem.[3,3 ] ] the argument used in the proof of proposition _ [ prop.branch ] _ shows that , if we were able to find a curve in @xmath309 with a singular point of type @xmath345 $ ] , then we could construct a surface @xmath7 with @xmath0 and @xmath344 .
unfortunately , at present we do not know whether such a curve exists .
[ prop.no product ] @xmath346 is not a product polarization .
assume by contradiction that @xmath346 is a product polarization .
then @xmath347 , with natural projection maps @xmath348 and @xmath349 , and @xmath350 .
let @xmath351 be the fibre of @xmath352 passing through @xmath95 .
since @xmath306 has a quadruple point at @xmath95 and @xmath353 , b@xmath354zout theorem implies that @xmath351 is a component of @xmath306 .
similarly , since @xmath355 has a triple point at @xmath95 and @xmath356 , it follows that @xmath351 is a component of @xmath355 .
therefore @xmath306 contains the curve @xmath351 with multiplicity at least @xmath12 , which is impossible since @xmath306 must be reduced . up to a translation
we can now suppose @xmath357 and by using corollary [ cor.2l-3 ] we can write @xmath358 , where @xmath359 and @xmath26 is a symmetric polarization , not of product type , such that @xmath360 . in the rest of this section
we assume for simplicity that @xmath306 contains no negligible singularities besides the quadruple point @xmath63 ; this is an open condition , equivalent to the ampleness of @xmath361 .
hence the map @xmath276 is just the blow - up at @xmath63 , we have @xmath362 and induces the following commutative diagram @xmath363^{\beta } \ar[dr]_{\alpha } & b \ar[r]^{\varphi } \ar[d]^{\sigma } & \mathbb{p}^1 \\ & a},\ ] ] where @xmath364 is a _
double cover and @xmath365 is the morphism induced by the base - point free pencil @xmath282 .
the double cover @xmath366 is branched along a _ smooth _
divisor @xmath367 hence it is defined by a square root of @xmath368 , namely by @xmath369 , where @xmath370 is a square root of @xmath31 . [ prop.invariants ] @xmath7 is a minimal surface of general type with @xmath0 and @xmath36 , unless @xmath371 . in the last case we have instead @xmath38 and @xmath36 .
standard formulae for double covers ( @xcite ) give @xmath372 and @xmath36 .
moreover we have @xmath373 , hence we obtain @xmath374 if @xmath370 is not trivial then @xmath375 , otherwise @xmath376 .
[ rem.pg=3 ] if @xmath37 then @xmath7 is the symmetric product of a smooth curve of genus @xmath6 , see _
@xcite _ and _ @xcite_. let us give an alternative construction of the double cover @xmath377 in this particular case . take a smooth curve @xmath11 of genus @xmath6 , admitting a double cover @xmath378 onto an elliptic curve @xmath109 .
let @xmath63 be the identity in the group law of @xmath109 and for all @xmath379 let us denote by @xmath380 the conjugate point of @xmath381 with respect to the involution @xmath382 induced by @xmath383 .
then @xmath384 contains the elliptic curve @xmath385 which is isomorphic to @xmath109 .
moreover , there is a morphism @xmath386 now take any point @xmath387 and let @xmath388 .
we have @xmath389 that is the induced map @xmath390 contracts @xmath391 to a point .
moreover @xmath13 has generic degree @xmath12 ; in fact @xmath392 for all @xmath381 , @xmath393 .
since we are interested in the case @xmath394 , in the sequel we always assume @xmath395 . summing up , we have proven the following result .
[ th.a ] given an abelian surface @xmath25 with a symmetric polarization @xmath26 of type @xmath18 , not of product type , for any @xmath396 there exists a curve @xmath397 whose unique non - negligible singularity is an ordinary quadruple point at the origin @xmath29 .
let @xmath30 be a square root of @xmath31 , and if @xmath32 assume moreover @xmath33 .
then the minimal desingularization @xmath7 of the double cover of @xmath25 branched over @xmath306 and defined by @xmath35 is a minimal surface of general type with @xmath0 , @xmath36 and albanese map of degree @xmath12 .
conversely , every minimal surface of general type with @xmath0 , @xmath36 and albanese map of degree @xmath12 can be constructed in this way . in order to proceed with the study of our surfaces ,
let us introduce the following [ def.type ] let @xmath7 be a minimal surface of general type with @xmath0 , @xmath36 and albanese map of degree @xmath12 .
* if @xmath32 we say that @xmath7 is a _ surface of type _ @xmath39 .
furthermore , if @xmath398 we say that @xmath7 is of type @xmath41 , whereas if @xmath42 we say that @xmath7 is of type @xmath47 .
* if @xmath44 we say that @xmath7 is a _ surface of type _ @xmath45 .
[ rem.types ] if @xmath7 is a surface of type @xmath39 , then @xmath306 is as in corollary @xmath399 , case @xmath230 or @xmath231 .
if @xmath7 is a surface of type @xmath45 , then @xmath306 is as in corollary _ [ cor.sing.curv ] _ , case @xmath241 .
see figures @xmath400 and @xmath401 . and
@xmath318 for a general surface of type @xmath39 ] and @xmath318 for a surface of type @xmath45 ] we denote by @xmath404 the ramification divisor of @xmath405 and by @xmath391 the divisor @xmath406 . then @xmath391 is an elliptic curve and @xmath407 .
+ [ prop.pencil.d ] the pullback via @xmath364 of the general curve @xmath34 in the pencil @xmath408 is reducible if and only if @xmath7 is of type @xmath47 .
the restriction of @xmath366 to @xmath34 is the trivial double cover if and only if @xmath409 , i.e. if and only if @xmath410 .
thus the result follows from corollary [ cor.restriction ] .
now we want to describe the canonical system of our surfaces .
let us analyze first surfaces of type @xmath39 .
then @xmath370 is a non - trivial , @xmath12-torsion line bundle and , for the general surface @xmath7 , the branch locus @xmath318 of @xmath364 is a smooth curve of genus @xmath6 belonging to the pencil @xmath408 .
[ prop.type i ] let @xmath7 be a surface of type @xmath411 then the following holds .
* if @xmath7 is of type @xmath41 , the pullback via @xmath364 of the pencil @xmath412 on @xmath279 is a base - point free pencil @xmath56 on @xmath7 , whose general element @xmath413 is a smooth curve of genus @xmath414 satisfying @xmath415 .
moreover , the canonical system @xmath54 has no fixed part , hence the general canonical curve of @xmath7 is irreducible . finally , @xmath416 . *
if @xmath7 is of type @xmath47 , i.e. @xmath417 for some @xmath418 , there is a commutative diagram @xmath419^{\beta } \ar[d]_{\phi } & b \ar[d]^{\varphi } \\
\mathbb{p}^1 \ar[r]^b & \mathbb{p}^1,}\ ] ] where @xmath420 is a double cover branched in two points , namely the point corresponding to the branch locus @xmath318 and the point corresponding to the curve @xmath421 , where @xmath174 is the unique curves in @xmath422 with an ordinary double point at @xmath63 and @xmath423 " stands for the strict transform in @xmath279 .
the general fibre @xmath413 of the map @xmath424 is a smooth curve of genus @xmath6 ; moreover @xmath391 is the fixed part of @xmath54 and @xmath55 , i.e. the canonical system is composed with the pencil @xmath56 .
finally , @xmath425 .
@xmath230 the fact that @xmath413 is a smooth curve of genus @xmath414 follows from proposition [ prop.pencil.d ] ; moreover @xmath426 .
we have @xmath427 and by hurwitz formula @xmath428 . since latexmath:[$\dim pencil
, we deduce that @xmath54 has no fixed part .
@xmath231 if @xmath7 is of type @xmath47 , then by proposition [ prop.pencil.d ] the pull - back via @xmath366 of a general element of @xmath412 is the disjoint union of two smooth curves of genus @xmath6 .
so there exists a base - point free genus @xmath6 pencil @xmath56 on @xmath7 and we obtain diagram . in this case @xmath430
is effective and it is no difficult to see that @xmath420 is branched only at the two points corresponding to @xmath318 and @xmath174 .
moreover @xmath431 , so we can write @xmath432 that is the fixed part of the canonical pencil of @xmath7 is @xmath433 and its movable part is @xmath56 .
let us consider now surfaces of type @xmath45 .
then @xmath434 for some @xmath418 , so @xmath30 is a degree @xmath177 line bundle whose order is exactly @xmath5 and the curve @xmath318 consists of two distinct half - fibres of @xmath412 , namely @xmath435 .
therefore @xmath7 is of type @xmath45 if and only if @xmath318 is _
disconnected_. proposition [ prop.pencil.d ] implies that the pullback via @xmath366 of a general curve in @xmath412 is irreducible , so we obtain [ prop.type ii ] if @xmath7 is a surface of type @xmath45 , then the pullback via @xmath436 of the pencil @xmath412 on @xmath279 is a base - point free pencil @xmath56 on @xmath7 , whose general element @xmath413 is a smooth curve of genus @xmath414 satisfying @xmath415 .
moreover , the canonical system @xmath54 has no fixed part , hence the general canonical curve of @xmath7 is irreducible . finally , @xmath437 with @xmath438 .
the first two parts of the statement follow from proposition [ prop.pencil.d ] by the same argument used in the proof of proposition [ prop.type i ] , part @xmath230 .
it remains only to prove the assertion about @xmath439 .
let @xmath440 , @xmath441 be the two effective curves in @xmath7 such that @xmath442 , @xmath443 ; then @xmath437 .
moreover , since @xmath444 and @xmath445 are both half - fibres of @xmath412 , it follows @xmath446 , @xmath447 and we are done .
[ rem.k ample ] the general surface of type @xmath39 has ample canonical divisor . in addition , all surfaces of type @xmath45 have ample canonical divisor .
let @xmath7 be a minimal surface of general type with @xmath0 , @xmath36 and albanese map of degree @xmath12 ; for a general choice of @xmath7 we may assume that @xmath361 is ample , see remark [ rem.k ample ] .
the following result can be found in ( * ? ? ?
* section 5 ) .
[ prop.degree.alb ] let @xmath7 be a minimal surface of general type with @xmath448 and albanese map @xmath302 , and assume that @xmath449 is a surface .
then this is a topological property .
if in addition @xmath450 , then the degree of the @xmath13 is a topological invariant .
by @xcite the albanese map @xmath13 induces a homomorphism of cohomology algebras @xmath451 and @xmath452 is isomorphic to the full exterior algebra @xmath453 . in particular , if @xmath9 the degree of the albanese map equals the index of the image of @xmath454 inside @xmath455 and it is therefore a topological invariant . by proposition [ prop.degree.alb ]
it follows that one may study the deformations of @xmath7 by relating them to those of the flat double cover @xmath364 .
by @xcite we have an exact sequence @xmath456 where @xmath457 is a coherent sheaf supported on @xmath439 called the _ normal sheaf of _ @xmath366 .
[ prop.n - beta ] assume that @xmath361 is ample .
if @xmath7 is a surface of type @xmath39 , then @xmath458 . if @xmath7 is a surface of type @xmath45 , then @xmath457 is a non - trivial , @xmath12-torsion element of @xmath459 .
since @xmath361 is ample , @xmath439 is smooth and we have an isomorphism @xmath460 see ( * ? ? ?
* lemma 3.2 ) . if @xmath7 is of type @xmath39 then either @xmath461 or @xmath462 , see proposition [ prop.type i ] , so @xmath457 is trivial . if @xmath7 is of type @xmath45 then @xmath463 , see proposition [ prop.type ii ] , so @xmath457 is a non - trivial , @xmath12-torsion line bundle .
[ prop.coh.tb ] assume that @xmath361 is ample .
then the sheaf @xmath464 satisfies @xmath465 since @xmath364 is a finite map , by using projection formula and the leray spectral sequence we deduce @xmath466 there is a short exact sequence @xmath467 see @xcite .
then a straightforward computation yields @xmath468 now let us tensor with @xmath469 .
since @xmath470 and @xmath471 , by taking cohomology we obtain @xmath472 moreover we have@xmath473 where the first equality comes from the fact that @xmath474 is an effective divisor , the third equality follows from serre duality and @xmath475 , since @xmath33 , and the second one is a consequence of riemann - roch .
therefore the claim follows using , , and .
we have a commutative diagram @xmath476 & 0 \ar[d ] & 0 \ar[d ] & \\ 0 \ar[r ] & t_s \ar[r ] \ar[d ] & \beta^*t_b \ar[r ] \ar[d ] & \mathcal{n}_{\beta } \ar[r ] \ar[d ] & 0\\ 0 \ar[r ] & t_s \ar[r]\ar[d ] & \alpha^*t_a \ar[r ] \ar[d ] & \mathcal{n}_{\alpha } \ar[r ] \ar[d ] & 0 \\ & 0 \ar[r ] & { \mathcal{o}}_{z}(-z ) \ar[r ] \ar[d ] & { \mathcal{o}}_{z}(-z ) \ar[r ] \ar[d ] & 0 \\ & & 0 & 0 & } \ ] ] whose central column is the pullback of via @xmath364 .
[ prop.h1-t ] let @xmath7 be a minimal surface with @xmath0 , @xmath36 and albanese map of degree @xmath12 , and assume that @xmath361 is ample .
then @xmath477 proposition [ prop.coh.tb ] yields @xmath478 , so looking at the central column of diagram we obtain the long exact sequence in cohomology @xmath479 since @xmath480 , it follows that the map @xmath481 is an isomorphism .
therefore the commutativity of implies that the image of @xmath482 in @xmath483 coincides with the image of @xmath484 in @xmath485 .
so we obtain the exact sequence @xmath486 we claim that the image of @xmath487 has dimension @xmath6 . in order to prove this
, we borrow an argument from ( * ? ? ?
* section 6 ) .
take a positive integer @xmath488 such that there exists a smooth pluricanonical divisor @xmath489 and let @xmath490 be the image of @xmath491 in @xmath25 . by (
* section 3.4.4 p. 177 ) , the first order deformations of a pair @xmath492 , where @xmath493 is a closed subscheme and @xmath494 is nonsingular , are parameterized by the vector space @xmath495 , where @xmath496 is the sheaf of germs of tangent vectors to @xmath494 which are tangent to @xmath90 .
notice that @xmath497 is usually denoted by @xmath498 when @xmath90 is a normal crossing divisor with smooth components .
in our situation , a first - order deformation of the pair @xmath499 induces a first - order deformation of the pair @xmath500 , because the the differential map @xmath501 sends vectors tangent to @xmath491 into vectors tangent to @xmath490 .
hence we have a commutative diagram @xmath502_{\epsilon } \ar[rr]^{\gamma ' } & & h^1(a , \ ,
t_a \langle \gamma ' \rangle ) \ar[d]^{\epsilon ' } \\
h^1(s , \ , t_s ) \ar[rr]^{\gamma } & & h^1(a , \ , t_a ) . \\
} \end{xy}\ ] ] let us observe now the following facts . *
since @xmath7 is smooth , the line bundle @xmath503 extends along any first - order deformation of @xmath7 , because the relative dualizing sheaf is locally free for any smooth morphism of schemes , see @xcite .
moreover , since @xmath7 is minimal of general type , we have @xmath504 , so every section of @xmath503 extends as well , see ( * ? ? ?
* section 3.3.4 ) .
this means that no first - order deformation of @xmath7 makes @xmath491 disappear , in other words @xmath505 is surjective
. therefore @xmath506 . * since @xmath507 , the line bundle
@xmath508 is ample on @xmath25 ; therefore it deforms along a subspace of @xmath509 of dimension @xmath6 , see @xcite .
since every first - order deformation of the pair @xmath510 induces a first - order deformation of the pair @xmath511 , it follows that the image of @xmath512 is at most @xmath6-dimensional . according to the above remarks ,
we obtain @xmath513 on the other hand , given any abelian surface @xmath25 with a @xmath130-polarization , not of product type , by the results of section [ sec.d=2 ] we can construct a surface @xmath7 of type @xmath39 or @xmath45 such that @xmath514 .
then the dimension of @xmath515 equals the dimension of the moduli space of @xmath516-polarized abelian surfaces , which is precisely @xmath6 .
so implies @xmath517 and by using proposition [ prop.n - beta ] we are done . by proposition [ prop.degree.alb ]
we may consider the moduli space @xmath46 of minimal surfaces @xmath7 of general type with @xmath0 , @xmath36 and albanese map of degree @xmath12 .
let @xmath2 , @xmath3 , @xmath4 be the subsets whose points parameterize isomorphism classes of surfaces of type @xmath41 , @xmath47 , @xmath45 , respectively . therefore @xmath46 can be written as the disjoint union @xmath48 set moreover @xmath518 .
[ prop.moduli ] the following holds : * @xmath2 and @xmath3 are irreducible , generically smooth of dimension @xmath519 * @xmath4 is irreducible , generically smooth of dimension @xmath6 .
@xmath520 the construction of a surface of type @xmath39 depends on the following data : * the choice of a @xmath125-polarized abelian surface @xmath521 , not of product type ; * the choice of a general divisor @xmath306 in the pencil @xmath522 ; * the choice of a non - trivial line bundle @xmath198 such that @xmath523 .
let @xmath524 $ ] be the space of pairs @xmath525 , where @xmath25 is a @xmath125-polarized abelian surface and @xmath162 is the isomorphism class of a non - trivial , @xmath12-torsion line bundle . in the appendix ( see proposition [ prop.components ] ) we show that @xmath524 $ ] is a quasi - projective variety , union of two connected , irreducible components of dimension @xmath6 @xmath526 \quad \textrm{and } \quad \mathcal{a}_{\delta}^{(b)}[2],\ ] ] which correspond to @xmath527 and @xmath528 , respectively .
therefore there are two generically finite , dominant maps @xmath529 where @xmath530 and @xmath531 are suitable projective bundles on @xmath532 $ ] and @xmath533 $ ] ; it follows that @xmath2 and @xmath3 are irreducible of dimension @xmath5 .
on the other hand , proposition [ prop.h1-t ] implies that for a general @xmath534 \in \mathcal{m}_i$ ] we have @xmath535 } \mathcal{m}_{i}=h^1(s , \ , t_s)=4.\ ] ] this shows that both @xmath2 and @xmath3 are generically smooth .
@xmath536 the construction of a surface of type @xmath45 depends on the following data : * the choice of a @xmath125-polarized abelian surface @xmath521 , not of product type ; * the choice of @xmath537 , which yields the unique curve @xmath538 ; * the choice of a square root @xmath370 of @xmath198 .
let @xmath539 $ ] be space of triplets @xmath540 , where @xmath25 is the isomorphism class of a @xmath516-polarized abelian surface , @xmath541 and @xmath30 is a square root of @xmath31 . in the appendix ( see proposition [ prop.monodromy.2 ] ) we show that @xmath539 $ ] is a @xmath6-dimensional , irreducible quasi - projective variety .
we have a generically finite , dominant map @xmath542 \longrightarrow \mathcal{m}_{ii},\ ] ] so @xmath4 is irreducible of dimension @xmath6 .
on the other hand , proposition [ prop.h1-t ] implies that for a general @xmath534 \in \mathcal{m}_{ii}$ ] we have @xmath535 } \mathcal{m}_{ii}=h^1(s , \
, t_s)=3,\ ] ] hence @xmath4 is generically smooth .
[ prop.disjoint ] @xmath2 , @xmath3 and @xmath4 are connected components of @xmath46 .
we proved that @xmath46 is the disjoint union of three irreducible , constructible sets @xmath543
so it is sufficient to show that @xmath2 , @xmath3 , @xmath4 are all open in @xmath46 . in other words ,
given a flat family @xmath544 over a small disk @xmath545 , such that @xmath546 ( resp .
@xmath3 , @xmath4 ) , we must show that @xmath548 ( resp .
@xmath3 , @xmath4 ) for @xmath550 .
we may associate to the family @xmath544 the family @xmath551 , whose fibre over @xmath552 is the stein factorization @xmath553 of @xmath554 , that is the contraction of the elliptic curve @xmath555 . by the previous results it follows that , up to a base change , the family @xmath544 is the double cover of a family @xmath556 of blow - ups @xmath557 of @xmath99-polarized abelian surfaces and the family @xmath551 is the double cover of the family @xmath558 , where @xmath559 is the minimal model of @xmath557 .
globalizing the results of section [ sec.d=2 ] we see that the polarizations @xmath560 on the abelian surfaces @xmath559 glue together in order to give an ample line bundle @xmath561 on @xmath562 and that there exists a divisor @xmath563 on @xmath564 whose restriction to the fibre @xmath557 is the branch locus @xmath565 of @xmath566 .
moreover , we find a commutative diagram @xmath567^{\beta } \ar[dr]_{\alpha } & \mathscr{b } \ar[d]^{\sigma } \\ & \mathscr{a}}\ ] ] and a line bundle @xmath568 of order @xmath12 such that @xmath569 where @xmath570 is the relative blow - down and @xmath571 is the exceptional divisor of @xmath572 .
we denote by @xmath573 the restriction of @xmath574 to @xmath559 .
now let us consider separately the three cases .
* @xmath4 _ is open in _ @xmath46 .
+ it is equivalent to prove that @xmath575 is closed in @xmath46 , namely that @xmath576 for @xmath550 implies @xmath577 .
the condition @xmath578 for @xmath550 implies that @xmath565 is connected for any @xmath550 ; it follows that @xmath579 is also connected , hence @xmath580 is again a surface of type @xmath39 .
* @xmath2 _ is open in _ @xmath46 .
+ assume that @xmath581 . by proposition [ prop.type i ] , this is equivalent to say that the branch locus @xmath579 of @xmath582 is connected and that @xmath583 is base - point free .
clearly these are both open conditions , so @xmath2 is open in @xmath46 . *
@xmath3 _ is open in _ @xmath46 .
+ assume that @xmath584
. then we have @xmath585 $ ] . by proposition [ prop.components ] in the appendix it follows that @xmath533 $ ] is a connected component of @xmath524 $ ] , in particular it is open therein . hence @xmath586 $ ] for @xmath587 , proving that @xmath3 is open in @xmath46 .
notice that the same argument gives an alternative proof of the fact that @xmath2 is open in @xmath46 , since @xmath532 $ ] is the other connected component of @xmath524 $ ] .
this completes the proof of proposition [ prop.disjoint ] . summing up , propositions [ prop.moduli ] , proposition [ prop.disjoint ] and remark [ rem.k ample ]
imply the following result .
[ thm.b ] let @xmath46 be the moduli space of minimal surfaces @xmath7 of general type with @xmath0 , @xmath36 and albanese map of degree @xmath12 .
then the following holds : * @xmath46 is the disjoint union of three connected components , namely @xmath588 * these are also irreducible components of the moduli space of minimal surfaces of general type@xmath50 * @xmath589 are generically smooth , of dimension @xmath52 , respectively@xmath50 * the general surface in @xmath2 and @xmath3 has ample canonical class@xmath50 all surfaces in @xmath4 have ample canonical class .
first let us recall some well - known facts about the moduli space of polarized abelian surfaces , that can be found for instance in ( * ? ? ?
* chapter 8) .
let us denote by @xmath122 the matrix @xmath590 and let @xmath591 be the siegel upper half - space .
we define a polarized abelian surface of type @xmath122 with symplectic basis to be a triplet @xmath592 with @xmath593 an abelian surface , @xmath594 a polarization of type @xmath122 on @xmath25 and @xmath595 a basis of the lattice @xmath596 , symplectic with respect to @xmath594 .
then any @xmath597 determines a polarized abelian surface of type @xmath122 with symplectic basis @xmath598 as follows : just set @xmath599 and let @xmath600 be the columns of the matrix @xmath601 .
moreover , there exists a universal family , that is a holomorphic family @xmath602 parameterizing these objects , see ( * ? ? ?
* section 8.7 ) . if @xmath603 , the polarized abelian surfaces @xmath604 and @xmath605 are isomorphic if and only if @xmath606 , where @xmath607 is the full paramodular group ( see ( * ? ? ?
* chapter 8) , @xcite ) and the action is defined as follows : for any @xmath608 and @xmath597 , we set @xmath609 notice that the following special matrices lie in @xmath57 : @xmath610 @xmath611 @xmath612 the action is properly discontinuos , so the moduli space @xmath124 of @xmath125-polarized abelian surfaces is a quasi - projective variety of dimension @xmath6 , obtained as the quotient @xmath613
. then @xmath57 is the orbifold fundamental group of @xmath124 and there is an induced monodromy action of @xmath57 on both @xmath66 $ ] and @xmath58 $ ] , see @xcite .
let us start by making a couple of observations .
first , the trivial line bundle @xmath615 is obviously invariant for the monodromy action .
second , for the computation of the monodromy we may assume that @xmath616 is @xmath617-dimensional , generated by the numerical class of @xmath26 .
then for any @xmath618 the monodromy transformation associated with @xmath619 sends @xmath67 to @xmath620 , with @xmath621 . since @xmath622
, it follows that @xmath149 is invariant under the monodromy action ; hence @xmath623 is invariant too .
it remains to show that @xmath58 \setminus \textrm{im } \ , \phi_2 $ ] forms a single orbit .
set @xmath624 and write the period matrix for @xmath25 as @xmath625 with @xmath626
. then the lattice @xmath596 is spanned by the four column vectors @xmath627 and the matrix of the alternating form @xmath628 with respect to this basis is @xmath629 therefore @xmath630 and all the other values are @xmath177 .
the finite subgroup @xmath58 $ ] of @xmath65 is isomorphic to @xmath631 and , by the appell - humbert theorem , its elements can be canonically identified with the @xmath632 characters @xmath633 with values in @xmath634 , see ( * ? ? ?
* chapter 2 ) .
since @xmath635 it follows @xmath636 and @xmath637 . in other words , @xmath21 corresponds to the four characters @xmath638 with @xmath639 .
we will denote a character @xmath640 by the vector + @xmath641 .
therefore @xmath150 consists of @xmath642 whereas the @xmath643 elements of @xmath58 \setminus \textrm{im } \ , \phi_2 $ ] correspond to @xmath644 now take @xmath645 , where @xmath646 and @xmath647 . by (
* proof of proposition 8.1.3 ) , the monodromy action of @xmath619 on @xmath596 is given by the matrix @xmath648 , so the induced action over a character @xmath649 is as follows : @xmath650 for instance we have @xmath651 hence the set @xmath652 is @xmath57-invariant ( and by using the matrices of type @xmath653 one checks that it is a single @xmath57-orbit , as expected ) . *
the monodromy permutation associated with a matrix of type @xmath653 is * * @xmath655 if @xmath656 is odd and @xmath657 , @xmath658 are even ; * * @xmath659 if @xmath657 is odd and @xmath656 , @xmath658 are even ; * * @xmath660 if @xmath658 is odd and @xmath656 , @xmath657 are even ; * the monodromy permutation associated with a matrix of type @xmath661 is * * @xmath662 if @xmath663 is odd and @xmath664 is even ; * * @xmath665 if @xmath664 is odd and @xmath663 is even ; * the monodromy permutation associated with the matrix @xmath666 is + @xmath667
. therefore , the subgroup of the symmetric group @xmath668 corresponding to the monodromy action of @xmath57 on @xmath669 contains @xmath670 a straightforward computation , for instance by using the computer algebra system ` gap4 ` ( see @xcite ) , shows that @xmath671 is a _ transitive _ subgroup of @xmath668 ; therefore @xmath654 form a single orbit for the @xmath57-action .
this completes the proof .
now let @xmath672 be a polarized abelian surface of type @xmath122 .
a symplectic basis @xmath673 of @xmath596 for @xmath594 determines the @xmath674 non - trivial characters @xmath675 , @xmath676 .
therefore we can consider the set of pairs @xmath677,\ ] ] which can be seen as a subscheme of the relative picard scheme @xmath678 .
the group @xmath57 acts on this set of pairs , the action being defined by on the first component and by the monodromy on the second one .
the corresponding quotient @xmath524 $ ] is a quasi - projective variety and by construction we have a degree @xmath674 cover @xmath679 { \longrightarrow}\mathcal{a}_{\delta}$ ] .
we can identify @xmath524 $ ] with the set of pairs @xmath525 , where @xmath25 is the isomorphism class of a @xmath125-polarized abelian variety and @xmath198 is a non - trivial , @xmath12-torsion line bundle on @xmath25 ; then the map @xmath680 is just the forgetful map @xmath681 .
[ prop.components ] @xmath524 $ ] is the disjoint union of two connected components : @xmath526
\quad \textrm{and } \quad \mathcal{a}_{\delta}^{(b)}[2],\ ] ] corresponding to @xmath682 and @xmath683 , respectively .
the forgetful maps
@xmath684 \longrightarrow \mathcal{a}_{\delta } , \quad \pi_2 \colon \mathcal{a}_{\delta}^{(b)}[2]\longrightarrow \mathcal{a}_{\delta}\ ] ] are finite covers of degree @xmath643 and @xmath6 .
finally , both @xmath532 $ ] and @xmath533 $ ] are irreducible and generically smooth . the first part of the statement follows immediately since the action of @xmath57 on the set of non - trivial characters @xmath685 has precisely two orbits , namely @xmath686 and @xmath654 ( proposition [ prop.monodromy ] ) .
moreover @xmath687 and @xmath688 are @xmath354tale covers on a smooth zariski open set @xmath689 ; then they are generically smooth . finally , by construction @xmath532 $ ] and @xmath533 $ ] are normal varieties , because they only have quotient singularities .
then , since they are connected , they must be also irreducible .
similarly , there is an action of @xmath690 on the set of triplets @xmath691 where @xmath597 , @xmath692 $ ] and @xmath693 is a character whose square is @xmath649 .
the corresponding quotient is a quasi - projective variety that can be identified with the space @xmath539 $ ] of triples @xmath694 , where @xmath25 is the isomorphism class of a @xmath99-polarized abelian surface , @xmath695 and @xmath30 is a square root of @xmath31 . there is forgetful map @xmath696 { \longrightarrow}\mathcal{a}_{\delta}$ ] , sending @xmath694 to @xmath25 ; it is a finite cover of degree @xmath697 .
it is sufficient to check that the monodromy action of @xmath57 is transitive on the set @xmath699 this is a straightforward computation which can be carried out as the one in the proof of proposition [ prop.monodromy ] , so it is left to the reader
. 9 w. barth , _
abelian surfaces with @xmath53-polarization_. adv .
pure math .
* 10 * ( 1987 ) , 4184 . c. birkenhake , h. lange , _ complex abelian varieties_. grundlehren der mathematischen wissenschaften , vol * 302 * , second edition , springer - verlag , berlin , 2004 . w. barth , k. hulek , c.a.m .
peters , a. van de ven , _ compact complex surfaces_. grundlehren der mathematischen wissenschaften , vol * 4 * , second enlarged edition , springer - verlag , berlin , 2003 . f. bastianelli , g. p. pirola , l. stoppino , _ galois closure and lagrangian varieties_. adv . math .
, * 225 * , ( 2010 ) , 34633501 .
f. catanese , _ footnotes to a theorem of i. reider_. algebraic geometry ( laquila , 1988 ) , lecture notes in math .
, vol * 1417*. berlin , 1990 , 6774 .
f. catanese , _ moduli and classification of irregular kaehler manifolds ( and algebraic varieties ) with albanese general type fibrations_. invent .
* 104 * ( 1991 ) , 263 - 289 .
f. catanese , _ a superficial working guide to deformations and moduli _ , e - print arxiv:1106.1368 , to appear in in the handbook of moduli , a volume in honour of david mumford , to be published by international press . c. hacon , r. pardini , _ surfaces with @xmath38_. trans .
soc . * 354 * ( 2002 ) , 26312638 .
e. horozov , p. van moerbeke , _ the full geometry of kowalewski s top and @xmath53-abelian surfaces_. comm .
pure appl . math .
* 42 * ( 1989 ) , no .
4 , 357407 .
m. manetti , _ smoothing of singularities and deformation types of surfaces _ , symplectic @xmath5-manifolds and algebraic surfaces , lecture notes in math .
1938 , springer 2008 , 169230 .
s. mukai , _ duality between @xmath700 and @xmath701 with its application to picard sheaves_. nagoya math . j.
* 81 * ( 1981 ) , 153175 .
s.mukai , _ moduli of abelian surfaces and regular polyhedral groups _ , in moduli of algebraic varieties and the monster " , proceedings , sapporo , january 1999 .
i. nakamura , hokkaido univ .
( 1999 ) , 68 - 74 .
m. penegini , _ the classification of isotrivial fibred surfaces with @xmath0 _ , with an appendix by s. roellenske .
* 62 * , no .
3 , ( 2011 ) , 239274 . m. penegini , f. polizzi , _ on surfaces with @xmath0 , @xmath702 and albanese map of degree @xmath6 _ , e - print arxiv:1011.4388 ( 2010 ) .
pirola , _ surfaces with @xmath38_. manuscripta math . * 108 * no . 2 ( 2002 ) , 163170 .
s. rollenske , _ compact moduli for certain kodaira fibrations _ , annali della scuola normale superiore di pisa , vol .
ix , issue 4 ( 2010 ) .
e. sernesi , _ deformations of algebraic schemes_. grundlehren der mathematischen wissenschaften , vol * 334 * , springer - verlag , berlin , 2006 .
f. zucconi , _ surfaces with @xmath0 and an irrational pencil_. canad . j. math .
* 55 * ( 2003 ) , no .
3 , 649672 . matteo penegini , lehrstuhl mathematik viii , universitt bayreuth , nwii , d-95440 bayreuth , + germany + _ e - mail address : _ ` [email protected] ` + + francesco polizzi , dipartimento di matematica , universit della calabria , cubo 30b , 87036 +
arcavacata di rende ( cosenza ) , italy + _ e - mail address : _ ` [email protected] ` | we classify minimal surfaces of general type with @xmath0 and @xmath1 whose albanese map is a generically finite double cover .
we show that the corresponding moduli space is the disjoint union of three generically smooth , irreducible components @xmath2 , @xmath3 , @xmath4 of dimension @xmath5 , @xmath5 , @xmath6 , respectively . |
extragalactic radio sources have been studied for many years , but it is still unclear how they are formed and evolve .
a crucial element in the study of their evolution is to identify the young counterparts of the `` old '' fri / frii extended objects .
good candidates for young radio sources are those with peaked spectra ( gigahertz peaked spectrum ( gps ) sources and compact steep spectrum ( css ) sources , fanti et al .
1995 , readhead et al .
1996 , odea and baum 1997 , odea 1998 ) , because they are small in angular size as expected for young sources .
gps sources are characterized by a convex radio spectrum peaking at a frequency of about 1 ghz and are typically 100 pc in size .
css sources have a peak in their spectrum at lower frequencies and have projected linear sizes of @xmath4 kpc . a commonly discussed alternative to them being young ,
is that they are small due to confinement by a particularly dense and clumpy interstellar medium that impedes the outward propagation of the jets ( odea et al 1991 ) .
important information about the nature of gigahertz peaked spectrum gps sources comes from the properties of their optical counterparts .
systematic optical identification programs on gps sources have been performed by odea et al .
( 1990 ) , stanghellini et al .
( 1993 ) and de vries et al .
( 1995 ) , and these show that the optical counterparts are a mixture of galaxies and quasars . odea et al .
( 1990 ) and stanghellini et al .
( 1993 ) found that disturbed optical morphologies are a common characteristic of gps galaxies , with many of them having close companions or appearing to lie in a group or cluster .
this may indicate that galaxy - galaxy interactions and/or mergers play an important role in the gps phenomenon .
gps galaxies appear to be a homogeneous class of giant ellipticals with old stellar populations ( snellen et al .
1996a , 1996b , odea et al .
their optical to near - infrared fluxes seem not to be contaminated by light produced by the active nucleus , and they are therefore potentially useful probes of galaxy evolution . in the light of the orientation unification scheme ( eg .
barthel 1989 ) , it would be expected that the redshift distributions of gps sources identified with galaxies and quasars are more or less the same .
however , the gps quasars tend to be at much higher redshift than the galaxies ( @xmath5 , odea et al .
in addition , the intrinsic peak frequency distributions and radio morphologies seem to be different for gps galaxies and quasars ( stanghellini et al . 1996 ) .
these results are probably influenced by several selection effects , but if true , they may indicate that gps galaxies and quasars are not unified by orientation , but form separate classes of object which happen to have similar radio spectra . to further address these issues ,
we have started to investigate gps sources at fainter flux density levels ( snellen et al .
1995 , snellen et al .
1998 ) than those in the samples of bright gps sources studied until now ( odea et al .
1991 , de vries et al .
1997 , stanghellini et al .
1998 ) . by comparing the properties of these new faint samples with those of bright samples in the literature
, we shall investigate properties of gps sources as a function of radio luminosity , redshift and rest frame peak frequency .
this paper describes the optical and near - infrared imaging of a sample of faint gps sources selected from the westerbork northern sky survey ( wenss rengelink et al .
the selection of the sample has been described in detail in snellen et al .
1998 ( paper i ) .
candidate gps sources selected from the westerbork northern sky survey , are those with an inverted spectrum between 325 mhz and higher frequencies .
the sources are located in two regions of the survey ; one at @xmath6 and @xmath7 , which is called the _ mini - survey _ region ( rengelink et al . 1997 ) , and the other at @xmath8 and @xmath7 .
additional observations at 1.4 , 5 , 8.4 and 15 ghz were carried out with the westerbork synthesis radio telescope ( wsrt ) and the very large array ( vla ) , yielding a sample of 47 genuine gps sources with peak frequencies ranging from 500 mhz to more than 15 ghz , and peak flux densities ranging from @xmath9 to @xmath10 mjy . to determine whether the gps sources have an optical counterpart on the palomar sky survey plates ( poss ) , the automatic plate measuring ( apm ) catalogue ( irwin et al .
1994 ) and the hst guide star catalogue ( russel et al . 1990 ) were used for the sources in the mini - survey region and in the region @xmath11 , respectively .
this information was used for finding charts , to determine the required exposure times for the observations , and for astrometric purposes .
the sample was imaged in the optical in the course of three observing sessions ( see table [ observations ] ) , with the 0.9 m jacobus kapteyn telescope ( jkt ) , the 2.5 m isaac newton telescope ( int ) and the 2.5 m nordic optical telescope ( not ) , all located at observatorio del roque de los muchachos on la palma , spain .
the sample was also observed in the near - infrared with the 4.2 m william herschel telescope ( wht ) , and with the not .
the int , not and wht observations were done as part of the international cci observing time awarded to the wenss team in 1995 . during the different observing sessions , several standard stars from landolt ( 1983 ) were observed for photometric calibration .
a summary of the relevant parameters for the observations is given in table [ observations ] . | a sample of 47 faint gigahertz peaked spectrum ( gps ) radio sources selected from the westerbork northern sky survey ( wenss , rengelink et al .
1997 ) , has been imaged in the optical and near infrared , resulting in an identification fraction of 87% . the @xmath0@xmath1@xmath2 and @xmath0@xmath1@xmath3 colours of the faint optical counterparts are as expected for passively evolving elliptical galaxies , assuming that they follow the @xmath0 band hubble diagram as determined for radio - bright gps galaxies .
we have found evidence that the radio spectral properties of the gps quasars are different from those of gps galaxies : the observed distribution of radio spectral peak frequencies for gps sources optically identified with bright stellar objects ( presumably quasars ) is shifted compared with gps sources identified with faint or extended optical objects ( presumably galaxies ) , in the sense that a gps quasar is likely to have a higher peak frequency than a gps galaxy .
this means that the true peak frequency distribution is different for the gps galaxies and quasars , because the sample selection effects are independent of optical identification .
the correlation between peak frequency and redshift as has been suggested for bright sources has not been found in this sample ; no correlation exists between @xmath0 magnitude ( and therefore redshift ) and peak frequency for the gps galaxies .
we therefore believe that the claimed correlation is actually caused by the dependence of the peak frequency on optical host , because the gps galaxies are in general at lower redshifts than the quasars .
the difference in the peak frequency distributions of the gps galaxies and quasars is further evidence against the hypothesis that they form a single class of object . |
0 the effect of curvature on the behaviour of the heat flow on a riemannian manifold is a classical problem .
ricci curvature manifests itself most directly in gradient formulas for solutions of heat equation .
gradient estimates for positive solutions of the heat equation serve as infinitesimal versions of harnack inequalities : by integrating along curves on the manifold local gradient estimates may be turned into local harnack type inequalities .
solutions to the heat equation @xmath0 on a riemannian manifold @xmath1 are well understood in probabilistic terms .
for instance , if @xmath2 denotes the minimal solution to , then a straightforward calculation using it s calculus leads to the stochastic representation of @xmath3 as @xmath4\ ] ] where @xmath5 is the initial condition in , @xmath6 denotes a brownian motion on @xmath1 , starting from @xmath7 at time @xmath8 , and @xmath9 its lifetime .
it is a remarkable fact that exact stochastic representation formulas for the derivative of solutions to the heat equation can be given , expressing the gradient @xmath10 of @xmath3 in terms of ricci curvature .
the following typical example for such a bismut type derivative formula is taken from @xcite . [
stochastic representation of the gradient][bismut ] let @xmath11 be a relatively compact open domain in a complete riemannian manifold @xmath1 , and let @xmath2 be a solution of the heat equation on @xmath12}$ ] which is continuous on @xmath13}$ ] . then , for any @xmath14 and @xmath15 , @xmath16 , \label{bismut1}\ ] ] where : 1 .
@xmath17 is a brownian motion on @xmath1 , starting at @xmath7 , and @xmath18 where @xmath19 is the first exit time from @xmath11 ; the stochastic integral is taken with respect to the brownian motion @xmath20 in @xmath21 , related to @xmath22 by the stratonovich equation @xmath23 , where @xmath24 denotes the stochastic parallel transport along @xmath22 .
2 . the process @xmath25 takes values in the group of linear automorphisms of @xmath21 and is defined by the pathwise covariant ordinary differential equation , @xmath26 where @xmath27 ( a linear transformation of @xmath21 ) , and @xmath28 for any @xmath29 , @xmath30 .
3 . finally , @xmath31 may be any adapted finite energy process taking values in @xmath21 such that @xmath32 , @xmath33 and @xmath34 formula is valid for any solution @xmath3 of the heat equation and does not require positivity of @xmath3 .
formula is easily adapted to more specific situations , for instance : \(i ) let @xmath2 be a solution of the heat equation on @xmath12}$ ] such that @xmath35 and @xmath36
. then @xmath37 , \quad \tau=\tau(x)\wedge t\,.\ ] ] \(ii ) let @xmath2 be a solution of the heat equation on @xmath12}$ ] such that @xmath35 and @xmath38 .
then @xmath39 \,.\ ] ] such formulas are interesting by several means . for instance , on a complete riemannian manifold @xmath1 , starting from the minimal solution to the heat equation @xmath40\ ] ] with bounded initial conditions @xmath41 , since for arbitrarily small @xmath42 brownian motion explores the whole manifold @xmath1 , we observe that the global structure of @xmath1 enters in formula ; lower ricci bounds may fail and thus `` brownian motion may travel arbitrarily fast '' .
nevertheless , looking at the formula for the gradient @xmath43 and taking into account that @xmath44,\ ] ] we see that eq .
reads as @xmath45\ ] ] where @xmath18 and @xmath46 the first exit time of @xmath6 from an arbitrarily small chosen neighbourhood of @xmath7 .
in other words , as far as the gradient at @xmath47 is concerned , ricci curvature of @xmath1 matters only locally about the point @xmath7 .
no derivative of the heat equation appears in the right - hand side of eq .
, thus gradient formulas are well suitable for estimates , see @xcite .
such inequalities easily allow to bound @xmath10 in terms of some uniform norm of @xmath3 .
for positive solutions of the heat equation however one wants to do better : typically one seeks for pointwise estimates which allow ( modulo additional terms if necessary ) to control @xmath43 by @xmath48 .
[ classical li - yau estimate @xcite][li - yau_class ] let @xmath1 be complete riemannian manifold and assume that @xmath49 where @xmath50 .
let @xmath3 be a strictly positive solution of @xmath51 and let @xmath52 . then @xmath53.\ ] ] if @xmath54 , i.e. @xmath55 , then the choice @xmath56 is possible . it is a surprising fact which has been noticed by many people that no straightforward way to pass from bismut type derivative formulas to li - yau type gradient estimates seems to exist .
li - yau type inequalities for positive solutions @xmath3 of the heat equation aim at estimating @xmath57 rather than @xmath10 .
this adds an interesting non - linearity to the problem which is better to deal with in terms of submartingales and bismut type inequalities than in terms of martingales which are the underlying concept for bismut formulas .
this point of view has been worked out in @xcite for local estimates in the elliptic case of positive harmonic functions .
such estimates in global form , i.e. , for positive harmonic functions on riemannian manifolds , are due to s.t .
yau @xcite ; local versions have been established by cheng and yau @xcite . in this paper
we pursue the approach via submartingales to study the parabolic case .
even if it is meanwhile quite standard to obtain li - yau type estimates in global form via analytic methods , local versions require often completely new arguments @xcite . in this paper
we derive various submartingales which lead to the wanted estimates in a surprisingly simple way .
0 the following formulas for solutions of the heat equation on a riemannian manifold depend on the fact that the solutions are strictly positive .
[ lemma:1.1 ] let @xmath1 be a riemannian manifold ( not necessarily complete ) and let @xmath2 be a positive solution of the heat equation @xmath58 on @xmath59 $ ] .
then the following equalities hold : @xmath60 all three equalities are easily checked by direct calculation .
eq . in lemma [ lemma:1.1 ] gives raise to some inequalities frequently used in the sequel and crucial for our approach .
most of our results are based on the following observation .
[ cor:1.1 ] let @xmath1 be a riemannian manifold of dimension @xmath61 ( not necessarily complete ) and let @xmath2 be a positive solution of the heat equation
. then we have : @xmath62 if @xmath49 on @xmath1 for some @xmath50 , then @xmath63 and in particular , @xmath64 this is a direct consequence of eq . .
[ lemma:2.1 ] let @xmath1 be a riemannian manifold ( not necessarily complete ) and let @xmath65 be a positive solution of the heat equation @xmath66 on @xmath59 $ ] . for any brownian motion @xmath22 on @xmath1 , the process @xmath67 where @xmath68 denotes the smallest eigenvalue of the ricci curvature at the point @xmath7 , is a local submartingale ( up to its natural lifetime ) . furthermore @xmath69 are local martingales ( up to their respective lifetimes ) .
the first claim is a consequence of ; the second part comes from eqs . and .
[ lemma:2.4 ] let @xmath1 be a riemannian manifold and @xmath65 be a positive solution of the heat equation on @xmath70}$ ] . if @xmath49 for some @xmath50 , then for any brownian motion @xmath22 on @xmath1 , the process @xmath71 is a local submartingale ( up to its lifetime ) .
the proof follows from it s formula using inequality , along with eq . .
0 in this section we explain how submartingales related to positive solutions of the heat equation can be turned into gradient estimates .
the resulting estimates of this section are classical inequalities ; our focus lies on the stochastic approach .
the main problem in the subsequent sections will then be to use methods of stochastic analysis to derive localized versions of the bounds .
the following gradient estimates follow immediately from lemma [ lemma:2.4 ] .
[ thm : entropy ] let @xmath65 be a positive solution of the heat equation on a compact manifold @xmath1 .
assume that @xmath49 for some @xmath50 .
then @xmath72 indeed if @xmath1 is compact , then the local submartingale @xmath73 in is a true submartingale .
let @xmath74 then , evaluating @xmath75\leq{\mathbb{e}}[n_t]$ ] , we obtain @xmath76 or in other words , @xmath77 dividing through @xmath78 completes the proof . keeping notation and assumptions of theorem
we observe that , for any @xmath79 , @xmath80 .
\label{eq : entropy1}\end{aligned}\ ] ] indeed with @xmath81 , we conclude from that @xmath82 let @xmath1 be a compact riemannian manifold such that @xmath49 throughout @xmath1 for some @xmath50 .
suppose that @xmath83 is a positive solution of the heat equation on @xmath70}$ ] , and let @xmath84}}u$ ] .
then @xmath85 in particular , if @xmath54 then @xmath86 the proof of is an application of theorem [ thm : entropy ] , along with the observation that @xmath87 when @xmath88 is normalized as @xmath89 for fixed @xmath7 . in the proofs above compactness of the manifold
has only been used to assure that the local submartingale is a true submartingale .
the results extend to bounded positive solution of the heat equation on complete manifolds with lower ricci bounds .
indeed , @xmath65 may be assumed to be bounded away from @xmath8 ( otherwise one may first pass to @xmath90 and let @xmath91 tend to @xmath8 in the obtained estimate ) .
the term @xmath92 may be bounded by bismut s formula , see @xcite .
we now turn to the classical li - yau estimate .
let @xmath1 be a given complete riemannian manifold of dimension @xmath61 such that @xmath49 for some @xmath50 .
suppose that @xmath3 is a positive solution to the heat equation on @xmath70}$ ] .
starting from in corollary [ cor:1.1 ] we have @xmath93 and thus @xmath94 where @xmath95 fixing @xmath96 , let @xmath97 be a brownian motion on @xmath1 starting at @xmath7 .
our first goal is to investigate the process @xmath98 where @xmath99 for some adapted continuous real - valued process @xmath100 with absolutely continuous paths such that @xmath101 and @xmath102 .
let @xmath103 , @xmath104 and @xmath105 .
using we find ( modulo differentials of local martingales ) @xmath106dt.\end{aligned}\ ] ] minimizing the term @xmath107 as a quadratic function of @xmath108 , we find @xmath109 thus , integrating from @xmath8 to @xmath110 and taking expectations , we obtain @xmath111\ ] ] [ thm : li_yau_ric0 ] let @xmath1 be a complete riemannian manifold of dimension @xmath61 such that @xmath54 .
let @xmath2 be a positive bounded solution of the heat equation on @xmath70}$ ] .
then , for each @xmath15 , @xmath112\ ] ] where @xmath113 is an adapted continuous real - valued process @xmath100 with absolutely continuous paths such that @xmath101 and @xmath102 . in particular , with the choice @xmath114 , we obtain @xmath115 which is the classical estimate of li - yau . by we have @xmath116\ ] ] first note that @xmath117 is a solution of the heat equation as well , hence @xmath118 a local martingale .
in particular , @xmath119 where @xmath120 stands for equality modulo differentials of local martingales , and hence @xmath121=-h_0\,(\delta u)_0=-\delta u(x , t).\ ] ] thus shows that @xmath122=n\,{\mathbb{e}}\left[\int_0^t\dot\ell_t^2\,dt\cdot \frac{u_t}{u_0}\right].\ ] ] \(i ) using , theorem [ thm : li_yau_ric0 ] is easily extended to the case of a lower ricci bound . for local versions of the li - yau s estimate one could try to modify the process @xmath100 in such a way that @xmath100 already vanishes as soon as the brownian motion @xmath123 reaches the boundary of @xmath11 .
we shall not pursue this approach here , but rather adopt an even simpler argument in the next section which leads to the local estimate .
\(ii ) of particular interest are localized versions of the entropy estimates and .
such estimates lead to harnack inequalities and heat kernel bounds , valid on arbitrary manifolds without bounded geometry , see @xcite .
results in this direction will be worked out elsewhere @xcite .
0 our main task of the remaining sections will be to localize the arguments of section [ section3 ] to cover local solutions of the heat equation on bounded domains .
we start with the li - yau estimate .
[ ass1 ] let @xmath1 be a complete riemannian manifold of dimension @xmath61 , and let @xmath2 be a solution of the heat equation on @xmath12}$ ] where @xmath124 is a relatively compact open subset of @xmath1 with nonempty smooth boundary .
assume that @xmath3 is positive and continuous on @xmath13}$ ] .
furthermore let @xmath125 be a lower bound for the ricci curvature on the domain @xmath11 .
finally let @xmath126 with @xmath127 in @xmath11 and @xmath128 .
[ ass2 ] let @xmath15 and @xmath129 be a brownian motion on @xmath1 starting at @xmath7 at time @xmath8 .
denote by @xmath46 its first exit time of @xmath11 .
now , for @xmath130 , consider the process @xmath131 where @xmath132 , @xmath133 are constants which will be specified later .
let @xmath134 be the solution of @xmath135 in other words , @xmath136 then , letting @xmath137 we find ( modulo differentials of local martingales ) @xmath138 dt.\ ] ] thus letting @xmath139 we get ( modulo differentials of local martingales ) @xmath140_t\\ & \ge \left[h_t\frac{q_t^2}{nu_t}+\left(\dot h_t - kh_t\right)q_t - kh_t(\delta u)_t\right]dt\\ & \quad+nu_t\,h_ty_t^2\,dt\\ & \quad - nu_th_t\left[c_1\,(t - t)^{-2}+c_2\,c_\varphi(x_t)\,\varphi^{-4}(x_t)\right]dt\\ & \quad - nc_2\,h_t\,d [ u,\varphi^{-2}(x(x))]_t\end{aligned}\ ] ] where the bracket @xmath141 $ ] stands for quadratic covariation on the space of continuous semimartingales and where @xmath142 on the other hand , denoting again @xmath143 we have for any @xmath144 , @xmath145_t\\ & = nc_2\,h_t2\varphi^{-3}(x_t)\langle \nabla u_t,\nabla \varphi(x_t)\rangle\,dt\\ & \ge -2nc_2\,h_t\varphi^{-3}(x_t)|\nabla u_t|\,|\nabla \varphi(x_t)|\,dt\\ & = -2nc_2\,h_t\left(\varphi^{-1}(x_t)\,(\alpha nu_t)^{-1/2}\,|\nabla u_t|\right)\\ & \qquad\times\left((\alpha nu_t)^{1/2}(\varphi^{-2}|\nabla \varphi|)(x_t)\right)\,dt\\ & \ge \left(-\alpha^{-1}c_2\,h_t\varphi^{-2}(x_t)(q_t+\delta u_t ) -c_2\,\alpha n^2u_th_t\varphi^{-4}|\nabla \varphi|^2(x_t)\right)\,dt\\ & \ge \left(-\alpha^{-1}h_t(y_t-\alpha k)(q_t+\delta u_t ) -c_2\,\alpha
n^2u_th_t\varphi^{-4}|\nabla \varphi|^2(x_t)\right)\,dt\\ & = \left(\left(\alpha^{-1}\dot h_t+kh_t\right)(q_t+\delta u_t ) -c_2\,\alpha n^2u_th_t\varphi^{-4}|\nabla \varphi|^2(x_t)\right)\,dt.\end{aligned}\ ] ] hence letting , for @xmath146 , @xmath147 we get @xmath148dt\\ & \quad+h_ty_t^2\,nu_t\,dt\\ & \quad - nu_th_t\big[c_1(t - t)^{-2}+c_2\,c_{\varphi,\a , n}\,\varphi^{-4}(x_t)\big]\,dt\end{aligned}\ ] ] where @xmath149 minimizing the first line on the right hand side , we find @xmath150\,dt.\end{aligned}\ ] ] putting things together , we arrive at the following result .
we keep notation and assumptions from above .
assume that @xmath151 and let @xmath152 and @xmath153 then @xmath154 is a local submartingale on @xmath155 .
using @xmath156 we get @xmath157 thus , under condition and , the right - hand side is nonnegative .
[ t1 ] we keep the assumptions from above .
let @xmath2 be a solution of the heat equation on @xmath12}$ ] which is positive and continuous on @xmath13}$ ] . for any @xmath1581,\infty[}$
] , we have @xmath159 where @xmath160 is a lower ricci bound on the domain @xmath11 , and where @xmath161 and @xmath162 are specified in , resp . .
the function @xmath163 is as in assumption .
recall that @xmath164 and @xmath165 according to , resp .
let @xmath166 consider @xmath167 we assume that @xmath168 and let @xmath169 be the first hitting time of @xmath8 by @xmath154 . then clearly @xmath170 , since @xmath171 converges to @xmath172 as @xmath173 tends to @xmath174 .
let @xmath175}(q-\alpha^{-1}\delta u)\quad\text{and } \quad m=\min_{\bar d\times [ 0,t]}u.\ ] ] then , for @xmath176 $ ] , we have @xmath177 , which implies @xmath178 . from this
we easily prove that on @xmath179 $ ] , the process @xmath180 is a submartingale with bounded local characteristics . as a consequence
, we have @xmath181.\ ] ] since @xmath182 , this contradicts the assumption @xmath168 .
hence we must have @xmath183 , which is the desired inequality . in the case of a global solution of the heat equation on a compact manifold
, we can take @xmath163 as a constant and then @xmath184 .
if moreover @xmath55 , then one can take @xmath185 and @xmath186 , and @xmath187 is a local submartingale .
this recovers one more time the usual li - yau estimate @xmath188 a similar reasoning applies for global solutions on complete riemannian manifolds with a lower ricci curvature bound .
0 we keep assumption [ ass1 ] and [ ass2 ] of section [ section4 ] as standing assumptions for the rest of the paper and study now the process @xmath189 where @xmath190 with constants @xmath191 to be specified later .
let @xmath192 then we have @xmath193dt\\ & \qquad - c_2\,nd\,[\varphi^{-2}(x),u]_t\\ & \ge \left[\frac{q_t^2}{nu_t}-nu_t\left(\frac{c_1}{(t - t)^2 } + \frac{c_2\,c_\varphi}{\varphi^4(x_t ) } + \frac{k}{n}\left\|\frac{|\nabla u|}{u}\right\|_{\bar d\times [ 0,t]}^2\right)\right]dt\\ & \qquad-2c_2\,nu_t\left\|\frac{|\nabla
u|}{u}\right\|_{\bar d\times [ 0,t ] } { \left\| \varphi\nabla \varphi\right\|_d^2}\,{\varphi^{-4}(x_t)}\ , dt\\ & \ge \left[\frac{q_t^2}{nu_t}-nu_t\left(\frac{c_1}{(t - t)^2}+\frac{c_2}{\varphi^{4}(x_t ) } \phantom{\sqrt\frac12}\right.\right.\\ & \qquad \times\left.\left.\left(c_\varphi+2\left\|\frac{|\nabla u|}{u}\right\|_{\bar d\times [ 0,t]}\left\| \varphi\nabla \varphi\right\|_d\right)\right)\right]dt\\ & \qquad - nu_t\,\frac{k}{n}\left\|\frac{|\nabla u|}{u}\right\|_{\bar d\times [ 0,t]}^2dt.\end{aligned}\ ] ] let @xmath194}\left\| \varphi\nabla \varphi\right\|_\infty\quad\text{and}\\ & c_3=\sqrt{\frac{k}{n}}\left\|\frac{|\nabla u|}{u}\right\|_{\bar d\times [ 0,t]}. \end{split}\end{aligned}\ ] ] then on @xmath195 , the process @xmath196 has nonnegative drift .
we have @xmath197 consequently , under condition , @xmath198 and the right - hand side is nonnegative on @xmath195 .
similarly to theorem [ t1 ] , we obtain the following result .
[ t2 ] we keep the notation from above , as well as assumption and from section .
let @xmath2 be a solution of the heat equation on @xmath12}$ ] which is positive and continuous on @xmath13}$ ]
. then @xmath199}.\ ] ]
0 we keep assumptions [ ass1 ] and [ ass2 ] of section and study now the process @xmath200 where @xmath201 for some constants @xmath132 , @xmath202 , @xmath203 .
assume that @xmath204 ( this assumption will be removed in theorem [ t3 ] through replacing @xmath3 by @xmath205}$ ] ) .
let @xmath206 again be given by .
then , denoting @xmath207 and using the fact that @xmath208 we get @xmath209 now from @xmath210 we get @xmath211 this together with @xmath212 yields ( modulo differentials of local martingales ) @xmath213 dt\\ & \quad - c_2\left[(1-\log u_t)\,2\sqrt 2\ ,
\varphi^{-2}(x_t)\,|\nabla \varphi(x_t)|\,\sqrt{u_t}\ , \varphi^{-1}(x_t)\,\sqrt{\frac{g_t}2\,}\,\right]dt\bigg\}\\ & \ge ( 1-\log u_t)\bigg\{\left[\left(\frac32 z_t - k\right)\frac12\,g_t\vphantom{\frac12}\right.\\ & \quad\left.\vphantom{\frac12}-u_t(1-\log u_t ) \left(c_1(t - t)^2+c_2c_\varphi(x_t)\varphi^{-4}(x_t)\right)\right ] dt\\ & \quad-\left[(1-\log u_t)^2u_t\,4c_2\varphi^{-4}(x_t)\,|\nabla \varphi(x_t)|^2 + \frac14 g_tc_2\varphi^{-2}(x_t)\right ] dt\bigg\}\\ & \ge ( 1-\log u_t)\biggl[\left ( z_t - k\right)\frac12\,g_t - u_t(1-\log u_t)^2\\ & \quad \times\big(c_1(t - t)^2+c_2\left(c_\varphi(x_t)+4|\nabla \varphi(x_t)|^2\right)\varphi^{-4}(x_t)\big)\bigg]dt.\end{aligned}\ ] ] letting @xmath214 we get @xmath215\,dt\\ & \ge ( 1-\log u_t)(z_t - k)\left [ \frac12\,g_t - u_t(1-\log u_t)^2z_t \right]\,dt.\end{aligned}\ ] ] this proves that @xmath187 has nonnegative drift on @xmath216 . on the other hand
, @xmath187 converges to @xmath172 as @xmath217 .
similarly to theorem [ t1 ] , we obtain the following result .
[ t3 ] we keep the assumptions as above .
assume that @xmath3 is a solution of the heat equation on @xmath12}$ ] which is positive and continuous on @xmath13}$ ]
. then @xmath218}}{u_0}\right)^2\!\ ] ] where @xmath163 is as above .
0 the estimates in theorems [ t1 ] , [ t2 ] and [ t3 ] have been given in terms of a function @xmath126 such that @xmath127 in @xmath11 and @xmath128 .
to specify the constants an explicit choice for @xmath163 has to be done .
we fix @xmath15 and let @xmath219 where @xmath220 denotes the riemannian distance .
we replace @xmath11 by the ball @xmath221 and consider on @xmath20 @xmath222 clearly @xmath223 , @xmath163 is nonnegative and bounded by @xmath224 , and @xmath163 vanishes on @xmath225 .
it is proven in @xcite that @xmath226 where by convention @xmath227 at points where @xmath228 is not differentiable . moreover , since the time spent by @xmath123 on the cut - locus of @xmath7 is a.s .
zero , the differential of the brackets @xmath229 $ ] may be taken as @xmath8 at points where @xmath228 is not differentiable . as a consequence ,
all estimates in theorems [ t1 ] , [ t2 ] and [ t3 ] remain valid with @xmath163 defined by .
we are now going to derive explicit expressions for the constants . to this end
we observe that @xmath230 from @xcite , we get @xmath231 which gives for any @xmath232 , @xmath233 this yields @xmath234}{4\d_x^2}+\frac{k}{4\b},\ ] ] @xmath235}{4\d_x^2}+\frac{k}{4\b},\ ] ] @xmath236}{4\d_x^2}+\frac{k}{4\b}.\ ] ] finally we replace @xmath237 by @xmath238 to obtain from theorems [ t1 ] , [ t2 ] and [ t3 ] the following explicit upper bounds .
[ c1 ] let @xmath3 be a solution of the heat equation on @xmath12}$ ] where @xmath11 is a relatively compact open subset of a riemannian manifold @xmath1 .
assume that @xmath3 is positive and continuous on @xmath13}$ ] .
furthermore let @xmath160 be a lower bound for the ricci curvature on the domain @xmath11 .
fix @xmath15 and let @xmath2391,2[}$ ] .
for any @xmath232 we have @xmath240}{(4-a^2)\d_x^2}\\ & \quad+\left(\frac{1}{(4-a^2)\b}+\frac{1}{a-1}\right ) nk,\\ \left|\frac{\nabla u_0}{u_0}\right|^2-\frac{\delta u_0}{u_0}&\le \frac{n}{t } + \frac{n\pi^2\left[(1+\b)n+3\right]}{4\d_x^2}+\frac{nk}{4\b}\\ & \quad+\sqrt{nk}\left\|\frac{|\nabla u|}{u}\right\|_{\bar d\times [ 0,t ] } + \frac{n\pi}{\d_x}\left\|\frac{|\nabla u|}{u}\right\|_{\bar d\times [ 0,t ] } \intertext{and } \left|\frac{\nabla u_0}{u_0}\right|^2 & \le 2\left(\frac1t+\frac{\pi^2\left[(1+\b)n+7\right]}{4\d_x^2}+\left(\frac{1}{4\b}+1\right)k\right)\\ & \qquad\qquad\qquad\qquad\times\left(4+\log \frac{\|u\|_{\bar d\times [ 0,t]}}{u_0}\right)^2\end{aligned}\ ] ] where
@xmath61 denotes the dimension of @xmath1 and @xmath241 the riemannian distance of @xmath7 to the boundary of @xmath11 .
recall that @xmath164 and @xmath165 . | in this paper we use methods from stochastic analysis to establish li - yau type estimates for positive solutions of the heat equation . in particular , we want to emphasize that stochastic analysis provides natural tools to derive local estimates in the sense that the gradient bound at given point depends only on universal constants and the geometry of the riemannian manifold locally about this point . |
in modern cosmology our notions concerning the early universe have introduced a new element , the inflationary phase of the early universe @xcite , which provides an attractive scenario for solving some of the fundamental puzzles of the standard big bang model , like the horizon and the flatness problems as well as providing a framework for sensible calculations of primordial density perturbations . even in the context of the inflationary scenario
however one encounters the initial singularity problem which remains unsolved , showing that the universe necessarily had a beginning for generic inflationary cosmological model@xcite .
recently models that can avoid those conclusions have been discovered @xcite .
the way to escape the singularity in these models is to violate the geometrical assumptions of these theorems , which assume i ) that the universe has open space sections ii ) the hubble expansion is always greater than zero in the past . in @xcite,@xcite the open space section condition is violated since closed robertson walker universes with @xmath1 are considered and the hubble expansion can become zero , so that both i ) and ii ) are avoided . in @xcite
even models based on standard general relativity , ordinary matter and minimally coupled scalar fields were considered and can provide indeed a non singular ( geodesically complete ) inflationary universe , with a past eternal einstein static universe that eventually evolves into an inflationary universe .
those most simple models suffer however from instabilities , associated with the instability of the einstein static universe .
the instability is possible to cure by going away from gr , considering non perturbative corrections to the einsteins field equations in the context of the loop quantum gravity@xcite , a brane world cosmology with a time like extra dimension@xcite , considering the starobinski model for radiative corrections ( which can not be derived from an effective action)@xcite , exotic matter@xcite or @xmath2 theories in the presence of perfect fluids with @xmath3 @xcite .
in addition to this , the consideration of a jordan brans dicke model also can provide a stable initial state for the emerging universe scenario @xcite . in this paper
we propose a different theoretical framework where such emerging universe scenario is realized in a natural way , where instabilities are avoided and a successful inflationary phase with a graceful exit can be achieved .
we work in the context of a two measures theory ( tmt ) @xcite and more specifically in the context of the scale invariant realization of such theories @xcite .
these theories can provide a new approach to the cosmological constant problem and can be generalized to obtain also a theory with a dynamical spacetime @xcite .
we will show how the stated goals can be achieved in the framework of the tmt models .
the paper will be organized as follows : in section ii the effective field eqs . in the einstein frame ,
give rise to an effective potential for a dilaton field ( needed to implement a model with global scale invariance ) which presents a flat region .
this model can describe an emerging universe scenario . in section iii , we look at the generalization of this model @xcite by adding a curvature square , `` @xmath0 '' , and show that the resulting model contain a flat region .
furthermore the universe can evolve into an inflationary state which can undergo a successful graceful exit .
we end with a discussion and conclusions section .
it is possible to obtain a model that through a spontaneous breaking of scale invariance can give us an emerging universe scenario .
we consider the scale invariant action of the form @xcite
@xmath4 here , @xmath5 and @xmath6 are given by @xmath7 and @xmath8 respectively . also , @xmath9 is defined as @xmath10 where @xmath11 is a 4-scalar fields . the last term in the previous action
is not only globally scale invariant , but also locally scale invariant , that is conformally invariant .
let us see which are the equations of motion .
the variation of the action respect to the 4-scalar fields , @xmath12 , yields to @xmath13 where @xmath14 , and since @xmath15 we get @xmath16 , for @xmath17 , and thus @xmath5 becomes @xmath18 where @xmath19 is an integration constant .
the variation of the action with respect to the metric @xmath20 gives @xmath21 it is interesting to notice that if we contract this equation with @xmath22 , the @xmath23 terms do not contribute . solving the scalar curvature from this and inserting in the other @xmath23 - independent equation @xmath24 we get the solution for the ratio of the measures i.e. @xmath25 .
we see that under a conformal transformation given by @xmath26 the new metric @xmath27 defines the known einstein frame . equations ( [ eq01 ] ) can now be expressed in this frame as @xmath28 where @xmath29 with @xmath30 the constant @xmath31 that appears in this latter expression could be obtained from the original lagrangian , where in it is added a term proportional to @xmath32 , where @xmath33 is the conformal factor .
this renders the action to be @xmath34 , \label{act.lambda}\ ] ] where @xmath35 with @xmath36 given by eq . ( [ eq04 ] ) .
also , eq.([eq00 ] ) expressed in terms of @xmath37 becomes @xmath38 , where we have written @xmath39 in place of @xmath40 .
this allows us to solve for @xmath41 and thus we get @xmath42 note that when we insert eq .
( [ eq05 ] ) into eq .
( [ eq04 ] ) , it depends on the derivatives of the scalar field .
it acts as a normal scalar field potential under the conditions of slow rolling or low gradients and in the case the scalar field is near the region @xmath43 . in the scale invariant case , where @xmath39 and @xmath44 are given by @xmath45 and @xmath46 , respectively@xcite ,
it is interesting to study the shape of @xmath47 as a function of @xmath48 in the case when @xmath48 change a little bit . in this case
, @xmath49 can be regarded as a scalar field potential . then from eq .
( [ eq05 ] ) we get @xmath50 , which inserted into eq .
( [ eq04 ] ) gives , @xmath51 in fig.[fig1 ] we have plotted the effective potential , @xmath36 as a function of the scalar field , @xmath48 for @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath56 and @xmath57 .
we should mention here that the choices of the values of the different parameters was done in virtue that they respect the conditions for a stable emerging universe .
notice that it shows a flat region of positive vacuum energy for large @xmath48 , a minimum obtained at zero without fine tuning . in the limit @xmath58 the effective scalar potential , @xmath47 , becomes @xmath59 .
notice that in all the above discussion it is fundamental that @xmath60 .
if @xmath61 the potential becomes just a flat one , @xmath62 everywhere ( not only at high values of @xmath63 ) .
all the non trivial features necessary for a graceful exit , the other flat region associated to the planck scale and the minimum at zero if @xmath64 are all lost .
the case @xmath65 implies that we are considering a situation with s.s.b . of scale invariance .
we now want to consider the detailed analysis of emerging universe solutions .
we start considering the closed friedmann - robertson - walker metric , @xmath66 where @xmath67 is the scale factor , @xmath68 represents the cosmic time and due to homogeneity and isotropy we take the scalar field @xmath48 to be a function of the cosmological time only , i.e. @xmath69
. we will consider a scenario where the scalar field @xmath48 is moving in the extreme right region , i.e. for asymptotic values of @xmath70 . in this case , the expressions for the energy density @xmath71 and pressure @xmath72 are given by , @xmath73 and @xmath74 where the constants @xmath75 and @xmath76 are given by @xmath77 @xmath78 and @xmath79 respectively .
it is interesting to notice that all terms proportional to @xmath80 behave like `` radiation '' , since @xmath81 and @xmath82 so that @xmath83 is satisfied .
this is also a consistency check , since the @xmath23 terms are associated to a conformally invariant @xmath84 term ( in the first order formalism ) and in the region @xmath85 that symmetry should hold , at least for those contributions .
the @xmath23 terms therefore do not contribute to the trace of the energy momentum tensor in the region @xmath85 .
the equations that determines such static universe @xmath86 , @xmath87 and @xmath88 give rise to a restriction on @xmath89 .
since @xmath88 and it is proportional to @xmath90 , we must require that @xmath91 , which leads to @xmath92 this equation has two roots , the first being given by @xmath93 and the second one is @xmath94 where @xmath95 from these expression we could get some specific range of the parameters in order that either @xmath96 and @xmath97 become positive . in the following section we will restrict these parameters in order to describe an emerging universe .
we will now consider the perturbation equations .
considering small deviations of @xmath98 from the static emerging solution value @xmath89 and also considering the perturbations of the scale factor @xmath99 , we obtain , from eq .
( [ eq06 ] ) , that @xmath100 at the same time @xmath101 can be obtained from the perturbation of the friedmann equation @xmath102 , which gives under a perturbation respect to the background static solution @xmath103 on the other hand , from the second order friedmann equation , @xmath104 , we get that @xmath105 where we have used @xmath106 at @xmath107 , and we have chosen to express our result in terms of @xmath108 , defined by @xmath109 , which for the emerging solution has the value @xmath110 . using this in eq .
( [ pert.fried.eq . ] ) , we obtain @xmath111 substituting this latter expression into eq .
( [ eq.density-pert . ] ) yields to a linear relation between @xmath112 and @xmath113 so that @xmath114 where @xmath115 since we could write that @xmath116 , with @xmath117 where , @xmath118 and therefore , the perturbation of the second order friedmann equation leads to , @xmath119 to evaluate this , we use ( [ omega - eq . ] ) , ( [ v - eq . ] ) and the expressions that relate the variations in @xmath99 and @xmath98 ( [ delta - delta ] ) . defining the `` small '' variable @xmath120 as @xmath121 we obtain , @xmath122 where , @xmath123,\label{eq08}\ ] ] with @xmath124 .
stability of the static solution requires that @xmath125 .
this static solution has to respect the conditions @xmath126 and @xmath127 .
only the second solution , i.e. @xmath128 ( see eq .
( [ eq09 ] ) ) , fulfills these requirements , providing that the parameters satisfy the following inequalities @xmath129 here , we have taken @xmath130 . versus the scalar field @xmath48 .
we have used @xmath52 , @xmath53 , @xmath131 , @xmath55 , @xmath56 and @xmath57.,width=480 ]
the study of the stability of the static solution and the properties of the different equilibrium points could be done in a more systematic way by using a dynamical system approach . in this scheme
we rewrite the friedmann and the conservation of energy equations as an autonomous system in terms of the variables @xmath132 and @xmath133 . in order to do so , we differentiating the fridmann equation and after using the expressions for @xmath71 and @xmath72 , eq .
( 16 , 17 ) , and the energy conservation equation we obtain : @xmath134 we can rewrite eq .
( [ dinamic0 ] ) and the energy conservation equation as the following two equations .
@xmath135 - h^2 , \label{dinamic1}\\ \nonumber \\
\dot{x } & = & - \frac{3a\,(1 + \kappa^2\,\epsilon \,x)\,x}{\frac{a}{2 } + a\,\kappa^2\,\epsilon \,x + 2b\,x}\,h \,\ , , \label{dinamic2}\end{aligned}\ ] ] where we have eliminated @xmath99 from eq .
( [ dinamic0 ] ) by using the friedmann equation and we have used the definition of the variable @xmath136 in order to get eq .
( [ dinamic2 ] ) from the energy conservation equation .
the equations ( [ dinamic1 ] ) and ( [ dinamic2 ] ) are a two - dimensional autonomous system on the variables @xmath132 and @xmath136 . in order to study the stability of the static solutions we look for critical points of the system ( [ dinamic1 ] ) and ( [ dinamic2 ] ) .
these points are @xmath137 the critical points have different properties depending on the values of the parameters of the model ( @xmath138 , @xmath31 ) . at this moment
we are not going to give an exhaustive description of these properties for all the critical points , instead , we are going to focus on the particular critical points which are related with static universe . from the definition of the variables
@xmath132 and @xmath136 we can note that only the first two critical points eqs .
( [ crit1 ] , [ crit2 ] ) correspond to a static universe . in order to study the nature of these two critical points we linearize the equations ( [ dinamic1 ] ) and ( [ dinamic2 ] ) near these critical points . from the study of the eigenvalues of the system we found that the first critical point , eq .
( [ crit1 ] ) , could be a center or a saddle point , depending on the values of the parameters of the model . on the other hand the second critical point eq .
( [ crit2 ] ) is a saddle .
stable static solutions correspond to a center .
then , by impose the critical point , eq.([crit1 ] ) , becomes a center , we recover the stability condition eqs.([cc1 ] ) , ( [ cc2 ] ) and ( [ cc3 ] ) . in order to satisfy the requirements of stability we are going take the values @xmath139 , @xmath54 , @xmath140 and @xmath141 . in fig .
[ fase1 ] it is shown a phase portrait for four numerical solution to eqs .
( [ dinamic1 ] ) and ( [ dinamic2 ] ) .
also , in this figure we have included the _ direction field _ of the system in order to have a picture of what a general solution look like . in this figure
are four of the six critical points described in eqs .
( [ crit1])-([crit3 ] ) .
one of this point is the center equilibrium point ( @xmath142 ) , the saddle point ( @xmath143 ) , the point ( @xmath144 is a future attractor and ( @xmath145 is a past attractor . the other equilibrium points are @xmath146 and @xmath147 which are a future attractor point and a past attractor point respectively .
as we have mentioned , only critical points with @xmath148 represent a static universe .
then in our case only the critical point which is a center correspond to a solution which represent a static and stable universe . in fig .
[ fase2 ] we show the _ direction field _ near this critical point ( @xmath142 ) together with two numerical solution .
the emerging phase owes its existence to a strictly constant vacuum energy ( which here is represented by the value of @xmath149 ) at very large values of the field @xmath48 . in fact , while for @xmath150 the effective potential of the scalar field is perfectly flat , for any @xmath151 the effective potential acquires a non trivial shape .
we consider then the relevant equations for the model in the slow roll regime , i.e. for @xmath98 small and when the scalar field @xmath48 is large , but finite and we consider the first corrections to the flatness to the effective potential . dropping higher powers of @xmath98 in the contributions for the kinetic energy and in the scalar curvature @xmath41 , we obtain @xmath152 here , as usual @xmath153 . in the slow roll approximation
, we can drop the second derivative term of @xmath48 and the second power of @xmath98 in the equation for @xmath154 and we get @xmath155 where @xmath156 . the relevant expression for @xmath36
will be that given by ( [ effpotslow ] ) , i.e. , where all higher derivatives are ignored in the potential , consistent with the slow roll approximation .
we now display the relevant expressions for the region of very large , but not infinite @xmath48 , these are : @xmath157 and @xmath158 the relevant constants that will affect our results are , @xmath76 , as given by ( [ c ] ) and @xmath159 and @xmath160 given by @xmath161 and @xmath162 respectively .
( [ effpotatlargephi ] ) we can calculates the key landmarks of the inflationary history : first , the value of the scalar field where inflation ends , @xmath163 and a value for the scalar field @xmath164 bigger than this ( @xmath165 ) and which happens earlier , which represents the `` horizon crossing point '' .
we must demand then that a typical number of e - foldings , like @xmath166 , takes place between @xmath164 , until the end of inflation at @xmath167 . to determine the end of inflation
, we consider the quantity @xmath168 and consider the point in the evolution of the universe where @xmath169 , only when @xmath170 , we have an accelerating universe , so the point @xmath169 represents indeed the end of inflation . calculating the derivative with respect to cosmic time of the hubble expansion using ( [ friedmannslow ] ) and ( [ slowroll ] ) , we obtain that the condition @xmath169 gives @xmath171 working to leading order , setting @xmath172 , @xmath173 and @xmath174 , this gives as a solution , @xmath175 notice that if @xmath19 and @xmath176 have different signs and if @xmath177 , @xmath178 for the allowed range of parameters the stable emerging solution , so @xmath179 represents the absolute value of @xmath180 .
we now consider @xmath164 and the requirement that this precedes @xmath163 by @xmath181 e - foldings , @xmath182 where in the last step we have used the slow roll equation of motion for the scalar field ( [ slowroll ] ) to solve for @xmath98 . solving @xmath154 in terms of @xmath36 using ( [ friedmannslow ] ) , working to leading order , setting @xmath183 and integrating ,
we obtain the relation between @xmath164 and @xmath163 , @xmath184 as we mentioned before @xmath178 for the allowed range of parameters the stable emerging solution , so that @xmath165 as it should be for everything to make sense . introducing eq .
( [ sol.endofinf.eq . ] ) into eq . (
[ sol.crossing . ] ) , we obtain , @xmath185 we finally calculate the power of the primordial scalar perturbations .
if the scalar field @xmath48 had a canonically normalized kinetic term , the spectrum of the primordial perturbations will be given by the equation @xmath186 however , as we can see from ( [ energydensityslow ] ) , the kinetic term is not canonically normalized because of the factor @xmath187 in that equation . in this point we will study the scalar and tensor perturbations for our model where the kinetic term is not canonically normalized .
the general expression for the perturbed metric about the friedmann - robertson - walker is @xmath188 dx^i dx^j,\end{aligned}\ ] ] where @xmath189 , @xmath190 , @xmath191 and @xmath192 are the scalar type metric perturbations and @xmath193 characterizes the transverse - traceless tensor perturbation .
the power spectrum of the curvature perturbation in the slow - roll approximation for a not - canonically kinetic term becomes ref.@xcite(see also refs.@xcite ) @xmath194 where it was defined `` speed of sound '' , @xmath195 , as @xmath196 with @xmath197 an function of the scalar field and of the kinetic term @xmath198 . here
@xmath199 denote the derivative with respect @xmath200 . in our case @xmath201 , with @xmath202 .
thus , from eq.([pec ] ) we get @xmath203 the scalar spectral index @xmath204 , is defined by @xmath205 where @xmath206 and @xmath207 , respectively . on the other hand ,
the generation of tensor perturbations during inflation would produce gravitational wave .
the amplitude of tensor perturbations was evaluated in ref.@xcite , where @xmath208 and the tensor spectral index @xmath209 , becomes @xmath210 and they satisfy a generalized consistency relation @xmath211 versus @xmath204 for three values of @xmath212 . for @xmath213 solid line , @xmath214 dash line and @xmath215 dots line , respectively . here
, we have fixed the values @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath56 and @xmath57 , respectively .
the seven - year wmap data places stronger limits on the tensor - scalar ratio ( shown in red ) than five - year data ( blue ) @xcite.[fig2],width=480 ] therefore , the scalar field ( to leading order ) that should appear in eq .
( [ primordialpert . ] ) should be @xmath216 and instead of eq .
( [ eq10 ] ) , we must use @xmath217 the power spectrum of the perturbations goes , up to a factor of order one , which we will denote @xmath218 as @xmath219 , so we have , @xmath220 this quantity should be evaluated at @xmath221 given by ( [ sol.crossing.final ] ) . solving for @xmath98 from the slow roll equation ( [ slowroll ] ) , evaluating the derivative of the effective potential using ( [ effpotatlargephi ] ) and solving for @xmath132 from ( [ friedmannslow ] ) , we obtain , to leading order , @xmath222 using then ( [ sol.crossing.final ] ) for @xmath223 , we obtain our final result , @xmath224 it is very interesting first of all that @xmath180 dependence has dropped out and with it all dependence on @xmath19 . in fact
this can be regarded as a non trivial consistency check of our estimates , since apart from its sign , the value @xmath19 should not affect the results .
this is due to the fact that from a different value of @xmath19 ( although with the same sign ) , we can recover the original potential by performing a shift of the scalar field @xmath48 . in fig.[fig2 ]
we show the dependence of the tensor - scalar ratio @xmath225 on the spectral index @xmath204 . from left to right @xmath213 ( solid line ) , @xmath214 ( dash line ) and @xmath215 ( dots line ) , respectively . from ref.@xcite , two - dimensional marginalized constraints ( 68@xmath226 and 95@xmath226 confidence levels ) on inflationary parameters @xmath225 and @xmath204 , the spectral index of fluctuations , defined at @xmath227 = 0.002 mpc@xmath228 .
the seven - year wmap data places stronger limits on @xmath225 ( shown in red ) than five - year data ( blue)@xcite . in order to write down values that relate @xmath204 and @xmath225 , we used eqs.([ns ] ) and ( [ ration ] )
. also we have used the values @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath56 and @xmath57 , respectively . from eqs.([n ] ) , ( [ ns ] ) and ( [ ration ] ) , we observed numerically that for @xmath229 , the curve @xmath230 ( see fig.[fig2 ] ) for wmap 7-years enters the 95@xmath226 confidence region where the ratio @xmath231 , which corresponds to the number of e - folds , @xmath232 . for @xmath214 , @xmath233 corresponds to @xmath234 and for @xmath215 , @xmath235 corresponds to @xmath236 . from 68@xmath226 confidence region for @xmath229 , @xmath237 , which corresponds to @xmath238 . for @xmath239
, @xmath240 corresponds to @xmath241 and for @xmath215 , @xmath242 corresponds to @xmath243 .
we noted that the parameter @xmath212 , which lies in the range @xmath244 , the model is well supported by the data as could be seen from fig.[fig2 ] .
the consideration of a scale invariant action leads to exactly the type of scalar field potentials needed to obtain an emerging universe scenario .
the addition of a @xmath0-term gives rise to solutions where @xmath245 , @xmath246 and @xmath247 and to the stability of the initial einstein universe solution .
we find that this is accomplished because the @xmath248 term is shown to have the effect of introducing nonlinearities in the theory that stabilize the solutions .
in addition in order for the whole idea to work we need to take @xmath151 , which means we are considering the s.s.b .
of the theory and this is what gives the effective potential a non trivial shape , which allows a graceful exit from the inflationary phase and can give the adequate power of the density perturbations .
it should be pointed out also that the @xmath0 theory studied here , in the context of an action that contains a measure @xmath9 independent of the metric and with the use of the first order formalism ( the connections are considered as independent degrees of freedom in the action principle ) , leads to equations of motion that are only second order , i.e. only second derivatives of the metric and the matter fields appear ( although higher powers of the derivatives of the scalar field do appear ) .
this after the new measure @xmath9 and the connections are solved ( through the equations of motion ) in terms of the metric and matter fields and after we express our results in the einstein frame .
this is in contrast to the usual @xmath0 theories in the second order formalism and with standard measure everywhere in the action . in this case , these theories lead to fourth order equations for the metric field . in ref.@xcite
it was shown that the fourth order structure of the equations can be reformulated as a system of second order equations which contain however one additional degree of freedom , a scalar field . for studies of inflation in the usual @xmath0 theories in the second order formalism ,
see ref.@xcite . in the case of the usual @xmath0 theories
this scalar field contains a potential with a flat region , as shown in fig.[fig1 ] . using the wmap seven - year data , we have found some constraints for the parameters appearing in our model .
in particular , fig.[fig2 ] shows that for the values of the parameter @xmath249 , the model is well supported by the wmap data .
this work was supported by comision nacional de ciencias y tecnologa through fondecyt grants 1070306 ( sdc ) , 1090613 ( rh and sdc ) . also it was supported by pontificia universidad catolica de valparaiso through grants 123.787 - 2007 ( sdc ) and 123 . 703 - 2009 ( rh ) .
one of us ( e.i.g ) would like to thank the astrophysics and cosmology group at the pontificia universidad catolica de valparaiso for hospitality , p. l. is supported by fondecyt grant n@xmath250 11090410 and by direccin de investigacin de la universidad del bo - bo through grant n@xmath250 096207 1/r .
guth , _ phys .
rev . _ * d23 * , 347 ( 1981 ) ; a.d .
linde , _ phys .
lett . _ * 108b * , 389 ( 1982 ) ; d. kazanas , _ astrophys.j . _ * 241 * , 1980 ( l59 ) ; a. albrecht and p.j .
steinhardt , phys .
lett , 48 , ( 1982 ) 1220 .
a.borde and a.vilenkin , _ phys .
lett . _ * 72 * , 3305 ( 1994 ) ; a.borde and a.vilenkin , _ phys .
rev . _ * d56 * , 717 ( 1997 ) ; a.h .
guth , arxiv : astro - ph/0101507 ; a.borde , a.h .
guth and a.vilenkin,_phys .
lett . _ * 90 * , 151301 ( 2003 ) ; a.vilenkin,arxiv:gr-qc/0204061 .
ellis and r. maartens , _ class .
quantum grav . _
* 21 * , 223 ( 2004 ) g.f.r .
ellis , r. maartens and c.g .
tsagas , _ class .
quantum grav . _
* 21 * , 233 ( 2004 ) d.j .
mulryne , r. tavakol , j.e .
lidsey and g.f.r .
ellis , _ phys .
rev . _ * d71 * , 123512 ( 2005 ) a. banerjee , t. bandyopadhyay and s. chaakraborty , grav .
cosmol .
* 13 * , 290 ( 2007 ) ; j.e .
lidsey and d.j .
rev . _ * d73 * , 083508 ( 2006 ) .
s. mukherjee , b.c.paul , s.d .
maharaj and a. beesham , arxiv : qr - qc/0505103 .
s. mukherjee , b.c.paul , n.k .
dadhich , s.d .
maharaj and a. beesham , _ class .
quantum grav . _ * 23 * , 6927 ( 2006 ) . c. g. boehmer , l. hollenstein , f. s. n. lobo , _ phys .
rev . _ * d76 * , 084005 ( 2007 ) , arxiv:0706.1663 ( gr - qc ) . s. del campo , r. herrera and p. labrana , jcap * 0907 * , 006 ( 2009 ) ; s. del campo , r. herrera and p. labrana , jcap * 0711 * , 030 ( 2007 ) .
basic idea is developed in e.i .
guendelman and a.b .
kaganovich , _ phys .
_ * d60 * , 065004 ( 1999 ) ; f. gronwald , u. muench , a. macias , f. w. hehl , _ phys . rev . _ * d58 * , 084021 ( 1998 ) , e - print : gr - qc/9712063 .
guendelman , a.b .
kaganovich , _ class .
quantum grav . _ * 25 * , 235015 ( 2008 ) , e - print : arxiv:0804.1278 [ gr - qc ] ; see aso related work by h. nishino , s. rajpoot , _ mod .
lett . _ * a21 * , 127 ( 2006 ) , e - print : hep - th/0404088 and by d. comelli , _ int .
* a23 * , 4133 ( 2008 ) , e - print : arxiv:0704.1802 [ gr - qc ] . for a recent review and further references
guendelman , a.b .
kaganovich , plenary talk given at the workshop on geometry , topology , qft and cosmology , paris , france , 28 - 30 may 2008 .
e - print : arxiv:0811.0793 [ gr - qc ] .
some previous works on this subject are : e.i .
guendelman and a.b .
kaganovich , _ phys .
rev . _ * d53 * , 7020 ( 1996 ) ; e.i .
guendelman and a.b .
kaganovich , proceedings of the third alexander friedmann international seminar on gravitation and cosmology , ed .
n. gneding , a.a .
grib and v.m .
mostepanenko ( friedmann laboratory publishing , st .
petersburg , 1995 ) ; ; e.i .
guendelman and a.b .
kaganovich , _ phys .
rev . _ * d55 * , 5970 ( 1997 ) ; e.i .
guendelman and a.b .
kaganovich , _ mod .
lett . _ * a12 * , 2421 ( 1997 ) ; e.i .
guendelman and a.b .
kaganovich , _ phys .
* d56 * , 3548 ( 1997 ) e.i .
guendelman and a.b .
kaganovich , _ hadron.journ . _
* 21 * , 1998 ( 19 ) ; e.i .
guendelman and a.b .
kaganovich , _ mod .
lett . _ * a13 * , 1583 ( 1998 ) ; f. gronwald , u. muench and f.w .
_ * 21 * , 1998 ( 3 ) ; e.i .
guendelman and a.b .
kaganovich , _ phys .
rev . _ * d57 * , 7200 ( 1998 ) ; e.i .
guendelman and a.b .
kaganovich , `` gravity cosmology and particle field dynamics without the cosmological constant problem '' , in the proceedings of the sixth international symposium on particle , strings and cosmology , pascos-98 , world scientific , singapore , 1999 ; e.i .
guendelman and a.b .
kaganovich , `` field theory models without the cosmological constant problem '' , plenary talk ( given by e.i .
guendelman ) on the fourth alexander friedmann international seminar on gravitation and cosmology , gr - qc/9809052 ; e.i .
guendelman , _ int .
phys . _ * a14 * , 3497 ( 1999 ) .
guendelman , _ mod .
* a14 * , 1043 ( 1999 ) , e - print : gr - qc/9901017 .
guendelman , _ mod .
lett . _ * a14 * , 1397 ( 1999 ) ; e.i .
guendelman , _ class .
quantum grav . _
* 17 * , 361 ( 2000 ) ; e.i.guendelman , `` scale invariance , mass and cosmology '' gr - qc/9901067 ; e.i.guendelman , `` scale invariance and the present vacuum energy of the universe '' , contribution to the 35-th rencontres de moriond , gr - qc/0004011 ; e.i.guendelman , `` measure fields , the cosmological constant and scale invariance '' contribution to the 30-th ichep conference , hep - th/0008122 . e.i .
guendelman , a.b .
kaganovich , _ int
_ * a17 * , 417 ( 2002),e - print : hep - th/0110040 ; e.i .
guendelman and a.b.kaganovich , _ mod .
* a17 * , 1227 ( 2002 ) ; e.i .
guendelman , a.b .
kaganovich , _ ann . of phys . _ * 323 * , 866 ( 2008 ) , e - print : arxiv:0704.1998 [ gr - qc ] ; e.i .
guendelman , a.b .
kaganovich , _ phys .
_ * d75 * , 083505 ( 2007 ) , e - print : gr - qc/0607111 ; e.i .
guendelman , a.b .
kaganovich , _ int . j. mod
. phys . _ * a21 * , 4373 ( 2006 ) .
guendelman and o. katz , _ class .
quantum grav . _ * 20 * , 1715 ( 2003 ) , e - print : gr - qc/0211095 .
j.d.barrow and s. cotsakis , _ phys .
lett . _ * 214b * , 515 ( 1988 ) ; j.d.barrow and a.c.otterwill , _ j. physics _ * a16 * , 2757 ( 1983 ) ; p.w.higgs , nuovo cimento * 11 * , 816 ( 1959 ) ; b.witt , _ phys .
lett . _ * 145b * , 176 ( 1984 ) . if and arbitrary function of the scalar curvature appears see g.mangano , m.ferraris and m.francaviglia , _ gen
* 19 * , 1987 ( 465 ) and m.ferraris , m. francaviglia and g. mangano _ class .
quantum grav . _ * 5 * , l95 ( 1988 ) .
m.b.mijic , m.m.morris and w - m suen , _ phys .
rev . _ * d34 * , 2934 ( 1986 ) .
the case when there is an additional scalar in the theory has been studied by c.contreras , r.herrera and s. del campo , _ class .
quantum grav . _
* 12 * , 1937 ( 1995 ) ; c. contreras , r. herrera and s. del campo , phys . rev .
d * 52 * , 4349 ( 1995 ) . the first inflationary model from a model with higher terms in the curvature was proposed in a.a .
starobinsky,_phys . lett .
_ * 91b * , 99 ( 1980 ) . | we consider a scale invariant model which includes a @xmath0 term in action and show that a stable `` emerging universe '' scenario is possible .
the model belongs to the general class of theories , where an integration measure independent of the metric is introduced . to implement scale invariance ( s.i . )
, a dilaton field is introduced .
the integration of the equations of motion associated with the new measure gives rise to the spontaneous symmetry breaking ( s.s.b ) of s.i .
after s.s.b . of s.i . in the model with the @xmath0 term ( and first order formalism applied ) , it is found that a non trivial potential for the dilaton is generated .
the dynamics of the scalar field becomes non linear and these non linearities are instrumental in the stability of some of the emerging universe solutions , which exists for a parameter range of the theory . |
what can be said about functions wth milnor fibre @xmath1 has the property @xmath2 for all @xmath3 ?
if @xmath1 is connected then @xmath4 is non - singular and equivalent to a linear function by acampo s trace formula .
the remaining question : _ what happens if @xmath1 is non - connected ? _ is only relevant for non - reduced plane curve singularities .
this question is related to a recent preprint @xcite . that preprint contains a statement about the so - called bobadilla conjecture in case of plane curves .
the invariant @xmath5 , used by massey @xcite should imply that the singular set of of @xmath4 is a smooth line . in this short note
we give a short topological proof of a stronger statement .
[ p : main ] if the first betti number of the milnor fibre of a plane curve singularity is zero , then the defining function is equivalent to @xmath0 .
[ c : line ] in the above case the singular set is a smooth line and the system of transversal singularities is trivial .
non - isolated plane curve isolated singularities have been thoroughly studied by rob schrauwen in his dissertation @xcite .
main parts of it are published as @xcite and @xcite .
the above proposition [ p : main ] is an easy consequence of his work .
we can assume that @xmath6 ( partition in powers of reduced irreducible components ) .
( folklore ) let @xmath7 * @xmath1 has @xmath8 components , each diffeomorphic to the milnor fibre @xmath9 of @xmath10 .
the milnor monodromy of @xmath4 permutes these components , * if @xmath11 then @xmath1 is connected .
( a. ) since @xmath12 the fibre @xmath1 consists of @xmath8 copies of @xmath9 .
+ ( b. ) we recall here the reasoning from @xcite .
deform the reduced factors @xmath13 into @xmath14 such that the product @xmath15 contains the maximal number of double points .
this is called a network deformation by schrauwen .
the corresponding deformation @xmath16 of @xmath4 near such a point has local equation are of the form @xmath17 ( point of type @xmath18 $ ] ) .
near every branch @xmath19 the milnor fibre is a @xmath20-sheeted covering of the zero - locus , except in the @xmath18$]-points .
we construct the milnor fibre @xmath1 of @xmath4 starting with @xmath21 copies of the affine line @xmath22 .
cover the @xmath23 branch with @xmath20 copies of @xmath22 and delete @xmath24 small discs around the @xmath18$]-points .
glue in the holes @xmath25 small annuli ( the milnor fibres of @xmath18 $ ] ) .
the resulting space is the milnor fibre @xmath1 of @xmath4 .
take a reference transversal fibre @xmath26 , which consits of @xmath27 cyclic ordered points .
as soon as @xmath28 intersects @xmath29 it connects the sheets of @xmath28 modulo @xmath30 . since @xmath31
we connect all sheets .
( of proposition ) .
if @xmath32 , then also @xmath33 .
the milnor monodromy has trace@xmath34 = 1 .
according to acampo s observation @xcite @xmath35 is regular : @xmath36 .
it follows that @xmath37 .
we consider first in any dimension @xmath38 with a 1-dimensional singular set , see especially the 1991-paper @xcite for definitions , notations and statements .
we focus on the group @xmath39 which occurs in two exact sequences on p. 468 of @xcite : here @xmath45 is the disjoint union of the transversal milnor fibres @xmath46 , one for each irreducible branch of the 1-dimensional singular set . was originally denoted by @xmath47 . in the second second sequence a misprint @xmath48 in the third term has been changed to @xmath49 . ]
the second part of @xcite contains an elegant classification of @xmath62 with the help of the acampo trace formula .
also the reduction of the generalized bobadilla conjecture to the ( ireducible ) bobadilla conjecture . as final remark : the great work ( the irreducible case ) has still has to be done ! together with the l - conjecture this seems to be an important question in the theory of hypersurfaces 1-dimensional singular sets . | if the first betti number of the milnor fibre of a plane curve singularity is zero , then the defining function is equivalent to @xmath0 . |
diverse fields of research including neurophysiology , sociology , geophysics , and economics considers heavy tails and scale - free distributions as the signature of complexity @xcite .
the brain as a complex dynamic system and brain activities including neuronal membrane potentials , noninvasive electroencephalography ( eeg ) , magnetoencephalography ( meg ) , and functional magnetic resonance imaging ( fmri ) signals observed at many spatiotemporal patterns exhibit power - law behavior @xcite . synchronized large amplitude bursts or neural avalanches , are power - law distributed @xcite and the widely varying profile of neural avalanche distribution in size and duration are described by a single universal scaling exponent , @xmath0 , in @xmath1 .
power law distribution may suggest that neural networks operate near a non - equilibrium critical point where the phase transition occurs .
however , power laws are not alone sufficient to establish criticality and also non - critical systems may produce power laws @xcite ; implying that many possible mechanisms generate power law distribution @xcite . in @xcite
, authors argue that the power laws observed in avalanche size distributions are actually artifacts , and in an _ in vivo _
study @xcite , no clear evidence of power - law scaling or self - organized critical states was found in the awake and sleeping brain of mammals in monkey , cat and human . according to the popular view of per bak @xcite criticality
is realized spontaneously by complex systems rather than requiring the fine tuning of a control parameter with a critical value .
they proposed the term self - organized criticality ( soc ) , as an operation of the system at criticality that generates the power law behavior in natural phenomena .
critical point sets a boundary between an ordered and a less ordered state with different scaling behaviors @xcite .
evidence for criticality in the brain has been found _ in vitro _
experiments @xcite , animals @xcite , and humans @xcite . in our previous study @xcite , using leaky integrate - and fire model ( lifm ) , we have shown that _ temporal complexity _ is a robust indicator of criticality with which a critical point is introduced .
we found that temporal complexity uncontaminated by periodicity occurs in a narrow range of values of the control parameter , in correspondence of which the information transfer from one network to another becomes maximal .
we argued that if this enhancement of information transport is interpreted as a signature of criticality , then the power law avalanches are a manifestation of super - criticality rather than criticality .
this led us to conclude that if power - law distribution in avalanches is not found in the brain of mammals , that would not conflict with the hypothesis that the brain works at criticality .
so how temporal complexity is connected to power - laws ? if we consider an event in a system , then the waiting time distribution densities , @xmath2 , have the time - asymptotic form @xmath3 it was proposed that the cumulative distribution function ( cdf ) is much more accurate to fit the power - law exponents , as well as to identify if the system obeys a power - law @xcite .
hence the corresponding cumulative distributions , or survival probabilities , is defined as @xmath4 thus the corresponding cdf also behaves as a power - law , but with a smaller exponent @xmath5 . in this work ,
we show that by replacing the discrete noise with a continuous gaussian noise in the model , avalanche probability distributions display power law behavior at criticality indicated by temporal complexity and phase transition .
the model used in the previous work and here is the popular leaky integrate - and fire model @xcite .
@xmath6 where @xmath7 is the membrane potential , @xmath8 is the membrane time constant of the neuron , @xmath9 is proportional to a constant input current , @xmath10 is the control parameter and @xmath11 describes the adjacency matrix .
note that @xmath12 where @xmath13 is the total number of neurons , @xmath14 is the time of firing of neuron @xmath15 , and @xmath16 is the time lag of firing neuron @xmath15 to the time of receiving the cooperation to the neuron @xmath17 .
each neuron starts from a random value or zero and fires when it reaches the threshold @xmath18 . when a neuron fires , it forces all the neurons linked to it to make a step ahead or backward by the quantity @xmath10 according to whether @xmath19 ( excitatory ) or @xmath20 ( inhibitory ) .
the parameter @xmath10 plays the all - important role of control parameter and is expected to generate criticality when the special value @xmath21 , to be determined using a suitable criticality indicator , is adopted . after firing , each neuron jumps back to the rest state @xmath22 .
when @xmath23 , the vanishing noise condition yields the following expression for the time distance between two consecutive firings of the same neuron @xmath24 if @xmath25 , after a few time steps all the neurons fire at the same time @xcite , thereby generating a sequence of quakes of intensity @xmath13 with the time period @xmath26 given by equation .
the parameter @xmath27 is the standard deviation of the noise which can also be considered as the noise intensity . in the previous solution of equation
, we adopted the integration time step @xmath28 and @xmath29 as a discontinuous random fluctuation taking with equal probability either the value of @xmath30 or @xmath31 @xcite . here , however , we treat the time continuously and then consider @xmath32 to be a continuous gaussian white noise with zero mean and unit variance , defined by @xmath33 to reduce the number of the parameters , we define the dimensionless time variable @xmath34 . in this way , setting the last term aside , equation can be rewritten in the following form @xmath35 in which @xmath36 is the dimensionless gaussian noise with zero mean and unit variance . to numerically integrate the stochastic differential equation
, we use the ito s interpretaion , which is @xmath37d{\cal t}+ \frac{\sigma}{\sqrt\gamma } \eta_i ( { \cal t } ) \sqrt{d{\cal t}}.\ ] ] here , we choose @xmath38 , @xmath39 , @xmath40 , and @xmath41 .
we assume that the neurons are residing on the nodes of a two - dimensional square lattice with periodic boundary condition with the size @xmath42 , where @xmath43 is the linear size of the lattice .
the adoption of periodic boundary conditions is to ensure the total equivalence of the cooperating units , so as to avoid the doubt that the onset of firing bursts may be triggered by units with a favorable topology .
however , numerical calculations not reported here show that the adoption of periodic boundary condition is not crucial for the results of this paper . here
, we show that temporal complexity detects a critical point where it is poised between a random and a regular state .
we then examine if avalanche distribution in size and time follow a power - law exponent as found in @xcite .
using scaling theory that expresses a relation between the exponents , we confirm that neural avalanche data collapses into critical exponent suggested by temporal complexity approach .
versus the dimensionless time @xmath34 , for three different cooperation parameters @xmath44 with @xmath45 , @xmath46 , @xmath39 in a @xmath47 lattice .
( b ) variation of the power index @xmath0 , and ( c ) scale of stretched exponential , @xmath48 , defined by eq . , with the cooperation parameter .
notice the abrupt change in @xmath49 .
, title="fig:",height=207 ] versus the dimensionless time @xmath34 , for three different cooperation parameters @xmath44 with @xmath45 , @xmath46 , @xmath39 in a @xmath47 lattice .
( b ) variation of the power index @xmath0 , and ( c ) scale of stretched exponential , @xmath48 , defined by eq .
, with the cooperation parameter .
notice the abrupt change in @xmath49 .
, title="fig:",height=207 ] versus the dimensionless time @xmath34 , for three different cooperation parameters @xmath44 with @xmath45 , @xmath46 , @xmath39 in a @xmath47 lattice .
( b ) variation of the power index @xmath0 , and ( c ) scale of stretched exponential , @xmath48 , defined by eq . , with the cooperation parameter .
notice the abrupt change in @xmath49 .
, title="fig:",height=207 ] in this section , we explore the temporal complexity of the lifm , given by , by calculating the cumulative survival probability , @xmath50 , which indicates the probability that no firing occurs up to the time @xmath51 from an earlier firing . in the lack of cooperation ,
the dynamics is poissonian process defined by an exponential @xmath52 according to @xcite , cooperation generates scale invariance , hence , increasing @xmath10 leads to a transition from the exponential form of equation to @xmath53 at criticality , with @xmath54 so small as to make equation virtually equivalent to the inverse power law of equation over the available time scale .
the time dependence of @xmath50 for different values of cooperation parameter @xmath10 is plotted in fig.[fig : f1].(a ) for a lattice of linear size @xmath55 . as shown by fig.[fig : f1].(a ) , we find instead that for small values of @xmath13 , the survival probability is identical to the mittag - leffler ( ml ) function @xcite @xmath56 ml function settles a bridge between the stretched exponential and inverse power law survival probabilities as important signs of complexity .
in fact , in the time region @xmath57 the ml survival probability is described by the stretched exponential function @xmath58 and in the large time region @xmath59 by a power - law of eq . the index @xmath0 of the stretched exponential function of equation is identical to the power index @xmath0 of the inverse power law of equation .
we show that the ml function of equation is the visible manifestation of a hidden survival probability with the inverse power law form of equation , which is thought to be a signature of complexity @xcite . , versus noise intensity
, @xmath27 , for @xmath46 , @xmath39 and @xmath60.,height=207 ] fig.[fig : f1].(a ) shows how neural network deviates from exponential to a regular behavior with changing control parameters , @xmath10 . in order to find the critical point
, we fit the survival probability functions with ml - function .
the corresponding fractal exponents of each curve , @xmath0 , and @xmath48 are plotted in figs .
[ fig : f1].(b ) and ( c ) , respectively . as can be seen from the figure
, a region represents a phase transitionary behavior that indicate criticality around a critical point of @xmath49 .
the model of the current paper is a generalization of the model proposed by mirollo and strogatz @xcite in which in the absence of noise , full synchronization is achieved after few steps in the absence of noise .
however , here , adding a noise to the model in order to generate temporal complexity along with an adjustable coupling parameter , @xmath10 , a competition is set between these two parameters .
we have explored how coupling has to be adjusted to the noise to maintain the criticality and phase transition .
hence , we run the model for different values of @xmath27 and arrive at fig . [ fig : f2 ] . as can be inferred from this figure , the critical coupling @xmath21 , approximately has a linear dependence on the noise intensity , and as we increase the noise in the system , cooperation has to level up in order to maintain the system at criticality . in the previous work
, we confirmed the criticality by aging experiment and information transfer . here , however , we encourage readers to see @xcite for further reading , and focus on neural avalanches . hence , we rely on previous results and take @xmath61 as indicated by fig.[fig : f2 ] ( b and c ) , and ask if avalanche data including avalanche size , duration and temporal profile fall within soc predictions according to @xcite .
, ( b ) avalanche duration @xmath62 .
( c ) scaling of the conditional average of avalanches with duration @xmath63 .
the results are obtained using @xmath64 at @xmath49 and the lattices with the linear size @xmath65 .
, title="fig:",height=207 ] , ( b ) avalanche duration @xmath62 .
( c ) scaling of the conditional average of avalanches with duration @xmath63 .
the results are obtained using @xmath64 at @xmath49 and the lattices with the linear size @xmath65 .
, title="fig:",height=207 ] , ( b ) avalanche duration @xmath62 .
( c ) scaling of the conditional average of avalanches with duration @xmath63 .
the results are obtained using @xmath64 at @xmath49 and the lattices with the linear size @xmath65 .
, title="fig : " ] in our previous study , we did not observed the coincidence of power - law distribution of neural avalanches at criticality indicated by temporal complexity .
this finding may cast doubt on the operation of the brain near criticality ; however , we hypothesized that temporal complexity as a promising detector of criticality may suffice . yet , considering many inconsistent results reported from recordings in awake animals and humans , we did not cease our attempt to find the source of presumptive conflict . in this section , we calculate the size and duration of the neuronal avalanches creating as the results of our lifm . for this purpose , we count the number of firing neurons in time bins of @xmath66 simulation steps ( @xmath67 ) .
an avalanche is recorded whenever a burst of neurons is followed by quiescent duration of minimum @xmath66 steps .
the number of neurons fired during this active region is called the avalanche size , @xmath9 , and the duration of this activity is called the avalanche duration @xmath63 . as predicted by renormalization group theory @xcite ,
avalanche data collapses onto universal scaling functions near a critical point and follow : @xmath68 where @xmath69 and @xmath70 are the probability density function of the avalanche size and duration , respectively .
@xmath71 is the average of avalanche size conditioned on a given duration @xcite .
@xmath72 , @xmath73 and @xmath74 are the critical exponents of the system and are independent of the details of the model or system @xcite .
the scaling theory requires the following relation between the exponents @xmath75 mean field theory predicts @xmath76 and @xmath77 based on soc . the results of our simulation for three lattice sizes @xmath78 , @xmath79 and @xmath80 at @xmath61 are illustrated in fig.[fig : f3 ] . in this figure , panels ( a ) and ( b ) show the scaling of the cumulative probabilities of the avalanche size ( @xmath81 ) and avalanche duration ( @xmath82 ) , respectively , and panel ( c ) is devoted to the scaling of conditional average avalanche size , @xmath83 in terms of the duration @xmath63 . according to equation , @xmath84 and @xmath62 scales as @xmath85 for the critical point suggested by temporal complexity ( @xmath61 for the noise intensity @xmath41 ) ,
we found the following exponents : @xmath86 , @xmath87 , and @xmath88 which satisfy the scaling relation , equation .
these exponents are in good agreement with the experimental values obtained recently @xcite .
this is another proof that the critical point suggested by temporal complexity is in fact the system s critical point .
it is notable that we have considered @xmath41 throughout the text , and as illustrated in fig .
[ fig : f2 ] , increasing noise , @xmath27 , comply increasing @xmath21 , in order to keep these results consistence .
here , by applying two modifications to our previous model @xcite , we explored if at the critical point indicated by temporal complexity , avalanche data collapses onto universal scaling function as predicted by the theory of dynamic critical phenomena .
our results provide compelling evidence that temporal complexity and neural avalanches both are consistent and robust indicators of criticality at which information processing , information storage , dynamic response , and computation is maximized @xcite .
we emphasize that the exponents of avalanche data collapses on scaling exponents that are model independent and identical for all systems in the same university class . on the other hand ,
the results of the paper indicate that a simple model using a regular lattice successfully sheds light into cooperation - induced criticality in neural system and establishes a connection between criticality and neural avalanches .
therefore , we conclude that neural avalanches are indicators of criticality .
the authors gratefully acknowledge financial support from iranian cognitive science and technologies council through grant no.2960 .
j. laherr , d. sornette , eur .
j. b * 2 * , 525 ( 1998 ) .
t. petermann , t. thiagarajan , m. lebedev , m. nicolelis , d. chialvo , and d. plenz , proc.natl.acad.sci.u.s.a .
* 106 * , 15921 ( 2009 ) . c. meisel , a. storch , s. hallmeyer - elgner , e. bullmore , and t. gross ( 2012 ) . ploscomput.biol . * 8 * , e1002312 ( 2012 ) .
a. politi a , and s. luccioli , `` _ dynamics of networks of leaky - integrate - and - fire neurons network science _ '' ed e estrada , mfox , dj higham and g - l oppo ( berlin : springer , 2010 ) .
note that the ml function is intimately related to the fractional calculus , widely used in different disciplines , ranging from physics to biology and to economics , and consequently is thought to be a important theoretical tool for the field of complexity . | neural avalanches in size and duration exhibit a power law distribution illustrating as a straight line when plotted on the logarithmic scales .
the power - law exponent is interpreted as the signature of criticality and it is assumed that the resting brain operates near criticality . however , there is no clear evidence that supports this assumption , and even there are extensive research studies conflicting one another .
the model of the current paper is an extension of a previous publication wherein we used an integrate - and - fire model on a regular lattice with periodic boundary conditions and introduced the temporal complexity as a genuine signature of criticality .
however , in that model the power - law distribution of neural avalanches were manifestation of super - criticality rather than criticality . here
, however , we show that replacing the discrete noise in the model with a gaussian noise and continuous time solution of the equation leads to coincidence of temporal complexity and spatiotemporal patterns of neural avalanches at a control parameter which is assumed to be the critical value of the model . |
a fourier-@xmath3 knot is one that can be parameterized as @xmath4 here the _ frequencies _ @xmath5 and @xmath6 are integers and the _ phase shifts _ @xmath7 and @xmath8 and _ amplitudes _ @xmath9 and @xmath10 are real numbers . because any function can be closely approximated by a sum of cosines , every knot is a fourier knot for some @xmath3 .
but a remarkable theorem of lamm @xcite states that in fact every knot is a fourier-@xmath11 knot for some @xmath12 . the special case of fourier-@xmath13 knots
are called _ lissajous _ knots .
lissajous knots are very highly symmetric : either strongly @xmath14 amphicheiral ( if all three frequencies are odd ) , or 2-periodic and linking the axis of rotation once ( if one frequency is even ) .
hence most knots , including all torus knots , are not lissajous knots .
most of what is known about the topic can be found in @xcite@xcite . while every knot can be expressed as a fourier-@xmath11 knot for some integer @xmath12 , it is not clear how big @xmath12 needs to be .
to date no one has a produced a knot that requires @xmath12 to be bigger than 2 .
could every knot be a fourier-@xmath11 knot with @xmath15 ?
it is known that this is true for all twist knots and for all 2-bridge knots to 14 crossings @xcite , @xcite . in this paper
it is shown to be true for all torus knots .
this was conjectured to be true in @xcite and answers the question posed by kauffman in @xcite .
the parameterization of the torus knot @xmath1 given in theorem [ main theorem ] was found after the methodical sampling of fourier-@xmath0 knots undertaken in @xcite yielded all torus knots to 16 crossings .
careful examination of these examples led to the general formula .
the standard parameterization of the @xmath16 torus knot as it winds round an unknotted torus in @xmath17 , @xmath18 times in the longitudinal direction and @xmath19 times in the meridional direction , is @xmath20 here the torus is defined by sweeping the circle of radius @xmath21 centered at the point @xmath22 and lying in the @xmath23-plane around the @xmath24-axis . by using the identity @xmath25 and converting sines to cosines , we may rewrite the parameterization as @xmath26 thus every torus knot can be expressed as fourier-@xmath27 knot .
this parameterization is given by both costa @xcite and kauffman @xcite , and kauffman asks if a simpler fourier representation can be found .
since torus knots can not be lissajous , the following theorem is best possible .
* proof : * the @xmath33-projection of the fourier-@xmath0 knot given by equations [ xparam]-[zparam ] with @xmath34 and @xmath35 is shown in figure [ 37 ] . the knot is oriented in the direction of increasing @xmath36 with the arrowhead placed at @xmath37 . to see that this is indeed a diagram of the torus knot @xmath38 consider only those arcs that move from left to right .
notice that each of these arcs lies completely above any arc that is moving from right to left .
thus the right - going arcs form a braid that lies entirely above the braid formed by the left - going arcs .
the two braids join to form a closed braid that we may think of as braided around the y - axis .
this figure is representative of the general case and it will suffices to prove the following two claims : suppose @xmath39 and @xmath40 are a pair of times that produce a double point in the projection of the knot into the @xmath33-plane .
it is not hard to determine that the pair @xmath41 must be one of the following two types .
the first possibility is that @xmath42 where @xmath43 and @xmath44 .
we call these _ type i _ crossings . at either type crossing
, both arcs are moving to the right , or both to the left , if and only if @xmath50 . evaluating @xmath51 at a type
i crossing we obtain @xmath52 \left [ -p \sin\left ( k\pi+\frac{j p\pi}{q}-\frac{\pi}{2 q}\right)\right]\\ & = & p^2\sin^2\left ( \frac{j p\pi}{q}-\frac{\pi}{2 q}\right)>0 .
\end{aligned}\ ] ] thus at a type i crossing both strings are going to the right , or both are going to the left .
a similar calculation reveals that at type ii crossings the strings are headed in opposite left - right directions . we may compute the sign of a crossing @xmath41 as follows .
if @xmath53 is the projection of the velocity vector into the @xmath33-plane , then the cross product @xmath54 will point up or down .
its dot product with the vector @xmath55 will be positive if and only if the crossing is right handed .
thus the crossing has the same sign as @xmath56\left[z(t_1)-z(t_2)\right ] .
\label{crossing sign formula}\ ] ] using equation [ trig identity ] we may simplify @xmath57 as follows : @xmath58 to prove claim 1 we want to show that at every type ii crossing it is the upper string that is moving to the right .
this is equivalent to showing that @xmath59 using [ z1-z2 ] we obtain @xmath60\\ & = & 2 p \sin^2(pk\pi / q)\left [ 1-(-1)^k\sin(jq\pi / p+\pi/(2p)-\pi/(4q))\right].\end{aligned}\ ] ] this quantity is always positive since within the final factor the sine term has magnitude less than 1 .
thus even if we are subtracting rather than adding , the difference is positive . to prove claim 2 , we must show that every type i crossing is left handed .
computing the expression in [ crossing sign formula ] , one finds that its sign is the same as the sign of @xmath61 which we want to show is negative .
notice that the two angles , @xmath62 and @xmath63 are very close together .
in fact , they can never be separated by a multiple of @xmath64 . for suppose that @xmath65 for some integer @xmath21 .
then , multiplying by @xmath66 gives @xmath67 which is clearly impossible .
thus the sines of the two angles always have the same sign and hence their product is positive .
claims 1 and 2 coupled with the fact that each string oscillates between 1 and @xmath68 in the @xmath69-direction as it moves to the left or the right completes the proof that the closed braid is indeed a torus knot . since there are a total of @xmath48 type i crossings , we see that it must be the knot @xmath1 .
the final assertion , that if @xmath18 is even the phase shift @xmath70 may be replaced with the simpler expression @xmath71 , can be proven by simply repeating the proof with the new phase shift .
an alternative proof , and one that offers more insight into the situation , is to consider the _ phase torus _
+ @xmath72 .
the type i crossing with indices @xmath73 becomes singular on the horizontal line @xmath74 where @xmath75 is an arbitrary integer .
the type ii crossing with indices @xmath73 becomes singular along the diagonal lines @xmath76 where @xmath75 is an arbitrary integer .
more detailed information on the phase torus and the kinds of singular curves that are associated to the type i and ii crossings can be found in @xcite .
each region complementary to the singular lines defines a knot type .
if @xmath18 is even , the two points @xmath77 and @xmath78 lie in the interior of the same region and therefore define the same knot type . if @xmath18 is odd , the second point no longer lies in the same complementary region as the first point .
in fact , it lies at the intersection of two singular lines associated to type ii crossings and on the boundary of the region containing the first point . @xmath79 | every torus knot can be represented as a fourier-@xmath0 knot which is the simplest possible fourier representation for such a knot .
this answers a question of kauffman and confirms the conjecture made by boocher , daigle , hoste and zheng .
in particular , the torus knot @xmath1 can be parameterized as @xmath2 |
the cubic nonlinear schrdinger equation ( nls ) arises naturally in physics and engineering , for instance , as a model of nonlinear optics , langmuir waves in plasmas , and bose - einstein condensation ( bec ) @xcite .
it turns out that the nls can be rigorously derived from the physics of interacting particles in bec via a scaling limit and statistical mechanics @xcite .
it also turns out that the probabilistic tools of statistical mechanics are key to understanding the dynamics of nonlinear schrdinger systems , especially why solitons are so common in certain regimes .
( they are so highly probable that they are the only structures observed macroscopically . ) consider a wavefunction @xmath0 , that is , @xmath0 mapping a time variable @xmath1 and space variable @xmath2 into the complex plane @xmath3 , which satisfies a @xmath4-dimensional nonlinear schrdinger equation ( nls ) : @xmath5 where @xmath6 is the laplacian operator in @xmath7 , @xmath8 is the nonlinearity parameter , and @xmath9 is a coupling parameter which is either @xmath10 ( repulsive / defocusing interaction ) or @xmath11 ( attractive / focusing ) . as a matter of notation , we will usually write a continuous wavefunction in the form @xmath12 and a discrete wavefunction in the form @xmath13 where @xmath14 is the discrete spatial variable . we will be interested in chiefly the cubic ( @xmath15 ) and focusing ( @xmath16 ) nls , and especially its discrete counterpart via the finite difference approximation , where the spatial variable @xmath14 is in a discrete set . also of interest
is a modification of with a fractional power of the laplacian as a non - local dispersive operator , an equation arising naturally in biophysics models of macro - molecules like dna @xcite .
the plan of this article is largely chronological , and background material for the continuous and discrete nls can be found in @xcite . in section [ dnls ] , we will discuss solitons and long - time results for the discrete nls based on analogues of dispersive pde methods .
then in section [ fnls ] there are similar but more nascent results for discrete nls - type systems with long - range interactions , again adapted from pde methods .
we introduce the gibbs measures for continuous nls in section [ sm ] and discuss applications to questions of well - posedness for low - regularity initial data ( [ smwp ] ) , applications to the typicality of solitons in the absence of wave collapse ( [ sms ] ) , and a controversial conjecture for the gibbs measures of the nls ( [ smconj ] ) .
there is a way to get around the problem of nonexistence of gibbs measures in the presence of wave collapse , so in section [ pt ] we discuss new results for the gibbs measures of the 3d discrete nls , where there is a striking phase transition to soliton - like behavior .
consider a lattice @xmath17 which is @xmath18 , for the present , with lattice spacing @xmath19 , i.e. , a positive real number denoting the `` distance '' between two neighboring vertices in @xmath17 @xcite .
the discrete nls on @xmath17 is a family of coupled odes governing the evolution of discrete wavefunctions @xmath20 , i.e. , maps from @xmath21 and @xmath22 into @xmath3 satisfying : @xmath23 here the discrete nearest - neighbor laplacian @xmath24 is defined by @xmath25 note that the sum is over all vertices @xmath26 that are neighbors of @xmath14 , and that division by @xmath27 means that the discrete laplacian converges to the continuous laplacian as @xmath19 goes to zero .
discrete nls equations can exhibit _ localized modes _ or _ discrete breathers _
, which are periodic - in - time and spatially localized solutions .
breathers have a long history in the literature , and flach , kladko , and mackay made a physics conjecture that the energies of breathers must be bounded away from zero in supercritical lattice models ( e.g. , discrete nls in 3d ) @xcite .
weinstein proved one of the conjectures in the supercritical case , that ground state standing waves exist if and only if the total power is larger than some strictly positive threshold @xcite , called the excitation threshold , which is characterized by variational methods . in order to make this precise ,
we define the _ total power _ or _ particle number _ of a wavefunction @xmath28 solving the discrete nls as : @xmath29 also the hamiltonian energy is well - known to be conserved in time : @xmath30 here the inner product @xmath31 is the usual one on @xmath32 .
the ansatz for breathers , or standing waves , is that they are of the form for @xmath33 : @xmath34 then @xmath35 solves the discrete euler - lagrange equation @xmath36 and the solution of , called the _ ground state _ , can be constructed as a minimizer of a variational problem : @xmath37 then for @xmath38 , the minimum is attained by @xmath35 ( which is not necessarily unique ) , with some @xmath39 such that both the constraint @xmath40 and the euler - lagrange system hold . then the main result of @xcite is the following .
[ w ] for supercritical nonlinearities @xmath41 ( for example , cubic nonlinearity in three dimensions ) , there exists an @xmath42-excitation threshold @xmath43 for the ground state of the discrete nls , meaning that : @xmath44 on the other hand , for subcritical @xmath45 ( cubic in one dimension ) , there is no excitation threshold , i.e. , @xmath46 for all @xmath47 . as a preview of the statistical mechanics perspective we will take later
, weinstein s result means that there exist localized modes at zero temperature , corresponding to the ground state .
( in the context of statistical mechanics , temperature is a scaling parameter that determines how likely states of various energies are . ) using statistical mechanics , we will be able to add that such localized modes exist for a whole range of temperatures and moreover are typical in a probabilistic sense .
moreover , there will be a phase transition as the mass and the temperature are varied : for low values of mass or inverse temperature , the states that are highly probable are small states with negligible energies .
but then there s a sudden jump at a critical threshold of mass or inverse temperature to highly probable localized states with negative energies , in fact bounded away from zero .
related work of malomed and weinstein used this same variational approach to analyze the dynamics of solitons in the 1d discrete nls : for nonlinearities up to about cubic , there is a unique soliton of large width and no excitation threshold . for nonlinearities in a medium range between cubic and quintic , there is no threshold and there are three solitons , one unstable and two stable ( in sharp contrast to the situation for the continuous nls ) . and for nonlinearities above quintic there is an excitation threshold for the solitons , one of which is stable and highly localized , and the other unstable and wide @xcite .
weinstein also made a conjecture in @xcite about the subthreshold behavior , that if @xmath48 for the solution @xmath28 of the discrete nls , then @xmath28 disperses to zero , in the sense that : @xmath49.\ ] ] this conjecture has been addressed by similar adaptations of pde methods to the discrete nls by stefanov and kevrekidis , who proved the following decay and strichartz - type estimates for discrete schrdinger ( and klein - gordon ) equations .
[ sk ] for the free discrete schrdinger equation @xmath50 with initial condition @xmath51 , we have the decay estimate for the wavefunction @xmath52 : @xmath53 and for the nonlinear discrete schrdinger equation , with initial data @xmath54 in the sequence space @xmath42 , we have the strichartz - type estimates for any admissible pairs @xmath55 and @xmath56 ( admissible means @xmath57 , @xmath58 , and @xmath59 ) , that @xmath60 where @xmath61 is the dual exponent of @xmath62 , and the mixed lebesgue spaces @xmath63 are defined by : @xmath64 moreover , the constant in the bound can be chosen uniformly in time for any compact interval @xmath65 $ ] .
the decay here is not as fast as for the continuous schrdinger decay estimates ( which is of order @xmath66 ) , but it is sharp in the discrete setting .
one consequence of this approach is the following result about decay of small solutions , addressing weinstein s conjecture and showing that standing wave solutions are impossible in the subthreshold regime .
[ skcor ] if @xmath67 , then there exists a positive @xmath68 depending on dimension only , such that for small enough initial data , @xmath69 , a unique solution exists for all time and vanishes in @xmath70 in the limit as @xmath71 , for all @xmath72 .
in fact : @xmath73 moreover , under certain conditions ( @xmath74 and @xmath75 ) , small solutions decay like the solution to the free schrdinger equation above : for all @xmath76,$]@xmath77 this is a scattering result , where the solution of the nonlinear equation is asymptotic to ( _ scatters to _ ) the solution of the linear ( free ) schrdinger equation as time goes to infinity @xcite .
but this uniform smallness of the solution for small initial data happens only when the nonlinearity is more than just supercritical ( @xmath78 rather than @xmath79 ) .
there are two scaling limits of the discrete schrdinger equation that are of interest .
one is the _ continuum limit _ alluded to earlier , of sending the lattice spacing @xmath19 to zero . in this limit ,
at least in one and two dimensions , one obtains the continuous nonlinear schrdinger equation .
another limit is the _ anti - integrable _ or _ anti - continuum _ limit , resulting in an infinite system of decoupled odes by sending @xmath19 to infinity , for instance , or equivalently sending the total power @xmath80 to infinity . in this case ,
ground states become concentrated on the lattice at a single site @xcite , and their spectral properties and stability issues have been studied extensively @xcite .
( an interesting thing will happen for the statistical mechanics of the continuum limit in three dimensions , however , because the nonlinearity is supercritical there and thus dominates the kinetic energy , making the coupling effectively negligible and producing behavior similar to the anti - continuum limit in that states typically localize at a single site . ) for instance , kevrekidis , pelinovsky , and stefanov @xcite build on the previous dispersive decay estimates of @xcite , in order to prove asymptotic stability of the ground states under some assumptions on the spectrum of the linearization of the dnls equation with external potential in 1d : @xmath81 because of the slower rate of decay in the @xmath82 norm of solutions to the discrete schrdinger equations [ sk ] , the critical power of the nonlinearity is @xmath83 in place of the continuous schrdinger equation s critical power @xmath84 , and @xcite does not handle the cubic case @xmath15 .
define the spaces @xmath85 and @xmath86 by the norms : @xmath87 @xmath88 decompose the solution into a family of solitons and a radiation part : @xmath89.\ ] ] [ kps ] assume that @xmath90 , that @xmath91 , that @xmath92 for some @xmath93 , that no solution @xmath94 of @xmath95 exists in @xmath96 for @xmath97 $ ] , and that @xmath98 has exactly one negative simple eigenvalue @xmath99 with eigenvector @xmath100 and no eigenvalues greater than @xmath101 .
also assume that there is no initial phase shift @xmath102 , that the initial soliton location @xmath103 is close to a fixed @xmath104 , and that the initial data @xmath105 is close to the soliton @xmath106 in the space @xmath42 .
then in the limit as @xmath107 goes to infinity , @xmath108 converges , and the radiation part @xmath109 vanishes in @xmath82 .
this is a nice asymptotic stability result , but what s more remarkable is that something similar happens in a continuum limit of high - dimensional ( @xmath110 ) finite lattices called the _ thermodynamic _ limit , sending @xmath19 to zero and simultaneously taking the total mass ( power divided by number of lattice sites ) to be fixed ( i.e. , sending power to infinity in particular way ) . in the thermodynamic limit ,
one can find a critical threshold above which one obtains states concentrated states at a single site , at least with high probability .
in pure physics applications like optics and bose - einstein condensation , the nls is the model , but in the context of biophysics where particle interactions can be long - range , the natural model turns out to be a schrdinger equation with fractional - order laplacian .
an example is modeling electron transport in dna @xcite . on dna , and other macromolecules with complicated helical structures , electrons or excitations
can propagate along the strand in a very non - local way .
mathematically , these phenomena are just beginning to be explored .
the usual questions about well - posedness and solitons are interesting for the fractional nls , and in fact , there is a corresponding discrete system whose solitons have been studied numerically and whose continuum limit is the predicted fractional nls .
work on the fractional pdes has been done by @xcite , on fractional nls specifically by @xcite , and on the related benjamin - ono equation by @xcite .
the discrete set - up is on a 1d lattice @xmath111 with mesh size @xmath112 , and some fixed @xmath113 such that @xmath114 .
writing @xmath115 with @xmath116 , we consider a family of discrete wavefunctions @xmath117 that satisfy the following discrete nls - type equation : @xmath118 here the sum is over all vertices @xmath26 distinct from fixed @xmath14 , and @xmath119 is a fixed parameter governing the decay of the long - range interaction .
these wavefunctions have conserved quantities similar to the traditional discrete nls for energy : @xmath120 and for mass : @xmath121 equivalent to this last , power is conserved : @xmath122 as a model of dna , the cubic nonlinearity describes a self - interaction for a base pair of the strand with itself , and the summation term models interactions between base pairs decaying like an inverse power of the distance along the strand @xcite .
the complex twisted structure of dna in three dimensions is what allows distant base pairs to interact , even at a long distance . under reasonable conditions , solutions of this fractional
" discrete schrdinger equation @xmath123 exist for all time and are unique by standard fixed - point methods @xcite , and solitary wave solutions of this system of highly coupled odes have been derived @xcite and extended to more general nonlinearities @xcite .
explicit solitary wave solutions for this discrete fractional nls have been studied in the biophysics literature .
for instance , gaididei et al .
@xcite make the ansatz @xmath124 , with localized state : @xmath125 the parameter @xmath126 representing the inverse `` width '' of the soliton and @xmath127 determining its location . numerically studying the associated variational problem , and minimizing @xmath128 under the power constraint @xmath129 ,
they find two kinds of localized states , one called on - site " with @xmath130 and one inter - site " with @xmath131 , depending on the value of the inverse power parameter @xmath132 .
for on - site states with cubic nonlinearity , there appears to be a critical value @xmath133 , estimated to be near @xmath134 , above which the dispersive interaction decays fast enough that the behavior is qualitatively like that of the nearest - neighbor approximation , or traditional discrete nls . and below the critical value @xmath133 , they find three bistable localized states , one low - frequency wide state and two high - frequency narrow states @xcite . on the other hand , inter - site states with cubic nonlinearity
are not so well understood , and they find a lower critical value @xmath135 and a smaller regime @xmath136 in which there are two stable inter - site states . in both cases , the dependence of the power on the frequency of the localized states is studied .
gaididei et al . also present the continuum version of these nonlocal models , with @xmath137 solving the fractional nls of the form : @xmath138 with @xmath139 , a constant @xmath140 depending only on @xmath132 , and @xmath141 depending on @xmath132 appropriately . in the biophysics literature , this limit @xmath142 is routinely taken without mathematically rigorous arguments , but rigor is forthcoming with the result as follows @xcite .
we consider the cubic interaction @xmath143 for simplicity , but this can be generalized to other nonlinearities @xmath8 , and we consider the inverse power law decay in , but this can be generalized to other interaction kernels in @xcite that fall into asymptotic classes with behavior substantively similar to the kernel in with parameter @xmath132 .
the result is that if we are given a family of initial data for the discrete evolution problems , @xmath144 , that approximate a sufficiently smooth initial datum for the fractional nls , then the solutions of the discrete equations converge in an appropriate weak sense to the solution of the fractional nls , with the fractional power of the laplacian : @xmath145 in other words , the critical value is exactly @xmath146 ( not too far from the numerical results in @xcite ) below which the long - range interactions produce a nonlocal fractional nls in the continuum limit , with laplacian of order @xmath147 . above the critical threshold ,
the interaction strength decays so quickly that only local effects survive in the continuum limit , which is exactly the
classical " local nls , @xmath148 . and at the threshold
, we get the classical nls in the continuum limit , with a logarithmic factor appearing in some scaling constants . for this continuum limit result
, we develop fractional discrete sobolev - type inequalities in the discrete setting , uniform embedding , interpolation , and _ a priori _ estimates , and finally for the limit itself , suitable weak compactness arguments .
it would be nice also to obtain rigorous results about the solitary wave states from the biophysics literature and their stability for both the discrete and continuous long - range systems , and these long - range nls systems should be amenable to the statistical mechanics approach discussed next .
adapting tools from the theory of pdes like variational methods and strichartz estimates has been useful for understanding the discrete nls and its localized soliton states , but it is nice to get beyond standard existence results to generic behavior . by setting up the gibbs measure , or statistical ensemble , which is a probability measure on the space of all solutions , we can tell which behaviors are typical in a probabilistic sense . _ gibbs measures _ , or _ invariant measures
_ , are based on normalizing the boltzmann distribution , so that the probability of finding the system in state @xmath0 is : @xmath149 here @xmath150 is the inverse temperature , sometimes replaced by @xmath151 , which makes temperature @xmath152 explicit using the boltzmann constant @xmath153 .
and @xmath154 is the normalizing constant , also known as the partition function , which encodes statistical properties of the system ( in particular the free energy as we ll see in section [ zf ] ) . for dynamical systems such as the nls
, the energy @xmath128 is conserved by the dynamics , so that we can invoke the liouville theorem to see that the measure is invariant .
the intuition behind the boltzmann distribution can be understood by thinking about the extreme cases : in the high temperature case , @xmath155 , the gibbs measure is essentially uniform , that is , the system could statistically be in almost any state . on the other hand , near zero temperature , @xmath156
, the gibbs measure is nearly vanishing for any state but the lowest - energy one , and thus the system has a strong preference for the ground state(s ) . the statistical ensemble associated to the gibbs measure
is called the _ canonical ensemble _ , and
the states that the system is likely to be in , according to the gibbs measure , are called _ canonical macrostates _ , so called because those are the states that are observed macroscopically .
there are variations on the standard gibbs measure and its canonical ensemble .
for instance , the _ microcanonical ensemble _ arises from conditioning the gibbs measure on the energy @xmath128 being a constant @xmath157 ( really being within @xmath68 of @xmath157 and sending @xmath68 to zero appropriately ) .
a natural question to ask is whether the microcanonical ensemble is equivalent to the canonical ensemble , meaning that they share the same macrostates .
the answer is yes and no : if certain conditions are satisfied , they are equivalent , and if not , the microcanonical ensemble has a richer set of macrostates .
another ensemble is the _ petit canonical _ ensemble , which results from conditioning on mass : this is also called a _ mixed ensemble _ because it is canonical in energy and microcanonical in mass .
an advantage of gibbs measures is that they ignore transient states and focus instead on the long - time probabilistic behavior of the system , and this approach was initiated in the work of lebowitz , rose , and speer @xcite for the 1d focusing nls on the circle @xmath158 : @xmath159 the canonical ensemble is nt the right thing to look at because it is not normalizable , so they conjectured that , with a mass cutoff , one can construct a formal invariant measure from the hamiltonian energy : @xmath160 the reason for the mass cutoff is that if one forms the gibbs measure with density @xmath161 , then it is not normalizable because @xmath154 is infinite .
this can be fixed by making @xmath162 the allowed mass : the partition function @xmath163 is then a ( finite ) function of both the thermodynamic @xmath150 and the allowed mass @xmath162 .
thus we form the following formal invariant measure : @xmath164 here the product needs to be interpreted properly , highlighting one of the difficulties in constructing gibbs measures for infinite - dimensional systems : a finite - dimensional approximation must be done , for instance , by spectral truncation or by spatial discretization .
bourgain continued the investigation by using spectral truncation to avoid the harmonic analysis on cyclic groups that would have been necessary for spatial discretization @xcite .
he introduced a discretization in fourier space by truncating the fourier sum at the @xmath165-th fourier coefficient : @xmath166 then the finite - dimensional hamiltonian system is : @xmath167 with the hamiltonian energy functional : @xmath168 where @xmath169 is the @xmath14-th fourier coefficient : @xmath170 the hamiltonian is conserved , as is the truncated mass : @xmath171 next consider a ball in this finite - dimensional phase space : @xmath172 and the natural gibbs measure on @xmath173 at fixed inverse temperature @xmath174 is then : @xmath175 where @xmath176 and the normalization constant , @xmath177 we note that this is a petit , or mixed , canonical ensemble : canonical in @xmath128 and microcanonical in the mass , i.e. , the @xmath178 norm is constrained to be no more than a constant @xmath162 .
the limit of @xmath179 as @xmath180 is @xmath181 , which turns out to be the image measure of the random fourier series map : @xmath182 the @xmath183 are independent identically distributed complex gaussian variables , and the image is almost surely in @xmath184 for any @xmath185 . and the limit of @xmath186 as @xmath180 is @xmath187 , the following weighted wiener measure : @xmath188 here @xmath189 can be represented either through the random fourier series leaving off the zero mode , or the formal weiner measure ( with its own normalization @xmath190 ) : @xmath191 then the invariance of the measures @xmath186 under the finite - dimensional flow @xmath192 of the spectrally truncated nls , combined with the uniform regularity of @xmath192 in @xmath193 and the approximation of the nls flow by @xmath192 , all give the following theorem for the limit measure : the measure @xmath187 in is invariant under the dynamics of the periodic 1d focusing cubic ( @xmath143 ) and quintic ( @xmath194 ) nls .
this is also parlayed into a global well - posedness result for the continuous nls : the cauchy problem for the 1d periodic nls with @xmath195 and with random initial data : @xmath196 is globally well - posed for almost all @xmath197 and @xmath198 satisfying the mass constraint : @xmath199 another advantage of gibbs measures has been illustrated in recent work on the 1d cubic nls with low - regularity initial data on the circle @xmath158 , namely that in spite of the ill - posedness results for the nls with initial data of lower regularity than @xmath178 @xcite , the bad initial data are quite rare , forming a set of vanishing measure with respect to the natural gibbs measure .
in fact , for the cubic nls on the circle , colliander and oh were able to prove almost - sure local well - posedness for initial data in @xmath200 , for every @xmath201 , and global well - posedness for data in @xmath202 , for @xmath203 @xcite .
oh and sulem proved weak continuity of the cubic nls solution map from @xmath204 to the space of distributions @xcite . both papers relied on a wick ordering of the nonlinearity , which seems to be an appropriate approach for the 1d cubic nls with rough initial data ; by contrast molinet proved that the standard solution map ( without wick ordering ) from @xmath204 to the space of distributions is not weakly continuous .
this approach has also been fruitful for almost sure global well - posedness for the periodic derivative nls @xcite .
jordan , josserand , and turkington have established the connection between gibbs measures and solitons for the 1d nls with non - traditional ( i.e. , not polynomial ) nonlinearities @xcite . consider the following nls on a finite interval @xmath205 $ ] with periodic boundary conditions : @xmath206 with focusing and bounded nonlinearity , i.e. , @xmath207 and @xmath208 , and as @xmath209 ; for example a nonlinearity that approximates behavior of large wave amplitudes : @xmath210 as usual the two conserved quantities are energy : @xmath211 where @xmath212 is the primitive of @xmath35 : @xmath213 , and mass : @xmath214 in this 1d setting , the energy appearing in the gibbs measure contains the kinetic term which captures the fine - grained fluctuations and the potential term which essentially ignores those fluctuations . using the same kind of spectral truncation as bourgain , the mixed ensemble measure is formed with a regularization with respect to mass as well as the usual conditioning on constrained mass : @xmath215 ) } \mathbf{1}_{\{\lvert u ^n\rvert^2_2 { \leqslant}b \ } } \prod du^n.\ ] ] they fix the mass @xmath162 and scale the inverse temperature as @xmath216 for large @xmath165 so that the mean energy has a finite limit , then rigorously analyze the limit using large deviations theory , comparing the result with numerical simulations @xcite .
the large deviations principle is that for appropriate sets @xmath217 : @xmath218 where @xmath219 is the large deviations rate function defined for @xmath220 : @xmath221 here @xmath157 is the coherent energy function @xmath222 what this means is that with high probability the random fields ( spectrally truncated wavefunctions ) @xmath223 converge exactly to the ground state solitons of the continuous nls .
this is because the minimizers of the large deviations rate function are the states that are most probable macroscopically , and all other states have exponentially vanishing probability . and since the large deviations rate function is equal to the hamiltonian energy up to a constant , the ground state solitons that solve the traditional variational problem are identical with the most probable macroscopic states .
the ground states also turn out to be stable both macroscopically and microscopically : the large deviations principle says that the macrostate stays close to the ground states with high probability , even if some of the energy goes into turbulent fluctuations ; and the typical microstate as sampled from the gibbs measure @xmath224 is close to a ground state in the @xmath225-norm .
two key things that jordan , josserand , and turkington need in order for their approach to work are non - integrability and the absence of wave collapse , so in particular , it does nt work for the cubic nls in one , two , or three dimensions .
it turns out that the 1d and 3d cases can be handled in the right frameworks , and possibly two dimensions as well .
numerical simulations led lebowitz , rose , and speer to make a second conjecture about a possible phase transition in the invariant measures for the 1d nls : at high temperature and low mass they saw uniformly small wavefunctions , but at low temperature and high mass there emerged sharply concentrated structures like solitons . by contrast , simulations in @xcite suggested that there is no phase transition , and then rider , following on the works of mckean and vaninsky @xcite , confirmed this for the 1d infinite - volume focusing nls by proving that the thermodynamic limit is trivial @xcite .
bourgain also studied invariant measures of the 1d infinite - volume defocusing nls @xcite and of the 2d defocusing nls @xcite ( see also the review article @xcite ) , as did tzvetkov @xcite .
these statistical mechanics methods that work so well for the subcritical nls break down at criticality .
for instance , brydges and slade studied the 2d focusing nls @xcite and saw that the natural gibbs measure construction can not produce an invariant measure for large coupling coefficients of the ( critical ) nonlinearity . similarly for 3d focusing nls , the natural construction is not normalizable , and it is thought to be impossible to make any reasonable gibbs construction .
this is the motivation for the alternative approach explained below , using a spatial discretization to construct the finite - dimensional approximation corresponding to the focusing cubic nls in three dimensions .
there are still unresolved issues about what information can be transferred to the continuum limit , but this approach has turned out to be very fruitful , especially as far as how much can be said about properties of the gibbs measures as it turns out , a lot more than the previous spectral approaches , and even about a phase transition to solitons .
because constructing gibbs measures for the 3d nls appears to be impossible , we consider instead the nls on a spatial discretization of the 3d unit torus @xmath226 ^ 3 $ ] , represented by : @xmath227 so that each vertex of @xmath17 has degree @xmath228 , @xmath17 has @xmath229 sites , and the inter - site distance is @xmath230 .
( we consider this interesting special case here for simplicity , but in @xcite the results actually handle general graphs @xmath231 that are finite and undirected without self - loops , and that are _ high - dimensional _ in the sense that @xmath232 goes to @xmath233 as @xmath142 , where @xmath234 is the distance between two neighboring vertices in @xmath212 and @xmath165 is the number of sites , @xmath180 .
this condition holds for the discrete 3d torus defined above , but does not hold for the discrete 2d torus in the usual continuum limit . )
solutions of are known to exist for all time @xcite , and the dynamics of this system conserves the power @xmath237 , or equivalently mass @xmath238 , and the energy , appropriately normalized by the number of sites @xmath165 : @xmath239 then by the liouville theorem , the natural gibbs measure @xmath240 is invariant under the dynamics of the discrete nls for any real @xmath150 ( inverse temperature ) .
however , this measure has infinite mass if @xmath241 , so we construct the gibbs measures with a mass cut - off to restrict the system to an allowed mass @xmath162 : @xmath242 here @xmath162 is arbitrary and positive , and the normalizing constant , or partition function , is @xmath243 .
this cut - off gibbs measure is still invariant under the nls dynamics because mass is a conserved quantity , and it turns out to be tractable in the limit as the grid size @xmath19 goes to zero , at least in dimensions three and higher .
let @xmath244 be a random element of @xmath245 with probability distribution @xmath246 : that is , @xmath244 is a random wavefunction on @xmath17 such that for each @xmath247 , @xmath248 where @xmath249 denotes the lebesgue differential element on @xmath245 . in order to understand the behavior of the random map @xmath244
, we first study the partition function @xmath163 .
the first theorem below shows that if we have a sequence of graphs with @xmath165 and @xmath232 both tending to infinity , the limit of @xmath250 can be exactly computed for any positive @xmath150 and @xmath162 .
the first result is that for positive temperature , and for the discrete torus in dimensions three and higher , there exists a function @xmath251 that we construct , such that as @xmath180 : @xmath252 more precisely , let @xmath253 be the function @xmath254 it can be checked that @xmath35 has a unique real zero that we call @xmath255 , because @xmath35 is strictly increasing in @xmath256 , @xmath257 and @xmath258 .
numerically , @xmath259 .
let@xmath260 ( figure [ fig1 ] shows a graph of @xmath251 versus @xmath150 when @xmath261 . )
[ zthm ] suppose @xmath262 .
let @xmath263 .
there exists a positive constant @xmath264 depending only on @xmath265 , @xmath150 , @xmath162 , @xmath19 and @xmath4 such that if @xmath266 , then @xmath267 and @xmath268 the proof of these asymptotics for the partition function is at the heart of the work @xcite : large deviations techniques for discrete random variables are used , first obtaining upper and lower bounds for @xmath269 and then optimizing which gives the free energy @xmath251 .
then , because @xmath251 is not differentiable at the critical threshold @xmath255 , that means there is a first - order ( discontinuous ) phase transition at @xmath255 .
the random map @xmath244 behaves quite differently in the two phases , and there is an important quantity which is discontinuous at the threshold . to roughly describe this phase transition ( depicted in figure [ impression ] ) , when @xmath270 there is a single site @xmath271 which bears an abnormally large fraction of the total mass of the random wavefunction @xmath244 .
this fraction is nearly deterministic , given by the ratio @xmath272 , where @xmath273 , where the heaviest site bears a negligible proportion of the mass compared to the total mass @xmath274 .
right : supercritical condensed phase , @xmath270 , where a significant proportion @xmath126 of the mass concentrates at a single site.,title="fig:",height=211 ] , where the heaviest site bears a negligible proportion of the mass compared to the total mass @xmath274 .
right : supercritical condensed phase , @xmath270 , where a significant proportion @xmath126 of the mass concentrates at a single site.,title="fig:",height=211 ] more precisely , if @xmath275 and @xmath276 are the largest and second largest components of the vector @xmath277 , then with high probability @xmath278 and @xmath279 .
moreover , @xmath280 with high probability .
a consequence is that the largest component carries more than half of the total mass : @xmath281 on the other hand , when @xmath282 , then @xmath283 , but still @xmath284 .
consequently @xmath285 the formula for @xmath126 ( see also the graph of @xmath126 in figure [ fig2 ] as a function of @xmath150 for @xmath261 ) shows that @xmath126 does not tend to zero as @xmath286 approaches @xmath255 from above , making it clear that this is a discontinuous ( in physics parlance , first - order ) phase transition . also , in the regime @xmath282 , the energy density is close to zero .
by contrast , when @xmath287 , the energy density @xmath288 is strictly negative and approximately equals @xmath289 , the exceptional site bearing nearly _ all _ of the energy of the system .
this is because the total energy @xmath290 is approximately @xmath291 , while the energy at @xmath14 is , summing over just the neighbors @xmath26 of @xmath14 : @xmath292 the equality by theorem [ phase ] , which makes precise this phase transition : 1 .
first , suppose @xmath296 .
for any @xmath297 , there is a constant @xmath264 depending only on @xmath150 , @xmath162 , @xmath4 , and @xmath62 such that if @xmath298 , then with probability @xmath299 , @xmath300 2 . next , suppose @xmath282 . for any @xmath301 , there is a constant @xmath264 depending only on @xmath150 , @xmath162 , @xmath4 , and @xmath62 such that whenever @xmath302 , with probability @xmath299 @xmath303 3 . finally ,
if @xmath304 and @xmath305 , then there is a constant @xmath264 depending only on @xmath150 , @xmath162 , @xmath4 , @xmath8 and @xmath62 such that whenever @xmath302 , with probability @xmath299 either or holds .
the proof proceeds by a multi - step approximation : first the gibbs measures are approximated by uniform measures on appropriate spaces ( intersections of @xmath42 and @xmath306 spheres ) , and then the uniform measures are approximated by i.i.d .
complex gaussian random variables on @xmath306 annuli . then the i.i.d .
gaussians must concentrate on the parts of the @xmath306 annuli closest to the origin , i.e. , where one coordinate is large and the others are uniformly small .
one consequence of this theorem is confirmation of the conjecture of flach , kladko , and mackay , because the energies of the localized states are bounded away from zero , made clear in figure [ fig4 ] where energy is plotted as a function of @xmath150 .
another straightforward consequence of this theorem is that typical discrete wavefunctions above the critical threshold blow up in the discrete @xmath225 norm : @xmath307 it also turns out that the discrete @xmath225 norm also diverges even for @xmath308 , just with different rates : [ blowup ] suppose the context of theorem [ phase ] holds , in dimension @xmath293 , with @xmath266 .
let @xmath309 be a discrete wavefunction picked randomly from the invariant probability measure @xmath246 defined in .
if @xmath296 , then there is a positive constant @xmath140 depending only on @xmath150 , @xmath162 , @xmath4 and @xmath8 such that @xmath310 whenever @xmath311 . on the other hand , if @xmath308 , then similarly : @xmath312 . since the measure @xmath246 of is invariant for the discrete nls
, one may expect that if the initial data comes from @xmath246 , localized modes will continue to exist for all time .
the question is whether the mode jumps around or stays in one place ( in which case we have a standing or stationary wave , also called a discrete breather ) .
the following theorem uses the statistical equilibrium result to deduce a dynamical result : the same site continues to be the localized mode for an exponentially long period of time .
[ expo ] suppose the previous context holds with @xmath287 and the function @xmath126 as defined in .
let @xmath313 be a discrete wavefunction evolving according to , where the initial data @xmath314 is picked randomly from the invariant probability measure @xmath246 defined in .
then for any @xmath315 , there is a constant @xmath264 depending only on @xmath150 , @xmath162 , @xmath4 , and @xmath62 such that if @xmath266 , then with probability @xmath316 the inequalities hold for @xmath317 for all @xmath318 , and moreover there is a single @xmath271 such that the maximum of @xmath319 is attained at @xmath320 for all @xmath321 . in particular , @xmath317 is approximately a standing wave with localized mode at @xmath14 for an exponentially long time .
this dynamical stability result combines pde and probabilistic methods .
first , with high probability the wavefunction is condensed at time @xmath322 , and then the invariance of the gibbs measure implies that the wavefunction is condensed for an overwhelming proportion of times up to the exponential time in the theorem . if the soliton were to jump from one site to another , then it would have to happen in a tiny window of time ; but that would result in a wavefunction that violates basic nls estimates , so in fact the soliton must stay put .
in other words , it is not only possible , but typical , for solutions of to have unique stable localized modes for exponentially long times if the initial inverse temperature or mass are above a threshold .
earlier results for localized modes were only for existence , and only for large mass , i.e. , tending to infinity ( e.g. , @xcite ) , while theorem [ expo ] proves typicality for finite mass .
our final result is about the probability distribution of the individual coordinates of a random map @xmath244 picked from the measure @xmath323
. we can give a precise description of the distribution for small collections of coordinates : in both phases , the joint distributions are approximately complex gaussian vectors with the appropriate variance ( @xmath162 below the threshold and @xmath324 above it , with @xmath126 defined in ) , provided the collection is sufficiently small compared to the number of sites in the grid and provided a symmetry assumption is satisfied .
the symmetry assumption on @xmath212 is that @xmath212 is _ translatable _ by some group @xmath325 : where @xmath325 is a group of automorphisms of @xmath212 such that : 1 ) @xmath326 , and 2 ) no element of @xmath325 except the identity has any fixed point .
for example , the discrete torus is translatable by the group of translations . [ distribution ] suppose the graph @xmath212 is translatable by some group of automorphisms , @xmath293 , and @xmath244 is a random wavefunction picked according to the measure @xmath246 . take any @xmath327 distinct points @xmath328 .
let @xmath329 be a vector of i.i.d .
standard complex gaussian random variables .
if @xmath282 , then there is a constant @xmath330 depending only on @xmath150 , @xmath162 , and @xmath4 such that if @xmath266 , then for all borel sets @xmath331 , @xmath332 if @xmath287 , the result holds after @xmath333 is replaced with @xmath334 where @xmath335 is defined in , and the error bound is changed to @xmath336 .
as for the future , one direction to push these kinds of results is into the two - dimensional setting .
there the nonlinear and kinetic energies are in a delicate balance , and the wavefunction ca nt be approximated by i.i.d .
gaussians ; instead one must consider correlations between sites , producing what is called a discrete gaussian free field .
another direction is to study the gibbs measures for the fractional nls , which will have non - local behavior coming from the underlying lvy processes .
there is also a possibility that this statistical mechanics approach could be useful for the soliton resolution conjecture : that in certain regimes , solutions generically tend asymptotically towards a soliton plus the small fluctuations of a radiation component .
j. bourgain , on nonlinear schrdinger equations , les relations entre les mathmatiques et la physique thorique : festschrift for the 40th anniversary of the ihs .
institut des hautes tudes scientifiques , 1121 ( 1998 ) . | we review some recent results concerning gibbs measures for nonlinear schrdinger equations ( nls ) , with implications for the theory of the nls , including stability and typicality of solitary wave structures .
in particular , we discuss the gibbs measures of the discrete nls in three dimensions , where there is a striking phase transition to soliton - like behavior . |
a compact riemann surface endowed with an abelian differential admits two natural flows , called , respectively , _ horizontal _ and _ vertical_. one of the main objects of this paper is the space @xmath0 of hlder cocycles over the vertical flow , invariant under the holonomy by the horizontal flow .
equivalently , cocycles in @xmath0 can be viewed , in the spirit of r. kenyon @xcite and f. bonahon @xcite , @xcite , as finitely - additive transverse invariant measures for the horizontal foliation of our abelian differential .
cocycles in @xmath0 are closely connected to the invariant distributions for translation flows in the sense of g.forni @xcite .
the space @xmath0 is finite - dimensional , and for a generic abelian differential the dimension of @xmath0 is equal to the genus of the underlying surface .
theorem [ multiplicmoduli ] , which extends earlier work of a.zorich @xcite and g. forni @xcite , states that the time integral of a lipschitz function under the vertical flow can be uniformly approximated by a suitably chosen cocycle from @xmath0 up to an error that grows more slowly than any power of time .
the renormalizing action of the teichmller flow on the space of hlder cocycles now allows one to obtain limit theorems for translation flows on flat surfaces ( theorem [ limthmmoduli ] ) .
the statement of theorem [ limthmmoduli ] can be informally summarized as follows .
taking the leading term in the asymptotic expansion of theorem [ multiplicmoduli ] , to a generic abelian differential one assigns a compactly supported probability measure on the space of continuous functions on the unit interval .
the normalized distribution of the time integral of a lipschitz function converges , with respect to weak topology , to the trajectory of the corresponding `` asymptotic distribution '' under the action of the teichmller flow .
convergence is exponential with respect to both the lvy - prohorov and the kantorovich - rubinstein metric .
let @xmath1 be an integer , let @xmath2 be a compact orientable surface of genus @xmath3 , and let @xmath4 be a holomorphic one - form on @xmath2
. denote by @xmath5 the area form induced by @xmath4 and assume that @xmath6 .
let @xmath7 be the _ vertical _ flow on @xmath2 ( i.e. , the flow corresponding to @xmath8 ) ; let @xmath9 be the _ horizontal _ flow on @xmath2 ( i.e. , the flow corresponding to @xmath10 ) .
the flows @xmath7 , @xmath9 preserve the area @xmath11 .
take @xmath12 , @xmath13 and assume that the closure of the set @xmath14 does not contain zeros of the form @xmath4 . the set ( [ admrectpsan ] )
is then called _ an admissible rectangle _ and denoted @xmath15 .
let @xmath16 be the semi - ring of admissible rectangles .
consider the linear space @xmath0 of hlder cocycles @xmath17 over the vertical flow @xmath7 which are invariant under horizontal holonomy .
more precisely , a function @xmath18 belongs to the space @xmath0 if it satisfies : [ bplusx ] 1 .
@xmath19 ; 2 .
there exists @xmath20 , @xmath21 such that @xmath22 for all @xmath12 and all @xmath23 satisfying @xmath24 ; 3 . if @xmath15 is an admissible rectangle , then @xmath25 .
a cocycle @xmath26 can equivalently be thought of as a finitely - additive hlder measure defined on all arcs @xmath27 $ ] of the vertical flow and invariant under the horizontal flow .
it will often be convenient to identify the cocycle with the corresponding finitely - additive measure .
for example , let @xmath28 be the lebesgue measure on leaves of the vertical foliation ; the corresponding cocycle @xmath29 defined by @xmath30 of course belongs to @xmath31 . in the same way
define the space @xmath32 of hlder cocycles @xmath33 over the horizontal flow @xmath9 which are invariant under vertical holonomy .
a cocycle @xmath34 can equivalently be thought of as a finitely - additive hlder measure defined on all arcs @xmath35 $ ] of the horizontal flow and invariant under the vertical flow .
let @xmath36 be the lebesgue measure on leaves of the horizontal foliation ; the corresponding cocycle @xmath37 is defined by the formula @xmath38 ; of course , @xmath39 . given @xmath40 , @xmath41 , a finitely additive measure @xmath42 on the semi - ring @xmath16 of admissible rectangles
is introduced by the formula @xmath43 in particular , for @xmath44 , set @xmath45 : @xmath46 for any @xmath41 the measure @xmath47 satisfies @xmath48 and is an invariant distribution in the sense of g. forni @xcite , @xcite .
for instance , @xmath49 .
an @xmath50-linear pairing between @xmath0 and @xmath51 is given , for @xmath40 , @xmath41 , by the formula @xmath52 for an abelian differential @xmath53 let @xmath54 be the space of continuous holonomy - invariant cocycles : more precisely , a function @xmath18 belongs to the space @xmath55 if it satisfies conditions 1 and 3 of assumption [ bplusx ] , while condition 2 is replaced by the following weaker version : for any @xmath56 there exists @xmath57 such that @xmath58 for all @xmath12 and all @xmath23 satisfying @xmath59
. given an abelian differential @xmath53 , we now construct , following katok @xcite , an explicit mapping of @xmath60 to @xmath61 .
a continuous closed curve @xmath62 on @xmath2 is called _ rectangular _ if @xmath63 where @xmath64 are arcs of the flow @xmath7 , @xmath65 are arcs of the flow @xmath9 . for @xmath66 define @xmath67 similarly , for @xmath68 write @xmath69
thus , a cocycle @xmath70 assigns a number @xmath71 to every closed rectangular curve @xmath62 .
it is shown in proposition [ acthomology ] below that if @xmath62 is homologous to @xmath72 , then @xmath73 . for an abelian differential @xmath53
, we thus obtain maps @xmath74 for a generic abelian differential , the image of @xmath31 under the map @xmath75 is the strictly unstable space of the kontsevich - zorich cocycle over the teichmller flow .
more precisely , let @xmath76 be a nonnegative integer vector such that @xmath77 .
denote by @xmath78 the moduli space of pairs @xmath79 , where @xmath2 is a riemann surface of genus @xmath3 and @xmath4 is a holomorphic differential of area @xmath80 with singularities of orders @xmath81 .
the space @xmath78 is often called the _ stratum _ in the moduli space of abelian differentials . the teichmller flow @xmath82 on @xmath78 sends the modulus of a pair @xmath79 to the modulus of the pair @xmath83 , where @xmath84 ; the new complex structure on @xmath2 is uniquely determined by the requirement that the form @xmath85 be holomorphic . as shown by veech ,
the space @xmath78 need not be connected ; let @xmath86 be a connected component of @xmath78 .
let @xmath87 be the fibre bundle over @xmath86 whose fibre at a point @xmath79 is the cohomology group @xmath61 .
the bundle @xmath87 carries the _ gauss - manin connection _ which declares continuous integer - valued sections of our bundle to be flat and is uniquely defined by that requirement .
parallel transport with respect to the gauss - manin connection along the orbits of the teichmller flow yields a cocycle over the teichmller flow , called the _ kontsevich - zorich cocycle _ and denoted @xmath88 .
let @xmath89 be a @xmath82-invariant ergodic probability measure on @xmath86 .
for @xmath90 , @xmath53 , let @xmath91 , @xmath92 be the corresponding spaces of hlder cocycles . denote by @xmath93 the space spanned by vectors corresponding to the positive lyapunov exponents of the kontsevich - zorich cocycle , by @xmath94 the space spanned by vectors corresponding to the negative exponents of the kontsevich - zorich cocycle . for @xmath89-almost
all @xmath90 the map @xmath75 takes @xmath91 isomorphically onto @xmath95 , the map @xmath96 takes @xmath92 isomorphically onto @xmath97 .
the pairing @xmath98 is nondegenerate and is taken by the isomorphisms @xmath99 , @xmath100 to the cup - product in the cohomology @xmath61 .
* remark . * in particular , if @xmath89 is the masur - veech smooth " measure @xcite , then @xmath101 . * remark . * the isomorphisms @xmath75 , @xmath96 are analogues of g. forni s isomorphism @xcite between his space of invariant distributions and the unstable space of the kontsevich - zorich cocycle .
now recall that to every cocycle @xmath102 we have assigned a finitely - additive hlder measure @xmath47 invariant under the flow @xmath7 . considering these measures as distributions in the sense of sobolev and schwartz , we arrive at the following proposition .
[ forni - classif ] let @xmath103 be an ergodic @xmath104 probability measure on @xmath105 the for @xmath106 every abelian differential @xmath107 the space @xmath108 coincides with the space of @xmath109 distributions belonging to the sobolev space @xmath110 by definition for any @xmath111 the distribution @xmath112 is @xmath113-invariant and belongs to the sobolev space @xmath114 .
g.forni has shown that for any @xmath115-invariant ergodic measure @xmath103 and @xmath103-almost every abelian differential @xmath116 the dimension of the space of @xmath9-invariant distributions belonging to the sobolev space @xmath114 @xmath117 the dimension of the strictly expanding oseledets subspace of the kontsevich - zorich cocycle ( under mild additional assumption on the measure @xmath103 g.forni proved that these dimensions are in fact equal , see theorem 8.3 and corollary @xmath118 in @xcite ; note , however , that the proof of the _ upper _ bound in forni s theorem only uses ergodicity of the measure ) . since the dimension of the space @xmath119 equals that of the strictly expanding space for the kontsevich - zorich cocycle for @xmath103-almost all @xmath116 the proposition is proved completely .
consider the inverse isomorphisms @xmath120 let @xmath121 be the distinct positive lyapunov exponents of the kontsevich - zorich cocycle @xmath122 , and let @xmath123 be the corresponding oseledets decomposition at @xmath124 .
[ hoeldergrowth ] let @xmath125 , @xmath126 , and denote @xmath127 .
then for any @xmath56 the cocycle @xmath128 satisfies the hlder condition with exponent @xmath129 and for any @xmath130 such that @xmath131 is defined for all @xmath23 we have @xmath132 [ cochyperb ] if the kontsevich - zorich cocycle does not have zero lyapunov exponent with respect to @xmath89 , then @xmath133 . * remark .
* the condition of the absence of zero lyapunov exponents can be weakened : it suffices to require that the kontsevich - zorich cocycle act isometrically on the neutral oseledets subspace corresponding to the lyapunov exponent zero .
isometric action means here that there exists an inner product which depends measurably on the point in the stratum and which is invariant under the kontsevich - zorich cocycle . in all known examples ( see , e.g. , @xcite )
the action of the kontsevich - zorich cocycle on its neutral lyapunov subspace is isometric ; note , however , that the examples of @xcite mainly concern measures invariant under the action of the whole group @xmath134
. * question .
* does there exist a @xmath82-invariant ergodic probability measure @xmath135 on @xmath86 such that the inclusion @xmath136 is proper almost surely with respect to @xmath135 ? * remark .
* g.forni has made the following remark . to a cocycle @xmath111 assign a @xmath80-current @xmath137 defined , for a smooth 1-form @xmath138 on the surface @xmath139 by the formula @xmath140 where the integral in the right hand side is defined as the limit of riemann sums .
the resulting current @xmath141 is a @xmath142 for the horizontal foliation .
the mapping of hlder cocycles into the cohomology @xmath143 of the surface corresponds to g. forni s map that to each basic current assigns its cohomology class ( the latter is well - defined by the de rham theorem ) .
in particular , it follows that for any ergodic @xmath115-invariant probability measure @xmath103 on @xmath144 and @xmath103-almost every abelian differential @xmath107 every basic current from the sobolev space @xmath114 is induced by a hlder cocycle @xmath145 the space of lipschitz functions is not invariant under @xmath7 , and a larger function space @xmath146 of weakly lipschitz functions is introduced as follows .
a bounded measurable function @xmath147 belongs to @xmath146 if there exists a constant @xmath148 , depending only on @xmath147 , such that for any admissible rectangle @xmath15 we have @xmath149 let @xmath150 be the infimum of all @xmath148 satisfying ( [ weaklippsan ] ) .
we norm @xmath146 by setting @xmath151 by definition , the space @xmath146 contains all lipschitz functions on @xmath2 and is invariant under @xmath7 .
if @xmath152 is an admissible rectangle , then its characteristic function @xmath153 is weakly lipschitz ( i am grateful to c. ulcigrai for this remark ) .
we denote by @xmath154 the subspace of @xmath146 of functions whose integral with respect to @xmath11 is @xmath155 . for any @xmath156 and any @xmath41
the integral @xmath157 can be defined as the limit of riemann sums .
if the pairing @xmath98 induces an isomorphism between @xmath0 and the dual @xmath158 , then one can assign to a function @xmath156 the functional @xmath159 by the formula @xmath160 by definition , @xmath161 .
we are now proceed to the formulation of the first main result of this paper , the approximation theorem [ multiplicmoduli ] .
[ multiplicmoduli ] let @xmath89 be an ergodic probability @xmath82-invariant measure on @xmath86 .
for any @xmath56 there exists a constant @xmath162 depending only on @xmath89 such that for @xmath89-almost every @xmath90 , any @xmath163 , any @xmath12 and any @xmath164 we have @xmath165 consider the case in which the lyapunov spectrum of the kontsevich - zorich cocycle is simple in restriction to the space @xmath166 ( as , by the avila - viana theorem @xcite , is the case with the masur - veech smooth measure ) .
let @xmath167 and let @xmath168 be the corresponding simple expanding lyapunov exponents .
let @xmath29 be given by the formula @xmath30 and introduce a basis @xmath169 in @xmath170 in such a way that @xmath171 lies in the lyapunov subspace with exponent @xmath172 . by proposition [ hoeldergrowth ] , for any @xmath56 the cocycle @xmath173 satisfies the hlder condition with exponent @xmath129 , and for any @xmath130 we have @xmath174 let @xmath175 be the dual basis in @xmath176 . clearly , @xmath38 . by definition , we have @xmath177 noting that by definition we have @xmath178 we derive from theorem [ multiplicmoduli ] the following corollary .
let @xmath89 be an invariant ergodic probability measure for the teichmller flow such that with respect to @xmath89 the lyapunov spectrum of the kontsevich - zorich cocycle is simple in restriction to its strictly expanding subspace .
then for any @xmath56 there exists a constant @xmath162 depending only on @xmath89 such that for @xmath89-almost every @xmath90 , any @xmath163 , any @xmath179 and any @xmath164 we have @xmath180 for horocycle flows a related asymptotic expansion has been obtained by flaminio and forni @xcite
. * remark . *
if @xmath89 is the masur - veech smooth measure on @xmath86 , then it follows from the work of g.forni @xcite , @xcite , @xcite and s. marmi , p. moussa , j .- c .
yoccoz @xcite that the left - hand side is bounded for any @xmath181 ( in fact , for any @xmath147 in the sobolev space @xmath182 ) . in particular ,
if @xmath181 and @xmath183 , then @xmath147 is a coboundary .
holonomy - invariant cocycles assigned to an abelian differential can be interpreted as transverse invariant measures for its foliations in the spirit of kenyon @xcite and bonahon @xcite , @xcite .
let @xmath2 be a compact oriented surface of genus at least two , and let @xmath184 be a minimal oriented measured foliation on @xmath2 .
denote by @xmath185 the transverse invariant measure of @xmath184 .
if @xmath186 $ ] is a smooth curve on @xmath2 , and @xmath187 satisfy @xmath188 , then we denote by @xmath189}\gamma$ ] the curve @xmath190 $ ] .
let @xmath191 be the space of uniformly continuous finitely - additive transverse invariant measures for @xmath184 .
in other words , a map @xmath192 which to every smooth arc @xmath62 transverse to @xmath184 assigns a real number @xmath193 belongs to the space @xmath194 if it satisfies the following : [ finaddmeasfoliation ] 1 .
( finite additivity ) for @xmath186 $ ] and any @xmath195 , we have @xmath196}\gamma)+\phi({\rm res}_{[s , t]}\gamma);\ ] ] 2 .
( uniform continuity ) for any @xmath56 there exists @xmath57 such that for any transverse arc @xmath62 satisfying @xmath197 we have @xmath198 ; 3 .
( holonomy invariance ) the value @xmath193 does not change if @xmath62 is deformed in such a way that it stays transverse to @xmath184 while the endpoints of @xmath62 stay on their respective leaves .
a measure @xmath199 is called hlder with exponent @xmath200 if there exists @xmath201 such that for any transverse arc @xmath62 satisfying @xmath202 we have @xmath203 let @xmath204 be the subspace of hlder transverse measures . as before
, we have a natural map @xmath205 defined as follows . for a smooth closed curve @xmath62 on @xmath2 and
a measure @xmath199 the integral @xmath206 is well - defined as the limit of riemann sums ; by holonomy - invariance and continuity of @xmath192 , this operation descends to homology and assigns to @xmath192 an element of @xmath61 .
now take an abelian differential @xmath53 and let @xmath207 be its horizontal foliation .
we have a `` tautological '' isomorphism between @xmath208 and @xmath55 : every transverse measure for the horizontal foliation induces a cocycle for the vertical foliation and vice versa ; to a hlder measure corresponds a hlder cocycle . for brevity , write @xmath209 .
denote by @xmath210 the unstable subspace of the kontsevich - zorich cocycle of the abelian differential @xmath53 .
theorem [ multiplicmoduli ] and proposition [ cochyperb ] yield the following [ finaddmeasfol ] let @xmath89 be a borel probability measure on @xmath86 invariant and ergodic under the action of the teichmller flow @xmath211 .
then for almost every abelian differential @xmath90 the map @xmath212 takes @xmath213 isomorphically onto @xmath214 .
if the kontsevich - zorich cocycle does not have zero lyapunov exponents with respect to @xmath89 , then for almost all @xmath90 we have @xmath215 .
in other words , in the absence of zero lyapunov exponents all continuous transverse finitely - additive invariant measures are in fact hlder
. * remark . * as before , the condition of the absence of zero lyapunov exponents can be weakened : it suffices to require that the kontsevich - zorich cocycle act isometrically on the oseledets subspace corresponding to the lyapunov exponent zero . by definition ,
the space @xmath213 only depends on the horizontal foliation of our abelian differential ; so does @xmath214 .
let @xmath216 .
let @xmath217 be the standard unit simplex @xmath218 let @xmath219 be a permutation of @xmath220 satisfying the _ irreducibility _ condition : we have @xmath221 if and only if @xmath222 . on the half - open interval @xmath223 consider the points @xmath224 and denote @xmath225 , @xmath226 .
the length of @xmath227 is @xmath228 , while the length of @xmath229 is @xmath230 .
set @xmath231 the map @xmath232 is called _ an interval exchange transformation _ corresponding to @xmath233 . by definition ,
the map @xmath232 is invertible and preserves the lebesgue measure on @xmath234 . by the theorem of masur @xcite and veech @xcite , for any irreducible permutation @xmath219 and for lebesgue - almost all @xmath235 ,
the corresponding interval exchange transformation @xmath232 is uniquely ergodic : the lebesgue measure is the only invariant probability measure for @xmath232 .
consider the space of complex - valued continuous finitely - additive invariant measures for @xmath232 .
more precisely , let @xmath236 be the space of all continuous functions @xmath237\to { \mathbb r}$ ] satisfying 1 .
@xmath238 ; 2 .
if @xmath239 and @xmath232 is continuous on @xmath240 $ ] , then @xmath241 .
each function @xmath192 induces a finitely - additive measure on @xmath242 $ ] defined on the semi - ring of subintervals ( for instance , the function @xmath243 yields the lebesgue measure on @xmath242 $ ] ) .
let @xmath244 be the subspace of hlder functions @xmath245 .
the classification of hlder cocycles over translation flows and the asymptotic formula of theorem [ multiplicmoduli ] now yield the classification of the space @xmath244 and an asymptotic expansion for time averages of almost all interval exchange maps .
let @xmath246 be an abelian differential , and let @xmath247 be a closed interval lying on a leaf of a horizontal foliation .
the vertical flow @xmath248 induces an interval exchange map @xmath249 on @xmath234 , namely , the poincar first return map of the flow . by definition
, there is a natural tautological identification of the spaces @xmath250 and @xmath251 , as well as of the spaces @xmath252 and @xmath253 . for @xmath12 ,
let @xmath254 .
note that the function @xmath255 is uniformly bounded on @xmath2 .
now take a lipschitz function @xmath147 on @xmath234 , and introduce a function @xmath256 on @xmath2 by the formula @xmath257 ( setting @xmath258 for points at which @xmath255 is not defined ) . by definition ,
the function @xmath256 is weakly lipschitz , and theorem [ multiplicmoduli ] is applicable to @xmath256 .
the ergodic integrals of @xmath256 under @xmath248 are of course closely related to ergodic sums of @xmath147 under @xmath249 , and for any @xmath259 , @xmath260 , there exists a time @xmath261 such that @xmath262 by the birkhoff
khintchine ergodic theorem we have @xmath263 where @xmath264 stands for the length of @xmath234 .
furthermore , theorem [ multiplicmoduli ] yields the existence of constants @xmath265 , @xmath266 , such that for all @xmath260 , @xmath259 , we have @xmath267 indeed , the interval @xmath234 induces a decomposition of our surface into weakly admissible rectangles @xmath268 ; denote by @xmath269 the height of the rectangle @xmath270 , and introduce a weakly lipschitz function that assumes the constant value @xmath271 on each rectangle @xmath270 . applying theorem [ multiplicmoduli ] to this function we arrive at desired estimate . in view of the estimate ( [ txn - est ] ) , theorem [ multiplicmoduli ]
applied to the function @xmath256 now yields the following corollary .
[ iet - approx ] let @xmath103 be a @xmath272 ergodic probability measure on @xmath273 for any @xmath56 there exists @xmath274 depending only on @xmath103 such that the following holds . for almost every abelian differential @xmath275 any horizontal closed interval @xmath247 , any lipschitz function @xmath276 , any @xmath260 and all @xmath277 we have @xmath278 * proof .
* applying theorem [ multiplicmoduli ] to @xmath256 , using the estimate @xmath279 and noting that the weakly lipschitz norm of @xmath256 is majorated by the lipschitz norm of @xmath280 we obtain the desired corollary .
let @xmath281 be the distinct positive lyapunov exponents of the measure @xmath282 and let @xmath283 be the dimensions of the corresponding subspaces .
the tautological identification of @xmath284 and @xmath285 together with the results of the previous corollary now implies zorich - type estimates for the growth of ergodic sums of @xmath286 more precisely , we have the following [ iet - logasympt ] in the assumptions of the preceding corollary , the space @xmath287 admits a flag of subspaces @xmath288 such that any finitely - additive measure @xmath289 in hlder with exponent @xmath290 for any @xmath56 and that for any lipschitz function @xmath276 and for any @xmath260 we have @xmath291 where @xmath292 for all @xmath293 and by convention we set @xmath294 if with respect to the measure @xmath89 the kontsevich - zorich cocycle acts isometrically on its neutral subspaces , then we also have @xmath295 * remark*. corollaries [ iet - approx ] , [ iet - logasympt ] thus yield the asymptotic expansion in terms of hlder cocycles as well as zorich - type logarithmic estimates for almost all interval exchange transformations with respect to any conservative ergodic measure @xmath296 on the space of interval exchange transformations , invariant under the rauzy - veech induction map and such that the kontsevich - zorich cocycle is log - integrable with respect to @xmath297 in particular , for the lebesgue measure , if we let @xmath298 be the rauzy class of the permutation @xmath219 , then , using the simplicity of the lyapunov spectrum given by the avila - viana theorem @xcite , we obtain for any irreducible permutation @xmath219 and for lebesgue - almost all @xmath299 all continuous finitely - additive measures are hlder : we have @xmath300 for any irreducible permutation @xmath219 there exists a natural number @xmath301 depending only on the rauzy class of @xmath219 and such that 1 . for lebesgue - almost all @xmath299 we have @xmath302 ; 2 .
all the spaces @xmath303 are one - dimensional and @xmath304 . the second statement of corollary [ iet - logasympt ] recovers , in the case of the lebesgue measure on the space of interval exchange transformations , the zorich logarithmic asymptotics for ergodic sums @xcite , @xcite
. * remark .
* objects related to finitely - additive measures for interval exchange transformations have been studied by x. bressaud , p. hubert and a. maass in @xcite and by s. marmi , p. moussa and j .- c .
yoccoz in @xcite .
in particular , the `` limit shapes '' of @xcite can be viewed as graphs of the cocycles @xmath17 considered as functions in @xmath305 . as before , @xmath79 is an abelian differential , and @xmath7 , @xmath9 are , respectively , its vertical and horizontal flows . take @xmath306 $ ] , @xmath307 , a real - valued @xmath308 and introduce the function @xmath309=\int_0^{\tau\exp(s ) } f\circ h^{+}_t(x)dt.\ ] ] for fixed @xmath147 , @xmath310 and @xmath311 the quantity @xmath312 $ ] is a continuous function of @xmath306 $ ] ; therefore , as @xmath311 varies in the probability space @xmath313 , we obtain a random element of @xmath314 $ ] . in other words , we have a random variable @xmath315 : ( m , \nu)\to c[0,1]\ ] ] defined by the formula ( [ sfstaux ] ) . for any fixed @xmath306 $ ] the formula ( [ sfstaux ] )
yields a real - valued random variable @xmath316 : ( m , \nu)\to { \mathbb r},\ ] ] whose expectation , by definition , is zero .
our first aim is to estimate the growth of its variance as @xmath317 . without losing generality ,
one may take @xmath318 .
let @xmath89 be an invariant ergodic probability measure for the teichmller flow such that with respect to @xmath89 the second lyapunov exponent @xmath319 of the kontsevich - zorich cocycle is positive and simple ( recall that , as veech and forni showed , the first one , @xmath320 , is always simple @xcite and that , by the avila - viana theorem @xcite , the second one is simple for the masur - veech smooth measure ) . for an abelian differential @xmath53 , denote by @xmath321 the one - dimensional subspace in @xmath61 corresponding to the second lyapunov exponent @xmath319 , and
. similarly , denote by @xmath323 the one - dimensional subspace in @xmath61 corresponding to the lyapunov exponent @xmath324 , and let @xmath325 .
recall that the space @xmath61 is endowed with the hodge norm @xmath326 ; the isomorphisms @xmath327 take the hodge norm to a norm on @xmath328 ; slightly abusing notation , we denote the latter norm by the same symbol .
introduce a multiplicative cocycle @xmath329 over the teichmller flow @xmath82 by taking @xmath330 , @xmath126 , and setting @xmath331 the hodge norm is chosen only for concreteness in ( [ h2sx ] ) ; any other norm can be used instead . by definition
, we have @xmath332 now take @xmath333 @xmath334 in such a way that @xmath335 .
[ varestfmoduli ] there exists @xmath336 depending only on @xmath89 and positive measurable functions @xmath337 such that the following is true for @xmath89-almost all @xmath90 .
if @xmath338 satisfies @xmath339 , then for all @xmath340 we have @xmath341 * remark .
* observe that the quantity @xmath342 does not depend on the specific choice of @xmath343 , @xmath344 such that @xmath345 .
* note that by theorems of egorov and luzin , the estimate ( [ varestest ] ) holds _ uniformly _ on compact subsets of @xmath86 of probability arbitrarily close to @xmath80 .
proposition [ varestfmoduli ] is based on there exists a positive measurable function @xmath346 such that for @xmath89-almost all @xmath90 , we have @xmath347 in particular @xmath348 for any @xmath307 .
the function @xmath349 is given by @xmath350 observe that the right - hand side does not depend on a particular choice of @xmath351 , @xmath352 .
go back to the @xmath314$]-valued random variable @xmath353 $ ] and denote by @xmath354 $ ] the distribution of the normalized random variable @xmath355}{{\sqrt{var_{\nu } { \mathfrak s}[f , s;1]}}}.\ ] ] the measure @xmath354 $ ] is thus a probability distribution on the space @xmath314 $ ] of continuous functions on the unit interval . for @xmath356 , @xmath357
, we also let @xmath358 $ ] be the distribution of the @xmath50-valued random variable @xmath359}{{\sqrt{var_{\nu } { \mathfrak s}[f , s ; \tau]}}}.\ ] ] if @xmath147 has zero average , then , by definition , @xmath358 $ ] is a measure on @xmath50 of expectation @xmath155 and variance @xmath80 . by definition , @xmath354 $ ] is a borel probability measure on @xmath314 $ ] ; furthermore , if @xmath360 $ ] , then the following natural normalization requirements hold for @xmath361 with respect to @xmath354 $ ] : 1 .
@xmath362 almost surely with respect to @xmath363 $ ] ; 2 . @xmath364}\xi(\tau)=0 $ ] for all @xmath306 $ ] ; 3 . @xmath365}\xi(1)=1 $ ] .
we are interested in the weak accumulation points of @xmath354 $ ] as @xmath317 .
consider the space @xmath366 given by the formula @xmath367 by definition , the space @xmath366 is a @xmath89-almost surely defined two - to - one cover of the space @xmath86 .
the skew - product flow of the kontsevich - zorich cocycle over the teichmller flow yields a flow @xmath368 on @xmath366 given by the formula @xmath369 given @xmath370 , set @xmath371 take @xmath372 such that @xmath373 and set @xmath374 let @xmath375 be the space of all probability distributions on @xmath314 $ ] and introduce a @xmath89-almost surely defined map @xmath376 by setting @xmath377 to be the distribution of the @xmath314$]-valued normalized random variable @xmath378.\ ] ] by definition , @xmath377 is a borel probability measure on the space @xmath314 $ ] ; it is , besides , a compactly supported measure as its support consists of equibounded hlder functions with exponent @xmath379 .
consider the set @xmath380 of probability measures @xmath381 on @xmath314 $ ] satisfying , for @xmath382 $ ] , @xmath383 , the conditions : 1 .
the equality @xmath362 holds @xmath381-almost surely ; 2 . for all @xmath384 we have @xmath385 : 3 .
we have @xmath386 and for any @xmath357 we have @xmath387 .
it will be proved in what follows that @xmath388 .
consider a semi - flow @xmath389 on the space @xmath314 $ ] defined by the formula @xmath390 introduce a semi - flow @xmath391 on @xmath380 by the formula @xmath392 by definition , the diagram @xmath393 is commutative .
let @xmath394 be the lvy - prohorov metric and let @xmath395 be the kantorovich - rubinstein metric on the space of probability measures on @xmath314 $ ] ( see @xcite , @xcite and the appendix ) .
we are now ready to formulate the main result of this subsection .
[ limthmmoduli - simple ] let @xmath89 be a @xmath82-invariant ergodic probability measure on @xmath396 such that the second lyapunov exponent of the kontsevich - zorich cocycle is positive and simple with respect to @xmath89 .
there exists a positive measurable function @xmath397 and a positive constant @xmath398 depending only on @xmath89 such that for @xmath89-almost every @xmath399 , @xmath400 , and any @xmath338 satisfying @xmath401 we have @xmath402 , { \mathcal d}_2^+({\bf g}_s^{\prime}{\bf x}^{\prime}))\leq c({\bf x } , { \bf g}_s{\bf x})\exp(-\alpha s).\ ] ] @xmath403 , { \mathcal d}_2^+({\bf g}_s^{\prime}{\bf x}^{\prime}))\leq c({\bf x } , { \bf g}_s{\bf x})\exp(-\alpha s).\ ] ] now fix @xmath356 and let @xmath404 be the distribution of the @xmath50-valued random variable @xmath405 for brevity , write @xmath406 . for @xmath89-almost any @xmath399 , the measure @xmath404 admits atoms for a dense set of times @xmath356 .
a more general proposition on the existence of atoms will be formulated in the following subsection .
proposition [ limthmmoduli - simple ] implies that the omega - limit set of the family @xmath407 $ ] can generically assume at most two values .
more precisely , the ergodic measure @xmath89 on @xmath86 is naturally lifted to its `` double cover '' on the space @xmath366 : each point in the fibre is assigned equal weight ; the resulting measure is denoted @xmath135 . by definition , the measure @xmath135 has no more than two ergodic components .
we therefore arrive at the following [ nonentwo ] let @xmath89 be a @xmath82-invariant ergodic probability measure on @xmath396 such that the second lyapunov exponent of the kontsevich - zorich cocycle is positive and simple with respect to @xmath89 .
there exist two closed sets @xmath408 such that for @xmath89-almost every @xmath409 and any @xmath338 satisfying @xmath410 the omega - limit set of the family @xmath407 $ ] either coincides with @xmath411 or with @xmath412 .
if , additionally , the measure @xmath135 is ergodic , then @xmath413 .
* question*. do the sets @xmath414 contain measures with non - compact support ? for horocycle flows on compact surfaces of constant negative curvature ,
compactness of support for all weak accumulation points of ergodic integrals has been obtained by flaminio and forni @xcite . * question*. is the measure @xmath135 ergodic when @xmath415 is the masur - veech measure ? as we shall see in the next subsection , in general , the omega - limit sets of the distributions of the @xmath50-valued random variables @xmath416 $ ] contain the delta - measure at zero . as a consequence
, it will develop that , under certain assumptions on the measure @xmath89 , which are satisfied , in particular , for the masur - veech smooth measure , for a generic abelian differential the random variables @xmath416 $ ] _ do not converge _ in distribution , as @xmath317 , for any function @xmath417 such that @xmath410 .
while , by the avila - viana theorem @xcite , the lyapunov spectrum of the masur - veech measure is simple , there are also natural examples of invariant measures with non - simple positive second lyapunov exponent due to eskin - kontsevich - zorich @xcite , g. forni @xcite , c. matheus ( see appendix a.1 in @xcite ) .
a slightly more elaborate , but similar , construction is needed to obtain limit theorems in this general case .
let @xmath103 be an invariant ergodic probability measure for the teichmller flow , and let @xmath418 be the distinct positive lyapunov exponents of the kontsevich - zorich cocycle with respect to @xmath419 we assume @xmath420 . as before , for @xmath409 and @xmath421 ,
let @xmath422 be the corresponding oseledets subspaces , and let @xmath423 be the corresponding spaces of cocycles . to make notation lighter , we omit the symbol @xmath124 when the abelian differential is held fixed . for @xmath424
we now write @xmath425 with @xmath426 and , of course , with @xmath427 where @xmath28 is the lebesgue measure on the vertical foliation . for each @xmath428 introduce a measurable fibre bundle @xmath429 the flow @xmath82
is naturally lifted to the space @xmath430 by the formula @xmath431 the growth of the norm of vectors @xmath432 is controlled by the multiplicative cocycle @xmath433 over the flow @xmath434 defined by the formula @xmath435 for @xmath409 and @xmath436 satisfying @xmath410 , denote @xmath437 define @xmath438 by the formula @xmath439 the growth of the variance of the ergodic integral of a weakly lipschitz function @xmath147 is also , similarly to the case of the simple second lyapunov exponent , described by the cocycle @xmath440 in the following way .
there exists @xmath336 depending only on @xmath103 and , for any @xmath441 positive measurable functions @xmath442 such that for @xmath103-almost every @xmath443 the following holds .
let @xmath436 satisfy @xmath410 .
then for all @xmath444 we have @xmath445 ) } { v^{(i(f))}(g_s^{{\bf{s}}^{(i(f))}}(\overline{\omega},v_{f}))(h_{i(f)}(s , ( { \bf x},v_{f } ) ) ) ^2}-1\right|\leqslant c^{(i(f))}({\bf x},{\bf g}_s{\bf x})e^{-\alpha s}.\ ] ] we proceed to the formulation and the proof of the limit theorem in the general case . for @xmath428 ,
introduce the map @xmath446 by setting @xmath447 to be the distribution of the @xmath314$]-valued random variable @xmath448.\ ] ] as before , we have a commutative diagram @xmath449 the measure @xmath450 $ ] , as before , stands for the distribution of the @xmath314$]-valued random variable @xmath451.\ ] ] [ limthmmoduli ] let @xmath103 be an invariant ergodic probability measure for the teichmller flow such that the kontsevich - zorich cocycle admits at least two distinct positive lyapunov exponents with respect to @xmath89 .
there exists a constant @xmath336 depending only on @xmath103 and a positive measurable map @xmath452 such that for @xmath106 every @xmath443 and any @xmath453 we have @xmath454,{\mathcal d}_{i(f)}^+(g_s^{{\bf{s}}^{(i(f))}}({\bf x},v_f ) ) ) \leqslant c({\bf x},{\bf g}_s { \bf x})e^{-\alpha s},\ ] ] @xmath455,{\mathcal d}_{i(f)}^+(g_s^{{\bf{s}}^{(i(f))}}({\bf x},v_f ) ) ) \leqslant c({\bf x},{\bf g}_s{\bf x})e^{-\alpha s}.\ ] ] for @xmath456 , let @xmath457 $ ] be the distribution of the @xmath458-valued random variable @xmath459 for @xmath106 every @xmath460 there exists a dense set @xmath461 such that if @xmath462 then for any @xmath456 , @xmath463 , the measure @xmath464 admits atoms .
our next aim is to show that along certain subsequences of times the ergodic integrals of translation flows converge in distribution to the measure @xmath465 , the delta - mass at zero .
weak convergence of probability measures will be denoted by the symbol @xmath466 .
we need the following additional assumption on the measure @xmath89 .
[ aczero ] for any @xmath56 the set of abelian differentials @xmath53 such that there exists an admissible rectangle @xmath467 with @xmath468 , @xmath469 has positive measure with respect to @xmath89 .
of course , this assumption holds for the masur - veech smooth measure .
let @xmath103 be an ergodic @xmath82-invariant measure on @xmath144 satisfying assumption [ aczero ] . then for @xmath103-almost
every @xmath470 there exists a sequence @xmath471 such that for any @xmath472 , @xmath463 , we have @xmath473\rightarrow\delta_0\ \mathrm{in } \ { \mathfrak m}({\mathbb r } ) \ \mathrm{as}\ n\to\infty.\ ] ] theorem [ limthmmoduli ] now implies the following let @xmath103 be an ergodic @xmath82-invariant measure on @xmath144 satisfying assumption [ aczero ] .
then for @xmath103-almost every @xmath470 there exists a sequence @xmath474 such that for any @xmath475 satisfying @xmath410 we have @xmath476\rightarrow\delta_0\ \mathrm{in } \ { \mathfrak m}({\mathbb r } ) \
\mathrm{as}\ n\to\infty.\ ] ] consequently , if @xmath477 satisfies @xmath478 then the family of measures @xmath479 $ ] does not converge in @xmath480 and the family of measures @xmath481 $ ] does not converge in @xmath482)$ ] as @xmath483 in this subsection we show that for an arbitrary abelian differential @xmath53 the map @xmath484 given by ( [ maptocohomology ] ) is indeed well - defined .
[ acthomology ] let @xmath485 , @xmath486 , be rectangular closed curves such that the cycle @xmath487 is homologous to @xmath155 .
then for any @xmath66 we have @xmath488 informally , proposition [ acthomology ] states that the _ relative _ homology of the surface with respect to zeros of the form is not needed for the description of cocycles .
arguments of this type for invariant measures of translation flows go back to katok s work @xcite .
we proceed to the formal proof .
take a fundamental polygon @xmath152 for @xmath2 such that all its sides are simple closed rectangular curves on @xmath2 .
let @xmath489 be the boundary of @xmath152 , oriented counterclockwise . by definition , @xmath490 since each curve of the boundary enters @xmath489 twice and with opposite signs .
we now deform the curves @xmath485 to the boundary @xmath489 of our fundamental polygon .
[ simplezero ] let @xmath491 be a simple rectangular closed curve .
then @xmath492 proof of proposition [ simplezero ] .
we may assume that @xmath62 is oriented counterclockwise and does not contain zeros of the form @xmath4 .
by jordan s theorem , @xmath62 is the boundary of a domain @xmath493 .
let @xmath494 be zeros of @xmath4 lying inside @xmath495 ; let @xmath496 be the order of @xmath497 .
choose an arbitrary @xmath56 , take @xmath57 such that @xmath498 as soon as the length of @xmath62 does not exceed @xmath499 and consider a partition of @xmath495 given by @xmath500 where all @xmath501 are admissible rectangles and @xmath502 is a @xmath503-gon containing @xmath497 and no other zeros and satisfying the additional assumption that all its sides are no longer than @xmath499 .
let @xmath504 , @xmath505 stand for the boundaries of our polygons oriented counterclockwise .
we have @xmath506 in the first sum , each term is equal to @xmath155 by definition of @xmath128 , whereas the second sum does not exceed , in absolute value , the quantity @xmath507 where @xmath508 is a positive constant depending only on @xmath509 . since @xmath510 may be chosen arbitrarily small , we have @xmath511 which is what we had to prove . for @xmath512 ,
let @xmath513 be the part of @xmath489 going counterclockwise from @xmath514 to @xmath515 .
let @xmath512 and let @xmath516 be an arbitrary rectangular curve going from @xmath514 to @xmath515 .
then @xmath517 we may assume that @xmath62 is simple in @xmath152 , since , by proposition [ simplezero ] , self - intersections of @xmath62 ( whose number is finite ) do not change the value of @xmath71 . if @xmath62 is simple , then @xmath62 and @xmath518 together form a simple closed curve , and the proposition follows from proposition [ simplezero ] . if @xmath491 is a rectangular curve which yields a closed curve in @xmath2 homologous to zero in @xmath2 , then @xmath492 indeed , by the previous proposition we need only consider the case when @xmath519 . since @xmath62 is homologous to @xmath155 by assumption , the cycle @xmath62 is in fact a multiple of the cycle @xmath489 , for which the statement follows from ( [ deltapizeroo ] ) .
let @xmath79 be an abelian differential .
a rectangle @xmath520 is called _ weakly admissible _ if for all sufficiently small @xmath56 the rectangle @xmath521 is admissible ( in other words , the boundary of @xmath152 may contain zeros of @xmath4 but the interior does not ) . assume that we are given a natural number @xmath522 and a sequence of partitions @xmath523 @xmath524 where @xmath525 are weakly admissible rectangles .
the sequence @xmath523 of partitions of @xmath2 into @xmath522 weakly admissible rectangles will be called a _
markovian sequence _ of partitions if for any @xmath526 , @xmath527 , the rectangles @xmath528 and @xmath529 intersect in a markov way in the following precise sense .
take a weakly admissible rectangle @xmath15 and decompose its boundary into four parts : @xmath530 @xmath531 @xmath532 @xmath533 the sequence of partitions @xmath523 has the _ markov property _ if for any @xmath534 and @xmath535 there exist @xmath536 such that @xmath537 @xmath538 @xmath539 @xmath540 to a markovian sequence of partitions we assign the sequence of @xmath541 _ adjacency matrices _
@xmath542 defined as follows : @xmath543 is the number of connected components of the intersection @xmath544 a markovian sequence of partitions @xmath523 will be called an _
markovian sequence of partitions if @xmath545 for an abelian differential both whose vertical and horizontal flows are minimal , there always exists @xmath216 and a sequence of partitions ( [ recpart ] ) having the markov property and satisfying the exactness condition ( [ minimality ] ) . a suitably chosen markovian sequence of partitions will be essential for the construction of finitely - additive measures in the following section .
* remark . *
an exact markovian sequence of partitions allows one to identify our surface @xmath2 with the space of trajectories of a non - autonomous markov chain , or , in other words , a markov compactum .
the horizontal and vertical foliations then become the asymptotic foliations of the corresponding markov compactum ; the finitely - additive measures become finitely - additive measures on one of the asymptotic foliations invariant under holonomy with respect to the complementary foliation ; the vertical and horizontal flow also admit a purely symbolic description as flows along the leaves of the asymptotic foliations according to an order induced by a vershik s ordering on the edges of the graphs forming the markov compactum .
the space of abelian differentials , or , more precisely , the veech space of zippered rectangles , is then represented as a subspace of the space of markov compacta .
the space of markov compacta is a space of bi - infinite sequences of graphs and is therefore endowed with a natural shift transformation . using rauzy - veech expansions of zippered rectangles ,
one represents the teichmller flow as a suspension flow over this shift .
the kontsevich - zorich cocycle is then a particular case of the cocycle over the shift given by consecutive adjacency matrices of the graphs forming our markov compactum . to an abelian differential ,
random with respect to a probability measure invariant under the teichmller flow , one can thus assign a markov compactum corresponding to a sequence of graphs generated according to a stationary process .
the relation between markov compacta and abelian differentials is summarized in the table below .
the main theorems of this paper , theorem [ multiplicmoduli ] and theorem [ limthmmoduli ] , are particular cases of general theorems on the asymptotic behaviour of ergodic averages of symbolic flows along asymptotic foliations of random markov compacta ; these generalizations , which will be published in the sequel to this paper , are proved in the preprint @xcite .
* markov compacta * & * abelian differentials * + asymptotic foliations of a markov compactum & horizontal and vertical foliations of an abelian differential + finitely - additive measures on asymptotic foliations of a markov compactum&the spaces @xmath546 and @xmath547 of hlder cocycles + vershik s automorphisms&interval exchange transformations + suspension flows over vershik s automorphisms&translation flows on flat surfaces + the space of markov compacta&the moduli space of abelian differentials + the shift on the space of markov compacta&the teichmller flow + the cocycle of adjacency matrices&the kontsevich - zorich cocycle + * acknowledgements . * w. a. veech made the suggestion that g. forni s invariant distributions for the vertical flow should admit a description in terms of cocycles for the horizontal flow , and i am greatly indebted to him .
g. forni pointed out that cocycles are dual objects to invariant distributions and suggested the analogy with f. bonahon s work @xcite ; i am deeply grateful to him .
i am deeply grateful to a. zorich for his kind explanations of kontsevich - zorich theory and for discussions of the relation between this paper and f. bonahon s work @xcite .
i am deeply grateful to h. nakada who pointed out the reference to s. ito s work @xcite and to b. solomyak who pointed out the reference to the work @xcite of p. dumont , t. kamae and s. takahashi and the work @xcite of t. kamae .
i am deeply grateful to j. chaika , p. hubert , yu.s .
ilyashenko , h. krger , e. lanneau , s. mkrtchyan and r. ryham for many helpful suggestions on improving the presentation .
i am deeply grateful to a. avila , x. bressaud , b.m .
gurevich , a.b .
katok , a.v .
klimenko , s.b .
kuksin , c. mcmullen , v.i .
oseledets , ya.g .
sinai , i.v .
vyugin , j .- c .
yoccoz for useful discussions .
i am deeply grateful to n. kozin , d. ong , s. sharahov and r. ulmaskulov for typesetting parts of the manuscript .
part of this paper was written while i was visiting the max planck institute of mathematics in bonn and the institut henri poincar in paris . during the work on this paper ,
i was supported in part by an alfred p. sloan research fellowship , by the dynasty foundation fellowship and the ium - simons fellowship , by grant mk-6734.2012.1 of the president of the russian federation , by the programme on dynamical systems and mathematical control theory of the presidium of the russian academy of sciences , by the edgar odell lovett fund at rice university , by the rfbr - cnrs grant 10 - 01 - 93115 , by the rfbr grant 11 - 01 - 00654 and by the national science foundation under grant dms 0604386 .
let @xmath548 be a bi - invariant sequence of invertible @xmath541-matrices with non - negative entries @xmath549 to a vector @xmath550 we assign the corresponding @xmath551-_equivariant _ sequence @xmath552 given by the formula @xmath553 we now consider subspaces in @xmath554 consisting of vectors @xmath555 such that the corresponding equivariant subsequence @xmath556 decays as @xmath557 tends to @xmath558 .
more formally , we write @xmath559 it will sometimes be convenient to identify a vector with the corresponding equivariant sequence , and , slightly abusing notation , we shall sometimes say that a given equivariant sequence belongs to the space @xmath560 or @xmath561 . as before ,
let @xmath79 be an abelian differential . by an arc of the vertical flow
we mean a set of the form @xmath562 such a set will also sometimes be denoted @xmath563,$ ] where @xmath564 let @xmath565 be an exact markovian sequence of partitions @xmath566 into weakly admissible rectangles .
take a rectangle @xmath567 and choose an arbitrary arc @xmath568 of the vertical foliation going all the way from the lower boundary of @xmath525 to the upper boundary .
more formally , take @xmath569 take @xmath305 such that @xmath570 and let @xmath568 be the vertical arc @xmath571 $ ] .
an arc @xmath568 of this form will be called a _ markovian _ arc .
the family of arcs @xmath572 will be called _ a canonical system of arcs _ assigned to the markov sequence of partitions @xmath523 .
of course , there is freedom in the choice of specific arcs @xmath568 , but our constructions will not depend on the specific choice of a canonical system .
given a finitely - additive measure @xmath573 introduce a sequence of vectors @xmath574 by setting @xmath575 now let @xmath576 @xmath577 be the sequence of adjacency matrices of the sequence of partitions @xmath523 , and assume all @xmath578 to be invertible . by the horizontal holonomy invariance , the value @xmath579 does not depend on the specific choice of the arc @xmath568 inside the rectangle @xmath525 .
finite additivity of the measure @xmath128 implies that the sequence @xmath580 , is @xmath551-equivariant .
exactness of the sequence of partitions @xmath581 , implies that that the equivariant sequence @xmath582 corresponding to a finitely - additive measure @xmath583 satisfies @xmath584 we have therefore obtained a map @xmath585 it will develop that under certain natural additional assumptions this map is indeed an isomorphism .
we now take an abelian differential @xmath107 whose vertical flow is uniquely ergodic and show that if the heights of the rectangles @xmath525 decay exponentially as @xmath586 , then the map @xmath587 sends @xmath588 to @xmath561 .
we proceed to precise formulations .
introduce a sequence @xmath589 by setting @xmath590 to be the height of the rectangle @xmath591 by the markov property , the sequence @xmath592 is @xmath551-equivariant ; unique ergodicity of the vertical flow and exactness of the sequence @xmath581 , imply that a positive @xmath551-equivariant sequence is unique up to scalar multiplication . by definition and , again , by exactness , we have @xmath593 if @xmath594 then @xmath595 . proof .
let a canonical family of vertical arcs @xmath568 corresponding to the markovian sequence of partitions @xmath523 be chosen as above .
the condition @xmath596 precisely means the existence of constants @xmath597 such that for all @xmath598 we have @xmath599 now if @xmath600 then there exists @xmath21 such that for all sufficiently large n and all @xmath601 we have @xmath602 consequently , @xmath603 for some @xmath604 and all @xmath598 which is what we had to prove .
the scheme of the proof of the reverse inclusion can informally be summarized as follows .
we start with an equivariant sequence @xmath605 and we wish to recover a measure @xmath606 the equivariant sequence itself determines the values of the finitely - additive measure @xmath128 on all _ markovian _ arcs , that is , arcs going from the lower horizontal to the upper horizontal boundary of a rectangle of one of the proposition @xmath523 . to extend the measure @xmath128 from markovian arcs to all vertical arcs , we approximate an arbitrary arc by markovian ones ( similar approximation lemmas were used by forni @xcite and zorich @xcite ) .
the approximating series will be seen to converge because the number of terms at each stage of the approximation grows at most sub - exponentially , while the contribution of each term decays exponentially . for this argument to work ,
we assume that the norms of the matrices @xmath578 grow sub - exponentially . the measure @xmath128 is thus extended to all vertical arcs . to check the hlder property for @xmath128 ,
one needs additionally to assume that the heights of the markovian rectangles @xmath525 decay not faster than exponentially .
more precise oseledets - type assumptions on the sequence @xmath607 of adjacency matrices are used in order to obtain lower bounds on the hlder exponent for @xmath128 and to derive the logarithmic asymptotics of the growth of @xmath128 at infinity .
all our assumptions are verified for markov sequences of partitions induced by rauzy - veech expansions of zippered rectangles as soon as one uses the veech method of considering expansions corresponding to occurrences of a fixed renormalization matrix with positive entries .
a sequence @xmath576 of @xmath608-matrices will be called _ balanced _
if all entries of all matrices @xmath578 are positive , and , furthermore , there exists a positive constant c such that for any @xmath609 and any @xmath610 we have @xmath611 a sequence @xmath576 of @xmath608-matrices with non - negative entries will be said to have _ sub - exponential growth _ if for any @xmath56 there exists @xmath162 such that for all @xmath609 we have @xmath612 in order to formulate our next group of assumptions , we need to consider @xmath551-_reverse equivariant _ sequences of vectors . to a vector v@xmath613
we assign the sequence @xmath614 given by the formula @xmath615 by definition , if @xmath616 is an @xmath551-equivariant sequence , while @xmath616 is an @xmath551-reverse equivariant sequence , then the inner product @xmath617 does not depend on @xmath609 .
in analogy to the spaces @xmath618 and @xmath619 , we introduce the spaces @xmath620 the unique ergodicity of the vertical and the horizontal flow admits the following reformulation in terms of the spaces @xmath619 , @xmath621 .
[ a - u - e ] .
the space @xmath619 contains an equivariant sequence @xmath622 such that @xmath623 for all @xmath609 and all @xmath624 the space @xmath621 contains a reverse equivariant sequence @xmath625 such that @xmath626 for all @xmath609 and all @xmath624 the sequences @xmath627 and @xmath628 are unique up to scalar multiplication . a convenient normalization for us will be : @xmath629
our next assumption is the requirement of lyapunov regularity for the sequence of matrices @xmath630 for renormalization matrices of rauzy - veech expansions this assumption will be seen to hold by the oseledets theorem applied to the kontsevich - zorich cocycle .
in fact , we will assume the validity of all the statements of the oseledets - pesin reduction theorem ( theorem 3.5.5 on p.77 in @xcite ) .
we proceed to the precise formulation .
[ a - lyap - reg ] there exists @xmath631 positive numbers @xmath632 and , for any @xmath633 direct - sum decompositions @xmath634 @xmath635 such that the following holds . 1 . for all @xmath609 we have @xmath636 2 .
for all @xmath609 and all @xmath637 we have + @xmath638 3 . for all @xmath609 , every @xmath637 , and any @xmath639
, we have @xmath640 and the convergence is uniform on the sphere @xmath641 4 . for all @xmath609 , every @xmath637 , and any @xmath642 , we have @xmath643 and the convergence is uniform on the sphere @xmath644 5 . for all @xmath609 we have @xmath645 6 . for any @xmath56 there
exists @xmath162 such that for any @xmath609 and @xmath646 we have + @xmath647 + @xmath648 + @xmath649 + @xmath650 7 . for all @xmath609 we have @xmath651 and , for any @xmath637
we also have @xmath652 .
if @xmath653 satisfy @xmath654 then @xmath655 implies @xmath656 while @xmath657 implies @xmath658 and vice versa .
a balanced sequence @xmath551 of @xmath541-matrices with positive entries , having sub - exponential growth and satisfying the unique ergodicity assumption as well as the lyapunov regularity assumption , will be called a _ strongly biregular _ sequence , or , for brevity , an sb - sequence . using markovian sequences of partitions induced by rauzy - veech expansions of zippered rectangles corresponding to consecutive occurrences of a fixed renormalization matrix with positive entries and applying the oseledets multiplicative ergodic theorem and the oseledets - pesin reduction theorem ( see theorem 3.5.5 on p.77 in @xcite ) to the kontsevich - zorich cocycle , we will establish in the next section the following simple let @xmath103 be an ergodic probability measure on a connected component @xmath144 of the moduli space @xmath659 of abelian differentials .
then @xmath103-almost every abelian differential @xmath660 admits an exact markov sequence of partitions whose sequence of adjacency matrices belongs to the class sb .
let @xmath107 be an abelian differential whose horizontal and vertical foliations are both uniquely ergodic .
assume that @xmath107 is endowed with an exact markov sequence @xmath581 , of partitions into weakly admissible rectangles such that the corresponding sequence @xmath551 of adjacency matrices belongs to the class sb .
note that if @xmath551 is an sb - sequence , then @xmath661 @xmath662 note also that there exists a constant @xmath663 such that the positive equivariant sequence @xmath592 satisfies @xmath664 for all @xmath665 .
it follows that for any @xmath56 there exists a constant @xmath162 such that for all @xmath666 we have @xmath667 given a partition @xmath219 of our surface @xmath2 into weakly admissible rectangles @xmath668 we consider the semi - ring @xmath669 of arcs of the form @xmath563 $ ] , where @xmath670 for some @xmath671 ( recall here that @xmath672 stands for the lower horizontal boundary of @xmath673 for the upper horizontal boundary ) .
our markov sequence @xmath523 thus induces a sequence of semi - rings @xmath674 we write @xmath675 for the ring generated by the semi - ring @xmath676 .
elements of @xmath675 are finite unions of arcs from @xmath676 .
for an arc @xmath62 of the vertical flow , let @xmath677 be the largest by inclusion arc from the ring @xmath678 contained in @xmath62 , and let @xmath679 be the smallest by inclusion arc from the ring @xmath678 containing @xmath62 .
the following lemma is immediate from the definitions ( note that similar arc approximation lemmas were used by forni in @xcite and zorich in @xcite ) .
[ arc - dec - lem ] let @xmath523 be an exact markovian sequence of partitions such that the corresponding sequence @xmath551 of adjacency matrices has sub - exponential growth .
then for any @xmath56 there exists @xmath274 such that for any arc @xmath62 of the vertical flow and any @xmath277 the set @xmath680 consists of at most @xmath681 arcs from the semi - ring @xmath682 .
+ informally , lemma [ arc - dec - lem ] says that any arc of our symbolic flow is approximable by markovian arcs with sub - exponential error ; we illustrate this by figure [ fig : five ] .
we are now ready to identify @xmath561 and @xmath588 .
[ sb - bplus ] let @xmath581 , be a markov sequence of partitions such that the corresponding sequence @xmath551 of adjacency matrices belongs to the class sb .
then for every equivariant sequence @xmath683 there exists a unique finitely - additive measure @xmath684 such that @xmath685 proof .
indeed , the sequence @xmath582 itself prescribes the values of the @xmath128 on all markovian arcs @xmath686 for a general arc @xmath62 of the vertical flow , set @xmath687 where the existence of both limits and the equality of their values immediately follow from lemma [ arc - dec - lem ] .
finite - additivity of @xmath128 is again a corollary of lemma [ arc - dec - lem ] .
we have thus obtained a finitely - additive measure @xmath128 defined on all vertical arcs .
the uniqueness of such a measure is clear by ( [ approx - arcs ] ) .
the invariance of the resulting measure under horizontal holonomy is also clear by definition .
to conclude the proof of lemma [ sb - bplus ] , it remains to check that the obtained finitely - additive measure @xmath128 satisfies the hlder property .
for markovian arcs the hlder upper bound is clear from the _ upper _ exponential bound @xmath688 and the _ lower _ exponential bound @xmath689 for general arcs , the hlder property follows by lemma [ arc - dec - lem ] .
lemma [ sb - bplus ] is proved completely , and we have thus shown that under its assumptions the map @xmath690 is indeed an isomorphism . *
* under stronger assumptions of lyapunov regularity we will also give a hlder _ lower _ bound for the cocycles @xmath26 , see proposition [ convhoeldprop ] .
let @xmath691 and let @xmath692 be the corresponding finitely - additive measures .
the definitions directly imply @xmath693 duality between the spaces @xmath588 and @xmath694 follows now from the duality between the spaces @xmath561 and @xmath695 , which holds by the lyapunov regularity assumption for the sequence @xmath551 .
we start with a sequence of matrices @xmath696 satisfying the following [ asexpdec ] there exists @xmath336 and , for every @xmath697 , a direct - sum decomposition @xmath698 satisfying the following . 1
. @xmath699 and @xmath700 is injective .
2 . @xmath701 3 . for any @xmath56 there exists @xmath274 such that for any @xmath702 we have @xmath703 + @xmath704 [ vngeneral ] let @xmath578 be a sequence of matrices satisfying assumption [ asexpdec ] .
let @xmath705 be a sequence of vectors such that for any @xmath56 a constant @xmath162 can be chosen in such a way that for all @xmath557 we have @xmath706 then there exists a unique vector @xmath707 such that @xmath708 proof : denote @xmath709 and decompose @xmath710 , where @xmath711 , @xmath712 .
let @xmath713 @xmath714 we have @xmath715 , @xmath716 , @xmath717
. now introduce a vector @xmath718 by our assumptions , the series defining @xmath555 converges exponentially fast and , moreover , we have @xmath719 for some constant @xmath720 . since , by our assumptions we also have @xmath721 , the lemma is proved completely .
uniqueness of the vector @xmath555 follows from the fact that , by our assumptions , for any @xmath722 , @xmath723 we have @xmath724 let @xmath725 be a weakly lipschitz function , and , as before , introduce a canonical family of markovian curves @xmath726 corresponding to the markovian sequence of partitions @xmath581 .
introduce a family of vectors @xmath727 by setting @xmath728 by definition of the adjacency matrices @xmath729 using the weak lipschitz property of the function @xmath147 and sub - exponential growth of the sequence @xmath730 , we arrive for all @xmath277 at the estimate @xmath731 by lemma [ vngeneral ] , there exists a unique vector @xmath732 such that for all @xmath277 we have @xmath733 we let @xmath734 be the finitely - additive measure corresponding to the vector @xmath735 , or , in other words , the unique finitely additive measure in @xmath588 satisfying @xmath736 the inequality @xmath737)\right|\leqslant c_{\varepsilon}''||f||_{lip}(1+t^{\varepsilon})\ ] ] now holds for all @xmath738 indeed , if @xmath571 $ ] is a markovian arc , then ( [ a - apr ] ) is clear by definition of the vector @xmath735 and the weak lipschitz property of the function @xmath147 , while for general arcs of the vertical flow the inequality ( [ a - apr ] ) follows by lemma [ arc - dec - lem ] .
our next step is to check that for every @xmath739 we have @xmath740 where we recall that @xmath741 as before , let @xmath742 be a canonical system of markovian arcs of the vertical flow , corresponding to the markov sequence of partitions @xmath743 and let @xmath744 be a canonical system of markovian arcs of the horizontal flow corresponding to the markov sequence of partitions @xmath745 by definition , for any @xmath609 we have @xmath746 we now write the riemann sum @xmath747 for the measure @xmath47 and let @xmath557 tend to @xmath748 . by definition ,
we have @xmath749 now for all @xmath750 we have @xmath751 while , by lyapunov regularity , the quantity @xmath752 decays exponentially as @xmath753 it follows that @xmath754 which is what we had to prove .
as was noted above , we identify a finitely - additive measure @xmath755 with a continuous cocycle over the vertical flow for which , slightly abusing notation , we keep the same symbol @xmath756 the identification is given by the formula @xmath757).\ ] ] the hlder property of a finitely - additive measure is equivalent to the hlder property of the cocycle , that is , to the requirement that the function @xmath758 be hlder in @xmath305 uniformly in @xmath311 .
our next aim is to give hlder lower bounds for the cocycles @xmath759 and to investigate the growth of @xmath760 as @xmath761 .
consider the direct - sum decomposition @xmath762 and the corresponding direct - sum decomposition @xmath763 with @xmath764 of course , we have @xmath765 take @xmath766 and write @xmath767 with @xmath768 .
take the smallest @xmath671 such that @xmath769 the exponent @xmath172 will then be called _ the top lyapunov exponent _ of @xmath770 similarly , if @xmath771 is the largest number such that @xmath772 then @xmath773 will be called _ the lower lyapunov exponent _ of @xmath774 we shall now see that the top lyapunov exponent controls the growth of @xmath758 as @xmath775 while the lower lyapunov exponent describes the local hlder behaviour of @xmath776 [ convhoeldprop ] let @xmath777 let @xmath778 , and let @xmath12 be such that @xmath779 is defined for all @xmath780 then @xmath781 proof .
we first let @xmath305 tend to @xmath782 let @xmath783 be the equivariant sequence corresponding to @xmath770 we have @xmath784 and , consequently , for every @xmath56 there exists a constant @xmath785 and , for every @xmath786 there exists @xmath787 such that @xmath788 now let @xmath789 informally , @xmath790 is the first such moment that the are @xmath791 $ ] contains a markovian arc going all the way through the rectangle @xmath792 .
it is clear from the @xmath793-property of the sequence @xmath551 that for any @xmath56 there exist constants @xmath794 such that @xmath795 now denote @xmath796 since @xmath797)=\phi^{+}([x , x^{\prime}(n)])+\phi^{+}([x^{\prime}(n ) , x^{\prime \prime}(n)]),\ ] ] and @xmath798)=v^{(n)}_{i(n)},\ ] ] it follows from ( [ est - tn ] ) , ( [ est - vn ] ) that we have @xmath799)| } , \log{|\phi^{+}([x , x^{\prime \prime}(n)])|}\}}{n}= \frac{\theta_r}{\theta_1},\ ] ] whence also @xmath800)|}}{\log{\nu^{+}([x , x^{\prime}(n ) ] ) } } , \frac{\log { |\phi^{+}([x^{\prime}(n ) , x^{\prime \prime}(n)])|}}{\log{\nu^{+}([x , x^{\prime \prime}(n ) ] ) } } \right ) = \theta_r,\ ] ] and , finally @xmath801 : @xmath802 $ ] , @xmath803 $ ] .
either @xmath804 or @xmath805 satisfies the lower bound .
, title="fig : " ] + the desired lower bound is established .
we illustrate the argument in figure [ fig : one ] .
the proof for @xmath806 is completely similar , while the case @xmath807 is obtained by taking @xmath808 and repeating the same argument .
an @xmath793 sequence @xmath551 will be called _ hyperbolic _ if @xmath809 it is clear from the definitions that if an abelian differential @xmath810 admits an exact markovian sequence of partitions such that the corresponding sequence @xmath551 is hyperbolic , then @xmath811 in what follows , we shall check that if @xmath103 is probability measure on @xmath144 , invariant under the teichmller flow and ergodic , and such that the kontsevich - zorich cocycle acts isometrically on its neutral oseledets subspace , then @xmath103-almost every abelian differential @xmath810 admits a markovian sequence of partitions @xmath523 such that the corresponding sequence @xmath551 of adjacency matrices is a hyperbolic @xmath793 - sequence .
it will follow that for @xmath103 - almost all @xmath4 we have @xmath812 * remark . *
if @xmath813 is strictly larger than @xmath814 , it does not follow that @xmath815 is strictly larger than @xmath816 our constructions do not allow us to assign a finitely additive measure defined on all arcs of the vertical flow to a general equivariant sequence @xmath783 .
[ phiexpzero ] for any @xmath40 and any @xmath817 we have @xmath818 proof : since the proposition is clearly valid for @xmath819 , it suffices to prove it in the case @xmath820 .
but indeed , if @xmath821 , then the ergodic theorem implies @xmath822 and then @xmath823 .
[ phivarnotzero ] for any @xmath40 not proportional to @xmath28 and any @xmath824 we have @xmath825 taking @xmath826 instead of @xmath128 , we may assume @xmath827 .
if @xmath828 , then @xmath829 identically , but then @xmath830 whence @xmath831 , and the proposition is proved .
* remark . * in the context of substitutions , cocycles related to the hlder cocycles from @xmath0 have been studied by p. dumont , t. kamae and s. takahashi in @xcite as well as by t. kamae in @xcite .
the renormalization action of the teichmller flow on the spaces @xmath0 and @xmath32 of hlder cocycles will play a main rle in the proof of limit theorems for translation flows .
we will use veech s representation of abelian differentials by zippered rectangles , and in this section we recall veech s construction using the notation of @xcite , @xcite . for a different presentation of the rauzy - veech formalism , see marmi - moussa - yoccoz @xcite .
we start by recalling the definition of the rauzy - veech induction .
let @xmath219 be a permutation of @xmath522 symbols , which will always be assumed irreducible in the sense that @xmath832 implies @xmath222 .
the rauzy operations @xmath833 and @xmath834 are defined by the formulas @xmath835 @xmath836 these operations preserve irreducibility .
rauzy class _ @xmath837 is defined as the set of all permutations that can be obtained from @xmath219 by application of the transformation group generated by @xmath833 and @xmath834 . from now on we fix a rauzy class @xmath298 and assume that it consists of irreducible permutations . for @xmath838 ,
denote by @xmath839 the @xmath541 matrix whose @xmath840 entry is @xmath80 , while all others are zeros .
let @xmath841 be the identity @xmath541-matrix .
following veech @xcite , introduce the unimodular matrices @xmath842 @xmath843 for a vector @xmath844 , we write @xmath845 let @xmath846 one can identify each pair @xmath847 , @xmath848 , with the _ interval exchange map _ of the interval @xmath849 as follows .
divide @xmath234 into the sub - intervals @xmath850 , where @xmath851 , @xmath852 , @xmath853 , and then place the intervals @xmath854 in @xmath234 in the following order ( from left to write ) : @xmath855 .
we obtain a piecewise linear transformation of @xmath234 that preserves the lebesgue measure .
the space @xmath856 of interval exchange maps corresponding to @xmath298 is defined by @xmath857 denote @xmath858 @xmath859 @xmath860 @xmath861 the _ rauzy - veech induction map _
@xmath862 is defined as follows : @xmath863 one can check that @xmath864 is the interval exchange map induced by @xmath865 on the interval @xmath866 $ ] , where @xmath867 ; the interval @xmath868 is then stretched to unit length .
denote @xmath869 every @xmath870-invariant probability measure is concentrated on @xmath871 . on the other hand , a natural lebesgue measure defined on @xmath872 , which is finite , but non - invariant ,
is also concentrated on @xmath871 .
veech @xcite showed that @xmath873 has an absolutely continuous ergodic invariant measure on @xmath856 , which is , however , infinite .
we have two matrix cocycles @xmath874 , @xmath875over @xmath873 defined by @xmath876 @xmath877 we introduce the corresponding skew - product transformations @xmath878 , @xmath879 , @xmath880 @xmath881 here we briefly recall the construction of the veech space of zippered rectangles .
we use the notation of @xcite . _ zippered rectangles _ associated to the rauzy class @xmath298 are triples @xmath882 , where @xmath844 , @xmath883 , @xmath884 , @xmath885 , and the vector @xmath499 satisfies the following inequalities : @xmath886 @xmath887 the set of all vectors @xmath499 satisfying ( [ deltaone ] ) , ( [ deltatwo ] ) is a cone in @xmath888 ; we denote it by @xmath889 . for any @xmath890 , set @xmath891
given a zippered rectangle @xmath892 , veech @xcite takes @xmath522 rectangles @xmath893 of girth @xmath228 and height @xmath269 , @xmath890 , and glues them together according to a rule determined by the permutation @xmath219 .
this procedure yields a riemann surface @xmath2 endowed with a holomorphic @xmath80-form @xmath4 which , in restriction to each @xmath270 , is simply the form @xmath894 .
the union of the bases of the rectangles is an interval @xmath895 of length @xmath896 on @xmath2 ; the first return map of the vertical flow of the form @xmath4 is precisely the interval exchange @xmath232 .
a zippered rectangle @xmath897 by definition carries a partition @xmath898 of the underlying surface @xmath899 into @xmath522 weakly admissible rectangles@xmath270 : @xmath900 the _ area _ of a zippered rectangle @xmath882 is given by the expression @xmath901 ( our convention is @xmath902 when @xmath903 . )
furthermore , to each rectangle @xmath270 veech @xcite assigns a cycle @xmath904 in the homology group @xmath905 : namely , if @xmath906 is the left bottom corner of @xmath270 and @xmath907 the left top corner , then the cycle is the union of the vertical interval @xmath908 and the horizontal subinterval of @xmath895 joining @xmath907 to @xmath906 .
it is clear that the cycles @xmath904 span @xmath905 .
denote by @xmath909 the space of all zippered rectangles corresponding to the rauzy class @xmath298 , i.e. , @xmath910 let also @xmath911 @xmath912 @xmath913 veech @xcite introduced the flow @xmath914 acting on @xmath915 by the formula @xmath916 and the map @xmath917 , where @xmath918 ( the inclusion @xmath919 is proved in @xcite . )
the map @xmath920 and the flow @xmath914 commute on @xmath921 .
they also preserve the area of a zippered rectangle ( see ( [ area ] ) ) and hence can be restricted to the set @xmath922 for @xmath923 , denote @xmath924 from ( [ mat_a ] ) , ( [ mat_b ] ) it follows that if @xmath925 , then @xmath926 where @xmath927 when @xmath928 , and @xmath929 when @xmath930 .
next denote @xmath931 @xmath932 @xmath933 let @xmath934 @xmath935 clearly @xmath936 is well - defined on @xmath937 for all @xmath938 .
we now set @xmath939 the above identification enables us to define on @xmath940 a natural flow , for which we retain the notation @xmath941 ( although the bounded positive function @xmath384 is not separated from zero , the flow @xmath941 is well - defined ) .
note that for any @xmath307 we have a natural tautological " map @xmath942 which on each rectangle @xmath270 is simply expansion by @xmath943 in the horizontal direction and contraction by @xmath943 in the vertical direction . by definition ,
the map @xmath944 sends the vertical and the horizontal foliations of @xmath897 to those of @xmath945 .
introduce the space @xmath946 and endow the space @xmath947 with the flow @xmath948 given by the formula @xmath949 the flow @xmath941 induces on the transversal @xmath950 the first - return map @xmath951 given by the formula @xmath952 observe that , by definition , if @xmath953 , then @xmath954 .
for @xmath955 , @xmath307 , let @xmath956 be defined by the formula @xmath957 endow the space @xmath958 with a matrix cocycle @xmath959 over the flow @xmath941 given by the formula @xmath960 and introduce the corresponding skew - product flow @xmath961 by the formula @xmath962 we also have a natural cocycle @xmath963 over the inverse flow @xmath964 given by the formula @xmath965 and the natural skew - product flow @xmath966 defined by the formula @xmath967 the strongly unstable oseledets bundle of the cocycle @xmath963 will be seen to describe all the measures @xmath41 in the same way in which the strongly unstable oseledets bundle of the cocycle @xmath968 describes all the measures @xmath40 .
* the kontsevich - zorich cocycle is isomorphic to the inverse of its dual ( see , e.g. , statement 2 in proposition [ veechtokz ] below ) .
this `` self - duality '' is , however , not used in the construction and characterization of finitely - additive invariant measures .
the duality between the spaces @xmath0 and @xmath32 corresponds to the duality between the cocycle and its transpose , that is , in our notation , between @xmath968 and @xmath969 : such duality takes place for any invertible matrix - valued cocycle over any measure - preserving flow . to a connected component @xmath86 of the space @xmath78 one can assign a rauzy class @xmath298 in such a way that the following is true @xcite .
[ zipmodule ] there exists a finite - to - one measurable map @xmath970 such that @xmath971 .
the image of @xmath972 contains all abelian differentials whose vertical and horizontal foliations are both minimal . as before ,
let @xmath87 be the fibre bundle over @xmath86 whose fibre at a point @xmath79 is the cohomology group @xmath61 .
the kontsevich - zorich cocycle @xmath122 induces a skew - product flow @xmath973 on @xmath87 given by the formula @xmath974 following veech @xcite , we now explain the connection between the kontsevich - zorich cocycle @xmath122 and the cocycle @xmath975 . for any irreducible permutation
@xmath219 veech @xcite defines an alternating matrix @xmath976 by setting @xmath977 if @xmath978 or if @xmath979 , @xmath980 if @xmath981 , @xmath982 if @xmath983 and denotes by @xmath984 the kernel of @xmath976 and by @xmath985 the image of @xmath976 .
the dimensions of @xmath984 and @xmath986 do not change as @xmath219 varies in @xmath298 , and , furthermore , veech @xcite establishes the following properties of the spaces @xmath984 , @xmath986 .
[ veechtokz ] let @xmath927 or @xmath834
. then 1 .
@xmath987 , @xmath988 ; 2 .
the diagram @xmath989 is commutative and each arrow is an isomorphism .
3 . for each @xmath219 there exists a basis @xmath990 in @xmath984 such that the map @xmath991 sends every element of @xmath990 to an element of @xmath992 .
each space @xmath993 is thus endowed with a natural anti - symmetric bilinear form @xmath994 defined , for @xmath995 , by the formula @xmath996 ( the vector @xmath997 lies in @xmath998 ; since for all @xmath999 , @xmath1000 by definition we have @xmath1001 , the right - hand side is well - defined . ) consider the @xmath1002-invariant subbundle @xmath1003 given by the formula @xmath1004 as well as a quotient bundle @xmath1005 the bundle map @xmath1006 given by @xmath1007 induces a bundle isomorphism between @xmath1008 and @xmath1009 .
both bundles can be naturally lifted to bundles @xmath1010 , @xmath1011 over the space @xmath1012 of zippered rectangles ; they are naturally invariant under the corresponding skew - product flows @xmath1013 , @xmath1014 , and the map @xmath1015 lifts to a bundle isomorphism between @xmath1010 and @xmath1016 .
take @xmath1017 and write @xmath1018 .
veech @xcite has shown that the map @xmath972 lifts to a bundle epimorphism @xmath1019 from @xmath1010 onto @xmath87 that intertwines the cocycle @xmath975 and the kontsevich - zorich cocycle @xmath122 : [ identif - coc - veech ] for almost every @xmath1017 , @xmath892 , there exists an isomorphism @xmath1020 such that 1 .
the map @xmath1021 given by @xmath1022 induces a measurable bundle epimorphism from @xmath1010 onto @xmath87 , which is an isomorphic on each fibre ; 2 .
the diagram @xmath1023 is commutative ; 3 . for @xmath892 , the isomorphism @xmath1024 takes the bilinear form @xmath994 on @xmath986 , defined by ( [ bilpi ] ) , to the cup - product on @xmath1025 . proof : recall that to each rectangle @xmath270 veech @xcite assigns a cycle @xmath904 in the homology group @xmath905 : if @xmath906 is the left bottom corner of @xmath270 and @xmath907 the left top corner , then the cycle is the union of the vertical interval @xmath908 and the horizontal subinterval of @xmath895 joining @xmath907 to @xmath906 .
it is clear that the cycles @xmath904 span @xmath905 ; furthermore , veech shows that the cycle @xmath1026 is homologous to @xmath155 if and only if @xmath1027 .
we thus obtain an identification of @xmath998 and @xmath1028 .
similarly , the subspace of @xmath888 spanned by the vectors @xmath1029 , @xmath1030 , is precisely @xmath986 .
the identification of the bilinear form @xmath994 with the cup - product is established in proposition 4.19 in @xcite .
let @xmath1031 be an ergodic @xmath941-invariant probability measure for the flow @xmath941 on @xmath909 and let @xmath1032 be the corresponding @xmath82-invariant measure on @xmath86 .
let @xmath1033 be the strongly unstable bundle of the cocycle @xmath975 . by proposition [ veechtokz ] ,
the bundle @xmath1033 is a subbundle of @xmath1010 .
it therefore follows from proposition [ identif - coc - veech ] that the map @xmath1019 isomorphically identifies the strongly unstable bundles of the cocycles @xmath1034 and @xmath1035 ; this identification is equivariant with respect to the natural actions of the skew - product flows @xmath1013 and @xmath973 on the corresponding bundles .
[ cor - mes ] let @xmath89 be an ergodic @xmath82invariant probability measure on @xmath86
. then there exists an ergodic @xmath941-invariant probability measure @xmath1031 on @xmath909 such that @xmath1036 this proposition is a corollary of the following general statement . [ finfibre ] let @xmath1037 be standard borel spaces , let @xmath1038 , @xmath1039 be measurable flows , and let @xmath1040 be a borel measurable map such that 1 . for any @xmath1041 the preimage @xmath1042 of @xmath1043 is finite ; 2 . the map @xmath1044 intertwines the flows @xmath1045 , @xmath1046 in the sense that the diagram @xmath1047 is commutative .
then for any borel @xmath1048-invariant ergodic probability measure @xmath1049 on @xmath1050 , there exists a borel @xmath1051-invariant ergodic probability measure @xmath1052 on @xmath1053 such that @xmath1054 the proof of proposition [ finfibre ] is routine : first , note that , by ergodicity , for @xmath1049-almost every @xmath1055 the cardinality of the preimage @xmath1042 of @xmath1043 is constant ; now consider the normalized product @xmath1056 of @xmath1049 and the counting measure in the preimage ; the measure @xmath1056 is by definition @xmath1051-invariant , and for the measure @xmath1052 one may take an ergodic component , in fact , almost every ergodic component , of the measure @xmath1056 . given a zippered rectangle @xmath897 , we shall speak of its vertical and horizontal foliations , hlder cocycles and so on , meaning the corresponding objects for the underlying abelian differential , and we shall use the notation @xmath1057 , @xmath1058 , @xmath1059 , @xmath1060 for the corresponding spaces of finitely - additive measures .
recall that , by construction , a zippered rectangle carries the partition @xmath1061 into weakly admissible rectangles .
the rauzy
veech expansion of a zippered rectangle now yields a markovian sequence of partitions @xmath523
, @xmath609 . to construct it ,
first take @xmath1062 and recall that we have a natural `` tautological '' identification map @xmath1063 now set @xmath1064 by definition , the sequence @xmath1065 , @xmath609 , is markovian .
minimality of the horizontal and vertical flows implies exactness of the sequence @xmath1065 . for a general zippered rectangle @xmath1066 ,
write @xmath1067 and set @xmath1068 ( informally , carry over sequence @xmath523 from the `` closest '' zippered rectangle lying on the transversal @xmath1069 . )
[ sb - zip ] for @xmath1070-almost every zippered rectangle @xmath897 there exists a sequence @xmath1071 , @xmath1072 , @xmath1073 , @xmath1074 , such that the sequence @xmath1075 of adjacency matrices of the exact markovian subsequence @xmath1076 , @xmath1074 satisfies the following . 1 .
@xmath551 is an sb - sequence , 2 .
the space @xmath561 coincides with the strongly unstable space of the cocycle @xmath968 at the point @xmath897 .
the proof of the lemma is routine : one chooses a rauzy - veech matrix @xmath1077 of the form @xmath1078 where @xmath1079 and @xmath1080 are rauzy - veech matrices all whose entries are positive and such that @xmath1070-almost all zippered rectangles @xmath897 contain infinitely many occurrences of the matrix @xmath1077 both in the past and in the future .
the sequence @xmath1081 is then the sequence of consecutive occurrences of the matrix @xmath1077 .
each adjacency matrix @xmath1082 now has the form @xmath1083 , where @xmath1084 is an integer matrix with non - negative entries .
it follows from the oseledets multiplicative ergodic theorem and the oseledets - pesin reduction theorem ( theorem 3.5.5 on p.77 in @xcite ) that @xmath551 is an sb - sequence and that @xmath561 coincides with the strongly unstable space of the cocycle @xmath968 at the zippered rectangle @xmath897 .
the proof of the lemma is complete .
we have the evaluation map @xmath1085 which to a finitely - additive measure @xmath111 assigns the vector of its values on vertical arcs of the rectangles @xmath1086 , @xmath1087 .
we must now check that the map @xmath1088 is indeed an isomorphism between the space @xmath1057 and the strongly unstable space of the cocycle @xmath1089 . introduce a measurable fibre bundle @xmath1090 over the veech space @xmath1091 by setting @xmath1092
extend the map @xmath1088 to a bundle morphism @xmath1093 given by the formula : @xmath1094 by definition , the map @xmath1095 intertwines the action of the flow @xmath1096 on the bundle @xmath1090 with that of the flow @xmath1097 on the trivial bundle @xmath1098 . recall that for any @xmath307 we have a natural tautological " map @xmath1099 given by ( [ def - t - s ] ) . the bundle @xmath1100 is now endowed with a natural renormalization flow @xmath1101 given by the formula
@xmath1102 we furthermore have a bundle morphism @xmath1103 given by the formula @xmath1104 the identification of cocycles now gives us the following [ hypzip ] let @xmath1031 be an ergodic @xmath941-invariant probability measure for the flow @xmath941 on @xmath909 .
we have a commutative diagram @xmath1105 the map @xmath1095 is injective in restriction each fibre . for @xmath1031-almost
every @xmath1106 , the map @xmath1095 induces an isomorphism between the space @xmath1107 and the strongly unstable oseledets subspace of the cocycle @xmath1108 at the point @xmath897 . proof .
let @xmath1109 be the sequence of partitions given by lemma [ sb - zip ] , and let @xmath607 be the corresponding sb - sequence of matrices . since @xmath551 is an sb - sequence , the map @xmath1088 induces an isomorphism between @xmath1057 and @xmath561 ( recall here that @xmath1072 ) .
since @xmath561 coincides with the unstable space of the cocycle @xmath968 , the proposition is proved completely . using proposition [ hypzip ]
, we will identify the action of @xmath1101 on @xmath1110 with the action of @xmath1013 on the strongly unstable oseledets subbundle of @xmath1111 , and speak of the action of the cocycle @xmath975 on the space of finitely - additive measures in this sense .
this renormalization action of the flow @xmath941 on the space of finitely - additive measures will play a key rle in the proof of the limit theorems in the next section .
we close this section by giving a sufficient condition for the equality @xmath1112 .
let @xmath1113 be a probability space endowed with a @xmath296-preserving transformation @xmath1114 or flow @xmath115 and an integrable linear cocycle @xmath514 over @xmath115 with values in @xmath1115 .
for @xmath1116 let @xmath1117 be the the _ neutral _ subspace of @xmath514 at @xmath1118 , i.e. , the lyapunov subspace of the cocycle @xmath1119 corresponding to the lyapunov exponent @xmath155 .
we say that @xmath1119 _ acts isometrically on its neutral subspaces _ if for almost any @xmath1118 there exists an inner product @xmath1120 on @xmath888 which depends on @xmath1118 measurably and satisfies @xmath1121 for all @xmath307 ( again , in the case of a transformation , @xmath115 should be replaced by @xmath1114 in this formula ) .
the third statement of proposition [ veechtokz ] has the following immediate [ hypzip1 ] let @xmath1031 be a borel ergodic @xmath941-invariant probability measure on @xmath909 and let @xmath1122 be the corresponding @xmath82-invariant measure on @xmath86 .
if the kontsevich - zorich cocycle acts isometrically on its neutral subspace with respect to @xmath89 , then the cocycle @xmath975 also acts isometrically on its neutral subspace with respect to @xmath1123 .
note that the hypothesis of corollary [ hypzip1 ] is satisfied , in particular , for the masur - veech smooth measure on the moduli space of abelian differentials .
the following proposition is clear from the definitions .
[ hyperboliccriterion ] let @xmath1031 be a borel ergodic @xmath941-invariant probability measure on @xmath909 such that the cocycle @xmath975 acts isometrically on its neutral subspace with respect to @xmath1123 .
let @xmath1122 be the corresponding @xmath82-invariant ergodic measure on @xmath86 . then for @xmath89-almost every abelian differential @xmath79
, we have the equality @xmath1124 in other words , if the cocycle @xmath975 acts isometrically on its neutral subspace with respect to @xmath1123 , then any continuous finitely - additive measure must in fact be hlder .
note that the assumptions of the proposition are verified , in particular , for the masur - veech smooth measure on the moduli space of abelian differentials . to prove proposition [ hyperboliccriterion ] we use proposition [ veechtokz ] , which implies that if the cocycle @xmath975 acts isometrically on its neutral subspace with respect to @xmath1123 , then @xmath89-almost every abelian differential @xmath79 admits an exact markovian sequence of partitions whose sequence of adjacency matrices is a hyperbolic sb - sequence : which , in turn , is sufficient for the equality @xmath1124
the main element in the proof of the limit theorem is the renormalization action of the teichmller flow @xmath941 on the bundle @xmath1125 .
start with the case when the second lyapunov exponent of the cocycle @xmath975 is positive and simple with respect to a @xmath941-invariant ergodic probability measure @xmath1126 on @xmath1127 .
then , by theorem [ multiplicmoduli ] , for @xmath1126-almost every zippered rectangle @xmath897 , and a generic weakly lipschitz function @xmath147 of zero average the ergodic integral @xmath1128 is approximated by an expression of the form @xmath1129 where the constant depends on @xmath147 , and @xmath1130 is a cocycle belonging to the second lyapunov subspace of the cocycle @xmath975 .
note that the cocycle @xmath1131 is defined up to multiplication by a scalar ; the double cover @xmath366 over the space @xmath86 in the formulation of the limit theorem is considered precisely in order to distinguish between positive and negative scalars .
now , proposition [ hypzip ] implies that the normalized distribution of the random variable @xmath1132 ( considered as a function of @xmath311 with fixed @xmath310 ) coincides with the normalized distribution of the random variable @xmath1133 .
assigning to a zippered rectangle @xmath897 the normalized distribution of the random variable @xmath1132 ( considered as a function of @xmath311 with fixed @xmath310 ) now yields the desired map @xmath1134 from the space of zippered rectangles ( more precisely , from its double cover ) to the space of distributions .
the fact that the normalized distributions of the ergodic integrals are approximated by the image under the map @xmath1134 of the orbit of our zippered rectangle under the action of the teichmller flow @xmath941 follows now from the asymptotic expansion of theorem [ multiplicmoduli ] .
we fix a @xmath941-invariant ergodic probability measure @xmath1126 on @xmath1127 and start with the case in which the second lyapunov exponent of the cocycle @xmath975 is positive and simple with respect to the measure @xmath1126 .
consider the oseledets subspace @xmath1135 corresponding to the top lyapunov exponent @xmath80 and the one - dimensional oseledets subspace @xmath1136 corresponding to the second lyapunov exponent .
furthermore , let @xmath1137 be the subspace corresponding to the remaining lyapunov exponents .
we have then the decomposition @xmath1138 denote @xmath1139 , @xmath1140 , @xmath1141 the corresponding spaces of hlder cocycles .
a similar decomposition holds for the dual space @xmath1142 , the strongly unstable space of the cocycle @xmath963 : @xmath1143 again , denote @xmath1144 , @xmath1145 , @xmath1146 the corresponding spaces of hlder cocycles .
choose @xmath1147 , @xmath1148 in such a way that @xmath1149 take @xmath1150 , @xmath1151 , @xmath1152 and observe that the expression @xmath1153 does not depend on the precise choice of @xmath1154 ( we have the freedom of multiplying @xmath1155 by an arbitrary scalar , but then @xmath1156 is divided by the same scalar ) .
now for @xmath1150 write @xmath1157 where @xmath1158 . in particular ,
there exist two positive constants @xmath148 and @xmath398 depending only on @xmath89 such that for any function @xmath147 satisfying @xmath1159 we have the estimate @xmath1160 in order to estimate the variance of the random variable @xmath1161 , we start by studying the growth of the variance of the random variable @xmath1162 as @xmath1163 .
recall that @xmath1164 for all @xmath1114 , while @xmath1165 for @xmath1166 .
recall that for a cocycle @xmath1167 , @xmath1168 , we have defined its norm @xmath1169 by the formula @xmath1170 .
introduce a multiplicative cocycle @xmath1171 over the flow @xmath941 by the formula @xmath1172 observe that the right - hand side does not depend on the specific choice of @xmath126 .
by definition , we now have @xmath1173 there exists a positive measurable function @xmath1174 such that the following equality holds for @xmath1126-almost all @xmath1175 : @xmath1176 indeed , the function @xmath1177 is given by @xmath1178 and the proposition is an immediate corollary of proposition [ hypzip ] .
observe that the right - hand side does not depend on a particular choice of @xmath1179 , @xmath352 . using ( [ phitwoapprox ] ) , we now proceed to estimating the growth of the variance of the ergodic integral @xmath1180 we use the same notation as in the introduction : for @xmath306 $ ] , @xmath307 , a real - valued @xmath1181 we write @xmath1182=\int_0^{\tau\exp(s ) } f\circ h^{+}_t(x)dt.\ ] ] as before , let @xmath11 be the lebesgue measure on the surface @xmath1183 corresponding to the zippered rectangle @xmath897 .
as before , as @xmath311 varies in the probability space @xmath1184 , we obtain a random element of @xmath314 $ ] . in other words
, we have a random variable @xmath315 : ( m({\mathscr x } ) , \nu)\to c[0,1]\ ] ] defined by the formula ( [ sfstaux - symb ] ) . for any fixed @xmath306 $ ] the formula ( [ sfstaux - symb ] )
yields a real - valued random variable @xmath316 : ( m({\mathscr x } ) , \nu)\to { \mathbb r},\ ] ] whose expectation , by definition , is zero .
[ varf ] there exists @xmath336 depending only on @xmath1123 and a positive measurable function @xmath1185 such that the following holds for @xmath1123-almost all @xmath1175 and all @xmath444 .
let @xmath1186 , @xmath1187 be chosen in such a way that @xmath1188 .
let @xmath1189 be such that @xmath1190 then @xmath1191}{v(p^s{{\mathscr x } } ) ( m_{\phi_2 ^ -}(f)|\phi^+_2|h_2(s , { { \mathscr x}}))^2}-1\right|\leq c({{\mathscr x } } , p^s{{\mathscr x}})\exp(-\alpha s).\ ] ] * remark . *
observe that the quantity @xmath1192 does not depend on the specific choice of @xmath1193 , @xmath1194 such that @xmath1195 . indeed , the proposition is immediate from theorem [ multiplicmoduli ] , the inequality @xmath1196 which holds for any two bounded random variables @xmath1197 on any probability space , and the following clear proposition , which , again , is an immediate corollary of theorem [ multiplicmoduli ] .
there exists a constant @xmath336 depending only on @xmath1126 , a positive measurable function @xmath1185 and a positive measurable function @xmath1198 such that for all @xmath444 we have @xmath1199 @xmath1200(x)}{v^{\prime}(p^s{{\mathscr x } } ) ( m_{\phi_2 ^ -}(f)|\phi^+|h_2(s , { { \mathscr x}}))^2}-1\right|\leq c({{\mathscr x } } , p^s{{\mathscr x}})\exp(-\alpha s).\ ] ] we now turn to the asymptotic behaviour of the distribution of the random variable @xmath353 $ ] as @xmath317 .
again , we will use the notation @xmath354 $ ] for the distribution of the normalized random variable @xmath355}{{\sqrt{var_{\nu } { \mathfrak s}[f , s;1]}}}.\ ] ] the measure @xmath354 $ ] is thus a probability distribution on the space @xmath314 $ ] of continuous functions on the unit interval . for @xmath356 , @xmath357
, we again let @xmath358 $ ] be the distribution of the @xmath50-valued random variable @xmath359}{{\sqrt{var_{\nu } { \mathfrak s}[f , s ; \tau]}}}.\ ] ] if @xmath147 has zero average , then , by definition , @xmath358 $ ] is a measure on @xmath50 of expectation @xmath155 and variance @xmath80 . again , as in the introduction , we take the space @xmath314 $ ] of continuous functions on the unit interval endowed with the tchebyshev topology , we let @xmath375 be the space of borel probability measures on the space @xmath314 $ ] endowed with the weak topology ( see @xcite or the appendix ) .
consider the space @xmath1201 given by the formula @xmath1202 the flow @xmath941 is lifted to @xmath1201 by the formula @xmath1203 given @xmath1204 , @xmath1205 , write @xmath1206 as before , write @xmath1207 now introduce the map @xmath1208 by setting @xmath1209 to be the distribution of the @xmath314$]-valued normalized random variable @xmath1210.\ ] ] note here that , by proposition [ phivarnotzero ] , for any @xmath1211 we have @xmath1212 , so , by definition , we have @xmath1213 .
now , as before , we take a function @xmath1214 such that @xmath1215 as before , @xmath394 stands for the lvy - prohorov metric on @xmath375 , @xmath395 for the kantorovich - rubinstein metric on @xmath375 .
[ limthmmarkcomp - simple ] let @xmath1126 be a @xmath941-invariant ergodic borel probability measure on @xmath1127 such that the second lyapunov exponent of the cocycle @xmath1034 is positive and simple with respect to @xmath1126 .
there exists a positive measurable function @xmath1216 and a positive constant @xmath398 depending only on @xmath1126 such that for @xmath1126-almost every @xmath1217 , @xmath1205 , and any @xmath1181 satisfying @xmath1218 we have @xmath1219 , { \mathcal d}_2^+(p^{s , \prime}{{\mathscr x}}^{\prime}))\leq c({{\mathscr x } } , p^s{{\mathscr x}})\exp(-\alpha s).\ ] ] @xmath1220 , { \mathcal d}_2^+(p^{s,\prime}{{\mathscr x}}^{\prime}))\leq c({{\mathscr x } } , p^s{{\mathscr x}})\exp(-\alpha s).\ ] ] proof : we start with the simple inequality @xmath1221 . for any pair of random variables @xmath1197 taking values in an arbitrary banach space and any positive real numbers
we consequently have @xmath1223 we apply the inequality ( [ ineqlimthm ] ) to the @xmath314$]-valued random variables @xmath1224 , \
\xi_2=\phi_{2 , p^s{{\mathscr x}}}^+(x,\tau\cdot e^s),\ ] ] letting @xmath1225 , @xmath1226 be the corresponding normalizing variances : @xmath1227 $ ] , @xmath1228 $ ] .
now take @xmath56 and let @xmath1229 be two random variables on an arbitrary probability space @xmath1230 taking values in a complete metric space and such that the distance between their values does not exceed @xmath510 . in this case
both the lvy - prohorov and the kantorovich - rubinstein distance between their distributions @xmath1231
, @xmath1232 also does not exceed @xmath510 ( see lemma [ dist - images ] in the appendix ) .
proposition [ limthmmarkcomp - simple ] is now immediate from equation ( [ phitwoapprox ] ) and proposition [ varf ] .
it remains to derive proposition [ limthmmoduli - simple ] from proposition [ limthmmarkcomp - simple ] .
to do so , note that the map @xmath1134 , originally defined on the double cover @xmath1233 of the space of zippered rectangles , naturally descends to a map , for which we keep the same symbol @xmath1134 , defined on the double cover @xmath366 of the connected component @xmath86 of the moduli space of abelian differentials .
indeed , it is immediate from the definitions that the image @xmath1234 of an element @xmath1235 , @xmath1236 only depends on the underlying element @xmath1237 of the space @xmath366 .
proposition [ limthmmarkcomp - simple ] is now proved completely .
for @xmath1217 , @xmath1238 let @xmath1239 $ ] be the distribution of the normalized @xmath50-valued random variable @xmath1240 let @xmath1126 be a @xmath941-invariant ergodic borel probability measure on @xmath1127 .
for @xmath1126-almost every @xmath1241 and any @xmath1242 , @xmath463 , the correspondence @xmath1243\ ] ] yields a continuous map from @xmath1244 to @xmath1245 . proof .
this is immediate from the hlder property of the cocycle @xmath128 and the nonvanishing of the variance @xmath1246 for @xmath357 , which is guaranteed by proposition [ phivarnotzero ] . as usual , by the
omega - limit set _ of a parameterized curve @xmath1247 , @xmath307 , taking values in a metric space , we mean the set of all accumulation points of our curve as @xmath317 .
we now use the following general statement .
[ omlimset ] let @xmath1248 be a standard borel space , and let @xmath115 be a measurable flow on @xmath1249 preserving an ergodic borel probability measure @xmath296 .
let @xmath1250 be a separable metric space , and let @xmath1251be a measurable map such that for @xmath296-almost every @xmath1252 the curve @xmath1253 is continuous in @xmath1254 .
then there exists a closed set @xmath1255 such that for @xmath296-almost every @xmath1252 the set @xmath1256 is the omega - limit set of the curve @xmath1253 , @xmath1254 .
the proof of proposition [ omlimset ] is routine .
we choose a countable base @xmath1257 of open sets in @xmath1250 . by ergodicity of @xmath115 , continuity of the curves @xmath1253 and countability of the family @xmath1258 , there exists a subset of full measure @xmath1259 , @xmath1260 , such that for any @xmath1261 and any @xmath1262 the following conditions are satisfied : 1 .
if @xmath1263 , then there exists an infinite sequence @xmath1264 such that @xmath1265 ; 2 . if @xmath1266 , then there exists @xmath1267 such that @xmath1268 for all @xmath1269
. now let @xmath1256 be the set of all points @xmath1270 such that @xmath1263 for any open set @xmath1261 containing the point @xmath1271 . by construction , for any @xmath1262
, the set @xmath1256 is precisely the omega - limit set of the curve @xmath1253 .
the proposition is proved .
proposition [ omlimset ] with @xmath1272 , @xmath1273 and @xmath296 an ergodic component of @xmath135 together with the limit theorem given by propositions [ limthmmoduli - simple ] , [ limthmmarkcomp - simple ] immediately implies corollary [ nonentwo ] .
let @xmath1276 be an ergodic @xmath941-invariant probability measure on @xmath1277 and let @xmath1278 be the distinct positive lyapunov exponents of @xmath968 with respect to @xmath103 .
we assume @xmath420 . for @xmath1279 let @xmath1280 be the corresponding direct - sum decomposition into oseledets subspaces , and
let @xmath1281 be the corresponding direct sum decomposition of the space @xmath1282 for @xmath1283 we now write @xmath425 where @xmath1284 and , of course , @xmath1285 for each @xmath428 introduce a measurable fibre bundle @xmath1286 the flow @xmath941 is naturally lifted to the space @xmath1274 by the formula @xmath1287 the growth of the norm of vectors @xmath432 is controlled by the multiplicative cocycle @xmath433 over the flow @xmath1288 defined by the formula @xmath1289 the growth of the variance of ergodic integrals is also , similarly to the previous case , described by the cocycle @xmath1290 for @xmath1175 and @xmath1291 we write @xmath1292 we now define a vector @xmath1293 by the formula @xmath1294 there exists @xmath336 depending only on @xmath1276 and , for any @xmath441 positive measurable functions @xmath1295 such that for @xmath1276-almost every @xmath1175 , any @xmath1291 and all @xmath444 we have @xmath1296 ) } { v^{(i(f))}(p^{s,{\bf{s}}^{(i)}}({{\mathscr x}},v_{f}))(h_i(s , ( { { \mathscr x}},v_{f } ) ) ) ^2}-1\right|\leqslant c^{(i)}({{\mathscr x}},p^s{{\mathscr x}})e^{-\alpha s}.\ ] ] indeed , similarly to the case of a simple lyapunov exponent , for @xmath1297 we write @xmath1298 and set @xmath1299 the proposition follows now in the same way as in the case of the simple second lyapunov exponent : the pointwise approximation of the ergodic integral by the corresponding hlder cocycle implies also that the variances of these random variables are exponentially close .
for @xmath441 introduce a map @xmath1300 by setting @xmath1301 to be the distribution of the @xmath314$]-valued random variable @xmath1302.\ ] ] as before , by definition we have @xmath1303 the measure @xmath450\in { { \mathfrak m}}$ ] is , as before , the distribution of the @xmath314$]-valued random variable @xmath1304.\ ] ] as before , let @xmath1305 be the number of distinct positive lyapunov exponents of the measure @xmath1276 . for @xmath1291
we define the number @xmath1306 by ( [ defif ] ) and the vector @xmath1307 by ( [ defvf ] ) .
[ limthmmarkcomp ] let @xmath1276 be a borel @xmath941-invariant ergodic probability measure on @xmath1277 satisfying @xmath1308 .
there exists a constant @xmath336 depending only on @xmath103 and a positive measurable map @xmath1309 such that for @xmath1276-almost every @xmath1175 and any @xmath1291 we have @xmath454,d_{i(f)}^+(p^{s , { \bf{s}}^{(i(f))}}({{\mathscr x}},v_f ) ) ) \leqslant c({{\mathscr x}},p^s{{\mathscr x}})e^{-\alpha s},\ ] ] @xmath455,d_{i(f)}^+(p^{s,{\bf{s}}^{(i(f))}}({{\mathscr x}},v_f ) ) ) \leqslant c({{\mathscr x}},p^s{{\mathscr x}})e^{-\alpha s}.\ ] ] the proof is similar to the proof of proposition [ limthmmarkcomp - simple ] .
again , the ergodic integral is uniformly approximated by the corresponding cocycle ; the uniform bound on the difference yields the uniform bound on the difference and the ratio of variances of the ergodic integral and the cocycle considered as random variables ; we proceed , as before , by using the inequality ( [ ineqlimthm ] ) with @xmath1310,\ \xi_2=\phi^+_{f , i(f)}(x , \tau)$ ] , and @xmath1311 the corresponding normalizing variances .
we conclude , again , by noting that a uniform bound on the difference between two random variables implies the same bound on the lvy - prohorov or kantorovich - rubinstein distance between the distributions of the random variables ( using lemma [ dist - images ] in the appendix ) .
theorem [ limthmmarkcomp ] now implies theorem [ limthmmoduli ] in the same way in which proposition [ limthmmarkcomp - simple ] implies proposition [ limthmmoduli - simple ] .
let @xmath897 be a zippered rectangle , and let @xmath1312 be the underlying abelian differential .
for @xmath1313 , let @xmath1314 stand for the leaf of the vertical foliation containing @xmath311 , and let @xmath1315 stand for the leaf of the horizontal foliation containing @xmath311 .
our next aim is to show that atoms of limit distributions occur at all `` homoclinic times '' , that is , moments of time @xmath1316 such that there exists a point @xmath1317 satisfying @xmath1318 . [ atom - eq - pr ] let @xmath897 be a zippered rectangle , and let @xmath1312 be the underlying abelian differential .
let @xmath1317 and assume that @xmath1319 does lies neither on a horizontal nor on a vertical leaf passing through a singularity of the abelian differential @xmath1320 .
let @xmath817 be such that @xmath1321 .
then there exists a rectangle @xmath152 of positive area such that for any @xmath1322 and any @xmath1323 we have @xmath1324 proof .
let @xmath1325 and write @xmath1326
. start with the case @xmath1327 . by our assumptions , for sufficiently small positive @xmath1328 , @xmath1329 , the rectangles @xmath1330 are both admissible .
the desired rectangle @xmath152 can now be taken of the form @xmath1331 indeed , take @xmath1322 .
our aim is to check the equality ( [ atomeq ] ) .
write @xmath1332 , where @xmath1333 .
we first check the equality @xmath1334 but indeed , @xmath1335 since @xmath1336 is admissible , whence @xmath1337 as desired .
the equality @xmath1338 is a direct corollary of admissibility of @xmath1339 . combining ( [ atomeq2 ] ) with ( [ atomeq3 ] )
, we arrive at the desired equality ( [ atomeq2 ] ) , and proposition [ atom - eq - pr ] is proved . for a fixed zippered rectangle @xmath1241 both whose vertical and horizontal flows are minimal , the set of `` homoclinic times '' @xmath1316 for which there exist @xmath1340 satisfying @xmath1341 , @xmath1342 , @xmath1343 , is countable and dense in @xmath50 .
proposition [ atom - eq - pr ] now implies the following let @xmath1276 be a borel @xmath941-invariant ergodic probability measure on @xmath1277 . for @xmath1123-almost
every @xmath1175 , there exists a dense set of times @xmath817 such that for any @xmath1344 the distribution of the random variable @xmath1345 has an atom .
our next step is to show that atoms of weight arbitrarily close to @xmath80 occur for limit distributions of our hlder cocycles .
informally , such atoms exist when one admissible rectangle occupies most of our surface .
more precisely , we have the following [ bigatom ] let @xmath1175 satisfy @xmath1346 .
then there exists a set @xmath1347 such that 1 .
@xmath1348 ; 2 . for any @xmath1349 ,
the function @xmath1350 is constant on @xmath1351 .
we consider @xmath1241 fixed and omit it from notation .
consider the partition @xmath1352 of the zippered rectangle @xmath897 .
let @xmath854 be the interval forming lower horizontal boundaries of the rectangles @xmath1353 , @xmath1354 , and set @xmath1355 the flow transversal @xmath234 carries the lebesgue measure @xmath1356 invariant under the first - return map of the flow @xmath7 on @xmath234 .
we recall that the first return map is simply the interval exchange transformation @xmath865 of the zippered rectangle @xmath892 .
we recall that @xmath1357 is the length of @xmath854 , and that @xmath1358 is the height of @xmath1353 . for brevity , denote @xmath1359 . by definition , @xmath1360 and we have @xmath1361 introduce the set @xmath1362 the first statement of the proposition is clear , and we proceed to the proof of the second .
note first that for any @xmath1349 and any @xmath1363 the quantity @xmath1364 is constant as long as @xmath311 varies in @xmath1365 .
fix @xmath1349 and take an arbitrary @xmath1366 . write @xmath1367 , where @xmath1368 , @xmath1369 .
we have @xmath1370 , whence @xmath1371 and @xmath1372 which concludes the proof of the proposition .
+ we illustrate the proof by figure [ fig : two ] .
recall that for @xmath1217 , @xmath1238 , @xmath463 , and @xmath356 , @xmath357 , the measure @xmath1239 $ ] is the distribution of the normalized @xmath50-valued random variable @xmath1240 as before , let @xmath1245 be the space of probability measures on @xmath50 endowed with the weak topology , and let @xmath1373 stand for the delta - measure at zero . similarly to the introduction , we need the following additional assumption on our @xmath941-invariant ergodic probability measure @xmath1276 on @xmath1277 .
[ fatzip ] for any @xmath56 we have @xmath1374 by proposition [ bigatom ] , in view of the ergodicity of @xmath1123 , for almost every @xmath1175 and every @xmath1238 , @xmath463 , the sequence of measures @xmath1239 $ ] admits atoms of weight arbitrarily close to @xmath80 .
the next simple proposition shows that the corresponding measures must then accumulate at _ zero _ ( rather than at another point of the real line ) .
[ atomatzero ] let @xmath1375 be a probability measure on @xmath458 such that @xmath1376 let @xmath1377 and assume that @xmath1378 then @xmath1379 proof . if @xmath1380 , then there is nothing to prove , so assume @xmath1381 ( the remaining case @xmath1382 follows by symmetry ) .
we have @xmath1383 and , consequently , @xmath1384 using the cauchy - bunyakovsky - schwarz inequality , write @xmath1385 whence , recalling that the variance of @xmath1375 is equal to @xmath80 , we obtain @xmath1386 and , finally , @xmath1387 which is what we had to prove . as before ,
the symbol @xmath466 denotes weak convergence of probability measures .
[ convtodelta ] let @xmath1276 be a borel ergodic @xmath941-invariant probability measure on @xmath1388 satisfying assumption [ fatzip ] .
then for @xmath1276-almost every @xmath1389 there exists a sequence @xmath471 such that for any @xmath1390 we have @xmath1391\rightarrow\delta_0\ \mathrm{as}\ n\to\infty.\ ] ] this is immediate from proposition [ bigatom ] and proposition [ atomatzero ] .
[ nonconvergence ] let @xmath1276 be a borel ergodic @xmath941-invariant probability measure on @xmath1388 satisfying assumption [ fatzip ] .
then for @xmath1276-almost every @xmath1389 there exists a sequence @xmath474 such that for any @xmath1392 satisfying @xmath1393 we have @xmath1394\rightarrow\delta_0\ \mathrm{as}\ n\to\infty.\ ] ] consequently , if @xmath1395 satisfies @xmath478 then the family of measures @xmath1396 $ ] does not converge in the weak topology on @xmath1245 as @xmath317 and the family of measures @xmath450 $ ] does not converge in the weak topology on @xmath1397)$ ] as @xmath317 . proof .
the first claim is clear from proposition [ convtodelta ] and the limit theorem [ limthmmarkcomp ] .
the second claim is obtained from the limit theorem [ limthmmarkcomp ] in the following way .
first note that the set @xmath1398 , \phi^+\in\mathfrak{b}^+({{{\mathscr x}}}),|\phi^+|=1\}\ ] ] is compact in the weak topology ( indeed , it is clear from the _ uniform _ convergence on spheres in the oseledets multiplicative ergodic theorem that the map @xmath1399\ ] ] is continuous in restriction to the set @xmath1400 whose image is therefore compact ) . in particular ,
the set ( [ phitausphere ] ) is bounded away from @xmath465 , and the function @xmath1401 , \delta_0)\ ] ] is a positive measurable function on @xmath1388 .
consequently , there exists @xmath1402 such that @xmath1403 from ergodicity of the measure @xmath1123 and the limit theorem [ limthmmarkcomp ] it follows that the family @xmath1396 $ ] , @xmath307 , does not converge to @xmath465 . on the other hand , as we have seen ,
the measure @xmath465 is an accumulation point for the family .
it follows that the measures @xmath1396 $ ] do not converge in @xmath1245 as @xmath317 , and , a fortiori , that the measures @xmath450 $ ] do not converge in @xmath1397)$ ] as @xmath317 .
corollary [ nonconvergence ] is proved completely .
let @xmath1404 be a complete separable metric space , and let @xmath1405 be the space of borel probability measures on _ x_. the _ weak topology _ on @xmath1405 is defined as follows .
let @xmath56 , @xmath1406 , and let @xmath1407 be bounded continuous functions . introduce the set @xmath1408 the kantorovich - rubinstein metric induces the weak topology on @xmath1416 by the kantorovich - rubinstein theorem , for bounded metric spaces , the kantorovich - rubinstein metric admits the following equivalent dual description .
given @xmath1414 let @xmath1417 be the set of probability measures @xmath138 on @xmath1418 such that projection of @xmath138 on the first coordinate is equal to @xmath1419 the projection of @xmath138 on the second coordinate is equal to @xmath1420 the kantorovich - rubinstein theorem claims that @xmath1421 in the proof of the limit theorems , we use the following simple estimate on the lvy - prohorov and the kantorovich - rubinstein distance between the push - forwards @xmath1432 of the measure @xmath103 under the mappings @xmath1433 aranson , s. h. ; grines , v. z. certain invariants of dynamical systems on two - dimensional manifolds ( necessary and sufficient conditions for the topological equivalence of transitive systems ) .
( russian ) mat .
90(132 ) ( 1973 ) , 372402 .
barreira , luis ; pesin , yakov nonuniform hyperbolicity .
dynamics of systems with nonzero lyapunov exponents .
encyclopedia of mathematics and its applications , 115 .
cambridge university press , cambridge , 2007 .
bufetov a.i . ,
decay of correlations for the rauzy
zorich induction map on the space of interval exchange transformations and the central limit theorem for the teichmller flow on the moduli space of abelian differentials .
soc . , 19 ( 2006 ) , no .
3 , 579623 .
bufetov , a.i . ,
gurevich b.m . ,
existence and uniqueness of the measure of maximal entropy for the teichmller flow on the moduli space of abelian differentials , matematicheskij sbornik , 2011 , volume 202 , number 7 , pp .
342 .
vershik , a. m. the adic realizations of the ergodic actions with the homeomorphisms of the markov compact and the ordered bratteli diagrams .
nauchn . sem .
pomi 223 ( 1995 ) , 120126 , 338 ; translation in j. math .
( new york ) 87 ( 1997 ) , no .
6 , 40544058 .
vershik , a. m. ; livshits , a. n. adic models of ergodic transformations , spectral theory , substitutions , and related topics .
representation theory and dynamical systems , 185204 , adv .
soviet math .
, 9 , amer .
soc . , providence , ri , 1992 . | the aim of this paper is to obtain an asymptotic expansion for ergodic integrals of translation flows on flat surfaces of higher genus ( theorem [ multiplicmoduli ] ) and to give a limit theorem for these flows ( theorem [ limthmmoduli ] ) . |
recent findings by the lhcb @xcite , belle @xcite , cdf @xcite , and d0 @xcite collaboration , have shown that the @xmath9 is produced in the @xmath10 decays into @xmath11 and @xmath12 and no trace of the @xmath13 is seen .
at the same time the complementary reactions on the @xmath14 decay into @xmath11 and @xmath12 @xcite , have shown that a signal is seen for the @xmath13 production and only a small fraction is observed for the @xmath9 .
these findings have generated excitement in the hadron community .
studies of weak @xmath15 and @xmath16 decays into two final mesons had got some attention @xcite in recent years .
the order of magnitude of the @xmath9 production was predicted in ref .
@xcite using light cone qcd sum rules under the factorization assumption .
after the experiments quoted above , some studies were done based on the @xmath17 or tetraquark structure of the scalar mesons in ref .
the @xmath14 decay into @xmath11 and @xmath12 branching ratio was calculated in the qcd - improved factorization approach in ref .
@xcite , where a two - meson distribution amplitude for the pion pair and final state interaction were considered .
the chiral unitary approach was used in ref .
@xcite , where predictions with no parameters fitted to the data were done for ratios of invariant mass distributions and the basic features of these experiments were well reproduced .
other work has also been done using the perturbative qcd approach in ref . @xcite .
more recently another approach has been used in ref .
@xcite using effective hamiltonians , transversity form factors and implementing the final sate interaction of the pions via the omnes representation . in this latter paper the @xmath10 decay into @xmath11 and @xmath18
is also studied and compared to experiment .
the method uses experimental phase shifts for the omnes representation and a few parameters fitted to the data . on the other hand the @xmath14 decay into @xmath11 and @xmath18
is not addressed there since `` this has an isovector component ( corresponding e.g. to the @xmath8 resonance ) and would have to be described by a coupled - channel treatment of the @xmath19 and @xmath18 @xmath20-waves . ''
some work along these lines has been recently done in ref .
@xcite . the chiral unitary approach for mesons @xcite is particularly suited for this purpose and has been used in many processes where
the @xmath21 are produced , including some related to @xmath16 and @xmath15 weak decays @xcite .
the purpose of this paper is to address the @xmath18 production in the @xmath10 decays into @xmath11 and two kaons and the @xmath18 and @xmath19 production in the @xmath14 decay into @xmath11 and this pair of mesons .
simultaneously , we shall also address the @xmath2 and @xmath4 decays .
experimentally , we have information in refs . @xcite . in ref .
@xcite the @xmath14 decay into @xmath11 and @xmath18 is studied .
rates for this branching ratio are provided and also for the @xmath22 .
however , the @xmath23 decay channel of the @xmath8 is not investigated . in ref .
@xcite a full phase space partial wave analysis of the @xmath24 is done , including @xmath25 and @xmath16 waves .
the @xmath10 decay into @xmath11 and @xmath26 with @xmath26 decaying into @xmath18 is addressed there @xcite . in this case
, the @xmath18 is in @xmath16-wave and in the present paper we only consider @xmath20-wave .
the work of ref .
@xcite on @xmath0 separates the @xmath20-wave @xmath18 below the @xmath27 region and this allows for a comparison with our predictions .
note that the results of refs .
@xcite are based on @xmath28 of @xmath29 collision data accumulated by lhcb during 2011 .
further improvement can be anticipated from studies including the @xmath30 of data accumulated during 2012 and with new data now being taken in the lhc run 2 . in ref .
@xcite the @xmath31 is studied , and concerning the present work , one measures the @xmath18 distribution close to threshold and the amount of @xmath20-wave @xmath18 below the @xmath27 peak .
more devoted measurements are under way .
a good thing of the use of the chiral unitary approach for these reactions is that one can obtain mass distributions up to an arbitrary normalization and relate the different distributions with no parameters fitted to the data .
this is due to the unified picture that the chiral unitary approach provides for the scattering of mesons . in this sense
, predictions on the coming measurements should be most welcome , and if supported by experiment , it can give us elements to dig into the nature of the low lying scalar mesons , which in this picture emerge as dynamically generated from the interaction of pseudoscalar mesons using a meson - meson interaction derived from the chiral lagrangians @xcite .
following refs .
@xcite , in fig .
[ fig : fig1 ] we draw the diagrams at the quark level that are responsible for the @xmath14 and @xmath10 decays into @xmath11 and another pair of quarks : @xmath32 in the case of the @xmath14 decay and @xmath33 in the case of @xmath10 decay .
the first process involves the @xmath34 , cabibbo suppressed cabibbo - kobayashi - maskawa matrix element , and the second one the @xmath35 cabibbo allowed one , which makes the widths large in the second case compared to the first one . following the work of ref .
@xcite we put together in a factor @xmath36 all elements which are common in the two decays .
next , in order to produce two mesons the final @xmath17 state has to hadronize , which we implement adding a @xmath37 pair with the quantum numbers of the vacuum @xmath38 , see fig .
[ fig : fig2 ] .
then we define the matrix @xmath39 for the @xmath17 elements @xmath40 which has the property @xmath41 and @xmath10 into @xmath11 and a primary @xmath42 pair , @xmath43 for @xmath14 and @xmath44 for @xmath10 .
[ fig : fig1],title="fig:",height=113 ] and @xmath10 into @xmath11 and a primary @xmath42 pair , @xmath43 for @xmath14 and @xmath44 for @xmath10 .
[ fig : fig1],title="fig:",height=120 ] .[fig : fig2],height=79 ] next we write the @xmath45 matrix elements of @xmath39 in terms of the physical mesons and then we can replace the matrix @xmath39 by the matrix @xmath46 given by @xmath47 where we are taking the standard mixing of the @xmath48 and @xmath49 in terms of a singlet and an octet of su(3 ) , @xmath50 the hadronization leads us to matrix was ignored . ]
@xmath51 where the @xmath49 terms have been neglected because the @xmath49 is too massive and has not effect here .
the decomposition of the former formulas tell us the weight by which a pair of pions are produced in the first step .
the next step consists in taking into account the interaction of the mesons produced to finally obtain the desired couple of mesons .
this is represented in fig .
[ fig : pipi_production ] . , @xmath52 and @xmath18 via direct plus rescattering mechanisms in @xmath14 ( a ) and
@xmath53 ( b ) decays .
[ fig : pipi_production],title="fig : " ] , @xmath52 and @xmath18 via direct plus rescattering mechanisms in @xmath14 ( a ) and @xmath53 ( b ) decays .
[ fig : pipi_production],title="fig : " ] the amplitudes for a final production of the different meson pairs are given by and @xmath54 terms with respect to ref .
@xcite because of the two possibilities to create the two identical particles with the @xmath55 and @xmath54 operators of eq .
( [ eq : phiphi_elements ] ) .
the changes in the numerical results for @xmath14 decay are of the order of 15% and there are no changes for the @xmath10 decay . ]
@xmath56 @xmath57 @xmath58 @xmath59 @xmath60 where @xmath61 are the loop functions of two meson propagators . in eqs .
( [ eq : t_b2jpsipipi])-([eq : t_bs2jpsikk ] ) we have taken into account that for the identical particles one has a factor of two from the two ways to match the two identical particles of the operator in eqs .
( [ eq : phiphi_elements ] ) with the two mesons produced and a factor @xmath62 in the @xmath63 function .
the @xmath64 are the scattering matrices and they are calculated in ref .
@xcite following ref .
@xcite . in the case of @xmath1 decay
, we have contributions from @xmath65 and @xmath66 .
it is interesting to split the contributions .
this is easily done taking all the @xmath67 equal , and also @xmath68 and rewriting @xmath69 as @xmath70 @xmath71 where the first two terms are in @xmath65 while the last two terms are in @xmath66 .
then we split eq .
( [ eq : t_b2jpsikk ] ) into @xmath72 with @xmath73,\end{aligned}\ ] ] @xmath74.\end{aligned}\ ] ] the @xmath75 decays proceed in an analogous way . the basic quark diagram is now given in fig .
[ fig : figb - decay - quarklevel ] . into @xmath11 and a primary @xmath76 pair .
[ fig : figb - decay - quarklevel],height=113 ] the hadronization leads us now , neglecting @xmath49 , to @xmath77 from where , taking into account the final state interaction of the mesons , we obtain @xmath78 @xmath79 there is no need to recalculate the meson - meson scattering matrix @xmath64 , since using isospin symmetry and the fact that @xmath80 of isospin , we have @xmath81 in particular one can see that the @xmath82 amplitude of eq .
( [ eq : t_b-2jpsipieta ] ) is @xmath83 times the amplitude for @xmath84 and hence its rate of production will be a factor of two bigger .
one final observation is the fact that in a @xmath85 transition we shall need a @xmath86 for the @xmath11 to match angular momentum conservation .
hence , @xmath87 , and we assume @xmath88 to be constant ( equal @xmath89 in the calculations ) .
thus , @xmath90 where the factor @xmath91 is coming from the integral of @xmath92 and @xmath93 is @xmath94 , which depends on the @xmath12 invariant mass . in eq .
( [ eq : dgamma ] ) @xmath95 is the @xmath11 momentum in the global cm frame ( @xmath96 at rest ) and @xmath97 is the pion momentum in the @xmath12 rest frame .
, @xmath18 invariant mass distributions for the @xmath98 , @xmath99 decays ; ( b ) @xmath12 , @xmath52 , @xmath18 invariant mass distributions for the @xmath6 , @xmath99 , @xmath100 decays .
[ fig : mass_distribution],title="fig : " ] , @xmath18 invariant mass distributions for the @xmath98 , @xmath99 decays ; ( b ) @xmath12 , @xmath52 , @xmath18 invariant mass distributions for the @xmath6 , @xmath99 , @xmath100 decays . [
fig : mass_distribution],title="fig : " ] and @xmath101 invariant mass distributions for the @xmath4 and @xmath2 decays .
[ fig : mass_distribution_pi2eta ] ] our results are summarized in figs . [
fig : mass_distribution ] ( a ) and ( b ) , and fig .
[ fig : mass_distribution_pi2eta ] . in fig .
[ fig : mass_distribution ] ( a ) we show the results for the @xmath5 together with those for @xmath0
. the former results are those calculated in ref .
the latter are new .
as we can see , the @xmath18 distribution gets maximum strength close to the @xmath18 threshold and then falls down gradually .
this is due to the effect of the @xmath9 resonance below threshold . in this case
we started from an @xmath33 quark state , which has isospin zero , and the strong interaction hadronization conserves it .
so , even if @xmath18 could be in @xmath102 , the process of formation guarantees that this is an @xmath65 state and the shape of the distribution is due to the @xmath9 .
the strength is small compared to the one of the @xmath9 at its peak , but the integrated strength over the invariant mass of @xmath18 is of the same order of magnitude as that for the strength below the peak of the @xmath9 going to @xmath103 .
however , we should take into account that we are calculating only the @xmath20-wave part of the @xmath18 spectrum , hence , contributions from @xmath27 ( @xmath104-wave ) , @xmath26 ( @xmath16-wave ) etc , are not evaluated .
it is interesting to compare this with experiment .
first by integrating the strength of the @xmath18 distribution over its invariant mass we find a ratio @xmath105}{\mathcal{b } [ \bar b^0_s \to j/\psi f_0(980 ) ; f_0(980 ) \to \pi^+ \pi^- ] } = 0.34 \pm 0.03,\ ] ] where we have added an estimated 10 % theoretical uncertainty . by taking into account that the rates for @xmath9 and @xmath27 are @xmath106 = ( 1.39 \pm 0.14 ) \times 10^{-4},\nonumber \\ & \mathcal{b } [ \bar b^0_s \to j/\psi \phi ] = ( 1.07 \pm 0.09 ) \times 10 ^{-3},\end{aligned}\ ] ] we find @xmath107}{\mathcal{b } [ \bar b^0_s \to j/\psi \phi]}=0.044 \pm 0.007 , \ ] ] where we have summed the relative errors in quadrature .
if we stick to a band of energies @xmath108 mev we find @xmath109(s{\text{-wave}})}{\mathcal{b } [ \bar b^0_s \to j/\psi \phi ; \phi \to k^+ k^- ] } = 0.017 \pm 0.003 , \ ] ] where the branching fraction of @xmath110 for @xmath27 decay into @xmath18 has been taken .
this value is in agreement with experiment @xcite , @xmath111 within errors .
these experimental numbers are consistent with , and improve , earlier measurements from cdf @xcite and atlas @xcite .
they are also in agreement with the results of ref .
@xcite and the estimations prior to the experiment in ref . @xcite . on the other hand , by looking at fig .
17 of ref .
@xcite one can compare our results for the @xmath18 mass distribution with the contribution of the @xmath9 in that figure , and the accumulated strength around threshold and the fall down with increasing invariant mass agree fairly well , although our distribution falls faster .
we come back now to the decays of the @xmath14 . in fig .
[ fig : mass_distribution ] ( b ) we show the results for the @xmath14 into @xmath11 and @xmath12 , @xmath11 and @xmath18 and @xmath11 plus @xmath52 . in this case
we had the hadronization of a @xmath32 state , which contains @xmath102 .
the @xmath12 in @xmath20-wave , however , can only be in @xmath65 , hence the peaks for this distribution reflect again the @xmath13 and @xmath9 excitation .
we should note that the normalization in figs .
[ fig : mass_distribution ] ( a ) and ( b ) is the same .
hence , the difference in size mostly reflects the differences between the ckm matrix elements .
we should note that because of the experimental resolution this peak would not appear so narrow in the experiments , but the integrated strength should be comparable , and this was already done in ref .
note that in any case , the strength of the @xmath9 excitation is very small compared to that of the @xmath13 ( the broad peak to the left ) as was already noted in the experiments .
the new results in this paper are for the @xmath18 and @xmath52 distributions .
the @xmath52 distribution has a sizeable strength , much bigger than that for the @xmath9 and reflects the @xmath8 excitation .
this prediction is tied exclusively to the weights of the starting meson meson channels in eqs .
( [ eq : phiphi_elements ] ) and the scattering matrices appearing in eqs .
( [ eq : t_b2jpsipipi ] ) .
hence , this is a prediction of this approach , not tied to any experimental input . in fig .
[ fig : mass_distribution](b ) we have plotted only the @xmath20-wave contribution . in the case of @xmath52 , there is no relevant @xmath104-wave contribution in the @xmath112 region of the figure @xcite .
however , this is not the case for @xmath12 , which can couple to the @xmath113 meson in @xmath104-wave and give a large contribution .
this is indeed the case and it was evaluated in ref .
@xcite ( see fig . 5 of that paper ) .
the @xmath113 contribution peaks around 770 mev , and has larger strength than the @xmath13 contribution , but at invariant masses around 500 mev and bellow , the strength of the @xmath13 dominates the one of the @xmath113 .
the @xmath18 distribution in the @xmath14 decay is now both in @xmath65 and @xmath66 , hence it reflect the effects of both the @xmath9 and the @xmath8 resonances . in this case , if we draw a smooth curve below the @xmath9 peak to separate it from the @xmath13 contribution , we find , @xmath114}{\mathcal{b } [ \bar b^0 \to j/\psi f_0(980 ) ; f_0(980 ) \to \pi^+ \pi^-]}=0.53 \pm 0.05 .\ ] ] the strength of this ratio is a bit larger now than the corresponding one for @xmath10 decay , given in eq .
( [ eq : ratio_bskk ] ) . in ref .
@xcite we also find the branching ratio for @xmath115 with @xmath8 decaying into @xmath18 , @xmath116 = ( 4.7\pm 3.31\pm 0.72)\times 10^{-7}.\ ] ] it is unclear how this is obtained because , as discussed above , both the @xmath9 and @xmath8 resonances contribute to the @xmath18 distribution .
with the caveat about not comparing exactly the same magnitude we can make an estimate of the @xmath8 production rate by taking a background below the @xmath23 distribution as done in ref . @xcite and using the former equations we get a rate @xmath117 if we multiply the latter value by the ratio @xmath118 @xcite , we obtain @xmath119 , which is agreement with the value of eq .
( [ eq : br_b2jpsia0 ] ) within errors .
we would also like to compare our @xmath18 mass distribution with the one of fig .
13 of ref .
@xcite . the experimental distribution after subtraction of combinatorial background and misidentified events has in principle large errors , which will be improved in coming experiments , but we can see that the shape of the distribution agrees qualitatively with ours , with a relatively faster fall down of the experimental one .
we come now to the results for @xmath75 decay which we plot in fig .
[ fig : mass_distribution_pi2eta ] .
the scale of the figure is the same as in fig .
[ fig : mass_distribution ] . as discussed in the former section ,
the strength for the @xmath120 mass distribution in @xmath4 is twice as big as the one of @xmath3 .
the strength of the @xmath101 mass distribution at the peak is however about four times bigger than the one for @xmath18 in the @xmath14 decay .
we also observe that the position of the peak has moved to higher invariant masses compared to the @xmath14 or @xmath121 cases .
both features find a natural explanation in the fact the the @xmath101 distribution in the @xmath75 decay is due to the @xmath8 , which is seen in the figures , is much wider than that of the @xmath9 .
we should also note that the shape of the @xmath18 distribution in the @xmath14 case is also a bit different , sticking more towards the @xmath122 threshold .
we also see that the @xmath9 distribution in this decay has a different shape to that in the @xmath121 decay , with zero strength around 1000 mev .
it is clear that there are now interferences of the different terms contributing to the amplitude in eq .
( [ eq : t_b2jpsikki01 ] ) . in order to see the interferences commented above in more detail
, we evaluate separately the contributions to @xmath1 decay from the @xmath65 and @xmath66 parts of the amplitude .
this can be seen in fig .
[ fig : kk_distribution_i01 ] .
what we observe there is that both the @xmath65 ( @xmath9 ) and @xmath66 ( @xmath8 ) individual contributions are larger than the total , and we also see that the shape of either of them resembles move the one of fig .
[ fig : mass_distribution_pi2eta ] for @xmath101 .
the strength of the @xmath8 contribution in the @xmath1 decay is still smaller than that of @xmath2 decay ( fig .
[ fig : mass_distribution_pi2eta ] ) because the latter contains also a tree level contribution , which is absent in the @xmath1 case .
we also observe in fig .
[ fig : kk_distribution_i01 ] , that when we add the @xmath65 and @xmath66 parts of the amplitude , the total strength becomes smaller , indicating a strong cancellation , which is also responsible for the change in the shape of the distribution
. it would be most instructive to see all these features in actual experiments .
invariant mass distributions for the @xmath1 decay .
the @xmath65 , @xmath66 individual contribution and the total , given by the sum of the amplitudes are shown by the dashed , dotted and full lines , respectively .
[ fig : kk_distribution_i01 ] ]
we have studied the decay rates for @xmath0 , @xmath1 , @xmath3 , @xmath4 , @xmath2 , and have compared them to the rates obtained for the @xmath5 and @xmath6 .
new measurements are underway in the lhcb collaboration and predictions prior to the experiment are most opportune to give strength to the claims of the chiral unitary theory on the interpretation of resonances and the ability of the approach to describe the meson - meson interactions .
we find that the rates of @xmath20-wave @xmath18 production are small compared to those of @xmath12 production but when integrated over the invariant mass they are still smaller but of the same order of magnitude
. the interesting feature of the results is that the @xmath18 distribution peaks around the @xmath18 threshold as a consequence of the presence of the @xmath9 and @xmath8 resonances . in the case of the @xmath0 ,
only the @xmath9 resonance influences this distribution , but in the case of the @xmath1 , both the @xmath9 and @xmath8 resonances contribute to its strength . in the case of the @xmath3 ,
one finds a peak for the @xmath8 resonance and its strength is much larger than the one for the @xmath6 reaction at the @xmath9 peak .
we also calculated the @xmath120 and @xmath101 mass distributions for the decays @xmath4 and @xmath2 .
we found in this case that the strength of the @xmath120 distribution is twice the one of @xmath52 in @xmath14 decay , and the strength of the @xmath101 distribution about four times bigger than that of the @xmath123 for the @xmath14 decay .
we could find an easy explanation based on the fact that the @xmath101 production , in @xmath66 , is only influenced by the @xmath8 state which has a wider distribution than the @xmath9 .
one interesting aspect of the calculations is that we could predict all these mass distributions with no free parameters , up to a global normalization which is the same for all processes .
the method relies on the constancy of the @xmath36 factor which summarizes the weak amplitudes and the hadronization procedure .
the only thing demanded is that this factor is smooth and practically constant as a function of the invariant masses in the limited range where the predictions are made .
a discussion on this issue and support for this approximation is found in refs .
@xcite .
the predictions made here compare reasonably well with present experimental information , but more precise data are coming from lhcb and comparison with these data will be useful to make progress in our understanding of the meson - meson interaction and the nature of the low lying scalar mesons .
we would like to thank tim gershon for a careful reading of the manuscript and useful information concerning the lhcb experiments .
one of us , e. o. wishes to acknowledge support from the chinese academy of science in the program of visiting professorship for senior international scientists ( grant no .
2013t2j0012 ) .
this work is partly supported by the spanish ministerio de economia y competitividad and european feder funds under the contract number fis2011 - 28853-c02 - 01 , fis2011- 28853-c02 - 02 , fis2014 - 57026-redt , fis2014 - 51948-c2- 1-p , and fis2014 - 51948-c2 - 2-p , and the generalitat valenciana in the program prometeo ii-2014/068 .
this work is also partly supported by the national natural science foundation of china under grant nos .
11165005 , 11565007 , and 11475227 .
it is also supported by the open project program of state key laboratory of theoretical physics , institute of theoretical physics , chinese academy of sciences , china ( no.y5kf151cj1 ) .
we acknowledge the support of the european community - research infrastructure integrating activity study of strongly interacting matter ( acronym hadronphysics3 , grant agreement n. 283286 ) under the seventh framework programme of eu .
t. aaltonen _ et al .
_ [ cdf collaboration ] , phys . rev .
d * 84 * , 052012 ( 2011 ) [ arxiv:1106.3682 [ hep - ex ] ] .
v. m. abazov _ et al .
_ [ d0 collaboration ] , phys . rev .
d * 85 * , 011103 ( 2012 ) [ arxiv:1110.4272 [ hep - ex ] ] .
b. el - bennich , o. leitner , j. -p .
dedonder and b. loiseau , phys . rev .
d * 79 * , 076004 ( 2009 ) [ arxiv:0810.5771 [ hep - ph ] ] .
o. leitner , j. -p .
dedonder , b. loiseau and b. el - bennich , phys .
d * 82 * , 076006 ( 2010 ) [ arxiv:1003.5980 [ hep - ph ] ] .
d. delepine , j. l. lucio m. and c. a. ramirez , eur .
j. c * 45 * , 693 ( 2006 ) [ hep - ph/0501022 ] .
w. f. wang , h. n. li , w. wang and c. d. lu , phys .
d * 91 * , 094024 ( 2015 ) [ arxiv:1502.05483 [ hep - ph ] ] .
j. t. daub , c. hanhart and b. kubis , arxiv:1508.06841 [ hep - ph ] .
m. albaladejo and b. moussallam , eur .
j. c * 75 * , no .
10 , 488 ( 2015 ) [ arxiv:1507.04526 [ hep - ph ] ] .
j. a. oller , e. oset and j. r. pelaez , phys .
d * 59 * , 074001 ( 1999 ) [ erratum - ibid .
d * 60 * , 099906 ( 1999 ) ] [ erratum - ibid .
d * 75 * , 099903 ( 2007 ) ] [ hep - ph/9804209 ] .
n. kaiser , eur .
j. a * 3 * , 307 ( 1998 ) . m. p. locher , v. e. markushin and h. q. zheng , eur .
j. c * 4 * , 317 ( 1998 ) [ hep - ph/9705230 ] .
j. nieves and e. ruiz arriola , nucl .
a * 679 * , 57 ( 2000 ) [ hep - ph/9907469 ] .
j. r. pelaez and g. rios , phys .
lett . * 97 * , 242002 ( 2006 ) [ hep - ph/0610397 ] .
j. j. xie , l. r. dai and e. oset , phys .
b * 742 * , 363 ( 2015 ) [ arxiv:1409.0401 [ hep - ph ] ] .
w. h. liang , j. j. xie and e. oset , phys . rev .
d * 92 * , 034008 ( 2015 ) [ arxiv:1501.00088 [ hep - ph ] ] . | we study the @xmath0 , @xmath1 , @xmath2 , @xmath3 and @xmath4 decays and compare their mass distributions with those obtained for the @xmath5 and @xmath6 .
the approach followed consist in a factorization of the weak part and the hadronization part into a factor which is common to all the processes .
then what makes the reactions different are some trivial cabibbo - kobayashi - maskawa matrix elements and the weight by which the different pairs of mesons appear in a primary step plus their final state interaction .
these elements are part of the theory and thus , up to a global normalization factor , all the invariant mass distributions are predicted with no free parameters .
comparison is made with the limited experimental information available .
further comparison of these results with coming lhcb measurements will be very valuable to make progress in our understanding of the meson - meson interaction and the nature of the low lying scalar meson resonances , @xmath7 and @xmath8 . |
the unknotting number @xmath2 of a knot @xmath3 is the minimum number of crossing changes required to convert @xmath3 into a trivial knot taken over all knot diagrams representing @xmath3 .
an unknotting operation involves switching the under / over crossing strand at a crossing point of a knot diagram .
one important tool to represent knots is that of the closure of braids .
an @xmath4-string braid is a set of @xmath4 smooth non - intersecting and non - self intersecting strings in @xmath5 , whose initial points are @xmath6 and final points are @xmath7 such that the third projection function increases while moving along a string from its initial point to its final point .
the @xmath4-string braids form a group , denoted by @xmath8 .
this group is generated by @xmath9 elementary braids @xmath10 , each of which contains a single crossing . an important result that connects knots and braids
is the alexander s theorem @xcite , which states that all knot isotopy classes can be represented by closure of braids . using generators of the braid group
, we observe that @xmath11 , denoted by @xmath0 , is a toric braid as its closure is a torus knot @xmath12 . in @xcite , it was proved that if @xmath13 , then the closure of the braid @xmath0 is a minimum crossing diagram for @xmath12 . in @xcite , kronheimer and mrowka used gauge theory to prove that the unknotting number of an algebraic knot is equal to the genus of the milnor fiber .
a consequence of this result provides the unknotting number of a torus knot of type @xmath14 to be @xmath15 . in general
, it is not true that some minimum crossing diagram contains @xmath2 crossings whose under / over crossing change makes it unknot .
but it is interesting to observe that in @xmath0 , there exist @xmath16 crossings whose change , from over to under , gives a braid whose closure is a trivial knot .
a well - established technique for unknotting a knot diagram is that of converting the diagram to an ascending or descending diagram .
if a knot @xmath3 with c crossings has a bridge with b crossings , then we can convert @xmath3 into an ascending or descending diagram with less than equal to @xmath17 crossing changes since switching of all the descending(ascending ) crossings starting from the end(beginning ) of the bridge gives a totally ascending(descending ) diagram . in the case of torus knot @xmath12 given by the closure of @xmath18 ,
we have @xmath19 and @xmath20 . thus we can convert @xmath12 into an ascending or descending diagram with less than equal to @xmath16 crossings .
since unknotting number of @xmath12 is @xmath16 , both ascending number and descending number are equal to the unknotting number .
this method is very easy to understand but very time consuming in deciding the exact position of these @xmath2 crossings as there is no symmetry / pattern in the selected crossings .
also the same procedure is not valid in unknotting torus links . in this paper , we present a method to unknot torus knots and extend the same to torus links
. we also present a proof for the symmetry of @xmath21 and @xmath22 in torus knot @xmath12 . in section [ pre ] , we introduce a few definitions and some preliminaries that we use in this paper . in section [ mucdftk ] ,
we provide a method of finding minimal unknotting crossing data for all torus knots .
the main results of this section are as follows : [ thm3 ] let @xmath23 be a torus knot with @xmath24 . then
, the following statements are equivalent : 1 .
the unknotting number of @xmath23 is equal to the number of elements in the @xmath25crossing data of @xmath26 , 2 .
@xmath27 , 3 .
the @xmath25crossing data of @xmath23 is a minimal unknotting crossing data for @xmath23 .
[ thm4 ] for every @xmath21 and @xmath28 , where @xmath29 , the @xmath30braid @xmath31 where @xmath32 , with @xmath33 , for @xmath34 and @xmath35 , and @xmath36 , for @xmath37 . in section [ pqsym ]
, we present a simple proof for symmetry of @xmath21 and @xmath22 in @xmath12 .
[ thm6 ] let @xmath38 be a @xmath21-braid , where @xmath29 , then if we change crossings in @xmath38 , based on @xmath39 , we get a braid which is markov equivalent to @xmath40 . in section [ mucdftl ] ,
we extend the method of finding minimal unknotting crossing data to torus links . in appendix
a we write a matlab program(taking @xmath21 and @xmath22 as input ) which gives two different sets of minimal unknotting crossing data for any torus knot @xmath12 .
a crossing data for any @xmath4-braid @xmath41 , denoted by @xmath42 $ ] , is a finite sequence of natural numbers enclosed in a bracket , given to the crossings starting from the first crossing from the top to the last crossing at the bottom , based on the braid representation of @xmath41 using elementary braids @xmath43 .
for example , the crossing data for the braid @xmath44 is @xmath45 $ ] ( figure 1 ) . from the definition of crossing data for a braid , we observe the following : 1
. two different @xmath46braids may have the same crossing data .
this is mainly because the crossing data does not provide under / over crossing information at a crossing .
crossing data for any two braids , having equal number of crossings , is same .
3 . crossing data for any two equivalent braids need not be the same .
the crossing data for @xmath26 is @xmath47.$ ] , width=113,height=151 ] an unknotting crossing data for an @xmath46braid , @xmath41 , is a subsequence to the crossing data of @xmath41 such that if we make crossing change at these crossing positions then the closure of this braid is equivalent to unknot or unlink .
if the number of elements in an unknotting crossing data of a braid @xmath41 is equal to the unknotting number of @xmath3 ( where @xmath3 is the closure of @xmath41 ) , then this unknotting crossing data is known as minimal unknotting crossing data for both the braid @xmath41 and the knot @xmath3 .
[ lem1 ] for every @xmath4 , the @xmath48-braid @xmath49 is equal to @xmath50 , where @xmath51 , @xmath52 , @xmath53 .
we prove this by means of mathematical induction .
let @xmath54 be the statement that @xmath49 is equal to @xmath50 .
+ basis step : @xmath55 is true , because the @xmath56-braid @xmath57 . also @xmath58 is true , because the @xmath59-braid @xmath60 . here
, we have used the fundamental relation @xmath61 .
+ inductive step : let @xmath3 be some fixed positive integer and @xmath62 is true , that is , the @xmath63-braid @xmath64 is equal to @xmath65 , where @xmath66 , @xmath67 , @xmath68 , is true . + to complete the inductive step , we must show that @xmath69 is also true .
+ consider @xmath70-braid @xmath71 using simple fundamental relations such as @xmath72 for @xmath73 , @xmath74 , we observe that the above equation is equal to @xmath75 @xmath76 @xmath77 @xmath78 @xmath79 @xmath80 finally , by symmetry , the above equation is equal to @xmath81 now , by induction hypothesis @xmath82 we have @xmath83 therefore , by the principle of mathematical induction , we have shown that @xmath84 is true for all positive integers @xmath4 .
in this section , we present a method for determining the minimal unknotting crossing data for the braids of type @xmath85 , where @xmath86 , so that , after switching the crossings at this selected crossing data , we obtain a braid whose closure is isotopically equivalent to the trivial knot . though this method does not enable us to determine the minimal unknotting crossing data for all toric braids of type @xmath26 , it ultimately provides us a way to find out the minimal unknotting crossing data for @xmath26 .
+ the following procedure directly provides the minimal unknotting crossing data for a large class of toric braids , known as toric braids @xmath26 , where @xmath24 , and @xmath87 , and helps to find out the minimal unknotting crossing data for every other toric braid . *
procedure 1 : * + consider a toric braid of type @xmath26 , such that @xmath88 . +
* step 1 * : first represent @xmath26 as @xmath89 * step 2 * : the crossing data for the braid @xmath26 , based on step 1 , is + @xmath90.\ ] ] * step 3 * : 1 .
when @xmath91 : consider the crossing data @xmath92\ ] ] if @xmath93 and @xmath94 $ ] if @xmath95 .
2 . when @xmath96 : let @xmath97 , where @xmath98 .
let @xmath99 if @xmath100 and @xmath101 if @xmath102 ; @xmath103 and @xmath104 .
now , consider the crossing data @xmath105 .\ ] ] * step 4 * : denote the crossing data obtained in step 3 as @xmath25crossing data for @xmath26 .
represent the @xmath25crossing data for @xmath26 as @xmath106 .
+ we observe that for toric braids @xmath26 , where @xmath24 , and @xmath87 , the @xmath106 is same as minimal unknotting crossing data for @xmath26 ( or @xmath23 ) .
using this procedure , we can also find minimal unknotting crossing data for all torus knots of type @xmath12 with @xmath24 and @xmath107 or @xmath108 .
in general , the @xmath25crossing data for any toric braid @xmath26 will help us to determine the minimal unknotting crossing data for that braid .
[ thm1 ] for every @xmath4 , the @xmath48-braid @xmath109 is a trivial @xmath48-braid .
proof directly follows from lemma [ lem1 ] .
[ thm2 ] let @xmath23 be a torus knot with @xmath24 . if @xmath110 then the @xmath25crossing data for @xmath26 is a minimal unknotting crossing data for @xmath26 ( or @xmath23 ) .
consider @xmath23 .
+ * case 1 .
* when @xmath111 .
+ then , @xmath112 for some @xmath113 .
the toric braid representation of @xmath114 is @xmath115 by making crossing changes at the @xmath25crossing data for @xmath26 , we obtain : @xmath116 now , we show that the number of elements in the @xmath25crossing data is equal to the unknotting number for these torus knots . from @xcite , we know that the unknotting number for these torus knots is equal to @xmath117 the number of elements in the @xmath25crossing data is equal to @xmath118 which is same as the unknotting number . + * case 2 . * when @xmath119 . + then , @xmath120 for some @xmath113 .
the toric braid representation of @xmath114 is @xmath121 . by using procedure 1 , making the crossing changes at the @xmath25crossing data ,
we obtain : @xmath122 @xmath123 , where : @xmath124 @xmath125 by theorem [ thm1 ] , we get @xmath126 by lemma [ lem1 ] , the above equation is equal to @xmath127 , whose closure is a trivial knot .
+ now , we show that the number of elements in the @xmath25crossing data is equal to the unknotting number for these torus knots .
observe that the unknotting number is equal to @xmath128 and the number of elements in the @xmath25crossing data is equal to @xmath129 now , we have sufficient data to prove the main theorem of this section .
@xmath130 : for given torus knot @xmath23 , we can write @xmath131 for some integer @xmath113 and for some @xmath132 .
observe that the unknotting number of @xmath23 is equal to @xmath133 and the number of elements in the @xmath25crossing data is @xmath134 . by hypothesis , these two are equal .
therefore , we have two solutions for @xmath135 , i.e. , either @xmath102 or @xmath136 .
hence , @xmath27 .
@xmath137 is nothing but theorem [ thm2 ] and @xmath138 directly follows from the definition of minimal unknotting crossing data .
the braid @xmath139 is the braid obtained from @xmath140 by rotating @xmath140 in @xmath5 about the @xmath141-axis through an angle of @xmath142 . [ rmk2 ] ( a ) .
it is easy to observe that the closure of the braids @xmath140 and @xmath139 are equivalent . actually @xmath140 and @xmath139 are same because we obtain @xmath139 from @xmath140 without any braid operation .
however we can not say that for any braid @xmath143 , @xmath144 . also , observe that if @xmath145 , then @xmath146 .
if @xmath147 is the @xmath25crossing data of @xmath0 , then the elements in @xmath25crossing data of @xmath148 are @xmath149 .
[ lem2 ] if @xmath150 , then @xmath151 holds if either @xmath152 or @xmath153 , where @xmath154 have values @xmath155 .
proof directly follows by using the braid relations @xmath156 if either @xmath157 or @xmath158 , where @xmath159 + we can verify the last relation for different possible values of @xmath160 .
now , we prove second main theorem of this section .
consider , the @xmath30braid @xmath161 @xmath162 @xmath163 @xmath164 @xmath165 @xmath166 .
+ similarly , we obtain @xmath167 .
+ finally , we have @xmath168 .
+ [ rmk3 ] if we take @xmath169 for @xmath34 and @xmath170 we obtain @xmath171 + @xmath172 that is , if we change crossings in @xmath38 , based on @xmath39 , we obtain a diagram which is markov equivalent to @xmath173 [ rmk4 ] for every @xmath174 , with @xmath175 , @xmath176 then , by changing the crossings in @xmath0 , based on @xmath177 , we get + @xmath178 , where @xmath179 and @xmath180 . + by theorem [ thm1 ] and remark [ rmk3 ] , @xmath181 is a trivial @xmath30braid and @xmath140 is markov equivalent to @xmath182 .
thus , we can find a minimal unknotting crossing data for @xmath12 , if we can find the unknotting crossing data for @xmath182 such that the number of elements in this unknotting crossing data is equal to the unknotting number of @xmath183 , which is equal to @xmath184 observe that , this is the unknotting number of @xmath185 .
now we will provide the unknotting process .
consider any torus knot of type @xmath23 , where @xmath186 .
let @xmath187 and @xmath188 .
since @xmath189 , euclid s algorithm ensures that @xmath190 such that if we define @xmath191 as @xmath192 @xmath193 we get @xmath194 .
+ now , if we select @xmath195-crossing data for @xmath196 for @xmath197 , then by theorem [ thm4 ] and remark [ rmk4 ] , there are corresponding crossings in @xmath12 which provides minimal unknotting crossing data for @xmath12 .
these corresponding crossings are determined as below : for finding the minimal unknotting crossing data for a torus knot we consider the following steps .
let @xmath12 be a torus knot .
then , ( say ) * step 1 * : @xmath198 * step 2 * : @xmath199 * step 3 * : @xmath200 * step 4 * : @xmath201 @xmath202 * step i * : @xmath203 @xmath202 * step n * : @xmath204 + is our last step .
+ now , consider @xmath205 ; and @xmath206 .
then the minimal unknotting crossing data for @xmath12 is @xmath207\ ] ] if n is odd ; and @xmath208\ ] ] if n is even .
here @xmath209 , where @xmath210 ; + and @xmath211 , width=556,height=476 ] [ ex2 ] to find the unknotting procedure and minimal unknotting crossing data for torus knot @xmath212 . + by the above described procedure , we have \1 .
@xmath213 where @xmath214 \2 .
@xmath215 where @xmath216 \3 .
@xmath217 where @xmath218 + here @xmath219 + for @xmath220 , the minimal unknotting crossing data is @xmath221=[8,12,13,14,17,18,22,23,24]$ ] , where @xmath222 and @xmath223 .
[ ex3 ] minimal unknotting crossing data for torus knot @xmath224 is @xmath225 $ ] . in this case
in this section we show that @xmath227 , by showing that both the braids @xmath0 and @xmath228 are markov equivalent to a braid @xmath140 , which is obtained by changing the crossings in @xmath229 , based on @xmath230 .
[ thm5 ] for any two natural numbers @xmath28 and @xmath4 , @xmath231 i.e. , the braid product of two toric braids @xmath232 and @xmath233 is markov equivalent to the toric braid @xmath234 .
to prove this theorem , we use markov moves and the following simple straight forward results : 1 . for any @xmath235 , @xmath236 .
2 . if @xmath237 , then @xmath238 .
it is easy to prove the above two results by using the fundamental braid relations @xmath239 observe that for @xmath240 , we have @xmath241 by @xmath242 , we have @xmath243 similarly , for @xmath244 , we have @xmath245 by @xmath242 , we have @xmath246 @xmath247 @xmath248 now , in general : @xmath249 @xmath250 @xmath251 @xmath252 @xmath164 @xmath164 @xmath164 @xmath164 @xmath164 @xmath164 @xmath164 @xmath164 @xmath164 + @xmath253 + @xmath254 @xmath255 .
[ rmk5 ] for @xmath256 , by theorem [ thm5 ] , the @xmath21-braid @xmath38 , where @xmath29 , is markov equivalent to a @xmath28-braid @xmath257 .
[ ex1 ] consider a toric braid of type @xmath258 .
then , by theorem [ thm5 ] , @xmath259 . here ,
in this example , we show @xmath260 .
observe that @xmath261 the same is shown in figure 4 @xmath262 .
using the equality @xmath263 in the above equation , we get + @xmath264 .
this is shown in figure 4 @xmath265 . again , using @xmath266 , we get + @xmath267 .
this expression is shown in figure 4 @xmath268 .
figure 4 @xmath269 represents the following expression , which is also obtained by a similar argument , + @xmath270 now using the markov move @xmath271 , @xmath272 , where @xmath273 and @xmath274 , we observe that @xmath275 .
this is shown in figure 4 @xmath276 .
+ using markov move @xmath242 , we remove @xmath277 from the above expression , so that we obtain @xmath278 , which is shown in figure 4 @xmath279 .
+ finally , again by using markov move @xmath271 , we obtain that + @xmath280 .
this is shown in figure 4 @xmath281 .
, width=548,height=302 ] now we prove the main theorem of this section .
let @xmath140 is the braid obtained from @xmath38 after changing the crossings in @xmath38 , based on @xmath39 .
then + @xmath180 @xmath282 @xmath283 @xmath164 @xmath284 + @xmath285 + by theorem [ thm1 ] + @xmath286 + @xmath287 + @xmath288 + @xmath289 + @xmath290 + @xmath291 + @xmath292 + @xmath293 + @xmath164 @xmath294 + @xmath164 @xmath295 + @xmath296 + @xmath297 + @xmath298 .
in this section , we extend the method of finding minimal unknotting crossing data for torus knots to torus links , by small changes in the method .
[ thm7 ] let @xmath299 be a torus link .
then the @xmath25crossing data for @xmath300 is a minimal unknotting crossing data for @xmath300 ( or @xmath299 ) .
it follows from theorem [ thm1 ] , that @xmath301 is a unknotting crossing data for @xmath302 .
note that , @xmath303 and the unknotting number of @xmath304 is @xmath305 thus , @xmath301 is the minimal unknotting crossing data @xmath302 . , width=521,height=309 ] [ rmk6 ] for every @xmath174 , with @xmath306 , then @xmath307 and @xmath308 then ,
after changing the crossings in @xmath0 , based on @xmath177 , we obtain + @xmath178 , where @xmath179 and @xmath180 . + by theorem [ thm4 ] and theorem [ thm1 ]
, @xmath181 is a trivial @xmath21-braid and @xmath140 is markov equivalent to @xmath182 .
thus , we can find a minimal unknotting crossing data for @xmath12 , if we can find the unknotting crossing data for @xmath182 such that the number of elements in this unknotting crossing data is equal to the unknotting number of @xmath309 , i.e. , @xmath310 observe that , this is the unknotting number of @xmath185 .
now we will provide the unknotting process .
consider a torus link of type @xmath12 , where @xmath311 and @xmath312 .
let @xmath187 and @xmath188 .
since @xmath313 , euclid s algorithm ensures that @xmath190 such that if we define @xmath191 as @xmath192 @xmath193 we get @xmath314 .
+ now , if we select @xmath195-crossing data for @xmath196 for @xmath197 , then by theorem [ thm4 ] , theorem [ thm7 ] and remark [ rmk6 ] , there are corresponding crossings in @xmath12 which provides minimal unknotting crossing data for @xmath12 .
[ ex4 ] by the above said method , minimal unknotting crossing data for @xmath315 is @xmath316 $ ] .
* acknowledgements * + authors thank professor akio kawauchi for his valuable comments and suggestions .
also the first author thanks csir , new delhi and iit ropar for providing financial assistance and research facilities .
+ 99 kunio murasugi , knot theory and its applications , translated from the 1993 japanese original by bohdan kurpita , birkhuser boston inc . ,
boston , ma , 2008 .
p. kronheimer and t. mrowka , gauge theory for embedded surfaces , i , topology , 32 ( 1993 ) , 773 - 826 .
p. kronheimer and t. mrowka , gauge theory for embedded surfaces , ii , topology , 34 ( 1995 ) , 37 - 97 .
k. murasugi , on the braid index of alternating links , trans .
soc , 326 ( 1991 ) , 237 - 260 .
% to find minimal unknotting crossing data for torus knot @xmath12 + % assume p = p(3 ) and q = q(3)- + clc + clear all + p(3 ) = input(`enter value of p : ' ) ; + q(3 ) = input(`enter value of q : ' ) ; + if @xmath317 % to find number of digits in q(3 ) + str = num2str(q(3 ) ) ; + digits = double(str ) - 48 ; + nod = sum@xmath318 + else % to find number of digits in p(3 ) + str = num2str(p(3 ) ) ; + digits = double(str ) - 48 ; + nod = sum@xmath318 % nod is number of digits in min\{p(3),q(3 ) } + end + a(2 ) = p(3 ) ; + a(1 ) = q(3 ) ; + v = zeros(nod*5,ceil((p(3)-1)*(q(3)-1)+gcd(p(3),q(3))-1)/2 ) ; + % nod*5 is upper bound for n + d = zeros(nod*5,1 ) ; + count = 3 ; + for i = 3:nod*5 + count = count+1 ; + if mod(i,2 ) = = 0 + a(i ) = mod(p(i),q(i ) ) ; + m(i ) = ( p(i)-a(i))/q(i ) ; + if @xmath319 + b3 = zeros((m(i)-1)*q(i)*(q(i)-1)/2,1 ) ; + for j = 1:(m(i)-1 ) + for k = 1:(q(i)-1 ) + for g = 1:k + b3(((j-1)*(q(i)-1)*q(i)/2)+(k*(k-1)/2)+g,1)=(k*(p(3)-1))+(p(i)-j*q(i)-g ) ; + end + end + end + else + b3 = [ ] ; + end + v(i-2,1:size(b3,1 ) ) = b3 ; % for even i , b(i ) is @xmath320 row of v + d(i-2,1 ) = size(b3,1 ) ; + if mod(p(i),q(i ) ) @xmath321 @xmath322 mod(p(i),q(i ) ) @xmath3230 + p(i+1 ) = a(i ) ; + q(i+1 ) = a(i-1 ) ; + else break + end + else % if mod(i,2 ) = = 1 + a(i ) = mod(q(i),p(i ) ) ; + m(i ) = ( q(i)-a(i))/p(i ) ; + if m(i)@xmath324 1 + b1 = zeros((p(i)-1)*p(i)*m(i)/2,1 ) ; + for j = 1:m(i ) + for k = 1:(p(i)-1 ) + for g = 1:k + b1(((j-1)*(p(i)-1)*p(i)/2)+(k*(k-1)/2)+g,1)= ( ( j-1)*p(i)+k)*(p(3)-1)+p(i)-g ; + end + end + end + else + b1 = [ ] ; + end + if a(i)@xmath3251 + b2 = zeros((a(i)-1)*a(i)/2,1 ) ; + for k = 1:(a(i)-1 ) + for g = 1:k + b2(((k*(k-1))/2)+g,1)=(m(i)*p(i)*(p(3)-1))+(k*(p(3)-1))+(p(i)-g ) ; + end + end + else + b2 = [ ] ; + end + b = zeros((size(b1,1)+size(b2,1)),1 ) ; + b = union(b1,b2 ) ; % b is @xmath326 + v(i-2,1:size(b,1 ) ) = b ; % for odd i , b(i ) is @xmath320 row of v + d(i-2,1 ) = size(b,1 ) ; + if mod(q(i),p(i))@xmath327= 1 & & mod(q(i),p(i))@xmath327= p(i)-1 & & mod(q(i),p(i))@xmath327=0 + p(i+1 ) = a(i-1 ) ; + q(i+1 ) = a(i ) ; + else break + end + end + end + % -to combine all @xmath328- + n = count-1 + w = zeros(1,((p(3)-1)*(q(3)-1)+gcd(p(3),q(3))-1)/2 ) ; + if mod(n,2 ) = = 1 + w1old = 0 ; + e1old = d(1,1 ) ; + for zc = 3:2:n + w1(zc ) = m(zc)*p(zc ) ; + w1new = w1(zc)+w1old ; + w1old = w1new ; + hh = e1old ; + e1(zc ) = ( d((zc-1),1)+d(zc,1 ) ) ; + e1new = e1(zc)+e1old ; + e1old = e1new ; + ww = ( w1old*(p(3)-1))+union(v((zc-1),1:d((zc-1),1)),v((zc),1:d(zc,1 ) ) ) ; + w(1,1:d(1,1 ) ) = v(1,1:d(1,1 ) ) ; + w(1,hh+1:e1new ) = ww ; + end + else + for zc = 1 + ww = v(1,1:d(1,1 ) ) ; + w(1,1:d(1,1 ) ) = ww ; + end + w1old = 0 ; + e1old = d(1,1 ) ; + for zc = 3:2:n-1 + w1(zc ) = m(zc)*p(zc ) ; + w1new = w1(zc)+w1old ; + w1old = w1new ; + jj = e1old ; + e1(zc ) = ( d((zc-1),1)+d(zc,1 ) ) ; + e1new = e1(zc)+e1old ; + e1old = e1new ; + ww = ( w1old*(p(3)-1))+ union(v((zc-1),1:d((zc-1),1)),v((zc),1:d(zc,1 ) ) ) ; + w(1,jj+1:e1new ) = ww ; + end + mukd1 = w + mukd2 = ( ( p(3)-1)*(q(3)-1)/2)+1-w + fprintf(`mukd1 and mukd2 are minimal unknotting crossing data ' ) + end | unknotting numbers for torus knots and links are well known . in this paper , we present a method for determining the position of unknotting number crossing changes in a toric braid @xmath0 such that the closure of the resultant braid is equivalent to the trivial knot or link .
also , we provide a simple proof for the important result @xmath1 using the results obtained from this method
. [ section ] [ section ] torus knots , unknotting number , braids 57m25 |
stellar magnetism at the top of the main sequence is not typical , but neither is it an extremely rare phenomenon .
in fact about 10% of the ob - type stars display strong and stable magnetic fields @xcite .
the hot magnetic stars are mainly characterized as oblique magnetic rotators ( omr ) : a dipolar magnetic field topology with field axis misaligned with respect to the rotation axis @xcite .
the existence of such well - ordered magnetic fields are a cause of inhomogeneous photospheres , giving rise to observable photometric , spectroscopic and magnetic variability that can be explained in the framework of the omr .
early type magnetic stars are sufficiently hot to produce a radiatively driven stellar wind , that in the presence of their large - scale magnetic fields may be strongly aspherical .
the wind plasma accumulates at low magnetic latitudes ( inner magnetosphere ) , whereas it can freely propagate along directions near the magnetic poles @xcite .
observable signatures of plasma trapped inside stellar magnetospheres can be recognized in the uv spectra @xcite and in the h@xmath0 line @xcite .
the interaction of a radiatively driven wind with the stellar magnetosphere has been well studied . as examples ,
see @xcite for an application of the @xcite model to radiatively driven winds , wind compression models with magnetic fields ( wcfields ) @xcite , and the magnetically torqued disk ( mtd ) model @xcite .
such interaction causes an accumulation of hot material close to the magnetic equatorial plane , as described by the magnetically confined wind shock ( mcws ) model @xcite .
the wind plasma arising from the two opposite hemispheres collides close to the magnetic equatorial plane , shocking the plasma to radiate x - rays ( @xcite has made a recent review of the x - ray emission from magnetic hot stars ) . in the presence of a strong magnetic field
, the stellar wind plasma is confined within the stellar magnetosphere and forced to rigidly co - rotate with the star @xcite .
the prototype of a rigidly rotating magnetosphere ( rrm ) is @xmath1orie @xcite .
evidence of circumstellar matter bound to the strong stellar magnetic field has been reported in a few other cases @xcite .
the stellar rotation plays an important role in establishing the size of the rrm , and the density of the plasma trapped inside .
the rotation works in opposition to the gravitational infall of the magnetospheric plasma , leading to a large centrifugal magnetosphere ( cm ) @xcite .
the existence of a cm filled by stellar wind material is a suitable condition to give rise to non - thermal radio continuum emission that was first measured for peculiar magnetic b and a stars @xcite . in accord with the omr model ,
their radio emission is cyclic owing to stellar rotation @xcite , suggestive of optically thick emission arising from a stable rrm .
the radio emission features are characterized by a simple dipolar magnetic field topology and have been successfully reproduced using a 3d model that computes the gyrosynchrotron emission @xcite . in this model ,
the non - thermal electrons responsible for the radio emission originate in magnetospheric regions far from the stellar surface , where the kinetic energy density of the gas is high enough to brake the magnetic field lines forming current sheets .
these regions are the sites where the mildly relativistic electrons originate . in the magnetic equator , at around the alfvn radius , there is a transitional magnetospheric `` layer '' between the inner confined plasma and the escaping wind .
energetic electrons that recirculate through this layer back to the inner magnetosphere radiate radio by the gyrosynchrotron emission mechanism . the analysis of radio emissions from magnetic early - type stars is a powerful diagnostic tool for the study of the topology of their magnetospheres .
the radio radiation at different frequencies probes the physical conditions of the stellar magnetosphere at different depths , even topologies are complex @xcite .
hence , the radio emission of the hot magnetic stars provides a favored window to study the global magnetic field topology , the spatial stratification of the thermal electron density , the non - thermal electron number density , and interactions between stellar rotation , wind , and magnetic field .
in fact , the above physical parameters can be derived by comparing the multi - wavelength radio light curves , for the total and circularly polarized flux density , with synthetic light curves using our 3d theoretical model .
it is then possible , to study how such stellar properties as rotation , wind , magnetic field geometry affect the efficiency of the electron acceleration mechanism .
in particular , it is important to apply the radio diagnostic techniques on a sample of magnetic early - type stars that differ in their stellar rotation periods , magnetic field strengths , and field geometries . to this end
, we conducted a radio survey of a representative sample of hot magnetic stars using the karl g. jansky very large array ( vla ) .
these stars probe different combinations of source parameters owing to their different physical properties .
this paper presents the first results of this extensive study . here
we present the analysis of the radio emission from the fast rotating , hot magnetic star hr7355 .
we were able to reproduce multi - wavelength radio light curves for the total and the circularly polarized flux density .
the model simulation of the radio light curves , along with a simulation of the x - ray spectrum of hr7355 , are used to significantly constrain the physical parameters of its stellar magnetosphere . on this basis
, we suggest a scenario that simultaneously explains the behavior of hr7355 at both radio and x - ray wavelengths . in section[sec
: uv ] we briefly introduce the object of this study , hr7355 .
the observations used in our analyses are presented in section[sec : obs ] .
the radio properties of hr7355 are discussed in detail in section[sec : radio ] .
section[model ] describes the model , while stellar magnetosphere is presented in section[sec : magnet ] .
analysis of x - ray emission of hr7355 is provided in section[sec_xray ] .
the considerations on auroral radio emission in hr7355 are given in section section[sec : aur ] , while section[sec : sum ] summarizes the results of our work .
the early - type main sequence star ( b2v ) hr7355 ( hd182180 ) is characterized by a surface overabundance of helium @xcite .
this star evidences also a very strong and variable magnetic field @xcite .
the magnetic curve of hr7355 changes polarity twice per period , and was modeled in the framework of the omr by a mainly dipolar field , with the magnetic axis significantly misaligned with respect to the rotation axis . among the class of the magnetic early - type stars , hr7355 is an extraordinarily fast rotator .
only the b2.5v type star hr5907 @xcite has a shorter rotation period .
the rotation period of hr7355 ( @xmath2 days ) sets it close to the point break - up , giving rise to a strong deformation from spherical @xcite .
the main stellar parameters of hr7355 are listed in table [ par_star ] .
hr7355 hosts a strong and steady magnetic field , indicating the existence of a co - rotating magnetosphere @xcite suitable for giving rise to non - thermal radio continuum emission .
hr7355 has a flux density at 1.4 ghz of @xmath3 [ mjy ] as listed by the nvss @xcite . at the tabulated stellar distance
, we estimate a radio luminosity of @xmath4 [ erg s@xmath5 hz@xmath5 ] , making hr7355 one of the most luminous magnetic hot stars at radio wavelengths .
thus , hr7355 is an ideal target to study the effects of fast rotation and the high magnetic field strength for magnetospheric radio emissions . .summary of stellar parameters for hr7355 [ cols="<,^,^,>",options="header " , ] under the reasonable assumption that hr7355 and cuvir have similar non - thermal acceleration efficiencies , the higher non - thermal electron column density of hr7355
could be explained if it is characterized by a more extended acceleration region as compared to cuvir .
the magnetosphere of hr7355 is bigger than cuvir , with @xmath7 versus 2.06 r@xmath8 for cuvir @xcite , and so the linear size of the acceleration region ( @xmath9 ) will be consequently wider for hr7355 . furthermore , the b2 type star hr7355 has a higher stellar mass compared to the ap star cuvir , 6 versus 3.06 m@xmath8 @xcite .
their kepler corotation radii ( @xmath10 ) are respectively : 1.3 r@xmath11 for hr7355 , 1.9 r@xmath11 for cuvir . comparing the above estimated values of @xmath12 with the average alfvn radii , respectively 15.5 for hr7355 , and 14.5 [ r@xmath11 ] for cuvir .
the ratio @xmath14 for the hr7355 cm magnetosphere is @xmath15 , versus @xmath16 in the case of cuvir .
the above estimation highlights that hr7355 is characterized by a larger magnetospheric volume maintained in rigid co - rotation compared with that of cuvir .
we also compare the magnetic field strength at the alfvn radius for both stars .
the two stars have similar rotation periods ( @xmath2 d ) , but hr7355 has a larger size , and a stronger polar magnetic field strength ( 11.6 kg versus 3.8 kg , @xcite ) .
the radial dependence for a simple magnetic dipole at the equatorial plane is described by @xmath17 [ gauss ] , and is plotted in fig .
the ranges of the allowed @xmath18 values , given in units of solar radii , are shown for both stars .
the corresponding magnetic field strengths are derived .
[ b_r ] makes clear that the current sheet region of hr7355 is characterized by a magnetic field strength of roughly double the value of the case for cuvir . from a purely qualitative point - of - view
, it is reasonable to assume the non - thermal acceleration process operates within a thicker middle magnetosphere for hr7355 . at the distances of the two analyzed stars , @xmath19 pc for hr7355 and @xmath20 pc for cuvir , their radio luminosities are respectively
, @xmath21 [ erg s@xmath5 hz@xmath5 ] , obtained using the average radio flux density measured in this paper ; and @xmath22 [ erg s@xmath5 hz@xmath5 ] , using the mean of the measured flux densities listed in @xcite .
as discussed above , the two stars are characterized by different radio emitting volumes .
for hr7355 the non - thermal electrons also travel within magnetospheric regions at higher magnetic field strength . using a model for gyrosynchrotron emission ,
the magnetic field strength directly affects the observed radio flux density level @xcite . taking into account the various physical differences , we are able to explain qualitatively why hr7355 is a brighter radio source as compared to cuvir .
the x - ray flux of hr7355 in the 0.210kev band measured by the _ xmm - newton _ is @xmath23erg @xmath24 s@xmath5 ( see table[par_xray ] ) .
this is orders of magnitude higher that may be expected if plasma emitting in radio regime would be solely responsible for the x - ray generation using the average temperature and emission measure listed in table [ tab_wind ] , the expected x - ray flux is only @xmath25 erg @xmath24 s@xmath5 .
thus , a cold thermal plasma component responsible for the radio emission alone can not explain the observed x - rays from hr7355 . lc @xmath26 [ @xmath27 @xmath24 ] & 3.2 + + @xmath29 [ kev ] & @xmath30 + @xmath31 [ @xmath32 @xmath33 & @xmath34 + @xmath35 [ kev ] & @xmath36 + @xmath37 [ @xmath32 @xmath33 & @xmath38 + @xmath39 [ kev ] & 3.6 + flux@xmath40 [ @xmath41 erg @xmath24 s@xmath5 ] & 1.6 + + @xmath29 [ kev ] & @xmath42 + @xmath31 [ @xmath32 @xmath33 & @xmath43 + @xmath0 & @xmath44 + @xmath45 [ photons kev@xmath5 @xmath24 s@xmath5 at 1 kev ] & @xmath46 + flux@xmath40 [ @xmath41 erg @xmath24 s@xmath5 ] & 1.7 + @xmath47 [ erg s@xmath5 ] & @xmath48 + @xmath49 & @xmath50 + @xmath51 [ hz ] & @xmath52 + correspond to the ism hydrogen column density dereddened ; in the 0.210 kev band comparing the x - ray and the radio emission of hr7355 to that of late - type stars reveals significant differences ( see table [ par_xray ] ) .
hr7355 violates the empirical relation coupling the x - ray and radio luminosities of magnetically active stars ( @xmath53 hz , @xcite ) , which is valid among stars distributed within a wide range of spectral classes ( from f to early m stars ) .
this is clear evidence that the physical mechanisms for the radio and x - ray emissions operating in an early b - star like hr7355 are distinct from coronal mechanisms operating in the intermediate- and low - mass main sequence stars . yet , somewhat surprisingly , the deviation of the early type hr7355 from the gudel
benz relation is similar to that for the stars at the bottom of the main sequence the ultra cool dwarfs with spectral type later than m7 @xcite .
these important similarities between active ultra cool dwarfs and a strongly magnetic b star indicate that radio and x - ray emission in their magnetospheres may be produced by related physical mechanisms and provide useful hints for the latter .
according to the mcws model , the thermal plasma responsible for x - ray emissions from magnetic b - type stars is produced by stellar wind streams colliding at the magnetic equator .
the radio wavelengths are instead sensitive to only the cold thermal plasma that accumulates at the higher magnetic latitudes .
consequently , the x - ray emission provides a different set of constraints on the physical conditions in the magnetosphere of hr7355 .
we have analyzed the archival _ xmm - newton _ measurements .
first , the observed spectrum in 0.2 - 10.0kev band was fit with a thermal two - temperature spectral model that assumes optically thin plasma in collisional equilibrium .
the fit is statistically significant , with reduced @xmath54 for 88 degrees of freedom .
the model fit parameters are shown in table [ par_xray ] .
the two - temperature components are well in accord with the values listed by @xcite .
the thermal plasma is extraordinarily hot , with the bulk of the plasma at a temperature 3.6kev ( 40mk ) .
this is significantly hotter than usually found in magnetic b - stars @xcite . in the framework of the mcws model ,
the wind plasma streams that collide at the magnetic equator give rise to a shock that heats the plasma .
hence , the maximum temperature follows from a rankine - hugoniot condition and can not exceed a value determined by the maximum stellar wind velocity . in the analysis presented in sect.[sec : radiodiagnostic ] , we assumed two distinct wind velocities , 500kms@xmath5 and 1000kms@xmath5 , that encompass values plausible for a main sequence b - type star @xcite . using eq.10 from @xcite
, we estimate the maximum plasma temperatures that can be produced via a magnetically confined wind shock for these two wind speeds ; these temperatures are respectively : 3.5mk and 14mk significantly lower than that deduced from the x - ray spectral analysis
. this led us to conclude that the assumption of the hard part of x - ray spectrum being produced by the hot thermal plasma is not realistic .
therefore , as a next step , we attempted to fit the hr7355 spectra with an absorbed power - law model , however no satisfactory fit could be obtained .
finally , we fit the observed x - ray spectrum by combining thermal and power - law models .
the resulting fit , corresponding to the power - law plus thermal model , is shown in fig .
[ fig_xray ] . a high - quality fit
with reduced @xmath55 for 89 degrees of freedom was obtained .
based on spectral fitting , the 2 t thermal model has no preference over a model that combines thermal and power - law ( non - thermal ) components .
the model fit parameters are shown in table [ par_xray ] .
the temperature of the thermal x - ray plasma in this combined model , @xmath56 mk , is easier to reconcile with a typical wind velocity of a b2v star .
a more complex model involving two temperatures plus a power - law , can also be fit to the observed spectra , yielding marginally better fitting statistics , however we choose the simplest models . to investigate on the origin of the power - law x - ray component in the hr7355 spectrum , it is useful to consider the sun .
the solar flare x - ray spectrum usually shows a hard x - ray component with a power - law energy distribution @xcite , well explained as bremsstrahlung from a non - thermal electron population @xcite . for the non - thermal bremsstrahlung emission , the observed x - ray spectral index ( @xmath0 )
is related to the spectral index ( @xmath57 ) of the injected non - thermal electrons . for a thick - target bremsstrahlung emission , @xmath57 and @xmath0
are related as follow : @xmath58 . in the case of hr7355 , if we assume that its power - law x - ray emission is generated by the impact with the stellar surface of the same non - thermal electron population responsible of the gyro - synchrotron radio emission , it is possible to estimate a value @xmath59 of the spectral index of these energetic electrons .
the simulation of the radio emission of hr7355 indicates that the non - thermal electrons have a spectral index of @xmath602.5 .
the above range of values is not quite consistent with the value derived by the spectrum of the x - ray photons , but is however close .
this suggests that the thick - target bremsstrahlung emission from a non - thermal electron population that impacts with the stellar surface is a plausible explanation for the origin of the power - law component detected in the x - ray spectrum of hr7355 .
similarly to the solar case , it is likely that not only electrons , but also protons are accelerated in hr7355 . in some solar flares the observable effects of these two different populations of high energy particles have been recognized in the x- and the gamma - ray domain .
the electrons radiate hard x - rays by thick - target bremsstrahlung emission , whereas the protons , that interact with the ions at the stellar surface , radiate gamma - rays ( see @xcite and references here reported ) .
these effects , if present , will open a new high energy observational window to the hot magnetic stars .
the mechanism responsible for this power - law x - ray spectral component has an important difference with the case of solar flares . during a solar flare ,
the energetic electrons are impulsively injected , whereas for hot star magnetospheres , the non - thermal electrons are continuously accelerated .
the latter mechanism is similar to the auroras from the magnetized planets in the solar system .
thus , we suggest that the x - ray emission from hr7355 is physically analogous to the x - ray from jupiter s aurora as measured by _ xmm - newton _ in november 2003 . the jovian x - ray spectrum has been modeled using a combination of thermal and power - law components @xcite .
the power - law x - ray component dominates at the high energies of the _ xmm - newton _ spectral range ( x - ray photons with energies higher then 2 kev ) , and is explained in terms of bremsstrahlung emission of the precipitating electrons with energies of @xmath61 kev @xcite .
this high - energy electron population is generated far from the planet ( 2030 jupiter radii ) , in the co - rotation breakdown region of jupiter s magnetosphere .
the mechanism we propose for the non - thermal electron acceleration in hot magnetic stars resembles the acceleration of high - energy electrons in jupiter s atmosphere . for the stellar magnetospheres , the co - rotation background region coincides with the equatorial current sheet outside the alvn radius ( see fig .
[ modello ] ) .
the non - thermal electrons are responsible for the gyro - synchrotron radio emission of hr7355 and have a power - law energy distribution with a low - energy cutoff at 100kev .
these particles can be , also , responsible for the power - law component in the hr7355 x - ray spectrum , as a consequence of the bremsstrahlung at the stellar surface . in this case
, the bremsstrahlung x - ray emission arises from an annular region around the pole .
the existence of a well defined spatial location of the hard x - ray source region could produce a smooth modulation of the x - ray emission as the star rotates .
unfortunately the short archival _ xmm - newton _ observations did not sample the stellar rotation period , so no conclusions can be drawn yet about the variability of the power - law component in the hr7355 x - ray spectrum .
we intend to remedy this observational shortcoming in future . in figure [ sezione_x ] we present the model sketch that may explain simultaneously the radio and the x - ray emission of hr7355 .
the thermal x - ray component arises from the hot plasma that is shocked by the impact between stellar wind flows from opposing stellar hemispheres .
the colder thermal plasma that accumulates in the the inner - magnetosphere does not significantly contribute to the x - rays , but explains the rotational modulation of the radio emission .
the stellar radio emission originates from a non - thermal electron population moving inside the magnetic cavity that is defined by the field lines that intercept the magnetic equator outside the alvn radius , coinciding with the magnetospheric regions where the co - rotation breakdown take place , and where the current sheet systems accelerate the electrons up to relativistic energies .
these precipitating non - thermal electrons could give rise to stellar auroral signatures , like the non - thermal x - ray emission from the polar caps .
the non - thermal electron population injected in the stellar magnetosphere close to the alfvn radius , and which moves toward the star , is responsible for the incoherent gyro - synchrotron radio emission of hr7355 .
the electrons with velocities almost parallel with the magnetic field lines ( i.e. , low pitch angle ) can deeply penetrate within the stellar magnetosphere , and are responsible for the x - ray auroral spectral features , as discussed in sec .
[ sec_xray ] .
non - thermal electrons that impact the stellar surface are lost to the magnetosphere . as a consequence ,
the distribution of the non - thermal electrons reflected outside ( by the magnetic mirroring ) will be deprived of electrons with quite low pitch - angle values .
this is a suitable condition to develop the unstable electron energy distribution , known as the loss - cone distribution @xcite .
this inverted electron velocity distribution can trigger the coherent electron cyclotron maser ( ecm ) emission mechanism .
the ecm amplifies the extraordinary magneto - ionic mode , producing highly circularly polarized radiation ( @xmath62 ) , at frequencies close to the first few harmonics of the local gyro - frequency ( @xmath63 ghz ) ; however , the fundamental harmonic is probably suppressed by gyro - magnetic absorption effects @xcite .
the ecm is the process that generates the broad - band auroral radio emission of the magnetized planets of the solar system @xcite , of the two magnetic stars cuvir @xcite and hd133880 @xcite , and , of the ultra cool dwarfs @xcite .
the auroral radio emission arises from the thin density - depleted magnetic cavity related to the auroral oval at the polar caps .
this kind of coherent radio emission is efficiently amplified within a narrow beam pattern tangentially directed along the cavity wall ( i.e. , the laminar source model , @xcite ) , giving rise to a radio light house effect . in the case of the earth auroral kilometric radiation ( akr ) , this highly directional coherent radiation is upward refracted by the dense thermal plasma trapped outside the auroral cavity @xcite . the auroral radio emission from cuvir was first discovered at 1.4 ghz as intense ( @xmath64 order of magnitude brighter then the incoherent radio emission ) radio pulses that were 100% circularly polarized @xcite .
this amplified emission from cuvir has also been detected at 2.5 ghz @xcite , and at 600 mhz @xcite .
the pulse arrival times are observed to be a function of the observing frequency @xcite , a signature of frequency - dependent refractive effects suffered by the cuvir auroral radio emission .
the radio pulses , related to the auroral radio emission from a dipole - shaped magnetosphere , occur when the magnetic dipole axis lies in the plane of the sky ( cross - over phases ) , with a duration of 510% of the rotational period .
the features of the auroral radio emission arising from the cuvir magnetosphere has been successfully modeled using a simple dipole shape @xcite .
our source hr7355 has a stellar geometry , rotation , and magnetic dipole obliquity suitable for detection of its auroral radio emission .
in fact , as evident from the top panel of fig . [ fase_coverage ] , the curve for the magnetic field shows a change of net polarity twice per rotation .
the phase coverage of the radio observations presented in this paper has good sampling at the cross - over phases at each observing frequency ( see bottom panel of fig . [ fase_coverage ] ) . at the frequencies ranging from 6 to 44 ghz
, the radio measurements of hr7355 do not show any hint of auroral radio emission .
as previously explained , the stellar auroral radio emission at a given frequency originates in the magnetospheric regions where the second harmonic of the local gyro - frequency is close to the observing frequency . in the case of cuvir ( @xmath65 [ gauss ] , @xcite ) ,
emission from the ecm arises from magnetospheric layers located between @xmath64 and @xmath66 stellar radii above the stellar surface . scaling to the polar strength of hr7355 ( @xmath67 [ gauss ] , @xcite ) ,
the selected observing bands , tuned at the frequencies of the second harmonic of the gyro - frequency , arise from magnetospheric layers with @xmath68 ranging between 1.14 and 2.21 r@xmath11 ( see the right @xmath69-axis of the bottom panel of fig . [ fase_coverage ] ) , that correspond to layers that are located at heights lower than @xmath70 stellar radii from the surface of the star . on the other hand , from the model simulations of the radio light curves of hr7355 ( this paper ) and cuvir @xcite
, we note that the thermal plasma trapped within the magnetosphere of hr7355 has a higher density as compared to cuvir .
the auroral radio emission arising from the deeper magnetospheric layers of hr7355 may suffer absorption effects .
up to now , we have no radio measurements at frequencies below 6 ghz , corresponding to layers further out in radius where conditions would be more conducive for the detection of auroral radio emission from hr7355 .
consequently , it is not possible at this time to draw firm conclusions regarding the production of auroral radio emission at hr7355 .
in this paper we have presented an extensive analysis of the rigidly rotating magnetosphere of the fast rotating b2v star , hr7355 .
this study has been made using radio ( vla ) and x - ray ( _ xmm - newton _ ) observations .
the radio measurements of hr7355 cover a large frequency range , from 6 to 44 ghz .
the total ( stokes i ) and the circularly polarized ( stokes v ) flux density are variable .
the radio data have been phase folded using the stellar rotation period , demonstrating that the radio variability is a consequence of the rotation .
the rotational phases have been well sampled , allowing us to build multi - wavelength radio light curves separately in stokes i and v. modeling of the stellar radio emission , using a 3d model that simulates the radio emission from a rigidly rotating stellar magnetosphere shaped by a simple dipole @xcite , allows us to constrain the physical conditions in the magnetosphere of hr7355 . as result of the present analysis , the wind mass - loss rate of hr7355 has been indirectly derived .
independently , we obtained constrains on mass - loss rate from the analysis of archival uv spectra of hr7355 by means of the non - lte stellar atmosphere model powr .
the radio and uv values of mass - loss rate are in good agreement , and are in accord with estimates of mass - loss rates derived from the uv spectra of other stars with similar spectral types .
the average radio luminosity of hr7355 is about @xmath71ergs@xmath5hz@xmath5 , in the range 644 ghz , making it one of the brightest radio sources among the class of the mcp stars ( mean radio luminosity @xmath72 [ ergs s@xmath5 hz@xmath5 ] @xcite ) . to investigate further ,
the magnetosphere of hr7355 was compared with that of cuvir , another magnetic star studied with the same modeling approach .
the comparison reveals that both these stars are characterized by centrifugal magnetospheres but , the same time , have significant differences .
the cm of hr7355 , as normalized to the stellar radius is larger , and the regions where the non - thermal electrons are generated are characterized by a stronger local magnetic field strength , with respect to the case of cuvir , with a consequent effect on the radio brightness of the two stars . the analysis presented in this paper allows us to estimate the average physical conditions of the thermal plasma confined within the magnetospheric region .
absorption effects by the trapped plasma influences the emerging stellar radio emission and plays a key role for the modeling of the radio light curves .
the measured thermal x - ray emission from hr7355 could be explained as a consequence of the shock heating of the colliding wind streams arising from the two opposite stellar hemispheres .
the fit to the x - ray spectrum of hr7355 suggests a presence of a non - thermal x - ray component described by a power law .
the spectral index of the non - thermal x - ray photons is compatible with the thick target bremsstrahlung emission generated by the non - thermal electron population , which are responsible for the observed radio emission , that impact with the stellar surface close to the polar caps .
this could be the signature for auroral x - ray emission from hr7355 .
stellar rotation can lead to greater x - ray emissions than predicted by the scaling laws in the framework of the mcws model @xcite .
however , @xcite point out that even taking rotation into account , there are some hot magnetic stars that are too bright in the x - ray band , one of them being hr7355 .
our new model suggests that auroral x - ray emission is a likely additional mechanism that increases x - ray production and can account for the strong x - rays from hr7355 .
we speculate that auroral mechanism operates in other hot magnetic stars that display hard and bright x - ray emission .
the auroral phenomenon also gives rise to features at radio wavelengths , such as coherent pulses with @xmath73 100% circular polarization that occur at predictable rotational phases .
this radio lighthouse phenomena has been recognized in the magnetized planets of the solar system , among some ultra cool dwarfs , and in two hot magnetic stars , with the prototype cuvir . for the stellar auroral radio emission to be detectable
, the stellar magnetic dipole must to be oriented with the axis lying in the plane of the sky ( null effective magnetic field ) .
the stellar geometry of hr7355 is favorable for detecting this coherent emission , in fact the magnetic field curve inverts the sign twice per period .
we do not however find any signature of auroral radio emission from hr7355 , at least in the frequency range 644 ghz .
for this frequency range , we suggest that the auroral radio emission originates deep inside the stellar magnetosphere and is strongly absorbed . new observations at lower frequencies , corresponding to less deep layers , could reveal whether auroral radio emissions are in fact produced the magnetosphere of hr7355 .
we want to underscore that the synergistic radio and x - ray analysis is a powerful combination that can led to strong constraints for the stellar magnetospheric conditions of hot magnetic stars . from the results of the radio modeling simulations and the x - ray spectral analysis of hr7355
, we have been able to outline a physical scenario that simultaneously explains features detected at opposite ends of the source spectrum .
we thank the referee for their very useful remarks that helped to improve the paper .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
this work has extensively used the nasa s astrophysics data system , and the simbad database , operated at cds , strasbourg , france .
this publication used data products provided by the xmm - newton science archive .
lmo acknowledges support by the dlr grant 50or1302 .
branduardi - raymont g. , bhardwaj a. , elsner r.f . , gladstone g.r .
, ramsay g. , et al . , 2007 , a&a , 463 , 761 branduardi - raymont g. , elsner r.f . , galand m. , grodent d. , cravens t.e . , et al . , 2008 , j. geophys .
res . , 113 , a2202 | in this paper we investigate the multiwavelengths properties of the magnetic early b - type star hr7355 .
we present its radio light curves at several frequencies , taken with the jansky very large array , and x - ray spectra , taken with the _ xmm - newton _ x - ray telescope .
modeling of the radio light curves for the stokes i and v provides a quantitative analysis of the hr7355 magnetosphere .
a comparison between hr7355 and a similar analysis for the ap star cuvir , allows us to study how the different physical parameters of the two stars affect the structure of the respective magnetospheres where the non - thermal electrons originate .
our model includes a cold thermal plasma component that accumulates at high magnetic latitudes that influences the radio regime , but does not give rise to x - ray emission .
instead , the thermal x - ray emission arises from shocks generated by wind stream collisions close to the magnetic equatorial plane .
the analysis of the x - ray spectrum of hr7355 also suggests the presence of a non - thermal radiation .
comparison between the spectral index of the power - law x - ray energy distribution with the non - thermal electron energy distribution indicates that the non - thermal x - ray component could be the auroral signature of the non - thermal electrons that impact the stellar surface , the same non - thermal electrons that are responsible for the observed radio emission . on the basis of our analysis
, we suggest a novel model that simultaneously explains the x - ray and the radio features of hr7355 and is likely relevant for magnetospheres of other magnetic early type stars .
[ firstpage ] stars : early - type stars : chemically peculiar stars : individual : hr7355 stars : magnetic field
radio continuum : stars x - rays : stars . |
integrals over spaces of connections play an important role in modern gauge theory , but as these spaces are infinite - dimensional , it is often difficult to make heuristic computations involving such integrals rigorous .
suppose one has a smooth principal @xmath0-bundle @xmath8 , and let @xmath2 be the space of smooth connections on @xmath5 .
then @xmath2 is an affine space , and becomes a vector space after an arbitrary choice of some point as origin , so initially it may be tempting to integrate functions using some sort of ` lebesgue measure ' on @xmath2 .
unfortunately , various theorems @xcite indicate that there are no well - behaved translation - invariant measures on an infinite - dimensional vector space .
one might then restrict ones ambition to integrating ` cylinder functions ' and certain limits thereof .
a cylinder function on @xmath2 is one that depends on finitely many coordinates , that is , one of the form @xmath9 where @xmath10 are continuous linear functionals and @xmath11 is bounded and continuous .
to integrate these all one needs is a ` cylinder measure ' ; the theory of these is well - developed and widely used in probability theory , quantum mechanics and quantum field theory @xcite .
however , in gauge theory the fact that @xmath2 is an affine space is in many ways less important than the fact that the group @xmath7 of gauge transformations acts on it .
for example , one is often interested in integrating gauge - invariant functions on @xmath2 .
this amounts to doing integrals on @xmath12 , an infinite - dimensional space which is not an affine space , nor even a manifold , but a kind of stratified space . in applications to physics
@xcite one is often interested in integrating ` wilson loops ' , gauge - invariant functions of the form @xmath13 where @xmath14 is a smooth loop in @xmath5 , @xmath15 denotes the holonomy of @xmath4 around @xmath14 , and the trace is taken in some finite - dimensional representation of @xmath0 .
wilson loops are typically not easy to approximate by cylinder functions unless @xmath0 is abelian , so it is difficult to integrate them against cylinder measures . motivated by rovelli and smolin s work @xcite on the loop representation of quantum gravity , ashtekar and isham @xcite introduced a nonlinear version of the cylinder measure idea which is specially adapted to gauge theory . taking advantage of subsequent reworkings , we may describe their idea as follows .
_ redefines _ a ` cylinder function ' on @xmath2 to be one of the form @xmath16 where @xmath17 are piecewise smooth paths in @xmath5 , the holonomy @xmath18 of the connection @xmath4 along @xmath17 is identified with an element of @xmath0 by means of a trivialization of @xmath3 over the endpoints of @xmath17 , and @xmath19 is continuous .
taking the completion of this algebra in the @xmath20 norm @xmath21 one obtains a c*-algebra of bounded continuous functions on @xmath2 .
then one defines a ` generalized measure ' @xmath22 on @xmath2 to be a bounded linear functional on this c*-algebra . using the gelfand - naimark spectral theory
, this c*-algebra can be identified with the c*-algebra of all continuous functions on a compact space @xmath23 of which @xmath2 is a dense subset .
elements of @xmath24 are called ` generalized connections ' , and the holonomy of one of these generalized connections along a piecewise smooth path is still well - defined . by the riesz - markov theorem ,
generalized measures on @xmath2 can be identified with finite regular borel measures on @xmath24 .
one might hope for some generalized measure on @xmath2 to serve as a substitute for the nonexistent ` lebesgue measure ' on @xmath2 . at the bare minimum
one would like this generalized measure to be invariant under all automorphisms of the bundle @xmath3 e.g. , gauge transformations and lifts of diffeomorphisms of @xmath5 . in a search for something along these lines , ashtekar and lewandowski
@xcite discovered that the study of generalized measures becomes more manageable when one works with cylinder functions defined using piecewise analytic paths .
technically , the difficulty with piecewise smooth paths is that they can intersect in very complicated ways even in a cantor set .
piecewise analytic paths , on the other hand , can only intersect in an infinite set if they overlap for some closed interval .
this turns out to greatly simplify matters .
after further work by ashtekar , lewandowski and baez @xcite , the theory of generalized measures in the analytic context now looks as follows .
one assumes @xmath5 is a real - analytic manifold , @xmath0 is a compact connected lie group , and @xmath8 is a smooth principal @xmath0-bundle .
one works only with cylinder functions for which the paths @xmath17 are piecewise real - analytic . letting @xmath25 denote the completion of this space of cylinder functions in the @xmath20 norm , one then defines a ` generalized measure ' to be a bounded linear functional on @xmath25 .
the results of ashtekar and isham still hold : @xmath25 is isomorphic to the c*-algebra of continuous functions on a space @xmath24 containing @xmath2 as a dense subset , and generalized measures on @xmath2 are the same as finite regular borel measures on @xmath24 . in the analytic context , however , it is not too hard to construct a canonical generalized measure on @xmath2 , the ` uniform ' generalized measure .
this generalized measure is invariant under all automorphisms of the bundle @xmath3 that act on the base manifold @xmath5 as real - analytic diffeomorphisms .
the uniform generalized measure is not the only one invariant under all these automorphisms .
in fact , many such generalized measures exist , and they can be constructed and in a rather abstract sense classified using the notion of an ` embedded graph ' .
an embedded graph @xmath26 is a finite set of analytic paths @xmath27 \to m$ ] that are 1 - 1 , embeddings when restricted to @xmath28 , and nonintersecting except possibly at their endpoints .
these paths are called the ` edges ' of the graph .
one can reduce the study of holonomies along any finite set of real - analytic paths to the case of a graph , because given any such set of paths , one can write them as finite products of the edges of some graph ( and their inverses ) .
given an embedded graph @xmath26 with @xmath29 edges , and trivializing @xmath3 over the endpoints of all the edges , a generalized measure @xmath22 on @xmath2 determines a measure @xmath30 on @xmath31 by @xmath32 where @xmath33 is related to @xmath34 by equation ( [ cyl ] ) .
the measures @xmath30 for all embedded graphs @xmath26 determine the generalized measure @xmath22 , and the uniform generalized measure on @xmath2 is the unique one for which all the measures @xmath30 are normalized haar measure on some product of copies of @xmath0 .
it is natural to wonder whether these results depend crucially on the use of analytic paths .
this is not a question of merely technical interest .
one might argue that the analyticity assumptions are not so bad , since every paracompact smooth manifold admits a real - analytic structure , which is unique up to smooth diffeomorphism @xcite .
however , in applications to topological quantum field theory and the loop representation of quantum gravity , diffeomorphism - invariance plays a key role , and real - analytic diffeomorphisms do behave very differently from smooth ones .
after all , a real - analytic diffeomorphism of a connected manifold is completely determined by its restriction to an arbitrarily small neighborhood . to see how this impinges on questions of real physical interest , it is interesting to read the recent work of ashtekar , lewandowski , marolf , mouro and thiemann on diffeomorphism - invariant gauge theories @xcite .
the goal of this paper is to treat the case where @xmath5 is merely _ smooth_. we work with cylinder functions on @xmath2 for which the paths @xmath17 are ` curves ' , that is , _
piecewise smoothly immersed _ , and we let @xmath35 denote the completion of the space of these cylinder functions in the @xmath20 norm . for us ,
a generalized measure will be a continuous linear functional on @xmath35 . note that if @xmath5 is real - analytic then @xmath36
, so any of our generalized measures restricts to a generalized measure as defined in the real - analytic context .
in particular , we construct a generalized measure that is invariant under all automorphisms of the bundle @xmath3 , and which restricts to the uniform generalized measure when @xmath5 is real - analytic .
we again call this the ` uniform ' generalized measure .
in fact , this uniform generalized measure was already constructed by ashtekar and lewandowski @xcite in the case @xmath37 , using special properties of abelian lie groups . from this point of view , the advance of the present paper consists of being able to handle nonabelian groups .
but our work also establishes a framework for handling other generalized measures in the smooth context .
the main ideas behind this framework are as follows .
in analogy with the analytic case , for every family of curves @xmath38 a generalized measure @xmath22 on @xmath2 determines a measure @xmath30 on @xmath31 by @xmath32 where @xmath33 is related to @xmath34 by equation ( [ cyl ] ) .
the goal is thus to reconstruct a generalized measure @xmath22 starting from such a measure @xmath30 for every family @xmath26 . of course
, some conditions must hold for such a collection of measures @xmath30 to come from a generalized measure on @xmath2 .
in particular , not all @xmath29-tuples of elements of @xmath0 can be simultaneously attained as the holonomies of some fixed connection along the curves in @xmath26 , but only those lying in some subset @xmath39 .
to come from a generalized measure , for each family @xmath26 the measure @xmath30 will need to be supported on this ` attainable subset ' @xmath40 , so we need a good understanding of this subset . in particular , in contrast to the analytic case , we can not reduce the problem to considering nice families such as embedded graphs for which @xmath41 . the reason why @xmath40 may not be all of @xmath31
is that there may be relations among the holonomies along the curves in the family @xmath26 .
these relations occur when the curves overlap for some open interval , so we need to introduce a notion of a ` type ' of possible overlap .
due to the complicated ways curves can intersect in the smooth context , a given type may occur infinitely often in a family @xmath26 ; for a simple example see figure [ fg : type ] .
@xmath42 the goal of section 2 is to describe the possible holonomies of a family of curves .
this is done first for especially well - behaved families of curves called ` tassels ' .
roughly , a tassel based on a point @xmath43 is a family of curves for which , when it is restricted to any neighborhood of @xmath44 , the same types of overlap still occur . this self - similarity forces @xmath40 to be a subgroup of @xmath31 , in fact a subgroup easily presented in terms of the types of overlap occurring in @xmath26 .
then we introduce the notion of a ` web ' .
this is a family @xmath45 of curves that can be written as a union of tassels @xmath46 , sufficiently separated so that @xmath47 .
we will show that the holonomies along any family can be expressed in terms of the holonomies along some web , thus giving an algebraic description of the possible holonomies .
the proofs of these facts , that is of propositions 1 and 2 , can safely be skipped by a reader looking for an initial overview of the results of this paper .
in fact , if the family one started with consisted of the edges of an embedded graph , the tassels this construction would produce would be the restriction of the edges to each cell in a cell decomposition dual to the graph .
thus each tassel would contain one vertex @xmath44 of the graph , and would in fact be based at @xmath44 .
one should therefore think of a web as a generalization of a finite graph , and a tassel based at @xmath44 as a generalization of a neighborhood of a vertex @xmath44 .
in section 3 we give a criterion for a collection of measures @xmath48 , one for each web @xmath45 , to arise from a generalized measure @xmath22 on @xmath2 .
we also show that @xmath22 is uniquely determined by the measures @xmath48 , so that we have a tool for constructing generalized measures .
in section 4 we apply this tool to construct the uniform generalized measure . in recent work on the loop representation of quantum gravity , ` spin network states ' play an important role @xcite .
these have already been dealt with rigorously in the analytic context , and in section 5 we describe how they work in the smooth context .
the basic idea is as follows .
using the uniform generalized measure @xmath49 on @xmath2 , one can define a hilbert space @xmath50 by completing @xmath35 in the norm associated to the inner product @xmath51 the group @xmath7 of gauge transformations acts on @xmath2 , and this gives rise to a unitary representation of @xmath7 on @xmath50 .
we define @xmath6 to be the subspace of @xmath7-invariant vectors in @xmath50 . the `
spin network states ' form a very explicit ` local ' orthonormal basis of @xmath6 , which is to say an orthonormal basis of the subspace associated to each web @xmath45 . in the analytic context , they are formed using embedded graphs whose edges are labeled with irreducible unitary representations of @xmath0 , and whose vertices are labeled with intertwining operators from the tensor product of the representations labeling the ` incoming ' edges , to the tensor product of the representations labeling ` outgoing ' edges . in the smooth context
spin networks are formed using webs equipped with similar , but more subtle , representation - theoretic data .
an embedded graph is a special case of a web , and in this case our spin network states reduce to the spin network states as defined in the analytic context .
however , it is not yet clear whether the spin networks can be combined in a simple fashion to give an orthonormal basis of all of @xmath6 simultaneously , as in the analytic case .
fix a connected compact lie group @xmath0 , a smooth ( paracompact ) @xmath52-dimensional manifold @xmath5 , and a smooth principal @xmath0-bundle @xmath53 . by a _ curve _ we mean a piecewise smooth map from a finite closed interval of @xmath54 to @xmath5 that is an immersion on each piece .
two curves are considered _ equivalent _ if one is the composition of the other with a positive diffeomorphism between their domains .
a _ family _ of curves is a finite set of curves with a chosen ordering @xmath55 .
if @xmath26 is such a family , let @xmath56 be the union of the ranges of the individual curves .
if @xmath57 \to m$ ] and @xmath58 \to m$ ] are two curves such that @xmath59 , we can form the _ product
_ @xmath60 by gluing them together at this common point .
of course this is defined only up to equivalence of curves .
it is exactly like the product in the fundamental groupoid , except that we do not identify homotopic curves .
it is still associative , however , and there is a category whose objects are points in @xmath5 and whose morphisms ( other than identity morphisms ) are equivalence classes of curves .
define the _ inverse _ @xmath61 of a curve @xmath62 to be @xmath62 reparametrized by an order - reversing map , again up to equivalence .
this is not truly an inverse for the product , but merely a contravariant functor .
if every curve in the family @xmath26 is equivalent to a ( finite ) product of curves in the family @xmath63 and their inverses , we say that @xmath26 _ depends _ on @xmath63 .
we say that a collection of families of curves @xmath64 is _ independent _ if when @xmath65 , any curve in the family @xmath66 intersects any curve in the family @xmath67 , if at all , only at their endpoints , and there is a neighborhood of each such intersection point whose intersection with @xmath68 is an embedded interval .
obviously even if two families are not independent , one may not depend on the other .
the above definitions are motivated by considering holonomies of connections along these curves .
the map from curves to holonomies given by such a connection sends product to product and inverse to inverse .
if one family of curves depends on another , one can compute the holonomy of a connection along all the curves in the first from the same information about the second . if two families are independent , knowing the holonomies along one family tells one nothing about the holonomies along the other .
a _ subcurve _ of a curve @xmath62 is a curve equivalent to the restriction of @xmath62 to a subinterval of its domain .
restriction _ of a family @xmath26 to a closed set @xmath69 is the family gotten by restricting each @xmath17 to each interval of @xmath70 $ ] .
a point @xmath71 is a _ regular point _ if it is not the image of an endpoint or nondifferentiable point of @xmath26 , and there is a neighborhood of it whose intersection with @xmath56 is an embedded interval .
a family of curves @xmath26 is _ parametrized consistently _ if each curve is parametrized so that @xmath72 implies @xmath73 .
thus each of the curves is actually an embedding , and each point @xmath44 in the range of the family is associated to a unique value of the parameter , which we call @xmath74 . if a family @xmath75 is parametrized consistently and @xmath44 is a point in @xmath56 , define the _ type _ of a regular point @xmath44 , @xmath76 , to be the lie subgroup of @xmath31 consisting of all @xmath29-tuples @xmath77 such that for some @xmath78 we have @xmath79 if @xmath44 lies on @xmath17 , and @xmath80 otherwise .
this gives a canonical isomorphism between any type and @xmath0 .
a fundamental concept in all that follows is that of a ` tassel ' . a family of curves @xmath81 is a _ tassel based on _
@xmath82 if : @xmath83 lies in a contractible open subset of @xmath5 @xmath81 can be consistently parametrized in such a way that @xmath84 is the left endpoint of every curve @xmath17 two curves in @xmath81 that intersect at a point other than @xmath44 intersect at a point other than @xmath44 in every neighborhood of @xmath44 any type which occurs at some point in @xmath83 occurs in every neighborhood of @xmath44 .
one may visualize the curves of the tassel as radiating outwards from the base @xmath44 .
see figure [ fg : example ] for an example .
@xmath85 finally , a _ web _
@xmath45 is a finite independent collection of tassels @xmath86 such that no tassel contains the base of another .
we frequently apply concepts defined for families of curves to webs without comment , using the fact that the web @xmath45 has an associated family @xmath87 . for example
, we say that a family depends on a web @xmath45 if it depends on the family @xmath88 .
our first main result about webs is : [ web1 ] any family of curves @xmath26 depends on a web @xmath45 .
we begin with a technical lemma .
[ lm : tech ] let @xmath26 be a family of smooth curves @xmath89 .
the preimage of any point in @xmath5 under any @xmath17 is finite .
every point @xmath71 has a contractible open neighborhood @xmath90 admitting coordinates @xmath91 such that for each @xmath92 , @xmath93 on @xmath94 $ ] .
given any point @xmath71 and any open neighborhood @xmath95 of @xmath44 , there is an open subneighborhood @xmath52 of @xmath44 such that for each @xmath92 , @xmath96 $ ] is a finite union of intervals , each containing a point of @xmath97 $ ] .
the set of regular points is open and dense in @xmath56 .
given any point @xmath71 and any open neighborhood @xmath95 of @xmath44 , there is an open subneighborhood @xmath52 with the properties in part ( c ) such that every point of @xmath56 lying on the boundary of @xmath52 is a regular point .
_ if not , the preimage would have an accumulation point , and at that point @xmath17 would not be an immersion .
we can choose an open neighborhood @xmath95 about @xmath44 with coordinates @xmath98 such that for all @xmath92 we have @xmath99 at all of the finitely many points in the preimage of @xmath44 under @xmath17 .
each such point has an open interval around it such that @xmath100 throughout that interval .
the union of the images of the complements of these intervals is a compact set @xmath101 not containing @xmath44 .
it follows that any contractible open neighborhood @xmath90 of @xmath44 contained in @xmath102 has the desired properties .
choose a coordinate patch @xmath90 around @xmath44 as in part ( b ) of this lemma , and consider the hyperplane through @xmath44 on which @xmath103 is constant .
the points of intersection of @xmath56 with this hyperplane are all transverse , so they have no accumulation points .
thus a small open neighborhood of @xmath44 in the hyperplane only intersects @xmath56 at @xmath44 .
shrinking this neighborhood to a sufficiently small subneighborhood , its product with a sufficiently small open interval in the @xmath103 axis is an open neighborhood @xmath52 of @xmath44 that only intersects each @xmath17 in finitely many embedded open intervals containing @xmath44 .
this choice of @xmath52 has the desired properties .
see figure [ fg : nbhd ] for an illustration .
@xmath104 consider any neighborhood @xmath95 of a point @xmath71 .
for each point in @xmath95 , consider the total number of points in the preimages of all the curves @xmath17 . by part ( a )
, we can pick a point @xmath105 for which this number is minimal .
we will show @xmath106 is regular . choose a subneighborhood @xmath52 of @xmath95 as in part ( c ) of this lemma , and small enough that each @xmath17 is 1 - 1 on each component of the preimage of @xmath52 .
each @xmath107 has at most one preimage point in each of the intervals comprising @xmath96 $ ] , and since @xmath106 was minimal and has exactly one preimage point in each of these intervals by part ( c ) , @xmath108 must have exactly one in each .
thus the images of these intervals in @xmath52 must coincide , and hence @xmath109 must be an embedded open interval in @xmath5 . since @xmath106 can not be the image of an endpoint , @xmath106 is regular
. choose @xmath52 as in the proof of part ( c ) .
recall that @xmath99 for all @xmath92 and all @xmath110 in the preimage of @xmath44 under @xmath17 . if we choose the interval @xmath111 in the @xmath103 axis used to define @xmath52 sufficiently small , each curve @xmath17 intersects the boundary of @xmath52 only at the planes where @xmath103 equals @xmath112 or @xmath113 .
moreover , choosing this interval sufficiently small guarantees that the intersection points are transversal . by part ( d ) , it follows that we can choose @xmath112 and @xmath113 such that these intersection points are all regular points .
3em 2ex _ proof of proposition [ web1 ] .
_ let @xmath26 be a family of curves ; we may assume all the curves in @xmath26 are smooth , since any family depends on a family of smooth curves . by lemma [ lm : tech](b ) and the compactness of @xmath56 , we can cover @xmath56 with open sets @xmath114 such that each @xmath115 is contractible and admits coordinates as in the lemma .
we claim that each @xmath71 has an open neighborhood @xmath116 with the following properties : @xmath116 is contained in every @xmath115 containing @xmath44 .
@xmath117 $ ] is a finite union of intervals .
@xmath26 restricted to @xmath118 depends on a tassel based at @xmath44 .
the points of @xmath56 lying on the boundary of @xmath116 are all regular points . to see this , we first tentatively take as @xmath116 the subneighborhood given by lemma [ lm : tech](e ) of the intersection of the @xmath115 s containing @xmath44
. then @xmath117 $ ] is a finite union of closed intervals , and the points of @xmath56 lying on the boundary of @xmath116 are all regular points .
use as coordinates on @xmath116 any of the coordinates on the @xmath115 s containing @xmath44 ; without loss of generality we assume that @xmath119 . by lemma [ lm : tech](b )
the restriction of @xmath26 to @xmath118 is a family consistently parametrized by the coordinate @xmath103 , and by lemma [ lm : tech](e ) each curve @xmath62 in this family has @xmath120 in its domain and @xmath121 .
take each curve in this family , and if its domain @xmath122 $ ] contains @xmath120 in its interior , replace it with the two subcurves given by its restriction to @xmath123 $ ] and the inverse of its restriction to @xmath124 $ ] . denote the resulting family by @xmath125 .
we can parametrize each curve in @xmath125 by @xmath126 , and then not only will @xmath125 be consistently parametrized , but also each curve @xmath62 in @xmath125 will have @xmath121 as its left endpoint .
note that the family @xmath26 restricted to @xmath118 depends on the family @xmath125 .
thus @xmath116 has all the properties claimed except that @xmath125 might not be a tassel . by the previous paragraph , and since @xmath127 lies in some contractible open set @xmath115 , the only way @xmath125 can fail to be a tassel is by violating conditions ( c ) or ( d ) in the definition of a tassel . to get condition ( c ) to hold ,
choose a neighborhood of @xmath44 in @xmath116 small enough that any two curves which intersect do so arbitrarily close to @xmath44 , and choose a subneighborhood as in lemma [ lm : tech](e ) .
use this subneighborhood as a new choice of @xmath116 , and restrict @xmath125 to the new @xmath118 .
this leaves us with a neighborhood @xmath116 with all the properties claimed except that @xmath125 might violate condition ( d ) in the definition of a tassel .
to get condition ( d ) to hold , note that for each type @xmath128 occurring in @xmath125 , either @xmath128 occurs at a sequence of points approaching @xmath44 , or it does not .
if it does not , choose a neighborhood of @xmath44 in @xmath116 which excludes all points of type @xmath128 , and choose a subneighborhood as in lemma [ lm : tech](e ) .
use this subneighborhood as a new choice of @xmath116 , and restrict the original family @xmath125 to the new @xmath118 , obtaining a new family @xmath125 .
since there were only finitely many distinct types in the original @xmath125 , and the new @xmath125 has fewer types , this process must end .
now @xmath125 is a tassel and @xmath116 is a neighborhood of @xmath44 having all the properties claimed .
see figure [ fg : maketassel ] for an example of this construction .
@xmath129 now cover @xmath56 with finitely many of these open sets @xmath116 .
call them @xmath130 @xmath131 , call the associated tassels @xmath132 , and call the points at which they are based @xmath133 .
we claim that if @xmath134 for some @xmath135 , then @xmath136 is still a neighborhood of @xmath137 with properties ( i - iv ) .
properties ( ii ) and ( iv ) are clear . for ( i ) , note that @xmath138 is contained in any @xmath115 containing @xmath137 , because @xmath139 is and @xmath140 so @xmath141 is as well . for ( iii ) , coordinatize @xmath138 using the coordinates on some @xmath115 containing @xmath137 , and construct a family as before , breaking
the restriction of each curve in @xmath26 to @xmath142 into two subcurves with @xmath143 and @xmath144 if necessary , and parametrizing them consistently by the value of @xmath126 . to see that this family is a tassel , the only nontrivial thing to check is condition ( d ) .
notice that any type occurring in this family in @xmath145 corresponds to a type occurring in @xmath146 , and therefore occurs arbitrarily close to @xmath137 .
any type occurring in the family in @xmath147 corresponds to a type occurring in @xmath148 , and thus occurs arbitrarily close to @xmath149 .
but then it also occurs in @xmath139 , and thus arbitrarily close to @xmath137 .
see figure [ fg : union ] for an illustration . here
bold curves are in @xmath148 , light curves are in @xmath146 , and medium weight curves are in their union .
@xmath150 using this fact , we can replace the @xmath139 by unions thereof until no @xmath139 contains @xmath149 for @xmath135 , and succeed in covering @xmath56 with open neighborhoods @xmath151 of points @xmath152 such that ( i - iv ) hold and such that @xmath153 only for @xmath154 . to prove the theorem
, we will now shrink each @xmath139 to a smaller neighborhood of @xmath137 while maintaining properties ( i - iv ) , so that the resulting neighborhoods no longer intersect , but their closures still cover @xmath56 .
when we have done this , the restriction of @xmath26 to the closure of each neighborhood will depend on a tassel by ( iii ) .
moreover , these tassels will form a web by ( iv ) .
finally , @xmath26 will depend on this web by ( ii ) . to shrink the @xmath139 in this way , first replace each @xmath141 for @xmath155 with @xmath156 , leaving @xmath157 the same
. properties ( i ) and ( iv ) for the original @xmath141 s easily imply those properties for the new @xmath141 s .
property ( ii ) holds because the new @xmath141 has only finitely many boundary points . as for property ( iii ) , recall that @xmath153 only for @xmath154 .
the only way @xmath148 restricted to the new @xmath158 could fail to be a tassel is by having a component that does not pass through @xmath149 ; but this could only happen if a curve of @xmath159 had been a subcurve of @xmath148 , in which case @xmath160 .
next replace @xmath141 with @xmath161 for each @xmath162 , and so on .
when we are done , we find that @xmath26 depends on the modified @xmath132 , which are all tassels , contain neighborhoods of the @xmath137 s , and intersect only at boundary points
. since all boundary points are boundary points of the original @xmath141 s , they are regular points of @xmath26 , and therefore satisfy the condition for boundary points of a web .
we thus obtain tassels @xmath86 forming a web on which @xmath26 depends .
3em 2ex let @xmath2 be the space of smooth connections on @xmath3 , equipped with its @xmath163 topology .
given a curve @xmath164 \to m$ ] , let @xmath165 be the set of functions @xmath166 compatible with the right action of @xmath0 on @xmath3 : @xmath167 given @xmath168 , the holonomy @xmath169 of @xmath4 along @xmath62 is an element of this set @xmath165 .
of course , a trivialization of @xmath3 at the endpoints of @xmath62 allows us to identify @xmath165 with @xmath0 , and this gives @xmath165 the structure of a smooth manifold in a manner independent of the trivialization .
note also that @xmath165 and the holonomy @xmath170 only depend on the equivalence class of @xmath62 .
more generally , if @xmath171 is a family , let @xmath172 be the subspace consisting of all elements of the form @xmath173 for some connection @xmath168 .
we call @xmath40 the space of _ connections on _ @xmath26 , and give it the subspace topology . if @xmath45 is a web consisting of tassels @xmath86 , we define @xmath174 to be @xmath175 , and again call this the space of connections on @xmath45 . note that the map @xmath176 given by @xmath177 is continuous and onto . furthermore
, if the product of curves @xmath60 exists there is a smooth map @xmath178 there is also for any curve @xmath62 a smooth map @xmath179 thus if @xmath26 depends on @xmath63 , a particular choice of a way to write each curve in @xmath26 as a product of curves in @xmath63 and their inverses gives a smooth map @xmath180 note that @xmath181 since @xmath182 is onto , it follows that @xmath183 is independent of how we write curves in @xmath26 in terms of curves in @xmath63 .
since @xmath184 is onto , it also follows that @xmath183 is onto .
now suppose @xmath22 is a finite borel measure on @xmath2 .
then for any family @xmath26 , @xmath22 pushes forward by the map @xmath184 to a finite borel measure @xmath30 on @xmath40 .
the collection @xmath185 satisfies an obvious consistency condition : whenever @xmath26 depends on @xmath63 , the measure @xmath186 pushes forward by the map @xmath183 to the measure @xmath30 .
not all collections of measures @xmath185 satisfying this consistency condition arise from finite borel measures on @xmath2 in this way , but as we shall see in section 3 , if such a collection satisfies a certain uniform bound , it arises from a generalized measure on @xmath2 .
this is essentially how we construct the uniform generalized measure .
however , it is easier in practice to construct generalized measures from collections @xmath187 where @xmath45 ranges over all webs , rather than all families .
proposition 1 is one of the results we need for this , since it allows us to express any family in terms of a web .
the second result we need is a good description of @xmath174 when @xmath45 is a web .
if @xmath81 is a tassel consisting of the curves @xmath188 , then we let @xmath189 be the smallest closed subgroup of @xmath31 containing all the types occurring in @xmath81 , which of course is a lie group .
then we have : [ web2 ] if @xmath190 is a tassel , and we fix a trivialization of @xmath3 over the endpoints of the curves @xmath17 to identify @xmath191 with a subset of @xmath31 , then we have @xmath192 . if @xmath45 is a web consisting of tassels @xmath86 , then @xmath193 . _ proof . _
first suppose @xmath81 is a tassel .
since @xmath83 is contained in a contractible @xmath95 , we can trivialize @xmath3 over @xmath83 , and by a suitable gauge transformation we choose this trivialization so that it agrees with the specified trivialization of @xmath3 over the endpoints of the curves @xmath17 .
this allows us to treat the holonomy of a connection along any of these curves from any point @xmath194 to any point @xmath195 as an element of @xmath0 .
it also allows us to treat a connection on @xmath196 as a @xmath197-valued one - form .
we claim that given finitely many disjoint neighborhoods @xmath198 intersecting @xmath83 in open intervals @xmath199 which contain no endpoints or nondifferentiable points of curves in @xmath81 , there is a connection @xmath200 on @xmath196 whose holonomy along @xmath199 is @xmath201 for any @xmath202 .
to see this , map @xmath199 to @xmath0 smoothly so that a neighborhood of its left endpoint gets sent to @xmath203 and a neighborhood of its right endpoint gets sent to @xmath201 .
pull the derivative back to @xmath199 , extend it to a smooth @xmath197-valued 1-form on @xmath204 which is trivial in a neighborhood @xmath205 of the two endpoints , and multiply it by a smooth function which is @xmath203 on @xmath199 and @xmath120 near the boundary of @xmath204 outside of @xmath205 .
this gives a connection on @xmath206 whose holonomy along @xmath199 is @xmath201 . defining @xmath200 this way on each @xmath204 and setting @xmath207
outside @xmath208 proves the claim .
notice if one of these intervals @xmath199 is of type @xmath128 , @xmath209\subset [ a , b]$ ] , and @xmath122 $ ] is in the domain of every @xmath17 , then the sequence of holonomies along the @xmath17 restricted to @xmath122 $ ] can be made to be any element of @xmath128 by the above procedure .
consider any element of @xmath189 and write it as @xmath210 where each @xmath211 is in @xmath212 , a type occurring in @xmath83 . by the definition of tassel
, we can choose a decreasing sequence ( with respect to the parameter @xmath110 ) of regular points @xmath137 of type @xmath212 , for @xmath213 , such that each @xmath214 is in the interior of the domain of every curve in @xmath26 .
choose nonintersecting neighborhoods @xmath139 , and construct a connection @xmath200 on @xmath196 which is trivial outside the @xmath139 and with holonomy @xmath210 .
thus every element of @xmath189 is the holonomy of some connection @xmath200 on @xmath196 .
moreover , since @xmath83 is closed , we can find a connection @xmath168 on all of @xmath3 which equals @xmath200 on @xmath83 , and thus has the same holonomy along each curve @xmath17 .
it follows that @xmath215 . on the other hand , consider the map @xmath216 sending each @xmath110 to @xmath217 . if @xmath195 is not defined , use @xmath218 , where @xmath219 is the right endpoint of @xmath17 , and if @xmath220 is not defined , use @xmath221 .
this is continuous except at finitely many points , namely endpoints or points of nondifferentiability of any @xmath17 .
since @xmath4 gives a @xmath197-valued one - form , we can interpret it as a map @xmath222 , so that @xmath223 is continuous except at finitely many points .
the set of @xmath110 such that @xmath224 is regular for all @xmath224 with @xmath225 is open dense , by lemma [ lm : tech](d ) .
for such a @xmath110 , @xmath226 is a sum of elements in the lie algebras of the types occurring with parameter value @xmath110 , and thus is in @xmath227 . by continuity
the range of @xmath228 is in @xmath227 except for finitely many points .
but @xmath229 is the endpoint of a curve in @xmath31 starting at the identity and having derivative @xmath230 at @xmath110 . since this curve lies entirely in @xmath189 , @xmath231
, so @xmath232 .
now suppose @xmath45 is a web consisting of tassels @xmath86 .
clearly @xmath233 , so we need merely prove the reverse inclusion .
in fact , we shall fix a trivialization of @xmath3 over the subset of @xmath5 consisting of all the endpoints of the curves in @xmath45 , so as to identify each space @xmath234 with @xmath235 , and we shall show that given @xmath236 , there is a connection @xmath168 with @xmath237 for all @xmath92 .
let @xmath238 be a contractible neighborhood containing the tassel @xmath148 .
we can choose a trivialization of @xmath3 over each @xmath238 which agrees with the above trivialization over all the endpoints of the curves in @xmath45 .
moreover , we can choose these trivializations so that they agree in a small neighborhood @xmath90 of all these endpoints .
the construction above then gives for each tassel @xmath148 a connection @xmath239 on @xmath240 such that @xmath239 has the desired holonomies along all curves in @xmath148 .
moreover , given any neighborhood @xmath241 of @xmath242 with @xmath243 , we can choose @xmath239 such that @xmath239 vanishes in @xmath241 except in an arbitrarily small neighborhood @xmath141 of finitely many points in the interiors of the curves in @xmath148 .
( here we use the trivialization of @xmath240 to think of @xmath239 as a @xmath197-valued 1-form . )
if we choose the @xmath241 s small enough that @xmath244 for @xmath245 , and choose the @xmath141 s small enough that @xmath246 for all @xmath245 , then the connections @xmath239 agree on all the overlaps @xmath247 so there exists a connection @xmath200 on @xmath248 having the desired holonomies on all curves in every tassel @xmath148 . since @xmath249 is closed there exists a connection @xmath168 that equals @xmath200 over @xmath250 , so @xmath251 for all @xmath252 as desired .
let @xmath253 be the algebra of _ cylinder functions _ on @xmath2 , that is functions of the form @xmath254 where @xmath26 is some family of curves and @xmath255 is continuous .
let @xmath35 be the completion of @xmath253 in the @xmath20 norm .
we define a _ generalized measure _ on @xmath2 to be a continuous linear functional on @xmath35 . given a generalized measure @xmath22 on @xmath2 , for any family ( or web ) @xmath26 we can define a bounded linear functional @xmath256 on the algebra of continuous functions on @xmath40 by : @xmath257 by the riesz - markov theorem , such a bounded linear functional is just a finite regular borel measure on @xmath2 .
( henceforth when we write simply ` measure ' we shall always mean a finite regular borel measure . ) in short , a generalized measure on @xmath2 determines a collection of measures on the spaces @xmath40 for all families @xmath26 , and in fact such a collection satisfying certain conditions uniquely determines a generalized measure . in light of propositions [ web1 ] and [ web2 ] , however , it is natural to translate this into the language of webs .
[ thm : gen ] given a generalized measure @xmath22 on @xmath2 and setting @xmath258 for any web @xmath45 , the collection @xmath187 is : consistent : if the web @xmath45 depends on the web @xmath259 then @xmath260
. uniformly bounded : the linear functionals @xmath261 are uniformly bounded as @xmath45 ranges over all webs .
conversely , given any such consistent and uniformly bounded collection @xmath187 of measures on the spaces @xmath174 , there exists a unique generalized measure @xmath22 on @xmath2 for which @xmath262 for all webs @xmath45 . _ proof .
_ it is clear that given a generalized measure @xmath22 on @xmath2 , the measures @xmath263 are consistent , and are uniformly bounded by the norm of @xmath22 . for the converse , suppose @xmath187 is a collection of measures on the space @xmath174 satisfying ( a ) and ( b ) .
first we define a linear functional @xmath264 on @xmath253 as follows . for any @xmath265 , choose a family @xmath26 and let @xmath266 be a continuous function with @xmath267 . by proposition [ web1 ] ,
there is a web @xmath45 upon which @xmath26 depends , so defining @xmath268 , we have @xmath269 .
now define @xmath270 of course , we need to check that @xmath264 is well - defined and linear .
suppose that @xmath271 is also continuous and @xmath272 . by proposition [ web1 ] again choose @xmath259 upon which @xmath273 , and hence @xmath45 and @xmath274 , depend .
then by ( a ) @xmath275 also , since @xmath276 we have @xmath277 , but since @xmath278 is onto this implies @xmath279 thus we have @xmath280 so @xmath264 is well - defined .
the linearity of @xmath264 then follows from the linearity of each of the @xmath48 s . by ( b )
we can choose @xmath281 such that @xmath282 for all @xmath45 , and this implies that @xmath283 for all @xmath284 . since @xmath253 is dense in @xmath35 , @xmath264 extends uniquely to a bounded linear functional @xmath22 on @xmath35 . by construction , @xmath262 for all @xmath45 .
the uniqueness of @xmath22 with this property also follows from the fact that @xmath253 is dense in @xmath35 .
3em 2ex in fact , generalized measures on @xmath2 are the same thing as measures on the projective limit @xmath24 of the spaces @xmath40 , where the families @xmath26 are ordered by dependence . in these terms , proposition [ web1 ] says that webs are cofinal in the set of all families , and theorem [ thm : gen ] is seen as a special case of a very general result , namely that a measure on a projective limit of spaces can be constructed from a consistent and uniformly bounded collection of measures on any cofinal set of these spaces .
ashtekar and lewandowski have given a clear exposition of this approach in the analytic context @xcite , but here we chose to prove everything ` from scratch . ' elements of @xmath24 may be called _ generalized connections _ on @xmath3 .
abstractly , @xmath24 is simply the gelfand spectrum of the c*-algebra @xmath35 .
the space @xmath2 of smooth connections on @xmath3 naturally maps into @xmath24 in a one - to - one and continuous way , and the image is dense in @xmath24 .
thus generalized connections may be regarded as limits of smooth connections .
in this section we construct a generalized measure @xmath49 on @xmath2 which we call the ` uniform measure ' . theorem
[ thm : gen ] suggests that we do this by choosing for each web @xmath45 a measure @xmath285 in some canonical way . in the special case of a web consisting of a single tassel @xmath81 , fixing a trivialization over the endpoints
lets us think of @xmath286 as a measure on @xmath189 .
since @xmath189 is a compact lie group , an obvious choice is haar measure on @xmath189 . for more general webs it is natural to use a product of haar measures .
this in fact gives a generalized measure .
[ thm : uni ] there exists a unique generalized measure @xmath49 on @xmath2 such that @xmath286 is haar measure on @xmath189 for any tassel @xmath81 and any choice of trivialization of the endpoints , and @xmath287 for any web @xmath45 consisting of tassels @xmath288 .
_ we first must prove that @xmath286 , for a tassel @xmath81 based at a point @xmath44 , is independent of the choice of trivialization . a change in the trivialization
would effectively replace the holonomy @xmath78 of a given connection along @xmath17 by @xmath289 , where @xmath290 and @xmath291 are elements of @xmath0 expressing the change of trivialization at the point @xmath44 and the right endpoint of @xmath17 respectively .
thus @xmath189 gets sent to @xmath292 , where @xmath293 and @xmath294 , and @xmath295 if @xmath17 and @xmath296 have the same right endpoint .
now consider any point @xmath224 in @xmath83 .
the set of @xmath110 such that all points in @xmath83 with parameter value @xmath110 are regular is open and dense , so there are such @xmath297 and @xmath298 arbitrarily close to @xmath299 . for @xmath297 sufficiently close ,
every curve that goes through @xmath224 goes through exactly one of the regular points with parameter value @xmath110 , and none shares a regular point with a curve that does not go through @xmath224 .
thus the group generated by their types includes points in @xmath31 with a @xmath300 in the @xmath92th entry if @xmath17 goes through @xmath224 and @xmath301 if it does not .
likewise , taking @xmath298 and small enough , we can find in the group generated by the types elements in @xmath31 with a @xmath300 in the @xmath92th entry if @xmath17 goes through @xmath224 and does not end there , and a @xmath203 otherwise . putting these together
, we see that @xmath189 contains every element of @xmath31 with a @xmath300 in the @xmath92th entry if @xmath17 has @xmath224 as its endpoint and a @xmath203 otherwise . thus @xmath302 is an element of @xmath189 .
likewise @xmath303 .
but since haar measure on a lie group is invariant under left and right multiplication , it gets sent to itself under the map @xmath304 .
thus the assignment of measures to tassels , and hence to webs , is independent of the choice of trivialization , and therefore well - defined .
now , to check condition ( a ) of theorem [ thm : gen ] , first consider a tassel @xmath81 based on @xmath44 , and let @xmath305 be a web on which @xmath81 depends .
we will show that @xmath306 in four cases .
these are illustrated in figure [ fg : cases ] , where the curves of @xmath81 are represented in bold and the curves of @xmath45 are represented in medium weight . _
@xmath45 consists of a single tassel @xmath159 based at @xmath307 .
_ since each curve in @xmath159 has @xmath44 as a left endpoint , and each curve in @xmath81 contains @xmath44 only as its left endpoint , every curve in @xmath81 is a curve in @xmath159 . thus writing @xmath308 and @xmath309 in the standard way
, @xmath310 sends @xmath311 to @xmath312 for some integers @xmath313 .
in particular , @xmath314 is an onto homomorphism from @xmath315 to @xmath189 .
the image of haar measure under an onto homomorphism is haar measure again , so @xmath316 . _
@xmath45 consists of a single tassel @xmath159 , and @xmath317 .
_ let @xmath274 be the set of curves in @xmath159 which contain @xmath44 , and @xmath318 be the set of curves which do not .
then since every curve in @xmath81 contains @xmath44 only as its left endpoint , every curve in @xmath81 can be written either as @xmath319 for @xmath320 or as @xmath321 for @xmath320 , @xmath322 .
clearly @xmath323 is a probability measure on @xmath189 , so it suffices to show that @xmath323 is invariant under right multiplication by elements of @xmath189 .
equivalently , since @xmath189 is generated by the types in @xmath81 , we must show that @xmath323 is invariant under right multiplication by any element of @xmath128 , for @xmath128 a type of @xmath81 .
for this , choose a point @xmath324 with @xmath325 and with @xmath299 small enough that @xmath224 is not on any curve of @xmath318 .
we can identify @xmath128 with @xmath0 by the canonical isomorphism , and hence with @xmath326 , the type of @xmath224 in @xmath274 .
then we have , for @xmath327 and @xmath328 @xmath329 so @xmath330 _ each @xmath146 contains a curve containing @xmath44 that is a subcurve of a curve in @xmath81 .
_ if there is a @xmath252 with @xmath331 , then since in a web no tassel is based on a point of intersection with other tassels , there is only one @xmath148 in @xmath45 , and we are in situation ( i ) .
so assume @xmath332 for all @xmath252 .
suppose @xmath62 in @xmath81 is product of curves including one in @xmath148 and one in @xmath146 , the one in @xmath148 being the one which contains @xmath44 .
then @xmath149 and @xmath137 lie on @xmath62 .
since @xmath146 contains a subcurve of some curve @xmath333 in @xmath81 containing @xmath149 and @xmath44 , we have that @xmath62 intersects @xmath333 at a point @xmath332 .
thus they intersect infinitely many times , arbitrarily near @xmath44 , and thus the curves in @xmath146 and @xmath148 intersect infinitely many times . since this is impossible in a web ,
we conclude that each curve @xmath62 in @xmath81 depends on one @xmath146 .
further , a curve depending on @xmath148 can not intersect a curve depending on @xmath146 for @xmath334 ( except at @xmath44 ) , because then they would intersect infinitely many times , arbitrarily near @xmath44 , and so would curves in @xmath148 and @xmath146 .
thus the @xmath148 s separate @xmath81 into families of curves @xmath335 that intersect only at @xmath44 , with each @xmath336 depending on @xmath148 .
but then each @xmath336 is a tassel based at @xmath44 .
any type of @xmath81 is a type of some @xmath336 , and commutes with all types of all other @xmath337 .
thus @xmath338 , and @xmath339 ( of course , @xmath340 is not a web , because they all intersect at their bases ) . thus @xmath341 can be written as a product of maps @xmath342 , so it suffices to show that @xmath343 in order to conclude that @xmath344 .
but @xmath336 is a single tassel depending on the single tassel @xmath148 , so by ( i ) and ( ii ) we have @xmath345 . _
@xmath45 is arbitrary_. let @xmath346 be the set of all @xmath148 which contain a curve containing @xmath44 that is a subcurve of a curve in @xmath81 , and let @xmath347 be the set of all other @xmath148 .
let @xmath348 be @xmath81
restricted to @xmath349 , and @xmath350 be @xmath81 restricted to @xmath351 . since every curve in @xmath348 contains @xmath44 , every curve in @xmath350 does not , and every curve in @xmath81 contains @xmath44 exactly once , it follows that every curve in @xmath81 is either a curve in @xmath348 or a product of a curve in @xmath350 and a curve in @xmath348 .
now @xmath350 depends on @xmath347 , so by trivializing @xmath3 over a neighborhood containing @xmath81 the set @xmath352 may be thought of a subset of a product of copies of @xmath0 .
since this subset consists of products of types in @xmath81 , it is contained in @xmath189 , so @xmath353 may be viewed as a probability measure @xmath22 on @xmath189 . since in this interpretation @xmath354 , the measure @xmath355 is the convolution of @xmath22 and @xmath356 .
now @xmath348 is a tassel based on @xmath44 , and @xmath357 , because every type occurring in @xmath81 occurs arbitrarily close to @xmath44 , and hence in @xmath348 .
so by ( iii ) @xmath358 .
but it is well known that the convolution of a probability measure on a group with haar measure is again haar measure , so @xmath359 .
[ cols="^,^,^,^ " , ] to finish checking condition ( a ) of theorem [ thm : gen ] , we suppose that @xmath360 is a web depending on the web @xmath361 , and show that @xmath362 . to see this , note that any @xmath363 can be divided into equivalence classes @xmath364 of curves which are parallel at @xmath137 , and that curves from different equivalence classes do not intersect except at @xmath137 ( this is essentially the argument in point ( iv ) above ) .
thus every type of @xmath363 is a type of some @xmath365 , and commutes with all types of any other @xmath366 .
in particular @xmath367 and @xmath368 . by ( i - iv )
, it suffices to show that @xmath369 assigns an independent measure to each @xmath370 , and by the above it suffices to show that curves in different @xmath370 s do not depend under @xmath371 on curves in the same @xmath365 .
this is clear , because if they did then @xmath137 would be in the range of both of these @xmath370 s , but no neighborhood of it could be an interval because the curves from the two different tassels would be parallel at @xmath137 .
condition ( b ) of theorem [ thm : gen ] is immediate .
each @xmath285 is a probability measure , so as a linear functional it has norm @xmath203 .
3em 2ex we call this generalized measure @xmath49 the _ uniform generalized measure_. this generalized measure has a number of important properties .
first notice that the group @xmath372 of automorphisms of the bundle @xmath3 acts on the space @xmath2 , and thus acts as automorphisms of the c*-algebra @xmath35 via @xmath373 as a consequence it acts dually on the space @xmath374 of generalized measures on @xmath2 .
we shall show that @xmath49 is invariant under this action .
moreover , since a generalized measure @xmath22 on @xmath2 is equivalent to a measure on @xmath23 , it is natural to speak of @xmath22 being a _ probability measure _
if @xmath375 and for all @xmath376 , @xmath377 implies @xmath378 . borrowing some terminology from c*-algebra theory , we also say that a probability measure @xmath22 is _ faithful _
if @xmath379 and @xmath380 imply @xmath381 for all @xmath382 .
[ cr : nu ] the uniform generalized measure @xmath49 is a faithful probability measure , invariant under the action of @xmath372 . _
_ to see that @xmath49 is a faithful probability measure it suffices to check that @xmath285 is a faithful probability measure for each web @xmath45
. for this , in turn , it suffices to check it for a tassel , and haar measure is clearly a faithful probability measure . to see that @xmath49 is invariant , note that every step in its construction was manifestly invariant except the choice of trivialization , and we showed that @xmath49 was independent of that .
since @xmath49 is a faithful probability measure , we may define @xmath50 as the hilbert space completion of the space @xmath35 with the inner product @xmath383 .
equivalently , we could set @xmath384 . since @xmath49 is invariant under @xmath372
, there is a unitary representation of @xmath372 on @xmath50 , and thus a unitary representation of the subgroup @xmath385 consisting of gauge transformations .
we define @xmath6 to be the closed subspace consisting of vectors in @xmath50 invariant under the action of @xmath7 . in this section
we describe an explicit set of functions on @xmath2 spanning the hilbert space @xmath6 ; by analogy with the analytic case @xcite we call these ` spin networks ' . given any family @xmath26 ,
let @xmath386 be the hilbert space of square - integrable functions on @xmath40 with respect to the measure @xmath387 .
of course , the map @xmath388 from @xmath389 to @xmath253 extends to an isometry of @xmath386 into @xmath50 , and the union of the images of these isometries over all families @xmath26 is dense in @xmath50 . in fact
if @xmath26 depends on @xmath63 then the embedding of @xmath386 in @xmath50 factors through that of @xmath390 , so the union of the images of @xmath391 over all webs @xmath45 is also dense . in keeping with the philosophy of this paper
, one can try to understand @xmath50 by understanding @xmath391 for all webs @xmath45 .
if @xmath305 is a web , then @xmath391 is fairly simple to describe .
fixing a trivialization of @xmath3 over the endpoints of the curves , @xmath392 .
note however that this isomorphism changes when we change the trivialization .
understanding how it changes is a part of what we need to describe the gauge invariant subspace . for each family @xmath26 ,
the group @xmath7 acts on @xmath40 .
the quotient of @xmath7 by the subgroup which acts trivially on @xmath40 is a finite - dimensional lie group @xmath393 , which is actually the product over all endpoints @xmath224 of curves in @xmath26 of the groups @xmath394 of gauge transformations of the fibers @xmath395 .
fixing a trivialization of @xmath395 gives an isomorphism between @xmath394 and @xmath0 , so we can think of @xmath393 as a product of copies of @xmath0 .
the action of @xmath7 on @xmath40 gives a unitary representation on @xmath386 , and when @xmath26 depends on @xmath63 the natural embedding @xmath396 is an intertwining operator . let @xmath397 be the subspace of @xmath393-invariant vectors in @xmath386 .
as before , @xmath397 embeds into @xmath398 if @xmath26 depends on @xmath63 , they both embed into @xmath6 in a consistent fashion , and the image of all such embeddings is dense in @xmath6 as @xmath26 ranges over all families , or all webs .
we will construct an orthonormal basis of @xmath399 for each web @xmath45 .
the resulting set of vectors will thus give a set spanning @xmath6 . to do this , we need an understanding of the action of @xmath400 on @xmath391 .
we begin by considering the action of @xmath401 on @xmath191 when @xmath81 is a tassel . if @xmath81 is a tassel based at @xmath44 , then @xmath402 is the group @xmath0 , with action inherited from the left action of @xmath189 on @xmath403 by the map @xmath404 .
if @xmath224 is any right endpoint of curves in @xmath81 , then @xmath394 is the group @xmath0 , with action inherited from the right action of @xmath189 on @xmath403 by the map @xmath405 , where @xmath211 equals @xmath300 in every entry corresponding to a curve with endpoint @xmath224 , and equals @xmath203 otherwise ( see the proof of theorem [ thm : uni ] ) .
more precisely , the peter - weyl theorem states that @xmath403 as a left and right @xmath189-module decomposes as @xmath406 where @xmath407 is the set of all isomorphism classes of irreducible unitary representations of @xmath189 , @xmath408 is an element of the isomorphism class @xmath409 as a left representation , and @xmath410 is the dual space of @xmath408 , as a right representation .
if @xmath44 is the base of @xmath81 , and @xmath411 is the subgroup of @xmath189 consisting of all @xmath412 , then the action of @xmath402 on @xmath413 is the left action of @xmath414 . likewise if @xmath415 is the subgroup of @xmath189 consisting of all @xmath416 with @xmath79 if the @xmath92th curve in @xmath81 has @xmath224 as its right endpoint , and @xmath80 otherwise , then the action of @xmath394 is the right action of @xmath417 on @xmath418 . if @xmath45 is a web , we can write @xmath419 where @xmath420 is shorthand for @xmath421
. the action of the gauge group will be the same , except if a point @xmath224 is the right endpoint of more than one tassel , in which case it is the right endpoint of two tassels , say @xmath148 and @xmath146 . in this case
@xmath422 acts on @xmath423 by the tensor product of the actions on each individually .
invariant vectors under this action come from invariant elements of @xmath424 for some choice of @xmath425 and @xmath426 . since the actions of different groups @xmath394 commute
, we can decompose each @xmath427 into an orthogonal direct sum of tensor products , over every @xmath224 an endpoint for @xmath148 , of irreducible unitary right representations of @xmath394 . to construct actual @xmath400-invariant elements of @xmath391 , recall that the peter - weyl isomorphism is given by sending the element @xmath428 to the function @xmath429 for @xmath430 , @xmath431 being the usual pairing of a vector space with its dual , but multiplied by the square root of @xmath432 to make the isomorphism unitary .
so choose a representation @xmath433 for @xmath434 , choose an @xmath435-invariant vector @xmath436 in @xmath437 , choose a term in the direct sum decomposition of each @xmath427 , such that the representation assigned to each endpoint @xmath224 which is only an endpoint for @xmath148 is assigned the trivial representation and the representations assigned to a @xmath224 which is an endpoint for @xmath146 and @xmath148 respectively are dual representations .
also choose a vector @xmath438 in the trivial representation chosen for each @xmath224 bounding one tassel , and an invariant element @xmath438 of the representation @xmath439 chosen for each @xmath224 bounding two tassels .
notice that @xmath440 for @xmath441 is an element of @xmath442 , and that @xmath443 , the product being over all endpoints @xmath224 , is after reordering appropriately an element of @xmath444 , and thus they can be paired ( by the rescaled pairing ) to get a number , which we call @xmath445 .
we call such a function a _ spin network_. choosing @xmath436 from an orthonormal basis of the subspace of @xmath437 of @xmath448-invariant vectors , and choosing fixed unit vectors @xmath449 for each term in the direct sum decomposition of @xmath427 , we get an orthonormal basis for @xmath399 . they are clearly orthonormal , and they certainly span the space of spin networks .
but by the peter - weyl theorem , every @xmath400-invariant element of @xmath391 is spanned by those of the form @xmath452 , with @xmath436 and @xmath453 invariant elements of @xmath437 and @xmath427 for some @xmath425 .
since such @xmath454 are certainly spanned by all the tensor products @xmath455 used to construct spin - networks , it is clear that the spin networks span @xmath399 .
3em 2ex the authors would like to thank greg kuperberg and richard palais for information on real - analytic structures , lisa raphals for the tassel terminology , and david vogan for sharing his knowledge of lie groups and haar measure .
j. b. would also like to thank jerzy lewandowski for inviting him to speak on this subject at the workshop on canonical quantum gravity in warsaw .
a. ashtekar and j. lewandowski , projective techniques and functional integration for gauge theories , to appear in _ jour .
_ , pennsylvania state university preprint available as gr - qc/9411046 .
a. ashtekar , j. lewandowski , d. marolf , j. mourao and t. thiemann , quantization of diffeomorphism invariant theories of connections with local degrees of freedom , to appear in _ jour.math .
_ , pennsylvania state university preprint available as gr - qc/9504018 .
j. baez , diffeomorphism - invariant generalized measures on the space of connections modulo gauge transformations , in _ proceedings of the conference on quantum topology , _ ed .
d. yetter , world scientific , singapore , 1994 , pp . 21 - 43 .
j. baez , knots and quantum gravity : progress and prospects , to appear in the proceedings of the seventh marcel grossman meeting on general relativity , university of california at riverside preprint available as gr - qc/9410018 . | let @xmath0 be a compact connected lie group and @xmath1 a smooth principal @xmath0-bundle . let a ` cylinder function ' on the space @xmath2 of smooth connections on @xmath3 be a continuous function of the holonomies of @xmath4 along finitely many piecewise smoothly immersed curves in @xmath5 , and let a generalized measure on @xmath2 be a bounded linear functional on cylinder functions .
we construct a generalized measure on the space of connections that extends the uniform measure of ashtekar , lewandowski and baez to the smooth case , and prove it is invariant under all automorphisms of @xmath3 , not necessarily the identity on the base space @xmath5 . using `
spin networks ' we construct explicit functions spanning the corresponding hilbert space @xmath6 , where @xmath7 is the group of gauge transformations .
a g r c tr range domain lie fun aut u [ 1]|#1 | [ 1 ] # 1 [ 1]#1 [ 2]#1 # 2 [ 1][#1 ] y [ 5 ] 8.5in.25 in .25 in .25 in .5 in * functional integration on spaces of connections + * _ stephen sawin + _ |
according to the general relativity principle , the most general classical theory of gravitation in @xmath9 dimensions is lovelock gravity .
it is a higher dimensional generalization of einstein gravity which can extend gravity to other consistent theories such as string theory .
an interesting case which is second order lovelock gravity is gauss - bonnet ( gb ) theory with @xmath12 dimensions . in string theory context , when assuming that the tension of a string is large as compared to the energy scale of other variables , gb term is the first curvature correction to gravitation @xcite .
exactly the same as einstein gravity , the einstein - gb ( egb ) field equation contains up to second order of metric tensor and is free of ghost .
this theory has been extended to ads and ds spacetimes @xcite .
taub - nut / bolt - ads ( tn / bads ) is a class of interesting spacetimes with negative cosmological constant which were investigated in egb gravity @xcite.the original four - dimensional solution of taub - nut / bolt spacetime @xcite is only locally asymptotically flat .
the spacetime develops a boundary at infinity by twisting @xmath13 bundle over @xmath14 , instead of simply being @xmath15 .
the boundary properties of tnads spacetimes has been discussed in einstein gravity @xcite . in general , the killing vector which corresponds to the coordinate that parameterizes the fibre @xmath13 can have a zero - dimensional fixed point set ( called a nut solution ) or a two - dimensional fixed point set ( referred to as a ` bolt ' solution ) .
generalizations to higher dimensions and higher derivative gravity have been done @xcite .
these solutions of einstein gravity play a central role in the construction of m - theory configurations .
indeed , the 4-dimensional tnads solution provided the first test for the ads / cft correspondence in spacetimes that are only locally asymptotically ads @xcite , and the taub - nut metric is central to the supergravity realization of the @xmath16-brane of type iia string theory @xcite .
it is therefore natural to suppose that the generalization of these solutions to the case of egb gravity , which is the low energy limit of supergravity , might provide us with a window into some interesting new corners of m - theory moduli space . the ads / cft correspondence , is a famous conjecture that relates a ( @xmath17)-dimensional conformal field theory to @xmath9-dimensional gravity theory with a negative cosmological constant @xcite .
in light of this conjecture , similar to quantum field theory , there are some counterterms which remove infinities in the gravity action .
the counterterm contains some terms which are functions of the curvature invariants of the induced metric on the boundary . for egb gravity ,
similar to einstein gravity @xcite , in computing the action and total mass , one encounters infrared divergences associated with the infinite volume of the spacetime .
recently , the counterterm for egb gravity was obtained for @xmath18 dimensions and applied to static and rotating black objects @xcite .
also a counterterm for charged black holes ( bhs ) in gb gravity was proposed @xcite .
however , an alternative regularization prescription with any odd - even dimensions of ads asymptotic spacetimes for any lovelock theory has been proposed @xcite .
this approach is known as kounterterms and uses boundary terms with explicit dependence on the extrinsic curvature @xmath19 .
the thermodynamics of black holes in lovelock gravity attracted a lot of attention in the last decade @xcite . in dealing with this subject
, a regularized action can be obtained by using the counterterm renormalization method . for taub - nut /
bolt black holes the area law of entropy is broken and one may obtain the entropy by using gibbs - duhem relation @xcite .
thermodynamics of taub - nut / bolt black holes is investigated in einstein and einstein - maxwell gravity in @xcite . by extending the dimension of spacetimes higher than seven ,
one encounters so many terms in the counterterm that the computation of thermodynamic quantities would be very difficult .
also a counterterm for @xmath20 in egb gravity has not been worked out . in this paper
we give a new method to obtain the finite action based on the gibbs - duhem relation and gibbs free energy for nut solutions . according to this method
, all thermodynamic quantities in arbitrary even dimensions can be obtained .
the validity of the first law of thermodynamics of black holes has been tested for taub - nut spacetimes in einstein gravity @xcite .
we also check the validity of this law in egb gravity .
the outline of our paper is as follows .
we give a brief review of the general formalism of egb gravity , the counterterm method and taub - nut / bolt spacetime in section [ fiel ] . in section [ 6d ] , two possible taub - nut / bolt solutions of egb gravity in six dimensions
are reviewed .
then , in sections [ thermnut ] and [ thermbolt ] , all thermodynamic quantities for nut and bolt solutions are obtained and thermodynamical stability is investigated .
finaly , in section [ thermnutk - d ] , a generalization of the thermodynamics to @xmath21-dimensional taub - nut - ads spacetimes in egb gravity is made .
we finish the paper with some concluding remarks .
the gravitational action of egb gravity in @xmath9 dimension is @xmath22 \label{ig}\ ] ] where @xmath23 is the cosmological constant ; @xmath24 , @xmath25 , and @xmath26 are the ricci scalar , the riemann and the ricci tensors of the metric @xmath27 ; and @xmath0 is the gb coefficient with dimension of @xmath28 .
@xmath29 is the gb term , @xmath30 in order to have a well - defined variational principle , we must consider two surface terms , the gibbons - hawking term and its counterpart for gb gravity , which are @xmath31 @xmath32 where @xmath33 is the induced metric on the boundary , @xmath34 is the trace of the extrinsic curvature of the boundary , @xmath35 is the einstein tensor on the induced metric , and @xmath36 is the trace of the tensor , @xmath37 we restrict ourselves to the case of @xmath38 , which is consistent with heterotic string theory @xcite .
since this term is a topological invariant in four dimensions , we therefore apply this term to higher dimensions which consist of only second order derivatives of the metric that produces second order field equations . varying the action with respect to the metric tensor @xmath39 , the vacuum field equation is @xmath40where@xmath41 we want to consider asymptotic ads spacetime which has negative scalar curvature at infinity .
this implies the following asymptotic expression for the riemann tensor : @xmath42where @xmath43 is the effective radius of ads spacetime in egb gravity @xcite .
this parameter is @xmath44where @xmath45with @xmath46 @xmath47 it is seen from eq . ( [ ueff ] ) that the parameter @xmath0 must have an upper bound as @xmath48 in the computation of the total action @xmath49 and conserved charges of egb solutions , one encounters infrared divergences , associated with the infinite volume of the spacetime manifold .
this quantities are computed by using a counterterm method which was proposed by balasubramanian and kraus @xcite for ads - einstein gravity .
the counterterm that regularizes the action for @xmath50 solutions , for egb gravity @xcite is @xmath51% { \large \ } } , \nonumber\end{aligned}\]]where @xmath24 , @xmath52 and @xmath53 are the curvature , the ricci tensor and the gb term ( [ lgb ] ) associated with the induced metric @xmath54 .
also , @xmath55 is the step - function with @xmath56 provided that @xmath57 , and zero otherwise .
it can be seen that as @xmath58 ( @xmath59 ) , this action reduces to the familiar counterterm expression in the einstein gravity @xcite . having the total finite action
, one can construct a divergence - free stress - energy tensor by varying the total action with respect to the boundary metric @xmath60 , ( brown and york s definition of energy - momentum tensor @xcite ) .
the total finite stress - energy tensor for @xmath50 @xcite is @xmath61}-\frac{u-1}{2(d-3)(d-4)(d-5)}h_{ab}% { \large \}}+\dots , \nonumber\end{aligned}\]]where @xcite @xmath62and @xmath63 is given by ( [ hgb ] ) in terms of the boundary metric @xmath64 . from the finite stress - energy tensor
, one can compute conserved charges . to do this job
, we must choose a spacelike surface @xmath65 in the boundary @xmath66 with metric @xmath67 and write the boundary metric in the arnowitt - deser - misner form , @xmath68where @xmath69 and @xmath70 are the lapse function and shift vector , respectively .
the coordinates @xmath71 , @xmath72 are the angular variables , parameterizing the hypersurface of constant @xmath73 .
the conserved charge associated with a killing vector @xmath74 is @xmath75where @xmath76 is the determinant of the metric @xmath67 and @xmath77 is the normal to quasilocal boundary hypersurface @xmath78 for example the conserved mass @xmath79 , is the charge associated with the timelike killing vector @xmath80 .
the taub - nut / bolt - ads metric is one of the solutions of egb gravity which was investigated in ( @xmath21 ) dimensions @xcite .
this metric is constructed on a base space endowed with an einstein - k@xmath81hler metric @xmath82 .
the euclidean section of the @xmath83-dimensional taub - nut / bolt spacetime can be written as : @xmath84where @xmath85 is the coordinate on the fibre @xmath13 and @xmath86 is the k@xmath81hler form of the base space @xmath87 , which is proportional to some covariantly constant two - form .
here @xmath8 is the nut charge and @xmath88 is a function of @xmath73 which is obtained by solving the egb field equation ( [ geq ] ) .
the solution will describe a ` nut ' if the fixed point set of the u(1 ) isometry @xmath89 ( i.e. the points where @xmath90 ) is less than @xmath91-dimension and a ` bolt ' if the fixed point set is @xmath91-dimensions .
in this section we give a revision of the six - dimensional taub - nut / bolt solutions([tn ] ) of gb gravity will be done .
the function @xmath88 for all nonextremal choices of the base space @xmath92 can be written @xcite in the form of @xmath93where @xmath94and the function @xmath95 depends on the choice of the curved base space @xmath96 .
for the case @xmath97 , the k@xmath81hler form @xmath86 and the @xmath3 metric are @xmath98the function @xmath95 in this case is @xmath99the other possibility is @xmath10 , where @xmath14 is the 2-sphere with @xmath100 , and the one - form @xmath86 is @xmath101 in this case @xmath95 is @xmath102for asymptotically ads spacetimes [ @xmath103 provided that @xmath104 , the function @xmath88 is not real for all values of @xmath73 in the range @xmath105 and @xmath106 .
the ricci scalar and @xmath29 in the bulk ( [ ig ] ) are @xmath107 , \label{rscal}\]]@xmath108 + 3rf(r)^{2}-2rf(r)+\gamma r\right\ } , \label{lgbcps}\]]where @xmath109 for @xmath97 and @xmath110 for @xmath10 respectively . a. nut solutions : in brief , the conditions for having a nut solution , which have been described completely in some previous works @xcite , are as follows : \(i ) @xmath111 ; ( ii ) @xmath112 ; ( iii ) @xmath88 should have no positive roots at @xmath113 .
the first condition comes from the fact that all extra dimensions should collapse to zero at the fixed point set of @xmath114 .
the second condition allows one to avoid a conical singularity with a smoothly closed fiber at @xmath2 . using the first condition ,
the gb gravity may admit nut solutions with @xmath3 and @xmath5 base spaces when the mass parameter @xmath115 is fixed to be @xmath116 @xmath117by inserting @xmath118 from eqs .
( [ mncp ] ) and ( [ mns ] ) in ( [ f6 ] ) , the function @xmath119 can be obtained .
but as it has been mentioned in ref .
@xcite , the metric with @xmath10 has a curvature singularity at @xmath2 .
although the second condition to have a nut solution is satisfied numerically , by having a curvature singularity at @xmath2 , this condition is basically violated .
therefore , the metric does not admit a nut solution in this case . in fig .
[ fig1 ] , the function @xmath119 is plotted versus @xmath73 for @xmath120 and @xmath121 for both base spaces . as it is shown in this figure , at the small vicinity of @xmath2 , @xmath119 is complex with a continues real part . following section [ thermnut ]
, we will show that if one does not consider this fact and investigate the thermodynamics of nut solutions for spacetimes with both base spaces , no obvious difference will be seen .
in the limit of @xmath58 , @xmath119 for both base factors goes to @xmath122 in einstein gravity @xcite .
the allowed range of @xmath88 in the @xmath3 case is more extended than @xmath5 case .
also if @xmath6 goes to @xmath123 , or @xmath124 , @xmath119 would be complex in some range of @xmath73 so that in both base spaces , we find that @xmath119 is not real at @xmath125 for all values of r in the range of @xmath126 and @xmath38 .
this fact is true for nut solutions of arbitrary even dimensions .
the boundary of this kind of spacetime will be an asymptotic surface at some large radius @xmath73 .
the boundary metric @xmath33 diverges at infinity . by rescaling @xmath60 upon a conformal factor @xmath127 ( see eq.([leff ] ) ) , the dual boundary metric @xmath128 converges and the general line element of the dual field theory for even - dimensional tnads spacetimes on the boundary can be obtained as @xmath129 in fact , @xmath130 . as it has been shown in ( [ boundmetric ] ) ,
the gauss - bonnet coefficient @xmath0 enters the metric on the boundary due to @xmath43 .
b. bolt solutions : the conditions for having bolt solutions are ( i ) @xmath131 and ( ii)@xmath132 with @xmath133 . these conditions give the following relation for @xmath134 @xmath135 - 40\alpha ( r_{b}^{2}+n^{2})}{15r_{b } } \label{mbcp}\]]and an equation for @xmath136 with the base space @xmath3 @xmath137 which at least has one real solution . this real solution for @xmath138 yields @xmath133 .
for the case of @xmath10 , the mass parameter is @xmath139 - 60\alpha ( r_{b}^{2}+n^{2})}{15r_{b}},\]]and @xmath136 can be found by solving the following equation : @xmath140this equation will have two real roots if @xmath141 . at @xmath142
there will be only one @xmath133 . therefore in order to have bolt solution in this case , there is a maximum value for @xmath8 @xcite .
it is found that the allowed range of @xmath136 for this case is more extended than in the @xmath3 case .
a. @xmath144 case : from relations ( [ rscal ] ) and ( [ lgbcps ] ) , we can simply compute the bulk action ( [ ig ] ) . by computing the other actions for the boundary and counterterm ( [ ibe ] ) , ( [ ibgb ] ) , and ( [ lagrangianct ] )
, the finite action @xmath145 can be obtained as @xmath146 in eq .
( [ finaction ] ) , @xmath147 is the inverse of the hawking temperature @xmath148 which is found by demanding regularity of the euclidean manifold ( [ tn ] ) . the total mass can be calculated as @xmath149the total angular momentum is zero , which is the same as einstein gravity . the gibbs free energy and entropy
may be computed as @xmath150@xmath151also the entropy can be obtained , as the einstein case @xcite , by using gibbs - duhem relation , which is @xmath152where @xmath153 is chemical potential and @xmath154 is the conserved charge ( such as mass , angular momentum , ... ) . the smarr type formula for mass versus extensive quantity @xmath155 may be written as@xmath156where @xmath157 is a function of @xmath155 , which may be obtained from the following equation : @xmath158the hawking temperature @xmath148 may be calculated as @xmath159now the specific heat can be obtained as@xmath160 note that the entropy become negative for @xmath161 and specific heat is negative for @xmath162 .
thus , for thermally stable solution for @xmath97 case , the value of @xmath6 parameter must be in the range of @xmath163 provided that @xmath164 , @xmath165 and @xmath103 . in the limit of @xmath166 ( einstein gravity ) , this spacetime is completely unstable @xcite .
b. @xmath167 case : as in the previous case , the finite action , gibbs free energy , entropy , and specific heat may be obtained as @xmath168@xmath169@xmath170@xmath171 by comparing all quantities in the @xmath97 and the @xmath5 case , we can see that by substituting @xmath172 in all equations in the previous case , all thermodynamic quantities can be obtained in this base space . by expanding computations in @xmath21 dimensions
, it can be seen that @xmath173according to this fact , there is no difference between these two base spaces and if we do not consider the metric with @xmath1 , as a nut solution of egb gravity , nothing would be missed .
it is worthwhile to mention that , for all thermodynamic quantities which are calculated in these cases , the first law of thermodynamics , @xmath174 , is satisfied .
a. @xmath167 case : the finite action can be calculated as@xmath175}}\,{\large \{}(15\,\lambda \,{n}% ^{8}+30\,{n}^{6 } \nonumber \\ & & + 30\,{n}^{4}{r_{b}}^{2}-70\,{n}^{2}{r_{b}}^{4}-10\,{n}^{4}\lambda \,{r_{b}}% ^{4}-8\,{n}^{2}\lambda \,{r_{b}}^{6}+10\,{r_{b}}^{6 } \nonumber \\ & & + 3\,{r_{b}}^{8}\lambda ) + ( 120\,{n}^{4}{r_{b}}^{2}\lambda -300\,{n}% ^{4}+100\,{r_{b}}^{4}-120\,{n}^{2}{r_{b}}^{2 } \nonumber \\ & & -200\,{n}^{2}{r_{b}}^{4}\lambda -120\,\lambda \,{n}^{6}+72\,\lambda \,{% r_{b}}^{6})\alpha + 480(\,{n}^{2}+\,{r_{b}}^{2}){\alpha } ^{2}{\large \}. } \label{ibs}\end{aligned}\ ] ] in the limit of @xmath58 the action ( [ ibs ] ) goes to the corresponding action in einstein gravity @xcite . the entropy and specific heat may be written as @xmath176 @xmath177[\lambda ( r_{b}^{2}-n^{2})^{2}-2(r_{b}^{2}-n^{2})-4\alpha ] } \nonumber \\ & & { \large \{}-192\,{\alpha } ^{3}+\left ( 576\,{r_{b}}^{2}\lambda \,{n}% ^{2}-32\,\lambda \,{r_{b}}^{4}+336\,{n}^{2}+96\,\lambda \,{n}^{4}+144\,{r_{b}% } ^{2}\right ) { \alpha } ^{2 } \nonumber \\ & & + ( 20\,{r_{b}}^{8}{\lambda } ^{2}-12\,{\lambda } ^{2}{n}^{8}+376\,{n}^{2}{% r_{b}}^{4}\lambda + 48\,{n}^{6}{r_{b}}^{2}{\lambda } ^{2}-120\,\lambda \,{n}% ^{6 } \nonumber \\ & & -144\,{n}^{2}{r_{b}}^{6}{\lambda } ^{2}+192\,{r_{b}}^{4}+96\,{r_{b}}^{2}{n}% ^{2}-72\,{n}^{4}{r_{b}}^{2}\lambda -120\,{r_{b}}^{6}\lambda -192\,{n}^{4 } \nonumber \\ & & + 88\,{n}^{4}{r_{b}}^{4}{\lambda } ^{2})\alpha -88\,{n}^{4}\lambda \,{r_{b}}% ^{4}+104\,{n}^{2}\lambda \,{r_{b}}^{6}+36\,{n}^{6}+36\,{r_{b}}^{6}+36\,{n}% ^{8}\lambda \nonumber \\ & & + 74\,{r_{b}}^{6}{\lambda } ^{2}{n}^{4}-30\,{\lambda } ^{2}{n}^{6}{r_{b}}% ^{4}-15\,{\lambda } ^{2}{n}^{8}{r_{b}}^{2}-43\,{r_{b}}^{8}{\lambda } ^{2}{n}% ^{2}-36\,{r_{b}}^{2}{n}^{4 } \nonumber \\ & & -36\,{r_{b}}^{4}{n}^{2}-28\,{r_{b}}^{8}\lambda + 5\,{r_{b}}^{10}{\lambda } % ^{2}+9\,{\lambda } ^{2}{n}^{10}-24\,{r_{b}}^{2}\lambda \,{n}^{6}{\large \ } } \label{cbs}\end{aligned}\ ] ] in fig .
[ fig2 ] the entropy and specific heat are plotted for @xmath178 and @xmath179 as a function of @xmath8 in the allowed range of @xmath180 , for the outer horizon of the bolt solution .
as it can be see from this figure , the bolt solution in this case is completely stable .
b. @xmath144 case : the finite action , entropy , and heat capacity can be obtained as @xmath181}{\large \{}% \left ( -40\,{r_{b}}^{2}-40\,{n}^{2}\right ) \alpha \nonumber \\ & & -40\,{n}^{3}r_{b}-32\,{n}^{5}r_{b}\lambda + 15\,\lambda \,{n}^{6}+45\,{n}% ^{4}{r_{b}}^{2}\lambda + 30\,{n}^{4 } \nonumber \\ & & -10\,{r_{b}}^{4}-15\,{n}^{2}{r_{b}}^{4}\lambda + 60\,{r_{b}}^{2}{n}^{2}+3\,{% r_{b}}^{6}\lambda { \large \ } } \label{ibcp}\end{aligned}\ ] ] @xmath182}{\large \{}% \left ( -40\,{r_{b}}^{2}-40\,{n}^{2}\right ) \alpha \nonumber \\ & & + 15\,\lambda \,{n}^{6}+45\,{n}^{4}{r_{b}}^{2}\lambda + 30\,{n}^{4}-10\,{% r_{b}}^{4}-15\,{n}^{2}{r_{b}}^{4}\lambda \nonumber \\ & & + 60\,{r_{b}}^{2}{n}^{2}+32\,{n}^{5}r_{b}\lambda + 3\,{r_{b}}^{6}\lambda + 40\,{n}^{3}r_{b}{\large \ } } \label{sbcp}\end{aligned}\ ] ] @xmath183[3\lambda ( r_{b}^{2}-n^{2})^{2}-6(r_{b}^{2}-n^{2})-8\alpha ] } \nonumber \\ & & { \large \{}-180\,{r_{b}}^{2}\lambda \,{n}^{6}+450\,{n}^{3}{r_{b}}% ^{5}\lambda -630\,{n}^{5}{r_{b}}^{3}\lambda + 180\,{r_{b}}^{5}n-300\,{n}% ^{4}\lambda \,{r_{b}}^{4 } \nonumber \\ & & + 420\,{n}^{2}\lambda \,{r_{b}}^{6}+270\,{n}^{7}r_{b}\lambda -90\,{\lambda } % ^{2}{n}^{8}{r_{b}}^{2}-120\,{r_{b}}^{8}\lambda -360\,{n}^{3}{r_{b}}^{3}+ \nonumber \\ & & + 180\,{n}^{6}+180\,{r_{b}}^{6}+180\,{n}^{8}\lambda + 45\,{\lambda } ^{2}{n}% ^{10}+312\,{n}^{5}{r_{b}}^{5}{\lambda } ^{2}+180\,{n}^{5}r_{b } \nonumber \\ & & -264\,{n}^{7}{r_{b}}^{3}{\lambda } ^{2}-90\,{\lambda } ^{2}{n}^{6}{r_{b}}% ^{4}-171\,{r_{b}}^{8}{\lambda } ^{2}{n}^{2}+18\,{r_{b}}^{10}{\lambda } % ^{2}+72\,{n}^{9}r_{b}{\lambda } ^{2 } \nonumber \\ & & + 288\,{r_{b}}^{6}{\lambda } ^{2}{n}^{4}-180\,{r_{b}}^{2}{n}^{4}-180\,{r_{b}}% ^{4}{n}^{2}-90\,{r_{b}}^{7}n\lambda -120\,{r_{b}}^{7}{n}^{3}{\lambda } ^{2 } \nonumber \\ & & + 10\left ( 128\,{n}^{2}+48\lambda \,{r_{b}}^{4}+96nr_{b}+128{n}% ^{3}r_{b}\lambda + 96{r_{b}}^{2}\lambda \,{n}^{2}+48\lambda \,{n}^{4}\right ) { % \alpha } ^{2 } \nonumber \\ & & + ( -90\,{\lambda } ^{2}{n}^{8}+480\,{n}^{4}{r_{b}}^{2}\lambda -640\,{n}% ^{3}r_{b}-840\,{n}^{4}+54\,{r_{b}}^{8}{\lambda } ^{2}-480\,{n}^{3}{r_{b}}^{5}{% \lambda } ^{2 } \nonumber \\ & & -360\,n{r_{b}}^{5}\lambda + 1520\,{r_{b}}^{3}{n}^{3}\lambda + 960\,{r_{b}}% ^{3}n+720\,{r_{b}}^{2}{n}^{2}+360\,{n}^{6}{r_{b}}^{2}{\lambda } ^{2 } \nonumber \\ & & -600\,\lambda \,{n}^{6}-504\,{n}^{2}{r_{b}}^{6}{\lambda } ^{2}+180\,{n}^{4}{% r_{b}}^{4}{\lambda } ^{2}+840\,{n}^{2}{r_{b}}^{4}\lambda -96\,{n}^{7}r_{b}{% \lambda } ^{2 } \nonumber \\ & & -336\,{r_{b}}^{6}\lambda + 576\,{n}^{5}{r_{b}}^{3}{\lambda } ^{2}+760\,{r_{b}% } ^{4}-904\,{n}^{5}r_{b}\lambda ) \alpha -640\,{\alpha } ^{3}{\large \}. } \label{cbcp}\end{aligned}\ ] ] fig .
[ fig3 ] shows the entropy and specific heat for @xmath184 and @xmath179 as a function of @xmath8 in the allowed range .
it is shown that the bolt solution in this case is completely unstable .
in this section a new procedure to achieve thermodynamic quantities by using the gibbs - duhem relation and gibbs free energy will be introduced .
the generalized form of even - dimensional tnads spacetimes in egb gravity has been constructed @xcite .
the mass parameter in @xmath185 dimensions for @xmath186 can be calculated as @xmath187 .\label{mnkcp}\end{aligned}\]]as it has been mentioned in @xmath188 nut solutions , all quantities go to the einstein gravity case in the limit of @xmath58 .
also the gb parameter @xmath0 , has not been appear in the calculation of the total mass and inverse hawking temperature @xmath189 .
the similar case has been shown in ref @xcite .
therefore we may consider the total mass and @xmath190 , similar to d - dimensional nut solution in einstein gravity @xcite which have been obtained as @xmath191@xmath192what about finite action ?
up to now , the main method for computing the finite action has been the counterterm method , which has been used in sec .
[ thermnut ] for six dimensions . in order to extend this method to higher dimensions
, one needs more terms than ( [ lagrangianct ] ) .
however recently , a new counterterm action has been achieved for @xmath193 in egb gravity which has so many terms @xcite . according to this method ,
the existence of the counterterm action is necessary for computing the thermodynamic quantities .
now we give a different method based on the gibbs - duhem relation ( [ gibbsduhem ] ) and gibbs free energy for nut solutions that can calculate regularized action independent of the counterterm method . according to this method , first , first write the finite action in @xmath21 -dimensional einstein gravity @xcite with an additional term as a function of @xmath0
. this term must be considered in such a way that in the limit of @xmath58 , it goes to zero for consistency to einstein gravity .
therefore the finite action in egb gravity can be written as @xmath194 .
\label{ink}\end{aligned}\]]after that , calculate the gibbs free energy as a function of @xmath148 and then calculate the entropy @xmath195 similar to eq .
( [ sncp ] ) .
the other way of computing the entropy , which is independent of the previous method , is to use the gibbs - duhem relation ( [ gibbsduhem ] ) . by calculating the entropy in the latter way ( @xmath196 ) , the first order differential equation @xmath197 gives us @xmath198 as @xmath199the results in @xmath188 may be considered as an initial condition .
the function ( [ fk ] ) is independent of @xmath8 and is the same as the last term in ( [ mnkcp ] ) .
now the finite action may be calculated in egb gravity by substituting ( [ fk ] ) in the action ( [ ink ] ) . for @xmath200 ( @xmath188 ) , the action ( [ ink ] ) is equal to ( [ incp ] ) in section [ thermnut ] . in the appendix ,
computations for @xmath201 are done using the counterterm method and we see that there is a good coincidence with our prescription .
the gibbs free energy , entropy , and heat capacity may be obtained as @xmath202 } , \label{gnk}\end{aligned}\]]@xmath203,\end{aligned}\]]@xmath204 .\end{aligned}\ ] ] the spacetime for @xmath186 is stable in the following range : @xmath205}{k(4k^{2}-8k+3)}\leq \alpha \leq \frac{n^{2}[2(k+1)^{2}n^{2}\lambda + k(2k^{2}+k-1)]}{% k(4k^{3}-12k^{2}+11k-3)}.\ ] ] provided that @xmath206 ( [ alphamax ] ) , @xmath165 , and @xmath103 .
according to all calculations in this section , the validity of the first law of thermodynamics of black holes can be simply verified .
in this paper , a review of the existence of non extremal asymptotically ads taub - nut / bolt solutions in egb gravity with curved base spaces has been discussed .
the spacetime with base space @xmath5 has a curvature singularity at @xmath2 .
this fact violates the second condition for having a nut solution .
therefore , the metric basically does not admit nut solutions .
but , as this condition has been satisfied numerically , we may proceed with computations the same as in the @xmath3 case .
we show that in the limit of @xmath58 , @xmath207 for both base spaces , goes to @xmath122 in einstein gravity .
also if @xmath6 goes to @xmath123 , or @xmath124 , @xmath119 would be complex in some range of @xmath73 so that in both base spaces , we find that @xmath208 is not real at @xmath125 for all values of r in the range @xmath209 and @xmath38 .
this fact is true for nut solutions of arbitrary even dimensions .
the boundary of taub - nut - ads spacetime at some large radius @xmath73 is obtained , and it was shown that the gauss - bonnet coefficient @xmath0 enters the metric on the boundary due to @xmath43 . the investigation of the thermodynamics of nut / bolt solutions in six dimensions is carried out . by obtaining the regularized action and stress tensor ,
the total mass is derived .
the smarr - type formula for the mass as a function of the extensive parameter s is calculated . by computing the hawking temperature ,
the validity of the first law of thermodynamics is demonstrated .
it is shown that in nut solutions all thermodynamic quantities for both base spaces are related to each other by substituting @xmath210\alpha^{s^{2}\times s^{2}\times ... s_{k}^{2}}$ ] .
this relation is not true for bolt solutions .
so , no further information is given by investigating nut solutions in the @xmath5 case , and we can remove this metric from nut solutions . a generalization of the thermodynamics of black holes in arbitrary even dimensions is made using a new method based on the gibbs- duhem relation and gibbs free energy for nut solutions .
according to this method , the finite action for tnads spacetimes in egb gravity is obtained by considering the generalized finite action for tnads spacetimes in einstein gravity with an additional term as a function of @xmath6 .
this term should vanished if @xmath58 .
the great importance of this method is the computation of the finite action , which is easier than the counterterm prescription , for some spacetimes higher than seven dimensions in higher derivative gravity .
however , this is just a conjecture that its validity must be checked as an open problem .
finally , we perform the stability analysis by investigating the heat capacity and entropy in the allowed range of @xmath6 , @xmath7 , and @xmath8 . for nut solutions in @xmath9 dimensions , there exists a stable phase at a narrow range of @xmath6 for both base factors . for bolt solutions in six dimensions ,
the metric is completely stable in the allowed range of @xmath6 and@xmath8 for the base factor @xmath10 , and is completely unstable for @xmath97 case .
an interesting open problem would be to compare the results of different prescriptions ( the counterterm method , kounterterm , and our procedure ) in computing the finite action .
the ricci scalar and @xmath29 for @xmath211 in the bulk are : @xmath212,\ ] ] @xmath213.\end{aligned}\ ] ] in this case counterterm has been obtained @xcite as @xmath214 \nonumber \\ & & + \mathsf{\theta } \left ( d-8\right ) \bigg[\bigg(1 + 31/30(u-1)\bigg)l_{e}-\frac{19\ell _ { c}^{5}}{57600}(u-1)\mathcal{l}_{(3)}^{(in ) } \bigg ] \bigg\ } , \label{ict8}\end{aligned}\ ] ] where @xmath215 and the third lovelock term is @xmath216 all quantities in ( [ ict8 ] ) , ( [ le8 ] ) and ( [ l3 ] ) must be calculated in the induced metric @xmath217 . by considering all actions in the bulk , boundary , and counterterm , the finite action can be obtained as @xmath218 it has a good coincidence with the action ( [ ink ] ) for @xmath219 in section ( [ thermnutk - d ] ) .
99 d. j. gross and e. witten , nucl .
b * 277 * , 1 ( 1986 ) ; r. r. metsaev and a. a. tseytlin , phys .
b * 191 * , 354 ( 1987 ) ; c. g. callan , r. c. myers and m. j. perry , nucl . phys
b * 311 * , 673 ( 1989 ) . s. cherkis and a. hashimoto , jhep * 0211 * , 036 ( 2002 ) [ hep - th/0210105 ] ; r. clarkson , a. m. ghezelbash and r. b. mann , jhep * 0404 * , 063 ( 2004 ) [ hep - th/0404071 ] ; _ ibid_. * 0408 * , 025 ( 2004 ) [ hep - th/0405148 ] . g. kofinas and r. olea , phys .
d * 74 * , 084035 ( 2006 ) [ hep - th/0606253 ] ; jhep * 0711 * , 069 , ( 2007 ) [ arxiv:0708.0782 ] ; o. miskovic and r. olea , jhep * 0710 * , 028 ( 2007 ) [ arxiv:0706.4460 ] ; r. olea , jhep * 0704 * , 073 ( 2007 ) [ hep - th/0610230 ] . | we give a review of the existence of taub - nut / bolt solutions in einstein gauss - bonnet gravity with the parameter @xmath0 in six dimensions .
although the spacetime with base space @xmath1 has curvature singularity at @xmath2 , which does not admit nut solutions , we may proceed with the same computations as in the @xmath3 case . the investigation of thermodynamics of nut / bolt solutions in six dimensions is carried out .
we compute the finite action , mass , entropy , and temperature of the black hole .
then the validity of the first law of thermodynamics is demonstrated .
it is shown that in nut solutions all thermodynamic quantities for both base spaces are related to each other by substituting @xmath4\alpha^{s^{2 } \times s^{2}\times ... s_{k}^{2}}$ ] .
so no further information is given by investigating nut solution in the @xmath5 case .
this relation is not true for bolt solutions .
a generalization of the thermodynamics of black holes to arbitrary even dimensions is made using a new method based on the gibbs - duhem relation and gibbs free energy for nut solutions . according to this method ,
the finite action in einstein gauss - bonnet is obtained by considering the generalized finite action in einstein gravity with an additional term as a function of @xmath6 .
stability analysis is done by investigating the heat capacity and entropy in the allowed range of @xmath6 , @xmath7 and @xmath8 . for nut solutions in @xmath9 dimensions ,
there exist a stable phase at a narrow range of @xmath6 . in six - dimensional bolt solutions ,
metric is completely stable for @xmath10 , and is completely unstable for @xmath11 case . |
compton scattering of real photons ( rcs ) is one of the simplest reactions for obtaining information on the structure of a stable composite system .
when expanded in the frequency of the photon , the leading - order term of the low - energy scattering amplitude is specified by the model - independent thomson limit in terms of the charge and the mass of the target .
genuine structure effects first appear at second order and can be parametrized in terms of the electric and magnetic polarizabilities ( for an overview see , e.g. , refs .
as there is no stable pion target , the empirical information on the electromagnetic polarizabilities has been extracted from high - energy pion - nucleus bremsstrahlung @xcite and radiative pion photoproduction off the nucleon @xcite . in principle , the electromagnetic polarizabilities of the pion also enter into the crossed process @xmath2 . however , there is some debate concerning the accuracy of extracting these quantities from the crossed channel @xcite . from a theoretical point of view ,
a precise determination of the pion polarizabilities is of great importance , since ( approximate ) chiral symmetry allows one to predict the electromagnetic polarizabilities of the charged pion in terms of the radiative decay @xmath3 @xcite .
corrections to the leading - order pcac result have been calculated at @xmath4 in chiral perturbation theory and turn out to be be rather small @xcite .
new experiments are presently being carried out @xcite or have been proposed @xcite to significantly reduce the uncertainties in the empirical results and thus subject the predictions of chiral symmetry to a stringent test .
clearly , the possibilities to investigate the structure of the target increase substantially if virtual photons are used , because energy and three - momentum can be varied independently and , furthermore , the longitudinal component of the transition current can be explored .
in particular , virtual compton scattering ( vcs ) off the nucleon , as tested in the reaction @xmath5 , has attracted considerable interest ( see , e.g. , refs .
the pion - vcs amplitude of @xmath0 can , in principle , be studied through the inelastic scattering of high - energy pions off atomic electrons , @xmath6 .
such events are presently analyzed as part of the selex e781 experiment @xcite . in this paper
, we will investigate the vcs reaction @xmath0 in the framework of chiral perturbation theory at @xmath1 .
we will first give a short survey of chiral perturbation theory and then define our conventions for the vcs invariant amplitude .
we then discuss the result for the soft - photon and residual amplitudes , respectively .
finally , the model - dependent residual amplitude is analyzed in terms of alternative definitions of generalized polarizabilities .
chiral perturbation theory ( chpt ) @xcite is based on the chiral @xmath7 symmetry of qcd in the limit of vanishing @xmath8- and @xmath9-quark masses .
the assumption of spontaneous symmetry breaking down to @xmath10 gives rise to three massless pseudoscalar goldstone bosons with vanishing interactions in the limit of zero energies .
these goldstone bosons are identified with the physical pion triplet , the nonzero pion masses resulting from an explicit symmetry breaking in qcd through the quark masses .
the effective lagrangian of the pion interaction is organized in a so - called momentum expansion , @xmath11 where the subscripts refer to the order in the expansion .
interactions with external fields , such as the electromagnetic field , as well as explicit symmetry breaking due to the finite quark masses , are systematically incorporated into the effective lagrangian .
covariant derivatives and quark - mass terms count as @xmath12 and @xmath13 , respectively .
weinberg s power counting scheme @xcite allows for a classification of the feynman diagrams by establishing a relation between the momentum expansion and the loop expansion .
the most general chiral lagrangian at @xmath13 is given by @xmath14,\ ] ] where @xmath15 is a unimodular unitary @xmath16 matrix , transforming as @xmath17 for @xmath18 .
as a parametrization of @xmath15 we will use @xmath19 where @xmath20 denotes the pion - decay constant in the chiral limit : @xmath21=92.4 $ ] mev .
we will work in the isospin - symmetric limit @xmath22 .
the quark mass is contained in @xmath23 at @xmath13 , where @xmath24 is related to the quark condensate @xmath25 .
the covariant derivative @xmath26 $ ] contains the coupling to the electromagnetic field @xmath27 . the most general structure of @xmath28 , first obtained by gasser and leutwyler ( see eq . ( 5.5 ) of ref .
@xcite ) , reads , in the standard trace notation , @xmath29 \right\}^2 + \frac{l_2}{4}\mbox{tr}[d_{\mu}u ( d_{\nu}u)^{\dagger } ] \mbox{tr}[d^{\mu}u ( d^{\nu}u)^{\dagger } ] + \frac{l_3}{16}\left[\mbox{tr}(\chi u^\dagger+ u\chi^\dagger)\right]^2 \nonumber\\ & & + \frac{l_4}{4}\mbox{tr}[d_\mu u(d^\mu\chi)^\dagger + d_\mu\chi(d^\mu u)^\dagger ] + l_5\left[\mbox{tr}(f^r_{\mu\nu}u f^{\mu\nu}_lu^\dagger ) -\frac{1}{2}\mbox{tr}(f_{\mu\nu}^l f^{\mu\nu}_l + f_{\mu\nu}^r f^{\mu\nu}_r)\right]\nonumber\\ & & + i\frac{l_6}{2}\mbox{tr } [ f^r_{\mu\nu } d^{\mu } u ( d^{\nu } u)^{\dagger } + f^l_{\mu\nu } ( d^{\mu } u)^{\dagger } d^{\nu } u ] -\frac{l_7}{16}\left[\mbox{tr}(\chi u^\dagger - u\chi^\dagger)\right]^2 + \cdots,\end{aligned}\ ] ] where three terms containing only external fields have been omitted . for the electromagnetic interaction ,
the field - strength tensors are given by @xmath30 .
in the following , we will discuss the vcs amplitude for @xmath31 ( @xmath32 , @xmath33 , @xmath34 ) . throughout the calculation we use the conventions of bjorken and drell @xcite with @xmath35 , @xmath36 .
for the isospin decomposition of the invariant amplitude we use @xmath37 where @xmath38 and @xmath39 denote the _ cartesian _ isospin indices of the initial and final pions , respectively . with the definition @xmath40|0\ ! > , \quad amplitudes in terms of the isospin amplitudes @xmath41 we split the contributions to @xmath42 into a pole piece ( @xmath43 ) and a one - particle - irreducible , residual part ( @xmath44 ) , @xmath45 , @xmath46 ( see fig .
[ vcsdiagrams.fig ] ) . since the @xmath47 is its own antiparticle , the electromagnetic vertex @xmath48 vanishes due to charge - conjugation invariance and hence @xmath49 . in general , the pole piece @xmath50 and the one - particle - irreducible piece @xmath51 are not separately gauge invariant .
according to weinberg s power counting , a calculation of the @xmath52- and @xmath8-channel pole terms at @xmath1 involves the renormalized irreducible vertex at @xmath1 , @xmath53,\quad q = p'-p,\ ] ] where @xmath54 is the prediction for the electromagnetic form factor of the pion ( see eq .
( 15.3 ) of ref .
@xcite ) . to that order ,
the renormalized propagator is simply given by @xmath55 with @xmath56 the @xmath1 result for the pion mass squared ( see eq . ( 12.2 ) of ref .
note that eqs .
( [ gammamu ] ) and ( [ prop ] ) satisfy the ward - takahashi identity @xcite @xmath57 . with these ingredients
the result for @xmath50 at @xmath1 reads @xmath58 + 2 q\cdot\epsilon\ , q\cdot\epsilon'^\ast \frac{1-f(q^2)}{q^2}\right\},\ ] ] which is easily seen not to be gauge invariant by itself .
the set of one - particle - irreducible diagrams is shown in fig .
[ classb.fig ] and gives rise to a residual part of the form @xmath59+\tilde{\cal b}_r.\ ] ] we combine eqs .
( [ bp ] ) and ( [ br ] ) into the form @xmath60 where @xmath61,\ ] ] with the result that @xmath62 and @xmath63 are now separately gauge invariant . in particular , @xmath62 has the form of the soft - photon result obtained in eq .
( 10 ) of ref .
@xcite . a somewhat different approach for obtaining the soft - photon result can be found in ref .
as has been discussed in detail in ref .
@xcite , a gauge - invariant parametrization of the residual amplitude for @xmath0 can be written in terms of three invariant functions @xmath64 , where @xmath65 . at @xmath1 ,
the result for the residual isospin amplitudes @xmath66 and @xmath63 reads : @xmath67,\end{aligned}\ ] ] where the combination @xmath68 is determined through the decay @xmath3 . in eqs .
( [ ar ] ) and ( [ tbr ] ) we have introduced the abbreviation @xmath69 -\frac{q^2}{2q\cdot q'}\left[j^{(0)}(a)-j^{(0)}(b)\right],\ ] ] where @xmath70\ ] ] and @xmath71 the one - loop integrals @xmath72 and @xmath73 are given by ( see appendix c of ref .
@xcite ) @xmath74 with @xmath75.\ ] ] a comparison of eqs .
( [ ar ] ) and ( [ tbr ] ) with eq .
( 18 ) of ref .
@xcite shows that , at @xmath1 , only one of the three functions @xmath64 contributes , i.e. , @xmath76 .
furthermore , at this order in the chiral expansion , the function @xmath77 does not depend on @xmath78 .
our result for @xmath66 is in agreement with ref .
@xcite , where the photoproduction of neutral pion pairs in the coulomb field of a nucleus was studied .
in order to discuss the generalized polarizabilities , we expand the function @xmath79 of eq .
( [ functiong ] ) for negative @xmath80 around @xmath81 , @xmath82,\quad j^{(0)'}(x)=\frac{d j^{(0)}(x)}{dx},\ ] ] where @xmath83 = \frac{1}{x}\left[1 + 2j^{(-1)'}(x)\right],\quad x < 0.\ ] ] for the charged and neutral pion we obtain , respectively , @xmath84 we will first discuss the generalized polarizabilities as defined in ref .
@xcite , where the residual amplitude was analyzed in the photon - pion center - of - mass frame in terms of a multipole expansion .
only terms _ linear _ in the frequency of the final photon were kept , and the result was parametrized in terms of `` generalized polarizabilities . ''
the connection with the covariant approach was established in ref.@xcite , where it was also found that only two of the three polarizabilities @xmath85 , @xmath86 , and @xmath87 of ref .
@xcite are independent , once the constraints due to charge conjugation are combined with particle - crossing symmetry . according to eqs .
( 35 ) and ( 36 ) of ref .
@xcite we define generalized electric and magnetic polarizabilities @xmath88 and @xmath89 , respectively , as @xmath90,\\ \label{beta1 } \beta(|\vec{q}|^2)&\equiv&-\frac{e^2}{4\pi}\sqrt{\frac{3}{8}}p^{(11,11)0 } ( |\vec{q}|)= \frac{e^2}{8\pi m_\pi}\sqrt{\frac{m_\pi}{e_i } } f_1(\omega_0 ^ 2-|\vec{q}|^2,0,0),\end{aligned}\ ] ] where @xmath91 .
a few remarks are in order at this point . 1 . in our present work
, we strictly stick to the convention of ref .
this is why eqs .
( [ alpha1 ] ) and ( [ beta1 ] ) differ by an overall factor @xmath92 from ref .
@xcite , where in eq .
( 1 ) an additional factor @xmath93 was introduced for the spin-0 case .
the variable @xmath80 only appears in the combination @xmath94 , resulting in @xmath95 3 .
the factor @xmath96 originates from an additional normalization factor @xmath97 in eq .
( 32 ) of ref.@xcite , such that @xmath98 using the results of eqs .
( [ f1pp ] ) and ( [ f1p0 ] ) together with @xmath99 ,
we then obtain @xmath100 , \nonumber\\ & & \\
\label{alphap01 } \alpha_{\pi^0}(|\vec{q}|^2)&=&-\beta_{\pi^0}(|\vec{q}|^2 ) = \frac{e^2}{4\pi } \frac{1}{(4\pi f_\pi)^2 m_\pi}\sqrt{\frac{m_\pi}{e_i } } \left(1 - 2\frac{m_\pi- e_i}{m_\pi}\right ) j^{(0)'}\left(2\frac{m_\pi- e_i}{m_\pi}\right).\end{aligned}\ ] ] at the one - loop level , the @xmath101 dependence is entirely given in terms of the pion mass @xmath102 and the pion - decay constant @xmath103 , i.e. , no additional @xmath1 low - energy constant enters . at @xmath104 , eqs .
( [ alphapp1 ] ) and ( [ alphap01 ] ) reduce to the rcs polarizabilities @xcite @xmath105 where we made use of @xmath106 . at @xmath4 ,
the rcs predictions for the charged pion read @xmath107 and @xmath108 @xcite . the corresponding corrections amount to a 12% ( 24% ) change of the @xmath1 result , indicating a good convergence .
we also note that the original degeneracy @xmath109 is lifted at @xmath4 .
the predictions of chpt have to be compared with the empirical results @xmath110 @xcite , @xmath111 @xcite , and @xmath112 @xcite .
clearly , an improved accuracy is required to test the chiral predictions . for
the neutral pion , the @xmath4 corrections turn out to be much larger , @xmath113 and @xmath114 @xcite .
another generalization of the rcs polarizabilities is obtained by parametrizing the invariant amplitude as @xmath115 where @xmath116 and @xmath117 refer , respectively , to the gauge - invariant combinations @xmath118 the functions @xmath119 , @xmath120 , and @xmath121 are _ even _ functions of @xmath43 . introducing the suggestive notation @xmath122 the structures of eq .
( [ mvcs2 ] ) are particularly simple when evaluated in the pion breit frame ( p.b.f . ) defined by @xmath123 , @xmath124 _ { p.b.f.},\\ p_\mu f^{\mu\nu } p^\rho f'_{\rho\nu}&= & [ -p^2_0 \vec{e}\cdot\vec{e}']_{p.b.f.},\\ p^\nu q^\mu f_{\mu\nu } p^\rho q^\sigma f'_{\sigma\rho}&= & [ p^2_0 \vec{q}\cdot \vec{e}\vec{q}\cdot\vec{e}']_{p.b.f.}.\end{aligned}\ ] ] note that by definition @xmath125_{p.b.f.}=p^2 $ ] . in the p.b.f . , eq . ( [ mvcs2 ] )
can thus be expressed as @xmath126_{p.b.f.}.\ ] ] since @xmath127 , @xmath128 contains both transverse and longitudinal components with respect to @xmath129 , for which reason we will introduce the quantities @xmath130 and @xmath131 below : @xmath132 \vec{e}_l\cdot\vec{e}'\right\}_{p.b.f.}.\ ] ] we now consider the limit @xmath133 of the residual amplitudes , for which @xmath134 , and define three generalized dipole polarizabilities in terms of the invariants of eq .
( [ mvcs2 ] ) , @xmath135,\end{aligned}\ ] ] the superscript @xmath136 referring to the residual amplitudes beyond the soft - photon result . in general , the transverse and longitudinal electric polarizabilities @xmath130 and @xmath131 will differ by a term , vanishing however in the rcs limit @xmath137 . comparing with eq .
( [ mvcsnbf2 ] ) , the generalized dipole polarizabilities are seen to be defined such that they multiply the structures @xmath138 , @xmath139 , and @xmath140 , respectively , as @xmath133 .
we note that @xmath141_{p.b.f.}$ ] and @xmath142_{p.b.f.}$ ] are of @xmath143 whereas @xmath144_{p.b.f.}= { \cal o}(\omega'^2)$ ] , i.e. , that different powers of @xmath145 have been kept . at @xmath137 ,
the usual rcs polarizabilities are recovered , @xmath146 the connection to the generalized polarizabilities of guichon _ et al .
_ @xcite can either be established by direct comparison or via the results of ref .
@xcite , @xmath147 with @xmath148 , @xmath149 , and @xmath150 , and all variables referring to the cm frame . using eqs .
( 35 ) - ( 37 ) of ref .
@xcite , we find that the transverse electric dipole polarizability is part of a second - order contribution in @xmath145 beyond the approximation of guichon _ et al .
_ , @xmath151 where @xmath152 . at @xmath1 , @xmath76 , with the result of particularly simple expressions for the generalized dipole polarizabilities , @xmath153 , \\
\label{alphap02 } \alpha^{\pi^0}_l(q^2)&= & \alpha^{\pi^0}_t(q^2)= -\beta^{\pi^0}(q^2 ) = \frac{e^2}{4\pi}\frac{1}{(4\pi f_\pi)^2m_\pi } \left(1-\frac{q^2}{m^2_\pi}\right ) j^{(0)'}\left(\frac{q^2}{m^2_\pi}\right).\end{aligned}\ ] ] the results for the generalized dipole polarizabilities are shown in fig .
[ alpha.fig ] .
even though chiral perturbation theory is only applicable for small external momenta , for the sake of completeness we also quote the asymptotic behavior as @xmath154 , @xmath155 as in the case of real compton scattering , we expect the degeneracy @xmath156 to be lifted at the two - loop level .
we have calculated the invariant amplitudes for virtual compton scattering off the pion , @xmath157 , at the one - loop level , @xmath1 , in chiral perturbation theory . for the charged pion , the result may be decomposed into a gauge - invariant soft - photon amplitude involving the electromagnetic form factor of the pion and a gauge - invariant residual amplitude . for the neutral pion ,
the soft - photon amplitude vanishes .
we have analyzed the low - energy behavior of the residual amplitudes in terms of generalized polarizabilities . in this context
we have introduced two alternative definitions of the generalized polarizabilities , a first one based on a multipole expansion in the center - of - mass frame , and a second one based on a covariant approach interpreted in the pion breit frame .
the connection between the different approaches has been established . in the framework of chpt at @xmath1 , the momentum dependence of the generalized polarizabilities
is entirely predicted in terms of the pion mass and the pion - decay constant , i.e. , no additional counter - term contribution appears . as in the case of real compton scattering , the results at @xmath1 show a degeneracy of the polarizabilities , @xmath156 , which we expect to be lifted at the two - loop level .
this work was supported by the deutsche forschungsgemeinschaft ( sfb 443 ) .
a. l. thanks the theory group of the institut fr kernphysik for the hospitality and support during his stay in mainz where part of his work was done .
b. r. holstein , comments nucl .
* 19 * , 239 ( 1990 ) .
a. i. lvov , int .
phys . a * 8 * , 5267 ( 1993 ) .
s. scherer , _ real and virtual compton scattering at low energies , lectures at the 11th indian - summer school on intermediate energy physics _ , prague , czech republic , september 1998 , to appear in czech .
, mainz university report no .
mkph - t-99 - 1 , nucl - th/9901056 , 1999 .
m. antipov _
121b * , 445 ( 1983 ) .
m. antipov _
et al . _ , z. phys .
c * 26 * , 495 ( 1985 ) .
t. a. aibergenov _
et al . _ ,
j. phys . *
b36 * , 948 ( 1986 ) .
d. babusci , s. bellucci , g. giordano , g. matone , a. m. sandorfi , and m. a. moinester , phys .
b * 277 * , 158 ( 1992 ) .
a. e. kaloshin and v. v. serebryakov , phys .
b * 278 * , 198 ( 1992 ) .
j. f. donoghue and b. r. holstein , phys .
d * 48 * , 137 ( 1993 ) .
a. e. kaloshin and v. v. serebryakov , z. phys .
c * 64 * , 689 ( 1994 ) .
j. portols and m. r. pennington , in _ the second da@xmath158ne physics handbook _ , edited by l. maiani , g. pancheri , and n. paver ( sis , frascati , italy , 1995 ) .
m. v. terentev , sov .
* 16 * , 87 ( 1973 ) .
u. brgi , phys .
b * 377 * , 147 ( 1996 ) ; nucl .
phys . * b479 * , 392 ( 1996 ) .
j. ahrens _ et al .
_ , few - body syst
* 9 * , 449 ( 1995 ) .
m. a. moinester and v. steiner , _ pion and kaon polarizabilities and radiative transitions _ , in proceedings of the workshop _
chiral dynamics : theory and experiment _ , mainz , germany , september 1997 , edited by a. m. bernstein , d. drechsel , and th .
walcher ( springer , berlin , 1998 ) .
t. gorringe ( spokesman ) , triumf expt .
e838 ( 1998 ) .
n. dhose , _ experiments on nucleon polarizabilities _ , in proceedings of the workshop _
chiral dynamics : theory and experiment _ , mainz , germany , september 1997 , edited by a. m. bernstein , d. drechsel , and th .
walcher ( springer , berlin , 1998 ) .
p. a. m. guichon and m. vanderhaeghen , prog .
. phys . * 41 * , 125 ( 1998 ) .
m. a. moinester et al .
( the selex collaboration ) , tel aviv university report no . taup-2568 - 99 , hep - ex/9903039 , 1999 .
s. weinberg , physica * 96a * , 327 ( 1979 ) .
j. gasser and h. leutwyler , ann .
( n.y . ) * 158 * , 142 ( 1984 ) .
j. gasser and h. leutwyler , nucl
. phys . * b250 * , 465 ( 1985 ) .
j. d. bjorken and s. d. drell , _ relativistic quantum mechanics _ ( mcgraw - hill , new york , 1964 ) .
j. c. ward , phys . rev .
* 78 * , 182 ( 1950 ) .
y. takahashi , nuovo cim . * 6 * , 371 ( 1957 ) .
h. w. fearing and s. scherer , few - body syst .
* 23 * , 111 ( 1998 ) .
a. i. lvov , s. scherer , b. pasquini , c. unkmeir , and d. drechsel , in preparation .
d. drechsel , g. knchlein , a. metz , and s. scherer , phys .
c * 55 * , 424 ( 1997 ) .
s. bellucci , j. gasser , and m. e. sainio , nucl . phys . * b423 * , 80 ( 1994 ) ; ibid . *
b431 * , 413 ( 1994 ) .
r. barbieri , j. a. mignaco , and e. remiddi , nuovo cim .
* 11 a * , 824 ( 1972 ) .
g. t hooft and m. veltman , nucl . phys . *
b153 * , 365 ( 1979 ) .
a. a. belkov , m. dillig , and a. v. lanyov , j. phys .
g * 23 * , 823 ( 1997 ) .
guichon , g.q .
liu , and a.w .
thomas , nucl .
a591 * , 606 ( 1995 ) . | we present a calculation of the virtual compton scattering amplitude for @xmath0 in the framework of chiral perturbation theory at @xmath1 .
we explicitly derive expressions for generalized electromagnetic polarizabilities and discuss alternative definitions of these quantities . |
let @xmath0 be _ i.i.d . _ random variables on @xmath1 with density function @xmath4 .
the points form a partially ordered set @xmath5 under the partial order @xmath6 let @xmath7 denote the length of a longest chain . ] in @xmath8 , and for @xmath9 let @xmath10 denote the length of a longest chain in @xmath8 consisting of points less than @xmath11 .
we are interested in the asymptotic properties of @xmath12 as @xmath13 .
when @xmath2 is a smooth density on @xmath14^d$ ] , hence @xmath15 , the problem of studying the asymptotics of @xmath7 has a long history .
it begins with ulam s famous problem @xcite of finding the length of a longest increasing subsequence of a random permutation .
hammersley @xcite made some of the first breakthroughs in understanding ulam s problem .
he observed that the distribution of the length of a longest increasing subsequence among @xmath16 numbers chosen uniformly at random is the same as the distribution of @xmath7 for uniformly distributed points on @xmath14 ^ 2 $ ] . using subadditive ergodic theory ,
hammersley showed that @xmath17 converges almost surely to a constant @xmath18 as @xmath13 , and he conjectured that @xmath19 . in subsequent papers , vershik and kerov @xcite and logan and shepp
@xcite showed that @xmath20 and @xmath21 , respectively .
hammersley s results were generalized by bollobs and winkler @xcite to uniformly distributed points on @xmath14^d$ ] ; they showed that there exist positive constants @xmath22 such that @xmath23 almost surely as @xmath24 , and @xmath25 as @xmath26 .
the only known values of @xmath22 are @xmath27 and @xmath28 .
deuschel and zeitouni @xcite generalized hammersley s results in another direction . for @xmath0 _ i.i.d .
_ on @xmath14 ^ 2 $ ] with @xmath29 density function @xmath30 ^ 2 \to { \mathbb{r}}$ ] , bounded away from zero , they showed that @xmath31 in probability , where @xmath32 is the supremum of the energy @xmath33 over all @xmath34\to[0,1]$ ] nondecreasing and right continuous .
there is another motivation for studying the asymptotics of @xmath12 that arises in multi - objective optimization problems .
such problems are of immense importance in many fields of science and engineering , including control theory and path planning @xcite , gene selection and ranking @xcite , data clustering @xcite , database systems @xcite and image processing and computer vision @xcite . in a discrete multi - objective optimization problem ,
one has several objective functions @xmath35 , where @xmath36 and @xmath37 is a finite set , and is tasked with finding an element @xmath38 that minimizes _ all _ of the functions simultaneously .
this is generally an impossible task , and instead , a family of solutions are obtained based on the notion of pareto - optimality . a feasible solution @xmath39 is called _ pareto - optimal _ if for every @xmath40 , we have @xmath41 for some @xmath42 , or @xmath43 for all @xmath42 ; in other words , no other feasible solution is better in every objective .
the collection of pareto - optimal elements is denoted @xmath44 and called the _ first pareto front_. it is the most natural notion of solution for a discrete multi - objective optimization problem .
if we set @xmath45 for @xmath46 , then assuming all @xmath3 are distinct , it is not hard to see that @xmath47 the second pareto front , @xmath48 , consists of the pareto - optimal elements of @xmath49 , and in general @xmath50 the pareto front that a particular feasible solution lies on is useful for ranking feasible solutions . as before , when the @xmath3 are all distinct we have @xmath51 this observation is essential .
it says that studying the asymptotic shapes of the pareto fronts @xmath52 is equivalent to studying the longest chain function @xmath12 .
figure [ fig : example - fronts ] shows the pareto fronts for @xmath53 points uniformly distributed on @xmath14 ^ 2 $ ] , and figure [ fig : example - fronts-10 ^ 6 ] shows the pareto fronts for @xmath54 points .
the points @xmath3 that are on the same pareto front are connected by a continuous staircase curve that represents the jump set of @xmath12 . in the multi - objective optimization literature ,
the process of computing the pareto fronts for a collection of points is called _ non - dominated sorting _ @xcite . in the combinatorics literature , the partition @xmath55
is called the _ canonical antichain partition _ @xcite .
although we have described non - dominated sorting in the context of a discrete optimization problem , it is a fundamental tool in continuous optimization as well .
many state of the art algorithms for continuous optimization involve a large number of discrete subproblems , each of which requires non - dominated sorting .
the most common examples are the so - called genetic and evolutionary algorithms for continuous multi - objective optimization @xcite .
the applications of non - dominated sorting are not restricted to optimization ; indeed , there are further striking applications in combinatorics @xcite , molecular biology @xcite , graph theory @xcite , young tableaux @xcite and even in physical layout problems in the design of integrated circuits @xcite .
the goal of this paper is to study the asymptotics of @xmath12 , and hence the asymptotics of non - dominated sorting .
our main result , theorem [ thm : linfty - conv ] , states that @xmath56 converges almost surely to a continuous function @xmath57 , which is the viscosity solution of a hamilton - jacobi equation .
our proof is based on linking the asymptotics of @xmath12 to a variational problem , which is a generalization of the variational problem discovered by deuschel and zeitouni @xcite to higher dimensions .
the hamilton - jacobi equation satisfied by @xmath57 is the hamilton - jacobi - bellman equation @xcite for the corresponding variational problem .
we describe our main result in section [ sec : main ] , and postpone the proofs to sections [ sec : var ] and [ sec : proof ] . in section
[ sec : num ] , we give a numerical scheme for computing @xmath57 , and show simulation results comparing the level sets of @xmath57 to pareto fronts for various density functions .
for @xmath58 , we write @xmath59 if @xmath60 and @xmath61 .
when @xmath62 for @xmath36 , we write @xmath63 , and we set @xmath64 .
we will always assume @xmath65 .
for @xmath66 , @xmath67 and @xmath68 will retain their usual definitions .
let @xmath69 and let @xmath70 .
we place the following assumptions on @xmath2 and @xmath71 .
* there exists a continuous nondecreasing function @xmath72 satisfying @xmath73 such that @xmath74 for @xmath75 , and @xmath76 for @xmath77 , * @xmath78 is open and bounded with lipschitz boundary .
set @xmath79 ; { \mathbb{r}}^d ) \ , : \ , { \gamma \ : } ' ( t ) \geq 0 \ \text { for all } \ t \in [ 0,1]\right\}.\ ] ] recall that @xmath80 means that @xmath81 for @xmath36 and @xmath82 .
define @xmath83 by @xmath84 and @xmath85 by @xmath86 we make the following definition .
[ def : pm ] given a domain @xmath87 , we say that a function @xmath88 is _ pareto - monotone _ if @xmath89 for all @xmath90 . in section [ sec : var ] ,
we show that @xmath57 is a pareto - monotone viscosity solution of the hamilton - jacobi partial differential equation ( pde ) @xmath91 the pde should be interpreted as the hamilton - jacobi - bellman equation for the value function @xmath57 .
we note that @xmath2 need only be borel - measurable , bounded and have compact support in @xmath92 for @xmath57 to be a viscosity solution of .
the stronger assumptions ( h1 ) and ( h2 ) are needed to prove that @xmath57 is the unique pareto - monotone viscosity solution of ( see theorem [ thm : uniq ] ) under an appropriate boundary condition at infinity .
our main result is [ thm : linfty - conv ] let @xmath2 satisfy ( h1 ) , let @xmath71 satisfy ( h2 ) , and let @xmath0 be _ i.i.d .
_ with density @xmath2
. then there exists a positive constant @xmath22 such that @xmath93 the constants @xmath22 are the same as those given by bollobs and winkler @xcite . in particular , @xmath27 , @xmath28 and @xmath94 as @xmath95 .
when @xmath2 is a product density , i.e. , @xmath96 , the value function @xmath57 is given by @xmath97 for the case @xmath98 and @xmath99 , aldous and diaconis @xcite provided a non - rigorous derivation of by viewing the problem as an interacting particle process .
they used this to motivate their proof that @xmath19 in ulam s problem , but make no rigorous statements about the relationship between and the longest chain problem . a similar ,
though tangentially related , pde also appears in growth models in multiple dimensions that are defined through the height of a random partial order @xcite .
theorem [ thm : linfty - conv ] provides a new tool with which to study the asymptotics of non - dominated sorting and the longest chain problem .
as an example of the applicability of this result , we show in theorem [ thm : stability ] that non - dominated sorting is asymptotically stable under bounded random perturbations .
evidently , theorem [ thm : linfty - conv ] reduces the problem of non - dominated sorting to solving a hamilton - jacobi equation . from an algorithmic perspective
, this may be useful in designing fast approximate algorithms for non - dominated sorting , or finding lengths of longest chains .
we study some of these applications in a subsequent paper @xcite . as motivation ,
let us give an informal derivation of the hamilton - jacobi pde .
suppose @xmath100 is continuous and @xmath101 uniformly .
fix @xmath16 large enough so that @xmath102
. then the @xmath103 pareto front should be well approximated by the level set @xmath104 .
it is not hard to see that @xmath57 should be pareto - monotone ( recall definition [ def : pm ] ) , and hence it is reasonable to assume that @xmath105 for all @xmath42 .
fix @xmath106 with @xmath107 , where @xmath108 denotes the gradient of @xmath57 at @xmath11 , and consider the quantity @xmath109 .
this is approximately the number of pareto fronts passing between @xmath11 and @xmath110 .
when counting these fronts , we may restrict ourselves to the region @xmath111 this is because any samples in @xmath112 will be on a previous pareto front and only samples that are less than @xmath110 can influence the pareto rank of @xmath110 .
see figure [ fig : derivation ] for a depiction of this region and some quantities from the derivation .
since @xmath113 for all @xmath42 , and @xmath57 is @xmath29 , @xmath114 is well approximated by a simplex for small @xmath115 , and furthermore , the samples within @xmath114 are approximately uniformly distributed .
let @xmath116 denote the number of samples falling in @xmath114 . by scaling the simplex into a standard simplex , without disrupting the pareto ordering within @xmath114 ,
it is reasonable to conjecture that the number of pareto fronts within @xmath114 ( or the length of a longest chain in @xmath114 ) is approximately @xmath117 for some constant @xmath18 , independent of @xmath11 . for simplicity
we take @xmath118 . by the law of large numbers
, we have @xmath119 .
hence when @xmath120 is small we have @xmath121 where @xmath122 denotes the lebesgue measure of @xmath114 .
let @xmath123 denote the side lengths of the simplex @xmath114 .
then @xmath124 for a constant @xmath18 which we again take to be @xmath125 . since @xmath126 lies approximately on the tangent plane to the level set @xmath127
, we see that @xmath128 rearranging the above we see that @xmath129 , and hence latexmath:[\[\label{eq : measurea } for small @xmath115 , we can combine and to obtain @xmath131 simplifying , we see that @xmath57 should satisfy @xmath132 up to scaling by a constant .
although this derivation is informal , it is straightforward and conveys the essence of the result .
it is difficult , however , to construct a rigorous proof based on these heuristics .
there are two main reasons for this .
first , it supposes that @xmath56 converges to a limit @xmath57 , which is not obvious .
second , it is essential that @xmath133 , as we require @xmath114 to be an approximate simplex .
solutions of are in general not smooth , and can have points of non - differentiability due to crossing characteristics .
this is true even in the case that @xmath2 is smooth , and is related to the geometry of @xmath71 .
before studying the variational problem , we recall some aspects of the theory of optimal control @xcite that are relevant to our problem
. we will describe the infinite horizon optimal control problem , but the discussion below applies with minor modifications to other variants of optimal control , such as finite horizon or undiscounted problems with exit times .
the state of the control problem , @xmath134 , is assumed to obey the dynamics @xmath135 where @xmath136 is the control , @xmath114 is a topological space , and @xmath137 . given an initial condition @xmath9 , the solution of is denoted @xmath138 .
let @xmath139 the goal in optimal control is to select the control @xmath140 to minimize the cost functional @xmath141 where @xmath142 and @xmath143 .
the value function for this problem is @xmath144 under sufficient regularity assumptions on @xmath18 and @xmath145 ( discussed below ) , the value function is a hlder- ( or lipschitz ) continuous viscosity solution of the hamilton - jacobi - bellman equation @xmath146 where @xmath147 although the variational problem can be cast in this framework , the assumptions on the running cost @xmath148 in the existing literature are too restrictive . for our variational problem , we have @xmath149 , @xmath150 , @xmath151 , @xmath152 and @xmath153 . in the proofs of theorems [ thm : pointwise - upper ] and [ thm : pointwise - lower ] , we require the standard optimal control theory to hold for @xmath2 piecewise constant on arbitrarily small grids . in the standard reference on optimal control @xcite , it is assumed that @xmath154 is uniformly continuous .
this assumption is then used to prove regularity of the value function @xmath155 .
there is relatively little research devoted to relaxing the regularity condition on @xmath18 .
there are some results for the optimal control problem associated with the eikonal equation @xcite , which allow @xmath18 to have discontinuities .
these results assume that @xmath156 and make essential use of either lipschitzness of @xmath155 , or uniform continuity and/or coercivity of @xmath157 , none of which hold for the variational problem
. soravia @xcite and garavello and soravia @xcite considered a running cost of the form @xmath158 , where @xmath159 is continuous and @xmath160 is borel - measurable , and showed that the standard optimal control results hold with minor modifications .
this is incompatible with when @xmath2 is not continuous .
a similar program is carried out for differential games here @xcite .
barles et al .
@xcite study optimal control on multi - domains , where the discontinuity in @xmath18 is assumed to lie in a half - space . under the assumption that @xmath2 is compactly supported , bounded and borel - measurable , the standard results on optimal control hold for the variational problem with minor modifications to the proofs .
in particular , in lemma [ lem : holder ] we show that @xmath57 is hlder - continuous with exponent @xmath161 , and in theorem [ thm : hjb ] we show that @xmath57 is a viscosity solution of .
the uniqueness of viscosity solutions of under the assumption that @xmath2 satisfies ( h1 ) is a more delicate problem .
this is addressed in section [ sec : comp ] . in our main result ,
theorem [ thm : linfty - conv ] , we assume that @xmath2 satisfies ( h1 ) , which is stronger than borel - measurability .
we assume borel - measurability in much of this section so that our results apply to piecewise constant densities , which are used to approximate @xmath2 in the proofs of theorems [ thm : pointwise - upper ] and [ thm : pointwise - lower ] . to be more precise , we set @xmath162^d\}.\ ] ] we note that the assumption @xmath163^d$ ] is not restrictive , as we can make a simple scaling argument to obtain the case where @xmath2 has compact support in @xmath1 .
we also note that borel - measurability of @xmath2 is necessary , as opposed to lebesgue - measurability , to guarantee that the composition @xmath164 is lebesgue measurable .
we now introduce some new notation .
we will write @xmath165 whenever @xmath166 for all @xmath167 $ ] .
we write @xmath168 whenever @xmath169 .
the same definitions apply to @xmath170 and @xmath171 with obvious modifications . for @xmath172 and @xmath173
we set @xmath174 .
for @xmath58 we set @xmath175 we establish here some basic properties of @xmath57 .
namely , in lemma [ lem : holder ] we establish hlder - continuity of @xmath57 , and in lemma [ lem : dpp ] , we establish a dynamic programming principle for @xmath57 .
[ lem : holder ] let @xmath176 .
then @xmath57 is hlder - continuous with exponent @xmath161 and hlder seminorm @xmath177_{\frac{1}{d}}\leq \|f\|^\frac{1}{d}_{l^\infty({\mathbb{r}}^d)}$ ] .
let @xmath178 and let @xmath179 .
choose @xmath180 with @xmath165 and @xmath181 .
since @xmath76 for @xmath182^d$ ] , we may assume that @xmath183^d$ ] for all @xmath184 $ ] .
set @xmath185 \ , : \ , \gamma(t ) \leqq z\}.\ ] ] if for all @xmath167 $ ] we have @xmath186 , then set @xmath187 .
we claim that @xmath188 to see this : in the case that @xmath189 , we have @xmath190 and hence @xmath191 in the case that @xmath187 , we have @xmath192 hence is established .
suppose @xmath193 .
then there must exist @xmath42 such that @xmath194 .
it follows that @xmath195 applying the generalized hlder inequality we see that @xmath196 inserting this into [ eq : holder - a ] we obtain @xmath197 if @xmath198 then inspecting , we see that @xmath199 , which implies .
sending @xmath200 we find that @xmath201 .
we can reverse the roles of @xmath11 and @xmath202 in the preceding argument to obtain the opposite inequality .
[ rem : wremark ] by a similar argument , we can show that @xmath203 is hlder - continuous with exponent @xmath204 and @xmath205_{\frac{1}{d } } \leq \|f\|^\frac{1}{d}_{l^\infty({\mathbb{r}}^d)}$ ] .
[ lem : dpp ] let @xmath176 .
then for any @xmath173 and @xmath206 we have @xmath207 let us denote the right hand side of by @xmath208 .
we first show that @xmath209 .
let @xmath179 and let @xmath180 such that @xmath210 and @xmath211 .
suppose that @xmath212 .
then there exists @xmath213 such that @xmath214 and hence @xmath215 if @xmath216 then there exists @xmath213 such that @xmath217 and hence @xmath218 finally , suppose that @xmath219 and @xmath220 .
then there exists @xmath221 such that @xmath222 . set @xmath223 and define @xmath224 by @xmath225.\ ] ] note that @xmath226 and @xmath227 . since @xmath228 is invariant under a change of parametrization of @xmath229
, we see that @xmath230 sending @xmath200 we obtain @xmath231 .
we now show that @xmath232 . by lemma [ lem : holder ] and remark [ rem : wremark ] , there exists @xmath213 with @xmath59 such that @xmath233 let @xmath179 and let @xmath224 with @xmath226 and @xmath227 such that @xmath234 since @xmath235 , we can concatenate @xmath236 and @xmath237 to find that @xmath238 .
thus we have @xmath239 sending @xmath200 yields @xmath232 .
we digress momentarily to recall the definition of viscosity solution of @xmath240 where @xmath87 is open , @xmath241 is continuous , @xmath242 is bounded , and @xmath243 is the unknown function . for more information on viscosity solutions of hamilton - jacobi equations ,
we refer the reader to @xcite .
the _ superdifferential _ of @xmath244 at @xmath245 , denoted @xmath246 , is the set of all @xmath247 satisfying @xmath248 similarly , the _ subdifferential _ of @xmath244 at @xmath249 , denoted @xmath250 , is the set of all @xmath247 satisfying @xmath251 equivalently , we may set @xmath252 and @xmath253 a _ viscosity subsolution _ of is a continuous function @xmath88 satisfying @xmath254 similarly , a _ viscosity supersolution _ of is a continuous function @xmath243 satisfying @xmath255 the functions @xmath256 and @xmath257 are the lower and upper semicontinuous envelopes of @xmath2 , respectively , defined by @xmath258 and @xmath259 . if @xmath244 is a viscosity subsolution and supersolution of , then we say that @xmath244 is a _ viscosity solution _ of . after a basic proposition
, we establish in theorem [ thm : hjb ] that @xmath57 is a _ pareto - monotone _ viscosity solution of .
[ prop : monotone ] let @xmath87 be open and let @xmath260 be continuous and pareto - monotone .
then @xmath261 let @xmath245 and @xmath262 .
for any index @xmath42 and small enough @xmath263 , we have @xmath264 .
since @xmath155 is pareto - monotone , we have @xmath265 hence @xmath266 as @xmath267 which implies that @xmath268 .
the proof for @xmath269 is similar .
[ thm : hjb ] let @xmath176 .
then the value function @xmath57 defined by is a pareto - monotone viscosity solution of the hamilton - jacobi equation @xmath270 furthermore , @xmath57 satisfies * whenever @xmath271 , we have @xmath272 * @xmath273 for every @xmath274 .
it follows from the definition of @xmath57 that @xmath57 is pareto - monotone , and ( ii ) follows from the fact that @xmath163^d$ ] . for ( i )
, let @xmath275 such that @xmath276 and let @xmath277 such that @xmath278 for some @xmath42 .
set @xmath279 .
since @xmath57 is pareto - monotone we have @xmath280 .
let @xmath281 and @xmath180 such that @xmath165 and @xmath282 .
let @xmath283 if for all @xmath284 $ ] we have @xmath285 , then set @xmath187 . if @xmath286 , then @xmath287 and hence @xmath288 .
if @xmath187 then for every @xmath284 $ ] , @xmath289 , and hence @xmath290 .
it follows that @xmath291 if @xmath221 , then for any @xmath292 , @xmath293 for some @xmath42 , and hence @xmath294 .
set @xmath295 for @xmath167 $ ]
. then @xmath296 and @xmath297 , hence @xmath298 .
sending @xmath200 we see that @xmath299 and hence @xmath300 .
we now show that @xmath57 is a viscosity supersolution of .
let @xmath206 , let @xmath301 , and set @xmath302 .
by lemma [ lem : dpp ] we have @xmath303 let @xmath304 .
since @xmath305 as @xmath306 , we have @xmath307 sending @xmath306 we obtain @xmath308 since @xmath309 was arbitrary , we obtain @xmath310 since @xmath57 is pareto - monotone , proposition [ prop : monotone ] yields @xmath311 . hence if @xmath312 then @xmath57 is trivially a viscosity supersolution of at @xmath313 .
we may therefore suppose that @xmath314 .
fix @xmath42 and set @xmath315 for @xmath316 .
by we have @xmath317 since @xmath318 can be arbitrarily large , we must have @xmath319 for the above to hold .
substituting @xmath320 into and simplifying we obtain @xmath321 thus @xmath57 is a viscosity supersolution of .
we now show that @xmath57 is a viscosity subsolution of .
let @xmath206 , let @xmath179 , and let @xmath322 . by lemmas [ lem : holder ] and
[ lem : dpp ] and remark [ rem : wremark ] , for every @xmath173 there exists @xmath323 with @xmath59 such that @xmath324 .
hence there exists @xmath180 with @xmath325 such that @xmath326 by hlder s inequality @xmath327 as @xmath328 . since @xmath329 as @xmath330
, we have @xmath331 as @xmath330 .
choose @xmath173 small enough so that @xmath332 , and set @xmath333
. then we have @xmath334 since @xmath179 was arbitrary , we see that @xmath335 since @xmath57 is pareto - monotone , we have @xmath311 . if @xmath336 for some @xmath42 , then @xmath337 .
thus we may assume that @xmath338 for all @xmath42 .
then the supremum above is attained at some @xmath339 with @xmath340 . by scaling @xmath341 so that @xmath342
, we see that @xmath343 since @xmath338 for all @xmath42 , we have that @xmath344 it follows that the supremum in is attained at some @xmath345 . introducing a lagrange multiplier @xmath142 , the necessary conditions for @xmath346 to be a maximizer of the above constrained optimization problem are @xmath347 it follows that @xmath348 and @xmath349 . substituting this into we have @xmath350 which completes the proof .
[ rem : truncation ] we remark that @xmath57 satisfies an important truncation property .
namely , if we fix @xmath275 and define @xmath351 by @xmath352 @xmath353 then we have @xmath354 in the viscosity sense . indeed , this follows directly from theorem [ thm : hjb ] by noting that @xmath355 and @xmath356 we aim here to establish that @xmath57 is the unique viscosity solution of under hypotheses ( h1 ) and ( h2 ) on @xmath2 and @xmath71 , which in general allow @xmath2 to be discontinuous .
the standard results on uniqueness of viscosity solutions @xcite assume uniformly continuous dependence on spatial variables .
there has been some recent work relaxing this condition , as it is important in many applications .
tourin @xcite considered hamilton - jacobi equations of the form @xmath357 , where @xmath358 is allowed to have a discontinuity along a smooth surface , and proved a comparison principle under the assumption that @xmath157 is convex and uniformly continuous .
neither assumption holds for , although the non - convexity can be easily remedied .
deckelnick and elliot @xcite prove a comparison principle for lipschitz viscosity solutions of eikonal - type equations of the form @xmath359 , where @xmath2 satisfies a regularity condition similar to ( h1 ) , but slightly more general .
as exhibited by the solution @xmath360 of for @xmath98 , solutions of are not in general lipschitz continuous .
camilli and siconolfi @xcite proposed a new notion of viscosity solution for hamilton - jacobi equations in which @xmath361 has measurable dependence on the spatial variable @xmath11 .
they obtain general uniqueness results under the assumption that @xmath362 is quasiconvex and coercive .
their results do not apply to due to the coercivity assumption .
the main result in this section , theorem [ thm : uniq ] , establishes uniqueness for under hypotheses ( h1 ) and ( h2 ) , and a boundary condition at infinity .
let us give a sketch of the proof now .
let @xmath244 be a pareto - monotone viscosity solution of .
we first prove a standard comparison principle , in theorem [ thm : comparison ] , for uniformly continuous @xmath2 .
we can then define the regularized value functions @xmath363 and @xmath364 by replacing @xmath2 by its inf and sup convolutions @xmath365 and @xmath366 , respectively , in .
since @xmath366 and @xmath365 are lipschitz continuous , the comparison principle from theorem [ thm : comparison ] yields @xmath367 . the proof is completed by showing that @xmath368 as @xmath200 , where @xmath57 is the value function defined by .
we establish a more general result in lemma [ lem : l1-pert ] , the proof of which relies on the second comparison principle , theorem [ thm : comparison2 ] .
this comparison principle holds for @xmath2 and @xmath71 satisfying ( h1 ) and ( h2 ) under the additional assumption that the subsolution is truncatable , as per definition [ def : trunc ] .
as pointed out in remark [ rem : truncation ] , the value function @xmath57 is truncatable , so theorem [ thm : comparison2 ] is applicable in the proof of lemma [ lem : l1-pert ] .
[ thm : comparison ] suppose @xmath369 is uniformly continuous with @xmath163^d$ ] .
let @xmath244 and @xmath155 be viscosity sub- and supersolutions , respectively , of @xmath370 and suppose that @xmath371 and @xmath155 is pareto - monotone . if @xmath372 on @xmath373 then @xmath372 on @xmath92 .
the proof of theorem [ thm : comparison ] utilizes the method of doubling the variables , which is standard in the theory of viscosity solutions @xcite , with appropriate modifications for the boundary condition . for @xmath374 , set @xmath375 , where @xmath376 .
fix @xmath277 and @xmath377 .
it is easy to see that @xmath378 .
hence we have @xmath379 since @xmath155 is pareto - monotone , we have @xmath380 for all @xmath42 , and therefore @xmath381 hence @xmath382 is a viscosity supersolution of @xmath383 and @xmath384 on @xmath373 .
suppose that @xmath385 .
for @xmath386 and @xmath387 , set @xmath388 and @xmath389 .
setting @xmath390 in @xmath391 we see that @xmath392 for all @xmath393 .
let @xmath394 note that @xmath395 , @xmath396 ( by ) and @xmath397 for all @xmath398 .
it follows that @xmath399 , and hence @xmath400 attains a maximum at some @xmath401^d\times [ 0,1]^d$ ] . since @xmath402
, we have @xmath403 it follows that latexmath:[\[\label{eq : xyalpha } since @xmath384 on @xmath373 and @xmath405 is continuous , it follows from and that @xmath406^d \times ( 0,1]^d$ ] for @xmath393 large enough . for
such @xmath393 we have @xmath407 , and hence @xmath408 subtracting the above equations yields @xmath409 , which contradicts the uniform continuity of @xmath2 and as @xmath410 .
therefore @xmath384 on @xmath92 .
sending @xmath411 completes the proof . we can prove a comparison principle for discontinuous @xmath2 by assuming that the subsolution satisfies the truncation property described in remark [ rem : truncation ] . for this , we make the following definition . [
def : trunc ] let @xmath244 be a viscosity subsolution of @xmath412 we say that @xmath244 is _ truncatable _ if for every @xmath413 , @xmath414 is a viscosity subsolution of @xmath415 where @xmath414 and @xmath416 are defined in remark [ rem : truncation ] .
we note that that truncatability is well - defined , i.e. , it depends only on @xmath257 . by remark [ rem : truncation ]
, the value function @xmath57 is truncatable .
it is easy to see that every @xmath29 pareto - monotone subsolution of is truncatable .
it turns out , thanks to theorem [ thm : uniq ] , that every pareto - monotone viscosity solution of satisfying is truncatable .
we place the following assumptions on @xmath417 and @xmath71 .
* there exists a continuous nondecreasing function @xmath72 satisfying @xmath73 such that @xmath74 for @xmath75 , and @xmath76 for @xmath77 . *
@xmath418 is open and bounded with lipschitz boundary . in particular
, since @xmath71 is open ( h1 ) implies that @xmath419 on @xmath92 .
the assumptions on @xmath71 imply that the following cone condition is satisfied .
* for every @xmath420 , there exists a cone @xmath421 with nonempty interior and a neighborhood @xmath422 of @xmath11 such that @xmath423 to see this : for any @xmath420 there exists , by lipchitzness of @xmath424 , a real number @xmath173 and a lipschitz continuous function @xmath425 such that , upon relabelling and reorienting the coordinate axes if necessary , we have @xmath426 one can check that the cone @xmath427 satisfies ( h2@xmath428 ) . as it is more useful in the comparison principle proof , we will assume that ( h2@xmath428 ) holds instead of lipschitzness of the boundary .
we note that the cone condition ( h2@xmath428 ) is similar to the one used by deckelnick and elliot @xcite .
[ thm : comparison2 ] suppose that @xmath71 satisfies ( h2@xmath428 ) and @xmath2 satisfies ( h1 ) .
let @xmath244 and @xmath155 be viscosity sub- and supersolutions , respectively , of @xmath429 and assume that @xmath244 is truncatable and @xmath155 is pareto - monotone .
then @xmath372 on @xmath373 implies that @xmath372 on @xmath92 . as in the proof of theorem
[ thm : comparison ] , the proof below is based on the standard technique of doubling the variables @xcite . the proof is similar to ( * ? ? ?
* theorem 2.3 ) in the way that ( h2@xmath428 ) is used , however , we can not assume lipschitzness of @xmath155 .
the truncatability condition on @xmath244 in a sense replaces the lipschitz condition on @xmath155 in ( * ? ? ?
* theorem 2.3 ) . for @xmath374 , set @xmath430 .
then @xmath431 on @xmath373 . as in the proof of theorem [ thm : comparison ] , @xmath382 is a viscosity supersolution of @xmath432 now suppose that @xmath433 and let @xmath434 since @xmath431 on @xmath435 and @xmath436 is continuous , we see that @xmath437 . by the definition of @xmath438
, there exists @xmath439 such that @xmath440 and every neighborhood of @xmath441 contains a point @xmath313 such that @xmath442 . for @xmath173 set @xmath443 and note that @xmath444 for any @xmath173 .
notice that we may assume ( h2@xmath428 ) holds at any @xmath277 . indeed ,
if @xmath445 then we may set @xmath446 and choose @xmath447 small enough so that @xmath448 . then any cone @xmath421 will suffice as either @xmath449 or @xmath450 .
let @xmath451 be in the interior of @xmath452
. for @xmath453 set @xmath454 for @xmath387 and @xmath455 , set @xmath456 and @xmath457 . by continuity of @xmath414 and @xmath382
we have @xmath458 set @xmath459
let @xmath460 such that @xmath461 .
then there exists @xmath42 such that @xmath462 and we have @xmath463 since @xmath464 and @xmath465 , we can choose @xmath173 small enough so that @xmath466 .
fixing such an @xmath173 we have @xmath467 since @xmath414 and @xmath382 are uniformly continuous on compact sets , it follows from that there exists @xmath468 such that @xmath469 for every @xmath386 .
now let @xmath470 and @xmath471 with @xmath472 , and set @xmath473 then @xmath474 , @xmath475 , @xmath476 , @xmath477 and @xmath478 .
we conclude that @xmath479 hence we have shown that for any @xmath480 @xmath481 fix @xmath387 large enough so that @xmath482 and let @xmath483 such that @xmath484
. then we have @xmath485 in particular , we have @xmath486 , and @xmath487 taking @xmath488 large enough so that @xmath472 , we see from that @xmath489 . now set @xmath490 and @xmath491 .
then we have @xmath492 , @xmath493 , @xmath494 , and @xmath397 .
it follows that @xmath399 .
hence for @xmath495 large enough , @xmath400 attains a global maximum at @xmath496 satisfying @xmath497 and @xmath498 . sending @xmath499 and extracting a subsequence , if necessary
, we may assume that @xmath500 for some @xmath501 .
since @xmath502 we see , by the continuity of @xmath414 and @xmath382 , that @xmath503 it follows that @xmath504 let @xmath505 since @xmath506 is a local max of @xmath400 , we have @xmath507 . since @xmath414 is a truncatable , we have @xmath508 where @xmath509 for @xmath510 and @xmath511 otherwise . by we have @xmath512 combining these we see that @xmath513 if @xmath514 , then for @xmath393 large enough we have either @xmath515 or @xmath516 .
hence by ( h1 ) we have @xmath517 suppose now that @xmath518 .
we have two cases ; either ( 1 ) @xmath519 or ( 2 ) @xmath520 .
case 1 . in this case
we have @xmath521 . if @xmath522 then @xmath523 and hence holds .
if @xmath524 then @xmath525 and holds trivially .
case 2 . set @xmath526 and
note that @xmath527 by we have @xmath528 as @xmath410 . since @xmath529 is in the interior of @xmath452
, we see that @xmath530 for @xmath393 large enough , and hence @xmath531 .
we can take @xmath173 smaller , if necessary , so that @xmath532 . since @xmath497 , we can choose @xmath393 large enough so that @xmath533 .
since @xmath534 for @xmath393 large enough , we have by ( h2@xmath428 ) that @xmath535 and hence @xmath536 and holds . sending @xmath537 in contradicts ,
hence @xmath384 on @xmath92 .
sending @xmath411 we find that @xmath372 on @xmath92 . in order to prove a general uniqueness result , theorem [ thm : uniq ]
, we require a perturbation result for the value function @xmath57 with respect to sup and inf convolutions of the density @xmath2 .
since a similar result , for a different type of perturbation , is required in the proof of our main result , theorem [ thm : linfty - conv ] , we state a more general result in lemma [ lem : l1-pert ] .
we first recall some notation standard in the theory of viscosity solutions . for a sequence of bounded functions @xmath538 ,
the upper and lower limits are defined by @xmath539 and @xmath540 [ lem : l1-pert ] let @xmath541 and suppose that @xmath71 satisfies ( h2@xmath428 ) , @xmath2 satisfies ( h1 ) , and @xmath542 for each @xmath16 , set @xmath543 then @xmath544 uniformly where @xmath57 is the value function given by .
we claim that @xmath545 is a uniformly bounded sequence . to see this ,
suppose to the contrary that there exists a sequence @xmath546 in @xmath14^d$ ] such that @xmath547 as @xmath24 . by passing to a subsequence , if necessary
, we may assume that @xmath548 ^ 2 $ ] as @xmath13 . by the definition of the upper limit and
, we have @xmath549 which contradicts the assumption that @xmath2 satisfies ( h1 ) and establishes the claim .
since @xmath545 is uniformly bounded , there exists ( by lemma [ lem : holder ] ) a constant @xmath550 such that @xmath551_{\frac{1}{d } } \leq c$ ] for all @xmath16 .
the sequence @xmath552 is therefore bounded and equicontinuous , and by the arzela - ascoli theorem there exists a subsequence @xmath553 and a hlder - continuous function @xmath554 such that @xmath555 uniformly on compact sets in @xmath1 as @xmath556 . by theorem [ thm : hjb ] ( i ) , ( ii ) , we conclude that the convergence is actually uniform on @xmath1 . by theorem [ thm : hjb ] , each @xmath552 is a pareto - montone truncatable viscosity solution of @xmath557 by standard results on viscosity solutions ( see ( * ? ? ?
* remark 6.3 ) ) and , we have that @xmath155 is a pareto - monotone viscosity solution of @xmath558 by the assumption that @xmath559^d$ ] , we have that @xmath560 for all @xmath561 , hence @xmath562 for all @xmath561 .
we claim that @xmath155 is truncatable . to see this , fix @xmath413 and define @xmath563 , @xmath564 , @xmath416 and @xmath565 as in remark [ rem : truncation ] . since @xmath552 is truncatable , @xmath564 is a viscosity solution of @xmath566 by the definition of @xmath565 , we have @xmath567 , with @xmath568 for @xmath569 .
it follows that @xmath570 for @xmath571 , there exists a neighborhood @xmath572 of @xmath11 on which @xmath565 is identically zero for all @xmath16 .
it follows that @xmath573 and therefore @xmath574 . since @xmath575 uniformly
, we can again apply standard results on viscosity solutions @xcite to find that @xmath563 is a viscosity subsolution of @xmath576 which proves the claim .
by theorem [ thm : comparison2 ] we have @xmath577 on @xmath92 . since @xmath578 for @xmath561 we have @xmath577 on @xmath1 . the above argument
can be used to show that every subsequence of @xmath552 contains a uniformly convergent subsequence converging to @xmath57 .
it follows that @xmath544 uniformly in @xmath1 as @xmath579 .
we now establish uniqueness of viscosity solutions of .
[ thm : uniq ] suppose that @xmath71 satisfies ( h2@xmath428 ) and @xmath2 satisfies ( h1 ) .
then there exists a unique pareto - monotone viscosity solution @xmath244 of @xmath580 satisfying the additional boundary condition @xmath581 by theorem [ thm : hjb ] there exists a hlder - continuous pareto - monotone viscosity solution @xmath244 of . to prove uniqueness
, we will show that @xmath582 , where @xmath57 is the value function defined by .
let @xmath179 and consider the inf and sup convolutions of @xmath2 , defined for @xmath277 by @xmath583 recall that @xmath365 and @xmath366 are lipschitz continuous with constant @xmath584 and @xmath585 .
without loss of generality , we may assume that @xmath586 , and hence for @xmath281 small enough , we have @xmath587^d$ ] .
for @xmath588 , set @xmath589 and @xmath590 by theorem [ thm : hjb ] , @xmath591 is a viscosity solution of @xmath592 and satisfies the boundary condition . since @xmath593 and @xmath244 is a viscosity solution of
, we see that @xmath244 is a viscosity subsolution of
. since @xmath594 on @xmath595 and @xmath244 satisfies , we can apply theorem [ thm : comparison ] to find that @xmath596 . by a similar argument
, we have that @xmath597 .
since @xmath598 and is satisfied for the sequences @xmath599 and @xmath600 , we have by lemma [ lem : l1-pert ] that @xmath368 uniformly in @xmath1 as @xmath200 , and hence @xmath582 .
the proof of theorem [ thm : linfty - conv ] is split into several steps . in section [ sec : piecewise ] , we prove a basic convergence result for piecewise constant density functions , which is a generalization of the results of deuschel and zeitouni @xcite . in section [ sec : cont ] , we extend the convergence result to densities that are continuous on @xmath71 and vanish on @xmath601 by considering a sequence of piecewise constant approximations to @xmath2 , applying the results from section [ sec : piecewise ] , and passing to the limit . this requires a perturbation result for the energy @xmath228 , which we obtained from the comparison principle for the associated hamilton - jacobi pde in lemma [ lem : l1-pert ] .
we aim to prove a basic convergence result for piecewise constant densities here .
the proof is split into a lower bound , theorem [ thm : lower - bound ] , and an upper bound , theorem [ thm : upper - bound ] .
we should note that the techniques used here are similar to those used by deuschel and zeitouni @xcite , who showed the same convergence result for @xmath29 densities on the unit hypercube in dimension @xmath99 .
let us introduce some notation .
for a finite set @xmath602 , let @xmath603 denote the length of a longest increasing chain in @xmath604 .
the set function @xmath605 has an important invariance . if @xmath606 is a mapping that preserves the partial order @xmath607 , i.e. , @xmath608 , then @xmath609 for @xmath610 we denote by @xmath611 the characteristic function of the set @xmath114 , which takes the value @xmath125 on @xmath114 and @xmath612 on @xmath613 .
when @xmath114 is lebesgue measurable , we denote by @xmath122 the lebesgue measure of @xmath114 .
we set @xmath614 and @xmath376 .
given an integer @xmath615 , we partition @xmath616 into @xmath617 hypercubes of side length @xmath618 .
more precisely , for a multiindex @xmath619 with @xmath620 , where @xmath621 , we set @xmath622 we say that @xmath623 is @xmath615-_piecewise constant _ if @xmath2 is constant on @xmath624 for all @xmath393 . if @xmath2 is @xmath615-piecewise constant then @xmath2 is @xmath625-piecewise constant for all @xmath626 . for convenience
, we also set @xmath627 we now establish an asymptotic lower bound on @xmath628 .
[ thm : lower - bound ] let @xmath623 be @xmath615-piecewise constant , and let @xmath0 be _ i.i.d . _ with density @xmath2
. then @xmath629 . , scaledwidth=60.0% ] let @xmath179 and select @xmath180 with @xmath630 . without loss of generality , we may assume that @xmath631 for all @xmath284 $ ] . let @xmath632 denote the @xmath633 times at which @xmath229 intersects the set @xmath634 set @xmath635 and @xmath636 . for @xmath637 set @xmath638 and @xmath639 for every @xmath640 we have @xmath641 for some @xmath393 .
recalling the definition of @xmath228 we have @xmath642 where the second equality follows from the fact that @xmath2 is constant on @xmath641 . applying the generalized hlder inequality to we
have @xmath643 setting @xmath644 we have @xmath645 fix @xmath646 .
let @xmath647 denote the number of points from @xmath0 falling in @xmath648 and set @xmath649 then @xmath647 is binomially distributed with parameters @xmath16 and @xmath650 . if @xmath2 is identically zero on @xmath648 then @xmath651 with probability one for all @xmath16 , and @xmath652 , hence @xmath653 if @xmath2 is not identically zero on @xmath648 , then since @xmath631 for all @xmath654 , we have @xmath655 and hence @xmath656 .
the conditional law @xmath657 is then uniform on @xmath648 .
let @xmath658 be the indices of the @xmath647 random variables out of @xmath0 that belong to @xmath648 .
let @xmath659 be the injective affine transformation mapping @xmath648 onto @xmath616
. then @xmath660 are independent and uniformly distributed on @xmath616 . by ( * ? ? ?
* remark 1 ) , we have @xmath661 since @xmath662 preserves the partial order @xmath607 , we have by that @xmath663 since @xmath664 almost surely we have @xmath665 combining this , and , we see that @xmath666 since @xmath229 is a monotone curve ( i.e. , @xmath667 ) , we can connect longest chains from each rectangle @xmath648 together to form a chain in @xmath616 .
it follows that @xmath668 combining this with we have @xmath669 which completes the proof . for the proof of the upper bound ,
we need to introduce some new notation .
let @xmath670 be an integer and set @xmath671 .
let @xmath672 be another integer and set @xmath673 . for given @xmath674
, we say that a sequence of multiindices @xmath675 is _ admissible _ if @xmath676 and @xmath677 for all @xmath640 .
we denote the set of admissible multiindices by @xmath678 . for @xmath679 , define @xmath680 in @xmath14^d$ ] by @xmath681 and @xmath682 for @xmath683 . since @xmath684 is admissible , @xmath685 defines a chain in @xmath14^d$ ] .
define @xmath686 \to [ 0,1]^d$ ] to be the polygonal curve connecting the points @xmath685 , i.e. , @xmath687 for @xmath688 $ ] . for @xmath679 and @xmath689 ,
set @xmath690 for each rectangle @xmath691 , we set @xmath692 .
we say that a chain @xmath693 in @xmath616 is @xmath684-_increasing _ if @xmath694 it is not hard to see that for any @xmath674 , every chain in @xmath616 is @xmath684-increasing for some @xmath679 .
see figure [ fig : grid - example ] for an illustration of the above definitions .
[ fig : grid - example ] , @xmath695 , @xmath624 and @xmath696 in two dimensions with @xmath697 . in this case , the unit square is partitioned into four squares , @xmath698 , @xmath699 , @xmath700 and @xmath701 , which are separated by dotted lines in the figure.,title="fig:",scaledwidth=75.0% ] we first need a preliminary lemma which bounds the length of a longest chain within the narrow strip @xmath702^{d-1 } \times [ ( j-1)\delta x , j\delta x),\ ] ] for any @xmath703 .
we note that the following lemma is a generalization of ( * ? ? ?
* lemma 7 ) .
the proof is based on the same idea of using a mixing process to embed @xmath0 into another set of _ i.i.d .
_ random variables that are uniform when restricted to the strip @xmath704 .
[ lem : strip ] let @xmath623 be @xmath615-piecewise constant , and let @xmath0 be _ i.i.d . _ with density @xmath2 .
fix an integer @xmath703 and let @xmath705 .
then @xmath706 set @xmath707 and let @xmath708 .
let @xmath709 be _ i.i.d . _ according to the conditional density @xmath710 where @xmath711 .
let @xmath712 be bernoulli zero - one random variables with parameter @xmath713 and set @xmath714 define @xmath715 through the mixture process @xmath716
then @xmath715 are _ i.i.d . _ with density @xmath717
let @xmath718 denote the cardinality of the set @xmath719 .
then @xmath718 is binomially distributed with parameters @xmath16 and @xmath720 . since @xmath145 is constant on @xmath704
, we can use a similar argument to that in theorem [ thm : lower - bound ] to show that @xmath721 let @xmath722 and note that @xmath723 and that @xmath724 . combining this with we have @xmath725
since @xmath116 is binomially distributed with parameters @xmath16 and @xmath713 , we have @xmath726 almost surely and hence @xmath727 since @xmath728 we have @xmath729 the desired result follows from noting that @xmath730 is monotone nondecreasing along every sample path and @xmath731 as @xmath13 with probability one .
the following short technical lemma is essential in the proof of theorem [ thm : upper - bound ] [ lem : approx ] let @xmath623 be @xmath615-piecewise constant . for every @xmath179 and @xmath732 we have @xmath733 for all @xmath679 , the admissible multiindices , and @xmath734 , where @xmath735 let @xmath736 , and @xmath679 .
set @xmath737 and fix @xmath738 and @xmath739 .
note that @xmath740 and @xmath741 where we set @xmath742 for convenience .
a short computation shows that @xmath743 . since @xmath2 is @xmath615-piecewise constant we have @xmath744 for all @xmath745 and @xmath739 .
noting that @xmath746 and recalling the definition of @xmath228 we have @xmath747 taking @xmath748 completes the proof .
we now establish an asymptotic upper bound on @xmath749 .
[ thm : upper - bound ] let @xmath623 be @xmath615-piecewise constant , and let @xmath0 be _ i.i.d . _ with density @xmath2
. then @xmath750 let @xmath751 , and @xmath679 .
we suppose that @xmath732 is a multiple of @xmath615 so that @xmath2 is @xmath670-piecewise constant .
let @xmath752 denote the length of a longest @xmath684-increasing chain .
let @xmath647 denote the number of @xmath0 that belong @xmath691 and set @xmath753 due to the monotonicity of @xmath685 , at most @xmath754 of @xmath755 can have a non - empty intersection with more than one hypercube @xmath624 .
it follows that @xmath756 , where @xmath757 .
since each @xmath684-increasing chain is the union of chains in @xmath755 , we have @xmath758 we will deal with each of the above sums separately .
for the first term , set @xmath707 and let @xmath670 be large enough so that @xmath759 . since @xmath760 for each @xmath640 , we have by lemma [ lem : strip ] that @xmath761 choose @xmath670 large enough so that @xmath762 we now bound the second sum in . by lemma [ lem : approx ] ,
choose @xmath763 so that @xmath764 for all @xmath679 .
for any @xmath745 , the conditional density @xmath765 on @xmath691 is uniform . by a similar argument as in the proof of theorem [ thm :
lower - bound ] , we have that @xmath766 combining this with , , and we have @xmath767 since every chain in @xmath0 is @xmath684-increasing for some @xmath679 , we have @xmath768 for every @xmath16 .
it follows that @xmath769 which completes the proof .
we now generalize the convergence results on piecewise constant densities , theorems [ thm : lower - bound ] and [ thm : upper - bound ] , to continuous densities on @xmath71 .
our main result , theorem [ thm : linfty - conv ] , is proved at the end of the section .
the idea of our approach is to divide @xmath616 into a large number of hypercubes , and to flatten @xmath2 on each sub - cube .
we can then apply the results from section [ sec : piecewise ] and take the limit as the size of the sub - cubes tends to zero . in order to pass to the limit
, we apply the perturbation result given in lemma [ lem : l1-pert ] .
let @xmath0 be _ i.i.d .
_ with density @xmath2 .
we recall that @xmath10 denotes the length of a longest chain among @xmath0 consisting of points less than or equal to @xmath11 under the partial order @xmath607 . in other words
@xmath770 we also recall the definition of the value function @xmath57 , defined in by @xmath771 we now establish pointwise asymptotic upper and lower bounds on @xmath12 .
[ thm : pointwise - upper ] let @xmath772 satisfy ( h1 ) and let @xmath418 satisfy ( h2 ) .
then for every @xmath275 we have @xmath773 set @xmath774 and @xmath775 .
suppose that @xmath776 .
it follows from ( h1 ) that @xmath2 is lower semicontinuous , and hence @xmath777 for @xmath778 .
thus @xmath779 almost surely .
suppose that @xmath780 and let @xmath281 .
let @xmath626 and partition @xmath781 into @xmath782 hypercubes @xmath783 for multiindices @xmath393 with @xmath784 .
define @xmath785 by @xmath786 and set @xmath787 .
for every integer @xmath788 , @xmath789 is @xmath788-piecewise constant on @xmath781 and @xmath790 .
define @xmath791 by @xmath792 note that the sequence @xmath793 is uniformly bounded , borel - measurable , and has compact support in @xmath14^d$ ] .
furthermore , it follows from that holds for the sequence @xmath789 .
hence by lemma [ lem : l1-pert ] we have that @xmath794 uniformly as @xmath795 .
now fix @xmath788 large enough so that @xmath796 and define @xmath797
. then @xmath798 and we can write @xmath145 as a convex combination of two distributions as follows : @xmath799 let @xmath709 be _ i.i.d .
_ with density @xmath800 , let @xmath712 be bernoulli random variables with parameter @xmath801 , and set @xmath802 define @xmath803 then a simple computation shows that @xmath715 are _ i.i.d .
_ with density @xmath145 .
let @xmath718 denote the cardinality of @xmath804 .
since @xmath145 is @xmath788-piecewise constant on @xmath781 , we can apply theorems [ thm : lower - bound ] and [ thm : upper - bound ] to obtain @xmath805 note that @xmath718 is binomially distributed with parameters @xmath16 and @xmath806 , hence @xmath807 almost surely . applying this
to we have @xmath808 set @xmath722 .
note that @xmath116 is binomially distributed with parameters @xmath16 and @xmath801 , and @xmath809 combining with and the fact that @xmath810 as @xmath13 we have @xmath811 recalling we have @xmath812 as in the proof of lemma [ lem : strip ] , the proof is completed by noting that @xmath730 is monotone nondecreasing along every sample path and @xmath731 as @xmath13 with probability one .
[ thm : pointwise - lower ] let @xmath772 satisfy ( h1 ) and let @xmath418 satisfy ( h2 ) . then for every @xmath275 we have @xmath813 let @xmath179 . as in the proof of theorem [ thm : pointwise - upper ] , we set @xmath814 and we may suppose that @xmath815 . as before , let @xmath626 and partition @xmath781 into @xmath782 hypercubes @xmath783 for multiindices @xmath393 with @xmath784 .
define @xmath785 by @xmath816 and set @xmath787 .
define @xmath817 for any @xmath393 such that @xmath818 , we have by ( h1 ) and that @xmath819 it follows that @xmath820 as @xmath556 .
as in the proof of theorem [ thm : pointwise - upper ] , we have that @xmath794 uniformly as @xmath795 , where @xmath821 is defined by .
we can therefore fix @xmath788 large enough so that @xmath822 for @xmath46 , let @xmath823 be a bernoulli zero - one random variable with parameter @xmath824 .
let @xmath825 and let @xmath826 denote the indices for which @xmath827 .
we claim that @xmath828 are _ i.i.d . _ with density @xmath829 where @xmath801 is defined in . to see this , first note since @xmath777 implies @xmath830 , we have @xmath831 for all @xmath832 .
thus @xmath833 where @xmath834 .
let @xmath683 and let @xmath835 be measurable .
we have @xmath836 by the construction of @xmath828 , they are independent random variables , hence the claim is established .
let @xmath718 denote the cardinality of @xmath837 . by theorems [ thm :
lower - bound ] and [ thm : upper - bound ] , we have @xmath838 define @xmath839 then @xmath840 .
each @xmath841 is a bernoulli zero - one random variable with parameter @xmath842 it follows that @xmath718 is binomially distributed with parameters @xmath16 and @xmath843 , and hence @xmath844 almost surely . combining this with yields @xmath845 noting that @xmath846 we have @xmath847 recalling we have @xmath848 which completes the proof .
we now have the proof of theorem [ thm : linfty - conv ] .
let @xmath179 .
let @xmath626 and for a multiindex @xmath849 , set @xmath850 . since @xmath57 is uniformly continuous ( by lemma [ lem : holder ] ) we can choose @xmath788 large enough so that @xmath851 .
let @xmath852 be the set of multiindices @xmath393 for which @xmath853^d$ ] .
note that the cardinality of @xmath852 is @xmath854 .
since @xmath852 is finite with cardinality independent of @xmath16 , theorems [ thm : pointwise - upper ] and [ thm : pointwise - lower ] yield @xmath855 let @xmath856^d$ ] .
then there exists @xmath857 such that @xmath858 . by the pareto - monotonicity of @xmath12
and we have @xmath859 by a similar argument , we have @xmath860 and hence @xmath861 combining and we have @xmath862 and hence @xmath863 almost surely .
the desired result now follows immediately from the boundary conditions on @xmath57 proved in theorem [ thm : hjb ] ( i ) , ( ii ) and the fact that there are almost surely no samples in @xmath864 . as a straightforward application of theorem [ thm : linfty - conv ]
, we can show that non - dominated sorting is stable under bounded random perturbations in the samples @xmath0 . for @xmath865 , we set @xmath866 where @xmath709 are _ i.i.d .
_ with a continuous compactly supported density function @xmath867 . for @xmath9
, set @xmath868 [ thm : stability ] let @xmath772 satisfy ( h1 ) and let @xmath78 satisfy ( h2 ) .
there exist constants @xmath869 , depending only on @xmath870 , @xmath2 , and @xmath145 , such that @xmath871 and @xmath872 as @xmath873 .
set @xmath874 .
then @xmath715 are _ i.i.d . _ with density @xmath875 .
set @xmath876 without loss of generality , we may suppose that @xmath586 .
since @xmath877 and @xmath145 has compact support , we can take @xmath468 small enough so that @xmath878^d$ ] .
it is not hard to see that holds for the sequence @xmath879 .
since each @xmath880 is continuous and bounded with compact support in @xmath14^d$ ] , it follows from lemma [ lem : l1-pert ] that @xmath881 uniformly on @xmath1 .
note that @xmath882 for every @xmath16 . since @xmath880 is continuous on @xmath883 and @xmath884 for @xmath885 , ( h1 ) is satisfied for @xmath880 by taking @xmath886
. we can therefore apply theorem [ thm : linfty - conv ] to obtain @xmath887 the proof is completed by setting @xmath888 .
theorem [ thm : linfty - conv ] guarantees that the level sets of @xmath889 will provide good approximations to the pareto fronts for large @xmath16 . in this section
, we present a numerical scheme for computing @xmath57 and show examples comparing the level sets of @xmath890 to the pareto fronts for various density functions . to compute @xmath57 , we need to compute the pareto - monotone viscosity solution of @xmath891 that satisfies @xmath892 on @xmath373 .
let @xmath893 and @xmath894 denote the values of @xmath57 and @xmath2 on a grid with spacing @xmath895 , where @xmath393 is a multi - index .
for a given grid point @xmath393 , the domain of dependence for is @xmath896 .
hence an upwind scheme will use backward difference quotients . now consider substituting backward difference quotients into .
we have @xmath897 we intend to solve the above equation for @xmath893 in terms of @xmath898 .
since we intend to compute the pareto - monotone solution of , we should look for a solution with @xmath899 for all @xmath42 .
consider the mapping @xmath900 where @xmath901 for all @xmath42 .
note that this mapping is strictly increasing for @xmath902 and @xmath903 for @xmath904 .
hence , for any non - negative @xmath905 , there exists a unique @xmath906 satisfying @xmath907 we denote this solution @xmath908 by @xmath909 and define our numerical scheme by @xmath910 with the boundary condition @xmath911 for @xmath393 with @xmath912 .
the numerical solution can be computed by any sweeping pattern respecting the partial order @xmath607 .
this requires visiting each grid point exactly once , and hence has linear complexity .
the value of @xmath909 can be computed numerically by either a binary search and/or newton s method restricted to the interval @xmath913 $ ] . in the case of @xmath99
, we have a closed form expression for @xmath914 @xmath915 note that we have chosen the positive square root to obtain the pareto - monotone solution .
we prove in a subsequent paper @xcite , that the numerical solutions , defined as above , converge to the unique pareto - monotone viscosity solution of as @xmath916 .
+ for @xmath99 , we have @xmath28 , hence the level sets of @xmath57 will approximate the pareto fronts , where @xmath917 .
we now show examples of pareto fronts alongside the level sets of @xmath57 for @xmath0 sampled according to different density functions . in figure
[ fig : uniform ] , we consider a uniform density on a portion of the unit square and show the pareto fronts for @xmath918 , @xmath919 and @xmath54 independent samples alongside the corresponding level sets of @xmath57 . observing the figure , we see that the pareto fronts are well approximated by the level sets of @xmath57 for large @xmath16 .
we also notice that the level sets of @xmath57 appear to yield a consistent underestimate of the pareto fronts .
bollobs and brightwell @xcite showed that the normalized expectation of the longest increasing subsequence among @xmath16 points chosen independently from the uniform distribution on @xmath14^d$ ] is always bounded above by @xmath22 , which is the limit of these normalized expectations as @xmath13 . in light of this result ,
our observation is not surprising and merely confirms the results in @xcite .
we also observe that although the boundary of @xmath71 is smooth , the solution @xmath57 develops shocks , or kinks , which are visible in the level sets of @xmath57
. + in figure [ fig : face ] , we show the same comparison for a multi - modal density function on @xmath14 ^ 2 $ ] .
the density function is depicted by the plot in figure [ fig : face ] and we have the same expected underestimation present here as well .
we thank john duchi for pointing out an error in the proof of theorem [ thm : linfty - conv ] in an earlier version of the manuscript d. kossmann , f. ramsak , s. rost , et al .
shooting stars in the sky : an online algorithm for skyline queries . in _ proceedings of the 28th international conference on very large data bases
_ , pages 275286 , 2002 .
t. sepplinen . .
in _ asymptotics : particles , processes and inverse problems _ , volume 55 of _ ims lecture notes monograph series _ , pages 204233 .
institute for mathematical statistics , beachwood , oh , 2007 . | we show that non - dominated sorting of a sequence @xmath0 of _ i.i.d .
_ random variables in @xmath1 has a continuum limit that corresponds to solving a hamilton - jacobi equation involving the probability density function @xmath2 of @xmath3 .
non - dominated sorting is a fundamental problem in multi - objective optimization , and is equivalent to finding the canonical antichain partition and to problems involving the longest chain among euclidean points . as an application of this result
, we show that non - dominated sorting is asymptotically stable under bounded random perturbations in @xmath0 .
we give a numerical scheme for computing the viscosity solution of this hamilton - jacobi equation and present some numerical simulations for various density functions . |
voids can be thought of as large volumes with very low galaxy density surrounded by the walls and filaments of the cosmic web .
these objects were first detected by gregory & thompson ( 1978 ) , joeever , einasto & tago ( 1978 ) , kirshner et al .
( 1981 ) and geller & huchra ( 1989 ) .
soon after their discovery , white ( 1979 ) , hoffman & shaham ( 1982 ) and peebles ( 1982 ) derived theoretical statistical properties of voids and related them to the gravitational instability scenario .
einasto , einasto & gramann ( 1989 ) also estimated sizes of voids in different samples of galaxies in redshift surveys and compared their results with numerical simulations .
the characteristics of voids in the galaxy distribution , such as the void probability distribution were studied by vogeley et al .
( 1994 ) , ghigna et al .
( 1996 ) , and mller et al .
( 2000 ) , whereas the fraction of the total volume in the universe occupied by voids , which was found to add up to about @xmath12 , was measured by el - ad & piran ( 1997 , 2000 ) , plionis & basilakos ( 2002 ) , and hoyle & vogeley ( 2002 ) .
peebles ( 2001 ) presents a detailed review of the observational works devoted to void statistics .
more recently , theoretical studies on the characteristics of voids and their surrounding structure have concentrated on several different aspects .
for instance , van de weygaert ( 2004 ) presented an excursion set approach ( bond et al .
1991 ) for predicting the distribution function of void sizes and its evolution with redshift .
patiri et al .
( 2004 ) also carry out an excursion set study on voids , and are able to characterize the number density of voids as a function of void radius as well as the mass function of haloes in voids .
their model takes into account the fact that the properties of voids depend on the haloes used to identify them ( gottlber et al . , 2003 , and references therein ) .
based on this result , patiri et al .
provide a framework which allows a comparison between theoretical predictions and statistical measures of voids found in the distribution of the visible galaxies that populate haloes .
also , taking into account the nearly spherical average shape of voids , ryden ( 1995 ) showed that it is possible to use the cosmologically induced redshift - space distortion in the apparent shape of intermediate redshift voids to measure the deceleration parameter , @xmath13 . on the observational side ,
only very recently it has become possible to perform accurate statistical studies of voids using real data .
for instance , croton et al .
( 2004 ) used the 2-degree field galaxy redshift survey ( 2dfgrs ) to measure the reduced void probability function ( rvpf ) , which connects the distribution of voids to the higher moments of galaxy counts in cells .
these new redshift surveys also allow the study of the population of galaxies inside voids .
hoyle et al . ( 2003 ) use the sloan digitized sky survey ( sdss ) to study the luminosity function of void galaxies , which they find by setting an upper limit on the local density contrast at the galaxy position , @xmath14 .
they find that galaxies are fainter in voids , but that the faint - end slope of the luminosity function , is consistent with that of galaxies in higher density environments . also using the sdss , goldberg et al .
( 2004 ) study the mass function of void galaxies selected using the same density threshold as hoyle et al .
( 2003 ) , finding that galaxies in voids are nearly unbiased with respect to the mass . in this paper
we carry out a comprehensive numerical study of the dynamical and spatial properties of galaxies , dark - matter particles and haloes with respect to void centres .
we study how the properties of voids identified from the distribution of dark - matter particles and haloes in a simulation relate to those of voids identified from the galaxy distribution .
goldberg et al .
find that galaxies trace the distribution of mass within voids , a fact that contrasts with the large biases expected in higher density environments , such as walls , filaments and clusters .
since these objects set the boundaries of voids , this study provides a useful insight on the global segregation pattern of dark - matter and galaxies .
the study of the effects of observational biases occurring in flux limited redshift surveys , and an application of the statistics presented here to observational catalogues are carried out in a forthcoming paper , ceccarelli et al .
( 2004a ) .
this paper is organised as follows .
section [ sec : voids ] describes the semi - analytical numerical simulation used in this work and the procedure adopted for identifying voids in the simulation box .
section [ sec : vpecs ] analyses the outflow signatures that characterise the regions surrounding voids . in section
[ sec : correlations ] we study the spatial 2-point cross - correlation functions between voids and different tracers , dark - matter particles , haloes and galaxies .
finally , section [ sec : conc ] summarises our results and presents the main conclusions extracted from this work .
in this work we study void statistics using a @xmath1cdm numerical simulation which contains @xmath15 million dark matter particles and @xmath16 million galaxies from the galform semi - analytic galaxy formation model ( cole et al .
2000 ) , kindly provided by the durham group .
there are approximately @xmath17 dark - matter haloes with at least @xmath18 members , identified from the distribution of datk - matter particles in the simulation using a friends - of - friends ( fof from now on ) algorithm with a linking length @xmath19 .
the resulting haloes are characterised by a minimum mass of @xmath20h@xmath3m@xmath4 , a median mass @xmath21h@xmath3m@xmath4 , and a maximum @xmath22h@xmath3m@xmath4 .
it should be noticed that the fof algorithm does not resolve substructure within the dark - matter haloes .
the semi - analytic model output consists of approximately @xmath23 million galaxies with @xmath0 at @xmath24 ( from now on , we will refer to all galaxies above this magnitude limit , simply as galaxies ) in the simulation .
the @xmath25 , @xmath26 and @xmath27 percentiles of the haloes hosting at least one galaxy are @xmath28h@xmath3m@xmath4 , @xmath29h@xmath3m@xmath4 , and @xmath30h@xmath3m@xmath4 respectively .
we now describe the simulation parameters : the box side measures @xmath31h@xmath3mpc , the matter density parameter corresponds to @xmath32 , the value of the vacuum density parameter is @xmath33 , the hubble constant , @xmath34kms@xmath3mpc@xmath3 , with @xmath35 , and the primordial power spectrum slope is @xmath36 .
the present day amplitude of fluctuations in spheres of @xmath37h@xmath3mpc is set to @xmath38 .
this particular cosmology is in line with recent cosmic microwave background anisotropy ( wmap , spergel et al .
2003 ) and large scale structure ( 2dfgrs , percival et al .
2004 ) measurements .
we use this simulation to study the dynamical and spatial characteristics of galaxies , dark - matter haloes and dark - matter particles around voids . throughout this paper ,
errors in the statistics are calculated using the jackknife method , which has been shown to provide equivalent errors to those found by applying the statistics to a large number of independent simulations and measuring a variance ( see for instance , croton et al .
2004 , padilla & baugh , 2003 ) .
these results indicate that jackknife errors provide a reasonable estimate of statistical uncertainties and cosmic variance .
the use of semi - analytic numerical simulations allows a comparison of galaxy- , halo- and mass - defined voids , since the population of galaxies in the simulation box is physically motivated , subject to evolution via mergers , metal enrichment , dust evolution , and other important astrophysical processes ( for full details see cole et al .
2000 ) . throughout this paper
we compare results from real- and redshift - space derived quantities .
when referring to voids this indicates that , when identifying voids in the simulation , the positions of particles , haloes or galaxies were considered either in real- or redshift - space .
we implement redshift - space positioning in this simulation by displacing the @xmath39coordinates of particles and galaxies by the @xmath39component of their velocities in units of h@xmath3mpc ( distant observer looking down the @xmath39axis ) . in the next subsections
we describe the void finding algorithm , present the resulting number density of voids as a function of void radius , and calculate the void auto - correlation function .
= 8.truecm = 11.truecm we find voids in the simulation by identifying spherical volumes where the overall density contrast satisfies @xmath40 , which corresponds to a usual definition of an under - dense region .
later in this section we find that small variations on this condition do not affect significantly our statistics .
we use the mass , dark - matter haloes , and the galaxies in the simulation to find voids , and we also investigate in a simple way the effects of redshift space distortions in the identification of voids .
therefore , we produce ten different catalogues of voids in the numerical simulation .
we first have @xmath41 sets of voids , found in the distribution of dark - matter ( which we will refer to as mass - defined voids ) , haloes with masses above @xmath42 , @xmath43 and @xmath44h@xmath3m@xmath45 ( halo - defined voids , also referred to as h@xmath46- , h@xmath47- and h@xmath48-defined voids respectively ) , h11.5 by @xmath49 , and h12 haloes by @xmath50 , and voids found in the distribution of galaxies ( referred to as galaxy - defined voids ) in real space . repeating the identification in redshift - space provides us with the remaining @xmath41 samples . for each of these different samples of voids , the procedure followed is exactly the same , and comprises the following steps : * we produce a large number of random positions which will be used to test whether the density contrast in spheres centred on these positions satisfies the condition @xmath51
the higher the number of random positions the higher the accuracy in finding the true centre of a void in the simulation . *
we calculate the density contrast as a function of radius of a sphere centred on the random positions generated in the first step .
* we identify the radii for which the void density criterion @xmath51 is satisfied . in the event where this condition is satisfied more than once at different radii
, we only consider the largest radius . * up to this point , the outcome of the identification process will likely consist of a large number of underdense overlaping spheres .
therefore , the last step consists of removing all spheres overlapping with a larger void .
for instance , this last step will make a large underdense spherical region of radius @xmath52 be represented by a single sphere of radius @xmath52 , instead of by many spheres of smaller radius that could be able to fill the underdense volume .
the nature of the finding algorithm makes voids of comparable sizes at similar positions be simply replaced by a larger void which naturally occupies the volume of all the smaller voids together .
we have made several checks to make sure the void finding algorithm is reliable .
we have varied the upper limit in @xmath53 up to @xmath54 , and find that in general results are not significantly affected in the case of galaxy- and halo - defined voids .
the main difference resides in the appearance of large spurious ( @xmath55h@xmath3mpc ) voids .
we have also checked that the removal of overlapping voids does not affect significantly the mean values of the statistics we investigate in this work by allowing for instance , voids within voids .
it is important , however , to take this effect into account when calculating errors , since many slightly de - centred voids representing only one physical void in the simulation artificially lowers the size of statistical errors . given that dark - matter particles outnumber galaxies by a factor @xmath56 in the simulation , we tested for the impact of the number of particles in the identification procedure . in order to do this
we select @xmath57th of the total dark - matter particles and run our void identification algorithm .
the resulting void catalogue is almost identical to the one found from the full simulation .
figure [ fig : pos ] shows a slice of the simulation box with the distribution of semi - analytic galaxies and the positions of voids found by our procedure . in order to improve clarity , in this particular plot we only show voids with radii in the range @xmath58h@xmath3mpc . as can be seen , the void identification algorithm is quite satisfactory , picking out low density regions with void radii compatible with the observed galaxy distribution .
the empty spaces not filled with voids are expected for instance from voids with @xmath59h@xmath3mpc , or voids with centres outside the slice shown in this figure .
the first statistic we measure using the voids identified in the previous subsection is the number density as a function of void radius , @xmath60 .
we show the measured results from the simulation box in the upper panel of figure [ fig : numdens ] , where the red ( gray ) lines correspond to voids found from the spatial distribution of galaxies and the blue ( black ) lines from the mass .
solid lines correspond to real - space , dashed lines to redshift - space .
symbols show the results from the distribution of voids identified from haloes of different masses as indicated in the key .
it is noticeable the good agreement between these estimates for @xmath61h@xmath3mpc , where all the void number densities are virtually indistinguishable . at larger
object / void - centre separations the different estimates start to diverge
. for instance the difference between real- and redshift - space estimates can diverge by as much as a factor of @xmath62 , as is the case at large mass - void radii @xmath63h@xmath3mpc .
this is attributable to the fingers of god effect which takes place in the void walls and displaces dark - matter particles from the walls and dilutes the void boundaries .
galaxy- and halo - void number densities are not significantly affected by a redshift - space identification of voids as expected , since the peculiar velocities of mass and galaxies are small ( @xmath64 ) compared to the void sizes .
the difference between number densities of voids found from the distribution of galaxies and mass starts to be noticeable at separations @xmath65h@xmath3mpc ; eventually there are no longer mass - defined voids with @xmath63h@xmath3mpc whereas it is still possible to identify galaxy - defined voids with @xmath66h@xmath3mpc .
the number density of halo - defined voids is more comparable to that of galaxy - defined voids .
the increase in the number density of voids with the mass of the haloes used to identify them can be understood in terms of the bias factor between haloes and the mass .
higher mass haloes are more strongly biased with respect to the mass , and therefore are preferably found in high mass density regions , which correspond preferably to the centres of void walls and filaments .
the net effect this would have is to increase the radius of the void when identified from a biased population of objects , and would explain the apparent increase in number density observed in figure [ fig : numdens ] . from the number density of voids , it is straight - forward to obtain the fraction of volume inside voids in the simulation .
the lower panel of fig .
[ fig : numdens ] , shows the volume inside voids with @xmath67 .
as can be seen , the difference between mass- and galaxy - defined voids is even more evident here . also , it is now possible to recognise a good agreement between galaxy - defined voids and h@xmath47- for @xmath68h@xmath3mpc and h@xmath48-defined with @xmath69h@xmath3mpc voids .
this demonstrates that caution should be taken when analysing properties of galaxy - defined voids , which can differ from those of halo- and mass - defined voids . as mentioned above
, the reason for this discrepancy is the slightly biased nature of the galaxy distribution in this simulation with respect to the dark - matter particles and haloes . in particular , the spatial distribution of galaxies shows larger and more numerous voids than the distributions of mass or low mass haloes ( less biased populations ) .
given that the redshift - space effect enhances the number density of mass - defined voids more strongly , the differences between results from mass- and galaxy - defined voids are less important in redshift - space .
the volume fraction inside voids can be as large as @xmath70 , a value which is slightly lower than estimates from redshift surveys of @xmath12 by el - ad & piran , 1997 , 2000 , plionis & basilakos , 2002 , and hoyle & vogeley , 2002 . as pointed out above
, our estimate of the volume fraction is only an order of magnitude estimate at best , since our simulated volume is limited and we might be badly affected by sampling statistics in the large - volume end of the void population .
( 230,440 ) ( 0,0 ) ( -7,165 ) a simple statistics that can be applied to the void positions is the 2-point auto - correlation function . the advantages in studying the correlation function relies in its widespread use and in its direct connection to the power spectrum of density fluctuations .
the upper panel of figure [ fig : xivoids ] shows the auto - correlation function of galaxy- ( symbols ) and h@xmath48- and h@xmath47-defined voids ( dashed , @xmath71h@xmath3m@xmath45 , solid , @xmath2h@xmath3m@xmath45 ) for different void radii ( gray scale and different symbols as indicated in the figure caption ) . ,
@xmath72 , and @xmath73 , for the three ranges of void radii shown in figure [ fig : xivoids ] ( smaller to larger void radii , respectively ) , which explains the noise present in the correlation functions of voids of large radius .
] as can be seen , larger voids are more strongly clustered , and show a positive correlation signal out to larger separations .
the amplitude of these correlations , however , is quite low and we only find a significant signal out to scales @xmath74h@xmath3mpc .
it is interesting to note that at scales larger than @xmath75h@xmath3mpc , the slopes of the correlation functions for voids of different radii are similar .
the auto - correlation function of voids defined using the most massive haloes ( h@xmath48 , dashed lines ) provides the best match to the clustering of galaxy - defined voids . in order to assess whether the correlation function responds to an effect of the void identification procedure , we show in the lower panel of figure [ fig : xivoids ] the average correlation function ( and @xmath76 confidence region ) of @xmath77 sets of randomly placed spheres ( random - voids ) .
each of these sets has the same number density as a function of radius than galaxy - defined voids , and also follows the exclusion constraints set for the identification of galaxy - defined voids .
also , random - voids are placed in a volume with the same size as the simulation box .
as can be seen , the exclusion constraint alone can produce a non - zero correlation function consistent with a power law of index @xmath78 .
the comparison with the galaxy - defined voids correlation function , however , shows that the amplitude of the galaxy - void correlation function is still significantly higher than that of the random - voids .
this could be interpreted as evidence of the influence of hierarchical clustering in the spatial distribution of galaxy - defined voids .
we also calculated the correlation function of sets of random - voids with a different number density as a function of void radius , in particular , similar to that of h@xmath46-defined voids .
we find that the comparison between h@xmath46-defined voids and this new set of random - voids is identical to the comparison we carry out above , namely that the correlation function of h@xmath46-defined voids is higher in amplitude than that of random - voids for small separations .
.5 cm our preliminary conclusion from the study of the number density of voids and void auto - correlation functions in the numerical simulation is that galaxy - defined voids show good agreement with voids identified from moderate to high mass haloes ( @xmath2h@xmath3m@xmath45 ) . from now on ,
we will mostly compare results from these two sets of voids , unless we state otherwise .
( 430,250 ) ( 0,-20 ) ( 215,-20 ) ( 430,250 ) ( 0,-20 ) ( 215,-20 ) ( 430,270 ) ( 0,-10 ) ( 215,-10 )
just as it has recently become possible to study the peculiar velocity field around groups ( ceccarelli et al .
2004b , c05 from now on ) , which is characterised by infall galaxy motions toward the group centres , it is also possible to study outflows of material escaping the void volumes .
a further motivation to study the outflow motions is that as has been shown by c05 , the infall motions of mass and galaxies are almost identical , being the galaxy infalls only slightly higher than the mass infalls .
it may be possible then , to obtain a reliable measurement of the bulk outflows of mass emptying the void volumes by studying the velocities of galaxies around voids .
we set out to compare the outflows of mass , dark - matter haloes , and galaxies from the simulated void catalogues constructed in this work . in order to do this
we simply project the peculiar velocity of galaxies and dark - matter haloes and particles onto the radial direction measured towards the void centres . adding the contribution from all the individual particles , haloes or galaxies together , and averaging the result
, we obtain a dependence of the outflow motions as a function of distance to the void , @xmath79 , and void radius , @xmath60 .
defined in this way , a positive velocity indicates an infall to - wards the void centre , and a negative velocity , an outflow .
figure [ fig : outflow2d ] shows the 2-dimensional diagrams of outflow velocities as a function of void size ( x - axis ) and distance ( y - axis ) .
the lines show constant outflow velocity levels ; solid lines correspond to positive velocities ( infalls ) , and dashed and dotted lines show negative velocities or outflows ( see the figure caption for velocity level values ) .
lower panels in this figure show results for halo velocities around h@xmath47-defined voids .
upper panels show results for galaxy velocities around galaxy - defined voids . as can be seen from the comparison of the iso - velocity contours in the upper and lower panels , the velocities around halo - defined voids
reach higher values of both , infall and outflow motions , than around galaxy - defined voids by approximately @xmath80kms@xmath3 ( see also figure [ fig : outflowd ] ) . however
, if we concentrate on the velocity minimum for large @xmath60 voids , we see that there is an agreement between results from haloes and galaxy - defined voids .
also , small voids , @xmath81h@xmath3mpc , show infall motions at separations @xmath82h@xmath3mpc .
this indicates the existence of overdense regions surrounding small voids , which would be responsible for producing the observed infall motions .
we provide a fit for the distance at which the outflow motions reach @xmath83kms@xmath3 , @xmath84 in the case of h@xmath47 velocities and h@xmath47-defined voids we find @xmath85h@xmath3mpc and @xmath86 , and for galaxy - defined voids and galaxies , @xmath87h@xmath3mpc and @xmath88 .
( for values of the parameters @xmath89 and @xmath90 for the other simulated samples , and in redshift - space , see table [ table : fits ] ) . at larger distances from the void centre ,
velocities become positive , indicating an opposite infall motion . the distance from the void centre where the minimum in outflow velocity occurs
, @xmath91 , is very well approximated by @xmath92 motivated by the results for @xmath93 , we can also fit eq .
[ eq : vfit ] to @xmath91 .
this equation could prove to be more accurate for larger voids . in the case of h@xmath47 velocities in real - space ,
the best fit parameters are @xmath94h@xmath3mpc and @xmath86 . for galaxy - defined voids and galaxies , @xmath95h@xmath3mpc and @xmath88 .
the best fit parameters for all the samples studied here can be found in table [ table : fits ] .
the fits corresponding to @xmath91 and @xmath93 as a function of the void radius are shown in all panels of figure [ fig : outflow2d ] in thick solid gray lines ( lower and upper lines respectively ) .
the fit parameters corresponding to different halo masses correlate reasonably well with halo mass , and can therefore be used to interpolate for samples of masses different than the ones used in this paper .
alternatively , this can also be used when analysing a sample of galaxies for which the masses of haloes in which they reside are known .
this is also applicable to other fits provided in this work ( equations [ eq : vmin ] and [ eq : rho ] ) .
( 230,240 ) ( 0,-10 ) ( 230,330 ) ( 0,-10 ) ccccc statistic & real / redshift - space & gx./mass & @xmath89hmpc@xmath3 & @xmath90 + @xmath91 & real & mass & 2.30 & 0.80 + & & h11 & 2.50 & 0.60 + & & h11.5 & 2.50 & 0.60 + & & h12 & 2.50 & 0.60 + & & gx . & 3.60 & 0.70 + & redshift & mass & 3.00 & 0.80 + & & h11 & 2.80 & 0.60 + & & h11.5 & 3.00 & 0.60 + & & h12 & 3.00 & 0.50 + & & gx . & 3.50 & 0.70 + @xmath93 & real & mass & 0.10 & 0.38 + & & h11 & 1.00 & 0.65 + & & h11.5 & 1.20 & 0.65 + & & h12 & 1.70 & 0.65 + & & gx . & 1.00 & 0.55 + & redshift & mass & 0.10 & 0.37 + & & h11 & 1.00 & 0.65 + & & h11.5 & 1.30 & 0.65 + & & h12 & 1.70 & 0.65 + & & gx . &
0.95 & 0.55 + [ table : fits ] we now focus on the dependence of the outflow velocities on void radius at fixed distances from the void centres .
we show this dependence in figure [ fig : outflowr ] where the left panels show results in real - space and right panels in redshift - space .
upper panels correspond to galaxies and galaxy - defined voids , and lower panels to h@xmath47 haloes and h@xmath47-defined voids .
the different line types show the outflow velocities at different distances from the void centres ( the key in the figure specifies their values ) .
the outflow velocities at different distances show a complicated variation as we go from regions near the void centres to large distances .
for instance , it is easy to find mild infall motions @xmath96kms@xmath3 near the centres of moderate sized voids with @xmath97h@xmath3mpc
. also , the smaller the distance to the void centre , the smaller the void at which an outflow motion ( negative velocities ) can be found .
the trend of velocities around voids with increasing void radius is negative for @xmath98h@xmath3mpc .
a much simpler dependence characterizes the outflow velocity as a function of distance for different void sizes , as can be seen in figure [ fig : outflowd ] . in this figure ,
the left- and right - hand panels show results in real- and redshift - space respectively ; upper panels show results from the galaxy distribution , lower panels from the h@xmath47 distribution .
the different line types in these plots correspond to different void radii . in general
, it can be seen that larger voids reach a minimum in velocity ( or a maximum outflow ) at larger separations .
the outflow is also stronger for larger voids . at larger distances from void centres ,
velocities tend to @xmath99 as expected given the lack of correlation of structure at these scales .
still , it should be noted that the outflow signal continues to be present at distances larger than the void size .
figure [ fig : outflowd ] also shows that smaller voids ( @xmath59h@xmath3mpc ) show a rapid transition in the very nearby regions , going from outflow to infall motions , revealing that these may be immersed in the middle of high density regions responsible for the infalls .
note however , that the amplitude of such infall motions is small compared to the infall motions around groups of galaxies ( c05 ) .
another contribution to this infall could originate in outflows from neighbouring voids .
we can also find out about the differences arising from identifying voids in the distribution of galaxies and dark - matter haloes in figure [ fig : outflowd ] . in this visualization
, the galaxy - defined voids also tend to show slightly stronger outflows than the h@xmath47-defined voids , and small galaxy - defined voids show a smaller degree of infall than their h@xmath47 counterparts .
the effects of redshift - space selection simply tend to decrease the range of velocities measured around voids , that is , infall and outflow extremes are lower in absolute value , both for galaxy and h@xmath47-defined voids .
these effects must be taken into account when comparying observational data on the dynamics of galaxies around voids with the corresponding theoretical predictions .
( 200,400 ) ( 0,-20 ) ( 0,190 ) cc gx./halo , mass & @xmath100km@xmath3sh@xmath3mpc + mass & -22.1 + h11 & -16.2 + h11.5 & -14.5 + h12 & -13.2 + gx . & -14.5 + [ table : vmin ] figure [ fig : outflowmin ] shows the dependence of the maximum outflow velocity with void radius .
as can be seen , the dependence is well approximated by a linear relation with the void radius .
the best fit parameters of the relation @xmath101 can be found in table [ table : vmin ] for the different choices shown in the figure .
it can be seen that mass - defined voids show a much lower and steeper maximum velocity relation when compared to galaxies , which show a difference of @xmath102kms@xmath3 at @xmath103h@xmath3mpc with respect to the results from mass - defined voids . in this case , it is not possible to decide which sample of halo - defined voids provides the closest outflow velocities to the galaxies , since the data from each halo sample are consistent with the galaxy - void data ( which shows considerable scatter around the best fit line ) .
still , the results from h@xmath47-defined voids are characterised by the same best fit @xmath104kms@xmath3hmpc@xmath3 than the galaxy - void data .
we have also investigated the effect of using redshift - space positioning and found that the linear relation between @xmath105 and void radius steepens considerably @xmath106kms@xmath3hmpc@xmath3 , independently of the sample of voids used .
finally , we also studied the possible differences in the velocity field in the directions parallel and perpendicular to the void walls . in order to do this
we calculated the 1-d velocity dispersion of galaxies in both directions ( @xmath107 and @xmath108 , respectively ) , removing the local outflow velocity when calculating the radial velocity dispersion .
our findings are summarised in the upper panel of figure [ fig : wallv ] , where the black lines show @xmath107 and the gray lines show @xmath108 .
different line types correspond to different void radii as indicated in the key , and we only show results for voids identified using galaxies in redshift - space since there is little variation when using haloes or real - space positions .
as can be seen , the velocities aligned with the radial direction from the void centre are about @xmath109 lower than the velocities on the direction of the void walls , regardless of void size and distance to void centre , indicating that galaxies have a slight but systematic ( @xmath110 to @xmath111 level ) tendency to follow trajectories along the void surfaces .
the inset in the upper panel shows the dependence of the velocity dispersions in the directions parallel and perpendicular to the line of sight as a function of the distance to the void centre normalised by the void radius .
as can be seen , at @xmath112 , the curves corresponding to different void radii converge . on the other hand ,
smaller voids show much higher velocity dispersions at @xmath113 when compared to larger voids .
this is in agreement with the indication that small voids are embedded in higher density regions , which could be respondible for the high velocity dispersions in the small voids inner regions .
the lower panel of this figure shows the ratio between the velocity dispersions in the directions parallel and perpendicular to the void walls , and as can be seen there is little dependence on the distance to the void centre and on the void radius , being consistent with a roughly constant @xmath114 .
.5 cm the observational study of the peculiar velocity field around voids is carried out in a forthcoming paper , ceccarelli et al .
( 2004 ) , where the different observational biases present in peculiar velocity data are analysed , and the results are compared to the theoretical expectations presented here .
we now focus our attention on the spatial statistics of galaxies , haloes and mass around voids .
we use the 2-point cross correlation function between void centres , and galaxies , haloes or dark - matter particles to characterise the spatial clustering properties of the vicinities of voids . in the following subsections we study the 1-dimensional cross - correlation function in real- and redshift - space , and the cross - correlation function as a function of the coordinates perpendicular and parallel to the line of sight . in this subsection
we study the spatial correlations between the void centres and the objects surrounding them .
we will compare once more the resulting statistics from considering voids in the distribution of galaxies , haloes and mass , using real- and redshift - space positions .
we start by analysing our expectations for the outcome of a void - object cross - correlation function , which we compute using the following estimator , @xmath115 where @xmath116 corresponds to the number of center - object pairs , and @xmath117 is the number of center - random pairs which would be measured if the positions of objects were uniformly distributed in space . the latter is calculated using the mean density of objects in the numerical simulation box .
what we expect to find at small separations is an anti - correlation , lower in amplitude than the maximum overdensity set for identifying voids , that is , @xmath118 . on the other hand , at sufficiently large separations , @xmath119 , we would expect to find departures from this value and to gradually tend towards a small but positive value for the correlation function , indicating a population of galaxies at the void walls .
we also expect @xmath120 at sufficiently large separations from the void centres .
figure [ fig : xis ] shows the cross - correlation functions between galaxy - defined voids and galaxies ( upper panel ) , and galaxy- , halo- and mass - defined voids with their respective counterparts , in real space ( lower panel ) . solid and dashed lines in the upper panel represent the results in real- and redshift - space respectively . concentrating on the results for galaxy - defined voids
, it can be seen that we obtain the expected value @xmath118 at low separations for void radii @xmath121h@xmath3mpc in real- and redshift - space ( smaller voids may show this effect at lower separations than those analysed in this work ) , and also that at the void radius , @xmath122 ( open squares ) .
this relation holds regardless of the tracer used to identify voids . the general effect produced by changing from real- to redshift - space data , where the voids are identified in redshift - space and the tracers
are also in redshift - space , is that of lowering very slightly the maximum clustering amplitude .
another important feature of the correlation functions shown in the upper panel of figure [ fig : xis ] is that at separations , @xmath119 , @xmath11 increases rapidly in value reaching @xmath123 , but always lower than the galaxy auto - correlation function .
it is also noticeable the fact that the correlation function around voids can be positive outside the void radius , particularly for the smaller voids , @xmath124h@xmath3mpc , which can be interpreted as small voids being embedded in extended , reasonably high density regions , in accordance with the results from the velocity field around small voids , which shows systematic infalls instead of outflows . note that this peak in @xmath11 is still quite low compared to the galaxy correlation function . from the comparison between results for different void populations in the lower panel of figure [ fig : xis ] , we see that mass - defined voids show a higher cross - correlation function than haloes and galaxies at @xmath125
. haloes , on the other hand , show the same correlation levels than galaxies at small separations from the void centres . at larger separations , @xmath126 , individual dark - matter particles
are less clustered than galaxies .
haloes show a range of clustering amplitudes which correlates with the halo mass .
in particular , h@xmath47 show consistent correlation amplitudes than galaxies , and h@xmath48 show higher @xmath11s .
it is interesting to note though , that independently of the object used to identify voids , the cross - correlation function shows a similar qualitative shape , starting off at a constant negative value at low separations , independent of void radius , and increasing at separations of the order of the void radii , approaching @xmath127 . note that this change from negative to @xmath128 is sharper for haloes and galaxies than for mass - defined voids .
for instance , for @xmath129h@xmath3mpc , the galaxy - void - galaxy cross - correlation function values are as high as the galaxy @xmath11 . given the clear correlation between the void - size and the distance at which the correlation amplitude departs from its constant negative value
, we have also calculated the density profiles around voids as a function of the reduced distance to the void centres , @xmath130 ( figure [ fig : rho ] ) .
as can be seen , density profiles for voids with @xmath129h@xmath3mpc are marginally consistent with one another ( @xmath131 ) .
( 210,570 ) ( 0,-10 ) ( 0,180 ) ( 0,360 ) the comparison between the density profiles of different void populations , shown in the middle panel of figure [ fig : rho ] , is similar to what was found for cross - correlation functions .
namely , we find that the density profiles of mass - defined voids are lower in amplitude than galaxy - defined voids .
the latter are found to be of similar amplitudes than the density profiles of halo - defined voids , in agreement with the numerical simulation results by gotlber et al . 2003 . on large separations
@xmath132 , small voids with @xmath133h@xmath3mpc show density profiles much higher than larger voids with @xmath129h@xmath3mpc .
the differences in @xmath134 for the different types of void can be alleviated significantly by dividing the density profile by the correlation function of the objects used to identify the voids ( see the lower panel of figure [ fig : rho ] ) . in this case , the density profiles divided by the correlation function of galaxies are in better agreement with one another for a wider range of void sizes , @xmath135h@xmath3mpc .
in addition , this ratii can be fitted by the empirical function , @xmath136 .
\label{eq : rho}\ ] ] the parameter @xmath137 is related to the density contrast threshold used for defining the voids ( @xmath40 in our case ) , and to whether galaxies , haloes or mass are used as the tracers of the density field .
@xmath138 depends strongly on the tracer and on the maximum density contrast used to identify the voids .
the parameter @xmath139 indicates how rapid is the change from constant negative density contrast to the mean density of the universe and also depends strongly on the tracer .
table [ table : rho ] summarises the parameter values corresponding to the different choices we show in the figure .
the use of different void - finding algorithms will also change the values of the quoted best - fit parameters .
for instance , the use of the distance to the closest object ( as in gotlber et al .
2003 ) , will lower the value of @xmath138 to compensate for the typically smaller radii found by this void - finding method .
the thick dot - dashed line in the inset shows again the fit in the main panel , this time divided by the galaxy correlation function ; the @xmath140 per degree of freedom is always lower than @xmath141 for @xmath121h@xmath3mpc in this case . as a final check
, we have corroborated that this level of coincidence between density profiles divided by correlation functions also holds for h@xmath46- , h@xmath47- , h@xmath48- , and mass voids . in sum
, we find that voids with @xmath129h@xmath3mpc , show a nearly universal radial profile for the ratio involving the density profile of objects and the 2-point auto - correlation function , @xmath142 .
this is true for voids identified from both , galaxies and dark - matter haloes .
the best fit parameters for the density profiles are presented in table [ table : rho ] . since voids of any size ( within the limits imposed by the numerical simulation available for
this work , @xmath143h@xmath3mpc@xmath144 ) show similar radial @xmath142 profiles .
the parameters in the table also describe this profile provided the fit is divided by the correlation function of the tracers used when constructing the sample of voids .
ccccc real / redshift - space&gx./mass & @xmath137 & @xmath138 & c + real & mass & 0.90 & 0.60 & 5.0 + & h11 & 0.96 & 0.80 & 5.5 + & h11.5 & 0.97 & 0.85 & 7.0 + & h12 & 0.98 & 0.89 & 8.0 + & gx . & 0.96 & 0.80 & 6.0 + redshift & mass & 0.90 & 0.65 & 5.0 + & h11 & 0.95 & 0.85 & 5.5 + & h11.5 & 0.96 & 0.90 & 6.0 + & h12 & 0.98 & 0.93 & 7.0 + & gx . & 0.95 & 0.80 & 5.0 + [ table : rho ] we study the galaxy - void - galaxy and halo - void - halo cross - correlation functions , @xmath10 , as a function of the coordinates parallel ( @xmath145 ) and perpendicular ( @xmath146 ) to the line of sight , which we take to be the @xmath39axis of the numerical simulation . in order to simulate the redshift - space distortions
we again displace the positions of dark - matter haloes and galaxies by the projection onto the @xmath147-axis of their peculiar velocities and measure @xmath10 . in order to visualize this 2-dimensional distorted function , we study the the contours of constant correlation amplitude in the @xmath148 plane .
we remind the reader that the iso - correlation contours for a @xmath10 measured using real - space positions would correspond to circles , since there are no preferred directions in the simulation and the correlation function must therefore be isotropic .
however , as we use redshift - space positions , we can learn about the dynamical properties of galaxies , dark - matter haloes and individual particles around voids from the departure of contour shapes from perfect circles .
such departures are commonly known as redshift - space distortions .
redshift - space distortions have been extensively studied for all kinds of objects , including quasars ( hoyle et al .
2002 ) , galaxies ( loveday et al . 1995 , rattclife et al . 1998 ) , groups ( padilla et al . 2001 ) , and clusters of galaxies ( bahcall & cen 1992 , padilla & lambas 2003a and 2003b ) . also , assuming that redshift - space distortions around voids are not important , ryden ( 1995 ) proposed using the cosmological distortions around voids for measuring @xmath13
. at small separations , the galaxy correlation function , @xmath10 , shows elongations in the direction of the line of sight due to their random motions in the interiors of clusters of galaxies ( loveday et al .
1995 ) . at larger separations , it is possible to see infall motions toward large mass concentrations ( hawkins et al .
the effect of infall motions was first measured using galaxy groups in the uzc by padilla et al .
the advantage of using groups relies in the fact that they only show infall signatures in their distortion patterns , which induces a flattening of the correlation function contours in the direction of the line of sight .
( 390,600 ) ( 0,-50 ) ( 195,-50 ) ( 0,145 ) ( 195,145 ) ( 0,340 ) ( 195,340 ) cll statistic & real or redshift - space & gx . , haloes and mass + number density & redshift space is higher & the number density of galaxy - defined voids is + & by up to factor @xmath149 at & higher than that of mass - defined voids by as + & large void radius .
consistent & much as two orders of magnitude : consistent at + & at @xmath124h@xmath3mpc for mass- & @xmath124h@xmath3mpc , but + & voids . halo- and
galaxy - defined voids show & no mass - haloes at @xmath150h@xmath3mpc .
+ & smaller changes from real- to redshift- & h@xmath47- and h@xmath48 haloes in agreement + & space .
& with galaxy - defined voids .
+ & & + void @xmath11 & & galaxy - defined voids show stronger + & & auto - correlation functions than + & & halo - void @xmath11s .
+ & & shape of void @xmath11 for voids + & & with different radii are similar , + & & and amplitude increases with @xmath60 . + & & + outflow velocity & slower outflows and infalls & galaxy - defined voids show similar + & in redshift - space ( by & infall and outflow motions + & about @xmath151kms@xmath3 . & to h@xmath47-defined voids ( differences @xmath152kms@xmath3 ) + & & + 1d @xmath11 & increase in @xmath11 at @xmath153 & galaxy @xmath11 is quite higher + & is milder when voids and tracers & than mass @xmath11 outside the void + & are in redshift - space . & radius . inside the void radius both + & & @xmath11s are almost indistinguishable .
good + & & agreement between galaxy- and h@xmath46-defined voids .
+ & & + 2d @xmath11 & real - space shows no anisotropies & galaxy - defined voids show larger elongations + & as expected .
redshift - space & in @xmath10 than h@xmath47-defined voids .
+ & elongations extending out to & qualitatively , the shapes of contours are + & @xmath154 . &
similar for both types of voids .
+ [ table : summary ] voids are the opposite of galaxy groups in terms of their surrounding peculiar velocity field , so we expect to find an elongation of the correlation function contours along the line of sight .
this is expected to affect not only small separations in the @xmath145 direction , as is the case with the fingers of god signatures in the galaxy @xmath10 , but also larger @xmath145 separations .
figure [ fig : xisp ] shows the correlation function contours corresponding to h@xmath47- and galaxy - defined voids ( left- and right - hand panels respectively ) as a function of normalised separation in the directions parallel and perpendicular to the line of sight ( in our case , the line of sight coincides with the z - axis in the simulation box ) .
the void radius increases from bottom to top panels , with values in the ranges shown in the key on each panel .
the different line types represent different correlation function amplitudes ; solid lines correspond to @xmath155 and dotted lines to @xmath156 . in order to facilitate the interpretation of this figure , and the comparison to results from the auto - correlation function of galaxies ,
the reader is referred to figure 4 in hawkins et al .
( 2003 ) , which shows estimates of @xmath10 for different galaxy types in the 2dfgrs .
as expected , the void - halo cross - correlations show elongated iso - correlation contours along the line of sight , specially for the @xmath157 contours .
this can also be found in the void - galaxy cross correlations , in particular for the lower correlation function contours , @xmath158 .
also , as can be seen in the figure , the contour patterns contain an unexpected wealth of information which we analyse in detail in what follows . *
the iso-@xmath11 contours around voids differ from the elongations seen in galaxy auto - correlation functions since here the elongated patterns extend out to separations @xmath159 , perpendicular to the line of sight ( ranging from @xmath160 to @xmath161h@xmath3mpc from bottom to top panels ) . on the other hand , the galaxy auto - correlation function is significantly distorted only out to @xmath162h@xmath3mpc ( hawkins et al .
* as expected , galaxy- and h@xmath47-void @xmath10 show a positive correlation signal in agreement with the results from the 1-dimensional correlation functions from the previous section .
the patterns corresponding to positive correlations show the finger - of - god effect taking place at the void boundaries . *
\(iii ) the flattening of the patterns beyond this radius is originated in random motions taking place in the void walls , which correspond to a radial velocity dispersion @xmath163kms@xmath3 ( see figure [ fig : wallv ] ) .
* there are usually two @xmath127 contours in the void @xmath10 diagrams .
the first one ( closer to the origin ) corresponds to the increase in density at the void walls .
the second @xmath127 contour ( further away from the origin ) is determined by the end of large scale fluctuations .
* h@xmath47 haloes show an outflow motion of lower amplitude than that of galaxies .
however , since halos have a larger velocity dispersion in the void edges , they appear with a smaller radial elongation than galaxies . the elongations along the line of sight seen in the 2-dimensional correlation function of voids can be used to infer the peculiar velocity field around voids without the need of measuring peculiar velocities .
the latter is a difficult task due to the large errors involved in the measurement of distances by means independent of redshift .
we tackle this problem in a forthcoming paper , ceccarelli et al .
( 2005a ) .
in this work we studied the statistical properties of voids selected from the distribution of mass , haloes and @xmath0 galaxies in a @xmath1cdm numerical simulation populated with galaxies using a galform semi - analytic galaxy formation model ( cole et al .
the aim of this work was to understand the systematic biases between these different void populations ( patiri et al .
2004 , gottlber et al .
2003 ) .
we find that the number density of galaxy - defined voids departs from that of mass - defined voids at @xmath164h@xmath3mpc .
however , the fraction of the volume of the simulation box in galaxy - defined voids and voids identified from haloes with @xmath2h@xmath165 m@xmath45 ( h@xmath47-defined voids ) are compatible with one another .
the effect of identifying voids in redshift space increases the number of voids in the simulation , but this effect is almost negligible for galaxy- and halo - defined voids .
the fraction of total volume occupied by galaxy voids in the simulation reaches the @xmath70 level when considering voids with @xmath166h@xmath3mpc .
this fraction is considerably smaller than results obtained from redshift surveys by hoyle & vogeley ( 2002 ) , who find that about @xmath12 of the volume in the universe is in larger voids . in order to properly assess whether this disagreement is significant ,
it must be taken into account that the volume fraction in voids is very sensitive to the abundance of large voids in the sample analysed .
therefore , it would be necessary to analyse larger simulations than the one used in this work to be able to make a more reliable comparison .
another factor that influences the volume fraction in voids is the density contrast used to identify voids ; a higher density contrast increases slightly the fraction of volume .
in addition , the identification of voids in redshift space also produces this effect , bringing the results from our simulation even closer to observational estimates .
the study of the auto - correlation function of voids shows that voids identified from more massive haloes show a higher auto - correlation function .
in particular , h@xmath48-defined voids show a correlation function which is in agreement with that of galaxy - defined voids .
still , the clustering of voids is very weak , with amplitudes almost always lower than @xmath167 .
we found that the shape of the auto - correlation function of voids is consistent with being independent of void radius .
the amplitude , however , is seen to increase with void size . by using 100 sets of random catalogues of voids which satisfy the same exclusion constraints as galaxy- and halo - defined voids ,
we demonstrate that the void auto - correlation function can not be associated solely to an artifact of the void identification procedure .
the average random - voids correlation function shows a similar shape than that of galaxy - defined voids , but also shows a lower amplitude indicating a possible influence from hierarchical clustering on the galaxy - defined void auto - correlations .
the study of the outflow velocities around voids shows outflow velocities which reach a maximum at a distance , @xmath168 . at larger separations , @xmath93 ,
the outflow velocities become moderate infall motions due to the higher density of the void walls and outflow motions from neighbouring voids .
we find that a very simple relation that describes @xmath91 and @xmath93 as a function of void radius , with parameters that can be found in table 1 . table 2 shows the linear fit parameters for the relation between void radius and maximum outflow motions , @xmath60 vs. @xmath105 .
we searched for differences in the velocity field of galaxies and dark - matter haloes around voids , and found that galaxy - defined voids show similar outflows to h@xmath47-defined voids .
also , the use of redshift space data lowers the extreme values of velocities around voids , that is , both outflows and infalls appear less important than in real - space .
we also analysed the galaxy velocity dispersion at the void walls , and find a systematically larger velocity dispersion in the direction parallel to the void walls with respect to the radial direction ( a @xmath169 difference ) .
we find that the cross - correlation functions between voids and galaxies , haloes and mass show negative constant values out to separations comparable with the void size . at larger separations the galaxy - void - galaxy correlation function level increases to approach the auto - correlation function of the galaxies .
the amplitude of the correlation function of galaxies near the centres of galaxy - defined voids is slightly lower than that of the mass around mass - defined voids .
we also studied the distortion pattern observed in @xmath10 and found an elongation along the line of sight .
this elongation differs from that found in the galaxy auto - correlation function , as this effect extends out to large separations perpendicular to the line of sight .
the study of these elongations would allow us to obtain properties of the velocity field around voids using only measured redshifts , which are much easier to obtain observationally than galaxy peculiar velocities .
the study of @xmath10 not only provides information about the peculiar velocity field around voids , but also on finger - of - god motions at the void walls through the elongation of @xmath170 contours along the line of sight , at separations @xmath171 .
we find that the study of voids selected from the distribution of galaxies can provide biased results which can not easily be reconciled with results from dark - matter haloes .
for instance , the number density of voids as a function of radius for galaxy - defined voids is in agreement with that of h@xmath48-defined voids for large void radii , and with h@xmath47 voids at small radii .
also , the void auto - correlation function of galaxy - defined voids is in better agreement with h@xmath48-defined voids , than with h@xmath47-defined voids ; the velocity field of galaxies is in better agreement with h@xmath47-defined voids .
however , the elongations in the galaxy - void @xmath10 are qualitatively matched by h@xmath47-haloes .
therefore care must be taken when making an interpretation based on galaxy - void data .
we present table [ table : summary ] as an overall summary of the differences between properties of voids selected from galaxies and dark - matter haloes in real- and redshift - space .
it can be argued that , observationally , the easiest way to obtain the velocity field surrounding voids is the 2-dimensional correlation function @xmath10 .
we analyse whether it is possible to estimate directly the peculiar velocity field around voids from peculiar velocity data , and also measure @xmath10 using large observational datasets such as 2dfgrs and sdss in the next step of this work , which at present is being carried out in a forthcoming paper ( ceccarelli et al . , 2005a ) , where we pay special attention to many observational biases such as distance measurement errors , irregular angular completeness masks and flux limit effects .
this work was supported in part by the eso - chile joint committee , ndp was supported by a proyecto postdoctoral fondecyt no . 3040038 .
dgl and lc are supported by conicet .
we thank the durham group for providing the semi - analytic galaxy formation and simulation dark - matter outputs used in this work .
we thank the referee for helpful comments and suggestions .
bahcall , n.a .
, & cen , r. 1992 , apj , 398 , 81 .
ceccarelli , l. , et al . , 2005a , in preparation ( c05 ) .
ceccarelli , l. , valotto , c.a . ,
lambas , d.g . ,
padilla , n.d . ,
giovanelli , r. , & haynes , m. , 2005b , accepted for publication in apj .
cole , s. , lacey , c.g . ,
baugh , c.m . , & frenk , c.s . , 2000 ,
mnras , 319 , 168 .
croton , d.j . ,
et al . ( the 2dfgrs team ) , 2004 , submitted to mnras , astro - ph/0401406 .
einasto , j. , einasto , m. , & gramman , m. , 1989 , mnras , 238 , 155 .
el - ad h. , & piran , t. , 1997 , apj , 491 , 421 .
el - ad h. , & piran , t. , 2000 , mnras , 313 , 553 .
geller , m.j .
, & huchra , j.p . , 1989 , sci , 246 , 897 .
ghigna , s. , bonometto , s.a . ,
retzlaff , j. , gottloeber , s. , & murante , g. , 1996 , apj , 469 , 40 .
goldberg , d.m . ,
jones , t.d . , hoyle , f. , rojas , r.r .
, vogeley , m.s .
, & blanton , m.r . , 2004 , submitted to apj , astro - ph/0406527 .
gottlber s. , lokas , e.l . ,
klypin , a. , & hoffman , y. , 2003 , mnras , 344 , 715 .
gregory s.a . , thompson , l.a .
, 1978 , apj , 222 , 784 .
hawkins , e. , et al .
( the 2dfgrs team ) , 2003 , mnras , 346 , 78 .
hoffman , y. , & shaham , j. , 1982 , apj , 262 , l23 .
hoyle , f. , outram , p.j .
, shanks , t. , boyle , b.j . , croom , s.m .
, & smith , r.j . , 2002 , mnras , 332 , 311 .
hoyle , f. , & vogeley , m.s . , 2002 , apj , 566 , 641 .
hoyle , f. , rojas , r.r .
, vogeley , m.s . , & brinkmann , j. , 2003 , submitted to apj , astro - ph/0309728 .
joeveer , m. , einasto , j. , tago , e. , 1978 , mnras , 185 , 357 .
kirshner , r. p. , oemler , a. , schechter , p. l. , shectman , s. a. , 1981 , apj , 248 , 57 .
loveday , j. , maddox , s.j . , efstathiou , g. , & peterson , b.a . , 1995 ,
apj , 442 , 457 .
mller , v. , arbabi - bidgoli , s. , einasto , j. , & tucker , d. , 2000 , mnras , 318 , 280 .
padilla , n.d . ,
merchn , m. , valotto , c. , lambas , d.g . , & maia , m. , 2001 , apj .
padilla , n.d . & baugh , c.m .
, 2003 , 343 , 796 .
padilla , n.d . &
lambas , d.g .
, 2003a , mnras , 342 , 532 .
padilla , n.d . &
lambas , d.g . , 2003b , mnras , 342 , 519 .
patiri , s.g .
, betancort - rijo , j.e . , & prada , f. , 2004 , submitted to apj , astro - ph/0407513 .
peebles , p.j.e . , 1982 , apj , 257 , 438 .
peebles , p.j.e . , 2001 , apj , 557 , 495 .
percival , w. , et al .
( the 2dfgrs team ) , 2004 , mnras accepted , astro - ph/0406513 .
plionis , m. , & basilakos , s. , 2002 , mnras , 330 , 399 .
ryden , b.s . , 1995 ,
apj , 452 , 25 .
spergel , d.n . , et al .
( the wmap team ) , 2003 , apjs , 148 , 175 .
van de weygaert , r. , sheth , r. , & platen , e. , 2004 , proceedings iau colloquium , 195 .
vogeley , m.s . ,
geller , m.j . , park , c. , & huchra , j.p . , 1994 ,
aj , 108 , 745 .
white , s. d. m. , 1979 , mnras , 186 , 145 . | we study statistical properties of voids in the distribution of mass , dark - matter haloes and galaxies ( @xmath0 ) in a @xmath1cdm numerical simulation populated with galaxies using a semi - analytic galaxy formation model(galform , cole et al .
2000 ) .
we find that the properties of voids selected from galform galaxies are compatible with those of voids identified from a population of haloes with mass @xmath2h@xmath3m@xmath4 , similar to the median halo mass , @xmath5h@xmath3m@xmath4 .
we also find that the number density of galaxy- and halo - defined voids can be up to two orders of magnitude higher than mass - defined voids for large void radii , however , we observe that this difference is reduced to about half an order of magnitude when the positions are considered in redshift - space . as expected , there are outflow velocities which show their maximum at larger void - centric distances for larger voids .
we find a linear relation for the maximum outflow velocity , @xmath6 .
the void - centric distance where this maximum occurs , follows a suitable power law fit of the form , @xmath7 . at sufficiently large distances
, we find mild infall motions onto the sub - dense regions .
the galaxy velocity field around galaxy - defined voids is consistent with the results of haloes with masses above the median , showing milder outflows than the mass around mass - defined voids .
we find that a similar analysis in redshift space would make both outflows and infalls to appear with a lower amplitude .
we also find that the velocity dispersion of galaxies and haloes is larger in the direction parallel to the void walls by @xmath8 . given that voids are by definition sub - dense regions , the cross - correlation function between galaxy - defined voids and galaxies are close to @xmath9 out to separations comparable to the void size , and at larger separations the correlation function level increases approaching the values of the auto - correlation function of galaxies .
the cross - correlation amplitude of mass - defined voids vs. mass has a more gentle behaviour remaining negative at larger distances .
the cross- to auto - correlation function ratio as a function of the distance normalised to the void radius , shows a small scatter around a relation that depends only on the object used to define the voids ( galaxies or haloes for instance ) .
the distortion pattern observed in @xmath10 is that of an elongation along the line of sight which extends out to large separations .
positive @xmath11 contours evidence finger - of - god motions at the void walls .
elongations along the line of sight are roughly comparable between galaxy- , halo- and mass - defined voids
. large - scale structure of universe , methods : n - body simulations , galaxies : kinematics and dynamics cosmology : theory |
our considerations here will be connected with the geometry of the quasiclassical electron trajectories on the fermi surface in the presence of a strong magnetic field .
the main goal of the paper is to give a detailed consideration of the contribution of the stable open trajectories to the conductivity tensor in the limit @xmath3 .
as we will see below , in spite of rather regular geometric properties of the stable ( generic ) open trajectories , their contribution to magneto - conductivity is quite non - trivial from analytical point of view , which is caused by nontrivial statistical properties of the trajectories of this kind . here
we will try to represent a general picture of the magneto - conductivity behavior in the case of presence of such trajectories on the fermi - surface , including the description of the dependence of conductivity on both the magnitude and the direction of @xmath4 .
as we will see , both the structure of a `` stability zone '' on the angle diagram and the conductivity behavior in strong magnetic fields will demonstrate actually rather non - trivial properties .
we will base our considerations on the topological picture for the electron dynamics in the space of the quasimomenta , arising for general dispersion relation @xmath5 under the presence of a magnetic field .
so , according to standard approach we will assume that the electron states in the conductivity zone are parametrized by the values of the quasimomenta @xmath1 which in fact should be considered as points of the three - dimensional space , factorized over the vectors of the reciprocal lattice : @xmath6 the basis vectors of the reciprocal lattice are connected with the basis vectors @xmath7 of the direct lattice by the formulae @xmath8 @xmath9 and define the fundamental ( brillouen ) zone in the @xmath10 - space .
the opposite faces of the parallelepiped formed by @xmath11 should be identified with each other , so the space of the electron states ( for a given conductivity zone ) represents in fact the three - dimensional torus @xmath12 .
the dispersion relation @xmath5 can be considered either as a 3-periodic function in the @xmath10 - space or just as a smooth function on the compact space @xmath12 , given by the factorization of the @xmath10 - space over the reciprocal lattice @xmath13 . in the same way ,
the constant energy levels @xmath14 can be considered either as 3-periodic two - dimensional surfaces in the @xmath10 - space or as smooth compact surfaces @xmath15 embedded in the three - dimensional torus @xmath12 .
being considered in the last way , every nonsingular energy level represents a smooth compact orientable surface , which is topologically isomorphic to one of the canonical surfaces defined by genus @xmath16 ( fig .
[ genus ] ) . 0 , 1 , 2 @xmath17 ( etc ... ) .
] the genus @xmath18 of the fermi surface represents an important topological characteristic of the electron spectrum in metal .
another important topological characteristic of the electron spectrum is the way of embedding of the fermi surface in @xmath12 . as an example , fig .
[ maximalrank ] @xmath17 represents an embedding of a genus 3 surface in the first brillouen zone having `` maximal rank '' . .
] the quasiclassical evolution of the electron states in the presence of external homogeneous magnetic field @xmath0 can be described by the adiabatic equation ( see e.g. @xcite ) : @xmath19 \,\,\,\ , = \,\,\,\ , { e \over c } \,\ ,
\left [ \nabla \epsilon ( { \bf p } ) \ , \times \ , { \bf b } \right]\ ] ] in the space of the quasimomenta .
the equation ( [ quasiclassicalevolution ] ) is analytically integrable in @xmath10 - space and the quasiclassical trajectories are given by the intersections of the constant energy surfaces @xmath14 with the planes orthogonal to @xmath20 .
however , despite this analytical property of system ( [ quasiclassicalevolution ] ) , the global geometry of the quasiclassical trajectories in @xmath10 - space can be rather complicated .
the reason for this circumstance is that the total @xmath1-space is not compact while in @xmath21 the function @xmath5 represents the only one - valued integral of motion of ( [ quasiclassicalevolution ] ) . in general
, we can expect rather complicated picture of intersection of the 3-periodic surface @xmath14 with the planes orthogonal to general ( totally irrational ) direction of @xmath22 ( see e.g. fig . [ generalsurface ] ) . .
] as it is usual for the case of normal metals , we will assume here that all the electron states with energies below the fermi energy @xmath23 are occupied by electrons , while all the electron states with energies above the fermi energy are empty .
the fermi surface @xmath24 represents a two - dimensional surface in @xmath21 separating occupied electron states from the empty ones .
we have naturally @xmath25 where @xmath26 and @xmath27 are the minimal and maximal possible electron energies in the conductivity zone . system ( [ quasiclassicalevolution ] ) conserves the energy of a particle and also the phase volume element @xmath28 . as a consequence , the evolution according to system ( [ quasiclassicalevolution ] ) does not change the fermi distribution of the quasiparticles or more general temperature dependent equilibrium distributions @xmath29 however , it appears that the form of the trajectories of ( [ quasiclassicalevolution ] ) plays an important role for the response of the electron system to the small electric fields @xmath30 , which defines the magneto - conductivity properties of normal metals . due to the high degeneracy of the electron gas in metals , all the properties of such transport phenomena
are defined in fact by the properties of system ( [ quasiclassicalevolution ] ) just on the one energy level @xmath31 i.e. just on the one ( fermi ) energy surface @xmath32 .
the role of the global geometry of the quasiclassical electron trajectories in the transport phenomena in strong magnetic fields was first revealed by the school of i.m .
lifshitz ( i.m .
lifshitz , m.ya .
azbel , m.i .
kaganov , v.g .
peschansky ) in 1950 s ( see @xcite ) .
thus , in paper @xcite a crucial difference in contribution to magneto - conductivity of closed ( fig .
[ closedandper ] , a ) and open periodic ( fig .
[ closedandper ] , b ) quasiclassical trajectories in the limit @xmath33 was pointed out .
-space ( only one brillouen zone is shown ) . ]
as was shown in @xcite , the contribution to the magneto - conductivity of the trajectories of the first kind is almost isotropic in the plane orthogonal to @xmath22 and has a rapidly decreasing character in this plane as @xmath34 .
in contrast , the presence of the open periodic trajectories gives an additional contribution to the conductivity in the plane orthogonal to @xmath4 , which is strongly anisotropic in the limit @xmath34 .
the asymptotic form of the conductivity tensor in the plane orthogonal to @xmath22 in the described cases can be represented as @xmath35 ( closed trajectories ) , @xmath36 ( open periodic trajectories ) , after an appropriate choice of coordinates .
( here and below the notation @xmath37 means just a constant of the order of unity ) .
the formulae ( [ closedtraj ] ) - ( [ periodictraj ] ) represent the asymptotic form of the conductivity tensor , so all the equalities give actually just the order of the absolute value of the corresponding quantities .
the value @xmath38 represents here just some ( approximate ) effective mass of electron in crystal and @xmath39 denotes the mean free path electron time defined by the intensity of the scattering processes .
the value @xmath40 can be considered as an `` approximate value '' of the cyclotron frequency in the crystal . from geometrical point of view
@xmath41 represents approximately the inverse time of motion along a closed trajectory ( fig .
[ closedandper ] , a ) or the inverse time of motion through one brillouen zone for periodic open trajectories ( fig .
[ closedandper ] , b ) .
it becomes clear from this point of view that the global geometry of the quasiclassical trajectories reveals exactly in the limit @xmath42 when the mean free path electron time exceeds the time of motion inside one brillouen zone .
let us say here that for experimental observation of the `` geometric strong magnetic field limit '' rather pure materials under rather low temperatures ( @xmath43 ) and rather strong magnetic fields ( @xmath44 ) should be usually used . in formula ( [ periodictraj ] )
the @xmath45 - axis is chosen along the mean direction of the periodic trajectories in @xmath10 - space . as can be easily derived from system ( [ quasiclassicalevolution ] ) ,
the projection of the corresponding trajectory in coordinate space onto the plane orthogonal to @xmath22 is given just by the rotation of the trajectory in @xmath10 - space by @xmath46 ( fig .
[ pandxtraject ] ) . - space ( a ) @xmath17 and its projection on the plane , orthogonal to @xmath4 , in @xmath47 space ( b ) . ]
we can see then that even very strong magnetic field does not block the electron motion along the @xmath48 - axis in the coordinate space in the presence of the trajectories shown at fig .
[ closedandper ] , b. as a result , the @xmath48 - component of conductivity in the plane orthogonal to @xmath22 remains finite in the limit @xmath3 .
the strong anisotropy of the conductivity tensor in this situation gives an experimental possibility to observe this phenomenon as well as to determine the mean direction of the open trajectories in @xmath10 - space .
the conductivity along the direction of @xmath22 remains finite in the limit @xmath42 in both the cases described above .
the asymptotic forms of the full conductivity tensor in @xmath49 - space can be represented as : @xmath50 ( closed trajectories ) , @xmath51 ( open periodic trajectories ) .
let us note here that both the trajectories shown at fig .
[ closedandper ] @xmath17 are actually closed in the torus @xmath12 .
however , their embeddings in @xmath21 are completely different from topological point of view .
thus , the embedding of the trajectory shown at fig .
[ closedandper ] , a @xmath17 is `` homologous '' to zero , while the embedding of the trajectory shown at fig .
[ closedandper ] , b @xmath17 represents a nonzero homology class in @xmath12 . certainly , the difference between the trajectories shown at fig .
[ closedandper ] , a @xmath17 and @xmath17 fig .
[ closedandper ] , b @xmath17 is more evident in the covering @xmath1 - space where they have different global geometry .
let us say here also , that both the situations shown at fig .
[ closedandper ] a , b @xmath17 represent the special case when the conductivity tensor can be represented as a regular series in powers of @xmath52 in the limit @xmath3 .
thus , we can actually write here @xmath53 ( closed trajectories ) , @xmath54 ( open periodic trajectories ) , where the even powers of @xmath55 correspond to the symmetric part of @xmath56 , while the odd powers of @xmath55 represent the anti - symmetric part in accordance with the onsager relations . as we will see below
, this situation does not take place for general open electron trajectories on the fermi surface . in papers @xcite different examples of complicated fermi surfaces and more general types of open electron trajectories in strong magnetic fields
were considered .
the trajectories considered in @xcite are in general not periodic and are non - closed both in @xmath21 and the @xmath1 - space and represent the first examples of the stable open electron trajectories on a complicated fermi surface .
the form of the trajectories found in @xcite demonstrates also strongly anisotropic properties , so the strong anisotropy of the conductivity tensor in the limit @xmath42 is also expected in this case .
as was pointed out in @xcite , the analytical behavior of conductivity in these examples can in fact be different from ( [ 3dimperiodictr ] ) and demonstrate a slower approach to its limiting form in the interval of not extremely strong magnetic fields . here
we will try to study this question in general case using the topological description of the stable open trajectories on the fermi surface . in the survey articles @xcite and also in the book @xcite a remarkable review of both the theoretical and experimental investigations of this branch of electron theory of metals made at that time can be found .
let us give here also a reference on a survey article @xcite where the same topics were revisited after forty years and which contains also the aspects appeared in the later period .
the general problem of classification of all possible types of trajectories of dynamical system ( [ quasiclassicalevolution ] ) with arbitrary ( periodic ) dispersion law @xmath5 was set by s.p .
novikov ( @xcite ) and was intensively investigated in his topological school ( s.p .
novikov , a.v .
zorich , s.p .
tsarev , i.a .
dynnikov ) . the topological investigation of system ( [ quasiclassicalevolution ] ) has led to rather detailed understanding of the geometry of the trajectories of different types and gave finally the full classification of all possible types of trajectories of ( [ quasiclassicalevolution ] ) .
one of the most important parts of the mathematical theory of the trajectories of ( [ quasiclassicalevolution ] ) , which will be also of great importance in the present paper , is connected with the detailed description of the stable open trajectories of system ( [ quasiclassicalevolution ] ) .
let us say here that the most important breakthroughs in this problem were made in papers @xcite where rather deep topological theorems about non - closed trajectories of system ( [ quasiclassicalevolution ] ) were proved . using the topological description of the stable open trajectories of system ( [ quasiclassicalevolution ] ) it was possible to introduce important topological characteristics of the stable nontrivial regimes of the magneto - conductivity behavior in the limit @xmath42 , which were introduced in paper @xcite . in general
, the characteristics introduced in @xcite can be described in the following way : let us exclude now the `` trivial '' cases when we have just closed electron trajectories on the fermi - surface in @xmath1 - space and consider the situation when open trajectories ( in @xmath1 - space ) are present on the fermi level .
besides that , let us require that the open trajectories are stable with respect to small rotations of @xmath22 such that we have a `` stability zone '' in the space of directions of @xmath20 , where similar open trajectories exist on the fermi level . as can be extracted from the topological description of the stable open trajectories of system ( [ quasiclassicalevolution ] )
, the trajectories of this type demonstrate the following remarkable properties : \1 ) every stable open ( in @xmath1 - space ) trajectory of ( [ quasiclassicalevolution ] ) lies in a straight strip of a finite width in the plane orthogonal to @xmath20 , passing through it from @xmath57 to @xmath58 ( fig .
[ stableopentr ] ) ; ) passing through a straight strip of a finite width in the plane orthogonal to @xmath20 . ] \2 ) the mean direction of all open stable trajectories in @xmath1 - space is given by the intersection of the plane orthogonal to @xmath22 and some integral ( generated by two reciprocal lattice vectors ) plane @xmath59 , which is the same for a given `` stability zone '' in the space of directions of @xmath20 .
let us note here that the statement ( 1 ) was first formulated by s.p .
novikov in the form of a conjecture , which was proved later for the stable open trajectories of ( [ quasiclassicalevolution ] ) ( @xcite ) .
as was pointed out in @xcite , the integral planes @xmath60 represent experimentally observable objects view the remarkable geometric properties of the open stable trajectories , described above .
thus , measuring the conductivity in the limit @xmath3 , we can observe strongly anisotropic behavior of conductivity in the planes orthogonal to @xmath22 for all @xmath61 and extract the integral plane @xmath62 , which is swept by the directions of the highest decreasing of conductivity in the @xmath63 - space .
more precisely , let us introduce the values @xmath64 then , we can chose a basis in @xmath63 - space @xmath65 smoothly depending on @xmath22 in @xmath66 , where the values @xmath67 will have the form @xmath68 as was said above , the direction of the vector @xmath69 is given by the intersection of the plane orthogonal to @xmath22 with an integral plane , i.e. the plane generated by two reciprocal lattice vectors @xmath70 , @xmath71 , which is the same for the zone @xmath66 .
the integral plane @xmath60 can be represented by an indivisible triple of integer numbers @xmath72 according to the equation @xmath73 where @xmath74 are the basis direct lattice vectors .
the numbers @xmath72 were called in @xcite the topological quantum numbers observable in conductivity of normal metals .
let us say here that the triples @xmath72 can be rather nontrivial for complicated fermi surfaces .
the complete zone @xmath75 can be called here the mathematical stability zone , corresponding to a given fermi energy .
we have to say here , however , that in experimental study of magneto - conductivity it is natural to introduce extended `` experimentally observable stability zone '' @xmath76 @xmath77 due to specific behavior of trajectories of system ( [ quasiclassicalevolution ] ) near the boundary of @xmath66 .
everywhere in @xmath78 the behavior of conductivity is experimentally indistinguishable from that in @xmath75 even under good formal implementation of the condition @xmath79 until the magnitude @xmath80 of magnetic field becomes extremely high .
the conventional form of @xmath78 depends actually on the experimental conditions and is determined mainly by the maximal values of @xmath80 in experiment .
the difference between the zones @xmath78 and @xmath66 , which can be detected only in extremely strong magnetic fields , represents in fact one of essential features of analytical behavior of conductivity , considered here .
let us note now that the stable open quasiclassical trajectories do not represent all possible types of non - closed trajectories of system ( [ quasiclassicalevolution ] ) and trajectories with more complicated geometry can arise on complicated fermi surfaces for some special directions of @xmath4 .
thus , the first example of trajectory which can not be restricted by any straight strip of a finite width was constructed by s.p .
tsarev ( @xcite ) .
the trajectory of tsarev demonstrates explicit chaotic behavior on the fermi surface and is essentially different from the stable open trajectories from this point of view . on the other hand ,
the tsarev trajectory is also strongly anisotropic in the @xmath1 - space and has an asymptotic direction in the plane orthogonal to @xmath4 .
so , from experimental point of view the contribution of tsarev trajectories to the magneto - conductivity is also strongly anisotropic in the limit @xmath42 and corresponds to the regime ( [ 3dimanisotropiclimit ] ) .
let us say , however , that the tsarev chaotic trajectories are completely unstable under small rotations of @xmath22 and are not connected with topologically stable characteristics unlike the stable open trajectories discussed above .
we have to say also , that the trajectories of tsarev type ( if they present ) can appear just for the set of zero measure in the space of directions of @xmath4 .
another example of non - closed trajectory of ( [ quasiclassicalevolution ] ) which demonstrates strongly chaotic behavior both in @xmath21 and in @xmath1 - space was constructed by i.a .
dynnikov in @xcite .
the trajectories of dynnikov are rather different from the tsarev chaotic trajectories and require the maximal irrationality of the direction of @xmath22 to be observed .
the behavior of the conductivity tensor in the presence of the chaotic trajectories constructed in @xcite was investigated in @xcite and appeared to be rather different from both the regimes given by ( [ 3dimclosedtr ] ) and ( [ 3dimperiodictr ] ) .
the most interesting feature of the contribution to the conductivity tensor of the trajectories of this kind is that the presence of such trajectories suppresses the conductivity in all directions , including the direction of @xmath4 , in the limit @xmath3 .
in general , the tsarev and dynnikov chaotic trajectories demonstrate two different general types of possible chaotic behavior of trajectories of system ( [ quasiclassicalevolution ] ) , so in this sense all chaotic trajectories of ( [ quasiclassicalevolution ] ) can be attributed to the tsarev or dynnikov type .
let us say here that different properties of chaotic trajectories of system ( [ quasiclassicalevolution ] ) are also intensively investigated in modern mathematical literature ( see @xcite ) .
the directions of @xmath22 , corresponding to the chaotic open trajectories , can appear only outside the `` stability zones '' , corresponding to the stable open trajectories of system ( [ quasiclassicalevolution ] ) . in the same way , the chaotic open trajectories of different types ( tsarev or dynnikov type ) can not appear at the same direction of @xmath20 .
let us say , however , that in general the contribution of non - closed trajectories to the conductivity tensor should be added with the contribution ( [ 3dimclosedtr ] ) of closed trajectories of system ( [ quasiclassicalevolution ] ) , which usually present together with open trajectories ( of any kind ) at the same fermi surface
. for more rigorous description of possible situations in the behavior of trajectories of ( [ quasiclassicalevolution ] ) with arbitrary periodic dispersion law let us give here a reference on a detailed mathematical survey @xcite , devoted to this question . the detailed consideration of different physical results based on the results of topological investigations of system ( [ quasiclassicalevolution ] )
can be found in @xcite .
our main goal in this paper will be more detailed investigation of the conductivity behavior in the `` stability zones '' , based on the detailed topological description of the stable open electron trajectories arising for the corresponding directions of @xmath4 . as we said already ,
the corresponding stable open trajectories demonstrate rather regular geometric properties in the planes orthogonal to @xmath22 .
thus , all the stable open trajectories have the form shown at fig . [ stableopentr ] and the same mean direction in @xmath10 - space , given by the intersection of the plane orthogonal to @xmath22 with an integral plane @xmath81 , which is fixed for the corresponding stability zone @xmath66 . as a result , the conductivity tensor has the asymptotic form ( [ 3dimanisotropiclimit ] ) in the limit @xmath42 and demonstrates the specific geometric properties described above .
these geometric ( or topological ) properties of the conductivity tensor give the most stable ( topological ) characteristics of the conductivity for a given `` stability zone '' and are connected with the `` purely geometric '' limit @xmath3 . at the same time
, we will see below , that the conductivity tensor demonstrates in general rather nontrivial analytical dependence on @xmath22 in different parts of `` experimentally observable '' stability zones , which is caused by specific features of different trajectories of system ( [ quasiclassicalevolution ] ) in these regions . here
we will investigate the analytical properties of conductivity for the case of rather big ( but finite ) values of the parameter @xmath55 using topological description of `` carriers of open trajectories '' on the fermi - surface , defined for the corresponding directions of @xmath4 .
as we will see below , both the dependence on the parameter @xmath55 and on the direction of @xmath22 demonstrate here rather nontrivial properties . in the next section we will give a description of `` experimentally observable '' stability zone and try to describe the main features of the analytical behavior of conductivity in its different parts , which gives from our point of view the main points of the picture arising in general case . in sections iii and
iv we will present more detailed consideration of the contribution of the stable open trajectories to the magneto - conductivity , based on the topological description of such trajectories on complicated fermi surfaces .
sections iii and iv will have more mathematical character in comparison with section ii .
to describe briefly the general picture , let us point out now the main points , which will play the most essential role in our considerations below
. this section will be mostly descriptive , while more detailed analysis will be presented in sections iii , iv . as we said already ,
the stable open trajectories demonstrate quasiperiodic properties for generic directions of @xmath82 ( @xmath83 ) . however , they become purely periodic for some special directions of the magnetic field .
it s not difficult to see that the last situation arises every time when the plane orthogonal to @xmath22 intersects the plane @xmath60 along an integer vector in @xmath1 - space : @xmath84 ( @xmath85 ) .
both the cases of the quasiperiodic and the periodic trajectories lead in this situation to the conductivity tensor ( [ 3dimanisotropiclimit ] ) in the limit @xmath3 , having the same stable geometric properties for a given `` stability zone '' .
however , as we will see below , the limiting values of the components @xmath86 , @xmath87 , @xmath88 , @xmath89 , @xmath17 i.e. the values @xmath90 ( @xmath91 is fixed ) , will have here sharp `` jumps '' with respect to the same values , defined for close generic directions of @xmath4 .
more precisely , we can introduce the functions @xmath92 , @xmath93 which represent continuous functions of @xmath94 in @xmath75 and coincide with the corresponding values @xmath95 for generic directions @xmath96 .
however , for special directions of @xmath22 described above we will have @xmath97 which is caused by special features of statistical averaging over the open trajectories of ( [ quasiclassicalevolution ] ) near the fermi surface .
it s not difficult to show also that we will always have the inequalities @xmath98 @xmath99 for these special directions of @xmath4 , while the signs of @xmath100 are in general indefinite .
the directions @xmath96 , corresponding to a fixed rational mean direction @xmath101 of open trajectories , represent a one - dimensional curve @xmath102 ( on @xmath103 ) given by the intersection of the big circle , orthogonal to @xmath104 , and the stability zone @xmath75 ( fig .
[ specdirmagnfield ] a ) .
we have to say , however , that according to a more detailed consideration the periodic trajectories exist actually on some bigger curve @xmath105 representing extension of the curve @xmath102 outside the stability zone @xmath75 ( fig .
[ specdirmagnfield ] b ) .
all the open trajectories have the same mean direction @xmath101 for all @xmath106 and the area covered by the periodic trajectories on the fermi surface vanishes at the endpoints of @xmath107 .
the open periodic trajectories demonstrate stability properties inside the stability zone @xmath75 and become unstable outside @xmath75 on the curve @xmath107 .
let us note also here that the measure of generic ( quasiperiodic ) open trajectories on the fermi surface remains finite up to the boundary of a stability zone .
in general , the schematic angle diagram corresponding to a complicated fermi surface can be represented by fig .
[ anglediagramm ] .
the measure of the stable open trajectories on the fermi surface remains finite up to the boundary of the `` mathematical stability zone '' @xmath108 .
however , it is often assumed in the literature that this measure vanishes at the boundary of a stability zone on the angle diagram , which seems to be in accordance with experimental data .
as we will see below , the last circumstance is caused by the existence of an extended `` experimentally observable '' stability zone @xmath78 including an additional region @xmath109 , adjacent to the boundary of @xmath108 , with very special behavior of trajectories of system ( [ quasiclassicalevolution ] ) .
it will be also shown , that the presence of additional zone @xmath110 is responsible also for arising of rather nontrivial analytical regimes of conductivity behavior in the extended zone @xmath111 . ] of special directions of @xmath4 , corresponding to a rational mean direction @xmath101 of open trajectories , within a stability zone @xmath66 .
@xmath112 ( b ) the full one - dimensional set @xmath105 of directions of @xmath4 , corresponding to existence of periodic trajectories with mean direction @xmath104 , intersecting the stability zone @xmath66 . ] on the unit sphere . ]
we can see then that every stability zone @xmath75 is actually covered by an everywhere dense net representing the `` special directions '' of @xmath22 corresponding to different rational mean directions @xmath101 of the stable open trajectories ( fig .
[ densenet ] ) .
the set of all possible rational mean directions @xmath101 of the open trajectories is given by two conditions : \1 ) @xmath113 , \2 ) @xmath114 , where @xmath115 is the big circle orthogonal to the direction @xmath116 .
inside @xmath117 , corresponding to arising of periodic trajectories , in the cases when the direction orthogonal to @xmath60 does not belong to @xmath75 ( a ) and when this direction belongs to @xmath75 ( b ) . ] everywhere on the net we have the situation @xmath118 however , the difference @xmath119 decreases with the growth of the numbers @xmath120 in ( [ specdirbeq ] ) .
as we will see below , the approximate evaluation for this difference for complicated fermi surfaces can be written in the form @xmath121 ( @xmath122 ) . in the physical situation , when the values of @xmath22 are finite but rather big , the `` jumps '' on the curves @xmath102 become very sharp `` peaks '' on these curves , which can be called the `` rational peaks '' in the values of magneto - conductivity according to rationality of the mean direction of the corresponding open trajectories in @xmath1 - space . the second important point in our picture
is connected with the behavior of trajectories of system ( [ quasiclassicalevolution ] ) near the boundaries of the stability zones @xmath75 outside the stability zones .
as we have already said , the periodic trajectories of system ( [ quasiclassicalevolution ] ) on the fermi level exist in fact for directions of @xmath22 belonging to the extended curves @xmath107 , so only at the endpoints of @xmath105 the measure of the trajectories with the period @xmath101 on the fermi surface is equal to zero . as a result , the full set of directions of @xmath22 near @xmath75 , corresponding to the presence of open trajectories on the fermi level , represents the stability zone @xmath75 with a set of additional segments near @xmath108 , which is dense on the boundary of @xmath75 ( fig .
[ diroutstabzone ] ) .. we would like to emphasize here that the additional segments arise for every curve @xmath123 connected with a rational mean direction of open trajectories of ( [ quasiclassicalevolution ] ) . as a result ,
these segments form a dense set near the boundary of @xmath108 , which plays an important role in the formation of additional `` experimentally observable '' zone @xmath110 around the exact mathematical stability zone @xmath108 . ] near @xmath117 , corresponding to arising of periodic trajectories on the fermi surface . ] at the same time , the measure of open trajectories on the fermi surface for generic directions @xmath124 remains finite up to the boundary of the stability zone and jumps abruptly to zero outside @xmath108 .
the transformation of the open trajectories on the boundary of @xmath75 leads in this case to appearance of very long closed trajectories near the boundary of @xmath75 which are very similar to open trajectories from experimental point of view up to very ( extremely ) high values of @xmath125 .
we can state then that in experimental observation of conductivity we observe actually some `` extended stability zone '' @xmath111 , including the exact stability zone as a subset .
the boundaries of the `` extended zone '' @xmath78 are in a sense conditional and depend on the maximal values of @xmath80 used in the experiment .
we can claim , however , that the experimentally observable stability zone exceeds the exact stability zone by a set of a finite measure for any arbitrary big maximal value of magnetic field @xmath126 ( see fig .
[ extstabzone ] ) .
observable in experiments . ]
the set @xmath127 can be called a `` supplement '' to the stability zone @xmath128 arising in the experimental observation of conductivity in strong magnetic fields .
it is possible to introduce also the values @xmath129 and @xmath130 in the regions @xmath109 , which will have the same meaning as in the regions @xmath117 . in these regions
we will have identically : @xmath131 like in @xmath108 , we will have here @xmath132 for generic directions of @xmath22 and in general @xmath133 for @xmath134 both inside the regions @xmath75 and @xmath110 the approach of the values of @xmath135 to their limiting values @xmath67 is rather irregular ( but with some general trend ) , which makes the conductivity behavior rather nontrivial from the experimental point of view even for the values of @xmath125 , satisfying the formal criterion @xmath136 . first , as we have already noted , the values @xmath95 are different from the values @xmath92 on the curves @xmath105 both in the regions @xmath75 and @xmath109 .
these values are continuous along every curve @xmath105 up to the boundary of the zone @xmath78 ( more precisely , up to the endpoints of @xmath107 ) , at the same time , they are discontinuous in the transverse directions to @xmath107 .
formally speaking , all the open trajectories in the zone @xmath110 are periodic , so their contribution to magneto - conductivity can be represented in the regular form ( [ 3dimperiodictr ] ) . however , as can be easily seen , the period of the trajectories given by formula ( [ specdirbeq ] ) can be rather large for big numbers @xmath137 . as a result
, the features of the regular expansion can be observed actually just on the scales @xmath138 at the same time , on the scales of the values of @xmath80 , satisfying the condition @xmath139 the periodic open trajectories are indistinguishable from the generic ones if the values of @xmath80 satisfy restriction ( [ bintermediate ] ) .
let us say here also , that under the same restrictions on @xmath80 the periodic open trajectories are also indistinguishable from the long closed trajectories in the vicinity of the curve @xmath105 in the zone @xmath109 .
let us now make some remarks concerning the experimental study of the magneto - conductivity in ( monocrystalline ) metals having complicated fermi surfaces .
let us note first that the measuring the magneto - conductivity in presence of the open trajectories very often is being made for directions of @xmath22 belonging to the largest `` stability zones '' , characterized by the biggest area and the highest symmetry on the angle diagram .
rather often it is convenient also to fix just a small set of directions of @xmath22 inside the stability zone and to measure the magneto - conductivity changing the magnitude @xmath80 of the magnetic field . according to the design of the research facility it appears rather often that the corresponding set of directions of @xmath22 represents in fact some symmetric or `` rational '' points on the angle diagram , such that all ( or a part ) the directions belong actually to some curves @xmath140 corresponding to rather simple periods @xmath101 of the open trajectories
this is exactly the case when the magneto - conductivity demonstrates the regular behavior ( [ sigmaregexpperiodic ] ) under the simple condition @xmath136 .
the corresponding magneto - resistance tensor can be written here in the form @xmath141 in particular , the behavior of the transverse resistivity @xmath142 demonstrates a regular quadratic law : @xmath143 where @xmath144 is the angle with the @xmath145-axis in our coordinate system .
the situation is rather different in the case when the open trajectories have an irrational mean direction @xmath146 in the plane @xmath62 . from one point of view ,
the limiting values of the conductivity tensor are also given here by the relations ( [ 3dimanisotropiclimit ] ) in the limit @xmath3 .
at the same time , the regular series ( [ sigmaregexpperiodic ] ) , ( [ resistperiodtr ] ) can not be written in general for generic mean direction of the open trajectories .
it can be shown also that the asymptotic behavior of the values @xmath56 demonstrates here a slower ( irregular ) approach to their limiting values @xmath147 .
in general , the asymptotic behavior of @xmath56 inside @xmath75 for irrational mean direction of the open trajectories can be represented in the form : @xmath148 where the functions @xmath149 , @xmath150 , @xmath151 have the following asymptotic behavior @xmath152 at the same time , the functions @xmath153 , @xmath154 have the following `` general trend '' @xmath155 let us emphasize here that the relation @xmath156 in ( [ trend ] ) means just some general trend in the behavior of the corresponding functions which admits additional irregular corrections on the finite scales . in general , these corrections can be characterized as `` cascades of step - like perturbations '' with the structure defined by the properties of the number @xmath157 ( fig . [ steps ] ) .
let us note here that the behavior of @xmath153 , @xmath154 can demonstrate noticeable local deviations from the trend for special irrational @xmath157 which can be approximated by rational numbers with a high precision .
in particular , the last circumstance can be important for the directions of @xmath22 close to the special directions @xmath158 which were discussed above .
the behavior of the function @xmath159 is even more complicated compared with that of @xmath160 and @xmath154 . here
we would like to suggest just the following `` reliable '' restriction @xmath161 on the @xmath125 - dependence of the function @xmath159 which can demonstrate rather irregular behavior at different values of @xmath80 . and @xmath154 in the region @xmath162 .
] another essential feature of the relations ( [ trend ] ) - ( [ rbtrend ] ) is that the simple condition @xmath79 is not sufficient here to get the values @xmath67 with a good precision .
the latter circumstance can be understood from the fact that the values @xmath67 are given now by the averaging of certain dynamical values over a complicated part of the fermi - surface but not a one - dimensional circle . as a result
, relations ( [ trend ] ) - ( [ rbtrend ] ) can contain in fact a large dimensionless coefficient , determined by the geometry of the fermi surface .
we should expect then , that in the interval of not extremely high values of @xmath80 the geometric characteristics of the carriers of open trajectories will play important role in the behavior of @xmath56 and can give noticeable deviations from the asymptotic regime ( [ trend ] ) . as a consequence
, we should expect also that in the interval of not extremely strong magnetic fields the interplay between the limiting values of @xmath56 and the corrections depending on @xmath80 will play an essential role in conductivity behavior .
in general , the picture described above demonstrates the main features of the behavior of @xmath56 inside the mathematical stability zones @xmath75 .
in the zone @xmath110 we have to introduce the function @xmath163 , @xmath164 , characterizing the mean size of long closed trajectories for generic directions of @xmath4 .
in common , the function @xmath163 can be defined as the ratio of the length of such trajectories in the @xmath10 - space to the size of the brillouen zone .
let us note that the definition of the function @xmath163 has actually rather approximate character .
it is natural then to introduce the `` intermediate '' stable values of conductivity @xmath165 , @xmath166 , which are proportional to the measure of the long closed trajectories on the fermi surface . for the behavior of conductivity for not extremely strong magnetic fields @xmath167 and
generic directions of @xmath22 in @xmath110 we can then use the following relations @xmath168 ( @xmath169 ) , with the same remarks about the functions @xmath149 , @xmath150 , @xmath151 , @xmath153 , @xmath170 , @xmath171 . to describe the asymptotic behavior of conductivity in the region of extremely strong magnetic fields @xmath172 for the same directions of @xmath22 it is convenient to consider separately the symmetric @xmath173 and the anti - symmetric part @xmath174 of the tensor @xmath56 .
as will be shown below , the corresponding parts of @xmath56 can be approximated here by the following expressions @xmath175 where @xmath176 and @xmath177 respectively .
let us note , that the interplay between the symmetric and the anti - symmetric parts of @xmath56 has here rather nontrivial character .
thus , we can see that the values @xmath178 and @xmath179 are defined mostly by the symmetric part in the interval @xmath180 and by the anti - symmetric part for @xmath181 .
another important feature in the behavior of conductivity is given here by the relation ( [ sigmaprime33 ] ) , which indicates the suppression of the conductivity along the direction of @xmath22 in the region @xmath182 .
the values of @xmath163 are equal to @xmath183 on the exterior boundary of the region @xmath109 . at the same time
, we should put @xmath184 on the curves @xmath105 in @xmath110 and on the boundary between the regions @xmath75 and @xmath109 .
it s not difficult to see that the functions @xmath165 lose their sense and are not well defined on the exterior boundary of @xmath109 . at the same time , we can write @xmath185 , where @xmath186 are the boundary values of @xmath187 inside @xmath108 , on the boundary between @xmath75 and @xmath109 . according to the above remarks , we can see actually , that the exact boundary between the regions @xmath75 and @xmath110 is not usually observable in experiments .
we can see that the picture described above makes the analytic structure of tensor @xmath188 in the zone @xmath78 rather complicated .
the most complicated situation can arise in that part of the zone @xmath110 where we have the relation @xmath189 for typical values of @xmath80 in the experiment . in the last case
the behavior of @xmath56 can demonstrate the most complicated trends , which are intermediate between ( [ sigmalowbtau ] ) and ( [ sigmahighbtausymm ] ) - ( [ sigmahighbtauantisym ] ) .
in experimental study of the behavior of conductivity it can be convenient to approximate the corresponding regimes with the aid of intermediate powers of @xmath55 and write @xmath190 ( @xmath191 ) in the intervals of rather high values of @xmath80 . as an example , putting approximately @xmath192 , we can write the relation for the resistivity behavior in strong magnetic fields : @xmath193 in particular , we get the following approximation for the resistivity in the plane orthogonal to @xmath22 : @xmath194 where @xmath144 is the angle with the @xmath145-axis in our coordinate system .
we should say , however , that the powers @xmath195 and @xmath196 play here just the role of local approximating parameters and are unstable both with respect to rotations of @xmath22 and big changes of its absolute value .
let us note also that the corresponding part of @xmath110 has in general rather complicated structure view the complicated behavior of the function @xmath163 ( fig .
[ complbehaviorzone ] ) . .
] we can see then that the experimental study of galvano - magnetic phenomena in the presence of open trajectories can reveal rather non - trivial behavior of magneto - conductivity , which is caused both by the complicated dependence of conductivity on the value of @xmath80 and possible small variations of directions of @xmath22 in experiment .
the general analytical picture described above can be used for description of the dependence of the magneto - conductivity both on the value of @xmath80 and the direction of @xmath22 inside the `` experimentally observable '' stability zone @xmath111 .
let us say again that the picture described above is based just on geometric consideration of the trajectories of system ( [ quasiclassicalevolution ] ) and does not include many other essential effects which can be important in real metals . at the end of this section
, we would like to note that the picture described above is based on the assumption that the fermi surface of a metal has a complicated form .
this assumption means , in particular , that the carriers of the stable open trajectories have the form defined by general topological theorems for generic dispersion law .
this form will be described in the next section and has some special features for generic fermi - surfaces @xmath197 . however , for special simple fermi - surfaces the carriers of open trajectories can have much simpler form , which will result also in simpler behavior of conductivity in comparison with the described above . as an example of the simple fermi - surface
, we can consider just a pair of slightly deformed integral planes in the @xmath10 - space ( like for some classes of organic metals ) .
it is easy to see that the exact mathematical stability zone @xmath198 can be identified here with the whole sphere @xmath199 and coincides with the `` experimentally observable '' stability zone @xmath200 . as a result
, we can not observe here the effects , connected with the reconstruction of the open trajectories , typical for the zones @xmath110 .
the `` rational peaks '' in the conductivity can be observable here for the surfaces having essential deformations ( see e.g. @xcite ) , however , they decrease exponentially with the growth of the numbers @xmath201 . as a result , we can expect here the experimental evidence of just a finite number of the lines @xmath202 instead of the complicated angle dependence of conductivity , observable in the zones @xmath75 .
it is easy to see also , that for the case of vanishing amplitude of deformation of the planes the conductivity tensor demonstrates just a very fast approach to its limiting form given by formula ( [ 3dimanisotropiclimit ] ) . in the next sections we will make more detailed consideration of the effects described above using the kinetic approach to the transport phenomena . in general ,
all the statements above can be obtained within the @xmath203 - approximation for the kinetic equation for the electron gas .
let us start with the description of generic picture arising on the fermi surface in the case of presence of the stable open trajectories at @xmath204 ( @xcite ) .
everywhere below we will assume that the fermi surface represents a smooth non - selfintersecting 3-periodic surface in @xmath10 - space .
consider the general picture of intersection of a 3-periodic surface in @xmath10 - space by the planes orthogonal to @xmath22 ( fig .
[ generalsurface ] ) .
first , let us look at the closed trajectories in @xmath10 - space if they present on the fermi surface .
the non - singular closed trajectories in @xmath10 - space are always locally stable and can be combined into connected cylinders consisting of trajectories of this type . in the case when a connected component of the fermi surface does not consist of the closed trajectories only ,
these cylinders have finite heights and are bounded by singular closed trajectories ( fig . [ cylinderclosedtr ] ) .
let us make now the following reconstruction of the fermi surface : we remove all the cylinders of closed trajectories from the fermi surface and fill the created holes by 2-dimensional discs orthogonal to @xmath22 ( fig .
[ reconstruction ] ) . .
] it is easy to see that the reconstructed fermi surface carries the same open electron trajectories as the initial one , so we can investigate the behavior of open trajectories on the reconstructed surface .
the reconstructed fermi surface represents a 3-periodic surface which contains in general infinitely many connected components in @xmath10 - space .
one important theorem can be formulated about the connected components of the reconstructed fermi surface in the case of presence of stable open trajectories ( @xcite ) : every connected component of the reconstructed fermi surface carrying open electron trajectories represents a periodic deformation of an integral plane in @xmath10 - space .
let us remind here that the plane @xmath205 is called integral if it is generated by any two reciprocal lattice vectors .
so , every connected component of the reconstructed fermi surface has two independent periods represented by two vectors of the reciprocal lattice .
let us note also that the picture described above is evidently stable with respect to all small rotations of @xmath22 view the local stability of nonsingular closed trajectories on the fermi surface .
all the connected components of the reconstructed fermi surface are parallel to each other which leads to the coinciding of the mean directions of all stable open trajectories in @xmath10 - space .
the last property was called in @xcite the topological resonance and plays very important role in the experimental observation of the anisotropic regimes of conductivity behavior and of the topological numbers corresponding to a given stability zone . in this paper
we will use the picture described above as the general picture arising in the case of presence of stable non - closed electron trajectories on the fermi surface .
let us note here that this picture can have some additional features for some special directions of @xmath22 within the stability zone ( see @xcite ) which can be also detected from the experimental point of view ( @xcite ) .
these phenomena , however , have a non - generic character and will not be considered here .
after the factorization over the vectors of the reciprocal lattice every connected component of the reconstructed fermi surface becomes a two - dimensional torus @xmath206 embedded in @xmath207 .
the number of non - equivalent components in @xmath21 is always finite and represents an even number .
the open ( in @xmath208 - space ) trajectories of system ( [ quasiclassicalevolution ] ) on the fermi level are given by the intersections of the components of the reconstructed fermi surface with the planes orthogonal to @xmath22 and represent irrational coverings of the corresponding tori @xmath206 ( except the discs @xmath209 ) in generic situation ( fig .
[ trajectont2 ] a ) .
however , for special directions of @xmath22 the intersection of the plane orthogonal to @xmath22 with the integral plane @xmath60 ( @xmath210 ) can have rational ( integral ) direction , which means in fact that the corresponding trajectories of ( [ quasiclassicalevolution ] ) become periodic in @xmath208 - space .
the corresponding trajectories in @xmath21 become now closed curves on the tori @xmath206 and are not dense anymore on the sets @xmath211 , which can be called the `` carriers of open trajectories '' on the fermi surface ( fig .
[ trajectont2 ] b ) . and closed ( periodic in @xmath208 - space ) trajectory on the same manifold after the factorization over the reciprocal lattice vectors . ] as we can see ,
the generic and the periodic open trajectories demonstrate an obvious difference in their statistical properties on the `` carriers of open trajectories '' . as will see below
, this circumstance will play important role for the behavior of the conductivity tensor @xmath188 in the limit @xmath42 .
we can see here that the periodic trajectories can not appear for completely irrational directions of @xmath22 since the plane orthogonal to @xmath22 should contain at least one vector of the reciprocal lattice , giving the mean direction of such trajectories .
let us note that the class of rationality of a direction of magnetic field @xmath22 is defined as the number of linearly independent reciprocal lattice vectors lying in the plane orthogonal to @xmath4 . in particular , the direction @xmath212 , orthogonal to the plane @xmath213 , represents a purely rational direction according to this definition . in general ,
it is not difficult to see that the directions of @xmath22 in a stability zone @xmath108 , corresponding to periodic open trajectories , form a dense net of one - dimensional curves , which do not intersect each other in @xmath75 in the case when the direction @xmath214 does not belong to @xmath75 ( fig .
[ densenet ] a ) or all intersect at the point @xmath215 if @xmath216 ( fig .
[ densenet ] b ) . as we already said above , all the one - dimensional curves @xmath102
should be actually extended outside the stability zone @xmath75 since the periodic open trajectories exist in fact for a wider set of directions of @xmath22 surrounding @xmath108 .
the periodic open trajectories become unstable with respect to small rotations of @xmath22 outside @xmath108 .
we have to note also that the difference between the curves @xmath102 and @xmath105 decreases with increasing of the `` denominator '' of a rational mean direction of the periodic trajectories and vanishes in the limit @xmath217 ( fig .
[ diroutstabzone ] ) .
let us consider now the carriers of open trajectories in more detail .
the general form of connected components of the reconstructed fermi surface is shown at fig .
[ periodicplane ] and represents a periodically deformed integral plane in @xmath208 - space with two periods @xmath70 , @xmath71 , @xmath17 given by two independent reciprocal lattice vectors . - space with periods @xmath218 . ]
the carriers of open trajectories in @xmath208 - space represent the same periodically deformed integral planes without `` artificial '' two - dimensional discs @xmath209 , attached after the removal of closed trajectories .
let us exclude here the non - generic case when the magnetic field @xmath22 is orthogonal to the plane @xmath60 and consider the generic case @xmath219\ ] ] we will assume in our considerations that the @xmath220 - axis is always chosen along the direction of @xmath22 while the @xmath145 - @xmath17 and @xmath221 - axis are orthogonal to @xmath4 . in the generic situation described above the values of @xmath222 separate all different trajectories at a given component and @xmath222 can be chosen as a `` coordinate '' on the corresponding carrier of open trajectories .
the second coordinate on the carriers of open trajectories is naturally given by the parameter @xmath223 , measured along the trajectories according to system ( [ quasiclassicalevolution ] ) .
in general , a natural system of local coordinates in @xmath208 - space ( or @xmath207 ) is given by the triple @xmath224 if we consider system ( [ quasiclassicalevolution ] ) for all values of @xmath225 .
the coordinates @xmath224 are naturally connected with system ( [ quasiclassicalevolution ] ) and possess also one more remarkable property : @xmath226 the last property is caused in fact by the hamiltonian properties of system ( [ quasiclassicalevolution ] ) , which can be considered as a hamiltonian system with the hamiltonian @xmath5 and the poisson bracket @xmath227 ( @xmath228 ) .
it can be seen also that the integration over the layer restricted by the energy values @xmath229 and @xmath230 can be represented by the integration over the energy surface @xmath231 according to the formula @xmath232 formula ( [ layerintegration ] ) gives in fact very convenient way of statistical averaging of any value over the fermi distribution in presence of magnetic field .
it s not difficult to see that every connected component of the reconstructed fermi surface can be divided into a periodic set of the fundamental ( minimal ) domains identical to each other ( fig .
[ periodicset ] ) .
it can be seen also that the fundamental domains can be chosen in the form of `` curvilinear '' parallelograms with the sides , corresponding to the vectors @xmath233 and @xmath71 .
the division of the connected components of the reconstructed fermi surface gives also a natural division of the corresponding carriers of open trajectories into fundamental domains .
all the fundamental domains are physically equivalent to each other , so every domain represents in fact the full set of the physical states represented by the whole component . the two - dimensional tori @xmath234 , which we discussed above , can be considered also as any of the fundamental domains with the identified opposite sides . as we said already
, every nonsingular open trajectory becomes then a smooth curve on the torus @xmath206 which can be open or closed ( periodic in @xmath10 - space ) depending on the direction of @xmath22 ( fig .
[ trajectont2 ] ) .
the difference between the closed and the non - closed trajectories on the tori @xmath206 causes an evident difference in the averaging of any of the physical quantities over time on the trajectories of this kind .
thus , we can claim that any time average over generic ( non - closed on @xmath235 ) trajectory coincides with the averaging of the corresponding quantity over the carrier of open trajectories , which depends smoothly on the direction of @xmath22 inside the `` stability zone '' . on the other hand ,
the time average over a periodic ( closed on @xmath235 ) trajectory is in general different from the mean value of the same quantity on the carrier of open trajectories .
as can be seen , the last feature can lead to irregular dependence of the conductivity tensor on the direction of @xmath22 inside @xmath108 .
let us consider now the standard kinetic approach ( see e.g. @xcite ) to the magneto - transport phenomena , using the geometric picture described above .
the full dynamical system for the adiabatic evolution of the electron states both in presence of magnetic and electric fields can be written as @xmath236 \,\,\ , + \,\,\ , e \ , { \bf e}\ ] ] in system ( [ fulldynsyst ] ) the electric field @xmath237 will be considered as a small value while the value of magnetic field is supposed to satisfy the strong magnetic field limit condition @xmath238 . in the kinetic approach
we have to introduce the single - particle distribution function @xmath239 satisfying the boltzmann equation @xmath240^{l } \,\ , { \partial f \over \partial p^{l } } \,\,\ , + \,\,\ , e \ , \sum_{l=1}^{3 } \ , e^{l } \ , { \partial f \over \partial p^{l } } \,\,\,\ , = \\ = \,\,\,\ , i [ f ] ( { \bf p } , \ , t ) \quad , \end{gathered}\ ] ] where @xmath241 \ , $ ] represents the collision integral . here
we will be interested in the stationary solutions of the equation ( [ kineq ] ) , so we put now @xmath242 . in the absence of electric field
the distribution function @xmath243 is given by its equilibrium values @xmath244 \ , + \
, 1}\ ] ] and we have identically @xmath241 ( { \bf p } ) \ , \equiv \ , 0 $ ] .
the conductivity tensor @xmath188 is defined by the linear in @xmath237 correction to the function @xmath245 , satisfying the equation @xmath246^{l } \,\ , { \partial f_{(1 ) } \over \partial p^{l } } \,\,\ , + \,\,\ , e \ , \sum_{l=1}^{3 } \ , e^{l } \ , { \partial f_{0 } \over \partial p^{l } } \,\,\,\ , = \\ = \,\,\,\ , \left [ { \hat l}_{[f_{0 } ] } \ , \cdot \ , f_{(1 ) } \right ] ( { \bf p } ) \quad , \end{gathered}\ ] ] where @xmath247 } \ , $ ] is the linearization of the functional @xmath241 ( { \bf p } ) \ , $ ] on the corresponding function @xmath248 . the value @xmath247 } \ , \cdot \ , f_{(1 ) } \ , $ ] can be considered as a term , connected with the relaxation of non - equilibrium perturbations to the equilibrium state . in the so - called @xmath249 - approximation the right - hand part of system ( [ linkineq ] )
can be replaced by the value @xmath250 , where @xmath39 plays the role of the characteristic relaxation time .
it will be convenient here to use the @xmath249 - approximation for the equation ( [ linkineq ] ) , which is enough to catch all the main features of our consideration .
the characteristic relaxation time @xmath39 is usually identified with the mean free electron motion time .
let us introduce now the variable @xmath251 , where @xmath252 is the parameter along the trajectories of system ( [ quasiclassicalevolution ] ) , introduced above , and consider the coordinate system @xmath253 in the @xmath208 - space .
the parameter @xmath254 has a purely geometrical meaning in @xmath208 - space and does not depend on the value of @xmath125 . according to ( [ dpzdtdl ] ) - ( [ layerintegration ] ) , we can naturally write @xmath255 @xmath256 it s not difficult to see also , that under our assumptions above the equation on the function @xmath257 can be represented as @xmath258 in the coordinate system @xmath259 . after the substitution @xmath260 we get the linear systems @xmath261 for the functions @xmath262 .
the solutions of systems ( [ gvsyst ] ) which we need can be written in the form @xmath263 where the integration is taken along the whole part of a trajectory preceding the point @xmath259 .
the mean value of the density of the electric current is given by the formula @xmath264 ( in the linear approximation in @xmath265 ) .
the form ( [ f1subst ] ) of the correction @xmath266 gives in fact very strong concentration of @xmath267 near the fermi surface for most of the normal metals , so the term @xmath268 can be actually replaced by the delta - function @xmath269 in the integration over @xmath270 .
finally , we come to the following simple expression for the value of the electric current density : @xmath271 the general form of the conductivity tensor @xmath135 can be represented as @xmath272 at the same time , the contribution of the open trajectories @xmath273 is given by the restriction of the integral in ( [ gensigmakl ] ) to the set of the carriers of open trajectories @xmath274 instead of the whole fermi surface , so we can write : @xmath275 let us say , that the irregular behavior of @xmath188 in @xmath75 is completely determined by the functions @xmath276 , so we will be interested here mostly in the behavior of @xmath276 in the limit @xmath34 .
it s not difficult to see that for all types of trajectories in our case we can write in the limit @xmath33 for the conductivity tensor @xmath277 where @xmath278 is the mean value of the group velocity component on the trajectory passing through the point @xmath279 : @xmath280 to get the contribution of the open trajectories we have now to restrict the integration to the set of the carriers of open trajectories @xmath274 on the fermi surface , so we can write for the contribution of the stable open trajectories to the conductivity in the limit @xmath33 : @xmath281 let us note here that according to our choice of coordinate system @xmath282 we will always have @xmath283 and @xmath284 , @xmath285 for the stable open trajectories .
we can see also , that according to the definition above , the values @xmath286 do not depend in fact on the variable @xmath287 for the trajectories of our type . as a corollary ,
the first multiplier can be also replaced by its mean value on every trajectory of system ( [ quasiclassicalevolution ] ) . finally , we can write for the contribution of the stable open trajectories to the magneto - conductivity for @xmath33 : @xmath288 let us note that the tensor @xmath67 is purely symmetric according to ( [ genformsigmakl ] ) , which is not the case for finite values of @xmath125 .
let us give now qualitative derivation of the regimes of the conductivity behavior described in the previous section .
as it is not difficult to see , in the case of generic direction of @xmath22 ( non - periodic open trajectories ) the mean values @xmath289 can be identified with the mean values of the same quantities on the corresponding ( connected ) carrier of open trajectories @xmath290 : @xmath291 and are constant on each carrier of the stable open trajectories . in general , the mean values @xmath292 can be also expressed through the values @xmath289 by the formula @xmath293 let us introduce now the functions @xmath294 where the summation is taken over all the connected carriers @xmath290 of the stable open trajectories .
the functions @xmath295 are smooth functions of the direction of @xmath22 inside any stability zone @xmath108 . according to ( [ genformsigmakl ] ) , the functions @xmath295 coincide with the limiting values @xmath296 for generic directions of @xmath22 , corresponding to non - periodic form of the open trajectories . at the same time , for special directions of @xmath22 , corresponding to the periodic form of the stable open trajectories , the values ( [ genformsigmakl ] ) do not coincide with @xmath295 due to the dependence of the values @xmath289 on @xmath297 . just from the schwartz inequality we can write in this case
@xmath298 @xmath299 as it was written above .
the values @xmath300 can have in general arbitrary signs .
we have also @xmath301 if @xmath302 or @xmath303 is equal to @xmath304 , in our coordinate system .
let us point out an important role of the geometry of the carriers of open trajectories on the complicated fermi surfaces for the effect which we discuss here .
as it is not difficult to see , the difference @xmath305 is defined actually by the `` inconstancy '' of the functions @xmath289 on the carriers of the stable open trajectories . as can be seen at fig .
[ difftraject ] , the presence of the singular trajectories on the carriers of open trajectories makes the behavior of @xmath289 rather irregular in the case of periodicity of the open trajectories on the fermi surface .
thus , the periodic trajectories adjacent to the singular trajectory from different sides have rather different geometry in @xmath10 - space . as a result
, we can expect that the mean variations of the values @xmath289 on the carrier of open trajectories can be comparable with the mean values of the same quantities and the same can be stated also about the difference @xmath306 @xmath307 .
it should be also noted here , however , that this effect decreases with increasing of `` denominators '' of the rational mean direction of the open trajectories due to the increasing of the density of the periodic open trajectory on a carrier of open trajectories . as we also said
already , we have the identities @xmath308 in our coordinate system . to evaluate the dependence of @xmath309 on the numbers @xmath310 it is convenient to choose the fundamental domain on the carrier of open trajectories in the form of curvilinear parallelogram with two opposite sides , represented by two equivalent parts of periodic open trajectories ( fig .
[ newfunddom ] ) .
the values @xmath222 and @xmath254 give a good parametrization of the fundamental domain under the requirement ( [ nonparallel ] ) .
we can assume without loss of generality that @xmath22 is not orthogonal to the vector @xmath311 which will also mean that the mean direction of the open trajectories does not coincide with the direction of @xmath71 .
let us put also that the values of @xmath222 belong to the interval @xmath312 \ , $ ] for our fundamental domain , where @xmath313 .
the parameter @xmath254 is not continuous in our fundamental domain and has singularities on the levels of @xmath222 , containing singular points within the domain .
as we will see , this circumstance will not actually be important in our considerations .
let us put that the values of @xmath254 are restricted by the values @xmath314 and @xmath315 at a given value of @xmath222 for our fundamental domain .
the functions @xmath314 and @xmath315 also have singularities at the same @xmath222 , defined by the singular behavior of @xmath316 : @xmath317 have asymmetric character since the corresponding trajectories approach the singular point once or twice depending on the sign of @xmath318 .
let us note here that the last circumstance played important role in the proof of the fact of mixing in the dynamical systems analogous to those considered here .
this property was proved in paper @xcite where systems of this kind were considered in connection with a different problem of hamiltonian dynamics ( @xcite ) . every trajectory with the period @xmath319
is represented in our domain by @xmath320 separate lines , giving a closed curve after the identification of the equivalent points in the @xmath208 - space .
the values of the functions @xmath321 are the same on every set of @xmath320 components representing one periodic trajectory .
the presence of singular points plays also important role in the global motion of electrons along the non - singular trajectories since they appear many times at rather small distances from a periodic trajectory with a large period @xmath322 . as a result ,
the length @xmath323 of a periodic trajectory in the parameter @xmath254 ( @xmath324 ) has essential fluctuations , caused by the presence of singular points in the fundamental domain .
let us note , that according to system ( [ quasiclassicalevolution ] ) we can write in our coordinate system @xmath325 where @xmath326 is the length of the corresponding topological cycle in the @xmath10 - space , depending just on the numbers @xmath327 .
in general , we have @xmath328 singular points @xmath329 \ , $ ] of the function @xmath330 , where @xmath331 is the number of ( non - equivalent ) singular points of system ( [ quasiclassicalevolution ] ) on the carrier of open trajectories .
the function @xmath330 represents a periodic function with the period @xmath332 and integrable ( asymmetric ) singularities @xmath333 $ ] , restricted by two subsequent `` singular '' values of @xmath222 ( @xmath334 ) , we can write the contribution of the singularities at the points @xmath335 and @xmath336 to the value of @xmath330 in the form @xmath337 ( @xmath338 ) . besides that , the singularities at the points @xmath339 and @xmath340 , close to the interval @xmath341 , play also in fact important role for the variations of @xmath330 inside the interval @xmath342 for big values of @xmath343 and @xmath344 . the corresponding contribution to @xmath330
can be written as a sum of terms analogous to ( [ singularcontr ] ) , which number is proportional to @xmath345 .
we actually will be interested here in the fluctuations of @xmath330 caused by these singularities , so only the variation of this contribution with the variable @xmath222 is important for us .
it s not difficult to show also that the difference of this contribution at the points @xmath335 and @xmath336 has the order @xmath346 . for our purposes here
( we omit here rigorous estimations ) the corresponding contribution @xmath347 to the function @xmath330 can be actually approximated by a linear function @xmath348 ( @xmath349 ) , on every interval @xmath341 .
the functions @xmath350 and @xmath347 represent the main corrections to the `` basic value '' @xmath351 of the function @xmath352 , which are responsible for the main terms in the dependence of @xmath353 on the numbers @xmath310 ( or @xmath354 ) . the basic value @xmath351 of the function @xmath330 is given by a constant function on the whole interval @xmath312 \ , $ ] , which in the main order is proportional to the length @xmath326 ( or @xmath355 ) of the topological cycle @xmath356 in the @xmath10 - space . easy to see that for any sequence of rational directions in @xmath75 converging to some fixed ( irrational ) direction we can write in the main order @xmath357 .
the functions @xmath358 , @xmath359 are continuous functions of @xmath297 , however , they are not smooth on the curves , representing singular trajectories .
thus , for the component @xmath360 we can write the following relations @xmath361^{-1 } \,\,\ , \rightarrow \,\,\ , 0\ ] ] near these curves . in general , the values @xmath358 represent a periodic function of @xmath222 with the period @xmath332 and very narrow cusps at the values of @xmath222 , corresponding to singular electron trajectories ( fig .
[ graphvlgr ] ) .
it is easy to see also that the number of cusps on the period of the function @xmath358 is equal to the number @xmath331 of non - equivalent singular points of system ( [ quasiclassicalevolution ] ) on the corresponding carrier of open trajectories . on a carrier of open trajectories for the case of rational mean direction @xmath319 of the trajectories . ] in our situation we can put approximately @xmath362 and @xmath363 on every interval @xmath341 .
let us say that the approximation above should be used in our scheme also for the calculation of the approximate values of @xmath364 according to formula ( [ carriertrajavconn ] ) . using this approximation to compute the values @xmath365 for @xmath366 we finally get the relations @xmath367 to give the necessary approximation for the function @xmath368 we have to evaluate also the integral @xmath369 over a periodic trajectory .
first , we should note , that we have to prescribe here also the `` basic value '' @xmath370 to this integral , which is constant on the interval @xmath312 $ ] .
the essential corrections to the value @xmath371 on the intervals @xmath342 are caused again by the presence of singular points and can be represented by two main contributions : \1 ) the contribution of the endpoints of @xmath341 , which can be approximately written as @xmath372 where @xmath373 and @xmath374 are the values of @xmath375 at the singular points , adjacent to the corresponding singular trajectories ; \2 ) the contribution of other singular points close to the interval @xmath341 , which can again be approximated by the linear function @xmath376 where @xmath377 represents some weighted average of the values @xmath373 over all the singular points in the fundamental domain .
we can use now the following approximation for the function @xmath378 : @xmath379 we can see here that the function @xmath378 does not have cusps at the values @xmath335 and the values @xmath380 are equal to @xmath381 .
however , the structure of the function @xmath378 still has many common features with the structure of the function @xmath382 . in particular , using the approximation ( [ vzgravappr ] ) in the formula ( [ sigmaklformula ] ) , we obtain actually the same relations @xmath383 for @xmath384 , which implies also relations ( [ sigmasigmabardiff ] ) for a given stability zone @xmath108 .
the arguments used above can be applied also to the estimation of the functions @xmath153 , @xmath170 , @xmath154 in relations ( [ sigmaomega ] ) .
let us say at once , that we will not present here rigorous calculations , however , the approximate considerations below capture all the essential features which are necessary for our purposes .
thus , for the estimation of the functions ( [ gengl ] ) we can use the approximation when the integration in ( [ gengl ] ) is taken just over a part of a trajectory having length @xmath385 in the parameter @xmath386 .
in the same way , we can use for the averaged values @xmath387 the approximation @xmath388 let us take again the fundamental domain in the form shown at fig .
[ newfunddom ] .
the open trajectories have now irrational mean direction and are not periodic in the @xmath10 - space anymore .
the corresponding parts of open trajectories are represented by nets of approximately @xmath389 curves in the fundamental domain with approximate distances @xmath390 between the curves .
the functions @xmath391 are almost constant along the curves @xmath392 in the fundamental domain and change for the most part along the coordinate @xmath297 . for generic mean
direction of open trajectories the functions @xmath393 do not have periods smaller than @xmath394 in the fundamental domain , however , in other essential features the behavior of @xmath393 in our approximation is similar to the above behavior of the functions @xmath395 , considered for a rational mean direction of open trajectories with @xmath396 . in particular , we can also introduce here the values @xmath335 in the fundamental domain , defined by the requirement of presence of a singular point on the preceding part of the corresponding trajectory at a distance not exceeding the value @xmath397 in the @xmath10 - space .
the number of the values @xmath335 in the fundamental domain has the order @xmath398 for @xmath399 and we can consider again the division of the interval @xmath312 \ , $ ] to the segments @xmath342 analogous to those considered above .
the similar analysis shows also that the behavior of the functions @xmath393 demonstrates on the intervals @xmath342 the features analogous to those observed in the behavior of @xmath400 for a rational mean direction of the open trajectories with the numbers @xmath401 , @xmath344 having the order @xmath402 .
let us introduce now the symmetric and the anti - symmetric parts of the conductivity tensor @xmath403 according to the onsager relations we have the identities @xmath404 after some elementary calculations we can get from the formula ( [ deltasigmab ] ) : @xmath405 @xmath406 it is not difficult to see that for any irrational mean direction of the open trajectories the values @xmath407 , @xmath408 coincide with the values @xmath292 , @xmath409 in the limit @xmath410 .
at the same time , for large but finite values of @xmath80 the behavior of the functions @xmath411 is analogous to the behavior of @xmath412 for a rational mean direction of the open trajectories with @xmath413 . as a result , in full analogy with formula ( [ sigmaklformula ] )
, we can write the relations @xmath414 for generic directions of the open trajectories .
it is easy to see that the estimation ( [ lnomegatau ] ) gives now the relations ( [ trend ] ) for the functions @xmath160 and @xmath154 . at the same time , the relations ( [ rbtrend ] ) for the function @xmath159 are defined by the behavior of the anti - symmetric part of the tensor @xmath56 . using the same approximations for the functions @xmath411
, we get now the estimation @xmath415 using also the identity @xmath416 we then get relations ( [ rbtrend ] ) for the function @xmath159 .
let us say that a unified description of the function @xmath417 can not be actually obtained in general situation , since its the behavior admits really essential variations depending on a concrete form of the fermi surface .
as we said already , the relations ( [ lnomegatau ] ) can be considered just as a `` general trend '' in the behavior of conductivity in strong magnetic fields .
let us give here a more detailed explanation of our statement .
as we can see , the analogy between the relations ( [ lnk1k2 ] ) and ( [ lnomegatau ] ) requires in fact the assumption of the `` uniform covering '' of the carrier of open trajectories by parts of the trajectories having length @xmath418 .
this assumption is fulfilled `` in average '' , however , it can be violated for concrete values of @xmath55 for special directions of the open trajectories .
it is not difficult to see , that this violation definitely takes place when the direction of the open trajectories can be approximated by a rational direction with a very high precision ( @xmath419 ) . as we can see , the corresponding open trajectories will then be very close to the periodic trajectories on the lengths @xmath420 so the parameter @xmath320 should be actually used instead of @xmath55 in ( [ lnomegatau ] ) up to the values @xmath421 . as a result ,
the general trend ( [ lnomegatau ] ) can have essential corrections . for better description of the behavior of @xmath56 for generic mean directions of the stable open trajectories
let us just say some words about the problem of approximation of irrational numbers by rational ones . in our case we have to investigate the possibility of rational approximations of the number @xmath157 in relation ( [ dbarkappa ] ) , which correspond to rational approximations of the mean direction of the open trajectories in the plane @xmath62 .
according to classical ( dirichlet ) theorem we can state that for any number @xmath422 there exist two integer numbers @xmath401 , @xmath423 : @xmath424 such that @xmath425 let us note here that we can put without loss of generality @xmath426 , so we will assume also @xmath427 for precise approximations of @xmath428 . for the value @xmath429 we can rewrite the relation ( [ dirrelation ] ) in the form @xmath430 which means that the deviation of the open trajectory from the corresponding periodic trajectory does not exceed the value @xmath431 on the length @xmath432 .
the relation ( [ trajappr ] ) can be considered as the first step in the approximation of irrational directions of open trajectories by rational ones since the mean distance between the curves representing a periodic trajectory in the fundamental domain is of order of @xmath433 we have to say , however , that relation ( [ trajappr ] ) does not give a precise connection between the values @xmath55 and @xmath401 , so this connection should be specially studied for every concrete irrational number @xmath428 . using the theorem of dirichlet , we can see that for any irrational number @xmath157 there always exists an infinite sequence of rational approximations @xmath434 such that the following relations @xmath435 take place for every @xmath287 .
if we consider now a sequence @xmath436 of the values of @xmath125 , such that @xmath437 , we get the relations @xmath438 analogous to ( [ trajappr ] ) , for this sequence .
it can be seen , that the estimation above should not cause in general crucial deviations from the general trend according to our picture .
let us say , however , that the estimation ( [ quadraticappr ] ) can be actually improved if we consider `` generic '' irrational numbers @xmath157 , representing a set of the full measure in @xmath439 .
for example , the following theorem can be formulated for generic @xmath157 : theorem ( see e.g. @xcite ) : let @xmath440 be a positive function of positive @xmath441 , such that the function @xmath442 is non - increasing . then for almost all irrational @xmath157 the inequality @xmath443 has infinitely many solutions @xmath444 if for some positive @xmath445 the integral @xmath446 is divergent .
the theorem above gives an improvement of the estimation ( [ quadraticappr ] ) for `` almost all '' irrational numbers @xmath157 in the sense of the standard measure in @xmath447 .
easy to see , that the set of the functions @xmath440 , which can be used for the improvement of the estimation ( [ quadraticappr ] ) , is rather big .
for example , we can put @xmath448^{-1 } \quad , \quad \quad x \
, > \ , 1 \quad , \ ] ] @xmath449^{-1 } \quad , \quad \quad x \ , > \ , e \quad , \ ] ] etc .
in particular , using the function @xmath450 , we get for the corresponding sequence of @xmath436 the estimation @xmath451 ( @xmath452 ) .
the estimation above shows that the general trend ( [ trend ] ) for the functions @xmath153 , @xmath154 should have in fact an additional structure represented by the presence of `` plateaus '' near the values @xmath436 . as we said already
, the estimation ( [ betterest ] ) can be used for `` almost all '' irrational numbers , characterizing the mean direction of open trajectories . for special irrational numbers these estimations
can be much stronger and we should say in this case that the corresponding number has `` very good '' approximations by rational numbers .
as we can see then , the general trend in the behavior of @xmath173 @xmath453 can be understood as an implementation of the relations ( [ trend ] ) for the `` reference values '' of @xmath80 , which represent an infinite sequence on the interval @xmath454 . at the same time ,
near the values @xmath436 the behavior of @xmath173 demonstrates some kind of `` plateaus '' where the local dependence on @xmath80 is actually different from the general trend . as a result , the global behavior of @xmath173 , @xmath455 demonstrates a kind of `` cascade structure '' defined by the properties of the irrationality @xmath157 .
let us note also that the cascade structure is more distinct for the numbers @xmath157 , having better rational approximations , and is smoothed for generic irrational @xmath157 . according to the above remark
, we can also see that the trend ( [ trend ] ) has in some sense `` conventional '' character and can be also considered just as a convenient choice among close possible descriptions of the asymptotic behavior of conductivity .
we have to note also that in the special case when the carriers of open trajectories do not contain holes with singular points the relations ( [ trend ] ) should be actually changed to the simpler trend @xmath456 as it is not difficult to see , all the remarks above apply also to the estimation ( [ skewlnomegatau ] ) , which can be violated in the same situations as ( [ lnomegatau ] ) . for this reason , we should consider the relation ( [ rbtrend ] ) just as a `` reliable '' restriction on the function @xmath159 , defined by relation ( [ sigmaomega ] ) . at the end of this section ,
let us consider the behavior of the components @xmath56 , where @xmath457 or @xmath458 is equal to @xmath304 .
let us note here that relations ( [ sigmaomega ] ) represent in this case the main terms in the behavior of these components which determines the difference between the functions @xmath149 , @xmath150 , @xmath459 and @xmath153 , @xmath170 , @xmath154 . according to system ( [ quasiclassicalevolution ] )
the integral @xmath460 represents the value @xmath461 and is a bounded function in our coordinate system .
an elementary estimation of the integral ( [ vaveragedb ] ) gives now the main term @xmath462 in the asymptotic behavior of the function @xmath463 , which can be used in the formulae ( [ symmetricpart ] ) - ( [ antisympart ] ) .
in particular , we easily get from ( [ symmetricpart ] ) the relation @xmath464 which , being added also with the corresponding term of ( [ sigmaregexpclosed ] ) , gives the first part of the relation ( [ asymptbehavior ] ) .
let us note now that the value @xmath463 demonstrates also the property @xmath465 ( view the same property of @xmath466 ) . using the representation @xmath467 @xmath468 for generic mean direction of the open trajectories in the formula ( [ symmetricpart ] )
, we then get the relations @xmath469 for the symmetric part of @xmath470 .
we can see then , that the main terms in the asymptotic behavior of the values @xmath471 , @xmath472 , @xmath473 are defined by the anti - symmetric part of the conductivity tensor . to estimate the corresponding terms let us note now that the estimation ( [ v1grrel ] ) is also valid for the function @xmath474 for our type of open trajectories .
using now the change of the order of integration in formula ( [ antisympart ] ) we can write @xmath475 @xmath476 , which gives the necessary estimation for the functions @xmath477 . adding also the anti - symmetric part of the tensor ( [ sigmaregexpclosed ] )
, we get finally the remaining part of the relation ( [ asymptbehavior ] ) .
let us note also that the higher corrections to the components @xmath471 , @xmath472 have more complicated behavior for generic open trajectories in @xmath208 - space but usually are not considered in applications .
let us say here , that the structure of @xmath56 described above has in some sense abstract theoretical character and probably can not be observed in all features in most of the experiments .
however , we expect that the basic elements of the picture described above still can be detected in special studies of the conductivity behavior in strong magnetic fields . in general
, we can expect that in the most of the real experiments it can be just established that the values @xmath478 , @xmath455 demonstrate some general decreasing with the increasing of @xmath80 with not uniquely specified decreasing law .
quite often it is convenient to approximate the behavior of @xmath56 by some intermediate powers @xmath479 of the parameter @xmath55 , however , we should admit an irregular dependence of the parameter @xmath195 on the direction and the value of @xmath22 in this case .
let us say also that the right - hand parts in ( [ lnomegatau ] ) - ( [ skewlnomegatau ] ) can arise with big dimensionless coefficients having geometric origin , so for complicated fermi surfaces the interplay between the values @xmath480 and @xmath481 , @xmath159 , @xmath482 can be important even for rather big values of @xmath55 .
let us say that the questions considered above are connected mostly with the geometry of the stable open trajectories .
in the next section we will consider the aspects of the conductivity behavior connected with the reconstruction of the trajectories of this type .
to explain the behavior of conductivity in the zone @xmath110 we have to consider the reconstruction of the carriers of open trajectories after crossing the boundary of the zone @xmath75 on the unit sphere .
in general , the most essential features of such reconstruction can be demonstrated by the following simplified scheme : let us imagine the fermi - surface in @xmath483 in the form of an infinite set of integral planes ( with holes ) connected by parallel cylinders of finite heights ( fig .
[ schemfermisurf ] ) . it is assumed that all the planes are divided into the `` even - numbered '' and the `` odd - numbered '' planes , which gives also the separation of the cylinders into two different classes .
we will assume here , that all the planes of a given class represent the same object after the factorization over the reciprocal lattice vectors and the same is assumed also about the cylinders of a given class .
thus , the physical fermi - surface represents here two parallel two - dimensional tori ( with holes ) @xmath234 connected by two cylinders in @xmath207 . .
] it is easy to see that for those @xmath4 , which are almost parallel to the axes of the cylinders , the cylinders constructed above represent cylinders of closed electron trajectories , while the planes with holes play the role of carriers of open trajectories .
it can be seen also , that the open trajectories on the planes of different types represent the motion in two opposite directions , while two types of finite - sized cylinders carry closed trajectories of the `` electron type '' and the `` hole type '' respectively .
thus , we get a stability zone @xmath75 around the direction @xmath484 , parallel to the axes of the cylinders connecting the integral planes
. the cylinders of the closed trajectories almost coincide with the cylinders , connecting the planes , for directions of @xmath4 , close to @xmath485 , however , they become shorter and represent just parts of these cylinders for @xmath4 , deviating from the direction @xmath485 . the boundary of the zone @xmath75 is defined by the condition that the height of the cylinders of closed trajectories of one type ( say , electron type ) becomes zero , while the height of the cylinders of closed trajectories of another type remains finite . after crossing the boundary of @xmath75 the individual integral planes do not represent carriers of open trajectories anymore , however , we can still claim that near the boundary of @xmath75 the cylinders of closed trajectories of the second type cut our fermi - surface into pairs of connected integral planes in @xmath486 .
so , we can investigate here the behavior of trajectories of system ( [ quasiclassicalevolution ] ) separately on these separated parts of the fermi - surface , which gives actually rather simple description of the trajectories in @xmath486 : let us denote here by @xmath60 the integral plane , giving the common integral direction of the planes considered above , and by @xmath487 the plane orthogonal to @xmath4 .
it is not difficult to see that in the picture described above all the trajectories on our fermi - surface become closed if the intersection of @xmath60 and @xmath487 has irrational direction .
the appearance of the closed trajectories on every pair of connected planes can be considered as a result of a reconstruction of the open trajectories , which is caused by the `` jumps '' of the trajectories between two planes outside @xmath75 ( fig .
[ reconstrtraject ] ) .
it is easy to see that these closed trajectories will have very big length in the immediate vicinity of the boundary of @xmath75 due to the low probability of the `` jumps '' in this region . at the same time
, it can be seen also that for any fixed irrational direction of intersection of @xmath60 and @xmath487 all the open trajectories will necessarily pass through the reconstruction after the crossing the boundary of @xmath108 . on the angle diagram . ]
the situation is rather different when the intersection of @xmath60 and @xmath487 has a rational direction @xmath104 . in this case
just a part of open trajectories undergoes reconstruction after the crossing the boundary of @xmath75 by the direction of the magnetic field , while the other trajectories still remain open .
so , in this situation the long and the open periodic trajectories coexist on the integral planes , having different measure for different directions of @xmath4 .
it is not difficult to see also , that the open trajectories have the maximal measure near the boundary of @xmath75 and disappear at the endpoints of the curve @xmath488 . from a simple estimation of the integrals in ( [ gensigmakl ] ) we easily get that in the interval @xmath489 the contribution of the long closed trajectories to
the conductivity is similar to that of the stable open trajectories , so we get formula ( [ sigmalowbtau ] ) in this interval . to estimate the contribution of the trajectories shown at fig .
[ reconstrtraject ] in the limit @xmath490 let us just first note that we have in general the relations @xmath491 for this type of trajectories .
at the same time , we can see that the trajectory shown at fig .
[ reconstrtraject ] can be considered as consisting of two parts belonging to different former carriers of open trajectories with opposite values of @xmath492 and @xmath493 . as a result ,
the typical value of @xmath494 on the closed trajectories is rather small and is defined by the density of the trajectory on the ( former ) carriers in the brillouen zone .
the exact dependence of @xmath495 on the value @xmath163 can have in fact a complicated character , it s not difficult to see , however , that for our pretty rough definition of the function @xmath163 the simple approximation @xmath496 can be used .
it can be seen also that with the same accuracy we can also use the approximations @xmath497 for large but finite values of @xmath80 . at the same time , for the values @xmath498 we can use the relation @xmath499 for all @xmath500 . using the above relations , we then easily get from ( [ symmetricpart ] ) the following estimations @xmath501 for the contribution of the long closed trajectories to the symmetric part of the conductivity tensor .
let us note also here that the integration over a set of carriers of open trajectories @xmath502 in formula ( [ symmetricpart ] ) should be understood now as the integration over the part of the fermi surface @xmath503 occupied by the long closed trajectories .
let us remind now , that the tensor @xmath504 should be added also with the symmetric part of the short closed trajectories extracted from the relations ( [ sigmaregexpclosed ] ) . as a result
, the value @xmath505 will get in fact a finite contribution , which does not depend on the value of @xmath163 .
we can claim , however , that the conductivity behavior demonstrates here a partial suppression of the conductivity along the direction of @xmath22 , which is expressed by the formula ( [ sigmaprime33 ] ) .
to evaluate the contribution of the long closed trajectories to the anti - symmetric part of the conductivity tensor we need actually a more detailed consideration of the behavior of @xmath506 on the trajectories of this type .
thus , using formula ( [ antisympart ] ) in the form @xmath507 we can easily get the estimation @xmath508 from the relation ( [ vxlctest ] ) .
at the same time , the estimation of the value @xmath509 requires in fact a more detailed analysis of the formula ( [ lctansympart ] ) .
thus , the estimations ( [ vyzlctest ] ) can not be simply used in formula ( [ lctansympart ] ) to get the corresponding estimation of @xmath509 because of the specific form of the trajectories shown at fig .
[ reconstrtraject ] . indeed , as can be easily seen , the values @xmath510 and @xmath511 on the trajectories of this form are strongly correlated on big scales @xmath512 , so , their slow harmonics will in fact annihilate each other in the anti - symmetric part of @xmath56 .
more precisely , let us represent the values @xmath506 on the long closed trajectories in the form @xmath513 @xmath514 where @xmath515 it is not difficult to check then by direct calculation that the contribution of the long closed trajectories to the anti - symmetric part of the conductivity tensor can be written as @xmath516 where the integration over @xmath222 is taken along the heights of the cylinders of long closed trajectories .
let us note here that the net height of all non - equivalent cylinders of the long closed trajectories has the order of @xmath517 while the density of the frequencies @xmath518 is proportional to @xmath163 .
the correlation of the functions @xmath519 and @xmath520 on the trajectory means now that the coefficients @xmath521 and @xmath522 are phase correlated , i.e. have almost the same complex phases at small values of @xmath523 . in general , we can use the estimation @xmath524 in the interval @xmath525 at the same time , we can write for the same values of @xmath523 the relations @xmath526 for the trajectories of our type . as a result , we can write approximately @xmath527 in the interval @xmath528 . using the above estimations in the formula ( [ antisymmfourie ] ) we can get now the regular expansion for the value @xmath509 in the region @xmath490 , having the form @xmath529 where all @xmath530 have the order of @xmath531 . in particular , we can write in the main order @xmath532 finally , we can write the estimation @xmath533 for the contribution of the long closed trajectories to the anti - symmetric part of the conductivity tensor . easy to see also that after the addition of the contribution of the short closed trajectories we get finally the formula ( [ sigmahighbtauantisym ] ) for the anti - symmetric part of the conductivity tensor .
for the values @xmath534 formula ( [ gensigmakl ] ) gives some intermediate regimes between ( [ sigmalowbtau ] ) and ( [ sigmahighbtausymm ] ) - ( [ sigmahighbtauantisym ] ) which have in general some irregular form .
we do nt consider here these regimes in detail and just speak of the experimental approach to description of @xmath56 using intermediate powers of @xmath535 .
as we have already said , the corresponding powers of @xmath55 have in this case just a local character and are not well defined as stable characteristics of the conductivity behavior . the schematic fermi - surface considered above looks rather special from the general point of view .
however , it demonstrates all the main points of the behavior of open trajectories near the boundary of the stability zone @xmath75 which are important for us .
in general , the schematic picture represented above can be enriched by additional cylinders of closed trajectories , connecting the integral planes or just having a point as a cylinder base ( fig .
[ enrichedschemsurf ] ) .
still , the main features of the reconstruction of the carriers of open trajectories after crossing the boundary of @xmath75 for generic dispersion relation and the fermi level @xmath23 is given by the vanishing of the height of one particular cylinder of open trajectories and the division of the fermi - surface into the pairs of integral planes separated by other cylinders of closed trajectories .
the most general picture of the division of the fermi - surface for @xmath536 is given actually by the scheme containing more than two non - equivalent integral planes ( with the same integral direction ) , connected by cylinders of closed trajectories . in the last situation , which is theoretically possible for fermi - surfaces of very high genus , the pairs of integral planes and individual carriers of open trajectories can coexist after the reconstruction of a part of the open trajectories . mathematically we will have in this case an overlapping of two stability zones @xmath75 and @xmath537 with the same `` topological quantum numbers '' , which will result also in overlapping of the zones @xmath78 and @xmath538 in the experimental study of conductivity .
the full conductivity tensor in the overlappings @xmath539 is given in this case by the sum of the corresponding asymptotic expressions above , so we have here a theoretical possibility to observe very complicated analytical behavior of conductivity in the limit @xmath410 . belonging to a fixed stability zone @xmath108 . ]
let us point out here that the possibility of representation of the fermi - surface in the special form described above in the case of presence of stable open trajectories is based on rather deep topological theorems playing the most important role in the considerations of papers @xcite .
so , the picture described above represents in fact ( under some generic requirements ) the most general situation . at the same time
, we have to note that the statement above should be understood just as a topological characterization of system ( [ quasiclassicalevolution ] ) in the case of presence of stable open trajectories .
thus , the topological structure shown at fig . [ enrichedschemsurf ] can have in fact much more complicated ( visual ) geometric representation in the @xmath208 - space .
finally , we can briefly describe the structure of the `` experimentally observable stability zone '' @xmath540 , arising in the quasiclassical approximation , as a union of the following main parts : \i ) the central part of the zone @xmath78 ( containing the `` mathematical stability zone '' @xmath108 ) .
the conductivity behavior in this part demonstrates the `` most regular '' dependence on the value of @xmath80 , represented by the asymptotic regimes ( [ sigmaomega ] ) or ( [ sigmalowbtau ] ) . at the same time
, the angular dependence of @xmath188 demonstrates here rather irregular character due to the presence of the `` rational peaks '' in the values of @xmath188 on a dense set of directions of @xmath22 in this area ( fig .
[ complbehaviorzone ] ) .
\ii ) the zone of rather complicated form around the central part of the zone @xmath78 ( the black area at fig .
[ complbehaviorzone ] ) .
the tensor @xmath188 demonstrates here the most complicated dependence both on the value and the direction of @xmath22 , corresponding to a gradual transition from the regime ( [ sigmalowbtau ] ) to ( [ sigmahighbtausymm ] ) - ( [ sigmahighbtauantisym ] ) . for an approximation of the behavior of @xmath188
the intermediate powers @xmath479 can be locally used , in general the value of @xmath195 demonstrates unstable behavior . at the same time , the behavior of @xmath188 demonstrates a suppression of the conductivity along the direction of @xmath22 ( @xmath541 ) to some lower values on the exterior boundary of this zone .
\iii ) the `` boundary region '' of the zone @xmath78 .
the behavior of @xmath188 demonstrates here a gradual transition from the regime ( [ sigmahighbtausymm ] ) - ( [ sigmahighbtauantisym ] ) to the simpler behavior ( [ 3dimclosedtr ] ) .
the behavior of @xmath542 demonstrates here a gradual increasing of the conductivity along the direction of @xmath22 to its maximal value on the boundary of @xmath78 . at the end of the paper
, we would like to make also one additional substantial remark .
namely , it can be actually shown , that the topological structure of system ( [ quasiclassicalevolution ] ) admits also a similar description being considered for the whole dispersion relation @xmath5 and not just for a fixed fermi level @xmath204 ( see @xcite ) .
in particular , the `` stability zones '' ( corresponding to the presence of stable open trajectories at least at one energy level ) can be introduced for the whole spectrum @xmath5 . from this point of view , the `` experimentally observable stability zones '' @xmath78 have in fact a connection with exact mathematical stability zones , defined for the whole energy spectrum . in particular
, every `` experimentally observable stability zone '' @xmath78 always belongs to a bigger mathematical stability zone @xmath537 , defined for the whole spectrum @xmath5 .
we investigate the analytical behavior of conductivity of normal metals in strong magnetic fields in the case of presence of stable open electron trajectories on the fermi surface .
as it is shown in the paper , the behavior of conductivity can demonstrate rather nontrivial analytical properties in `` experimentally observable stability zones '' even under the condition @xmath79 .
in particular , in different parts of the experimentally observable stability zone the behavior of conductivity can demonstrate different types of `` regular '' or more complicated intermediate `` irregular '' regimes depending on the value of the magnetic field .
the results of the paper are based on topological description of the carriers of open trajectories on complicated fermi surfaces obtained in recent mathematical investigations of the corresponding geometric problem .
dynnikov . ,
semiclassical motion of the electron .
a proof of the novikov conjecture in general position and counterexamples .
, solitons , geometry , and topology : on the crossroad , amer . math .
ser . 2 , 179 , amer .
soc . , providence , ri , 1997 , 45 - 73 .
i. dynnikov , a. skripchenko . , on typical leaves of a measured foliated 2-complex of thin type . ,
topology , geometry , integrable systems , and mathematical physics : novikov s seminar 2012 - 2014 , advances in the mathematical sciences . ,
ser . 2 , 234 , eds .
v.m . buchstaber , b.a .
dubrovin , i.m .
krichever , amer .
providence , ri , 2014 , 173 - 200 , arxiv : 1309.4884 | we consider the electric conductivity in normal metals in presence of a strong magnetic field .
it is assumed here that the fermi surface of a metal has rather complicated form such that different types of quasiclassical electron trajectories can appear on the fermi level for different directions of @xmath0 .
the effects we consider are connected with the existence of regular ( stable ) open electron trajectories which arise in general on complicated fermi surfaces .
the trajectories of this type have a nice geometric description and represent quasiperiodic lines with a fixed mean direction in the @xmath1-space .
being stable geometric objects , the trajectories of this kind exist for some open regions in the space of directions of @xmath0 , which can be represented by `` stability zones '' on the unit sphere @xmath2 .
the main goal of the paper is to give a description of the analytical behavior of conductivity in the stability zones , which demonstrates in general rather nontrivial properties . |
the pseudocubic perovskite kcuf@xmath0 is a magnetic insulator that has attracted significant interest since the sixties , as a prototype material for orbital ordering , cooperative jahn - teller distortion , and low - dimensional magnetism.@xcite kcuf@xmath0 is also structurally related to high - t@xmath3 superconducting cuprates and colossal - magnetoresistance manganites . in recent years , renewed attention has been given to the problem of orbital ordering and to the study of the interactions between orbital , magnetic , and structural ordering in strongly correlated @xmath1-transition - metal compounds , stimulated in large part by the interest in colossal - magnetoresistance manganites.@xcite in the latter systems , the interplay between orbital , charge , spin , and lattice degrees of freedom is known to play an essential role in establishing the observed physical properties .
contrary to the case of charge , spin , and lattice ordering , however , which can be investigated by conventional x - ray , neutron , and electron diffraction techniques , direct experimental observation of orbital ordering has been and remains in most cases a challenge .
recently , resonant elastic x - ray scattering ( rxs ) techniques have been developed to probe orbital ordering .
rxs has been applied to a number of orbitally - ordered compounds , including various manganites@xcite and kcuf@xmath0.@xcite except for very recent measurements carried out on layered / doped manganites at the mn @xmath4 edges,@xcite most of the experiments performed to date have used excitations from the transition - metal 1@xmath5 core level , i.e. , k - edge excitations . in the case of kcuf@xmath0 and for related perovskite manganites such as lamno@xmath0 , the k edge excitation is the only possible core - level excitation for which resonant photons have a sufficiently large wavevector to access a bragg vector of the targetted orbitally ordered structure .
the k - edge scattering process in manganites and kcuf@xmath0 , however , derives from virtual electric dipole excitations to empty transition - metal 4@xmath6 states .
the resulting rxs amplitude is thus not directly related to the @xmath1 states that exhibit orbital ordering , but to the anisotropy induced in the surrounding 4@xmath6 states .
although the latter is due , in principle , to the orbital ordering and/or associated jahn - teller distortion , this indirect dependence complicates the interpretation of k - edge rxs data .
in fact , it has been first proposed that the measured anisotropy is determined by the 3@xmath7 orbital polarization via onsite 3@xmath74@xmath6 coulomb interaction.@xcite however , subsequent local - density - approximation ( lda ) calculations and lda plus onsite - coulomb interaction ( lda+u ) calculations , performed for a number of manganites@xcite and also for kcuf@xmath0,@xcite have provided rather strong evidence that the k - edge scattering amplitude in these systems is mostly sensitive to the jahn - teller distortion that accompanies the orbital ordering , rather than to the orbital ordering itself . in kcuf@xmath0 , an interesting feature has been observed concerning the temperature dependence of the k - edge rxs intensity for orbital bragg peaks.@xcite the measured scattering intensity shows a substantial and sudden increase ( within 4 - 5 k ) , with decreasing temperature , when approaching the nel temperature .
a somewhat similar phenomenon has been observed also in lamno@xmath0 at the k edge@xcite and in la@xmath8sr@xmath9mno@xmath10 at the @xmath4 edges,@xcite for temperatures approaching the nel temperature , but in these systems the changes in the orbital scattering intensity take place over a much wider temperature range , i.e. , one to two orders of magnitude wider .
these phenomena are therefore not necessarily related , but they are all very puzzling , because they take place in a temperature region where both the orbital - order parameter and the jahn - teller distortion are expected to be saturated . in all of these systems ,
the origin of the changes in the scattering intensity is an unresolved issue . for kcuf@xmath0 ,
two different interpretations have been proposed .
on one hand , the change in rxs intensity has been ascribed to a change in the orbital order parameter , and interpreted as indicative of a strong coupling between spin and orbital degrees of freedom.@xcite on the other hand , based on the results of lda + u calculations showing a dominant influence of the jahn - teller distortion on the rxs spectra , the jump in intensity has been attributed to an increase of the jahn - teller distortion in the low - temperature magnetic phase.@xcite these interpretations are also not necessarily mutually exclusive ; a change in the magnetic structure may induce a change in the orbital - order parameter , which in turn may modify the jahn - teller distortion . in principle , first - principle calculations could help discriminating between these various possibilities by providing quantitative information on the influence of the magnetic structure on both , the jahn - teller distortion and the orbital - order parameter . as
yet , however , to our knowledge these issues have not been addressed by such calculations for kcuf@xmath0 . in the present work ,
we use state - of - the - art lda+u calculations , based on _ ab initio _ pseudopotentials , to investigate the electronic , structural , and magnetic properties of kcuf@xmath0 .
our main targets are the orbital ordering and jahn - teller distortion , and we address the influence of the magnetic structure on these properties .
we also evaluate the cu k - edge rxs spectrum , and investigate the influence of the magnetic structure and jahn - teller distortion on the simulated spectrum . unlike other density - functional - related approaches , the lda+u method has not been extensively used to examine structural properties . in this study
we therefore also first demonstrate the ability of lda + u pseudopotential calculations to provide a realistic description of the ground - state properties of kcuf@xmath0 , when full structural optimization is performed .
kcuf@xmath0 crystallizes in a perovskite structure which is tetragonally distorted by a collective jahn - teller distortion of the cuf@xmath11 octahedra .
two different polytypes have been identified experimentally at room temperature , which differ in their atomic - plane stacking along the @xmath12 axis ( fig .
[ fig : cell]).@xcite the cu ions in kcuf@xmath0 are nominally in a cu@xmath13 ( d@xmath14 ) configuration , with completely filled @xmath1-@xmath15 orbitals and one hole in one of the cubic degenerate @xmath1-@xmath16 orbitals ( @xmath17 , @xmath18 ) .
the cubic degeneracy of the @xmath16 states is lifted in the paramagnetic phase , well above room temperature , by an orbital polarization and jahn - teller instability .
the octahedra elongate along the @xmath19 or @xmath20 principal axes , with an alternate order on neighboring cu a and b sites in the @xmath21 plane ( see fig .
[ fig : cell ] ) .
this distortion corresponds to an alternate distribution of @xmath17 and @xmath22 hole orbitals on cu a and b sites.@xcite along the @xmath12 axis , the stacking of the planes may be antiferromagnetic - like ( type @xmath19 ) or ferromagnetic - like ( type @xmath7 ) , giving rise to the two different polytypes .
( color online ) schematic views of the atomic structure and of the ordering of the cu-@xmath1 hole orbitals in the a- ( left ) and d- ( right ) polytypes of kcuf@xmath0 .
, width=302 ] below the nel temperature ( 38 k for polytype @xmath19 , 22 k for polytype @xmath7 ) , the spins arrange in an a - type antiferromagnetic ( af ) structure , with strong af superexchange interaction between adjacent cu atoms along the @xmath12 axis , and weak ferromagnetic ( fm ) superexchange coupling between cu nearest neighbors within the @xmath21 plane.@xcite this magnetic structure is consistent with the predictions of the goodenough - kanamori - anderson@xcite rules for the superexchange interactions associated with the orbital - ordered configurations displayed in fig .
[ fig : cell ] .
such orbital configurations have been confirmed by previous lda+u and hartree - fock calculations for kcuf@xmath0.@xcite previous calculations have also shown that the lda approach fails to correctly predict the stable orbitally - ordered af insulating structure of kcuf@xmath0 , while the lda+u corrects this feature.@xcite in this paper , we concentrate on the @xmath19-type structure of kcuf@xmath0 , which has been the focus of the experimental rxs studies.@xcite the work is organized as follows : in section [ sec : method ] , we briefly present the calculation method .
in section [ sec : lattice ] , we address the structural properties of the low - temperature a - af structure ; the corresponding orbital ordering and electronic properties are examined in section [ sec : orbital ] . in section
[ sec : spin ] , we investigate the energetics and superexchange couplings of several different ( meta)-stable spin structures of kcuf@xmath0 .
the influence of the magnetic structure on the orbital ordering and jahn - teller distortion is addressed in section [ sec : influence ] . in section
[ sec : rxs ] , we present our results for the rxs spectra , and we summarize our conclusions in section [ sec : summary ] .
the calculations are performed within the lda + u scheme , using the pseudopotential plane - wave method .
we employ the rotational invariant@xcite lda + u energy functional:@xcite
@xmath23 = e^{\text{lda } } [ \rho^{\uparrow},\rho^{\downarrow } ] & + & \nonumber \\ \frac{u - j}{2 } \sum_{i,\lambda,\sigma } n^{i,\sigma}_{\lambda } \left ( 1 - n^{i,\sigma}_{\lambda } \right ) , & & \end{aligned}\ ] ] where @xmath24 $ ] is the standard lda energy functional , @xmath25 is the electron spin density with spin polarization @xmath26 , @xmath27 ( @xmath28 ) is the onsite coulomb ( exchange ) interaction parameter for the localized electrons the cu-@xmath1 electrons , and @xmath29 are the occupation numbers of the localized orbitals .
i and @xmath30 are indexes for atomic sites and orbitals , respectively .
the occupation numbers @xmath29 are the eigenvalues of the density matrices : @xmath31 where @xmath32 are the rotational - degenerate atomic orbitals of the localized-@xmath7-electron shell , @xmath33 stands for the angular quantum number , @xmath34 are the crystal bloch states , and @xmath35 are the corresponding occupation numbers . in our pseudopotential calculations , the crystal bloch states , @xmath34 , are represented by bloch pseudo wavefunctions , and the matrix elements @xmath36 , in eq .
( [ eq : densmat ] ) , are evaluated by projecting the crystal pseudo wavefunctions on atomic pseudo wavefunctions.@xcite we employ the perdew - zunger parametrization of the lda exchange - correlation potential.@xcite for the effective onsite coulomb interaction parameter , @xmath37 , we use a value of 6.6 ev .
this value corresponds to the calculated values @xmath38 ev and @xmath39 ev obtained by constrained lda computations for kcuf@xmath0 in ref . .
we employ a vanderbilt@xcite ultrasoft pseudopotential for cu and norm - conserving troullier - martins@xcite pseudopotentials for k and f. some of the calculations have also been performed using a troullier - martins pseudopotential for cu,@xcite for comparison .
the pseudopotentials we use are scalar relativistic .
the vanderbilt pseudopotential has been generated in the non - magnetic cu @xmath40 atomic configuration , using for the core - cutoff radii : @xmath41 , @xmath42 , and @xmath43 a.u .. the troullier - martins pseudopotentials have been generated in the ground - state configuration of the non - spin - polarized atom , with the cutoff radii ( in a.u . ) : @xmath44 , @xmath45 for k and @xmath46 for f. the troullier - martins pseudopotentials have been cast into the kleinmann - bylander@xcite non - local form using the @xmath5-component as local part . for cu and k
, we have included the non - linear core correction@xcite to the exchange - correlation potential to account for the overlap between the valence and core charge ; for fluorine this is not necessary , since we treat the f @xmath47 semicore electrons as valence electrons .
all calculations are carried out in a 20-atom tetragonal unit cell .
we use a kinetic - energy cutoff of 80 ry for the plane - wave expansion of the electronic states .
the integrations in reciprocal space are performed using a ( 4,4,2 ) monkhorst - pack@xcite k - point grid for all structures , except the non - magnetic structure of kcuf@xmath0 which is metallic . for the latter structure
we employ a ( 6,6,4 ) monkhorst - pack grid together with a gaussian electronic - level broadening scheme,@xcite with a full width at half maximum of 0.01 ry . with the above parameters ,
the estimated convergence on the relative energies of the various magnetic structures considered in this work is @xmath48 mev per formula unit . concerning the pseudopotentials
, we have performed some comparative studies using the troullier - martins and vanderbilt pseudopotential for cu .
they indicate negligible differences , in general , for results concerning the electronic - structure properties . for the structural properties ,
the vanderbilt pseudopotential yields a larger ( @xmath49% larger ) jahn - teller distortion , and a smaller equilibrium volume ( 2% smaller ) than the troullier - martins pseudopotential .
this is related to a higher degree of localization within the core region of the cu-@xmath1 pseudo - charge density obtained with the vanderbilt compared to the troullier - martins pseudopotential .
the structural trends , however , and in particular the dependence on the magnetic state , are the same with the two types of pseudopotentials .
the effect of the k @xmath50 and @xmath51 semicore electrons has also been investigated .
potassium is known to have semicore states which are rather polarizable .
we have therefore performed test calculations treating the k @xmath50 and @xmath51 electrons as valence electrons .
they show that the inclusion of the k semicore electrons has virtually no effect on the electronic properties and changes the equilibrium values of the structural parameters by less than one percent .
in table [ tab : equilibrium ] , we show our results for the equilibrium lattice constant , @xmath19 , tetragonal lattice ratio @xmath52 , fluorine coordinate @xmath53 , and bulk modulus @xmath54 of the low - temperature a - af structure of kcuf@xmath0 .
the theoretical values of the structural parameters are compared to experimental values at low temperature when available and at room temperature . the theoretical equilibrium lattice constant @xmath19 is about 2% smaller than the measured low - temperature value .
this is comparable to lda results for related perovskite materials , such as the ferroelectric oxides studied in ref .
, for which the lda underestimation of the lattice constant is typically @xmath48 to 2% .
.[tab : equilibrium ] calculated values of the equilibrium lattice constant , @xmath19 , tetragonal lattice ratio , c / a internal structural parameter , x@xmath55 , and bulk modulus , b@xmath56 , of kcuf@xmath0 .
the experimental data are from ref .
( a ) and from ref .
( b ) . [ cols="<,^,^,^,^",options="header " , ]
resonant elastic x - ray scattering is a second order process in which a core electron is virtually promoted to some intermediate states above the fermi energy , and subsequently decays to the same core level .
we concentrate here on rxs near the cu k - absorption edge , and examine the resonant scattering intensity for bragg reflections that selectively probe the orbital order ( or equivalently the jahn - teller structural order ) .
the corresponding bragg vectors are @xmath57 , with @xmath58 odd , in units @xmath59 .
we consider dipole transitions to cu-@xmath60 band states , and neglect the effect of the core - hole potential on the intermediate states .
the rxs intensity for a bragg vector @xmath61 and incoming- ( outgoing- ) photon energy @xmath62 may be written as : @xmath63 where @xmath64 are the positions of the cu ions in the unit cell , @xmath65 @xmath66 are the components of the polarization vector of the incoming ( outgoing ) photon , @xmath67 , and @xmath68 are the components of the resonant atomic scattering tensor for the @xmath69th cu ion in the cell . in eq .
( [ eq : fdef ] ) , @xmath70 is the cu-1@xmath5 wave function at site @xmath69 , @xmath71 is the corresponding core level energy , @xmath72 is the position operator measured relative to @xmath64 , @xmath73 ( @xmath74 ) are the empty cu-@xmath60 band states ( energies ) with bloch vector @xmath75 and band index @xmath76 , and @xmath77 is the broadening corresponding to the inverse lifetime of the cu @xmath78-core - hole and @xmath60-electron excited states . for bragg vectors associated with the orbital order , the structure factors , in eq .
( [ eq : idef ] ) , read : @xmath79 where @xmath80 is the atomic scattering tensor of a cu ion at site a ( b ) , and the factor of two accounts for the two a ( b ) cu ions in the unit cell . because of the reflection symmetries @xmath81 , @xmath82 , and @xmath83 present at the cu sites in kcuf@xmath0 , the only non - vanishing components of the atomic scattering tensors are : @xmath84 , @xmath85 , and @xmath86 , and because of the 90-degree roto - translation symmetry that transforms the a and b site into each other , one has : @xmath87 and @xmath88 . the scattering intensity in eq .
( [ eq : idef ] ) may therefore be written as : @xmath89 except for the special case where the polarization - dependent term on the right - hand side of eq .
( [ eq : i ] ) vanishes , the scattering intensity is proportional thus to @xmath90 .
we also note that , within the extremely localized core region where the cu-@xmath78 wave function does not vanish , the radial dependence of each band state @xmath73 , in eq .
( [ eq : fdef ] ) , is essentially the same as that of the atomic @xmath60 state , except for a global scaling factor in the wave - function amplitude .
the scattering factors , in eq .
( [ eq : i ] ) , may therefore be written as : @xmath91 where @xmath92 is the projected atomic @xmath93 density of states and @xmath94 is the fermi energy . hence ,
in our pseudopotential calculations we evaluate the orbital scattering intensity , @xmath95 , as a function of photon energy , directly from the partial @xmath60 densities of states . the core - level energy @xmath71 ( not computed here )
is a value we adjust to align the main feature of the calculated and experimental rxs spectrum.@xcite in fig .
[ fig : resx](a ) , we confront our calculated orbital rxs spectrum for the a - af structure of kcuf@xmath0 to the experimental spectrum [ measured at the ( 3,3,1 ) orbital bragg reflection and for a @xmath96 polarization ] .
we also reported in this figure the theoretical rxs spectrum of the nm structure , for comparison .
both spectra have been computed using the experimental values of the structural parameters . a spectrum obtained for the a - af structure with a 50% increase in @xmath97
is also shown in fig .
[ fig : resx](a ) .
the corresponding calculated k - edge absorption spectra @xmath98 , evaluated as @xmath99 $ ] , are displayed in fig .
[ fig : resx](b ) , and also compared to experiment ( fluorescence data).@xcite all theoretical spectra have been evaluated using for @xmath77 0.5 ev , and we convoluted these spectra with a gaussian of full - width at half maximum of 1 ev .
orbital rxs intensity ( a ) and absorption coefficient ( b ) , as a function of photon energy , near the cu k edge in kcuf@xmath0 .
the calculated spectra of the a - af ( solid line ) and nm ( dashed line ) structures are displayed .
the effect of a 50% increase in the quadrupolar distortion , @xmath97 , in the a - af structure is also shown ( dashed - dotted line ) .
the experimental data are from ref . ; the measured rxs spectrum corresponds to a ( 3,3,1 ) orbital bragg reflection and a @xmath100 polarization .
, width=302 ] the calculated rxs and absorption spectra for the a - af structure are in reasonable agreement with the experimental data .
we note , however , that the energy separation between the two main features in the rxs spectrum is somewhat underestimated in our calculations , and the intensity of the high - energy feature near 9 kev is also not very well reproduced ( both in the rxs and in the absorption spectrum ) .
the change in the magnetic configuration , from the a - af structure to the nm structure , decreases somewhat the rxs intensity .
we note , however , that this change is very small compared to the factor - of - two decrease in intensity observed experimentally when the material transforms to the paramagnetic phase .
the weak dependence of the rxs spectrum on the magnetic configuration is consistent with the results of previous all - electron calculations for the a - af structure ( calculated within the lda+u ) and the nm structure ( calculated within the lda in previous studies).@xcite to assess the influence of the orbital ordering on the rxs signal , we have also computed the rxs intensity of the a - af structure without the quadrupolar distortion ( @xmath101 ) .
although the @xmath102/@xmath103 orbital ordering is still large in this case : @xmath104 , @xmath105 , the calculated rxs signal is about two orders of magnitude smaller than that obtained with the experimental jahn - teller distortion ; the orbital ordering by itself thus has a negligible influence on the rxs intensity .
we note that the fact that @xmath106 ( and @xmath107 ) are very similar with and without the quadrupolar distortion indicates that , already at vanishing @xmath97 , the orbital ordering has essentially reached its saturation value , given the hybridization present in the system . concerning the effect of the jahn - teller distortion , we observe that increasing @xmath97 by 50% increases noticeably the splitting between the two mean features in the rxs spectrum of the a - af structure , but does not change significantly the intensity of the dominant peak .
decreasing , instead , @xmath97 by the same amount has a larger influence on the rxs intensity,@xcite but corresponds to a situation where the two features are essentially merged .
a change in the jahn - teller distortion in the magnetic phase ( not supported by our lda+u calculations of the jahn - teller distortion ) has been suggested as a possible cause of the drastic change observed experimentally in the rxs signal near the nel temperature.@xcite we observe here , however , that a factor of two change in the rxs intensity would require a drastic change in @xmath97 , in the conventional crystallographic structure of kcuf@xmath0 . such a change is bound to induce also a substantial change in the @xmath52 ratio ( see section [ sec : lattice ] ) , which has not been observed experimentally .
the jahn - teller origin of the jump in intensity appears thus somewhat unlikely .
we therefore suggest that possibly a structural transformation occurs at low temperature in kcuf@xmath0 , which is reflected in the rxs signal .
we note that such a change in the crystal structure would not be inconsistent with some unexplained raman@xcite and nuclear magnetic resonance@xcite features observed experimentally at low temperature in this material .
we also note that recently a new crystal structure ( superstructure ) has been proposed for kcuf@xmath0 based on x - ray diffraction at room temperature.@xcite in our lda+u calculations , however , his structure is found to relax to the conventional crystallographic structure of kcuf@xmath0 .
we have studied by means of lda+u pseudopotential calculations , the structural , electronic , and magnetic properties of kcuf@xmath0 , and investigated the cu k - edge rxs spectrum for bragg reflections associated with orbital order . in our studies
, we considered several different ( meta)-stable spin structures for kcuf@xmath0 in order to assess the influence of the magnetic structure on the orbital ordering , jahn - teller distortion , and rxs spectrum . for kcuf@xmath0
, the lda fails to correctly predict the stable af insulating structure .
we have found here that lda+u pseudopotential calculations provide an accurate general description of the properties of this system . the ground state is correctly predicted to be an a - type af structure .
the structural parameters obtained from the lda+u calculations agree also well with experiment , with an accuracy comparable to that obtained by lda calculations for other perovskite materials .
the orbital ordering is predominantly of the @xmath102/@xmath103 type , consistent with previous theoretical predictions for this system .
we find , however , that the orbital - order parameter is significantly reduced with respect to its nominal value due to cu(@xmath1)f(@xmath2 ) hybridization .
we also find that , given this hybridization , the cu-@xmath1 orbital ordering in the a - af phase is already saturated at vanishing jahn - teller quadrupolar distortion of the f neighbours . the rxs spectrum calculated for the a - af structure agrees relatively well with the experimental spectrum .
we find that the resonant signal is dominated by the jahn - teller distortion , with a minor influence of the orbital ordering , in agreement with previous theoretical results .
our lda+u calculations , however , also indicate that a change in the magnetic structure has a small influence on the jahn - teller distortion , and hence on the resonant spectrum , in the commonly accepted crystallographic structure of kcuf@xmath0 .
we therefore suggest that the change observed experimentally in the rxs signal near the nel temperature may be related to a low - temperature structural transformation in kcuf@xmath0 .
we also note that preliminary results obtained from lda+u molecular - dynamic simulations for a - af kcuf@xmath0 indicate the existence of a superstructure , with tilted cuf@xmath11 octahedra , which is slightly lower in energy than the conventional crystal structure of kcuf@xmath0 .
it is a pleasure to thank r. caciuffo , s. de gironcoli , m. cococcioni , g. trimarchi , and n. stojic for useful discussions .
the calculations are based on the pwscf code.@xcite we acknowledge support for this work by the infm within the framework of the `` iniziativa trasversale di calcolo parallelo '' . k. i. kugel and d. i. khomskii , sov .
jetp * 37 * , 725 ( 1973 ) .
j. b. goodenough , _ magnetism and the chemical bond _
( interscience , new york , 1963 ) . m. t. hutchings , e. j. samuelsen , g. shirane , and k. hirakawa , phys . rev . * 188 * , 919 ( 1969 ) . for a theoretical overview of the role of orbital degeneracy in cuprates and manganites ,
see e.g. , a. m. ole , m. cuoco , and n. b. perkins , in _ lectures on the physics of highly correlated electron systems iv _ , edited by f. mancini , aip conf .
no 527 ( aip , melville , ny , 2000 ) , p. 226 .
y. murakami , j. p. hill , m. blume , i. koyama , m. tanaka , h. kawata , t. arima , y. tokura , k. hirota , nd y. endoh , phys .
lett . * 81 * , 582 ( 1998 ) .
m. v. zimmermann , c. s. nelson , j. p. hill , d. gibbs , m. blume , d. casa , b. keimer , y. murakami , c. c. kao , c. venkataraman , t. gog , y. tomioka , and y. tokura , phys .
b * 64 * , 195133 ( 2001 ) ; and references therein .
l. paolasini , r. caciuffo , a. sollier , p. ghigna , and m. altarelli , phys .
* 88 * , 106403 ( 2002 ) .
r. caciuffo , l. paolasini , a. sollier , p. ghigna , e. pavarini , j. van den brink , and m. altarelli , phys .
b * 65 * , 174425 ( 2002 ) .
s. b. wilkins , p. d. hatton , m. d. roper , d. prabhakaran , and a. t. boothroys , phys .
lett , * 90 * , 187201 ( 2003 ) .
s. b. wilkins , p. d. spencer , p. d. hatton , s. p. collins , m. d. roper , d. prabhakaran , and a. t. boothroys , phys .
lett , * 91 * , 167205 ( 2003 ) ; and s. b. wilkins , private communication .
s. s. dhesi , a. mirone , c. de nadai , p. ohresser , p. bencok , n. b. brookes , p. reuter , a. revcolevschi , a. tagliaferri , o. toulemonde , and g. van der laan , phys .
lett . * 92 * , 56403 ( 2004 ) .
k. j. thomas , j. p. hill , y. j. kim , s. grenier , p. abbamonte , l. venema , a. rusydi , y. tomioka , y. tokura , d. f. mcmorrow , and m. van veenendaal , cond - mat/0311553 . s. ishihara and s. maekawa , phys . rev . lett . * 80 * , 3799 ( 1998 ) . i. s. elfimov , v. i. anisimov , and g. a. sawatzky , phys .
lett . * 82 * , 4264 ( 1999 ) .
m. benfatto , y. joly , and c. r. natoli , phys .
rev . lett . * 83 * , 636 ( 1999 ) .
s. di matteo , t. chatterji , y. joly , a. stunault , j. a. paixao , r. suryanarayanan , g. dhalenne , and a. revcolevschi , phys .
b * 68 * , 024414 ( 2003 ) ; and references therein .
p. benedetti , j. van den brink , e. pavarini , a. vigliante , and p. wochner , phys .
b * 63 * , 60408 ( 2001 ) . m. takahashi , m. usuda , and j. igarashi , phys .
b * 67 * , 64425 ( 2003 ) .
a. okazaki , j. phys .
japan * 26 * , 870 ( 1969 ) .
s. k. satija , j. d. axe , g. shirane , h. yoshizawa , and k. hirakawa , phys .
b * 21 * , 2001 ( 1980 ) .
a. i. liechtenstein , v. i. anisimov , and z. zaanen , phys .
b * 52 * , 5467 ( 1995 )
. m. d. towler , r. dovesi , and v. r. saunder , phys .
b * 52 * , 10150 ( 1995 ) .
j. e. medvedeva , m. a. korotin , v. i. anisimov , and a. j. freeman , phys .
b * 65 * , 172413 ( 2002 ) .
m. cococcioni , ph .
d thesis , sissa , 2002 , http://www.sissa.it/cm/phd.php ; and m. cococcioni and s. de gironcoli , cond - mat/0405160 .
h. sawada , y. morikawa , k. terakura , and n. hamada , phys .
b * 56 * , 12154 ( 1997 ) .
a tight - binding bloch state , normalized over the crystal unit cell , is generated for each localized atomic function @xmath32 and used to perform the projection .
we note that with the ultrasoft - pseudopotential scheme , the projection includes also the augmentation pseudo - charge density ( see ref . ) .
perdew and a. zunger , phys .
b * 23 * , 5048 ( 1981 ) .
d. vanderbilt , phys .
b * 41 * , 7892 ( 1990 ) . n. troullier and j. l. martins , phys . rev .
b , * 43 * , 1993 ( 1991 ) . l. kleinman and d. m. bylander , phys .
* 48 * , 1425 ( 1982 ) .
s. g. louie , s. froyen , and m. l. cohen , phys .
b * 26 * , 1738 ( 1982 ) . h. j. monkhorst and j. d. pack , phys .
b * 13 * , 5188 ( 1976 ) .
fu and k .- m . ho , phys .
b * 28 * , 5480 ( 1983 ) .
r. d. king - smith and d. vanderbilt , phys .
b * 49 * , 5828 ( 1994 ) .
r. h. buttner , e. n. maslen , and n. spadaccini , acta cryst .
b * 46 * , 131 ( 1989 ) . a. e. nikiforov and s. y. shaahkin , phys
. solid state * 38 * , 1880 ( 1996 ) .
the projected atomic dos s are evaluated from projections within a small atomic - core sphere : @xmath108 where @xmath109 a.u .. changing @xmath110 to 1 a.u . or 0.2 a.u .
leaves the projected dos s unchanged ( except for a constant dos scaling factor ) .
we note that an accurate determination of the absolute value of the cu @xmath111 @xmath112 cu @xmath113 exciation energy is presently beyond the reach of state - of - the - art _ ab initio _ all - electron lda computations ( see , e.g. , ref . ) . t. ueda , k. sugawara , t. kondo , and i. yamada , solid state commun . *
80 * , 801 ( 1991 ) . c. mazzoli , g. allodi ,
r. de renzi , g. guidi , and p. ghigna , j. magn .
magn . mater . * 242 * , 935 ( 2002 ) .
m. hidaka , t. eguchi , and i. yamada , j. phys .
* 67 * , 2488 ( 1998 ) .
s. baroni , a. dal corso , s. de gironcoli , and p. giannozzi , http://www.pwscf.org . | the orbital , lattice , and spin ordering phenomena in kcuf@xmath0 are investigated by means of lda+u calculations , based on _ ab initio _ pseudopotentials .
we examine the cu-@xmath1 orbital ordering and the associated jahn - teller distortion in several different spin - ordered structures of kcuf@xmath0 .
the ground state is correctly predicted to be an a - type antiferromagnetic structure , and the calculated jahn - teller distortion agrees also well with experiment . concerning the orbital ordering , we find that even for a highly ionic compound such as kcuf@xmath0 , the orbital - order parameter is significantly reduced with respect to its nominal value due to cu(@xmath1)f(@xmath2 ) hybridization .
we also calculate the cu k - edge resonant x - ray scattering spectra for bragg reflections associated with orbital order .
consistent with previous studies , we find that the resonant signal is dominated by the structural anisotropy in the distribution of the f neighbors of the resonant cu atom , and that the cu-@xmath1 orbital ordering has only a minor influence on the spectra . our lda+u results , however , also indicate that a change in the magnetic structure has a small influence on the jahn - teller distortion , and hence on the resonant spectrum , in the conventional ( room - temperature ) crystallographic structure of kcuf@xmath0 .
this may indicate that the large change observed experimentally in the resonant signal near the nel temperature is related to a low - temperature structural transformation in kcuf@xmath0 . |
owing to rapidly advancing detector technology and faster computing speeds , optical and high energy astronomy have enjoyed a synoptic survey capability for over a decade .
a rich discovery phase space has been revealed through synoptic experiments such as the fermi gamma - ray satellite @xcite , the swift gamma - ray burst alert mission @xcite , and optical imagers such as the catalina real - time transient survey ( crts ) , the panoramic survey telescope rapid response system ( pan - starrs ) , and the palomar transient factory ( ptf ; * ? ? ?
a similar science yield likely awaits centimeter radio surveys . in table
[ tab : transients_summary ] we give a summary of known extragalactic and galactic slow transients ( timescale @xmath41s ) at ghz frequencies ( see also * ? ? ?
* ; * ? ? ?
these known sources highlight that radio observations trace high energy particles ; arise in locations of high magnetic field ; and probe the interaction of fast outflows with the surrounding medium .
they are complementary to optical and high energies in that they are not affected by extinction and the emission from fast outflows is often not as narrowly beamed ; providing reliable calorimetry and the true rates of cosmic explosions . in recognition of these discovery opportunities , a new generation of centimeter wavelength facilities have been built or are underway .
these include the karl g. jansky array ( jansky vla ; * ? ? ?
* ) , askap @xcite , meerkat @xcite , and apertif / wsrt @xcite .
the upgraded jansky vla is the first operational interferometer with the survey speed to routinely detect the extragalactic explosive population with modest time allocation . unlike the legacy vla ,
the jansky vla possesses frequency agility essential to characterize the broadband radio spectra of detected transients .
it has a much improved snapshot multi - frequency uv coverage ; imaging artifacts and contamination by sidelobes @xcite are no longer major impediments in transient search .
furthermore , unlike the ska pathfinders , the jansky vla has the spatial resolution to allow the accurate localization within a host galaxy ( nuclear vs non - nuclear ) necessary to rule out the agn activity that dominates the extragalactic transient and variable population .
llllllllll no & object & variability & location & & & rate & ref .
+ & & process & & radio & optical & radio & optical & ( @xmath40.3 mjy ) & + & & & & & & ( erg s@xmath5 hz@xmath5 ) & ( erg s@xmath5 ) & ( deg@xmath6 ) & + + 1 & agn & a ) quiescent & nuclear & years decades & years & 10@xmath710@xmath8 & 10@xmath910@xmath10 & 0.6 & 17 + & & b ) shock - in - jet & & days years & & 10@xmath710@xmath8 & & 1.25 & 610 + & & c ) iss & & minutes days & & & & 60 & 1113 + & & d ) ese & & months & & & & 0.001 & 1415 + & & e ) jet - precession & & years & & & & unknown & 1617 + 2 & sn - ii / ib / c & shocked csm & traces sf & months decades & days months & 10@xmath1110@xmath12 & 10@xmath1310@xmath14 & 0.04 ( ii ) , & 1823 + & & & & & & & & 6@xmath1510@xmath16 ( ib / c ) & + 3 & sn - ia & shocked csm & non - nuclear & days weeks & days months & @xmath1710@xmath18 & 10@xmath1910@xmath14 & @xmath2010@xmath21 & 2328 + 4 & long - grb & shocked csm / jet & traces sf & days years & days months & 10@xmath1210@xmath22 & 10@xmath1410@xmath10 & 6@xmath1510@xmath21 & 29 + & oa & shocked csm & traces sf & weeks years & days ? & 10@xmath1210@xmath23 & 10@xmath1410@xmath24 & 7@xmath1510@xmath25 & 3032 + 5 & short - grb & shocked csm / jet & non - nuclear & days years ? & days & @xmath010@xmath11 & @xmath26 & @xmath2010@xmath16 & 3335 + & bns merger & shocked csm & non - nuclear & weeks years & days & @xmath27 & @xmath28 & 2@xmath1510@xmath25 & 31,3637 + 6 & tde & shocked cnm / jet & nuclear & years ? & months years ?
& @xmath29 & 10@xmath30 & 0.005 & 31,3840 + & off - axis tde & shocked cnm & nuclear & years ? & unknown & @xmath010@xmath23 & unknown & 0.01 & 31 + 7 & aic & shocked csm & non - nuclear & months & day & unknown & @xmath31 & 6@xmath1510@xmath21 & 7072 + + 1 & active star & flaring ( magnetic ) & isotropic & hours days & seconds hours & 10@xmath3210@xmath33 & @xmath34 & 0.02 & 7,4144 + 2 & active binary & flaring ( magnetic ) & isotropic & hours days & hours days & 10@xmath3510@xmath36 & @xmath37&0.01 & 7,41,43,46 + 3 & x - ray binary & mass accretion / flare & gp & days weeks & minutes hours & 10@xmath38 & @xmath010@xmath39 & 0.005 & 7,44,4650 + 4 & cv & dwarf nova / jet & gp & hours days & days weeks & 10@xmath40 & @xmath41 & 0.001 & 44,5154 + 5 & yso & mass accretion / flare & gp & hours weeks & hours weeks & 10@xmath4210@xmath43 & @xmath44 & @xmath200.05 & 7,41,55 + 6 & pulsar & scattering & gp & weeks months & & 10@xmath45 & & 0.003 & 44,56,57 + 7 & magnetar & flaring ( magnetic ) & gp & weeks & unknown & @xmath010@xmath46 ? & unknown & @xmath200.05 & 7,5860 + 8 & brown dwarf & pulsing ( magnetic ) & isotropic & seconds hours & seconds hours & 10@xmath47 & @xmath010@xmath48 & @xmath200.05 & 7,61,62 + 9 & novae & ejecta / shocked csm & gp & years & days months & @xmath49 & 10@xmath5010@xmath51 & @xmath200.05 & 23,6365 + 10 & gcrt & unknown & unknown & minutes years & unknown & unknown & unknown & unknown & 6669 + + [ tab : transients_summary ] to date , there have been rather few radio surveys dedicated to slow variables and transients , and they all have a number of limitations .
the majority of these surveys were single , multi - epoch interferometric pointings with a limited field of view and as a result the number of variables and transients is low ( e.g. * ? ? ? * ; * ? ? ? * ) .
existing wide - area surveys are either based on archival data or the data reduction and candidate source identification was carried out significantly delayed from the observing dates .
this approach has drawbacks since without near - real - time data reduction and multi - wavelength follow - up , the candidate lists contain ambiguous transient classifications @xcite .
we note that wide field , shallow surveys are superior to narrow field , deep surveys , since they have the advantage of bringing the detectable population of transients closer in distance , thus improving the ability to find optical / infrared counterparts and to characterize host galaxies and/or progenitors . in appendix
a we show mathematically that wide and shallow surveys are ideal for radio transient searches . in the light of these factors and to address some of the limitations of past surveys ,
we have commenced a multi - year program , the caltech - nrao stripe 82 survey ( cnss ) , to search for slow radio transients with the jansky vla at 3 ghz in the sloan digital sky survey stripe 82 region .
the cnss is a dedicated transient survey carried out in five epochs ( cadence of days , months and years ) over the entire @xmath0270 deg@xmath1 of stripe 82 , with a uniform single - epoch rms noise of @xmath080 @xmath2jy and a spatial resolution of @xmath03 . in this paper
we present the cnss pilot survey , a sub - mjy jansky vla survey at 3 ghz in 50 deg@xmath1 of stripe 82 .
this is a prototype survey to demonstrate the fast imaging capabilities of the vla , and to develop near real - time data processing , source identification , and transient search .
another unique aspect of this radio survey is that it was undertaken with a contemporaneous high - cadence optical survey with the ptf so that a direct comparison could be made of the dynamic radio and optical skies .
technical details of the radio and the optical surveys are given in
[ sec : survey ] .
the calibration , rfi flagging , imaging and source cataloging of the jansky vla data carried out in near - real - time as well as during the final careful processing is described in [ sec : radio_processing ] .
[ sec : optical_processing ] details the optical data processing for the contemporaneous survey with the ptf .
a description of the radio transient search on timescales less than one week , one month , 1.5 years and longer timescales is provided in [ sec : radio_transients ] .
a subset of the radio variable and transient sources representative of the full sample found in the pilot radio survey are discussed in detail in
[ sec : radio_transients ] .
the optical counterparts of radio sources from the jansky vla survey , and the optical variability of the radio transients from [ sec : radio_transients ] , are studied in [ sec : optical_properties ] using photometry from ptf and sdss .
[ sec : optical_properties ] also gives a brief discussion of the optical transients found in ptf independently of the radio survey .
a summary of the results from our joint radio - optical experiment and their implications are discussed in [ sec : summary ] .
the caltech - nrao stripe 82 survey ( cnss ) was designed to : 1 ) logarithmically sample timescales roughly between one week and one year , 2 ) have a high survey speed ( which is a function of the bandwidth and the antenna system temperature ) , 3 ) have a relatively high angular resolution to facilitate a precise location of transients , 4 ) be carried out in a part of the sky where ample multi - wavelength archival data were available so as to enable the identification of host galaxies or progenitors of the radio transients , 5 ) find non - thermal transients relatively early - on in their evolution , and 6 ) be wide and sensitive enough to potentially detect or place strong constraints on binary neutron star ( bns ) mergers and orphan long - duration gamma - ray burst afterglows ( oas ) , for the pilot survey , we chose a @xmath050 deg@xmath1 region in sdss stripe 82 with similar characteristics .
the region had to satisfy other scheduling constraints in the radio and optical , so we specifically chose the region bounded in right ascension by 329.127and 353.158 , and in declination by -1.132and + 1.167 , approximately .
observations were carried out across four epochs with the jansky vla in b array configuration and s - band was chosen to maximize survey speed . to maximize the continuum imaging sensitivity ,
the observing setup chosen was : wideband interferometric digital architecture ( widar ) correlator with 16 spectral windows , 64 2-mhz - wide channels each to get 2 ghz of total bandwidth centered on 3.0 ghz , and 1-sec integrations .
the first three epochs were observed under the project code 12a-371 between jul
aug 2012 . due to telescope scheduling constraints ,
each of these epochs was divided into two nights of observing of 7 hours each .
all the six 7-hour observing blocks were centered on 23h local sidereal time ( lst ) .
we refer to the three epochs as e1/2/3 and the regions observed on the two nights of the first epoch as r1 and r2 respectively .
_ observations were carried out at the same lst in the three epochs in order to minimize systematic effects associated with sidelobes and beam squint . _ each of the regions , r1 and r2 , was divided into 485 pointings arranged in a hexagonal mosaic pattern ( figure [ fig : radio_obs_setup ] ) , optimized using the _ makeschedule _ task in casa written by andreas brunthaler . in order to maximize the volume probed by our survey , we compromised on uniform - sensitivity coverage , placing neighboring pointings at @xmath5215 .
the mosaic is therefore not precisely hexagonal close packed .
nearest neighbors having the same declination are separated by 15 , and those having offset in declination are @xmath53 apart . during each night of observing ,
one 39-second snapshot observation was obtained at each of the 485 target pointings .
j2212 + 0152 and j2323 - 0317 were chosen as the phase calibrators for the two regions respectively , and 3c48 as the flux and bandpass calibrator .
the phase calibrator was observed every 15 minutes .
a summary of the first three observing epochs is given in table [ tab : radio_observations ] .
for details on the pointing setup.,width=326 ] [ cols= " < , < , < , < , < , < , < , < , < " , ] the fourth epoch ( e4 ) was observed under the project code 13b-370 , and is essentially a co - addition of three epochs of observation carried out between nov 2013 and feb 2014 . for these observations , we took advantage of on - the - fly mosaicing ( otf ) , available as a resident shared risk observing ( rsro ) capability of the jansky vla . in the otf mode ,
data was taken while the telescopes were slewing at a speed of 1 per second along right ascension , and the correlator phase center being stepped every four seconds . in each observing block ,
9 in right ascension and 2.5 in declination centered on 0 were observed to get a total areal coverage of 22.5 deg@xmath1 .
observations were carried out with the telescopes moving in a zig - zag basket - weave pattern on the sky : slewing 9 in decreasing right ascension along a constant declination strip at -1.25 , then stepping up in declination by 10.6 and slewing along increasing right ascension , and so on . covering 2.5 along declination required 15 such zig - zag strips offset by 10.6 in declination from nearest neighbors .
this observing plan is designed to yield 2025 four - second - long scans , and a mosaic with a uniform rms noise of @xmath080 @xmath2jy across 22.5 deg@xmath1 .
since these observations were carried out as part of a proposal to observe 270 deg@xmath1 of sdss stripe 82 , the details will be provided in a later publication ( mooley et al .
2015 , in prep ) .
in this paper , we will be interested only in the region bounded in right ascension by 329.127 and 353.158 , and in declination by -1.132 and + 1.167 .
a summary of the fourth epoch of observations is given in table [ tab : radio_observations ] .
the optical survey , designed to be contemporaneous with the radio survey , was carried out with the 1.2-meter ( 48-inch ) samuel oschin telescope at the palomar observatory as part of the palomar transient factory ( ptf ; * ? ? ?
* ; * ? ? ?
ptf uses a large field camera ( cfh12k ) consisting of a @xmath54 array of @xmath55 pix@xmath1 ccds out of which one ccds are inactive .
the camera subtends @xmath07.2 deg@xmath1 on the sky .
the 50 deg@xmath1 survey region was covered in 14 pointings with each pointing overlapping by about 50% with its adjacent one ( necessitated by one inactive ccd ) .
the footprints of the 14 pointings ( fields ) are shown in figure [ fig : optical_obs_setup ] .
the ptf observations of the 50 deg@xmath1 region were carried out between 25 june 2012 and 25 september 2012 .
the survey was done primarily in the r - band , but g - band observations are available for some of the nights . the stripe 82 fields
were dynamically queued with the other ptf program fields , such that the frequency of observing was determined by the priorities of different ptf fields and the weather . as a result ,
each ptf field was observed a maximum of five times each night .
the log of ptf observations of stripe 82 carried out as part of our coordinated program is shown in figure [ fig : optical_obs_log ] .
standard exposure time per frame is 60 seconds , yielding 3-sigma limiting magnitudes of 20.5 and 21 in the r- and g - band respectively . )
are shown at the top for reference .
[ sec : survey : optical_observations ] for details.,width=326 ]
with the new wideband system in place for the jansky vla , the data rates are quite large , and this poses a computational challenge in terms of rapid data processing needed for triggered transient follow - up .
the size of each of our observation blocks ( listed in table [ tab : radio_observations ] ) is about 250 gb , and to enable near - real - time data processing , dedicated computational hardware was set in place at the nrao in socorro . in general , the calibration , source cataloging and transient search was carried out over a single cluster node at socorro , and the imaging was distributed across multiple ( up to six ) cluster nodes .
all this processing was done using a custom - developed , semi - automated aipslite / python - based pipeline developed at caltech ( e.g. * ? ? ?
details of the near - real - time data processing are given in
[ sec : radio_processing : nrt ] . after the completion of all observing epochs , a more detailed calibration and imaging of the raw data was carried out using casa 4.1 ( [ sec : radio_processing : final ] ) . for each observing block
, the data were downloaded onto the nrao lustre file system ( typical download time of 20 min ) immediately after they appeared on the vla archive ( only several minutes lag after the completion of the observations ) .
next , the raw data was calibrated on a single cluster node using aips / obit tasks accessed via obittalk , parseltongue and aipslite .
the calibration part of the pipeline follows the procedures outlined in chapter e of the aips cookbook .
the first three spectral windows ( spws ) , spanning 2.002.38 ghz in frequency , and as well as the last three spws , spanning 3.624.00 ghz , were completely flagged .
calibration as per the prescription is carried out on a per - spw basis and almost always failed for many of these spws .
this was caused by strong satellite - induced rfi in these bands which was not automatically removed by rflag .
therefore , to speedup the data processing and meet the goal of rapid follow - up , we flagged these spws .
for the fourth epoch , we recognized that the first spw was recoverable with minor manual flagging and the entire spw was not flagged .
post - calibration imaging was carried out somewhat differently for the first three epochs and the fourth epoch , since different observing modes were used .
for the first three epochs , the derived calibration was applied to each pointing during imaging .
the imaging stage of the pipeline was distributed over 16 cores across two cluster nodes such that , at any point in time , 16 pointings were being imaged simultaneously using the aips task imagr .
cleaning was done with 800 iterations per pointing and natural weighting . for each pointing ,
the cleaned @xmath56 pix@xmath1 , 0.58 pix@xmath5 images , uncorrected for the primary beam to get a uniform rms noise across the image , were sub - imaged to @xmath57 pix@xmath1 using the aips task subim .
after the imaging of each observing block , the 485 single - pointing sub - images were copied over to a local machine at caltech , where the source cataloging and transient detection part of the pipeline was run . the thermal noise in each pointing
is expected to be @xmath050 @xmath2jy , and our automated flagging produced data which is within 10% of this value , except for e1 where larger rfi as present .
for source - finding , the use of the _ sfind _ algorithm from miriad was motivated by the short runtime and reasonable completeness and reliability of catalogs @xcite .
we cataloged all sources above the 5@xmath58 threshold and used these for the variability and transient search described in detail in
[ sec : radio_transients : nrt ] .
the imaging for the fourth epoch was carried out with the casa task _ clean _ after the calibrated data from aipslite were exported into uvfits format and then imported into casa as measurement sets . before the imaging step , we clipped the visibility amplitudes of all spws at a threshold determined by the spw with the least rfi . for each spw ,
a rfi - proxy value was calculated as the sum of the mean and three times the standard deviation of the scalar - averaged amplitude .
the clipping amplitude was then chosen to be the least of the rfi - proxy values .
we carried out tests on several pointings to ensure that this clipping threshold was robust and did not alter the flux densities and spectral indices of sources .
after the clipping of visibility amplitudes , all 4-second - long scans , which have unique phase centers , were imaged individually .
imaging was distributed over 60 cores across five cluster nodes such that , at any given time , 60 pointings were being imaged simultaneously using the casa task _ clean_. cleaning was done with 500 iterations per scan and natural weighting with a _
cyclefactor _ of 4.5 chosen to mitigate imaging artifacts .
a common 3 circular restoring beam was chosen to facilitate the co - addition of scans in the image plane .
pix@xmath1 , 0.75 pix@xmath5 images , uncorrected for the primary beam , were produced and then sub - imaged to @xmath60 pix@xmath1 using the image analysis toolkit in casa .
1700 pixels corresponds to 20% of the primary beam at 3 ghz .
the casa images were exported to fits format and combined into a mosaic using the aips task _ flatn _ after correcting for the primary beam .
the mean rms noise is @xmath095 @xmath2jy while the theoretically expected value is 80 @xmath2jy .
the increased rms noise is likely due to low - level rfi and joint - deconvolution not being incorporated during the cleaning process .
sources beyond 5@xmath58 were found using rmsd and sad tasks in aips to get reasonably complete and reliable catalogs @xcite .
once all the e4 observations were complete , the images were combined using the aips task _ comb _ to make a final e4 mosaic .
this mosaic was used to carry out a deeper variability and transient search with respect to epoch e2 , the first survey , and the vla - stripe 82 survey ( [ sec : radio_transients ] ) .
since computing time required for the various stages of data processing and transient search are critical for any near - real - time search , we briefly describe this aspect here .
data i / o is the rate - determining factor in our near - real - time data processing . as a result ,
the processing time is a strong function of the percentage of free disk space on the nrao lustre file system at socorro which hosted the data - processing for our survey .
for the first three epochs , the lustre had 75% free disk space , and the calibration , imaging , cataloging and transient search for each observing block , covering 25 deg@xmath1 over 485 pointings or scans and amounting to 250 gb of visibility data , required about 4.5 hours , 0.5 hours , 0.5 hours , and 0.5 hours respectively .
thus , we were able to detect transient candidates within six hours from the completion of observations .
this fast turnaround allowed same - night triggering of optical follow - up and also rapid broadband radio follow - up , which are crucial for understanding transients that evolve on relatively short timescales . for the fourth epoch ,
the lustre had 1520% free disk space and imaging problem was much more lengthy owing to the 2025 scans to be imaged . the calibration and imaging for each observing block within e4 , which had 250 gb of visibility data and covered a 22.5 deg@xmath1 area on the sky , required 12 hours and 3 hours respectively , while the transient search was completed in 1 hour . during the near - real - time processing , the data editing and source identification ( [ sec : radio_processing : nrt ] )
was not done optimally in order to enable rapid follow - up of interesting radio transient candidates . also , some problems with a small subset of the data were discovered after the near - real - time processing .
specifically , 1 ) the automated rfi algorithm and manual flagging likely excised both terrestrial interference and valid astronomical data , 2 ) the gain calibration for several pointings was affected by strong satellite signals , and 3 ) a subset the initial calibrated data suffered from a systematic phase distortion ( [ sec : radio_processing : final : imaging ] ) .
thus , following the completion of the observing epochs e1 , e2 and e3 , we carried out a more detailed processing of the raw data using casa .
epoch e4 was not subjected to final detailed processing ; calibrated data from the near - real - time processing was directly used for analysis .
the final calibration was carried out using the nrao calibration pipeline ( pipeline version 1.2.0 implemented in casa 4.1 ) , modified to accommodate manual flagging and additional calibration diagnostic plots , and to bypass hanning smoothing .
the calibration was done with two iterative runs of the pipeline .
diagnostic plots for the flux and gain calibrators ( plots of the gain calibration tables and calibrated phases and amplitudes ) from the initial run were used to determine visibilities with incorrect amplitude and/or phase calibration or bad calibrator data .
antenna- , baseline- , correlation- , spw- , and time - based manual flagging of the flux and gain calibrators was incorporated in the second run to remove these visibilities .
especially , five spws having frequencies between 2.122.38 ghz and 3.624.00 ghz , for which the amplitude and phase calibration failed in most instances , were manually flagged before calibration .
after the second iteration of the nrao calibration pipeline followed by imaging , two issues persisted : 1 ) the amplitude gain solutions were not stable as a function of time likely due to rfi - induced non - linearities in the signal chain ( see [ sec : radio_processing : final : rfi ] ) , and 2 ) y - shaped imaging artifact in the first 95 scans of the 24jul2012 epoch , indicating bad raw phase data for the gain calibrator ( see [ sec : radio_processing : final : imaging ] ) . since stripe 82 is close to the clarke belt , radio observations are prone to severe rfi from satellites in geostationary and geosynchronous ( gso ) orbits .
the rfi in the frequency range 3.624.00 ghz is low - level in amplitude , but it distorts the phase information quite significantly .
those data that are irreparably affected are flagged for our final reduction .
gso satellites seen by the jansky vla have not been individually characterized in terms of downlink frequencies and polarizations , and our target data could potentially be affected .
the derived calibration also shows instability in the form of sporadically varying amplitude gains from the gain calibrator .
this effect is pronounced in those gain calibrator scans that are severely affected by rfi . in epochs
e1 , e2 and e3 , where the observations were carried out over the same lst range ( [ sec : survey : radio_observations ] ) , the aberration in gain is reproducible within a specific lst range .
this suggests that the aberrant amplitude gains occur at a particular range of azimuth : between 124 and 255 for region r1
( j2212 + 0152 used as gain calibrator ) and between 135 and 220 for region r2 ( gain calibrator j2323 - 0317 ) .
the scans of the southern calibrator , j2323 - 0317 , are more severely affected than those of j2212 + 0152 due to the proximity of the former to the clarke belt .
the upper panel of figure [ fig : gainamp ] shows the normalized gain amplitudes for j2323 - 0317 and the normalized rms noise in the target pointings as a function of the observing scan number ( as a proxy for time ) .
it is evident that the amplitude gains are correlated with the rms noise of the target pointings , suggestive of anomalous gain values .
our analysis of the calibrated data is complicated by the fact that , although the rfi is localized to typically a single spw , amplitude gains in spws that are free from rfi are also significantly affected ( irrespective of the baseband in which the rfi is present ; lower panel of figure [ fig : gainamp ] ) .
we refer to this as `` gain compression '' .
the `` compression '' signifies that the amplitude gain values are reduced with respect to their true values and result in increased rms noise in the target fields .
initial assessment attributes the aberrant gains with non - linearities in the amplification stage in the analog signal chain somewhere upstream from the correlator caused by the the high power of the satellite signals .
however , this hypothesis has not been thoroughly tested .
we devised a fix for the gain compression in the gain calibrator scans by altering the amplitude gain calibration table generated by the nrao calibration pipeline . towards the end of each epoch
there is a time interval lasting several minutes where the gains are relatively stable and close to unity ( to within a few percent ) .
we therefore applied constant amplitude gains of unity for the duration of each epoch , and for all epochs . with this correction ,
the rms noise values for epochs e1 , e2 and e3 are more stable with time and are usually within 10% of the thermal noise .
the spread in the baseline - based amplitude gains in the stable time duration is @xmath200.05 , and hence we estimate that this correction will affect the true amplitudes of the target sources by @xmath205% . _
given that gain calibrators in the vicinity of the clarke belt are susceptible to gain compression , we have avoided using such gain calibrators for the final cnss survey .
_ accordingly , there is no indication of severe gain compression in the gain calibrator scans from epoch e4 .
while the gain compression in the gain calibrator scans has a relatively straightforward workaround , the effect of gain compression on the target pointings , if present , is much more challenging to deduce .
we investigate the possible issue of gain compression in target fields in [ sec : radio_transients ] .
after calibration of the data using the nrao pipeline , we clipped the visibility amplitudes of all spws in the manner described in [ sec : radio_processing : nrt ] .
we then derived and applied a single - round phase - only self calibration solution to every pointing containing a source brighter than 10 mjy in the near - real - time image .
the model for the self calibration was constructed by making a preliminary image with 100 clean iterations using the casa task _ clean_. self calibration solutions were allowed to break at spectral window and scan boundaries .
each self calibrated pointing was then imaged by cleaning within the 1% power point of the primary beam of the lowest frequency , 2.0 ghz , using natural weighting in the _ clean _ task in casa , choosing a cellsize of 0.6 , and 500 clean iterations to minimize clean bias . a single taylor term ( @xmath61 )
was used , which was sufficient to achieve the desired dynamic range @xmath62 .
we also chose a _
cyclefactor _ of 4.5 to minimize imaging arifacts .
basic quality checks were made by visually inspecting the images for each pointing . using briggs weighting with
a _ robust _ parameter of zero , we re - imaged those pointings in which imaging artifacts persisted .
all the 4800@xmath154800 pix@xmath1 cleaned single - pointing images at 2.8 ghz were then sub - imaged to 2400@xmath152400 pix@xmath1 .
2400 pixels corresponds to 42% , 15% and 1% of the power point of the primary beam at 2.0 ghz , 2.8 ghz and 3.63 ghz respectively ( figure [ fig : pb ] ) .
the median synthesized beam obtained for all three epochs is @xmath63 , except for e1r1 , where the median synthesized beam size is @xmath64 likely due to increased rfi .
the mean rms noise per epoch is given in table [ tab : radio_observations ] .
the cumulative plot for rms noise ( flat sky ) as a function of survey area is shown in figure [ fig : sensitivity ] .
for epochs e2 and e3 , more than 90% of the survey area has rms noise lower than 105 @xmath2jy , while for e4 it is better than 75 @xmath2jy .
epoch e1 , having rms noise of 130 @xmath2jy or better over 90% of the survey area , is severely affected by low - level rfi compared to other epochs .
note that , for the first three epochs , neighboring pointings are quite far apart and do not contribute significantly to the sensitivity of the overlap region .
hence , single pointings were imaged separately and treated independently during the source finding step . for the 24aug2012 ( epoch e3 , region r1 )
epoch , the first 1.5 hours of observations have been affected by a systematic phase distortion .
about 80 target pointings and six phase calibrator ( j2212 + 0152 ) scans have been thus affected .
when the calibrated data is imaged , the phase distortion manifests as a three - pronged artifact such that point sources appear extended along three axes characteristic of the y - shape of the vla configuration ( left panel of figure [ fig : y ] ) .
we looked at the phase gain solutions per antenna derived from one of the affected phase calibrator scans and plotted it as a function of x- and y - positions ( east - west and north - south ) of the respective antennas in the array .
the center of the array has higher gain phase ( degrees ) than the arms , and along the arms , it decreases monotonically , in general .
this is equivalent to a bowl - shaped or lenticular phase screen above the array .
the exact cause of the phase distortion is unknown , but it may arise due to rfi , unfavorable weather , the ionosphere , or some internal error in the antenna phase recording .
while the true phases can be restored through self calibration , not all our affected pointings have bright sources to facilitate this .
therefore , for the affected pointings , we used the corresponding phases and amplitudes from the 18jul2012 ( epoch e2 , region r1 ) observations as a model for self calibration .
the image of a bright source after such a calibration is applied , is shown in the right panel of figure [ fig : y ] . .
the field of view in both panels is identical and equal to 35 . in the left panel ,
the colorbar runs from @xmath65 mjy to @xmath66 mjy , while in the right panel it runs from @xmath67 mjy to @xmath68 mjy.,title="fig:",width=161 ] .
the field of view in both panels is identical and equal to 35 . in the left panel ,
the colorbar runs from @xmath65 mjy to @xmath66 mjy , while in the right panel it runs from @xmath67 mjy to @xmath68 mjy.,title="fig:",width=161 ] the aips task rmsd was used to create an rms image for each individual pointing for epochs e1 , e2 and e3 , and the resulting background images were supplied as input to sad .
source - finding was then carried out with sad to find sources with snr@xmath45 . for sources present in adjacent pointings ,
only the catalog entries were retained in which the sources are closest to their respective pointing centers .
thus , the cataloged sources are within 9 ( within 40% of the 2.8 ghz primary beam ) from their nearest pointing centers .
approximately 10@xmath69 synthesized beams span the images from our 4-epoch survey , from which we expect 500 , 2 , and @xmath201 noise peaks above 5@xmath58 , 6@xmath58 , and 7@xmath58 respectively .
while we used 5@xmath58 source catalogs for the near - real - time transient search , we followed the recommendations of @xcite , to filter our new source catalogs and keep only snr@xmath47 sources during the final processing . with the goal of variable and transient search in mind
, we further retained only probable point sources10 mjy sources were rejected , retaining the @xmath410 mjy sources themselves . in order to filter out false positives with flux densities greater than 10 mjy , all cataloged objects within 1.6 of @xmath450 mjy sources were rejected , retaining the @xmath450 mjy sources themselves .
this reduces our transient search area by less than one square degree , but gets rid of almost most false positives . some sidelobes identified as such by eye
, were also rejected from the source catalogs .
this false positive rejection step eliminates 438 objects out of 4205 unique objects in our list of probable point sources .
additional 35 sources were discarded as being resolved after manual inspection of a subset of images from our survey . ] .
the resulting catalogs from each epoch were then merged into a single point - source catalog ( psc ) . in many cases , it was found that the peak flux densities of sources in the psc , as reported in the original sad catalogs , was not in agreement with the peak flux densities as seen in the images . since accurate peak flux densities are critical for any variability and transient search , we replaced the sad peak flux density values for all objects in the psc with the corresponding peak pixel values in the single - epoch images .
the uncertainties in the peak flux densities of sources in the psc were taken directly from their corresponding counterparts in the sad catalogs or by measuring the image rms noise in the vicinity of the sources .
the flux densities of all sources in the psc and their associated uncertainties were corrected for the primary beam at 2.8 ghz using parameters from the pbcor task in aips .
the psc thus contains peak flux densities and the uncertainties of 3652 point sources having snr@xmath47 in any one of the four epochs .
the histogram in the left panel of figure [ fig : s_hist ] shows the distribution of flux densities from epoch e2 of sources in the psc .
we used this psc for our variability and transient search ( [ sec : radio_transients ] ) .
as part of a preliminary data release , we provide to the astronomical community single - epoch images and our psc for the four epochs of radio observations of the survey region . additionally , we release a 5@xmath58 catalog ( e2cat ) of sources from epoch e2 .
e2cat consists of 6846 sources after a preliminary elimination of false positives ( probable sidelobes ) as done with the psc , and described in
[ sec : radio_processing : final : catalogs ] .
the histogram of the integrated flux densities recorded in e2cat are shown in the left panel of figure [ fig : s_hist ] .
the final data release , consisting of calibrated uv data , a deep mosaic made from the co - addition of all four epochs , and a deep source catalog , will be done in a later publication ( horesh et al .
, in prep ) .
the ptf near - real - time reduction and transient search was carried out using a pipeline hosted at the lawrence livermore berkeley laboratory .
transient search was done by subtracting an older reference ( deep co - add ) image from the new observations . for further details on the ptf near - real - time operations ,
see 2.1 of @xcite .
the stripe 82 data processed with the ipac pipeline @xcite and photometrically calibrated @xcite was used to make light curves reported in this work .
the processed data and catalogs are available online at the nasa / ipac infrared science archive as part of ptf first data release .
the histogram of the r band magnitudes of ptf sources considered in this work are shown in the right panel of figure [ fig : s_hist ] .
the distribution of r band magnitudes from the sdss stripe 82 co - add catalog are also given in figure [ fig : s_hist ] .
in this paper we use the terms `` variables '' and `` transients '' somewhat interchangeably , making note of the fact that the processes listed in table [ tab : transients_summary ] are all transient processes . here , we use our four survey epochs to carry out two - epoch transient searches on different timescales .
we are thus probing variability on all timescales that are approximately larger than the duration of each observation ( 40 seconds per pointing in our case ) but less than the separation between the two epochs being compared . accordingly , in the text below , we refer to a transient search on a `` timescale of x days '' synonymously with
timescale less than x days , but note that in table [ tab : transients_summary ] we refer to the precise `` timescale of evolution '' since the evolution of the light curves of these transients has been well studied through high - cadence targeted observations .
one of the primary goals of our survey was to understand the transient phenomena on timescales of one week , one month , and one year .
for this purpose , we performed a two - epoch comparison of flux densities of sources in our survey using the psc from [ sec : radio_processing : final : catalogs ] . since radio interferometric noise is gaussian distributed ( e.g. * ? ? ? * ) , we can compare the flux densities of a source between two epochs with the statistic @xmath70 under the null hypothesis that they are drawn form the same distribution ( and hence the source is non - variable between the two epochs that are being compared ) .
from statistical theory we know that this quantity is distributed according to the student - t distribution @xcite .
we define a source as being a variable if the t - statistic lies beyond the 95% confidence interval4@xmath58 . for the gaussian distribution
, 4@xmath58 corresponds to a probability of about 1/16,000 , while the number of measurements in our variability analysis ( few thousand point sources multiplied by four epochs ) is about 15,000 . for our variability analysis
we could have used the @xmath71 statistic as previous studies have done ( e.g. * ? ? ?
* ; * ? ? ?
* ) , but for cases where the number of degrees of freedom are small , student s t is preferred .
see also chapter 4 and appendix c of @xcite .
] , i.e. if , @xmath72 the choice of the confidence interval was motivated by the need for high reliability of the variable sources found , while making a possible compromise on the completeness .
this issue of completeness and reliability is taken into account while calculating the rates of variable sources in [ sec : summary : variability ] .
for the two - epoch comparison of variables selected using equation [ eqn : var_2epoch ] , we use the modulation index as the measure of variability .
@xmath73 where @xmath74 is the mean of the flux densities , @xmath75 and @xmath76 , in the two epochs being compared .
we note that the definition of modulation index used here is proportional to the one used by @xcite , but larger by a factor of @xmath77 .
the modulation index is related to the fractional variability , @xmath78 , by the following equation .
@xmath79 initially , during the near - real - time transient and variability search , we used variability statistic @xmath80 ( equation [ eqn : var_2epoch ] ) , and further selected sources with fractional variability larger than 50% .
as soon as a new epoch was observed , it was compared with all previous epochs . among @xmath06000 sources in our near - real - time 5@xmath58 single - epoch catalogs , we found @xmath0500 to be variables .
we note here , beforehand , that following the final data processing , we found that most of these candidates were falsely identified as variable due the rfi - induced amplitude calibration problem identified in
[ sec : radio_processing : final : rfi ] .
the shortlist of follow - up candidates was generated from the near - real - time catalogs .
for further filtering of variable candidates , we used multi - wavelength archival data and ptf , and inspected the radio image cutouts by eye .
it turned out that the ptf data were not very useful for this purpose . as we show in [ sec : optical_properties ] , there is little overlap between variable radio sources and sources that vary in the optical , and almost all these jointly variable sources are agn .
optical spectra from the sdss boss program @xcite were especially useful in filtering our list of candidates .
in general , for those radio transients having boss spectra , we filtered out sources showing agn features in their optical spectra .
wise colors were also used to eliminate possible agn .
preference was given to sources that are undetected in archival radio data .
we also compared our catalogs with those from past surveys at 1.4 ghz , viz .
first @xcite and vla - stripe 82 @xcite , and selected @xmath010 point sources with implied spectral indices more than 2.5 or less than -2.5 between 3 ghz and 1.4 ghz .
this was motivated by the search for transients such as supernovae and tidal disruption events ( swift j1644 + 57-like ; * ? ? ?
* ; * ? ? ?
* ) which evolve on a timescale of years .
however , most of these candidates turned out to be either gps sources or flaring agn ( see below ) .
our final list of candidates for follow - up consisted of 20 objects that displayed a factor of two or more fractional variability between any two epochs , or were new bright sources compared with previous surveys at 1.4 ghz . after correcting for the aberrant complex gains during the final data processing ( [ sec : radio_processing : final ] ) , we carried out a careful search for variables and transients bearing in mind the possibility of gain compression ( [ sec : radio_processing : final : rfi ] ) in the target fields . to maintain consistency during our variability and transient search , we only compared epochs e1 , e3 , and e4 with epoch e2 in order to probe variability on timescales of one month , one week , and 1.5 years respectively .
note that among the first three epochs , e2 was the least affected by rfi .
for each source in the psc , we calculated the statistical quantity @xmath80 using equation [ eqn : var_2epoch ] .
there appeared to be a small deviation of this statistic from the expected distribution ( student s t distribution ) .
this may be a result of low - level gain compression in the target fields .
although the epoch - to - epoch flux densities are not significantly discrepant , we addressed this issue by applying small , relative , corrections to the ratios of and the differences in the flux densities of sources between the two epochs being compared .
for the one - week and one - month comparison , we made multiplicative corrections of 2%@xmath813% ( median value with standard deviation ) to the ratios of flux densities and additive corrections of 20 @xmath2jy @xmath81 50 @xmath2jy to the differences in the flux densities of sources in the psc .
these corrections were derived for the two - epoch comparisons by plotting the ratios and differences of flux densities of sources as a function of time of observation and demanding the ratios and differences to be centered at unity and zero respectively . during the comparison between epochs e4 and e2 , we found it necessary to make two independent sets of corrections to the ratios of flux densities . in the first set ,
multiplicative corrections of 3%@xmath814% were seen to be correlated with the declination of the sources .
this discrepancy of flux densities in epoch e4 with respect to epoch e2 is a result of the the otf survey design and the near - real - time imaging process , and will be described in detail in cnss paper ii . in short , during the otf observations carried out for epoch e4 , the primary beam moved by 4 within every scan ; this was not accounted for in the imaging step .
the resulting fractional change in flux density of a source depends on its position in the primary beam .
this effect is manifested as a sinusoidal pattern when the flux density ratios are plotted as a function of source declinations .
the second set of corrections consist of multiplicative corrections of 4%@xmath814% to the ratios of flux densities and may be attributed to gain compression .
additive corrections of 10 @xmath2jy @xmath81 30 @xmath2jy were applied to the differences in the flux densities between epochs e4 and e2 . with the application of these small , first - order ,
corrections we ensure to not over - fit the data and remove real variability in flux densities of the sources in the psc . the histograms showing the resultant distributions of the variability statistic , @xmath80 , is shown in figure [ fig : varstat ] .
it is evident that , after the corrections , @xmath80 agrees well with the student s t distribution ( however , see note on the 1.5 year distribution below ) . , for sources in the psc . after the application of the corrections to the flux densities mentioned in
[ sec : radio_transients ] , @xmath80 agrees with the expected distribution , student s t.,width=336 ] to determine additional sources of artificial variability , we ran further intensive tests with manual flagging and imaging .
we found variations of up to a few percent in the flux densities of sources just by choosing different values for the input parameters to the casa _
clean _ task .
specifically , these parameters are : cleaning iterations ( _ niter _ ) and taylor terms ( nterms ) .
we also noticed variations by changing the values provided to the _ clipminmax _ parameter in the _ flagdata _ task .
additionally , we know that the flux density absolute calibration ( we used 3c48 as the flux density standard calibrator in our survey ) is usually accurate to only 35% ( e.g. * ? ? ? * ; * ? ? ?
* ; * ? ? ?
all in all , due to gain compression and other possible causes of artificial variability mentioned above , we adopt a conservative criterion for selection of our variable sources : fractional variability greater than 0.3 ( @xmath82 ) .
our final variability selection criteria are @xmath83 and @xmath82 .
the plots of the @xmath80 versus @xmath84 for sources in the psc after applying the corrections mentioned earlier in this subsection are shown in figure [ fig : varstat_m ] .
for the sources in the psc satisfying the variability selection criteria , we manually inspected the image cutouts and their fitted gaussian parameters , viz .
major axes ( @xmath85 ) , minor axes ( @xmath86 ) , the peak flux densities ( @xmath87 ) and the integrated flux densities ( @xmath88 ) , in order to ensure that they are not resolved . for sources detected at a high significance level ( snr@xmath8915 ) , we used the criteria : @xmath90 , @xmath91bmaj , and @xmath92bmin , where the variables in typewriter text denote the major and minor axes of the synthesized beam . for the sources
detected at low significance , we relaxed the integrated - to - peak flux density ratio to 1.31.5 .
this procedure for selecting unresolved is motivated by the standard techniques used for deep radio surveys ( e.g. * ? ? ? * ; * ? ? ?
while this approach ( c.f .
* ) errs on the conservative side , we prefer to identify real variables rather than false positives caused by resolution effects .
we thus find 142 variables among 3652 sources in our psc , or , in other words , about 3.8% of the sample has fractional variability more than 30% on timescales @xmath201.5 years .
the number of point sources varying on week , month , and 1.5 year timescales are 38 , 31 , and 96 respectively , or 1.0% , 0.8% and 2.6% respectively .
these sources are listed in table [ tab : transients ] .
a glance at figure [ fig : varstat ] suggests that the variability statistic , @xmath80 , comparing epochs e2 and e4 ( probing variability on a timescale of 1.5 years ) deviates from the student s t distribution in the tail
. this may be due to some calibration errors from the near - real - time processing that we have not tried correcting for , and therefore the fraction of variables on a 1.5-year timescale , 2.6% , is likely an upper limit .
note that the variable sources are distributed uniformly on the sky with no pattern indicative of particular region(s ) of the sky being affected by calibration errors . since our 50 deg@xmath1 survey region contains rich archival data from the first and vla - stripe 82 surveys , we extended our transient search to timescales of up to @xmath020 years .
we compared the 3 ghz peak flux densities of sources in the psc with the 1.4 ghz peak flux densities from these two surveys ( figure [ fig : var_cnss_first_hodge ] ) .
the mean spectral index , as @xmath93 .
] , @xmath94 , appears to be between @xmath67 and @xmath95 , as expected ( e.g. * ? ? ?
* ) . with the view of finding only the extreme variable and transient candidates
, we searched beyond implied spectral indices of @[email protected] ( i.e. less than @xmath97 or greater than 1.5 ) with respect to the first survey , and beyond @[email protected] with respect to the vla - stripe 82 survey .
the threshold of @xmath812.5 in spectral index is somewhat arbitrary ; it is motivated by our need for filtering out most of the sources and selecting only a few , extreme , objects .
our search resulted in 11 unique objects .
two candidates are resolved out ( closely - separated radio lobes ) in our data ; one candidate is identified as such due to its erroneously absent in vla - stripe 82 catalog ( but present in the image ) ; and one candidate , vtc220456 - 000147 , is identified as a variable also on a timescale of 1 month .
accordingly , out of the eight genuine variables , seven have been listed in table [ tab : transients ] in the timescale @xmath2020 years section . ) .
triangles denote upper limits to the peak flux density .
red color denotes sources selected as variable candidates by our search criteria .
[ sec : radio_transients : final ] for details.,width=336 ] below , we discuss five candidates , where `` vtc '' stands for `` vla transient candidate '' and the right ascension and declination are both rounded at the least significant digit . ] in detail , that are meant to be broadly representative of our complete sample of 142 variables and transients .
additionally , we discuss about two transients found serendipitously .
a summary of all the variable and transient sources found in this work is given in table [ tab : transients ] , and a summary of radio follow - up observations is given in table [ tab : radio_followup ] .
this variable source has flux densities [email protected] , [email protected] , [email protected] , and [email protected] mjy in epochs e1 to e4 , and in the first two epochs it has spectral indices @xmath98 and @xmath99 respectively at 3 ghz . at the location of vtc225411 - 010651 , the 1.4 ghz peak pixel values from the first and @xcite surveys are @xmath100 and @xmath101 mjy , respectively . to investigate further , we obtained follow - up observations of this object between 115 ghz in two epochs ( on 02 sep and 17 sep 2012 ; 10 days and 25 days respectively after the epoch e3r2 ) with the vla .
the continuum radio spectra are shown in figure [ fig : vtc225411 - 010651_radiospectrum ] .
the first follow - up observation reveals a flat spectrum source in the 215 ghz frequency range .
the second observation indicates a spectrum peaked at about 1.5 ghz with the optically thin part having a spectral index of @xmath102 .
the spectrum is flat beyond 4 ghz .
sub - mm ( 100 ghz ) follow - up observations on 09 sep 2012 with carma gave a non - detection with a 3@xmath58 upper limit of 2.1 mjy .
there is no optical counterpart in ptf , but a faint sdss dr7 @xcite source having @xmath103 mag .
lies 1.8 away from the radio source position of vtc225411 - 010651 .
the photometric redshift from sdss is @xmath104 . assuming a 1.4 ghz quiescent flux density of 0.2 mjy
, we can calculate the radio - to - optical flux density ratio for the host galaxy to be @xmath105 , which is typical of radio - loud agn ( e.g. * ? ? ?
wise colors ( @xmath106 mag , @xmath107 mag ) of the host galaxy are consistent with a liner / ( u)lirg / spiral galaxy @xcite .
upper limit at 1.4 ghz from @xcite is shown in red .
the measurements from the four survey epochs are shown in grey .
the first two epochs have spectral indices of @xmath98 and @xmath99 within the s band respectively and are indicated by grey crosses joined by straight lines between the two frequencies used for spectral index measurement .
[ sec : radio_transients : one_week : vtc225411 - 010651 ] for details.,width=326 ] the flat spectrum of this source in the first follow - up epoch indicates a jet with unresolved knots . assuming a 3 ghz quiescent flux density of 0.3 mjy
, we obtain of radio luminosity of @xmath108 erg s@xmath5 hz@xmath5 at the sdss photometric redshift .
it is thus likely that this source is a low - luminosity radio - loud agn , in which case , the peaked spectrum in the second follow - up epoch can be naturally explained as an intrinsic agn flaring phenomenon .
the flaring spectrum in figure [ fig : vtc225411 - 010651_radiospectrum ] is reminiscent of the shock - in - jet model that has been extensively used to explain flaring in bright quasars @xcite .
the strong variability of vtc225411 - 010651 in the s band between epochs e2 and e3 of our survey indicates that the flares in this agn evolve on a timescale of less than one week .
we note that current radio variability surveys are primarily focused towards the monitoring of blazars and bright quasars , where the flares at ghz frequencies evolve typically on timescales of a few months ( e.g. * ? ? ?
* ; * ? ? ?
* ) , and that vtc225411 - 010651 presents the first ever direct evidence of shocks propagating down the jet in sub - mjy agn on a relatively short timescale .
vtc223612 + 001006 is detected as a transient on 13 jul 2012 ( epoch e1 , regions r1 ) .
the peak pixel values at the location of the transient in epochs e1e4 are [email protected] , [email protected] , [email protected] , [email protected] mjy respectively .
the image cutouts near the source location for the three epochs are shown in figure [ fig:1222af ] .
a comparison source about 1.5 away from this transient candidate has relatively stable peak flux densities , 0.604 , 0.584 , 0.551 , and 0.488 mjy in epochs e1e4 respectively , indicating that the variability of the candidate is secure . in the first and @xcite surveys , the peak pixel values at the location of the transient candidate are @xmath109 and @xmath110 mjy respectively .
the nearest optical counterpart of vtc223612 is the 10th v - band magnitude star hd 214129 ( @xmath01 away ) , documented as a k0iv star by @xcite , and as a g5/6iii star by @xcite .
hd 214129 is a known visual binary and possibly a triple stellar system ( e.g. * ? ? ?
* ) , and has been detected in the x - rays by rosat ( 1rxs j223612.5 + 001008 ; * ? ? ?
hd 214129 is saturated in ptf and sdss , precluding the study of optical variability of this galactic radio transient .
we obtained an optical spectrum for this star using the echelle spectrograph ( esi ) at keckii on 13 sep 2012 , which suggests a spectral type similar to k0iv .
the spectrum does not show any strong emission lines within our spectral coverage between 400010000 .
figure [ fig:1222af ] shows the binned optical spectrum between 40007000 .
the narrow absorption line profiles and absence of lorentzian wings indicate that the star is a subgiant or a giant . fitting a blackbody to the photometric data from the sdss , nomad and wise catalogs
gives an effective temperature of @xmath111 k. for details .
, title="fig:",width=355 ] for details .
, title="fig:",width=336 ] if this were a main sequence star , it would have a photometric parallax distance of @xmath112 pc , while if it is a giant , @xmath113 pc .
a sub - giant can have any value between these extremes .
an independent constraint on the distance comes from the space velocity . using the proper motion and radial velocity , @xmath114 and @xmath115 @xcite
, we can calculate the space velocity ( heliocentric ) for @xmath112 pc and @xmath113 pc as @xmath116 and @xmath117 respectively @xcite .
typically for stars we expect heliocentric space velocities between 0100 ( e.g. * ? ? ?
therefore , we expect @xmath118 . here , we adopt a distance of 100 pc .
we used a 1 kev collisionally - excited plasma ( apec ) model and an absorbing galactic hydrogen column of @xmath119 ( derived from the _ nh _ task in _ heasoft _ ) to convert the rosat / pspc countrate from 1rxs ( 0.23 counts s@xmath5 ) in the 0.12.4 kev energy band to an unabsorbed flux of @xmath120 using webpimms .
this corresponds to an x - ray luminosity of @xmath121 at 100 pc .
the x - ray luminosity , the x - ray - to - optical flux ratio of 1.4@xmath1510@xmath122 , and the spectral type all suggest that hd 214129 is an rs cvn - type binary .
accordingly , we expect the quiescent 3 ghz radio luminosity of the transient to be @xmath123 erg s@xmath5 hz@xmath5 ( * ? ? ?
* ; * ? ? ?
* ; however , strictly speaking , their l@xmath124-l@xmath125 relationship is valid for 5 ghz flux density ) . from epoch
e4 we get the 3@xmath58 upper limit on the quiescent 3 ghz flux density of this transient as @xmath126 erg s@xmath5 hz@xmath5 .
using a distance of 100 pc , we can estimate the flaring radio luminosity in epoch e1 as @xmath127 erg s@xmath5 hz@xmath5 . from the binarity , the optical spectrum , radio and x - ray luminosities , x - ray hardness ratios from rosat , and the x - ray - to - optical flux ratio ,
we conclude that this radio transient is a flare from an active binary system .
vtc225707 - 010238 is a persistent but variable source with flux densities [email protected] , [email protected] , [email protected] , and [email protected] mjy in the four epochs of our survey .
the spectral indices within the s band for the first three epochs are @xmath128 , @xmath129 and @xmath130 .
the 1.4 ghz peak flux densities from the first and @xcite surveys are @xmath131 and @xmath132 mjy respectively .
we followed up this transient with the vla on 01 sep 2012 ( 215 ghz ) and 17 sep 2012 ( 14 ghz ) .
the continuum radio spectra are shown in the top panel of figure [ fig:1222y ] .
both follow - up spectra are peaked at a few ghz .
on 01 sep 2012 , the peak is 2.5 mjy at 7 ghz , while the 17 sep 2012 spectrum is likely peaked at 3 ghz with a flux density of 1.5 mjy .
this might suggest that this object is a flaring gigahertz - peaked spectrum ( gps ; young agn ) source .
sub - mm ( 100 ghz ) follow - up observations on 09 sep 2012 with carma gave a non - detection with 3@xmath58 supper limit of 2.1 mjy .
upper limit at 1.4 ghz from @xcite is shown in red .
the measurements from the first three survey epochs are shown in grey .
[ sec : radio_transients : one_month : vtc225707 - 010238 ] for details .
_ bottom _ : the optical spectrum of vtc225707 - 010238 observed with lris at kecki on 18 aug 2012 . the spectrum is typical of a blue quasar .
we derive a redshift of 1.56.,title="fig:",width=336 ] upper limit at 1.4 ghz from @xcite is shown in red .
the measurements from the first three survey epochs are shown in grey .
[ sec : radio_transients : one_month : vtc225707 - 010238 ] for details .
_ bottom _ : the optical spectrum of vtc225707 - 010238 observed with lris at kecki on 18 aug 2012 .
the spectrum is typical of a blue quasar .
we derive a redshift of 1.56.,title="fig:",width=336 ] the optical counterpart of vtc225707 - 010238 is a blue point source in sdss having @xmath133 mag .
the optical light curve from ptf shows intra - day small amplitude ( sub - magnitude ) variability .
follow - up spectroscopy with kecki lris on 18 aug 2012 ( bottom panel of figure [ fig:1222y ] ) shows that this is a quasar at a redshift of 1.56 .
we therefore conclude that this transient is a flaring type - i quasar .
vtc224730 + 000008 is a variable source with flux densities [email protected] , [email protected] , [email protected] , and [email protected] mjy in the four epochs of our survey .
its 1.4 ghz peak flux densities from the first and vla - stripe 82 surveys are @xmath134 mjy and @xmath135 mjy respectively .
the optical counterpart of vtc224730 + 000008 is a blue quasar , classified as a bl lac - type object .
blazar - like optical variability is indeed evident in the ptf and sdss photometric data .
optical spectra from the sdss boss program indicate that the redshift is z@xmath01 . given this information ,
we conclude that vtc224730 + 000008 is an intermediate - redshift blazar .
this is a transient with respect to the first survey .
its flux density is [email protected] , [email protected] , [email protected] , [email protected] mjy in our four survey epochs , e1e4 , with spectral indices @xmath136 , @xmath137 , and @xmath138 in epochs e1e3 , within the s band . the flux density in the first survey at the location of vtc233002 - 002736 is @xmath139 mjy ( mean epoch 1999.2 ) .
this implies an order of magnitude or more increase in flux density at 1.4 ghz over the past decade , and relatively stable flux density at 3 ghz over the past 2 years .
we obtained follow - up observations with the vla ( 115 ghz ) on 01 sep 2012 , 17 sep 2012 , and 29 may 2014 .
these observations reveal a gigahertz - peaked spectrum source with an optically thick spectral index of 2.1 between 13 ghz . on 01 sep 2012 ,
the spectrum peaks at 5 ghz with a flux density of about 10 mjy , and the optically thin spectral index is @xmath140 between 715 ghz . in the 29 may 2014 observations , spectral flattening
is observed between 215 ghz with respect to 01 sep 2012 and the spectral peak appears to be at 3 ghz , suggesting a significant evolution in the spectrum beyond 3 ghz .
the radio continuum spectra from the survey and follow - up observations are shown in the inset of the top panel of figure [ fig : vtc233002 - 002736 ] .
follow - up observations from carma at 100 ghz on 09 sep 2012 give a detection at @xmath141 mjy .
upper limit at 1.4 ghz from the first survey and the upper limit from wise w4 filter are shown as red arrows .
blue circles and green squares indicate the first and second follow - up observations .
black dashed and dotted lines are average spectra for radio - loud and radio - quiet quasars from @xcite .
the _ inset _ shows the zoom - in of the follow - up radio continuum spectra from the three epochs .
red triangles show data from the third follow - up epoch , and the measurements from the four survey epochs appear in grey . _
bottom _ : the optical spectrum of vtc233002 - 002736 observed with diemos at keckii on 14 sep 2012 . the y - axis is flux in arbitrary units .
broad emission lines indicate a quasar at a redshift of 1.65 .
[ sec : radio_transients : gt_one_year : vtc233002 - 002736 ] for details.,title="fig:",width=336 ] upper limit at 1.4 ghz from the first survey and the upper limit from wise w4 filter are shown as red arrows .
blue circles and green squares indicate the first and second follow - up observations .
black dashed and dotted lines are average spectra for radio - loud and radio - quiet quasars from @xcite .
the _ inset _ shows the zoom - in of the follow - up radio continuum spectra from the three epochs .
red triangles show data from the third follow - up epoch , and the measurements from the four survey epochs appear in grey . _
bottom _ : the optical spectrum of vtc233002 - 002736 observed with diemos at keckii on 14 sep 2012 . the y - axis is flux in arbitrary units .
broad emission lines indicate a quasar at a redshift of 1.65 .
[ sec : radio_transients : gt_one_year : vtc233002 - 002736 ] for details.,title="fig:",width=307 ] the sdss counterpart of vtc233002 - 002736 is a reddish point source with @xmath142 mag .
the detection in the ptf is only marginal , and no meaningful optical variability information can be obtained .
the sdss light curve between 19982008 reveals sub - magnitude variability , typical of agn , on shortest timescales of a few weeks .
vtc233002 - 002736 also has an x - ray counterpart in xmm - newton @xcite .
the multi - frequency continuum spectrum of vtc233002 - 002736 is shown in the top panel of figure [ fig : vtc233002 - 002736 ] .
the optical to mid - infrared part of the spectrum agrees well with a type - ii quasar template from @xcite .
a comparison between the radio and optical flux densities ( @xmath143 ) indicates that this is a radio loud quasar .
we obtained a follow - up optical spectrum on 14 sep 2012 with diemos at keck ii ( figure [ fig : vtc233002 - 002736 ] , bottom panel ) .
the spectrum shows broad ciii ] and mgii spectral lines allowing us to calculate the redshift , @xmath144 . from the nature of the radio spectrum , the timescale of its evolution , and the nature of the host galaxy ,
we conclude that vtc233002 - 002736 is due to renewed jet - activity from a type - ii radio loud quasar , where the gps spectrum is indicative of a young jet .
the order - of - magnitude increase in flux density with respect to the first survey could be indicative of an enhanced accretion phenomenon leading to an intensified jet .
the flattening of spectral index at ghz frequencies between 2012 and 2014 may imply cessation in the increased accretion episode and/or the interaction of the jet with the circum - nuclear material in the host galaxy .
this subsection describes two transients that were found while manually inspecting archival radio images and analyzing our 5@xmath58 source catalogs .
being below the cutoff values that we have used , our variability criteria are not able to recover them .
these two exemplars highlight the possibility that there are many more transients in our pilot survey data than have been reported in this work .
vtc221515 - 005028 has flux densities [email protected] , [email protected] , [email protected] , [email protected] mjy and spectral indices @xmath145 , @xmath146 , @xmath147 , and @xmath148 in epochs e1e4 of our survey .
it is absent in the first survey ( mean epoch 1996.1 ) , where the flux density at that location is @xmath149 mjy , but present in the 1.4 ghz stripe 82 survey catalog @xcite with a peak flux density of 0.79 mjy .
for the transient search on timescales @xmath1720 years ( [ sec : radio_transients : gt_one_year ] ) we assumed the upper limit to the flux density in first as 1 mjy , and hence did not recovered this source as a transient using our spectral index criterion .
we carried out radio follow - up observations with the vla ( 115 ghz ) on 29 may 2014 .
the radio continuum spectrum are shown in figure [ fig : vtc221515 - 005028 ] .
the source appears to have a flat spectrum between 15 ghz and a spectral index of @xmath150 between 515 ghz .
upper limit from the first survey is shown with the red arrow and the detection from the @xcite survey is marked by a red circle .
he measurements from the four survey epochs are shown in grey .
[ sec : radio_transients : serendipitous : vtc221515 - 005028 ] for details.,width=326 ] the optical counterpart of vtc221515 - 005028 is a faint red galaxy having @xmath151 mag . from the ptf light curve
it is seen to exhibit stochastic variability with a maximum amplitude of one magnitude in the r band , and having a shortest timescale of about three days .
the sdss light curve between 19992008 reveals sub - magnitude variability and intermittent flares with amplitudes upto @xmath152 mag on @xmath01 day timescale .
the photometric redshift from sdss is 0.37 .
the radio - to - optical flux density ratio for the host galaxy , @xmath153 , and the radio luminosity at 1.4 ghz , @xmath154 , suggest a low - luminosity radio - loud agn .
we thus conclude that vtc221515 - 005028 is an extremely variable flat - spectrum agn .
vtc230241 + 003450 is detected as a 6@xmath58 source on 15 aug 2012 ( epoch e2 , regions r2 ) and is not detected in the other epochs .
the peak pixel values at the location of this transient in epochs e1e4 are [email protected] , [email protected] , [email protected] , and [email protected] mjy respectively . in the first and @xcite surveys , the peak pixel values at the location of the transient candidate are @xmath109 and @xmath110 mjy respectively .
the optical counterpart of this transient is a 11th magnitude star , sdss j230241.41 + 003450.2 , classified as a k4ve star by @xcite .
there is no evidence for binarity .
fitting a blackbody to the photometric data from sdss , the nomad catalog , and wise gives an estimate of the effective temperature , 3800@xmath81500 k , and the distance , 70@xmath8120 pc ( assuming the radius of a main sequence star ) .
sdss j2302 + 00 derive a @xmath155 sin(@xmath156 ) of 85 km s@xmath5 .
this star has a rosat counterpart , 1rxs j230240.3 + 003453 , 17@xmath8112 away , whose hardness ratios are consistent with those of a coronal emitter .
we used a 1 kev apec model in webpimms to convert the rosat / pspc countrate to a flux of @xmath157 in the 0.12.4 kev energy band .
the x - ray - to - optical flux ratio is 0.002 , and the x - ray luminosity at a distance of 70 pc is @xmath158 , both values being consistent with a dke star .
the quiescent 3 ghz radio luminosity of vtc230241 + 003450 can therefore be estimated as @xmath159 erg s@xmath5 hz@xmath5 @xcite .
this can be compared with the flaring luminosity , @xmath160 erg s@xmath5 hz@xmath5 .
we conclude that this transient is a flare from a dke star .
ptf carried out a concurrent optical survey which resulted in the identification of a few hundred thousand sources per epoch down to a limiting magnitude of r@xmath16121 mag1011 mag .
this represents the lower limiting magnitude of the optical study carried out here . ] . our radio survey further benefits from the presence of sdss deep co - add images containing more than a million objects over the 50 deg@xmath1 to [email protected] mag @xcite . to find optical counterparts of radio sources we followed the procedure from @xcite .
we matched each source in the e2cat with the sdss and ptf catalogs using a 15 matching radius , and selected only the nearest match .
to understand the false matching rate , we repeated the search by offsetting the radio source positions by 1 in an arbitrary direction .
the resulting number of matches as a function of the matching radius is plotted as a histogram in figure [ fig : matching ] .
based on these results we choose a matching radius of 1 for sdss and 1.5 for ptf to achieve a false matching rate of less than 3% and completeness better than 85% .
the ptf source positions for the faintest objects are known to have a larger scatter than theoretically expected , and hence the larger matching radius of 1.5 is reasonable . using these matching radii
, we found the optical counterparts of sources in the psc .
the corresponding matching fraction in sdss as a function of the r - band magnitude is shown graphically in figure [ fig : radio_optical_quiescent ] .
49% of the radio sources in the psc have an optical counterpart down to the sdss r - band limit35% . ] .
@xcite , using the deep co - added sdss images ( limit [email protected] mag ) , found a matching ratio ( within a 1 radius ) of 44.4% .
for radio variable sources , we found a somewhat higher matching fraction of 63% ( figure [ fig : radio_optical_quiescent ] ) , which is also in agreement with @xcite . mag . ) and sdss ( stripe 82 co - add ; @xmath162 mag . ) are shown as red dashed lines.,width=345 ] the modest optical matching rate in our medium - wide , medium - deep survey was not unexpected . @xcite compared the sdss ( limit r@xmath022 mag ) and the first ( 5@xmath1631 mjy ) surveys over in 1230 deg@xmath1 of the sky and matched about 30% of the 108,000 first radio sources .
@xcite looked at the 382,892 first sources in the north galactic cap and looked for optical matches using the apm scans of poss - i plates ( r@xmath020 mag , [email protected] mag ) for a 18% identification rate ( 70,000 sources ) .
the optical match rate improves substantially in deep , narrow surveys .
for example , @xcite use australia telescope compact array ( atca ) data taken toward the hubble deep field ( hdf ) at 20 cm , 11 cm , 6 cm and 3.6 cm , establishing a 66% matching of optical counterparts to [email protected] mag .
there is a strong color dependence , with the matching rate increasing from the blue to the near infrared @xcite .
match rates approach 100% with the use of deep infrared data @xcite .
the near - real - time optical transient search carried out via image subtraction during the three - month high - cadence ptf observations resulted in approximately 8 million detections .
following standard practice , about 0.9 million of these ( corresponding to only @xmath050,000 unique optical sources ) were identified in machine learning software as unlikely to be image subtraction artifacts , and further filtering was carried out using several stringent selection criteria ( such as : at least two detections within one hour , no coincidence with stellar counterparts or agn , etc . ) and automated classifiers @xcite .
this was followed by human inspection of the subtracted images , light curves , and automated classifications , and the list was narrowed down to 193 candidates for further follow - up . out of the 193 candidates ,
only 10 sources were ultimately followed up spectroscopically , among which are 8 confirmed supernovae , ptf12gzk ( sn ic , [email protected] ) , 12jaa ( iib , 0.024 ) , 12giy ( ia , 0.029 ) , 12hwb ( ia , 0.056 ) , 12hmx ( ia , 0.085 ) , 12iet ( ii , 0.095 ) , 12ild ( ia , 0.17 ) , and 12itq ( ia , 0.22 ) .
the multi - wavelength observations of ptf12gzk have been discussed at length by @xcite . in figure
[ fig : sne ] we plot the optical light curves of these eight supernovae as well as the upper limits to their spectral luminosity in the radio .
the radio non - detections are in accordance with the expected flux densities of type ii , ic and ia supernovae ( table [ tab : transients_summary ] ) . the radio detection and subsequent non - detections of ptf12gzk from @xcite are also shown for reference , and these data emphasize the need for deep radio observations for optically - bright supernovae such as the ones generally found by ptf .
after all the survey observations were complete , we compiled optical light curves of sources using the catalogs hosted at ipac ( [ sec : optical_processing ] ) . in order to enable robust variability search , we selected only those optical sources having more than 16 reliable observations ( unflagged in sextractor ) and r band magnitudes between 10 and 23 .
there are 402,747 such sources in the ptf database with the 90% completeness of this sample corresponding to approximately an r band limiting magnitude of 21 .
802 ( 0.2% ) of these sources have counterparts in the psc ( i.e. 22% of the psc sources have optical counterparts ) . 13,667 ( 3.4% )
are optical variables probability less than 1 in 500,000 and variability @xmath430% , i.e. a standard deviation of @xmath40.28 mag .
were selected as optical variables .
these criteria were designed to be similar to the radio variability criteria . ] .
only 42 radio sources have optical variable counterparts among the sample of 3652 , while only two of these are variable also in the radio .
thus , given the limiting magnitude of r@xmath16421 mag . in optical and our source detection threshold of @xmath00.5 mjy in the radio , we find that the overlap between optical variables and radio variable sources is extremely small .
these demographics are succinctly presented in a venn diagram in figure [ fig : venn ] .
radio upper limits from the jansky vla survey ( four epochs ) are shown as downward - pointing triangles .
for ptf12gzk , the radio detection ( red circle and errorbar ) and 3@xmath58 upper limits ( downward - pointing red arrows ) at 5 ghz from @xcite are also marked .
[ sec : optical_properties ] for details.,width=336 ] here we are less concerned about optical - only transients and will focus instead on the optical variability properties of the radio transients and variables .
a study of optical - only variables and transients from synoptic surveys is better done elsewhere in literature ( e.g. * ? ? ?
* ; * ? ? ?
we have obtained photometric data from ptf and sdss after searching for counterparts as described above .
a compilation of the optical light curves of the radio variable sources is given in figure [ fig : optical_variability ] .
majority of the ptf light curves reveal sub - magnitude variability , while the sdss light curves show gradual sub - magnitude or magnitude - level variability on timescales of years .
this is characteristic of agn , where fluctuations in the accretion rate or other causes give rise to optical variability on a wide range of timescales .
coupled with radio variability information , it seems likely that all of these sources are agn , excepting the ones associated with known stars .
we have carried out a dedicated radio transient survey in a 50 deg@xmath1 region of the sdss stripe 82 .
this survey is a pilot for the caltech - nrao stripe 82 survey ( cnss ) , a multi - epoch survey of the entire @xmath0270 deg@xmath1 of stripe 82 , which is currently underway .
the pilot observations were carried out with the jansky vla between 24 ghz over four epochs spaced logarithmically in time , i.e. having cadences of one week , one month , and 1.5 years .
the median rms noise per epoch is between 50 @xmath2jy and 90 @xmath2jy . with this pilot
we have successfully demonstrated the near - real - time calibration , imaging , and transient search capability with the jansky vla data .
we have found 142 sources displaying fractional variability beyond 30% . based on radio follow - up observations , radio and optical variability , radio spectral indices , and multi - wavelength archival photometric data ,
almost all of these are likely to be due to shocks in agn jets .
we have also found two bona fide radio transients associated with galactic objects , a flare from an rs cvn binary and a dke flare .
comparison of our pilot survey data with the first and vla - stripe 82 surveys has ( unexpectedly ) revealed a few additional , highly variable and transient , sources on timescales of 520 years .
these sources , most of which are either previously - known radio sources or have optical agn hosts , are likely associated with renewed agn activity .
two other transients , found serendipitously , suggest that there are indeed many more transients in this dataset than the ones which we have found through our stringent selection criteria .
we encourage the astronomical community to take advantage of the data from this pilot radio survey ( see [ sec : radio_processing : final : dr ] for the data release website ) for finding these hidden transients and for other science .
a high - cadence optical survey using ptf was carried out over three months , contemporaneous with the first three epochs of the radio survey .
the motivation behind this joint survey was the selection of radio transients for rapid follow - up , preliminary classification of radio variables and transients based on optical light curves , and a direct comparison of the dynamic radio and optical skies .
as noted by previous radio and optical studies ( see [ sec : optical_properties ] ) , 20% of the persistent radio sources have an optical counterpart above the ptf limiting magnitude of 21 , while 50% have counterparts above the sdss limiting magnitude of 23.5 in the red filter .
we have found primarily slowly - evolving radio transients , and , within the sensitivity limits of our ptf and pilot jansky vla surveys , very little overlap between the optical and radio variable skies .
only 0.05% of the radio point sources and @xmath200.001% of the optical sources are variable at both frequencies .
the fraction of strong variables per square degree of the radio decimetric sky is at least an order of magnitude smaller than those in the optical sky .
we will now use the radio data to assess the degree of variability and the transient rate of the radio sky , and make recommendations for future surveys .
we found only a small fraction ( @xmath165 ) of the point sources varying by @xmath430% on week - month - year timescales .
majority of these are variable only on a 1.5 year timescale , and as described in
[ sec : radio_transients : final ] , this variability fraction is to be taken as the upper limit . several previous studies at 1.4 ghz @xcite have shown that the fraction of variables on timescales between minutes and years , and flux densities between 0.1 mjy and 100 mjy , is @xmath01% or less .
similar to these studies , we see that the 3 ghz sky is not highly variable on timescales longer than a few days .
radio follow - up observations , optical and radio light curves , and multi - wavelength archival data of the radio variable sources suggest that almost all of these are shock - related flaring in the jets of agn .
our choice of modulation index for measuring variability and selection threshold introduces a bias towards selecting sources with large flux densities .
a source in the psc would have to have a mean flux density of 1.4 mjy in order to detect a 30% fractional variability , while at the 7@xmath58 source detection limit of of @xmath00.5 mjy , a source would have to have @xmath166 ( @xmath167120% ) in order to be identified as a significant variable .
fifteen strong variables ( i.e. sources having @xmath168 , or @xmath167100% ) were identified in our survey , but only 3000 sources in the psc are bright enough ( @xmath40.55 mjy ) to have been identified as a strong variable .
we can therefore conclude that the fraction of strong variables is less than one percent .
radio variability appears to be a function of timescale ; the variability on hours@xmath0days timescales contrasts with the variability on weeks@xmath0years timescales .
@xcite found that @xmath030% of point sources brighter than 1.5 mjy at 5 ghz were variable , with majority of the sources varying on timescales @xmath2010 days .
this variability is primarily small amplitude ( modulation indices of less than 0.2 ) .
the structure function of the variable sources constructed by @xcite shows a sharp rise on timescales of @xmath201 day , a shallow but steady rise between 110 days , and a roughly constant value beyond 10 days .
a large fraction of variables displaying small variability amplitudes on short timescales were also found by the masiv survey @xcite conducted at 5 ghz .
the findings of @xcite and @xcite suggest that the low - amplitude variability on timescales of @xmath2010 days is mostly extrnisic . among the @xcite radio sources , only 0.3% have modulation indices
greater than 0.2 on timescales less than two years .
the variability fraction that we find on week - month timescales in our pilot survey is similar ( @xmath171% ) , and agrees with the variability fraction found in narrow - deep surveys by @xcite , @xcite , and @xcite , wide - field surveys such as @xcite , and other studies between 15 ghz on week - month - year timescales . on 1.5 year timescale
we find the variability fraction to be less than 4% .
recently , @xcite compared the 1.4 ghz vla - stripe 82 and first surveys , and found @xmath012% ( 6 percent per epoch @xmath15 2 epochs ) of the sources having fractional variablility larger than 0.3 on timescales between 7 years and 22 years , the majority of which were agn . while it is possible that , on these longer timescales , the fraction of variable sources in the radio sky is large , some of the @xcite variables could be artificial and solely due to angular resolution differences between the vla - stripe 82 and first surveys .
indeed , @xcite , @xcite , and @xcite find that the fraction of variables on similar timescales is a few percent or less .
radio variability ( especially extrinsic ) is expected to have a frequency dependence , but this aspect of variability has not been extensively studied .
we can only use the results of past blind surveys to conclude that the fraction of strong variables is less than a few percent between frequencies of 1 ghz and 5 ghz , between flux densities of @xmath00.3 mjy and @xmath0100 mjy , and over a wide gamut of timescales ( between one day and several years ) . in our pilot survey , we found a single agn , vtc233002 - 002736 , with flux density @xmath43 mjy at 1.4 ghz , that appears to have increased in flux density by more than a factor of ten over the past 15 years .
similar objects may have been found earlier , by @xcite , in the sumss survey ( e.g. j201524 - 395949 or j060938 - 333508 ) .
it is likely that such phenomena are a result of episodes of enhanced accretion leading to increased jet activity ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for discussions of possibly related phenomena ) .
assuming a timescale of @xmath020 years for an enhanced accretion episode , and given that fact that 50 deg@xmath1 of the sky has @xmath02000 agn with flux density @xmath43 mjy @xcite , we can estimate the period of occurrence of such episodes over the lifetime of an agn : @xmath040,000 years .
this is remarkably consistent with previous studies @xcite suggesting some young radio - loud agn to have short - lived jets operating on time scales of 10@xmath16910@xmath170 years .
we searched our four - epoch dataset for transients and found only a single source in the psc ( vtc223612 + 001006 ; rscvn ) that was present in one epoch and absent in the rest .
our transient search was carried out over a single - epoch area of @xmath052 deg@xmath1 , but the sensitivity is not uniform across this area .
for the first three epochs , our transient search was conducted on single - pointing images out to a radius ( @xmath171 ) of @xmath08 from the pointing center .
although the fourth epoch has fairly uniform rms noise , the comparison with the first three epochs for transient search diminishes the significance of the added sensitivity for this epoch .
we can approximate the transient rate larger than a threshold flux density , @xmath172 , using the formulation from appendix c of @xcite .
we use a gaussian primary beam response with half - width at half - maximum ( @xmath173 ) of 7.5 .
the 2@xmath58 upper and lower limits for the number of transient events , n@xmath175 , in 50 deg@xmath1 given that we found one event are 5.683 and 2.3@xmath1510@xmath6 @xcite . since the mean rms noise ( flat sky ) for the first three epochs of our survey is 70 @xmath2jy , we use a 7@xmath58 detection threshold of 0.5 mjy . for rs cvn variables ( active binaries )
we calculate a rate of @xmath176(@xmath40.5 mjy)@xmath1640.0081@xmath177 events deg@xmath6 . for all other types of transients
we derive a 95% confidence upper limit of @xmath176(@xmath40.5 mjy)@xmath200.024 events deg@xmath6 .
the logn - logs plots for galactic and extragalactic slow radio transients are shown in figure [ fig : lognlogs ] .
the top panel shows the upper limits to the transient rates derived from previous radio surveys ( colored wedges ) , the rates derived from radio transient detections ( filled circles with errorbars ) , and the theoretically - expected / empirically - estimated transient rates ( dashed gray lines ; see table [ tab : transients_summary ] for more details ) .
it should be noted that the expected transient rates are not sharp lines but are probability surfaces in the logn - logs diagram with the `` most - probable '' rates reported as dashed lines .
the dashed line labeled `` tde '' represents swift j1644 + 57-like @xcite events .
upper limits from a few radio surveys which do not probe any new part of the phase space are not shown in this figure .
all observed quantities are color - coded according to the observing frequency .
the solid gray line is the rate claimed by @xcite , plotted for reference .
the upper limit from our pilot survey and the phase space probed by the full cnss survey are shown as thick green wedges .
the phase space probed by the vla sky survey ( vlass ) all - sky tier is shown as a thin green wedge .
the source counts from the first survey are represented by the solid red line , and the dashed red line denotes 1% of these persistent sources , representing strong variable sources at 1.4 ghz ( e.g. * ? ? ? * and references therein ) .
@xcite report only a single extragalactic transient , sumss j060938 - 333508 , found to be a nuclear source from atca follow - up observations ( keith bannister , private communication ) .
hence the transient rate is @xmath178 events deg@xmath6 .
@xcite report 57 transients , but some of these are galactic and others have indefinite classifications .
hence , we adopt a 95% confidence level upper limit of 71 transients .
the lower panel of figure [ fig : lognlogs ] shows the galactic transient phase space .
symbols have similar meanings as for the extragalactic plot ( top panel ) . for reference ,
the source counts from the first and the magpis 1.4 ghz @xcite surveys are denoted by black solid lines .
the approximate source counts for variable galactic sources from @xcite are denoted by the blue dashed line .
the transient rate for active binaries derived from this work is shown by the green errorbar and the upper limit for all other classes of galactic transients is denoted by a thick green wedge .
it is evident from these logn - logs diagrams that our pilot survey is not sensitive and wide enough to discover extragalactic explosive transients , but it is already in the regime where stellar flares are expected .
our contemporaneous optical survey with ptf allows us to make a direct comparison between the transient optical and radio skies . within the limiting sensitivities of our experiment
, the transient skies at these two frequencies appear to be quite distinct . in the three months of high - cadence ptf search ,
eight spectroscopically - confirmed supernovae were discovered , none of which were recovered in our radio survey .
conversely , three stellar flares were detected in the radio , but their optical counterparts are saturated in ptf . we thus conclude that future joint radio and optical searches , such as meerkat and meerlicht , may only be beneficial if shallow optical surveys are combined with deep radio surveys or vice versa , or if both surveys are deep ( @xmath2jy - level sensitivity ) .
we find that deep multi - wavelength photometric data and spectroscopy are very important for host and progenitor classification of extragalactic radio transients , and this should inform future slow transient searches .
given the expensive radio observing time , we have attempted to distinguish between agn and other transients based on optical light curves from ptf and sdss .
although in the present study we have optical light curves only for a small fraction of our radio sources ( due to the completeness of optical counterparts ) , we expect to have 100% completeness in the lsst era , when radio transient classification based on optical light curves may be a feasible approach .
the radio sky at frequencies of a few ghz appears to be quiet , with less than a percent of the persistent sources being strong variables . from table
[ tab : transients_summary ] and figure [ fig : lognlogs ] we find that the rates for a majority of classes of slow radio transients is less than one transient per 10,000 persistent sources .
we also find that the fraction of strong variables and explosive transients among persistent sources in the optical sky is similar to the radio .
however , the large number of persistent optical sources per square degree of the sky ( down to the limiting sensitivity of ptf , @xmath179 mag . , for example ; see figure [ fig : venn ] ) implies that the number of varying optical sources is much larger than the number of varying radio sources ( down to the limiting sensitivity of our jansky vla survey , @xmath00.5 mjy ) .
accordingly , we conclude that the false positive foreground for future em - gw searches is lower in the radio than in the optical mag . corresponds to 12@xmath2jy . ] .
the rates for extragalactic transients , like core - collapse supernovae and binary neutron star mergers , are significantly low ( table [ tab : transients_summary ] ) , and it is not surprising that we found none of these transients in our pilot radio survey . surveys over large areas of the sky are thus motivated by the search for such exotica .
therefore , we recommend wide - field surveys like that of the entire @xmath0270 deg@xmath1 of stripe 82 with the jansky vla ( the cnss survey , which is currently underway ) , all - sky surveys like the vla sky survey ( vlass ; the all - sky tier has been recently approved ) , or similar surveys with askap , meerkat , and wsrt / apertif . also , shallow radio surveys are preferred to deep surveys due to the added advantage of restricting the transient population to a low redshift space , thus making optical counterpart search feasible .
_ acknowledgemets : the authors extend sincere thanks to joan wrobel and other scheduling staff at the nrao in socorro for extensive help with the scheduling of the vla observations , and to james robnett and other computing staff for their untiring assistance with the data storage and allocation of computing resources .
the authors also wish to thank luis ho , branimir sesar , eran ofek , sanjay bhatnagar , urvashi rau , kumar golap , vivek dhawan , craig walker , talvikki hovatta , tim pearson , anthony readhead , chuck steidel , and allison strom for insightful discussions . the contribution of ptf collaboration members to optical data processing and optical follow - up observations relevant for this project is acknowledged .
kpm is grateful to nrao for the grote reber fellowship , and to yamini jangir for going over this manuscript and providing useful suggestions .
we thank the anonymous referee for comments which helped in improving certain parts of the manuscript .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
srk s research in part is supported by nasa and nsf .
some of the data presented herein were obtained at the w.m .
keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w.m .
keck foundation .
this research has made extensive use of the ads , cds ( vizier and simbad ) , ned , sdss , and irsa online databases , and the python , matplotlib , and astropy software packages .
_ given a total observing time t , we can either do a wide - shallow or a narrow - deep survey , where wide / narrow refers to the area covered , and shallow / deep refers to the sensitivity in terms of limiting flux density . to find which of these will explore the largest volume , consider two surveys , one which observes a single pointing for time t , and the other which observes n pointings with time t / n alloted for each pointing .
we know that the limiting flux density ( s@xmath180 ) , integration time ( @xmath181 ) , maximum distance ( d@xmath180 ) and the total volume ( v@xmath180 ) probed by a survey are related as : @xmath182 thus we see that the volume probed by a wide - shallow survey ( v@xmath183 ) is larger than that seen by a narrow - deep survey ( v@xmath184 ) by a factor of the number of pointings to the power of 1/4 .
however , note that since the distance probed in these two kinds of surveys goes as @xmath185 .
this means that if we want to probe transient populations concentrated at large distances ( redshifts ) , it is best to do a narrow - deep survey .
also see @xcite for survey parameter optimization to maximize the yield of slow transients having well - defined distributions in space .
lrrrrrrrrrccccl + + 220210 - 000203 & 330.539900 & -0.034252 & [email protected] & [email protected] & [email protected] & [email protected] & -0.56 & @xmath200.50 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 220609 - 004654 & 331.536395 & -0.781711 & [email protected] & [email protected] & [email protected] & [email protected] & -0.33 & [email protected] & [email protected] & 20.7 & nv & 0.37 & 31.1 & agn + 221122 + 005202 & 332.842184 & 0.867327 & [email protected] & [email protected] & [email protected] & [email protected] & 0.55 & [email protected] & [email protected] & 20.0 & nv & 0.31 & 30.7 & agn + 221136 + 011142 & 332.898974 & 1.194886 & [email protected] & [email protected] & [email protected] & [email protected] & -0.36 & [email protected] & [email protected] & 24.1 & @xmath186 & @xmath186 & @xmath186 & agn + 221226 + 004322 & 333.107130 & 0.722873 & [email protected] & [email protected] & [email protected] & [email protected] & 0.31 & [email protected] & [email protected] & 23.1 & @xmath186 & @xmath186 & @xmath186 & agn + 221406 + 002305 & 333.524900 & 0.384618 & [email protected] & [email protected] & [email protected] & [email protected] & 0.58 & @xmath200.51 & @xmath200.32 & 18.4 & nv & 0.15 & 29.5 & agn + 221541 + 010856 & 333.921049 & 1.148981 & [email protected] & [email protected] & [email protected] & [email protected] & -0.48 & @xmath200.69 & [email protected] & 19.9 & @xmath186 & 0.32@xmath187 & 30.5 & agn + 221713 - 002216 & 334.306166 & -0.371232 & [email protected] & [email protected] & [email protected] & [email protected] & -0.33 & @xmath200.53 & [email protected] & 23.8 & @xmath186 & @xmath186 & @xmath186 & agn + 221913 + 011147 & 334.805979 & 1.196250 & [email protected] & [email protected] & [email protected] & [email protected] & 0.90 & [email protected] & [email protected] & 19.1 & @xmath186 & 0.20@xmath187 & 30.0 & agn + 222009 - 010814 & 335.036567 & -1.137346 & [email protected] & [email protected] & [email protected] & [email protected] & -0.38 & [email protected] & [email protected] & 20.6 & nv & @xmath186 & @xmath186 & qso + 222150 - 000325 & 335.460301 & -0.056997 & [email protected] & [email protected] & [email protected] & [email protected] & 0.29 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 222232 + 011104 & 335.634687 & 1.184480 & [email protected] & [email protected] & [email protected] & [email protected] & 0.80 & [email protected] & [email protected] & 19.5 & @xmath186 & 0.29 & 30.2 & agn + 222236 + 010556 & 335.648844 & 1.098840 & [email protected] & [email protected] & [email protected] & [email protected] & 0.96 & @xmath200.72 & [email protected] & 18.4 & nv & 0.20 & 29.9 & agn + 222330 + 010227 & 335.873148 & 1.040835 & [email protected] & [email protected] & [email protected] & [email protected] & 0.40 & [email protected] & [email protected] & 19.3 & nv & @xmath186 & @xmath186 & qso + 222520 + 004939 & 336.331883 & 0.827551 & [email protected] & [email protected] & [email protected] & [email protected] & -0.44 & [email protected] & [email protected] & 22.3 & @xmath186 & 0.80@xmath187 & 31.9 & agn + 222758 + 003706 & 336.992096 & 0.618294 & [email protected] & [email protected] & [email protected] & [email protected] & -0.30 & [email protected] & [email protected] & 18.1 & nv & @xmath186 & @xmath186 & qso + 223607 + 005259 & 339.028010 & 0.883015 & [email protected] & [email protected] & [email protected] & [email protected] & -0.37 & @xmath200.50 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 223624 + 004021 & 339.099104 & 0.672614 & [email protected] & [email protected] & [email protected] & [email protected] & 0.43 & @xmath200.50 & [email protected] & 22.3 & @xmath186 & 0.59@xmath187 & 30.9 & agn + 224843 - 005459 & 342.178945 & -0.916381 & [email protected] & [email protected] & [email protected] & [email protected] & -0.27 & @xmath200.72 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 224922 + 001805 & 342.343018 & 0.301369 & [email protected] & [email protected] & [email protected] & [email protected] & 0.27 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 225411 - 010651 & 343.544297 & -1.114119 & [email protected] & [email protected] & [email protected] & [email protected] & 0.85 & @xmath200.74 & @xmath200.60 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 230117 - 000204 & 345.322041 & -0.034390 & [email protected] & [email protected] & [email protected] & [email protected] & 0.37 & [email protected] & [email protected] & 19.6 & nv & 0.28@xmath187 & 30.6 & agn + 230133 - 002538 & 345.387209 & -0.427294 & [email protected] & [email protected] & [email protected] & [email protected] & -0.80 & @xmath200.61 & [email protected] & @xmath186 & @xmath186 & 0.84@xmath187 & 31.0 & agn + 230556 - 001652 & 346.481658 & -0.281193 & [email protected] & [email protected] & [email protected] & [email protected] & -0.46 & @xmath200.51 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 230636 - 002609 & 346.649268 & -0.435896 & [email protected] & [email protected] & [email protected] & [email protected] & -0.58 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 230918 + 002642 & 347.324430 & 0.445083 & [email protected] & [email protected] & [email protected] & [email protected] & -0.73 & @xmath200.50 & @xmath200.39 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 231437 + 003844 & 348.655518 & 0.645515 & [email protected] & [email protected] & [email protected] & [email protected] & -0.42 & @xmath200.53 & [email protected] & 24.9 & @xmath186 & @xmath186 & @xmath186 & agn + 231444 + 004026 & 348.685137 & 0.673960 & [email protected] & [email protected] & [email protected] & [email protected] & 0.89 & @xmath200.53 & [email protected] & 23.9 & @xmath186 & @xmath186 & @xmath186 & agn + 231559 - 001205 & 348.994359 & -0.201415 & [email protected] & [email protected] & [email protected] & [email protected] & -0.31 & [email protected] & [email protected] & 23.1 & @xmath186 & @xmath186 & @xmath186 & agn + 231746 - 005355 & 349.441332 & -0.898619 & [email protected] & [email protected] & [email protected] & [email protected] & -0.44 & @xmath200.69 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 231942 - 004547 & 349.925567 & -0.763046 & [email protected] & [email protected] & [email protected] & [email protected] & 0.29 & [email protected] & [email protected] & 23.9 & @xmath186 & @xmath186 & @xmath186 & agn + 232217 + 001252 & 350.569589 & 0.214481 & [email protected] & [email protected] & [email protected] & [email protected] & -0.59 & @xmath200.49 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 232226 + 010357 & 350.606463 & 1.065718 & [email protected] & [email protected] & [email protected] & [email protected] & -0.42 & [email protected] & @xmath186 & 17.4 & @xmath186 & 0.12 & 30.0 & agn + 232634 - 010513 & 351.642405 & -1.086815 & [email protected] & [email protected] & [email protected] & [email protected] & -0.40 & [email protected] & @xmath186 & 23.4 & @xmath186 & 0.79@xmath187 & 31.5 & agn + 232656 - 000438 & 351.734689 & -0.077164 & [email protected] & [email protected] & [email protected] & [email protected] & -0.37 & [email protected] & @xmath186 & 24.1 & @xmath186 & @xmath186 & @xmath186 & agn + 232723 - 000507 & 351.843984 & -0.085213 & [email protected] & [email protected] & [email protected] & [email protected] & 0.27 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 232933 - 004002 & 352.386845 & -0.667239 & [email protected] & [email protected] & [email protected] & [email protected] & 0.26 & [email protected] & @xmath186 & 22.4 & @xmath186 & @xmath186 & @xmath186 & agn + 233106 + 002607 & 352.774153 & 0.435175 & [email protected] & [email protected] & [email protected] & [email protected] & -0.40 & [email protected] & @xmath186 & 22.2 & @xmath186 & 0.74 & 31.1 & agn + + 220127 + 001402 & 330.360610 & 0.233772 & [email protected] & [email protected] & [email protected] & [email protected] & 0.82 & @xmath200.49 & @xmath186 & 23.9 & @xmath186 & @xmath186 & @xmath186 & agn + 220456 - 000147 & 331.232587 & -0.029821 & [email protected] & [email protected] & [email protected] & [email protected] & -0.27 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 220643 - 003102 & 331.680375 & -0.517180 & [email protected] & [email protected] & [email protected] & [email protected] & -0.39 & [email protected] & [email protected] & 19.5 & v & @xmath186 & @xmath186 & qso + 220910 - 001601 & 332.293354 & -0.267023 & [email protected] & [email protected] & [email protected] & [email protected] & -0.34 & [email protected] & [email protected] & 19.5 & nv & 1.11 & 32.4 & qso + 221308 - 010837 & 333.284558 & -1.143507 & [email protected] & [email protected] & [email protected] & [email protected] & 0.84 & @xmath200.72 & [email protected] & 23.2 & @xmath186 & @xmath186 & @xmath186 & agn + 221350 - 011130 & 333.460186 & -1.191627 & [email protected] & [email protected] & [email protected] & [email protected] & 0.59 & @xmath200.71 & @xmath200.28 & 20.0 & @xmath186 & 1.96 & 32.0 & qso + 221959 + 011045 & 334.994454 & 1.179204 & [email protected] & [email protected] & [email protected] & [email protected] & 0.27 & [email protected] & [email protected] & 21.0 & @xmath186 & 0.84 & 31.9 & qso + 222123 - 002509 & 335.345654 & -0.419074 & [email protected] & [email protected] & [email protected] & [email protected] & 0.37 & [email protected] & [email protected] & 23.4 & @xmath186 & @xmath186 & @xmath186 & agn + 222942 + 003556 & 337.423504 & 0.598817 & [email protected] & [email protected] & [email protected] & [email protected] & 0.31 & [email protected] & [email protected] & 24.5 & @xmath186 & @xmath186 & @xmath186 & agn + 223317 - 005009 & 338.321408 & -0.835723 & [email protected] & [email protected] & [email protected] & [email protected] & 0.38 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 223612 + 001007 & 339.050505 & 0.168559 & [email protected] & [email protected] & [email protected] & [email protected] & 0.99 & @xmath200.52 & @xmath200.39 & 9.7 & @xmath186 & @xmath186 & @xmath186 & rs cvn + 224036 + 010852 & 340.150070 & 1.147826 & [email protected] & [email protected] & [email protected] & [email protected] & -0.31 & [email protected] & [email protected] & 21.0 & @xmath186 & 2.99 & 32.7 & qso + 224448 - 000619 & 341.200443 & -0.105397 & [email protected] & [email protected] & [email protected] & [email protected] & 0.27 & [email protected] & [email protected] & 18.9 & nv & @xmath186 & @xmath186 & qso + 224657 + 005240 & 341.737102 & 0.877754 & [email protected] & [email protected] & [email protected] & [email protected] & 0.31 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 225649 - 005401 & 344.205944 & -0.900221 & [email protected] & [email protected] & [email protected] & [email protected] & 0.32 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 225707 - 010238 & 344.280373 & -1.043805 & [email protected] & [email protected] & [email protected] & [email protected] & -0.68 & @xmath200.74 & @xmath200.39 & 20.1 & nv & 1.56 & 31.8 & qso + 230236 + 005739 & 345.647977 & 0.960878 & [email protected] & [email protected] & [email protected] & [email protected] & 0.27 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 230803 - 000145 & 347.012572 & -0.029111 & [email protected] & [email protected] & [email protected] & [email protected] & -0.60 & @xmath200.52 & @xmath200.29 & 18.5 & nv & 0.15 & 29.5 & agn + 230810 + 002117 & 347.039702 & 0.354644 & [email protected] & [email protected] & [email protected] & [email protected] & -0.38 & [email protected] & [email protected] & 21.1 & nv & 0.40@xmath187 & 30.7 & agn + 231242 + 002451 & 348.174063 & 0.414044 & [email protected] & [email protected] & [email protected] & [email protected] & 0.30 & [email protected] & [email protected] & 18.7 & nv & 1.90 & 32.5 & qso + 231334 - 001645 & 348.390098 & -0.279113 & [email protected] & [email protected] & [email protected] & [email protected] & 0.48 & [email protected] & [email protected] & 22.6 & @xmath186 & 0.74 & 31.4 & agn + 231455 + 002456 & 348.728453 & 0.415556 & [email protected] & [email protected] & [email protected] & [email protected] & 0.43 & [email protected] & [email protected] & 24.8 & @xmath186 & @xmath186 & @xmath186 & agn + 232125 - 004845 & 350.352737 & -0.812522 & [email protected] & [email protected] & [email protected] & [email protected] & 0.32 & [email protected] & @xmath186 & 21.8 & @xmath186 & 0.58@xmath187 & 31.2 & agn + 232526 - 003623 & 351.359745 & -0.606393 & [email protected] & [email protected] & [email protected] & [email protected] & -0.27 & [email protected] & @xmath186 & 21.4 & nv & 0.63 & 31.4 & agn + 232548 - 011134 & 351.451958 & -1.192766 & [email protected] & [email protected] & [email protected] & [email protected] & -0.29 & [email protected] & @xmath186 & 20.7 & @xmath186 & 0.46 & 31.5 & agn + + 215701 + 005124 & 329.252431 & 0.856537 & [email protected] & [email protected] & [email protected] & [email protected] & 0.40 & [email protected] & @xmath186 & 22.9 & @xmath186 & 0.57@xmath187 & 30.9 & agn + 215732 + 010748 & 329.384654 & 1.130011 & [email protected] & [email protected] & [email protected] & [email protected] & 0.39 & @xmath200.78 & @xmath186 & 19.1 & nv & 0.30 & 30.4 & agn + 215929 + 004723 & 329.872536 & 0.789814 & [email protected] & [email protected] & [email protected] & [email protected] & -0.27 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 215951 + 010041 & 329.963762 & 1.011278 & [email protected] & [email protected] & [email protected] & [email protected] & -0.37 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 220005 + 002309 & 330.022192 & 0.385698 & [email protected] & [email protected] & [email protected] & [email protected] & 0.31 & @xmath200.51 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 220017 - 000134 & 330.072365 & -0.026093 & [email protected] & [email protected] & [email protected] & [email protected] & 0.26 & [email protected] & @xmath186 & 19.8 & nv & 0.61 & 31.7 & qso + 220109 + 010124 & 330.288600 & 1.023416 & [email protected] & [email protected] & [email protected] & [email protected] & -0.30 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 220110 + 002547 & 330.293140 & 0.429823 & [email protected] & [email protected] & [email protected] & [email protected] & 0.28 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 220221 + 001114 & 330.587782 & 0.187335 & [email protected] & [email protected] & [email protected] & [email protected] & 0.29 & [email protected] & @xmath186 & 24.2 & @xmath186 & @xmath186 & @xmath186 & agn + 220445 + 005129 & 331.187680 & 0.857985 & [email protected] & [email protected] & [email protected] & [email protected] & -0.26 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 220708 + 010126 & 331.784654 & 1.023799 & [email protected] & [email protected] & [email protected] & [email protected] & -0.28 & [email protected] & [email protected] & 18.8 & nv & 2.91 & 33.2 & qso + 220804 + 000556 & 332.018501 & 0.098983 & [email protected] & [email protected] & [email protected] & [email protected] & -0.26 & [email protected] & [email protected] & 23.2 & @xmath186 & @xmath186 & @xmath186 & agn + 220904 + 004607 & 332.266020 & 0.768647 & [email protected] & [email protected] & [email protected] & [email protected] & 0.52 & @xmath200.51 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 221160 - 003139 & 332.998688 & -0.527367 & [email protected] & [email protected] & [email protected] & [email protected] & -0.27 & [email protected] & [email protected] & 23.9 & @xmath186 & @xmath186 & @xmath186 & agn + 221257 - 005711 & 333.235700 & -0.953020 & [email protected] & [email protected] & [email protected] & [email protected] & -0.50 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 221548 - 001031 & 333.948344 & -0.175409 & [email protected] & [email protected] & [email protected] & [email protected] & -0.52 & [email protected] & [email protected] & 22.6 & @xmath186 & 0.54@xmath187 & 30.9 & agn + 221609 - 005708 & 334.037121 & -0.952234 & [email protected] & [email protected] & [email protected] & [email protected] & -0.46 & [email protected] & [email protected] & 17.5 & nv & 2.40 & 33.2 & qso + 221642 - 004904 & 334.174200 & -0.817899 & [email protected] & [email protected] & [email protected] & [email protected] & -0.26 & [email protected] & [email protected] & 22.8 & @xmath186 & 0.69@xmath187 & 31.4 & agn + 222038 - 001209 & 335.158511 & -0.202623 & [email protected] & [email protected] & [email protected] & [email protected] & -0.47 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 222109 - 001940 & 335.286275 & -0.327789 & [email protected] & [email protected] & [email protected] & [email protected] & -0.53 & [email protected] & [email protected] & 22.4 & @xmath186 & 2.36 & 32.2 & qso + 222127 - 001530 & 335.362235 & -0.258239 & [email protected] & [email protected] & [email protected] & [email protected] & -0.30 & [email protected] & [email protected] & 23.2 & @xmath186 & 0.46@xmath187 & 31.0 & agn + 222201 - 005008 & 335.505891 & -0.835608 & [email protected] & [email protected] & [email protected] & [email protected] & -0.51 & [email protected] & [email protected] & 20.1 & nv & 0.33@xmath187 & 31.4 & agn + 222359 + 011148 & 335.997306 & 1.196745 & [email protected] & [email protected] & [email protected] & [email protected] & -0.38 & [email protected] & [email protected] & 22.4 & @xmath186 & @xmath186 & @xmath186 & agn + 222443 + 002858 & 336.180822 & 0.482741 & [email protected] & [email protected] & [email protected] & [email protected] & -0.29 & [email protected] & [email protected] & 19.5 & nv & 0.26@xmath187 & 30.5 & agn + 222524 - 001837 & 336.348824 & -0.310389 & [email protected] & [email protected] & [email protected] & [email protected] & 0.40 & @xmath200.55 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 222546 + 004038 & 336.439976 & 0.677156 & [email protected] & [email protected] & [email protected] & [email protected] & 0.34 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 222605 - 010441 & 336.521610 & -1.078109 & [email protected] & [email protected] & [email protected] & [email protected] & 0.29 & [email protected] & [email protected] & 22.9 & @xmath186 & 0.42@xmath187 & 31.3 & agn + 222630 - 001248 & 336.625294 & -0.213342 & [email protected] & [email protected] & [email protected] & [email protected] & -0.28 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 222647 + 005211 & 336.694041 & 0.869751 & [email protected] & [email protected] & [email protected] & [email protected] & -0.29 & [email protected] & [email protected] & 21.3 & nv & 2.26 & 34.7 & qso + 222704 + 011055 & 336.766725 & 1.182021 & [email protected] & [email protected] & [email protected] & [email protected] & -0.42 & [email protected] & @xmath203.06 & 17.4 & @xmath186 & 0.06@xmath187 & 29.1 & agn + 222907 + 001908 & 337.277758 & 0.318936 & [email protected] & [email protected] & [email protected] & [email protected] & -0.43 & [email protected] & [email protected] & @xmath186 & nv & 1.80 & 32.0 & qso + 222930 - 000845 & 337.373252 & -0.145791 & [email protected] & [email protected] & [email protected] & [email protected] & -0.27 & [email protected] & [email protected] & 22.9 & @xmath186 & 0.56@xmath187 & 31.5 & agn + 223024 + 004334 & 337.598552 & 0.726077 & [email protected] & [email protected] & [email protected] & [email protected] & -0.29 & [email protected] & [email protected] & 18.3 & nv & 0.13 & 29.7 & agn + 223047 + 003160 & 337.694749 & 0.533246 & [email protected] & [email protected] & [email protected] & [email protected] & -0.41 & [email protected] & [email protected] & 17.4 & nv & 0.09 & 29.7 & agn + 223140 + 002305 & 337.917233 & 0.384585 & [email protected] & [email protected] & [email protected] & [email protected] & -0.55 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 223225 + 003431 & 338.103745 & 0.575289 & [email protected] & [email protected] & [email protected] & [email protected] & 0.31 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 223409 + 010618 & 338.537649 & 1.105054 & [email protected] & [email protected] & [email protected] & [email protected] & -0.28 & [email protected] & [email protected] & 22.0 & @xmath186 & @xmath186 & @xmath186 & agn + 223516 - 005849 & 338.817742 & -0.980367 & [email protected] & [email protected] & [email protected] & [email protected] & -0.30 & [email protected] & [email protected] & 19.1 & nv & 1.18 & 32.2 & qso + 223908 + 003232 & 339.784225 & 0.542353 & [email protected] & [email protected] & [email protected] & [email protected] & -0.35 & [email protected] & [email protected] & 19.3 & nv & 0.24@xmath187 & 30.9 & agn + 224023 - 003555 & 340.096938 & -0.598702 & [email protected] & [email protected] & [email protected] & [email protected] & -0.41 & [email protected] & [email protected] & 18.9 & nv & 1.16 & 32.0 & qso + 224628 - 001214 & 341.615430 & -0.203797 & [email protected] & [email protected] & [email protected] & [email protected] & -0.26 & [email protected] & [email protected] & 21.8 & nv & 0.55@xmath187 & 32.7 & agn + 224730 + 000006 & 341.875813 & 0.001783 & [email protected] & [email protected] & [email protected] & [email protected] & -0.67 & [email protected] & [email protected] & 18.1 & v & 0.97 & 34.1 & qso + 224733 + 010817 & 341.885647 & 1.138013 & [email protected] & [email protected] & [email protected] & [email protected] & 0.34 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 224803 + 003959 & 342.013923 & 0.666301 & [email protected] & [email protected] & [email protected] & [email protected] & -0.33 & [email protected] & [email protected] & 22.3 & @xmath186 & 0.82@xmath187 & 32.2 & agn + 225103 + 000156 & 342.760572 & 0.032354 & [email protected] & [email protected] & [email protected] & [email protected] & -0.35 & [email protected] & [email protected] & 23.3 & @xmath186 & @xmath186 & @xmath186 & agn + 225307 - 010950 & 343.277710 & -1.163845 & [email protected] & [email protected] & [email protected] & [email protected] & 0.37 & [email protected] & [email protected] & 20.5 & nv & 0.34 & 30.7 & agn + 225438 - 001641 & 343.658007 & -0.277990 & [email protected] & [email protected] & [email protected] & [email protected] & 0.31 & [email protected] & [email protected] & 22.0 & @xmath186 & 0.38@xmath187 & 31.1 & agn + 225510 + 002526 & 343.791849 & 0.423809 & [email protected] & [email protected] & [email protected] & [email protected] & 0.62 & @xmath200.53 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 225525 - 000956 & 343.854026 & -0.165458 & [email protected] & [email protected] & [email protected] & [email protected] & -0.31 & [email protected] & [email protected] & 22.6 & @xmath186 & 0.71@xmath187 & 31.8 & agn + 225621 - 004110 & 344.085666 & -0.686069 & [email protected] & [email protected] & [email protected] & [email protected] & 0.41 & [email protected] & [email protected] & 21.5 & nv & 0.56 & 30.9 & agn + 225934 + 010821 & 344.891408 & 1.139151 & [email protected] & [email protected] & [email protected] & [email protected] & 0.45 & @xmath200.69 & @xmath200.34 & 23.7 & @xmath186 & @xmath186 & @xmath186 & agn + 225936 - 003356 & 344.898924 & -0.565593 & [email protected] & [email protected] & [email protected] & [email protected] & -0.30 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 230112 - 002112 & 345.299432 & -0.353462 & [email protected] & [email protected] & [email protected] & [email protected] & -0.33 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 230132 - 010319 & 345.382529 & -1.055362 & [email protected] & [email protected] & [email protected] & [email protected] & -0.44 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 230158 + 000352 & 345.490838 & 0.064493 & [email protected] & [email protected] & [email protected] & [email protected] & -0.68 & [email protected] & [email protected] & 23.1 & @xmath186 & @xmath186 & @xmath186 & agn + 230218 - 005817 & 345.576721 & -0.971495 & [email protected] & [email protected] & [email protected] & [email protected] & -0.27 & [email protected] & [email protected] & @xmath186 & @xmath186 & 0.91@xmath187 & 31.8 & agn + 230334 - 004006 & 345.890039 & -0.668195 & [email protected] & [email protected] & [email protected] & [email protected] & -0.27 & [email protected] & [email protected] & 22.0 & @xmath186 & 0.70@xmath187 & 31.7 & agn + 230423 - 000417 & 346.096113 & -0.071504 & [email protected] & [email protected] & [email protected] & [email protected] & 0.27 & [email protected] & [email protected] & 19.7 & nv & 1.05 & 32.0 & qso + 230748 + 002213 & 346.950170 & 0.370314 & [email protected] & [email protected] & [email protected] & [email protected] & 0.45 & [email protected] & [email protected] & 22.4 & @xmath186 & 0.71@xmath187 & 31.8 & agn + 230847 + 010904 & 347.196936 & 1.151171 & [email protected] & [email protected] & [email protected] & [email protected] & 0.40 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 231014 + 002531 & 347.559929 & 0.425187 & [email protected] & [email protected] & [email protected] & [email protected] & 0.32 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 231210 - 003135 & 348.042870 & -0.526485 & [email protected] & [email protected] & [email protected] & [email protected] & 0.50 & [email protected] & [email protected] & 23.3 & @xmath186 & @xmath186 & @xmath186 & agn + 231517 + 002630 & 348.822686 & 0.441566 & [email protected] & [email protected] & [email protected] & [email protected] & 0.26 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 231537 - 003726 & 348.902796 & -0.623911 & [email protected] & [email protected] & [email protected] & [email protected] & -0.81 & [email protected] & [email protected] & @xmath186 & nv & @xmath186 & @xmath186 & agn + 231542 + 002937 & 348.923546 & 0.493578 & [email protected] & [email protected] & [email protected] & [email protected] & 0.39 & [email protected] & [email protected] & 21.0 & nv & 1.35 & 32.8 & qso + 231557 + 005001 & 348.986515 & 0.833474 & [email protected] & [email protected] & [email protected] & [email protected] & 0.35 & @xmath200.57 & [email protected] & 19.7 & @xmath186 & 2.52 & 32.5 & qso + 231713 + 000256 & 349.305229 & 0.048771 & [email protected] & [email protected] & [email protected] & [email protected] & -0.28 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 231846 - 000755 & 349.690889 & -0.131884 & [email protected] & [email protected] & [email protected] & [email protected] & -0.33 & [email protected] & [email protected] & 19.5 & nv & 0.86 & 31.9 & qso + 232025 + 002744 & 350.106120 & 0.462184 & [email protected] & [email protected] & [email protected] & [email protected] & -0.31 & [email protected] & [email protected] & @xmath186 & @xmath186 & 2.89 & 33.8 & qso + 232037 + 001335 & 350.153115 & 0.226407 & [email protected] & [email protected] & [email protected] & [email protected] & 0.55 & @xmath200.51 & @xmath200.39 & @xmath186 & nv & 0.42@xmath187 & 30.5 & agn + 232236 - 000712 & 350.650440 & -0.119883 & [email protected] & [email protected] & [email protected] & [email protected] & -0.29 & [email protected] & @xmath186 & 25.3 & @xmath186 & @xmath186 & @xmath186 & agn + 232311 - 003122 & 350.794080 & -0.522647 & [email protected] & [email protected] & [email protected] & [email protected] & 0.32 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 232324 + 003328 & 350.849824 & 0.557710 & [email protected] & [email protected] & [email protected] & [email protected] & 0.37 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 232656 + 000303 & 351.732481 & 0.050827 & [email protected] & [email protected] & [email protected] & [email protected] & -0.39 & [email protected] & @xmath186 & 23.6 & @xmath186 & @xmath186 & @xmath186 & agn + 232804 + 001904 & 352.017137 & 0.317754 & [email protected] & [email protected] & [email protected] & [email protected] & 0.47 & [email protected] & @xmath186 & 21.6 & @xmath186 & @xmath186 & @xmath186 & agn + 233210 - 003101 & 353.043262 & -0.516945 & [email protected] & [email protected] & [email protected] & [email protected] & 0.35 & [email protected] & @xmath186 & 20.3 & nv & 0.45 & 30.8 & agn + 233227 - 010443 & 353.111298 & -1.078566 & [email protected] & [email protected] & [email protected] & [email protected] & 0.36 & [email protected] & @xmath186 & @xmath186 & nv & 0.26 & 30.5 & agn + 233260 - 005129 & 353.249596 & -0.857978 & [email protected] & [email protected] & [email protected] & [email protected] & -0.41 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 233301 - 004501 & 353.255643 & -0.750179 & [email protected] & [email protected] & [email protected] & [email protected] & -0.57 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + + 221650 + 005429 & 334.210170 & 0.908094 & [email protected] & [email protected] & [email protected] & [email protected] & -0.09 & @xmath200.57 & @xmath200.31 & 21.4 & nv & 0.55@xmath187 & 31.1 & agn + 221711 + 011038 & 334.294275 & 1.177200 & [email protected] & [email protected] & [email protected] & [email protected] & -0.20 & [email protected] & [email protected] & 22.2 & @xmath186 & 0.49@xmath187 & 32.0 & agn + 221813 - 010344 & 334.554017 & -1.062315 & [email protected] & [email protected] & [email protected] & [email protected] & 0.02 & @xmath200.72 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 223041 - 001644 & 337.672656 & -0.278969 & [email protected] & [email protected] & [email protected] & [email protected] & 0.08 & @xmath200.53 & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 223514 - 001425 & 338.806545 & -0.240294 & [email protected] & [email protected] & [email protected] & [email protected] & -0.01 & [email protected] & [email protected] & @xmath186 & @xmath186 & 0.14 & 30.3 & agn + 230113 - 002941 & 345.306025 & -0.494588 & [email protected] & [email protected] & [email protected] & [email protected] & -0.12 & [email protected] & [email protected] & @xmath186 & @xmath186 & @xmath186 & @xmath186 & agn + 233002 - 002736 & 352.507328 & -0.460065 & [email protected] & [email protected] & [email protected] & [email protected] & -0.03 & @xmath200.52 & @xmath186 & 21.4 & @xmath186 & 1.65 & 32.5 & qso + + 221515 - 005028 & 333.811588 & -0.841078 & [email protected] & [email protected] & [email protected] & [email protected] & 0.11 & @xmath200.68 & [email protected] & 21.4 & nv & 0.44@xmath187 & 31.0 & agn + 223634 - 003352 & 339.141345 & -0.564383 & [email protected] & [email protected] & [email protected] & [email protected] & 0.78 & @xmath200.66 & @xmath200.63 & 8.9 & @xmath186 & @xmath186 & @xmath186 & rs cvn + 230241 + 003450 & 345.672648 & 0.580639 & [email protected] & [email protected] & [email protected] & [email protected] & 1.61 & @xmath200.51 & @xmath200.28 & 10.9 & @xmath186 & @xmath186 & @xmath186 & dke + + lrrr obs .
date & freq . &
s & @xmath188 + & ( ghz ) & ( mjy ) & ( mjy ) + + 01 sep 2012 & 2.4 & 0.504 & 0.039 + 01 sep 2012 & 3.2 & 0.604 & 0.028 + 01 sep 2012 & 3.8 & 0.569 & 0.033 + 01 sep 2012 & 4.5 & 0.596 & 0.024 + 01 sep 2012 & 5.1 & 0.623 & 0.023 + 01 sep 2012 & 7.1 & 0.641 & 0.021 + 01 sep 2012 & 7.7 & 0.646 & 0.021 + 01 sep 2012 & 13.2 & 0.584 & 0.035 + 01 sep 2012 & 13.8 & 0.672 & 0.032 + 01 sep 2012 & 14.2 & 0.639 & 0.035 + 17 sep 2012 & 1.2 & 1.130 & 0.079 + 17 sep 2012 & 1.8 & 1.363 & 0.077 + 17 sep 2012 & 2.4 & 1.109 & 0.051 + 17 sep 2012 & 3.1 & 0.841 & 0.040 + 17 sep 2012 & 3.8 & 0.671 & 0.068 + 17 sep 2012 & 4.5 & 0.596 & 0.049 + 17 sep 2012 & 5.1 & 0.623 & 0.047 + 17 sep 2012 & 7.1 & 0.641 & 0.044 + 17 sep 2012 & 7.7 & 0.646 & 0.045 + + 01 sep 2012 & 2.4 & 0.739 & 0.065 + 01 sep 2012 & 3.0 & 0.701 & 0.048 + 01 sep 2012 & 3.4 & 0.740 & 0.054 + 01 sep 2012 & 4.8 & 0.778 & 0.027 + 01 sep 2012 & 7.4 & 0.741 & 0.021 + 01 sep 2012 & 13.5 & 0.598 & 0.026 + 01 sep 2012 & 14.5 & 0.568 & 0.027 + + 01 sep 2012 & 2.4 & 6.846 & 0.095 + 01 sep 2012 & 3.2 & 9.294 & 0.071 + 01 sep 2012 & 3.8 & 9.641 & 0.108 + 01 sep 2012 & 4.5 & 10.519 & 0.073 + 01 sep 2012 & 5.1 & 10.558 & 0.071 + 01 sep 2012 & 7.1 & 9.612 & 0.070 + 01 sep 2012 & 7.7 & 9.381 & 0.073 + 01 sep 2012 & 13.2 & 7.269 & 0.102 + 01 sep 2012 & 13.8 & 6.836 & 0.097 + 01 sep 2012 & 14.2 & 6.487 & 0.103 + 01 sep 2012 & 14.8 & 6.217 & 0.098 + 17 sep 2012 & 1.2 & 1.829 & 0.189 + 17 sep 2012 & 1.8 & 3.791 & 0.179 + 17 sep 2012 & 2.4 & 7.437 & 0.082 + 17 sep 2012 & 3.1 & 9.327 & 0.067 + 17 sep 2012 & 3.8 & 10.380 & 0.123 + 29 may 2014 & 1.3 & 2.396 & 0.179 + 29 may 2014 & 1.9 & 4.253 & 0.240 + 29 may 2014 & 2.4 & 7.095 & 0.188 + 29 may 2014 & 3.0 & 7.511 & 0.109 + 29 may 2014 & 3.4 & 7.272 & 0.094 + 29 may 2014 & 8.4 & 4.984 & 0.051 + 29 may 2014 & 9.4 & 4.562 & 0.053 + 29 may 2014 & 10.4 & 3.955 & 0.055 + 29 may 2014 & 11.4 & 3.474 & 0.081 + 29 may 2014 & 13.2 & 3.295 & 0.074 + 29 may 2014 & 13.8 & 3.505 & 0.070 + 29 may 2014 & 15.7 & 2.937 & 0.076 + 29 may 2014 & 16.3 & 2.949 & 0.076 + + 29 may 2014 & 1.3 & 2.759 & 0.137 + 29 may 2014 & 1.9 & 2.549 & 0.175 + 29 may 2014 & 2.4 & 2.654 & 0.080 + 29 may 2014 & 3.0 & 2.373 & 0.069 + 29 may 2014 & 3.4 & 1.935 & 0.067 + 29 may 2014 & 4.5 & 2.155 & 0.061 + 29 may 2014 & 5.1 & 2.058 & 0.055 + 29 may 2014 & 7.1 & 1.337 & 0.047 + 29 may 2014 & 7.7 & 1.198 & 0.048 + 29 may 2014 & 13.2 & 0.487 & 0.059 + 29 may 2014 & 13.8 & 0.710 & 0.054 + 29 may 2014 & 15.7 & 0.554 & 0.060 + 29 may 2014 & 16.3 & 0.464 & 0.059 + | we have commenced a multi - year program , the caltech - nrao stripe 82 survey ( cnss ) , to search for radio transients with the jansky vla in the sloan digital sky survey stripe 82 region .
the cnss will deliver five epochs over the entire @xmath0270 deg@xmath1 of stripe 82 , an eventual deep combined map with a rms noise of @xmath040 @xmath2jy and catalogs at a frequency of 3 ghz , and having a spatial resolution of 3 .
this first paper presents the results from an initial pilot survey of a 50 deg@xmath1 region of stripe 82 , involving four epochs spanning logarithmic timescales between one week and 1.5 years , with the combined map having a median rms noise of 35@xmath2jy .
this pilot survey enabled the development of the hardware and software for rapid data processing , as well as transient detection and follow - up , necessary for the full 270 deg@xmath1 survey .
data editing , calibration , imaging , source extraction , cataloging and transient identification were completed in a semi - automated fashion within six hours of completion of each epoch of observations , using dedicated computational hardware at the nrao in socorro , and custom - developed data reduction and transient detection pipelines .
classification of variable and transient sources relied heavily on the wealth of multi - wavelength legacy survey data in the stripe 82 region , supplemented by repeated mapping of the region by the palomar transient factory ( ptf ) .
@xmath3% of the few thousand detected point sources were found to vary by greater than 30% , consistent with similar studies at 1.4 ghz and 5 ghz .
multi - wavelength photometric data and light curves suggest that the variability is mostly due to shock - induced flaring in the jets of agn .
although this was only a pilot survey , we detected two bona fide transients , associated with an rs cvn binary and a dke star .
comparison with existing legacy survey data ( first , vla - stripe 82 ) revealed additional highly variable and transient sources on timescales between 520 years , largely associated with renewed agn activity .
the rates of such agn possibly imply episodes of enhanced accretion and jet activity occurring once every @xmath040,000 years in these galaxies .
we compile the revised radio transient rates and make recommendations for future transient surveys and joint radio - optical experiments . |
it is known that cluster structures appear in various nuclei including unstable nuclei ( for instance , @xcite and references therein . ) for cluster states having an @xmath1 cluster around a core nucleus , well - known examples are @xmath8o+@xmath1 states in @xmath9ne and @xmath10c+@xmath1 states in @xmath8o @xcite .
recent experimental and theoretical studies have revealed many cluster resonances in highly excited states near the @xmath1-decay threshold also in unstable nuclei , for instance , @xmath11he+@xmath1 states in be isotopes @xcite , @xmath0be+@xmath1 states in @xmath2c @xcite , @xmath2c+@xmath1 states in @xmath12o and their mirror states @xcite , @xmath12o+@xmath1 states in @xmath13ne @xcite .
multi-@xmath1 cluster states such as cluster gas and linear - chain states of @xmath14 systems are also interesting topics .
the @xmath1-cluster gas was proposed by tohsaki _
_ to describe @xmath15 cluster structure of @xmath10c(@xmath16 ) @xcite and extended to excited states of @xmath10c and other nuclei @xcite .
the linear - chain @xmath14 state was originally proposed for @xmath10c(@xmath16 ) by morinaga in the 1950 - 60s @xcite . however , in the 1970s , this picture was excluded at least for @xmath10c(@xmath16 ) having a larger @xmath1-decay width than the one expected from the linear - chain structure @xcite . in spite of many discussions for several decades , existence of linear - chain @xmath14 states
have yet been confirmed and it is still an open problem to be solved .
it is naively expected that the linear - chain configuration is not favored in an @xmath14 system because it costs much kinetic energy to keep @xmath1 clusters in a row .
it means that some mechanism is necessary to form the linear - chain structure . in progress of physics of unstable nuclei since the 1990s
, it was proposed for neutron - rich c isotopes that excess neutrons may stabilize the linear - chain structure @xcite .
analyzed stability of a @xmath15-chain configuration surrounded by excess neutrons in molecular orbitals against the bending motion and suggested that the linear - chain structure can be stable in @xmath8c but unstable in @xmath10c and @xmath2c @xcite .
more recently , suhara and the author predicted a rotational band with a linear @xmath15 chain configuration in excited states of @xmath2c near the @xmath1-decay threshold @xcite .
they pointed out that the orthogonal condition to lower states is important for the stability of the linear - chain structure .
the linear - chain structure is expected to be more favored in high spin states because of stretching effect in rotating systems as suggested in @xmath17c @xcite and @xmath8o @xcite . according to analysis in refs .
@xcite , linear - chain states of @xmath2c are found to have a @xmath18 correlation and are interpreted as @xmath0be+@xmath1 structures , where the @xmath0be cluster is a prolately deformed state containing a 2@xmath1 core and an additional @xmath1 cluster is located in the longitudinal direction of the @xmath0b cluster .
similarly , the linear - chain state of @xmath17c suggested in ref .
@xcite also shows a @xmath19be+@xmath1 cluster structure with a prolately deformed @xmath19be cluster and an @xmath1 cluster in the longitudinal direction .
it means that , the linear - chain states in these neutron - rich c tend to have the @xmath20 correlation , and therefore @xmath15 linear - chain structures are expected to be found in be+@xmath1 cluster states . in this paper ,
i focus on @xmath0b+@xmath1 cluster states in excited states of @xmath2n . in experimental energy levels of @xmath2n near the @xmath1-decay threshold , @xmath21 and @xmath6 resonances
were observed by @xmath1 elastic scattering by @xmath0b @xcite .
these resonances are expected to be @xmath0b+@xmath1 cluster states because of significant @xmath1-decay widths . in analogy to @xmath0be+@xmath1 cluster states , it is interesting to investigate whether @xmath0b+@xmath1 cluster states with the dominant linear - chain structure exist .
the ground state ( @xmath5 ) and the first excited state ( @xmath6 ) of @xmath0b can be described by the deformed state with a 2@xmath1 core surrounded by @xmath22 as discussed in refs .
if a @xmath0b+@xmath1 cluster state has an @xmath1 cluster in the longitudinal direction of the deformed @xmath0b cluster , the @xmath0b+@xmath1 cluster state can be interpreted as a kind of linear - chain state that contains dominantly @xmath15 clusters arranged in a row .
my aim is to study @xmath0b+@xmath1 cluster states of @xmath2n and discuss @xmath15 configurations , in particular , the linear - chain component in @xmath0b+@xmath1 cluster states .
i calculate @xmath23 and @xmath24 components and evaluate partial @xmath1-decay widths of @xmath0b+@xmath1 cluster states . to discuss stability of the linear - chain @xmath0b+@xmath1 structure ,
i analyze angular motion of an @xmath1 cluster around the deformed @xmath0b cluster , i.e. , rotation of the @xmath0b cluster .
this paper is organized as follows . in sec .
[ sec : formulation ] , i explain the formulation of the present @xmath0b+@xmath1 cluster model . in sec . [
sec : results ] , calculated positive - parity states and @xmath25 transition strengths of @xmath2n are shown .
i discuss @xmath1 cluster motion around @xmath0b(@xmath5 ) and @xmath0b(@xmath6 ) in sec .
[ sec : discussion ] .
finally , a summary is given in sec .
[ sec : summary ] .
for the @xmath0b cluster in the present @xmath0b+@xmath1 cluster model , i adopt a @xmath26 wave function which can reasonably describe features of the ground ( @xmath21 ) and first excited ( @xmath6 ) states of @xmath0b as discussed in ref .
the @xmath26 wave function is given by a three - body cluster wave function , where @xmath1 clusters and a dinucleon @xmath27 cluster are written by @xmath28 and @xmath29 harmonic oscillator configurations , respectively , as @xmath30 where @xmath31 is the antisymmetrizer for all nucleons and @xmath32 is the spatial part of the single - particle wave function of the @xmath33 orbit around @xmath34 ; @xmath35 and @xmath36 is the spin - isospin wave function for @xmath37 , @xmath38 , @xmath39 , and @xmath40 . for the @xmath0b cluster
, i set 2 @xmath1 clusters in the @xmath41 direction as @xmath42 with @xmath43 fm and a spin - aligned @xmath22 cluster on the @xmath44-@xmath45 plane at the distance @xmath46 from the @xmath20 center as @xmath47 .
i write the @xmath0b wave function localized around @xmath48 as @xmath49 with the center position @xmath50 and the distance and angle parameters , @xmath46 and @xmath51 , for the @xmath22 cluster position . in the @xmath0b+@xmath1 cluster model , i superpose the @xmath0b wave functions with @xmath52 ( fm ) and @xmath53 ( @xmath54 ) .
parity ( @xmath55 ) and @xmath56 ( @xmath57 ) projections of the subsystem @xmath0b can be approximately done by the @xmath58 summation ; @xmath59 with @xmath60 and @xmath61 . here , @xmath57 is the @xmath41 component of the total angular momentum @xmath62 of @xmath0b and is given by a sum of the aligned intrinsic spin @xmath63 and the orbital @xmath51 rotation of the @xmath22 cluster .
it is clear that @xmath58 superposition with given coefficients @xmath64 is equivalent to @xmath57 mixing .
a @xmath65 wave function is written using the @xmath0b wave function @xmath66 and the @xmath1-cluster wave function @xmath67 as @xmath68 where @xmath69 . here
the distance @xmath70 and the angle @xmath71 indicate the @xmath1-cluster position relative to the deformed @xmath72 cluster ( see fig .
[ fig:10b-4he ] ) .
the center of mass position is taken to be @xmath73 so as to decouple the center of mass motion and the intrinsic wave function .
wave functions for the @xmath74 states of @xmath2n are expressed by superposition of the @xmath75-projected wave functions as @xmath76 where @xmath77 is the parity and total angular momentum projection operator .
coefficients @xmath78 are determined by diagonalizing hamiltonian and norm matrices .
i take @xmath79 ( fm ) , @xmath80 , @xmath81 ( fm ) , and @xmath82 ( @xmath54 ) . in the present paper ,
i calculate positive - parity ( @xmath83 ) states of @xmath2n . in eq .
, coupling of @xmath62 ( the spin of the @xmath0b cluster ) and @xmath84 ( the orbital angular momentum of the @xmath1 cluster relative to the @xmath0b cluster ) is implicitly described by the @xmath75 projection , @xmath56 mixing , and @xmath71 , @xmath51 summations . as shown in fig .
[ fig:10b-4he ] , @xmath84 couples with @xmath62 to the total angular momentum @xmath85 .
the @xmath41 component , @xmath86 , is the so - called @xmath56 quantum . strictly speaking , @xmath87 ( @xmath88-wave ) mixing
is approximately taken into account by the summation of @xmath80 but higher @xmath89 mixing can not be controlled in the present calculation because of the finite number of mesh points for @xmath71 .
schematic figures for @xmath0b+@xmath1 configurations .
( a ) parameters of the @xmath0b+@xmath1 cluster model .
( b ) longitudinal configuration for @xmath90 , ( c ) transverse configuration for @xmath91 without @xmath92 alignment , and ( d ) @xmath92-aligned transverse configuration . , width=264 ] in order to evaluate @xmath0b@xmath93 and @xmath0b@xmath94 components at a channel radius ( @xmath70 ) , i consider the @xmath95 projected @xmath96 wave function , @xmath97 with @xmath98 , @xmath83 , @xmath99 , @xmath69 , and @xmath73 .
@xmath100 is the spherical harmonics .
@xmath101 is the wave function for the @xmath1 cluster at @xmath102 around the @xmath57 projected @xmath0b cluster , for which i fix @xmath103 fm in the present analysis .
@xmath104 is determined from the normalization condition @xmath105 . in eq .
, the @xmath95 projection is approximately performed by the summation @xmath106 @xmath107 with the weight function @xmath108}_{\min[\theta_\alpha-\pi/2n_\theta,0]}\sin{\theta}d\theta$ ] .
i perform only @xmath109 projections because @xmath110 projections are not possible for the present @xmath111 case .
i calculate the squared overlap of the @xmath2n wave function with the above wave function , @xmath112 . assuming that the @xmath113 and @xmath114 states of the @xmath0b cluster are approximately described by the @xmath57 projected @xmath0b wave functions , @xmath0b(@xmath115 ) and @xmath0b(@xmath116 ) , respectively
, i approximately estimate the @xmath117 components as @xmath118 with @xmath119 , where @xmath120 is the clebsch - gordan coefficient .
i also calculate @xmath1-cluster probability at @xmath102 around the @xmath57 projected @xmath0b cluster as @xmath121 the probability @xmath122 is useful to discuss geometric configurations of @xmath15 clusters in @xmath0b+@xmath1 cluster states in the strong coupling picture . for instance , @xmath122 for @xmath90 means the component of the `` longitudinal '' configuration , where the @xmath1 cluster is localized in the longitudinal direction of the deformed @xmath123 cluster .
this configuration corresponds to the linear - chain structure as 3 @xmath1 clusters are arranged in a row as shown in fig .
[ fig:10b-4he](b ) .
because of the axial symmetry , the longitudinal configuration contains only @xmath124 @xmath125 component . for @xmath126
, @xmath122 indicates the component of the `` transverse configuration '' for the @xmath1 cluster in the transverse direction of the deformed @xmath123 cluster .
the transverse configuration contains @xmath127 components corresponding to the alignment of @xmath84 to the spin of the @xmath22-cluster ( @xmath57 ) in the @xmath72 cluster as well as the @xmath124 component ( see fig .
[ fig:10b-4he](b ) and ( c ) ) .
positive - parity energy levels of @xmath2n obtained by the @xmath0b+@xmath1 cluster model compared with experimental levels taken from @xcite .
@xmath0b+@xmath1 cluster states in the @xmath3 band and those in the @xmath4 band are labeled by asterisk and down - triangle symbols , respectively . ]
transition strengths calculated by the @xmath0b+@xmath1 cluster model for ( a ) @xmath128 and ( b ) @xmath129 transitions with @xmath130 . asterisk and
down - triangle symbols show @xmath0b+@xmath1 cluster states in the @xmath3 and @xmath4 bands , respectively .
[ fig : be2],width=264 ] i adopt the two - body effective nuclear interactions used in ref .
@xcite which are adjusted to describe low - lying energy levels of @xmath0b .
namely , i use the volkov central force @xcite with the bartlett , heisenberg , and majorana parameters @xmath131 and @xmath132 and the g3rs spin - orbit force @xcite with the strength @xmath133 mev , and the coulomb force approximated by seven gaussians . using these interactions , energies of @xmath0b
are obtained to be @xmath13454.3 mev for the ground state ( @xmath5 ) and @xmath135 mev for the first excite state ( @xmath6 ) with the @xmath136 cluster model by superposing @xmath137 with @xmath52 ( fm ) . though the calculation underestimates the experimental binding energy ( 64.75 mev ) , it reproduces the spin parity of the ground state ( @xmath0b@xmath138 ) , and also the calculated excitation energy @xmath139 mev of the @xmath6 state reasonably agrees to the experimental value @xmath140 mev for @xmath0b@xmath141 . using the @xmath0b+@xmath1 cluster wave function in eq .
, i calculate positive - parity states of @xmath2n .
properties of the ground state @xmath2n@xmath142 are reasonably reproduced by the present calculation .
namely , the binding energy b.e.=102.6 mev , the magnetic moment @xmath143 ( @xmath144 ) , and the electric quadrupole moment @xmath145 ( @xmath146 ) of @xmath2n(@xmath147 ) , obtained by the present calculation reasonably agree to the experimental data ( b.e.=104.66 mev , @xmath148 ( @xmath144 ) , and @xmath149 ( @xmath146 ) ) .
the calculated energy spectra are shown in fig .
[ fig : spe ] .
the @xmath1-decay threshold is much higher in the present calculation than the experimental threshold .
in other words , the ground and some low - lying states of @xmath2n show too deep binding from the @xmath1-decay threshold compared with the experimental data .
the significant overestimation of the @xmath1-decay threshold is a general problem in microscopic calculations with density - independent two - body effective interactions as found for @xmath2c and o isotopes @xcite .
one of the origins of this problem is a difficulty in reproducing systematics of binding energies in a wide mass - number region with such effective interactions . in this paper
, i mainly investigate @xmath65 cluster states near the @xmath1-decay threshold and discuss their features . in the calculated energy levels near the threshold ,
i obtain several excited states having significant component of a spatially developed @xmath1 cluster around the @xmath72 cluster . from remarkable @xmath25 transitions ,
i assign the @xmath65 cluster states to a @xmath3 band of @xmath21 , @xmath150 , and @xmath151 states , and a @xmath4 band of @xmath152 , @xmath153 , @xmath5 , @xmath150 , and @xmath151 states .
the former and the latter bands are shown by asterisk and down - triangle symbols in fig . [
fig : spe ] .
the @xmath3 band has the significant @xmath0b(@xmath5)+@xmath1 component , whereas the @xmath4 band contains the @xmath0b(@xmath6)+@xmath1 component .
more details of structure of these states are discussed in the next section .
figure [ fig : be2 ] shows @xmath25 transitions with @xmath154 @xmath155 for @xmath156 and @xmath157 transitions . in - band transitions for the @xmath3 and @xmath158 @xmath65 bands are rather strong because of the developed cluster structures , though @xmath25 strengths are somewhat fragmented into neighboring states .
@xmath0b+@xmath1 cluster states in the @xmath3 and @xmath4 bands have maximum amplitudes of @xmath1-cluster probability around @xmath159 fm . in this section ,
i focus on angular motion of the @xmath1 cluster at @xmath159 fm .
i first investigate the angular momentum coupling of the @xmath1-cluster ( @xmath92 ) and the @xmath0b cluster ( @xmath160 ) in a weak coupling picture and estimate @xmath1-decay widths .
then , i discuss geometric configurations of @xmath0b+@xmath1 cluster states in the strong coupling picture by analyzing @xmath71-dependence of the @xmath1-cluster probability around the deformed @xmath0b cluster . in the present calculation , radial motion of the @xmath1 cluster
is described by superposing @xmath0b+@xmath1 wave functions for @xmath161 fm . instead of the full model space in eq . including @xmath161 fm wave functions , i also perform a similar calculation using the @xmath70-fixed model space @xmath162 where i fix @xmath159 fm and take @xmath163 , @xmath81 ( fm ) , and @xmath82 ( @xmath54 ) . in the @xmath159 fm fixed calculation
, i find the states near the threshold energy corresponding to @xmath0b+@xmath1 cluster states in the @xmath3 and @xmath4 bands , but do not obtain lower states below the threshold because of the truncation of the model space .
energy levels of the @xmath3 and @xmath4 bands obtained with the full and @xmath70-fixed calculations are shown in fig .
[ fig : spe - d5 ] .
the calculated energies are measured from the @xmath1-decay threshold .
the experimental levels observed by @xmath1 elastic scattering by @xmath0b are also shown in the figure .
the level structures of the @xmath3 and @xmath4 bands are essentially consistent between the full and @xmath70-fixed calculations , though about 2 mev global shift is found for the @xmath3 band between two calculations .
b+@xmath1 cluster states obtained by the full and @xmath70-fixed calculations and those observed by the experiment of @xmath164 reactions @xcite .
energies are measured from the @xmath1-decay threshold .
[ fig : spe - d5],width=226 ] in table [ tab : s - fac ] , i show @xmath92 components ( @xmath165 in eq . ) at @xmath166 coupled with @xmath167 and @xmath168 in @xmath0b+@xmath1 cluster states obtained by the full and @xmath70-fixed calculations . in the result of the @xmath70-fixed calculation ,
@xmath3 band states are dominated by the @xmath23 component , whereas @xmath4 band states contain dominantly the @xmath169 component . in the result of the full calculation ,
the @xmath3 and @xmath4 band states still contain significant @xmath23 and @xmath169 components , respectively , except for the @xmath170 state , though the absolute amplitude of the dominant component decreases because of radial motion and state mixing .
the @xmath170 state obtained by the full calculation shows a feature quite different from that obtained by the @xmath70-fixed calculation . in the @xmath70-fixed calculation ,
the @xmath170 state is approximately described by the pure @xmath171 state , where the orbital angular momentum @xmath172 of the @xmath1 cluster weakly couples to the spin @xmath173 of the @xmath0b cluster . however , in the full calculation , the @xmath170 state does not show the weak coupling feature but has @xmath171 , @xmath174 , and @xmath175 components with the same order showing a strong coupling feature .
figure [ fig : s - fac - lalpha ] shows @xmath92 components ( @xmath165 ) at @xmath159 fm of @xmath75 states in the @xmath2n spectra obtained by the full calculation .
the @xmath0b@xmath176+@xmath177 and @xmath0b@xmath176+@xmath178 components concentrate at the @xmath179 and @xmath180 states , respectively , though the components are fragmented into other states .
the @xmath181 state shows rather strong state mixing .
the @xmath0b@xmath182+@xmath178 component concentrates at the @xmath183 and @xmath184 states , whereas , the @xmath0b@xmath182+@xmath177 component feeds lower @xmath6 states of @xmath2n .
@xmath185 components , @xmath186 , in positive - parity states of @xmath2n obtained by the @xmath0b+@xmath1-cluster model .
asterisk and down - triangle symbols show @xmath0b+@xmath1 cluster states in the @xmath3 and @xmath4 bands , respectively . ,
width=396 ] in the experiment of @xmath164 reactions @xcite , the @xmath5 state at @xmath187 mev ( @xmath188 mev ) with the width @xmath189 mev is strongly populated . in the analysis of ref .
@xcite , this state is described well by the dominant ( almost 100% ) @xmath190-wave @xmath1-decay indicating the significant @xmath191 component of the @xmath5 state .
the @xmath6 state at @xmath192 mev ( @xmath193 mev ) is weakly populated in @xmath194 reactions , whereas its @xmath1-decay into the fisrt excited state of @xmath168 was observed in @xmath195 reactions @xcite .
these experiments suggest that the @xmath6 state would contain @xmath171 and @xmath175 components . from the experimental @xmath1-decay properties ,
i tentatively assign the theoretical @xmath179 and @xmath170 states having @xmath0b+@xmath1 cluster structures to the experimental @xmath5 ( @xmath196 mev ) and @xmath6 ( @xmath197 mev ) states , though the band - head energies @xmath198 mev and @xmath199 mev obtained by the full calculation do not necessarily agree to the experimental energies ( see fig .
[ fig : spe - d5 ] ) .
i estimate partial @xmath1-decay widths for @xmath200 channels from @xmath201 ( @xmath202 is the channel radius ) as follows . using the approximate evaluation of the reduced width amplitude proposed in ref .
@xcite , the reduced width @xmath203 is calculated as @xmath204 and the partial @xmath1-decay width @xmath205 for @xmath206 is calculated as @xmath207 where @xmath208 , and @xmath209 and @xmath210 are the regular and irregular coulomb functions , respectively .
here i use the momentum @xmath211 of the energy @xmath212 which is phenomenologically adjusted to the experimental energy position because it is difficult to quantitatively predict the energy position in the present calculation .
namely , i adjust the band - head energies of the @xmath3 and @xmath4 bands to the experimental energy positions @xmath213 mev and @xmath214 mev , by a constant shift for each band as @xmath215 calculated partial @xmath1-decay widths are shown in table [ tab : s - fac ] .
i calculate widths for @xmath216 and @xmath217 channels .
@xmath1-decay widths obtained by the full calculation are several times smaller than those obtained by the @xmath70-fixed calculation because of the suppression of the @xmath1-cluster probability as shown previously . as a result
, the @xmath1-decay width of the @xmath179 state reduces to be @xmath218 mev with the dominant @xmath219 decay , which is quantitatively consistent with the experimental observation ( @xmath220 mev ) @xcite . for the @xmath170 state ,
i obtain a small @xmath1-decay width @xmath221 mev with the dominant @xmath222 decay .
this result seems consistent with the weak population in the @xmath1 elastic scattering @xcite and the fact that the @xmath6 state was observed in @xmath195 reaction @xcite . however , experimental information of partial @xmath1-decay widths is not enough to confirm the present assignment of the @xmath170 state .
the calculated @xmath1-decay width is much smaller than the experimental total width , @xmath223 mev , of the @xmath6 state at @xmath224 mev .
i should comment that , because the @xmath222 component is fragmented into neighboring states as shown in fig .
[ fig : s - fac - lalpha ] , an effectively large width could be observed for the @xmath170 state .
ccccc & & + @xmath75 & @xmath216 & @xmath217 & @xmath216 & @xmath217 + + @xmath5(@xmath3 ) & 0.21 & 0.10 & & 0.04 + @xmath150(@xmath3 ) & & 0.23 & & + @xmath151(@xmath3 ) & & 0.14 & & + @xmath6(@xmath4 ) & & 0.03 & 0.05 & 0.09 + @xmath153(@xmath4 ) & & 0.02 & & 0.25 + @xmath5(@xmath4 ) & 0.00 & 0.02 & & 0.37 + @xmath150(@xmath4 ) & & 0.01 & & + @xmath151(@xmath4 ) & & 0.14 & & + + @xmath5(@xmath3 ) & 0.57 & 0.25 & & 0.01 + @xmath150(@xmath3 ) & & 0.73 & & + @xmath151(@xmath3 ) & & 0.75 & & + @xmath6(@xmath4 ) & & 0.02 & 0.89 & 0.05 + @xmath153(@xmath4 ) & & 0.01 & & 0.78 + @xmath5(@xmath4 ) & 0.10 & 0.13 & & 0.74 + @xmath150(@xmath4 ) & & 0.00 & & + . [ tab : gamma ] partial @xmath1-decay widths of @xmath0b+@xmath1 cluster states in the @xmath3 and @xmath4 bands obtained by the full and @xmath70-fixed calculations .
energies of the band - head states of the @xmath3 and @xmath4 bands are adjusted to the experimental resonance energies of the @xmath5 state at 1.58 mev and the @xmath6 state at 2.11 mev .
the sum ( @xmath225 ) of partial widths of decay channels @xmath226 and @xmath227 is also shown .
the unit is mev . [
cols="^,^,^,^,^,^,^ " , ] i here discuss angular motion of the @xmath1-cluster around the deformed @xmath0b cluster by analyzing @xmath71 dependence of @xmath1-cluster probabilities .
discussions in this section are based on the strong coupling picture , which is somehow different from the previous discussion based on the @xmath92 decomposition in the weak coupling picture .
i show energies of @xmath101 , in which the @xmath1 cluster is localized at @xmath102 around the @xmath57 projected @xmath0b cluster . in fig .
[ fig : s - fac - lalpha ] , intrinsic energies before parity and angular - momentum projections of @xmath101 for @xmath115 and @xmath6 are plotted on the @xmath228 plane .
the energy curves for @xmath159 fm are also shown as functions of @xmath71 . in the @xmath229 fm region ,
the contour of the energy surface on the @xmath230 plane is deformed in the longitudinal ( @xmath231 ) direction because of the prolate deformation of the @xmath0b cluster meaning that the @xmath1 cluster at the fixed distance @xmath159 fm feels an attraction in the longitudinal direction .
in other words , in the intrinsic system , the @xmath1 cluster at @xmath159 fm energetically favors the longitudinal direction to form the linear @xmath15 configuration rather than the transverse direction to form the triangle @xmath15 configuration . in the @xmath232 fm region
, the @xmath1 cluster feels an effective repulsion in the longitudinal direction because of the pauli blocking from the @xmath0b cluster , whereas it feels an attraction in the transverse ( @xmath233 ) direction .
in contrast to the intrinsic energy behavior , @xmath71 dependence of @xmath75-projected energy is not trivial because the energy is affected by not only potential energy but also kinetic energy of angular motion , i.e. , rotational energy .
figure [ fig : ene - theta ] shows energies of @xmath234-projected states @xmath235 at @xmath159 fm for @xmath124 , which corresponds to the @xmath236 projection . in high @xmath237 states , the longitudinal direction ( @xmath238 ) is energetically favored than the transverse direction ( @xmath239 ) because of the larger moment of inertia ( m.o.i . ) of the longitudinal configuration than that of the transverse configuration for the @xmath236 projection .
however , in the lowest spin state ( @xmath240 ) , the energy almost degenerates in a wide region of @xmath71 because the kinetic energy is smaller for the transverse configuration than the longitudinal configuration because of the phase space factor @xmath241 in the @xmath236 projection .
this energy degeneracy results in the @xmath216 ( @xmath190-wave ) dominance in the @xmath170 state obtained by the @xmath70-fixed calculation
. figures [ fig : ene - theta - j3 ] and [ fig : ene - theta - j1 ] show energies of @xmath234-projected states at @xmath159 fm for @xmath127 .
note that the @xmath127 projection corresponds to the @xmath242 projection , and @xmath243 means the @xmath244 alignment to the @xmath41 direction ( see fig .
[ fig:10b-4he](c ) ) .
for instance , the @xmath92-aligned state for @xmath217 ( @xmath245-wave ) is the @xmath246 state . as shown in figs .
[ fig : ene - theta - j3](a)-(c ) and [ fig : ene - theta - j1](a)-(d ) , @xmath92-aligned states energetically favor the transverse configuration because of the larger m.o.i . than that of the longitudinal configuration in the @xmath247 projection .
figures [ fig : ene - theta - j3 ] and [ fig : ene - theta - j1 ] also show the @xmath1-cluster probability @xmath122 at @xmath159 fm in the @xmath0b+@xmath1 cluster states obtained by the @xmath70-fixed and full calculations .
let me first discuss the result obtained by the @xmath70-fixed calculation ( figs .
[ fig : ene - theta - j3](d)-(f ) and [ fig : ene - theta - j1](e)-(h ) ) . in the @xmath3 band states ( fig .
[ fig : ene - theta - j3](d)-(f ) ) , the @xmath21 state contains dominantly the longitudinal configuration ( @xmath238 ) rather than the transverse configuration ( @xmath239 ) as expected from the @xmath234-projected energy curve for @xmath124 .
as @xmath237 goes up to @xmath248 , the @xmath92-aligned component @xmath249 of the transverse configuration becomes large corresponding to the alignment of the orbital angular momentum @xmath92 of the @xmath1 cluster to @xmath250 ( the spin of @xmath27 cluster in the @xmath0b cluster ) . in the @xmath4 band states ( fig .
[ fig : ene - theta - j1](e)-(h ) ) , the @xmath152 state shows the @xmath1-cluster probability distributed widely in the @xmath251 region indicating the dominant @xmath216 ( @xmath190-wave ) component . as @xmath237 increases , the longitudinal component becomes dominant compared with the transverse component .
the alignment of @xmath92 ( the orbital angular momentum of the @xmath1 cluster ) and @xmath57 is not so remarkable for @xmath0b@xmath252 differently from the @xmath0b@xmath253 .
next , let me look into the result of the full calculation shown in figs .
[ fig : ene - theta - j3](g)-(i ) and [ fig : ene - theta - j1](i)-(l ) . compared with the @xmath70-fixed calculation ,
transverse components tend to be relatively more suppressed than longitudinal components in the full calculation .
note that the longitudinal ( @xmath231 ) component is not dominant but is @xmath254 , which is comparable to the @xmath255 component .
it indicates that @xmath0b+@xmath1 cluster states are different from the ideal linear configuration of a classical picture but they show significant quantum fluctuation in the angular ( @xmath71 ) motion and are regarded as the chain - like configuration that has relatively enhanced longitudinal components with suppressed transverse components .
intrinsic energies of @xmath256 and @xmath257 before the parity and angular - monentum projections .
energies for ( a ) @xmath256 and ( b ) @xmath257 plotted on @xmath228 , and ( c ) those at @xmath159 fm plotted as functions of @xmath71 . , width=207 ] energies of the @xmath234-projected @xmath258 wave function @xmath259 with @xmath124 for ( a ) @xmath260 and ( b ) @xmath261
. energies for @xmath159 fm are plotted as functions of @xmath71 . , width=302 ] ( a)(b)(c ) energies of the @xmath234-projected @xmath258 wave function @xmath259 for @xmath260 .
( d)(e)(f ) @xmath1-cluster probability @xmath122 for @xmath115 at @xmath159 fm in @xmath3 @xmath0b+@xmath1 cluster states obtained by the @xmath70-fixed calculation and ( g)(h)(i ) that obtained by the full calculation .
, width=453 ] ( a)-(d ) energies of the @xmath234-projected @xmath258 wave function @xmath259 for @xmath261 .
( e)-(h ) @xmath1-cluster probability @xmath122 for @xmath116 at @xmath159 fm in the @xmath4 @xmath0b+@xmath1 cluster states obtained by the @xmath70-fixed calculation and ( i)-(l ) that obtained by the full calculation .
, width=604 ] the origin of the suppression of transverse components in @xmath0b+@xmath1 cluster states in the full calculation can be described by orthogonality to lower states which contain transverse components with @xmath262 fm . as shown in fig .
[ fig : ene - nopro ] for the energy surface on the @xmath263 plane , an energy pocket exists in the transverse direction ( @xmath264 ) around @xmath265 , and therefore , transverse components contribute to low - lying @xmath2n states .
although the low - lying states are compact states containing mainly configurations with small @xmath70 , transverse components with @xmath159 fm somewhat feed the low - lying states . as a result of the feeding of lower states , transverse components in the @xmath0b+@xmath1 cluster states near the threshold are suppressed .
figures [ fig : s - theta - j3 ] and [ fig : s - theta - j1 ] show the @xmath1-cluster probability @xmath266 for @xmath231 at @xmath159 fm and that for @xmath255 and @xmath267 at @xmath268 fm . as seen in [ fig
: s - theta - j3](a)-(c ) for @xmath260 , the longitudinal ( @xmath231 ) component of @xmath256 shows the largest amplitude at the @xmath3 band states ( labeled by asterisks ) and some fragmentation into neighboring states .
similarly , the longitudinal component of @xmath257 concentrates on the @xmath4 band states ( see fig . [
fig : s - theta - j1](a)-(e ) ) . on the other hand , transverse components feed states lower than @xmath0b+@xmath1-cluster states as seen in fig .
[ fig : s - theta - j3](d)(f ) and fig .
[ fig : s - theta - j1](f)(g ) ) .
consequently the @xmath1 cluster in @xmath0b+@xmath1-cluster states near the threshold tends to avoid transverse configurations so as to satisfy orthogonality to lower states .
this mechanism is consistent with the discussion of ref .
@xcite for linear - chain 3@xmath1 states in @xmath2c .
@xmath1-cluster probability @xmath266 for @xmath115 .
@xmath70 is taken to be @xmath159 fm for @xmath231 , and @xmath268 fm for @xmath255 and @xmath267 .
asterisk and down - triangle symbols show @xmath0b+@xmath1 cluster states in the @xmath3 and @xmath4 bands , respectively .
, width=396 ] @xmath1-cluster probability @xmath269 for @xmath116 .
@xmath70 is taken to be @xmath159 fm for @xmath231 , and @xmath268 fm for @xmath255 and @xmath267 .
asterisk and down - triangle symbols show @xmath0b+@xmath1 cluster states in the @xmath3 band @xmath4 bands , respectively .
, width=396 ]
i calculated positive - parity states of @xmath2n with the @xmath0b+@xmath1 cluster model and investigated @xmath0b+@xmath1 cluster states . near the @xmath1-decay threshold energy , i obtained the @xmath3 and @xmath4 rotational bands having the developed @xmath1 cluster with the @xmath0b(@xmath5 ) and @xmath0b(@xmath6 )
cores , respectively .
i assigned the @xmath179 state in the present result to the experimental @xmath5 at @xmath187 mev observed in @xmath1 scattering reactions by @xmath0b , and showed that the calculated @xmath1-decay width agrees to the experimental width .
i analyzed the component of the longitudinal configuration having an @xmath1 cluster in the longitudinal direction of the deformed @xmath0b cluster , which corresponds to a linear - chain @xmath15 structure with valence nucleons . in the spectra of @xmath2n ,
the linear - chain component concentrates at the @xmath0b+@xmath1 cluster states in the @xmath3 and @xmath4 bands .
however , the @xmath0b+@xmath1 cluster states are different from the ideal linear configuration of a classical picture but they show significant quantum fluctuation in the angular ( @xmath71 ) motion and are regarded as the chain - like configuration that has relatively enhanced longitudinal components and suppressed transverse components .
the orthogonality to low - lying states plays an essential role in the suppression of the transverse component .
the authors would like to thank dr .
suhara for fruitful discussions .
the computational calculations of this work were performed by using the supercomputers at yitp .
this work was supported by jsps kakenhi grant number 26400270 .
y. kanada - enyo and h. horiuchi , prog .
phys . suppl . *
142 * , 205 ( 2001 ) .
y. kanada - enyo m. kimura and h. horiuchi , c. r. physique * 4 * , 497 ( 2003 ) .
y. kanada - enyo , m. kimura and a. ono , ptep * 2012 * 01a202 ( 2012 ) .
h. horiuchi , k. ikeda , and k. kat , prog .
. suppl . * 192 * , 1 ( 2012 ) .
m. seya , m. kohno , and s. nagata , prog .
. phys . * 65 * , 204 ( 1981 ) .
w. von oertzen , z. phys .
a * 354 * , 37 ( 1996 ) ; * 357 * , 355 ( 1997 ) ; w. von oertzen , nuovo cimento * 110 * , 895 ( 1997 ) . k. arai , y. ogawa , y. suzuki and k. varga , phys .
c * 54 * , 132 ( 1996 ) .
a. dote , h. horiuchi and y. kanada - enyo , phys .
c * 56 * , 1844 ( 1997 ) .
y. kanada - enyo , h. horiuchi and a. dot , phys .
c * 60 * , 064304 ( 1999 ) .
n. itagaki and s. okabe , phys .
c * 61 * , 044306 ( 2000 ) .
y.ogawa , k.arai , y.suzuki and k.varga , nucl . phys . *
a673 * , 122 ( 2000 ) .
y. kanada - enyo , phys .
c * 66 * , 011303 ( 2002 ) .
p. descouvemont , nucl .
a * 699 * , 463 ( 2002 ) .
m. ito , k. kato and k. ikeda , phys .
b * 588 * , 43 ( 2004 ) .
m. ito , phys .
b * 636 * , 293 ( 2006 ) .
m. freer _ et al .
lett . * 82 * , 1383 ( 1999 ) .
m. freer _ et al .
c * 63 * , 034301 ( 2001 ) .
a. saito,_et al . _ ,
* a738 * , 337 ( 2004 ) ; a. saito _ et al .
_ , mod .
a * 25 * , 1858 ( 2010 ) .
n. curtis _ et al .
_ , phys .
c * 70 * , 014305 ( 2004 ) .
et al . _ ,
* a753 * , 263 ( 2005 ) .
m. gai , r. keddy , d. a. bromley , j. w. olness and e. k. warburton , phys .
c * 36 * , 1256 ( 1987 ) .
n. furutachi , s. oryu , m. kimura , a. dote and y. kanada - enyo , prog .
phys . * 119 * , 403 ( 2008 ) . c. fu _ et al .
_ , phys . rev .
c * 77 * , 064314 ( 2008 ) .
e. d. johnson _ et al .
_ , eur .
phys . j. a * 42 * , 135 ( 2009 )
. w. von oertzen _
et al . _ ,
j. * a 43 * , 17 ( 2010 ) .
n. curtis _ et al .
_ , phys . rev .
c * 66 * , 024315 ( 2002 ) .
n. i. ashwood _ et al .
_ , j. phys .
g * 32 * , 463 ( 2006 ) .
s. yildiz _ et al . _ , phys .
c * 73 * , 034601 ( 2006 ) .
w. scholz , p. neogy , k. bethge and r. middleton , phys .
c * 6 * , 893 ( 1972 ) .
p. descouvemont , phys .
c * 38 * , 2397 ( 1988 ) .
g. v. rogachev _ et al .
_ , phys .
c * 64 * , 051302 ( 2001 ) . v. z. goldberg _ et al .
_ , phys .
c * 69 * , 024602 ( 2004 ) .
m. kimura , phys .
c * 75 * , 034312 ( 2007 ) .
a. tohsaki , h. horiuchi , p. schuck and g. ropke , phys . rev
. lett . * 87 * , 192501 ( 2001 ) . y. funaki , a. tohsaki , h. horiuchi , p. schuck and g. ropke , phys .
c * 67 * , 051306 ( 2003 ) . t. yamada and p. schuck , phys .
c * 69 * , 024309 ( 2004 ) .
y. funaki , t. yamada , h. horiuchi , g. ropke , p. schuck and a. tohsaki , phys .
* 101 * , 082502 ( 2008 ) .
h. morinaga , phys . rev . * 101 * , 254 ( 1956 ) .
h. morinaga , phys . lett . * 21 * , 78 ( 1966 ) .
y. suzuki , h. horiuchi and k. ikeda , prog .
. phys . * 47 * , 1517 ( 1972 ) .
n. itagaki , s. okabe , k. ikeda and i. tanihata , phys .
c * 64 * , 014301 ( 2001 ) .
t. suhara and y. kanada - enyo , phys .
c * 82 * , 044301 ( 2010 ) .
t. ichikawa , j. a. maruhn , n. itagaki and s. ohkubo , phys .
lett . *
107 * , 112501 ( 2011 ) .
t. suhara and y. kanada - enyo , phys . rev .
c * 84 * , 024328 ( 2011 ) .
t. mo and h. r. weller , phys .
c * 8 * , 972 ( 1973 ) .
y. kanada - enyo , h. morita and f. kobayashi , arxiv:1504.02594 [ nucl - th ] . | i investigate @xmath0b+@xmath1 cluster states of @xmath2n with a @xmath0b+@xmath1 cluster model . near the @xmath1-decay threshold energy ,
i obtain @xmath3 and @xmath4 rotational bands having @xmath0b(@xmath5)+@xmath1 and @xmath0b(@xmath6)+@xmath1 components , respectively .
i assign the band - head state of the @xmath3 band to the experimental @xmath5 at @xmath7=13.19 mev of @xmath2n observed in @xmath1 scattering reactions by @xmath0b and show that the calculated @xmath1-decay width is consistent with the experimental data .
i discuss an @xmath1-cluster motion around the @xmath0b cluster and show that @xmath0b+@xmath1 cluster states contain significant components of a linear - chain 3@xmath1 configuration , in which an @xmath1 cluster is localized in the longitudinal direction around the deformed @xmath0b cluster . |
strong spin - orbit coupling is the fundamental origin of several topological phases of matter , such as topological insulators and superconductors in two and three spatial dimensions @xcite .
the salient features of these systems are ( i ) an insulating bulk ( electrical or thermal ) and ( ii ) topologically protected metallic surface states ( which could be superconducting depending on the situation ) . among such gapless phases , the ones ( so - called weyl materials ) formed by chiral weyl fermions
have attracted considerable recent attention and may serve as an experimental platform where various exotic phenomena ( such as axionic electrodynamics and chiral anomaly ) , which were originally proposed in the context of high energy physics and quantum field theory , may be observed in the context of solid state physics @xcite . the current work is a comprehensive theoretical study of the strong - field magnetotransport properties of generalized three - dimensional weyl systems ( with arbitrary monopole charges ) in the quantum limit .
weyl fermions may result from complex band structures in strong spin - orbit coupled semiconductors @xcite , multilayer heterostructures @xcite and they can also be found inside a broken symmetry phase in strongly correlated materials , such as 227 pyrochlore iridates , as emergent quasiparticles @xcite . the weyl semimetals ( wsms ) can be classified into two broad categories .
( i ) inversion ( @xmath4 ) symmetry breaking wsms that are commonly found in weakly correlated semiconductors , and ( ii ) time - reversal ( @xmath5 ) odd wsms that can be found in strongly correlated materials with comparable strength of spin - orbit coupling and electronic interaction .
however , irrespective of the microscopic origin , wsms are composed of weyl nodes in the reciprocal space , where kramers non - degenerate valence and conduction bands touch each other at the so - called diabolic points in the momentum space @xcite .
the weyl nodes act as the source ( monopole ) and sink ( anti - monopole ) of abelian berry flux , and in its close vicinity , weyl fermions can be identified as left ( right ) chiral fermions .
a _ no - go theorem _ guarantees the existence of an equal number of left and right chiral fermions in the system @xcite .
the monopole charge ( @xmath0 ) also dictates the amount of berry flux enclosed by a plane perpendicular to the line joining these two points and in turn defines the _ integer topological invariant _ of the system .
thus monopole charge permits a topological classification of weyl semimetals , and for @xmath1 and @xmath2 we can call them single , double and triple wsm , respectively .
the most common weyl system has just @xmath6 whereas the system with arbitrary @xmath0 can be regarded as a generalized weyl system .
( in the current work we consider magnetotransport in single , double , and triple wsms although our theory is generalizable to even higher values of @xmath0 , but it is well - known that weyl systems with monopole charge @xmath7 is not allowed in three - dimensional lattice systems , and therefore our work applies to all possible physical weyl materials that can be studied in the laboratory ) .
the dispersion of weyl quasiparticles in these systems along various high symmetry directions is shown in fig .
[ dispersion ] . although the weyl nodes are often accompanied by _
monopole charge , it is nonetheless conceivable to realize weyl nodes with higher integer charge and the system with arbitrary @xmath0 can be regarded as a generalized weyl system , such as the proposed double wsm in hgcr@xmath8se@xmath9 and srsi@xmath8 @xcite .
even though , the material existence of a triple wsm has remained elusive so far , with the anticipated progress of materials science the discovery of such a generalized triple wsm is certainly possible .
it is well known that weyl systems with monopole charge @xmath7 is not allowed in three - dimensional lattice systems @xcite , and therefore our work applies to all possible physical weyl materials that can be studied in the laboratory .
most , if essentially not all , of the substantial theoretical literature on wsms have focused on the single wsm because the existing experimental materials are all wsms with @xmath6 .
on the other hand , there are significant differences in the properties of generalized wsms depending on their monopole charges , necessitating theoretical studies of multi - wsms with higher ( @xmath10 ) values of monopole charges , which is what we do in the current work considering @xmath11 and @xmath2 , and contrasting their properties with @xmath6 wsm ( and with each other ) .
the integer topological invariant also determines various thermodynamic , transport , and topological properties of a wsm .
for example , one of the hallmark signatures of a wsm is the presence of topologically protected _ fermi arc _ @xcite that has recently been seen through angle resolved photoemission spectroscopy ( arpes ) and scanning tunneling microscopy ( stm ) measurements @xcite . in a multi - wsm the fermi arc bears additional @xmath0 fold degeneracy , where @xmath0 is the monopole charge of weyl nodes .
therefore , in double and triple wsms , the fermi arc respectively possesses two and three fold orbital degeneracies , which , for example , can be detected in arpes and stm .
we here explore the ramifications of the underlying topological invariant of a pristine wsm , when the system is placed in a strong magnetic field .
our findings are the following . 1 .
when placed in a strong magnetic field , wsms undergo landau level ( ll ) quantization , supporting _ exactly _
@xmath0 number of zeroth landau levels ( zlls ) , irrespective of the field orientation ( within continuum approximation ) .
all lls , including the zeroth ones , disperse along the direction of the applied field .
thus , the orbital degeneracy of zll ( @xmath0 ) gets tied with the monopole charge or topological invariant of the weyl nodes .
2 . even though @xmath0 dispersive zlls always go through zero energy ( in the continuum limit ) , they are in general non - degenerate , unless the field is applied along the axis on which weyl nodes reside in the absence of the field [ see figs .
[ ll_dw_general ] and [ ll_tw_general ] ] .
3 . density of states ( dos ) displays nontrivial dependence ( such as its periodicity ) on the tilting angle of the magnetic field away from a high symmetry direction , separating the weyl nodes [ see figs . [
fig : dos_dw ] and [ fig : dos_tw ] ] .
such intriguing features can be observed in _ angle resolved _ quantum oscillation measurements @xcite .
direction in any weyl semimetal .
dispersion in the @xmath12 plane respectively for ( b ) single , ( c ) double and ( d ) triple weyl semimetals .
notice that @xmath13 , where @xmath0 is the monopole charge and @xmath14 , but @xmath15 in all weyl semimetals.,width=302,height=226 ] a strong magnetic field causes an _ effective _ reduction of the dimensionality of the system .
in particular , each branch of zlls cuts the zero energy ( where the fermi energy is pinned ) at two isolated points in the brillouin zone , yielding two field - induced weyl nodes , around which the quasiparticles are once again described as one - dimensional chiral ( left and right ) fermions ( we note that the effective reduction of a generic three dimensional electron system to an apparent one dimensional electron gas along the magnetic field direction is of course a well - known effect of the strong field limit where ll coupling can be ignored for wsms this dimensional reduction leads to effectively chiral one dimensional fermion systems ) . if an electric field ( @xmath16 ) is now applied in parallel to the magnetic field , it can cause charge transfer from left to right weyl nodes , causing _ violation of separate conservation laws for left and right chiral fermions _
, captured by the following anomaly equation @xcite @xmath17 here , @xmath18 is the electric charge , @xmath19 is the charge ( for @xmath20 ) and current ( for @xmath21 ) operator for left / right chiral fermion , and @xmath22 can be identified as the total degeneracy of effective one - dimensional system . the total charge and current density
is respectively given by @xmath23 and @xmath24 , where @xmath25 is a two component spinor , describing the emergent one - dimensional world .
due to the reduced dimensionality of the system , one can define @xmath26 , and accordingly we can define axial or chiral charge and current density as @xmath27 and @xmath28 .
hence , one can express eq .
( [ anomaly-1d ] ) as @xmath29 . in the language of quantum field theory , such violation of separate conservation laws for left and right fermions
is known as _ chiral anomaly _ , specific to odd spatial dimensions . in regular wsm
, the degeneracy of the zll plays the role of @xmath22 , whereas in wsms with @xmath30 , @xmath31 ll degeneracy .
therefore , upon identifying @xmath22 as the _ total _ degeneracy of the generalized wsm zll , we arrive at the quantum field theoretic _ adler - jackiw - bell _ anomaly equation @xcite @xmath32 now generalized for the generalized weyl system , constituted by weyl nodes with an arbitrary monopole charge @xmath0 .
notice that here electric ( @xmath33 ) and magnetic ( @xmath34 ) fields are strictly _ static abelian background fields_. we also stress that in the above expression @xmath35 can _ never _ be negative as the @xmath3-linear dependence arises from the ll dos , and insensitive to the direction of applied electric and magnetic field .
such a tantalizing quantum field theoretic chiral anomaly analogy has led to considerable theoretical @xcite and experimental @xcite activities aimed at demonstrating the manifestation of chiral anomaly through lmt .
it is , however , unclear that these chiral anomaly considerations apply directly in solid state materials since the nonconservation chiral current in the quantum field theoretic adler - jackiw - bell anomaly crucially depends on the existence of an _ unbounded _ linear dispersion of weyl fermions @xcite , whereas the energy dispersion in all condensed matter systems is unavoidably bounded by the natural lattice cut off .
however in a pioneer work nielsen and ninomiya showed that certain gapless semiconductors accommodating linear touching of valence and conduction bands can give rise to lmt arising essentially from the chiral anomaly physics of current nonconservation @xcite . in this regard , we must recall that almost thirty years before the proposal of nielsen and ninomiya , adam and argyres demonstrated the existence of lmt for conventional three - dimensional fermi liquids , i.e. , ordinary metals and doped semiconductors , without invoking the notion of chiral or axial anomaly @xcite .
their derivation of lmt solely relies on few generic features of a three - dimensional electronic system in a strong magnetic field : the existence of ( i ) one - dimensional dispersive lls since momentum along the applied magnetic field is a conserved quantity , and ( ii ) at least one partially filled ll that cuts fermi energy at two isolated points ( or in general even number of points in the brillouin zone ) . therefore , the existence of lmt in the absence of the classical lorentz force ( when @xmath33 and @xmath34 are parallel ) although necessarily points toward subtle quantum mechanical effects , yet its direct connection with chiral anomaly is not obvious .
specifically , the chiral anomaly implies an lmt through the nielsen - ninomiya mechanism , but the reverse is not true , i.e. , the existence of an lmt does not necessarily imply a chiral anomaly
lmt is necessary , but by no means sufficient , for the universal existence of the chiral anomaly .
recently goswami _ et .
_ in ref .
@xcite have demonstrated the generic existence of lmt for _ a generic three dimensional metal _ " .
specifically , the chiral anomaly implies an lmt through the _ nielsen - ninomiya mechanism _ , but the reverse remains to be established .
we here follow the philosophy of ref .
@xcite and show that any wsm with arbitrary monopole charge display lmt when placed in a strong magnetic field without invoking the physics of chiral anomaly . for the sake of technical simplicity
we assume that only the manifold of zlls is partially filled .
however , our analysis can be generalized to demonstrate generic lmt in weyl materials even when multiple lls are partially filled , i.e. , the system is in the strong field limit ( see sec .
[ generalanomaly ] ) .
the lmt we predict theoretically is a generic property of impurity scattering in the strong - field limit as was originally pointed out in ref .
@xcite a long time ago in the context of doped bulk semiconductors . ) and magnetic ( @xmath36 ) fields ( always parallel ) .
quasiparticle dispersion along @xmath37 direction is @xmath38 , while in the @xmath12 plane is proportional to @xmath39 , where @xmath40 . here
, @xmath41 parametrizes the tilting of magnetic field away from the @xmath37 direction .
weyl nodes are separated along the @xmath37 direction . ,
width=188,height=188 ] in this context few comments are due . in the presence of the periodic potential in a solid , arising from the ionic lattice ,
quasiparticles perform bloch oscillations and consequently the system can not sustain any finite steady state current @xcite .
nevertheless , in the presence of a momentum relaxation mechanism , which is naturally offered by impurity scattering , quasiparticles lose the freedom to perform a complete bloch oscillation and a metallic system can support finite steady state current , when the relaxation or transport lifetime ( @xmath42 ) is much shorter than the bloch oscillation period . in this work ,
we focus on the low temperature situation where impurity scattering is the primary scattering mechanism for transport as phonons are mostly frozen out .
furthermore , in the presence of strong magnetic fields , i.e. when @xmath43 where @xmath44 is the cyclotron frequency , the path of the carriers is so curved that the period of completing a cyclotron orbit becomes comparable to that between two successive collisions .
therefore , the transport / relaxation time acquires a strong field dependence and @xmath45 . in this strong - field limit , one can not assume the relaxation time to be field - independent which is the standard approximation in metallic transport since in metals one is often in the weak - field semiclassical limit where @xmath46 .
such a field dependence of @xmath42 is extracted here by using _ quantum boltzmann equation _ , but omitted in the semiclassical theory ( where the transport relaxation time is taken to be field - independent ) , which therefore is only applicable for extremely weak magnetic fields ( @xmath47 ) .
but , for sufficiently weak magnetic fields one needs to account for yet another pure quantum mechanical effect , the _ weak localization _ @xcite , which arises from quantum interference of electron paths our theory neglects the weak localization physics since the strong magnetic field completely breaks the time reversal invariance .
presently it is unknown how to incorporate all these effects in a unified theoretical framework , and restricting ourselves to the strong field quantum limit @xmath43 we establish the following theoretically : 1 . in the presence of both gaussian and coulomb impurities ,
we extract the magnetic field dependence of transport lifetime using quantum boltzmann equation , and show that all wsms manifest _ positive _ longitudinal magnetoconductivity ( lmc ) or negative longitudinal magnetoresistivity ( lmr ) , when the magnetic and electric fields are applied parallel to each other ( see fig . [ geometry ] ) .
2 . due to the carrier - induced screening of coulomb potential ( arising from finite dos of lls )
, we mainly focus on gaussian impurities and show that when the fields are applied along a high symmetry axis separating the weyl nodes , the lmc increases linearly with @xmath3 in all wsms ( see fig .
[ fig : conductivity - singleweyl ] ) .
the lmc in the presence of only coulomb impurities grows as @xmath48 , but only in extreme strong field limit ( we believe that the same result applies to screened coulomb impurity potential also since screening renders the coulomb potential into a short - range scattering potential similar to the gaussian case ) .
although such a linear increase of lmc is insensitive to the field orientation in single wsms , lmc develops _ nonlinear _ dependencies on the magnetic field in double and triple wsm when the field is tilted away from the high symmetry axis ( see figs .
[ fig : conductivity-1 ] and [ fig : conductivity-2 ] ) , which can be tested in proposed double wsms hgcr@xmath8se@xmath9 and srsi@xmath8 .
thus , the quantitative field dependence of lmc depends on the wsm monopole charge whereas the qualitative existence of the negative lmr itself ( positive lmc ) is generic in all wsms in the strong - field regime .
we find that typically the relaxation lifetime due to gaussian impurities @xmath49 , where @xmath50 is the same quantity but due to ionic impurities .
hence , the transport lifetime is dominated by gaussian impurities and in our analysis we mainly focus on short range elastic scatterer .
in addition , we also investigate the role of electronic interaction on lmt as the temperature ( magnetic field ) is gradually decreased ( increased ) .
the dos in a pristine wsm scales as @xmath51 and the weyl nodes are extremely robust against weak short - range electron - electron interaction .
however , due to landau quantization , the dos for the emergent one - dimensional system is _ constant _ , which in turn can trigger various density - wave instabilities even when the electronic interaction is sufficiently weak .
the quintessential effects of interactions can be summarized as the following . 1 . depending on the microscopic details
, zlls can undergo a weak coupling ( bcs - like ) instability in the charge - density - wave ( cdw ) or spin - density - wave ( sdw ) channel as common in one - dimensional electron systems .
irrespective of the actual nature of the ordering the density - wave orders break the translational symmetry and gap out the zlls .
when the field is applied along the high symmetry axis all @xmath0 number of zlls are expected to suffer density - wave ordering _
simultaneously_. thus , below the transition temperature ( @xmath52 ) lmc ( lmr ) becomes _ negative ( positive ) _ , reversing the trend compared with the normal phase with no density wave ordering .
however , upon tilting the field away from the high symmetry axis , the exact @xmath0 fold degeneracy of zlls is lifted , and the system can undergo a _ cascade _ of density - wave transitions with distinct transition temperatures . hence , lmc / lmr in a multi - wsm is expected to display @xmath0 fold discontinuity before it becomes negative / positive .
although such ordering in a clean system ( devoid of any impurity ) should take place for arbitrarily weak interaction , disorder naturally reduces the propensity towards such translational symmetry breaking ordering @xcite ( we mention that there is no anderson theorem here protecting the symmetry - broken phase against impurity scattering ) . thus clear signatures of cdw or sdw ordering can only be observed in clean systems at sufficiently low temperature and for strong enough magnetic fields .
the rest of the paper is organized in the following way . in sec .
[ weylintro ] , we discuss the possibility of realizing weyl semimetals with different monopole charge and address their topological properties ( such as bulk - boundary correspondence through degeneracy of fermi arc ) .
[ landaulevel ] is devoted to the demonstration of ll spectra in single , double and triple wsms for arbitrary orientation of the magnetic field .
dos in these systems in the presence of strong magnetic field is presented in sec . [ dos ] .
longitudinal magnetotransport for all members of the weyl family constitutes the central theme of sec .
[ transport ] .
effects of various density - wave ordering on lmt are discussed in sec .
[ densitywave ] .
we summarize the main findings in sec .
[ summary ] .
some technical details are relegated to appendices .
[ weylintro ] a wsm is realized when a solid state system lacks inversion and/or time - reversal symmetry , and the kramers non - degenerate valence and conduction bands touch each other at isolated points in the brillouin zone , known as weyl nodes . the simplest realization ( composed of only two weyl points )
arises from the following tight - binding model in a cubic lattice
@xmath53 \psi_{\boldsymbol k},\ ] ] where @xmath54 , @xmath55 is a two component spinor and @xmath56s are three pauli matrices . with the following choice @xmath57 -t \cos ( k_z a)+m_z , \notag\end{aligned}\ ] ] where @xmath58 is the lattice spacing , two weyl nodes are located along one of the @xmath59 axes , namely at @xmath60 $ ] . now ,
if we choose @xmath61 and @xmath62 , two weyl nodes with monopole charge @xmath63 are realized at @xmath64 , where @xmath65)$ ] .
the linearized hamiltonian for single wsm in the vicinity of the @xmath66 point is @xmath67 where @xmath68s are the fermi velocity of weyl quasiparticles , @xmath69 , @xmath70 , @xmath71 , and the momentum @xmath72 is measured from the @xmath73 points . in the close proximity to these two weyl points the low energy excitations are respectively described by left and right chiral fermion , constituting source and sink of the abelian berry curvature , respectively . in an appropriate crystallographic environment it is also conceivable to realize weyl points with higher monopole charge @xcite .
in particular weyl nodes with monopole charge _
two _ and _ three _ can respectively be stabilized when the underlying lattice possesses tetragonal @xmath59 and @xmath74 symmetries , respectively @xcite .
a double wsm , for example , can be realized from eq .
( [ lattice ] ) by choosing @xmath75 $ ] , @xmath76 , while leaving @xmath77 unchanged @xcite .
but , in three spatial dimensions it is impossible to realize weyl nodes with monopole charge larger than three from an underlying lattice model .
the above construction can also be viewed in the following way that will allow us to identify the topological invariant of these systems .
note that wsms can be constructed by appropriately stacking two - dimensional layers of _ quantum anomalous hall insulators _ ( qahi ) in the brillouin zone along the @xmath78-direction .
a two - dimensional qahi supports quantized hall conductivity @xmath79 , where @xmath0 represents the number of one - dimensional chiral edge states . for each @xmath78 ,
satisfying @xmath80 , the pseudospin texture is _ skyrmion _ , with skyrmion number @xmath0 , and two weyl nodes appear as _ singularities _ in the brillouin zone , across which the skyrmion number _ jumps _ by @xmath81 . when stacked along the @xmath78 direction
, the chiral edge modes from two - dimensional layers of qahi produce the fermi arc , _ possessing @xmath0-fold orbital degeneracy_. the topological invariant of a wsm can be assessed from the gauge invariant berry curvature defined as @xcite @xmath82,\end{aligned}\ ] ] where @xmath83 , and @xmath84 corresponds to valence and conduction band , respectively . upon evaluating the components of the abelian berry curvature we plot it in figs .
[ topology - sw ] , [ topology - dw ] and [ topology - tw ] respectively for single , double and triple wsm .
the components of berry curvature @xmath85 and @xmath86 are odd function of @xmath87 and @xmath88 respectively .
hence , the number of field lines coming in and out of the @xmath89 or @xmath90 planes are equal , and the net berry flux through these planes is zero . on the other hand
, @xmath91 is an even function of @xmath92 and the net berry flux through the @xmath12 plane is @xmath93 , where @xmath94 is the chern number of effective two - dimensional system for a fixed @xmath78 . for @xmath80 , @xmath95 , while @xmath96 when @xmath97 .
therefore , two weyl points in a general wsm act as source and sink of abelian berry curvature , across which the chern number of the underlying two - dimensional system jumps by an integer amount @xmath0 . in the low energy limit
the hamiltonian for an wsm , constituted by weyl nodes with monopole charge @xmath81 is compactly written as @xmath98 after linearizing the generalized version of the hamiltonian from eq .
( [ swsm_top ] ) for general wsm around @xmath99 , where @xmath100 . in the above equation ,
therefore , @xmath78 is measured from @xmath99 .
thus , @xmath101 and @xmath102 respectively bear the unit of fermi velocity and inverse of mass .
the majority of the known weyl materials only break the inversion symmetry and the weyl nodes are placed at different energies .
notice that the wsm arising from the tight - binding model , defined in eq .
( [ lattice ] ) , breaks both time - reversal and inversion symmetries , but the weyl nodes are still located at the same energy . with simple modifications in this tight - binding model , for example by adding a term @xmath103 \sigma_0 $ ] to eq.([lattice ] ) , the weyl nodes can be placed at different energies , namely at @xmath104 . here ,
@xmath105 and @xmath106 are respectively regular and chiral chemical potentials
. however , our following discussion on various aspects of weyl fermions in a strong magnetic field is qualitatively insensitive to such details of the system .
next we demonstrate how the underlying topological invariant of a wsm manifests through the ll spectrum when the system is subject to a strong magnetic field .
[ landaulevel ] ( color online ) landau level ( ll ) spectrum in single weyl semimetal , obtained from eq .
( [ hamillowenergy ] ) when @xmath6 . for details
see appendix [ landaulevel_append ] . here
the zeroth ll ( red curve ) is nondegenerate , and the energy @xmath107 ( in units of @xmath108 ) is measured about a reference or zero point energy @xmath109 . at
half - filling the zeroth ll cuts the fermi energy ( dotted line ) at @xmath110 ( we set @xmath111 for convenience ) @xcite . similar ll spectrum is found for double and triple weyl semimetals , for which , however , the zeroth ll has two and three fold degenerate , respectively . ] effects of the magnetic field applied in an arbitrary direction , can be captured through minimal coupling @xmath112 , where @xmath113 is the vector potential , @xmath18 is electronic charge and the magnetic field is given by @xmath114 . before delving with the situation when the magnetic field is tilted away from the @xmath37 direction ( parametrized by @xmath41 ,
see fig . [ geometry ] )
, let us first focus on a simpler situation when the field is applied along the @xmath37-direction , i.e. @xmath115 , for which momentum @xmath78 is a conserved quantity . we can analytically obtain the ll spectrum @xcite . for simplicity , we work with the landau gauge @xmath116 , and define the raising and lowering operators as @xmath117 and @xmath118 , where @xmath119 is the magnetic length , and @xmath120 = -i\hbar eb$ ] .
the low energy hamiltonian for a general weyl semimetal then reads as @xmath121 we here omit the term @xmath122 with respect to @xmath87 and @xmath88 for small momentum ( near the weyl nodes ) , mainly since inclusion of such term causes an overall shift in the ll energy at least when @xmath123 .
the ll spectrum can then be readily obtained , yielding @xmath124 for @xmath125 for single , double and triple wsm , respectively . for simple wsm
the zll is comprised of a single branch with energy @xmath126 , while for double wsm @xmath127 and for a triple wsm @xmath128 .
thus , the zll for double and triple wsm enjoys additional _ two _ and _ three _ fold orbital degeneracy , respectively .
we obtain exactly @xmath0 number of zll in the presence of bounded dispersion from a lattice model , as shown in fig .
[ fig : spectrum - lllattice ] when @xmath123 .
details of the calculation are provided in appendix [ landaulevel_append ] .
_ thus , the integer topological invariant of a wsm or monopole charge of the weyl nodes sets the orbital degeneracy of the zll , at least when the field is applied along the separation of the weyl nodes_. when the underlying lattice potential is taken into account , giving rise to bounded dispersion , the zll cuts the zero energy at @xmath129 , and in the following we focus near one such weyl point .
next we will show that the orbital degeneracy of zll remains unaffected with the tilting of the magnetic field away from the @xmath37-direction ( @xmath130 ) , at least within the continuum description . for concreteness , we take the following vector potential @xmath131 , giving rise to @xmath132 .
the commutation relations between different momentum operators are given by @xmath120 = -i\hbar eb \cos\theta$ ] and @xmath133 = i\hbar eb \sin\theta$ ] .
we can define a pair of raising and lowering operators as @xmath134 , \notag\\ a^\dagger & = \dfrac{\ell}{\sqrt{2}\hbar } [ \pi_x + i(\pi_y \cos\theta - \pi_z \sin\theta ) ] , \end{aligned}\ ] ] which satisfies the standard commutation relation @xmath135 = 1 $ ] .
the momentum along the magnetic field , defined as @xmath136 , is a conserved quantity , and we find @xmath137 the ll spectrum for an arbitrary orientation of the magnetic field @xmath138 in a multi - wsm can not be obtained analytically .
the numerically obtained ll spectrum for double and triple wsm are respectively shown in fig . [ ll_dw_general ] and
[ ll_tw_general ] , for various choices of @xmath41 . for simplicity ,
we have only shown the ll structure near one weyl node , hosting left chiral fermion .
a similar ll structure is also realized for right chiral fermion , and all lls ( including the zeroth one ) are bounded due to the underlying lattice .
therefore , when @xmath130 the exact degeneracy among the chiral zlls is lifted . however , there is always @xmath0 number of chiral zll crossing the zero energy .
such an outcome can be substantiated from the fact that in the absence of any magnetic field the monopole charge of the weyl nodes is @xmath81 , which in turn determines the orbital degeneracy of the zll .
although the ll spectrum obtained from the continuum description of wsm captures most of the essential features , a comment in this context is in order .
when the field is tilted from the @xmath37 direction the two copies of zlls in double wsm cuts the zero energy at momentum @xmath139 a^{-1}$ ] , where @xmath140 is dependent on the tilting angle , as shown in fig .
[ ll_dw_general ] , when the system is at half - filling .
similarly , in triple wsm the zll cuts the zero energy at @xmath141a^{-1}$ ] for @xmath142 . in figs .
[ ll_dw_general ] and [ ll_tw_general ] , the conserved momentum @xmath143 is measured from one of the emergent weyl nodes in magnetic fields , located at @xmath144 .
next we discuss the resulting dos in the presence of magnetic field .
[ dos ] it is instructive to separately study the dos in single , double and triple wsm , when placed in a magnetic field , as it can be directly tested in angle resolved quantum oscillation measurements @xcite . in the presence of strong magnetic fields the kinetic energy in the plane perpendicular to @xmath34
is completely quenched , while the lls remain dispersive along the applied magnetic field .
thus the magnetic field causes an effective dimensional reduction of the system , and the wsm in the external magnetic field can be viewed as a _ collection of one - dimensional systems with multiple subbands_. the dos of such a system can be found from the following definition @xmath145 where @xmath146 labels different subbands , and @xmath147 is the conserved momentum of the one - dimensional conducting channel , along the magnetic field direction , and @xmath148 is the linear dimension along the field . the dos for a single wsm can be obtained in a closed - form .
we first note that the dispersion for a single wsm can be cast in the following form @xmath149 where @xmath150 and @xmath151 .
in such a system each @xmath152 lls crosses the fermi energy @xmath153 twice , at @xmath154 , whereas the @xmath155 ll only cuts the fermi energy once , at @xmath156 .
therefore , the dos for this single wsm is given by @xmath157 , \end{aligned}\ ] ] where @xmath158 is the heaviside theta function .
( color online ) dos for landau levels in a double - weyl semimetal , in units of @xmath159 .
the parameters are the same as those in fig .
[ ll_dw_general ] .
the black curves are the dos when the applied magnetic field is along the @xmath37 direction ( @xmath160 ) , see eq . .
] ( color online ) dos for landau levels in a triple - weyl semimetal , in units of @xmath159 .
the parameters are the same as those in fig .
[ ll_tw_general ] .
the black curves are the dos when the applied magnetic field is along the @xmath37 direction ( @xmath160 ) , see eq . .
] similarly we can analytically find the dos for double wsm when the field is along the @xmath78 direction .
the dispersion relation is then readily obtained from eq . after replacing @xmath161 by @xmath162 , where @xmath163 , and the dos for such
a system is given by @xmath164 , \label{eq : dosexact}\end{aligned}\ ] ] where we have accounted for the two - fold degeneracy of the zll .
the above expression for dos in double wsm is plotted as black lines in fig . [
fig : dos_dw ] .
similarly , it is easy to show that when the field is applied along the @xmath37 direction the dos in triple wsm is given by @xmath165 , \label{eq : dosexact_tw}\end{aligned}\ ] ] where @xmath166 , as shown by the black lines in fig .
[ fig : dos_tw ] . for arbitrary orientations of the magnetic field , we compute the dos numerically . in fig .
[ fig : dos_dw ] we display the dos for three different orientations of the field in a double wsm .
we find that the ll spacing continues to increase when the field is gradually tilted from @xmath167 to @xmath168 .
we also notice that as @xmath169 an additional peak in dos gradually develops within the manifold of zlls , which , however , can be an artifact of continuum model .
a similar result also holds for triple wsm , as shown in fig .
[ fig : dos_tw ] .
the dos dictates the pattern of quantum oscillation measurements ( through shubnikov - de haas effect or de haas - van alphen effect ) of all transport and thermodynamic quantities .
therefore , the proposed anisotropic structure in dos for multi - wsm can be detected from angle resolved quantum oscillations @xcite .
such experiments can also test the robustness of the emergent peak in dos within the zll of double wsm by tilting the magnetic field from the @xmath37 direction toward the @xmath12 plane .
[ transport ] as shown in sec .
[ landaulevel ] , in the presence of strong magnetic fields the quasiparticle spectrum in any wsms breaks into a set of lls that remains dispersive along the direction of the applied field .
therefore , the application of an external magnetic field effectively breaks the system into a set of one - dimensional conducting wires , which is responsible for lmt . for the sake of simplicity and specificity
we assume that the external magnetic and electric ( @xmath33 ) fields are always parallel to each other ( see fig . [ geometry ] ) .
when @xmath34 and @xmath33 are not parallel , the transport occurs through both longitudinal and transverse components .
however , the analysis of the transport properties for an arbitrary relative orientation of @xmath34 and @xmath33 fields falls outside the scope of the current work . to sustain a steady - state current , we take into account impurity - induced backscattering between the magnetic field induced emergent one dimensional weyl nodes , while the forward scattering contributes to the _ dingle factor _ for quantum oscillations .
when such scattering is accounted for , the lmc ( intra - band ) can be written as @xmath170 where the sum is over all partially filled landau levels labeled by @xmath171 , @xmath172 is the additional degeneracy due to internal degrees of freedom ( such as spin ) , @xmath173 is the field dependent fermi velocity obtained by linearizing the ll spectrum around the emergent one - dimensional weyl nodes , and @xmath174 is the transport lifetime ( obtained from backscattering ) .
we here use the linearized boltzmann equation to compute the lifetime and numerically extract the effective field dependent fermi velocity .
we assume that only the subspace of zlls is partially filled and contributes to lmt .
generalization of our results is discussed in sec .
[ generalanomaly ] .
two sources of elastic scattering are taken into account : ( i ) gaussian disorder with @xmath175 where @xmath176 is the range of the impurity potential , and ( ii ) long - range ionic impurity scattering characterized by the screened coulomb potential @xmath177 where @xmath178 is the strength of the coulomb impurity , and @xmath179 is the thomas - fermi wave vector @xmath179 , @xmath180 which is also a function of the magnetic field .
therefore with an increasing magnetic field , the screened coulomb impurity potential gets more and more short - ranged as @xmath181 increases .
hence , in the strong field limit , the lmt is dominated by gaussian impurities .
the increase in screening with increasing magnetic field follow from the enhancement of the dos by the magnetic field .
note that in the extremely strong field limit the screened coulomb potential becomes very small as @xmath181 diverges . also , we find that typically the relaxation time due to gaussian impurities is much smaller than that due to coulomb impurities . therefore , transport lifetime is dominated by the former source of elastic scattering .
the transport lifetime @xmath174 can be obtained from the linearized boltzmann equation , which reads as @xcite @xmath182 , \notag\end{aligned}\ ] ] where the impurity matrix element squared is given by @xmath183 .
\notag\end{aligned}\ ] ] after some involved algebra , the linearized boltzmann equation can be cast in the following compact form @xmath184 \notag\\ & + \sum_{\beta\neq\alpha } \dfrac{\tau_{\beta}({k_{f,\beta}})}{|v_{\alpha}({k_{f,\alpha}})| } \biggl [ |u_{\alpha,\beta}({k_{f,\alpha}}+{k_{f,\beta}})|^2 - |u_{\alpha,\beta}({k_{f,\alpha}}-{k_{f,\beta}})|^2 \biggr ] , \label{eq : simplifiedboltzmann}\end{aligned}\ ] ] where @xmath185 is the magnitude of the one - dimensional fermi momentum in landau level @xmath171 , @xmath186 is the corresponding fermi velocity and @xmath187 is impurity density .
the effective one - dimensional interaction potential @xmath188 has the following form @xmath189 where @xmath190 is the impurity potential , and @xmath191 satisfies the relation @xmath192 one can then solve the coupled equations in eq . and obtain the transport lifetime @xmath193 for each ll contributing to lmc .
we can obtain the analytical expression for effective one - dimensional potentials from eq .
, when the field is applied along the @xmath37 direction ( along which the weyl nodes are separated in the absence of magnetic field ) .
let us first consider only the gaussian impurity potential , for which the effective one - dimensional potential in a single wsm is given by @xmath194 where @xmath195 .
however , with additional branches within the subspace of zlls , such as in double wsms , supporting two zlls , there are the following two additional components for the effective potential @xmath196 which respectively describes scattering potential within the @xmath197 zll and that between @xmath198 and @xmath197 zlls .
the above formalism can immediately be generalized for triple wsms , supporting three copies of zll . a comparative plot of longitudinal magnetoconductivity ( lmc ) in single ( black ) , double ( red ) and triple ( blue ) weyl semimetals in the presence of only gaussian impurities .
the field is applied along the axis separating the weyl nodes in pristine system ( the @xmath37 direction ) . in the shaded region landau level
are not sharp and one needs to account for weak localization effects . in this region
the semiclassical theory for magnetotransport may also be applicable . here , @xmath199 , and for discussion see eq . ( [ eq : lmc_singleweyl ] ) .
] similarly the effective one - dimensional potential arising from coulomb impurities is given by @xmath200 .
in single wsm @xmath201 where @xmath202 and @xmath203 is the _ upper incomplete gamma function_. for double wsms , the two additional components of the one - dimensional potential are @xmath204 , \notag\\ |u_{0,1}^{(c)}(q_z)|^2 & = \dfrac{u_c^2}{8\pi}\left[-e^{y}(y+1)\gamma(-1,y)+y^{-1}\right ] , \end{aligned}\ ] ] where the subscript notations are the same as that for gaussian impurities .
note that in the extreme strong field limit , i.e. , @xmath205 , all three effective potentials have an identical asymptotic form @xmath206 . for an arbitrary orientation of the magnetic field
, we extract the effective one - dimensional potentials numerically .
next we discuss the lmc in single , double and triple wsm separately . to set the stage ,
we begin with the discussion of lmt in single wsm . in the presence of only gaussian impurities ,
the scattering lifetime in single wsm is given by @xmath207 and the corresponding lmc is @xmath208 where @xmath199 and @xmath209 .
we have used the fact that the fermi wave vector for backscattering in a single wsm is @xmath210 .
therefore , lmc in single wsm , subject to only gaussian impurities , increases linearly with the magnetic field , as shown in fig .
[ fig : conductivity - singleweyl ] ( black curve ) .
the scattering lifetime due to coulomb impurities in a single wsm is given by @xmath211 where @xmath212/2 $ ] .
therefore , the total scattering lifetime ( @xmath42 ) in the presence of both gaussian and coulomb impurities is given by @xmath213 following the _
matthiessen s rule_. thus the total lmc in a single wsm reads @xmath214^{-1 } , \end{aligned}\ ] ] where @xmath215 .
notice that in the presence of only coulomb impurities the lmc grows as @xmath48 in the extremely strong field limit .
however , in the strong magnetic field limit gaussian impurities dominate the lmt ( since typically @xmath216 ) and from here onward we only take into account the scattering lifetime arising from gaussian impurities . a plot of longitudinal magnetoconductivity ( lmc ) in double weyl semimetal for various orientations of the applied magnetic field ( captured by the parameter @xmath41 ) .
we here only account for gaussian impurities . in the shaded region
weak localization effect dominates the lmc , where our analysis is qualitatively inapplicable . here , @xmath199 , and for discussion
see eq . ( [ eq : lmc_singleweyl ] ) . ] for a double wsm _ at half filling _ , the coupled equations in eq .
can be simplified since the fermi velocities of the two zlls are exactly the same by symmetry . as a result
the two coupled equations for the transport lifetime can be rewritten as @xmath217 + \tau_{l } \left[\mathcal{i}_{rl}^{(+)}-\mathcal{i}_{rl}^{(-)}\right ] , \notag\\ & = \tau_{l}\left[2\mathcal{i}_{ll}+\mathcal{i}_{lr}^{(-)}+\mathcal{i}_{lr}^{(+)}\right ] + \tau_{r } \left[\mathcal{i}_{lr}^{(+)}-\mathcal{i}_{lr}^{(-)}\right],\end{aligned}\ ] ] where @xmath218 represents the same fermi velocity for the two zlls , which we label @xmath219 and @xmath148 , as shown in fig .
[ ll_dw_general ] .
the backscattering potentials can be written as @xmath220 where @xmath221 is defined in eq .
( [ eq : effectivepotential ] ) and by symmetry @xmath222 . therefore , one of the scattering lifetimes is @xmath223 and @xmath224 is obtained after taking @xmath225 in the above expression .
since the fermi velocities of the two zlls are identical at half - filling , the total transport lifetime is @xmath226 and the lmc assumes the form @xmath227 where @xmath228 is a dimensionless quantity , which for @xmath160 reads as @xmath229}{2\lambda + 4\lambda^2(\lambda + 1)+1+e^{4\pi^2t^2}(1 + 2\lambda + 2\lambda^2)},\notag\\ & \simeq \dfrac{e^{4\pi^2 t^2}}{2 } \begin{cases } 1 + 4\lambda & \lambda \ll 1 , \\ 2 + 4\lambda & \lambda \gg 1 .
\end{cases}\end{aligned}\ ] ] we display @xmath230 for @xmath160 in fig .
[ fig : conductivity - singleweyl ] ( red curve ) .
thus for zero - range ( @xmath231 ) impurities the lmc is constant , whereas for gaussian impurities it increases monotonically and linearly with magnetic field @xmath3 , thus giving rise to _ positive _ lmc .
also notice that in the limit @xmath232 ( corresponding to fat gaussian impurities ) or @xmath233 ( corresponding to strong magnetic field ) the lmc for a double wsm is exactly twice that of a single - weyl semimetal [ see eq . and fig .
[ fig : conductivity - singleweyl ] ] , since the scattering between two landau levels with different indices
l " and r " will be suppressed in comparison to that within the same ll .
thus in the strong field limit ( @xmath234 ) or fat gaussian impurity ( @xmath232 ) limit the two zlls become effectively decoupled and conduct independently . for general orientations of the magnetic field we can not obtain the lmc analytically and we have to pursue a numerical approach .
the results are shown in fig .
[ lmc_dws_1 ] .
it is interesting to notice that when @xmath130 , lmc develops _ nonlinear _ dependence on the field , and only at sufficiently strong fields the @xmath3-linear lmc is recovered
. such non - trivial dependence on magnetic field and tilted angle should be experimentally observable in transport measurements . for lmc
in double wsm in the presence of only coulomb or ionic impurities see appendix [ lmt_coulomb_append ] .
finally we discuss lmc in triple wsm by restricting ourselves to gaussian impurities . unlike double wsm , the fermi velocities of the three zlls ( at and near half filling ) are no longer equal : in general we have @xmath235 , where `` l '' , `` c '' , and `` r '' are the labels for the three branches of zll , shown in fig .
[ ll_tw_general ] .
the lmc then assumes the following form , @xmath236 \notag\\ & = \sigma_0\left\{\dfrac{v_l^2(b)}{v_z^2}\left[\tilde{\tau}_l+\tilde{\tau}_r + w^{-1}\tilde{\tau}_c\right]\right\ } , \label{lmc_twsm } \end{aligned}\ ] ] where @xmath237 and @xmath238 is a dimensionless quantity .
the lmc in a triple wsm can be obtained analytically when the magnetic field is applied along the @xmath37 direction ( @xmath160 ) , see fig .
[ fig : conductivity - singleweyl ] ( blue curve ) .
notice all three fermi velocities are then equal ( @xmath239 ) .
for sufficiently strong magnetic fields ( @xmath234 ) or fat gaussian impurity potential ( @xmath232 ) , three zlls again behave as decoupled one - dimensional wires and the lmc in a triple wsm is three times that of a single wsm ; compare black and blue curves in fig .
[ fig : conductivity - singleweyl ] . a plot of longitudinal magnetoconductivity ( lmc ) in triple weyl semimetal for various orientations of the applied magnetic field ( captured by the parameter @xmath41 ) .
we here only account for gaussian impurities . in the shaded region
weak localization effect dominates the lmc , where our analysis is qualitatively inapplicable .
here , @xmath199 , and for discussion see eq . ( [ eq : lmc_singleweyl ] ) . ] for arbitrary orientations of the magnetic field , we compute various scattering elements , shown in eq .
( [ lmc_twsm ] ) , numerically .
in particular , notice that three fermi vectors for backscattering have the magnitudes of @xmath240 , and @xmath241 , respectively , where @xmath242 is again a function of the direction and strength of the applied magnetic field .
the resulting lmc in triple wsm is shown in fig .
[ fig : conductivity-2 ] . similar to the situation in double wsm ,
the lmc in triple wsm also exhibits an anisotropic angular dependence on the field orientations .
in particular , as the tilting angle @xmath41 increases lmc displays a smooth crossover from @xmath3-linear to _ nonlinear _ in @xmath3 dependence , but lmc always remains positive .
[ generalanomaly ] so far we discussed the lmt in general wsms when the system is at precisely half - filling and consequently only the subspace of zlls is partially filled .
however , our conclusions remain qualitatively valid even when ( i ) the system is away from half - filling , ( ii ) higher lls are also partially filled .
let us first focus on the former situation .
when the system is placed away from half - filling with increasing strength of the magnetic field , as the dos of lls increases the chemical potential continuously shifts toward the half - filling value ( since carrier density is fixed ) and at infinite field limit the chemical potential gets locked at band center or zero energy .
hence , our theoretical prediction regarding the fate of lmc and its scaling with the magnetic field at half - filling is achieved for a strong enough field , but the outcomes remain qualitatively operative , as long as only the subspace of zlls is partially field .
when multiple lls are partially filled ( including the zeroth ones ) , any wsm supports positive lmc , since its origin relies on the existence of partially filled lls . as a generic property of any three - dimensional metallic system , since for each partially filled lls there exists at least a pair of emergent one - dimensional weyl points ( characterized by one - dimensional chiral fermions ) . with the emergence of such weyl nodes
our analysis of lmt remains operative and the system continues to support positive lmc .
however , with increasing number of partially filled lls , the number of backscattering channels increases rapidly and the calculation of relaxation lifetime gets involved .
nonetheless , one can borrow our analysis ( at least qualitatively ) for multi wsms ( @xmath10 ) , in particular when the field is tilted away from the @xmath37 direction , for which we have established generic positive lmc .
hence , we believe that positive lmc in general wsm should be a generic situation when the system is in the quantum limit ( @xmath43 ) , irrespective of the location of the chemical potential .
the magnetotransport results presented in this section are valid only when the wsm remains in its normal ` fermi liquid ' phase , i.e. , there is no symmetry - breaking transition induced by interactions .
any quantitative interaction effects can be included in the theory simply by reinterpreting each parameter as the corresponding renormalized quantity within a fermi liquid phase . in particular
, the results from this section would not apply in the presence of any field - induced density wave ordering , which might happen due of electronic interactions .
next we will discuss the effect of various density - wave ordering within the manifold of zlls and its effects on lmt .
our conclusions regarding the effects of density - wave ordering on lmt can easily be generalized when the chemical potential placed deep inside the valence or conduction like lls , so that multiple lls ( including the zeroth one ) are partially filled .
[ densitywave ] the dos in a wsm , constituted by weyl nodes with monopole charges @xmath81 , vanishes as @xmath51 as one approaches the weyl points . here
the energy @xmath16 is measured from the weyl points .
therefore , any weak electron - electron interaction ( restricting ourselves to hubbard - like local interactions , as appropriate for tight - binding lattice models ) is an _ irrelevant _ perturbation around the noninteracting gaussian fixed point , and weyl semimetals describe an infrared stable fixed point in the language of renormalization group .
however , a strong magnetic field quenches the quasiparticle dispersion into a set of discrete lls , all of them dispersing along the applied magnetic field .
the dos for such effective one - dimensional systems is _ constant _ , significantly enhancing the effects of electron - electron interaction .
therefore , sufficiently strong magnetic fields can hybridize the emergent weyl nodes and gap them out through the formation of a density - wave order that breaks the _
translational symmetry_. such a mechanism of developing density - wave order remains operative in any three - dimensional interacting electronic system in the presence of a strong external magnetic field @xcite . in the continuum limit the cdw or sdw order breaks an emergent _ continuous _
@xmath243 translational symmetry @xcite , and disorder can couple to density - wave order as _
random field_. the imry - ma argument then prohibits such an order from acquiring a true long - range order @xcite .
however , in the presence of an underlying lattice , the density - wave orders only break _ discrete _ translational symmetry .
hence , our proposed density - wave order ( cdw or sdw ) can exhibit a true long range ordering and may be observable at moderately strong magnetic fields . for the sake of simplicity , we here assume that only the manifold of zlls are partially filled , and concomitantly , density - wave ordering develops within this manifold .
the following discussion can easily be generalized when multiple lls are partially filled , but the quantitative details will be different with the density wave ordering trend being suppressed as more lls participate in the energetics .
in addition , the actual nature of the density - wave ordering ( cdw or sdw ) within zlls depends on the microscopic details of the weyl system , as we discuss below . to capture the low energy physics in such wsms
we appropriately define a four component spinor as @xmath244 $ ] , where @xmath245 are two component spinors for left ( @xmath148 ) and right ( @xmath219 ) chiral fermions , organized as @xmath246 $ ] for @xmath247 . in this notation ,
weyl nodes are located at @xmath248 and @xmath249 are the kramers partners or two spin projections .
let us first focus on the low energy hamiltonian for wsms , shown in eq .
( [ hamillowenergy ] ) , which can be written compactly as @xmath250 \alpha_n k^{n}_\perp + \tau_3 \otimes \sigma_3 v_z k_z , \label{swm_1 } \end{aligned}\ ] ] where @xmath251 . for such weyl systems
@xmath252 , hence the weyl nodes are along the @xmath37 direction .
notice that two distinct types of density - wave orders , namely the cdw and sdw , can gap out the weyl nodes at the cost of the translational symmetry .
respectively the effective single - particle hamiltonian in the presence of these two orderings are @xmath253 in the continuum description @xmath254 is a continuous variable , and these two ordered phases break continuous @xmath243 chiral symmetry generated by @xmath255 , representing the generator of translational symmetry @xcite , and @xmath256=0 $ ] , whereas @xmath257 for @xmath258 and @xmath259 .
but , the underlying lattice potential lifts such continuous symmetry and prefers certain locking angle for the order parameter .
therefore , the ordered phase is not accompanied by true massless goldstone mode , similar to the situation with underlying valence bond solid or superconductor in two - dimensional graphene @xcite .
even though both cdw and sdw orders gap out the zlls , a subtle competition between these two orders may ultimately determine the actual pattern of the symmetry breaking .
notice that @xmath260 and the sdw order stands as a chiral symmetry breaking _ mass _ for weyl fermions .
thus , when sdw order sets in , besides gapping out the zlls , it also pushes down all filled lls , following the spirit of _ magnetic catalysis _ @xcite .
hence , the sdw order appears to be energetically favored over cdw , when weyl nodes are located along one of the @xmath59 axes of a cubic system , for example @xmath78 .
it is also conceivable to realize the weyl semimetals for which the low energy hamiltonian is given by @xmath261 + \sigma_3 v_z k_z \right ] , \label{swm_2}\end{aligned}\ ] ] when the weyl nodes are located along ( 1,1,1 ) direction in a cubic system , for example , and @xmath262 .
once again both cdw and sdw orders can generate a spectral gap within the manifold of zlls . however ,
the role of the cdw and sdw orders is now reversed ; cdw representing a chiral symmetry breaking mass .
hence , in a wsm for which the weyl nodes are placed along one of the @xmath263 axes of a cubic system and the low energy hamiltonian assumes the form in eq .
( [ swm_2 ] ) , the cdw order is energetically favored over the sdw one .
thus the nature of the density - wave ordering crucially depends on the microscopic details of the system .
however , the signature of such translation symmetry breaking ordering on lmt is qualitatively insensitive to its actual nature since in both cases spectral gaps open up within the zlls . as shown in the previous section , due to
partially filled subspace of zlls , all wsms manifest positive lmc or negative lmr , when these systems are placed in a strong magnetic field . on the other hand , electronic interaction can be conducive for translational symmetry breaking cdw or sdw order that produces a spectral gap within the zll . with the onset of such insulating ordering
there is no partially filled ll in the system and lmr ( lmc ) can now display an upturn ( downturn ) and become positive ( negative ) in a strong magnetic field .
although such an instability occurs for an infinitesimal strength of interaction in a clean system for @xmath264 , with @xmath52 being the transition temperature that scales as @xmath265 $ ] , where @xmath266 is the strength of interaction responsible for density - wave ordering and @xmath267 is the dos of zlls ( following the bcs scaling law ) , the presence of disorder may hinder its onset .
this can be seen most easily by realizing that the dos @xmath267 is suppressed by impurity scattering ( e.g. dingle temperature effect ) leading to a strong suppression of @xmath52 @xcite .
thus only for sufficiently strong magnetic fields and at low temperature these orderings can set in and the system can display an upturn in lmr . when the magnetic field is applied along @xmath37 direction ,
all zlls are degenerate in double and triple wsm ( in a single wsm there is always only one zll ) ; hence they simultaneously undergo manybody instability toward the formation of a density - wave order as temperature is gradually lowered below @xmath52 .
concomitantly , all wsms display negative lmc or positive lmr above a unique strength of magnetic field @xmath268 , when the field is applied along the @xmath37 direction .
the periodicity of such density - wave ordering is @xmath269 .
however , the situation changes dramatically when the field is tilted away from the @xmath37 direction in double and triple wsms . as shown in sec .
[ landaulevel ] , when the magnetic field is tilted away from the @xmath37 direction , the exact degeneracy among the zlls in double and triple wsm is lifted . as a result , the density - wave ordering in each zll is characterized by different transition temperatures , even though the degeneracy of the zll does not change with the field orientation , which can be qualitatively understood in the following way . for the sake of simplicity
we can assume that short - range interactions arise from its long range coulomb tail , which in three spatial dimensions scales as @xmath270 .
then it is natural to assume that electronic interaction is stronger for density - wave ordering with smaller wave vectors since @xmath271 increases for smaller @xmath272 .
therefore , in double wsms , as the temperature ( magnetic field ) is gradually lowered ( increased ) the fermi points marked by
r " get gapped out first , followed by the ones marked by
l " [ see fig .
[ ll_dw_general ] ] .
the periodicity of density - wave ordering in the r " and l " channels is respectively @xmath273/a$ ] where @xmath274 depends on the orientation of the magnetic field .
similarly , in triple wsm , the mass generation first takes place among the r " points , followed by the c " points and finally the l " points , see fig .
[ ll_tw_general ] .
the periodicity of density - wave ordering in these three channels are @xmath275/a$ ] , for @xmath142 respectively for r " , c " and l " channels .
thus as the temperature is gradually lowered in double and triple wsms ( with field tilted away from the @xmath37 direction ) , they undergo a cascade of transitions into density - wave phases with different periodicity , and below each such transition temperature the lmc displays _ discontinuity _ and its rate of growth decreases , before it becomes negative when all zlls are gapped out , as shown in fig .
[ lmt - density - wave ] ( b ) and ( c ) . in other words , in single wsm the lmr becomes positive once the density - wave ordering sets in , while in double ( triple ) weyl semimetal lmr should display a two ( three ) stage transition toward a positive value .
the separation among the transition temperature or onset magnetic field ( @xmath276 ) among various density - wave transitions ( hence the distance among various discontinuities in lmc / lmr ) gets bigger as the magnetic field is gradually tilted away from the @xmath37 direction , since the emergent one - dimensional fermi points gets further separated with increasing tilting of the magnetic field away from the axis separating two weyl nodes [ see figs .
[ ll_dw_general ] , [ ll_tw_general ] ] .
such steps in lmr can possibly be detected in experiments and may serve as an indication for density - wave ordering in the system . as a final remark , we would like to point out that recent magnetotransport measurements in graphite ( for band structure see refs .
@xcite ) is strongly suggestive of a density - wave ordering in strong magnetic fields @xcite .
[ summary ] to summarize , we here address several important aspects of weyl semimetals when these systems are placed in a strong magnetic field . in the absence of magnetic fields , three - dimensional weyl semimetals are constituted by weyl nodes that can be classified according to the monopole charge @xmath6 , @xmath277 and @xmath2 . the integer monopole charge ( @xmath0 )
determines the topological invariant of a weyl semimetal , and the surface fermi arc in such systems possesses an additional @xmath0-fold _ orbital _ degeneracy .
when placed in a magnetic field , the underlying topological invariant manifests itself through the number of chiral zeroth landau levels . thus double and triple weyl semimetals respectively support two and three copies of the zeroth landau level while a regular or single weyl semimetal is devoid of any flavor degeneracy when the magnetic field is applied in the @xmath37 direction , along which the weyl nodes are separated in the momentum space .
the same conclusion also follows from a calculation based on a lattice model , as shown in appendix [ landaulevel_append ] . within the framework of the continuum description of these system
, we always find @xmath0 zeroth landau levels , irrespective of the direction of the applied magnetic field , although their exact degeneracy breaks down when the field is tilted away from the @xmath37 direction .
also the density of states in the presence of landau levels displays a strong angular dependence , which can be observed in angle resolved quantum oscillation experiments .
we also show that the formation of landau levels in weyl materials can generically lead to longitudinal magnetotransport , since they are dispersive in the direction of the applied magnetic field . to obtain a steady state magnetoconductivity
we invoke the relaxation mechanism arising from the backscattering by impurities and take into account two sources of elastic scattering ( a ) gaussian impurity and ( b ) coulomb impurity .
however , due to finite density of states , the scattering potential in the effective one - dimensional world is always short ranged , and we mainly focus on the former source of elastic scattering since the coulomb disorder is strongly screened by the carriers . assuming that ( i ) the system is in the quantum limit ( @xmath278 ) , and ( ii ) only the subspace of zeroth landau level is partially filled , we calculate the magnetic field dependence of transport lifetime in the presence of strong backscattering using the quantum boltzmann equation and show that the longitudinal magnetoconductivity increases linearly with the applied magnetic field , when the field is applied along the separation of weyl nodes in the pristine system . within the low - energy approximation , we find that such linear-@xmath3 dependence of the longitudinal magnetoconductance in double and triple weyl semimetals smoothly crosses over to a nonlinear @xmath3 dependence when the field is tilted away from the @xmath37 direction , which can be tested experimentally .
the longitudinal magnetoconductivity in all weyl semimetals scales as @xmath48 when the magnetic field is applied along the separation of two pristine weyl nodes , in the presence of coulomb / ionic impurities , but in the extremely strong field limit .
in order to verify our predictions of lmc in systems dominant by gaussian ( short - ranged ) and coulomb ( long - ranged ) impurities , as well as the crossover between them , one can systematically introduce short - ranged or long - ranged impurities in a given sample , and vary their concentrations in a controlled way . for example , long - range disordered potentials can be generated by dopant charged impurities , while short - range disordered potentials can be introduced by radiation damage .
such an experimental technique provided the key evidence in finding out the relative importance of long - range versus short - range disorder potentials in the transport properties of graphene @xcite .
therefore , a similar technique could in principle be used in wsms to test our predictions with respect to longitudinal magnetotransport .
finally , we also address the effects of electron - electron interaction on longitudinal magnetotransport in weyl semimetals . due to the finite density of states of effective one - dimensional system , weyl fermions are susceptible to various types of translational symmetry breaking density - wave ordering ( charge or spin depending on microscopic details ) even for infinitesimal strength of interaction in the presence of strong external magnetic fields . since a single weyl semimetal hosts one zeroth landau level , the onset of a density wave order immediately gives rise to positive longitudinal magnetoresistance .
such upturn in lmr can be observed below the transition temperature for sufficiently strong magnetic fields .
a similar situation arises in double and triple weyl semimetals when the field is applied along the @xmath37 direction , since all zeroth landau levels are exactly degenerate .
however , the situation changes dramatically as the applied field is gradually tilted away from the @xmath37 direction . since , the two / three zeroth landau levels in double / triple weyl semimetals are no longer degenerate , they respectively undergo two and three transitions into density - wave phases , characterized by different periodicity as well as transition temperature .
thus , as the temperature is gradually lowered in the presence of a strong magnetic field ( applied at a finite angle with respect to the @xmath37 direction ) , the decrease of the longitudinal magnetoresistance with magnetic field gets slower across each such transition , before it finally becomes positive when all members of the zeroth landau level are gapped out by density - wave orderings . thus future experiments in various members of weyl semimetals , when placed in a magnetic field , can display intriguing confluence of landau quantization , longitudinal magnetotransport ( possibly manifesting one - dimensional chiral or axial anomaly ) , and interaction driven charge / spin - density - wave ordering .
this work was supported by jqi - nsf - pfc and lps - mpo - cmtc .
we thank pallab goswami , m. zahid hasan for discussion .
b. r. is thankful to nordita , center for quantum materials for hospitality where part of this work was finalized .
[ landaulevel_append ] in this appendix we show the derivation of ll spectrum for single , double and triple wsm from a lattice model ( bounded dispersion ) , when the magnetic field is applied along the separation of weyl nodes , i.e. along the @xmath37 direction .
let us begin with the following model for a general wsm @xmath279\sigma_z \notag\\ & + \begin{pmatrix } 0 & \alpha_n(\hbar k_x - i\hbar k_y)^n \\
\alpha_n(\hbar k_x + i\hbar k_y)^n & 0 \end{pmatrix},\end{aligned}\ ] ] where @xmath0 is the monopole charge .
if the applied magnetic field is along the @xmath37 direction , the landau level spectrum of this model can be expressed in a closed form .
for example , for a single wsm @xmath280 we have for @xmath281 @xmath282 ^ 2+\dfrac{2\alpha_1 ^ 2\hbar^2}{\ell_b^2}n},\end{aligned}\ ] ] where @xmath283 , and @xmath22 is the landau level index .
the energy dispersion for double @xmath284 and triple @xmath285 wsm can be similarly obtained as follows , @xmath286 ^ 2 + \dfrac{4\hbar^4\alpha_2 ^ 2}{\ell^4 } n(n-1)\right]^{\frac{1}{2 } } , \notag\\ & = \pm \left[\left[\epsilon_z+\dfrac{\hbar^2}{2m\ell_b^2}(2n+1)\right]^2 + \dfrac{8 \hbar^6\alpha_3 ^ 2}{\ell^6 } n(n-1)(n-2)\right]^{\frac{1}{2}},\notag\end{aligned}\ ] ] which is valid for @xmath287 and @xmath288 , respectively .
the resulting ll spectrum for single wsm has already been displayed in fig .
[ fig : spectrum - lllattice ] .
the energy of the zll ( respectively one , two and three fold degenerate in single , double and triple wsm ) is @xmath289 and throughout we measured the energy from the reference value or zero point energy @xmath290 . ) .
we here only account for coulomb impurities . in the shaded region
weak localization effect dominates the lmc , where our analysis is qualitatively inapplicable.[fig : lmc_coulomb ] ]
[ lmt_coulomb_append ] in this appendix we want to briefly discuss the lmc in double wsms only in the presence of coulomb impurities .
the results are displayed in fig .
[ fig : lmc_coulomb ] for various field orientations .
we note that typically the transport lifetime due to coulomb impurities are much longer than that for gaussian impurities . thus according to the matthiessen s rule , shown in eq . , the transport lifetime is dominated by gaussian impurities . however , in the complete absence of gaussian impurities , one can expect to observe the lmc shown in fig .
[ fig : lmc_coulomb ] .
we also want note that lmc driven by coulomb impurities will show a @xmath48 dependence only in the limit of @xmath291 .
we also find similar field dependence of lmc in the presence of only coulomb impurities in single as well as triple wsms .
huang , s .- y .
xu , i. belopolski , c .- c .
lee , g. chang , b. wang , n. alidoust , g. bian , m. neupane , a. bansil , h. lin , and m. z. hasan , _ an inversion breaking weyl semimetal state in the taas material class _ , arxiv:1501.00755 ( 2015 ) .
xu , i. belopolski , n. alidoust , m. neupane , c. zhang , r. sankar , s .-
huang , c .- c .
lee , g. chang , b. wang , g. bian , h. zheng , d. s. sanchez , f. chou , h. lin , s. jia , and m. z. hasan , _ experimental realization of a topological weyl semimetal phase with fermi arc surface states in taas _ , arxiv:1502.03807 ( 2015 ) .
xu , i. belopolski , n. alidoust , m. neupane , g. bian , c. zhang , r. sankar , g. chang , z. yuan , c .- c .
huang , h. zheng , j. ma , d. s. sanchez , b. wang , a. bansil , f. chou , p. p. shibayev , h. lin , s. jia , and m. z. hasan , _ discovery of a weyl fermion semimetal and topological fermi arcs _
, science * 349 * , 613 ( 2015 ) .
b. q. lv , h. m. weng , b. b. fu , x. p. wang , h. miao , j. ma , p. richard , x. c. huang , l. x. zhao , g. f. chen , z. fang , x. dai , t. qian , and h. ding , _ experimental discovery of weyl semimetal taas _
x * 5 * , 031013 ( 2015 ) .
xu , n. alidoust , i. belopolski , z. yuan , g. bian , t .-
chang , h. zheng , v. n. strocov , d. s. sanchez , g. chang , c. zhang , d. mou , y. wu , l. huang , c .- c .
huang , b. wang , a. bansil , h .-
jeng , t. neupert , a. kaminski , h. lin , s. jia , and m. zahid hasan , _ discovery of aweyl fermion state with fermi arcs in niobium arsenide _ , nat . phys . *
11 * , 748 ( 2015 ) .
n. xu , h. m. weng , b. q. lv , c. matt , j. park , f. bisti , v. n. strocov , d. gawryluk , e. pomjakushina , k. con- der , n. c. plumb , m. radovic , g. auts , o. v. yazyev , z. fang , x. dai , g. aeppli , t. qian , j. mesot , h. ding , m. shi , _ observation of weyl nodes and fermi arcs in tap _ , arxiv:1507.03983 ( 2015 ) .
h. inoue , a. gyenis , z. wang , j. li , s. w. oh , s. jiang , n. ni , b. a. bernevig , a. yazdani , _ quasiparticle interference of the fermi arcs and surface - bulk connectivity of a weyl semimetal _ , science * 351 * , 1184 ( 2016 ) . c. shekhar , a. k. nayak , y. sun , m. schmidt , m. nicklas , i. leermakers , u. zeitler , z. liu , y. chen , w. schnelle , j. grin , c. felser , b. yan , _ extremely large magnetoresistance and ultrahigh mobility in the topological weyl semimetal nbp _ , arxiv:1502.04361 ( 2015 ) .
g. chang , s - y .
xu , d. s. sanchez , s - m .
huang , c - c .
lee , t - r .
chang , h. zheng , g. bian , i. belopolski , n. alidoust , h - t .
jeng , a. bansil , h. lin , m. z. hasan , _ a strongly robust weyl fermion semimetal state in ta@xmath292s@xmath293 _ , arxiv:1512.08781 ( 2015 ) .
j. y. liu , j. hu , q. zhang , d. graf , h. b. cao , s. m. a. radmanesh , d. j. adams , y. l. zhu , g. f. cheng , x. liu , w. a. phelan , j. wei , d. a. tennant , j. f. ditusa , i. chiorescu , l. spinu , z.q .
mao , _ discovery of a topological semimetal phase coexisting with ferromagnetic behavior in sr@xmath294mn@xmath295sb@xmath8 _ , arxiv:1507.07978 ( 2015 ) .
q. d. gibson , l. m. schoop , l. muechler , l. s. xie , m. hirschberger , n. p. ong , r. car , and r. j. cava , _
3d dirac semimetals : current materials , design principles and predictions of new materials _ , arxiv:1411.0005 ( 2014 ) .
x. wan , a. turner , a. vishwanath , and s. y. savrasov , _ topological semimetal and fermi - arc surface states in the electronic structure of pyrochlore iridates _ , phys .
b * 83 * , 205101 ( 2011 ) .
moon , c. xu , y. b. kim , l. balents , _ non - fermi liquid and topological states with strong spin - orbit coupling _ , phys .
* 111 * , 206401 ( 2013 ) .
huang , s - y .
xu , i. belopolski , c - c .
lee , g. chang , b. wang , n. alidoust , m. neupane , h. zheng , d. sanchez , a. bansil , g. bian , h. lin , m. z. hasan , _ new type of weyl semimetal with quadratic double weyl fermions _ ,
pnas * 113 * , 1180 ( 2016 ) . m. e. peskin , and d. v. schroeder , _ an introduction to quantum field theory _ , ( addison - wesley , 1995 ) .
k. fujikawa , h. suzuki , _ path integrals and quantum anomalies _ , ( clarendon press , 2004 ) .
s. adler , _ axial - vector vertex in spinor electrodynamics _ ,
phys . rev . *
177 * , 2426 ( 1969 ) .
j. s. bell and r. a. jackiw , _ a pcac puzzle : @xmath297 in the @xmath298-model _ , nuovo cimento a * 60 * , 47 ( 1969 ) .
v. aji , _ adler - bell - jackiw anomaly in weyl semi - metals : application to pyrochlore iridates _ ,
b * 85 * , 241101 ( 2012 ) .
d. t. son , and b. z. spivak , _ chiral anomaly and classical negative magnetoresistance of weyl metals _ , phys . rev .
b * 88 * , 104412 ( 2013 ) .
s. a. parameswaran , t. grover , d. a. abanin , d. a. pesin , and a. vishwanath , _ probing the chiral anomaly with nonlocal transport in three - dimensional topological semimetals _ , phys .
x , * 4 * , 031035 ( 2014 ) .
e. v. gorbar , v. a. miransky , and i. a. shovkovy , _ chiral anomaly , dimensional reduction , and magnetoresistivity of weyl and dirac semimetals _ , phys .
b * 89 * , 085126 ( 2014 ) .
burkov , a. a. _ chiral anomaly and diffusive magnetotransport in weyl metals _ , phys .
lett . * 113 * , 247203 ( 2014 ) .
t. osada , _ negative interlayer magnetoresistance and zero - mode landau level in multilayer dirac electron systems _ , j. phys .
jpn . * 77 * , 084711 ( 2008 ) .
a. a. zyuzin , and a. a. burkov , _ topological response in weyl semimetals and the chiral anomaly _ ,
b * 86 * , 115133 ( 2012 ) .
a. g. grushin , _ consequences of a condensed matter realization of lorentz - violating qed in weyl semi - metals _ , phys .
d * 86 * , 045001 ( 2012 ) .
p. goswami , and s. tewari , _ axionic field theory of ( 3 + 1)-dimensional weyl semimetals _ , phys .
b * 88 * , 245107 ( 2013 ) .
k. fukushima , d. e. kharzeev , h. j. warringa , _ the chiral magnetic effect _
d * 78 * , 074033 ( 2008 ) .
m. m. vazifeh , m. franz , _
electromagnetic response of weyl semimetals _ , phys .
lett . * 111 * , 027201 ( 2013 ) .
y. chen , si wu , a. a. burkov , _ axion response in weyl semimetals _ , phys .
b * 88 * , 125105 ( 2013 )
. g. xu , h. weng , z. wang , x. dai , and z. fang , _ chern semimetal and the quantized anomalous hall effect in hgcr@xmath8se@xmath9 _ , phys .
107 * , 186806 ( 2011 ) .
g. y. cho , _ possible topological phases of bulk magnetically doped bi@xmath8se@xmath299 : turning a topological band insulator into the weyl semimetal _
, arxiv:1110.1939 ( 2011 ) .
kim , k .- s .
kim , j. f. wang , m. sasaki , n. satoh , a. ohnishi , m. kitaura , m. yang , and l. li , _ dirac versus weyl fermions in topological insulators : adler - bell - jackiw anomaly in transport phenomena _ , phys .
* 111 * , 246603 ( 2013 ) .
q. li , d. e. kharzeev , c. zhang , y. huang , i. pletikosic , a. v. fedorov , r. d. zhong , j. a. schneeloch , g. d. gu , t. valla , _ observation of the chiral magnetic effect in zrte@xmath300 _ , arxiv:1412.6543 ( 2014 ) .
n. tajima , s. sugawara , r. kato , y. nishio , and k. kajita , _ effects of the zero - mode landau level on inter - layer magnetoresistance in multilayer massless dirac fermion systems _ , phys .
* 102 * , 176403 ( 2009 ) .
n. kikugawa , p. goswami , a. kiswandhi , e. s. choi , d. graf , r. e. baumbach , j. s. brooks , k. sugii , y. iida , m. nishio , s. uji , t. terashima , p. m. c. rourke , n. e. hussey , h. takatsu , s. yonezawa , y. maeno , and l. balicas , _ interplanar coupling - dependent magnetoresistivity in high - purity layered metals _ , nature commun .
* 7 * , 10903 ( 2016 ) . c. shekhar , f. arnold , s .- c .
wu , y. sun , m. schmidt , n. kumar , a. g. grushin , j. h. bardarson , r. d. dos reis , m. naumann , m. baenitz , h. borrmann , m. nicklas , e. hassinger , c. felser , and b. yan , _ large and unsaturated negative magnetoresistance induced by the chiral anomaly in the weyl semimetal tap _
, arxiv:1506.06577 ( 2015 ) .
h. wang , c .- k .
li , h. liu , j. yan , j. wang , j. liu , z. lin , y. li , y. wang , l. li , d. mandrus , x. c. xie , j. feng , and j. wang , _ chiral anomaly and ultrahigh mobility in crystalline hfte@xmath300 _ phys .
b * 93 * , 165127 ( 2016 ) .
g. zheng , j. lu , x. zhu , w. ning , y. han , h. zhang , j. zhang , c. xi , j. yang , h. du , k. yang , y. zhang , and m. tian , _ transport evidence for the three - dimensional dirac semimetal phase in zrte@xmath300 _ , phys .
b * 93 * , 115414 ( 2016 ) .
y. zhao , h. liu , c. zhang , h. wang , j. wang , z. lin , y. xing , h. lu , j. liu , y. wang , s. jia , x. c. xie , and j. wang , _ anisotropic fermi surface and quantum limit transport in high mobility 3d dirac semimetal cd@xmath299as@xmath8 _ , arxiv:1412.0330 ( 2014 ) .
j. xiong , s. k. kushwaha , t. liang , j. w. krizan , w. wang , r. j. cava , and n. p. ong , _ signature of the chiral anomaly in a dirac semimetal : a current plume steered by a magnetic field _ , arxiv:1503.08179 ( 2015 ) .
t. liang , q. d. gibson , m. n. ali , m. liu , r. j. cava , and n. p. ong , _ ultrahigh mobility and giant magnetoresistance in the dirac semimetal cd@xmath299as@xmath8 _ , nat . mater . * 14 * , 280 ( 2015 ) .
x. huang , l. zhao , y. long , p. wang , d. chen , z. yang , h. liang , m. xue , h. weng , z. fang , x. dai , and g. chen , _ observation of the chiral anomaly induced negative magneto - resistance in 3d weyl semi - metal taas _
x * 5 * , 031023 ( 2015 ) . c. zhang , s .- y .
xu , i. belopolski , z. yuan , z. lin , b. tong , n. alidoust , c .- c .
huang , h. lin , m. neupane , d. s. sanchez , h. zheng , g. bian , j. wang , c. zhang , t. neupert , m. z. hasan , and s. jia , _ observation of the adler - bell - jackiw chiral anomaly in a weyl semimetal _
, arxiv:1503.02630 ( 2015 ) .
s. wiedmann , a. jost , b. fauqu , j. van dijk , m. j. meijer , t. khouri , s. pezzini , s. grauer , s. schreyeck , c. brne , h. buhmann , l. w. molenkamp , and n. e. hussey , _ anisotropic and strong negative magnetoresistance in the three - dimensional topological insulator bi@xmath8se@xmath299 _ , phys .
b * 94 * , 081302(r ) ( 2016 ) .
s. hikami , a. i. larkin , and y. nagaoka , _ spin - orbit interaction and magnetoresistance in the two - dimensional random system _ , prog .
. phys . * 63 * , 707 ( 1980 ) .
g. bergmann , _ weak localization in thin films a time - of - flight experiment with conduction electrons _
, phys . rep . *
107 * , 1 ( 1984 ) .
j. c. slonczewski and p. r. weiss , _ band structure of graphite _ , phys
* 109 * , 272 ( 1958 ) .
j. w. mcclure , _ band structure of graphite and de haas - van alphen effect _ ,
phys . rev . *
108 * , 612 ( 1957 ) .
b. fauqu , d. leboeuf , b. vignolle , m. nardone , c. proust , k. behnia , _ two phase transitions induced by a magnetic field in graphite _ , phys .
lett . * 110 * , 266601 ( 2013 ) .
e. h. hwang , s. adam , and s. das sarma , _ carrier transport in two - dimensional graphene layers _
lett . * 98 * , 186806 ( 2007 ) ; s. adam , e. h. hwang , v. m. galitski , and s. das sarma , _ a self - consistent theory for graphene transport _ , proc .
usa * 104 * , 18392 ( 2007 ) ; y .- w .
tan , y. zhang , k. bolotin , y. zhao , s. adam , e. h. hwang , s. das sarma , h. l. stormer , and p. kim , _ measurement of scattering rate and minimum conductivity in graphene _ ,
lett . * 99 * , 246803 ( 2007 ) ; j .-
h . chen , c. jang , s. adam , m. s. fuhrer , e. d. williams , and m. ishigami , _ charged - impurity scattering in graphene _
, nat . phys . * 4 * , 377 ( 2008 ) . | we theoretically address the effects of strong magnetic fields in three - dimensional weyl semimetals ( wsms ) built out of weyl nodes with a monopole charge @xmath0 . for @xmath1 and @xmath2
we realize single , double , and triple wsm , respectively and the monopole charge @xmath0 determines the integer topological invariant of the weyl systems . within the linearized continuum description , the quasiparticle spectrum
is then composed of landau levels ( lls ) , containing exactly @xmath0 number of chiral zeroth landau levels ( zlls ) , irrespective of the orientation of the magnetic field . in the presence of strong backscattering ( due to impurities )
, these systems generically give rise to longitudinal magnetotransport . restricting ourselves to the quantum limit ( and assuming only the subspace of the zlls is partially filled ) and mainly accounting for gaussian impurities , we show that the longitudinal magnetoconductivity ( lmc ) in all members of the weyl family displays a positive linear-@xmath3 scaling , when the field is applied along the axis that separates the weyl nodes .
but , in double and triple wsm , lmc displays a smooth crossover to a nonlinear @xmath3-dependence as the field is tilted away from such a high symmetry direction .
in addition , due to the enhanced density of states , the ll quantization can trigger instabilities toward the formation of translational symmetry breaking density - wave orderings for sufficiently weak interaction ( bcs instability ) , which gaps out the zlls .
concomitantly as the temperature ( magnetic field ) is gradually decreased ( increased ) the lmc becomes negative .
thus wsms with arbitrary monopole charge ( @xmath0 ) can host an intriguing interplay of ll quantization , longitudinal magnetotransport ( a possible manifestation of one - dimensional chiral or axial anomaly ) , and density - wave ordering , when placed in a strong magnetic field . |
let @xmath0 denote the space of smooth embeddings @xmath1 coinciding with a fixed linear embedding outside a compact subset of @xmath2 .
by @xmath3 we denote the homotopy fiber of the inclusion @xmath4 , where @xmath5 is the space of immersions with the same behavior at infinity . in @xcite arone and turchin study the rational homology and homotopy of @xmath0 and @xmath3 .
in particular in @xcite the authors define graph - complexes @xmath6 computing the rational homotopy @xmath7 , @xmath8 .
the graph - complexes up to a regrading depend only on the parities of @xmath9 and @xmath10 .
thus there are only 4 cases to consider .
the connection between @xmath7 and @xmath11 , @xmath8 , is established by ( * ? ? ?
* corollary 4.3 ) .
the graph - complexes @xmath6 split into a direct sum of complexes by the number of univalent vertices ( this number is called _ hodge degree _ ) and the number of loops ( first betti number ) in the graphs .
this paper computes the homology of the part of @xmath6 spanned by two - loop graphs .
the homology of the part spanned by trees and of the part spanned by one - loop graphs was computed in ( * ? ? ?
* section 3 ) .
theorems [ t_homology_m_odd_n_odd ] , [ t_homology_m_even_n_even ] , [ t_homology_m_even_n_odd ] , [ t_homology_m_odd_n_even ] describe generating functions of the homology ranks and of the euler characteristics in each of the four cases .
theorems [ t_rank_m_odd_n_odd]-[t_rank_m_odd_n_even ] give explicit formulae for those ranks .
we can make two conclusions from our computations .
firstly , for all parities of @xmath9 and @xmath10 , the homology of the two - loop part for any hodge degree is concentrated in only one grading .
thus the euler characteristics completely determine the homology ranks .
secondly , the ranks grow linearly with the hodge degree .
our method is borrowed from similar computations of the dimensions of the space of 2-loop uni - trivalent graphs modulo @xmath12 and @xmath13 relations @xcite , which appears in the study of finite type knot invariants , see also @xcite .
the latter space is the bottom degree part of our homology for the case when both @xmath9 and @xmath10 are odd .
the field of coefficients for the considered complexes is always @xmath14 .
in this section we recall definition of the complex @xmath6 of uni-@xmath153-valent graphs from @xcite . the homology of this complex is naturally isomorphic to @xmath7 , @xmath8 .
the complex @xmath6 is spanned by abstract connected graphs having a non - empty set of non - labeled _ external vertices _ of valence 1 , and a possibly empty set of non - labeled _ internal vertices _ of valence @xmath16 .
the graphs are allowed to have edges joining a vertex to itself and multiple edges . for such graph define its _ orientation set _ as the union of the set of its external vertices ( considered as elements of degree @xmath17 ) , the set of its internal vertices ( considered as elements of degree @xmath18 ) , and the set of its edges ( considered as elements of degree @xmath19 ) . by an _ orientation _ of a graph
we will understand ordering of its orientation set together with an orientation of all its edges .
two such graphs are _ equivalent _ if there is a bijection between their sets of vertices and edges respecting the adjacency structure of the graphs , orientation of the edges , and the order of the orientation sets .
the space of @xmath6 is the quotient space of the vector space freely spanned by equivalence classes of such oriented graphs modulo the orientation relations : \(1 ) @xmath20 if @xmath21 differs from @xmath22 by reversing the orientation of an edge .
\(2 ) @xmath23 , where @xmath22 is obtained from @xmath21 by a permutation of the orientation set .
the sign here is the koszul sign of the permutation taking into account the degrees of the elements .
notice that if a graph has a symmetry that produces a negative sign , then such a graph is zero modulo orientation relations ( 1 - 2 ) .
the total degree of a graph is the sum of degrees of all the elements from its orientation set .
the differential @xmath24 of a graph @xmath25 is defined as the sum of expansions of its internal vertices .
an expanded vertex is replaced by an edge .
the set of edges adjacent to the expanded vertex splits into two sets one containing the edges that go to one vertex of the new edge and the other set containing the edges that go to the other vertex .
an expansion of a vertex of valence @xmath26 is a sum of @xmath27 graphs obtained in such way .
one subtracts @xmath28 to exclude graphs with internal vertices of valence @xmath29 , and one divides by 2 because of the symmetry .
the orientation set of a new graph is obtained by adding the new vertex and the new edge as the first and second elements to the orientation set , and by orienting the new edge from the old vertex to the new one .
there is a freedom which of 2 vertices of the new edge is considered as a new one and which as an old one , but regardless of this choice , the orientation of the boundary graph is the same .
all the graphs in the differential appear with positive sign ( the sign is hidden in the way we order the orientation set and orient the new edge ) .
notice that the differential preserves the number of external vertices @xmath30 referred as _
hodge degree _ and also the first betti number @xmath31 ( number of loops in a graph ) . the first betti number @xmath32 of the graph obtained by gluing together all univalent vertices will be called _
complexity_. it is also preserved by the differential .
one obviously has @xmath33 the part of @xmath6 of hodge degree @xmath30 and complexity @xmath32 is denoted by @xmath34 .
we will also define _ defect _ of a graph as the sum @xmath35 , where @xmath36 runs through the internal vertices of the graph , and @xmath37 is the valence of @xmath36 .
the defect measures how much the graph is different from a uni - trivalent one .
the differential decreases the defect by one .
in this section we define a quasi - isomorphic subcomplex @xmath38 of @xmath6 which will require a few definitions .
an edge in a connected graph is called _ tree - type _ if when one removes it , the graph becomes disconnected , whose exactly one connected part is a tree . by the _ frame _ of a graph
we will mean its subgraph generated by edges which are not tree - type .
notice that the frame of a connected graph is always connected .
a tree - type edge is called a _ hair _ if one of its vertices is univalent and the other one belongs to the frame and has no other tree - type edges adjacent to it .
a vertex is called _ cut _ if when one removes it the graph becomes disconnected .
one defines the _
complex @xmath38 of hairy graphs _ as the subcomplex of @xmath6 spanned by trees with @xmath39 external vertices and by graphs whose frame has no cut - vertices and whose tree - type edges are all hairs . [ t_hairy ]
the inclusion @xmath40 is a quasi - isomorphism .
for the part @xmath41 , i.e. the part which is spanned by trees , see ( * ? ? ?
* proposition 3.2 ) . for the case
when the first betti number is positive the proof follows from the argument of ( * ? ? ?
* theorem 1.1 ) . in that paper ,
graphs do not have hairs , the orientation data is not as general , and the differential contracts , rather than expands edges .
however the argument still works with a little modification .
one writes @xmath42 , where @xmath43 expands cut vertices into separating edges and @xmath44 .
consider the subspace of graphs which have a cut vertex which is not a trivalent vertex incident to a hair , which is a complex with respect to @xmath43 . by the same argument as in that paper ,
the homology of this complex is trivial .
so in the spectral sequence for the double complex , only graphs without cut vertices survive , and the surviving boundary operator @xmath45 is the standard boundary restricted to the hairy graph subcomplex .
from theorem [ t_hairy ] one can immediately conclude that the loopless part of @xmath38 is always one - dimensional .
it is spanned by the tree with two vertices when @xmath9 and @xmath10 are of the same parity and by the tree with three external and one internal vertices when @xmath9 and @xmath10 are of opposite parity .
the differential is obviously zero .
the one - loop part of @xmath38 is spanned by wheels
graphs obtained from a circle by attaching several hairs .
* proposition 3.3 ) tells when such wheels survive their dihedral symmetry .
the differential is also zero in this situation .
let @xmath46 denote the space of two - loop hairy graphs of defect @xmath47 .
notice that the frame of any such graph has the shape of the @xmath48-graph .
the defect @xmath47 can be either 0 , 1 , or 2 , see the figure below .
[ 0][0][1][0]@xmath49 [ 0][0][1][0]@xmath50 [ 0][0][1][0]@xmath51 one gets the complex @xmath52 we will see that the differential @xmath53 is injective . which means that the homology must be concentrated in defect 0 and 1 .
let @xmath54 , respectively @xmath55 , be the rank of the homology of defect 0 , respectively 1 , in hodge degree @xmath56 .
let @xmath57 denote the euler characteristic of in hodge degree @xmath56 .
we will compute the generating functions @xmath58 obviously , @xmath59 and @xmath60 depend on the parities of @xmath9 and @xmath10 .
the total degree of a graph with @xmath56 hairs in @xmath61 is @xmath62 .
this implies the following lemma .
[ l_euler ] one has @xmath63.\ ] ] follows from injectivity of @xmath64 , see section [ s_homology ] .
the main idea for our computations is that the spaces of graphs will be encoded as certain spaces of polynomials .
denote by @xmath65 the graph from @xmath49 that has @xmath66 hairs on the upper edge of @xmath48 , @xmath67 hairs on the middle one , and @xmath68 hairs on the lower one .
its edges are oriented as in the figure below .
[ 0][0][1][0]@xmath69 [ 0][0][1][0]@xmath70 [ 0][0][1][0]@xmath71 [ 0][0][1][0]@xmath72 [ 0][0][1][0]@xmath73 [ 0][0][1][0]@xmath74 [ 0][0][1][0]@xmath75 [ 0][0][1][0]@xmath76 [ 0][0][1][0]@xmath77 [ 0][0][1][0]@xmath78 [ 0][0][1][0]@xmath79 [ 0][0][1][0]@xmath80 [ 0][0][1][0]@xmath81 [ 0][0][1][0]@xmath82 [ 0][0][1][0]@xmath83 [ 0][0][1][0]@xmath84 [ 0][0][1][0]@xmath85 [ 0][0][1][0]@xmath86 [ 0][0][1][0]@xmath87 [ 0][0][1][0]@xmath88 [ 0][0][1][0]@xmath89 [ 0][0][1][0]@xmath90 [ 0][0][1][0]@xmath91 [ 0][0][1][0]@xmath92 [ 0][0][1][0]@xmath93 [ 0][0][1][0]@xmath94 [ 0][0][1][0]@xmath95 [ 0][0][1][0]@xmath96 [ 0][0][1][0]@xmath97 [ 0][0][1][0]@xmath98 [ 0][0][1][0]@xmath99 [ 0][0][1][0]@xmath100 [ 0][0][1][0]@xmath101 [ 0][0][1][0]@xmath102 [ 0][0][1][0]@xmath103 [ 0][0][1][0]@xmath104 [ 0][0][1][0]@xmath105 [ 0][0][1][0]@xmath106 [ 0][0][1][0]@xmath107 [ 0][0][1][0]@xmath108 the ordering of its orientation set is as follows .
first one has @xmath109 .
then one puts @xmath110 , then a similar product of @xmath111 , @xmath112 , corresponding to the middle edge , then a similar product of @xmath113 , @xmath114 . abusing notation we will denote by @xmath65 similar graphs in @xmath50 and @xmath51 .
the ordering of the orientation sets will start in these cases as @xmath115 , @xmath116 respectively .
the two - loop graphs have two possible types of symmetry : symmetry with respect to the vertical line and transpositions of edges in the @xmath48-graph . if a graph has a symmetry that produces a negative sign , it means the graph is zero modulo orientation relations .
the lemma below computes the signs arising from these symmetries .
[ l_sign ] ( 1 ) the symmetry of the graph @xmath117 with respect to the vertical line produces the sign @xmath118
\(2 ) the symmetry of the graph @xmath119 with respect to the vertical line produces the sign @xmath120 \(3 ) the graph @xmath121 , @xmath122 , is equal to @xmath123 with the sign @xmath124 notice that the orientation of @xmath65 that we defined matters only for ( 3 ) . in the latter case the sign @xmath125 arises from the transposition of @xmath75 and @xmath76 .
the sign @xmath126 is due to the permutation of the factors @xmath111 with @xmath127 .
the sign in ( 1 ) is the product of * @xmath128 corresponding to the transpositions of @xmath69 with @xmath70 , @xmath73 with @xmath74 , @xmath71 with @xmath72 ; * @xmath129 corresponding to permutation of @xmath130 s , @xmath131 s , @xmath132 s , @xmath133 s ; * @xmath134 corresponding to permutation of @xmath135 s , @xmath136 s , @xmath137 s , @xmath138 s , @xmath139 s , @xmath140 s , @xmath141 s , @xmath142 s , @xmath143 s ; * @xmath144 corresponding to the change of orientation of edges .
the sign in ( 2 ) is the same as in ( 1 ) except that one should not count the transpositions of @xmath69 with @xmath70 and of @xmath73 with @xmath74 .
let @xmath145 $ ] denote the the free commutative algebra generated by @xmath146 , @xmath147 , @xmath148 .
we say that a polynomial is even ( odd , respectively ) if all of its monomials are of even ( odd , respectively ) degree . in the previous section we encoded two - loop graphs by monomials @xmath65 .
the symmetric group @xmath149 acts on the space @xmath145 $ ] by permuting the variables @xmath146 , @xmath147 , @xmath148 , which geometrically correspond to the reordering of the edges in the theta - graph .
lemma [ l_sign ] ( 3 ) tells us what sign should be used to obtain an equivalent monomial - graph . instead of taking a quotient by this action we will be taking invariants since in characteristic zero ,
invariants coincide with coinvariants . when both @xmath9 and @xmath10 are odd , a graph with @xmath150 , @xmath151 , hairs respectively , will be denoted instead by @xmath152 .
it follows from lemma [ l_sign ] that the space @xmath49 in this case is represented by odd symmetric polynomials , the space @xmath50 is represented by the space of all symmetric polynomials , and the space @xmath51 is represented by even symmetric polynomials whose all monomials are of strictly positive degree : @xmath153 \xrightarrow{\partial_2 } { \operatorname{sym}}[x_1 , x_2 , x_3 ] \xrightarrow{\partial_1 } { \operatorname{sym}}^{\mathrm{{\mathrm{even}}}}_{>0}[x_1 , x_2 , x_3 ] \xrightarrow{\partial_0 } 0.\ ] ] the following result is well known .
the algebra @xmath154 $ ] is isomorphic to the free commutative algebra generated by three letters : @xmath155 $ ] , where @xmath156 , @xmath157 , and @xmath158 .
so , we have as a basis for @xmath154 $ ] polynomials of the form @xmath159 , where @xmath160 .
also , every such symmetric polynomial has degree @xmath161 , since @xmath103 is of degree 1 , @xmath104 is of degree 2 , and @xmath162 is of degree 3 .
our next corollary follows , which we use freely throughout this section .
@xmath163 = \langle e_1^\alpha e_2^\beta e_3^\gamma |\ , \alpha + \gamma \equiv 0 \,\mathrm{mod}\ , 2\rangle\ ] ] and @xmath164 = \langle e_1^\alpha e_2^\beta e_3^\gamma |\ , \alpha + \gamma \equiv 1 \,\mathrm{mod}\ , 2\rangle.\ ] ] now we will describe the differentials . let @xmath165\to { { \mathbb{q}}}[x_1,x_2,x_3]$ ] be the automorphism of @xmath145 $ ] sending each generator @xmath166 to @xmath167 , @xmath151 .
geometrically this automorphism corresponds to the symmetry with respect to the vertical line that our hairy @xmath48-graphs have .
notice that the projection to the space of even , respectively odd , monomials is described by @xmath168 , respectively @xmath169 . sometimes the even , respectively odd , part of @xmath170
will also be denoted by @xmath171_{{\mathrm{even}}}$ ] , respectively @xmath171_{{\mathrm{odd}}}$ ] .
[ l_diff_m_odd_n_odd ] in the case both @xmath9 and @xmath10 are odd , the differentials of the hairy two - loop graph - complex are defined as follows @xmath172_{{\mathrm{even}}}.\end{aligned}\ ] ] the coefficient 2 in the definition of @xmath64 is due to the fact that there are two 4-valent vertices to expand , which produce the same result due to the symmetry .
such an expansion produces a new hair on one of the three edges of @xmath48 which is algebraicly expressed as a multiplication by @xmath173 .
the sign minus in both cases is due to the rule of signs for the differential , see section [ s_graph ] .
see also remark [ r_diff_general ] where general formulas for the differentials in all the four cases are given .
we now wish to compute homology .
notice @xmath174 since @xmath175 , @xmath176 secondly , we have @xmath177
so , we have @xmath178 we will now compute @xmath179 .
we have @xmath180 in other words this kernel consists of symmetric even polynomials .
similarly we have @xmath181 thus , @xmath182 we see that @xmath183 $ ] and @xmath184 $ ] . as a consequence we get the following result . [ t_homology_m_odd_n_odd ] in case both @xmath9 and @xmath10 are odd , the generating functions of the dimensions of the two - loop homology @xmath185 and @xmath179 , and of the euler characteristics , are @xmath186 although the underlying spaces of @xmath185 and @xmath179 are the same , we add the factor @xmath30 in @xmath187 because of the additional hair growing from the left vertex of the @xmath48-graph . the formula for @xmath188
was obtained using lemma [ l_euler ] .
this is the second case where @xmath9 and @xmath10 have the same parity .
we consider the same space of polynomials @xmath145 $ ] as in our previous case .
however , we obtain a negative sign when we transpose two edges , see lemma [ l_sign ] ( 3 ) .
thus , we can no longer use symmetric polynomials to describe our space ; instead , we turn to antisymmetric polynomials .
an antisymmetric polynomial is defined as a polynomial on @xmath189 variables such that any transposition of variables results in a negative sign .
the two - loop hairy graph - complex has now the following form : @xmath190 \xrightarrow{\partial_2 } { \operatorname{asym}}[x_1,x_2,x_3 ] \xrightarrow{\partial_1 } { \operatorname{asym}}^{{\mathrm{even}}}[x_1,x_2,x_3 ] \xrightarrow{\partial_0 } 0\ ] ] [ l_diff_m_even_n_even ] in the case both @xmath9 and @xmath10 are even the differentials in the two - loop hairy graph - complex are defined as follows @xmath191_{{\mathrm{even}}}.\end{aligned}\ ] ] see remark [ r_diff_general ] for a general formula .
one can clearly see that @xmath192 . the following result is well known .
the space @xmath193 $ ] of antisymmetric polynomials is a free module over the algebra @xmath194 $ ] of symmetric polynomials generated by the only element @xmath195 the terms @xmath46 of our complex are generated by the bases : @xmath196 we then compute @xmath185 and @xmath179 : @xmath197 following a similar process for @xmath179 : @xmath198 thus we immediately obtain the following [ t_homology_m_even_n_even ] in case both @xmath9 and @xmath10 are even , the generating functions of the dimensions of the two - loop homology @xmath185 and @xmath179 , and of the euler characteristics , are @xmath199 notice that both @xmath185 and @xmath179 are free modules over @xmath200 $ ] generated by @xmath201 .
the degree of @xmath201 is 6 and the graded dimension of @xmath200 $ ] is @xmath202 .
the result follows . again to compute @xmath188 we used lemma [ l_euler ] .
when @xmath9 and @xmath10 are of different parities , lemma [ l_sign ] ( 3 ) tells us that the generators @xmath166 should be odd to make the action of @xmath149 well defined . to emphasize that they are odd we will denote them by @xmath203 , @xmath151 instead .
let @xmath204 be a super graded algebra generated by odd @xmath203 s , @xmath151 , that anti - commute with each other , but not with themselves . to be precise we require @xmath205 when @xmath206 , but not @xmath207 .
notice that this algebra is neither commutative nor supercommutative .
the group @xmath149 acts on it by renumbering the variables .
we will denote by @xmath208 its @xmath149-invariant part .
notice however that a symmetrization of a monomial @xmath209 can be zero if for example @xmath210 and they are odd .
again this algebra is neither commutative nor supercommutative .
it is not hard to find its set of generators and relations , but we omit it here since it unfortunately does not help with our homology computations .
we will still denote by @xmath211 . to describe the symmetry with respect to the vertical line and applying the signs from lemma [ l_sign ] ( 1 - 2 )
, we define a linear map @xmath212 by @xmath213 given a super graded algebra @xmath214 one can define its opposite algebra @xmath215 by taking the same space and defining the new product @xmath216 as follows @xmath217 a homomorphism @xmath218 is called an antihomomorphism , and an isomorphism @xmath219 is called an antiautomorphism .
the following is easy to check .
[ l_anti_aut ] the linear map @xmath220 can also be described as the antiautomorphism of @xmath204 that sends the generators @xmath203 to @xmath221 , @xmath151 .
since @xmath222 this antiautomorphism is an _ anti - involution_. [ d_*parity ] we say that an element @xmath223 is @xmath224even ( @xmath224odd , respectively ) if @xmath225 ( @xmath226 , respectively ) . with this , we may describe our graph complex for this case : @xmath227 the superscript @xmath228 says that we take only subspace of @xmath224even polynomials ; the subscript @xmath229 means that we take polynomials without constant term .
[ l_diff_m_even_n_odd ] for the case @xmath9 even and @xmath10 odd , the differentials in the two - loop hairy complex are described as follows : @xmath230_{*even}.\end{aligned}\ ] ] [ r_diff_general ] in all the four cases the differentials in the two loop hairy graph - complex can be described as follows @xmath231 in the above @xmath232 refers to the supergrading of @xmath170 . in the case @xmath9 and @xmath10 of the same parity , @xmath232 is always even .
the signs @xmath233 and @xmath234 in the aforementioned formulas are not important for the homology computations for which reason we do nt show how they were calculated .
the following lemma proves that @xmath192 .
[ l_square ] let @xmath235 be a linear element of @xmath204 . if @xmath223 is @xmath224even , respectively @xmath224odd , then @xmath236 is @xmath224odd , respectively @xmath224even . using the fact that @xmath237 is an antiautomorphism and that @xmath238 , we get @xmath239 the result follows .
unfortunately the property of being @xmath224even or @xmath224odd is not preserved by a product .
thus in particular @xmath240 is not an algebra .
this makes our computations more difficult for the case when @xmath241 is odd . in order to compute the homology
we will introduce a good basis in our spaces @xmath46
, @xmath122 .
first denote by @xmath242 the symmetrization of @xmath209 that has coefficient 1 in front of @xmath209 .
for example , @xmath243 the elements @xmath242 satisfying @xmath244 and if @xmath245 , then @xmath150 is even , @xmath246 , form a basis of @xmath247 that will be denoted by @xmath248 .
this set comes with a natural order
first one compares the degree @xmath249 , then within the same degree one orders the elements lexicographically .
since the above elements form a basis , any symmetric polynomial @xmath170 is a unique linear combination of them .
the maximal element of @xmath248 that has a non - zero coefficient in the above sum , will be called the leading term of @xmath170 .
the subset of @xmath224even , respectively @xmath224odd , elements in @xmath248 will be denoted by @xmath250 , respectively @xmath251 .
for a basis of @xmath49 we will simply choose the set @xmath250 .
for a basis of @xmath50 we will take a union of the following sets @xmath252 the above union of sets forms a basis of @xmath253 since every element of @xmath248 appears exactly once as a leading term of one of the elements in the union .
since @xmath254 up to a sign is a multiplication by @xmath255 on the right , we have @xmath256 thus @xmath257 is a space spanned by a subset of basis elements .
now , we compute @xmath258 on the basis elements . applying lemmas [ l_diff_m_even_n_odd ] , [ l_square ] , @xmath259 notice that the second line produces elements whose leading terms form the subset of @xmath250 of elements @xmath242 such that @xmath260 , or @xmath261 and @xmath67 is even .
we continue our computation of @xmath262 on the basis elements : [ cols="<,<,<",options="header " , ] it is easy to see that the images from the fourth line are twice the images of the seventh line .
if one excludes the elements @xmath263 | k_2 \equiv 0 \bmod2 , k_3 \equiv
3\bmod4\}$ ] from the basis of @xmath50 ( i.e. the elements that appear on the fourth line above ) , then all the non - zero images of such set have different leading terms
. thus these images are linearly independent . to complete the obtained set to a basis of @xmath51
we add elements @xmath264 \,|\ , k_2 \equiv 0 \bmod2 , k_3 \equiv 1 \bmod4 , k_2 >
k_3\ } \cup \\ & \{[k_2 + 1 , k_2 , k_3 ] \,|\ , k_2 \equiv 0 \bmod2 , k_3 \equiv 2 \bmod4 , k_2 >
k_3 + 1 \ } \cup \{[k_3 + 1 , k_3 + 1 , k_3 ] \,|\ , k_3 \equiv 0 \bmod4\}.\end{aligned}\ ] ] this immediately implies the above set is a basis of cycles in @xmath185 , in other words , @xmath265 \,|\ , k_2 \equiv 0 \bmod2 , k_3 \equiv 1 \bmod4 ,
k_2 > k_3 \rangle \oplus \\ \langle [ k_2 + 1 , k_2 , k_3 ] | k_2 \equiv 0 \bmod2 , k_3 \equiv 2 \bmod4 , k_2 > k_3 \rangle \oplus \langle [ k_3 + 1 , k_3 + 1 , k_3 ] |
k_3 \equiv 0 \bmod4 \rangle . \label{eq_h0_m_odd_n_even}\end{gathered}\ ] ] we will now compute @xmath179 .
it is easy to see from above that @xmath266 - [ k_2 + 1 , k_2 , k_3 + 1 ] | k_2 \equiv 0 \bmod2 , k_3 \equiv 2 \bmod4 , k_2 > k_3 \rangle \oplus \\ \langle [ k_2 , k_2 , k_3 ] | k_2 \equiv 1 \bmod2 , k_3 \equiv 1 \bmod4 , k_2 \geq k_3 \rangle.\end{gathered}\ ] ] thus we get @xmath267 - [ k_2 + 1 , k_2 , k_3 + 1 ] | k_2 \equiv 0 \bmod2 , k_3 \equiv 2 \bmod4 , k_2 > k_3 \rangle \oplus \\ \langle [ k_2 , k_2 , k_3 ] | k_2 \equiv 1 \bmod2 , k_3 \equiv 1 \bmod4 , k_2 \geq k_3 \rangle . \label{eq_h1_m_odd_n_even}\end{gathered}\ ] ] [ t_homology_m_odd_n_even ] in case @xmath9 is odd and @xmath10 is even , the generating functions of the two - loop graph - homology @xmath185 and @xmath179 , and of the euler characteristics , are @xmath268 we start with @xmath59 .
the first summand in has the generating functions of dimensions @xmath269 in this formula @xmath270 in the numerator corresponds to @xmath271 $ ] .
the denominator is obtained exactly as described in the proof of theorem [ t_homology_m_even_n_odd ] .
the second summand has the generating function of the dimensions @xmath272 where @xmath273 comes from @xmath274 $ ] .
the last summand of has the generating function of dimensions @xmath275 where @xmath276 corresponds to @xmath277 $ ] .
+ for @xmath187 , the first summand of has the generating function of dimensions @xmath278 here the numerator @xmath279 corresponds to @xmath280 - [ 5 , 4 , 3]$ ] .
for the second summand we get the function @xmath281 here @xmath282 stands for the element @xmath283 $ ] .
+ to compute @xmath188 , again we use lemma [ l_euler ] .
we will now examine the concentration of homology in all four cases .
we will first focus on the case where both @xmath10 and @xmath9 are odd . recall that the formula for the graded dimension of @xmath185 is @xmath284 and the graded dimension for @xmath179 is @xmath285 [ t_rank_m_odd_n_odd ] let @xmath54 represent the sequence of coefficients generated by @xmath59 , and let @xmath55 represent the sequence of coefficients generated by @xmath60 .
when @xmath10 and @xmath9 are odd , we have @xmath286 and @xmath287 follows from the fact that the expansion of @xmath288 has coefficient zero in front of any odd exponent @xmath289 and coefficient @xmath290 in front of any even exponent .
the next case we examine is when both @xmath10 and @xmath9 are even .
we have @xmath291 [ t_rank_m_even_n_even ] as before , let @xmath54 represent the sequence of coefficients generated by @xmath59 , and let @xmath55 represent the sequence of coefficients generated by @xmath60 .
then , when @xmath10 and @xmath9 are even , @xmath292 similar to the previous theorem .
we now turn to the cases where @xmath293 is odd , beginning with @xmath10 odd and @xmath9 even .
our formulas for graded dimension are @xmath294 and the graded dimension for @xmath179 is @xmath295 [ t_rank_m_even_n_odd ] when @xmath10 is odd and @xmath9 is even , using the same notation as above , we have @xmath296 and @xmath297 follows from the fact that @xmath298 has coefficient zero in front of any exponent @xmath289 with @xmath56 non - divisible by 4 , and coefficient @xmath299 for @xmath56 divisible by 4 . finally ,
when @xmath10 is even and @xmath9 is odd , recall that the graded dimension of @xmath185 is @xmath300 and the graded dimension for @xmath179 is @xmath301 again , we expand these series with respect to @xmath30 to find a general formula to compute the coefficient for a given exponent .
[ t_rank_m_odd_n_even ] when @xmath10 is even and @xmath9 is odd , we have @xmath302 where @xmath54 represents the coefficient of @xmath185 , and @xmath303 where @xmath55 represents the coefficient of @xmath179 .
similar to the previous theorem .
notice that , for a given hodge grading , in all the four cases the homology is concentrated in a single degree .
all the authors acknowledge nsf for the support . the idea for this work appeared during conant s visit to k - state in 2009 supported by the midwest topology network grant dms 0844249 . at that moment
conant and turchin made computations in the easy case `` of @xmath241 even . later on costello and weed were working with turchin on this project supported by the summer reu program dms 1004336 .
costello and weed redid the computation that were previously done and finished them by computing the difficult case '' of @xmath241 odd .
costello and weed thank k - state for hospitality and also baltazar chavez - diaz another reu student , for discussions and his contribution to the project . finally turchin was also partially supported by dms 0967649 and by the max planck institute for mathematics in bonn , where he was working on the last version of this paper , and which he thanks for hospitality .
99 g. arone , v. turchin . on the rational homology of high dimensional analogues of spaces of long knots .
arxiv:1105.1576 .
g. arone , v. turchin .
graph - complexes computing the rational homotopy of high dimensional analogues of spaces of long knots . to appear in _ annales de linstitut fourier _ 64 ( 2014 ) , arxiv:1108.1001 .
j. conant , f. gerlits , k. vogtmann.cut vertices in commutative graphs._q .
_ 56 ( 2005 ) , no . 3 ,
pp 321336 . o. dasbach . on the combinatorial structure of primitive vassiliev invariants ii , _
j. combin .
theory ser .
a _ 81 ( 2 ) ( 1998 ) 127139 .
d. moskovich , t. ohtsuki .
vanishing of 3-loop jacobi diagrams of odd degree . _
j. combin .
theory ser .
a _ 114 ( 2007 ) , no . 5 , 919930 .
t. nakatsuru , vassiliev fuhenryou no jigen no keisan , masters thesis , tokyo institute of technology , january 1998 .
in this appendix section we present tables of ranks of the 2-loops graph - homology in small hodge degrees .
notice that the results of our computations of euler characteristics ( the third column in each table ) confirm previous computations the diagonal @xmath304 in ( * ? ? ?
* tables 1 , 3 , 5 , 7 ) . | arone and turchin defined graph - complexes computing the rational homotopy of the spaces of long embeddings .
the graph - complexes split into a direct sum by the number of loops in graphs . in this paper
we compute the homology of its two - loop part . |
0 let @xmath2 be a positive integer .
we consider the bilinear hilbert transform along a curve @xmath3 defined by @xmath4 where @xmath5 are schwartz functions on @xmath6 .
the main theorem we prove in this paper is [ thm1 ] the bilinear hilbert transform along the curve @xmath3 can be extended to a bounded operator from @xmath0 to @xmath1 .
it can be shown , with a little modification of our method , that the bilinear hilbert transforms along polynomial curves @xmath7 are bounded from @xmath8 to @xmath9 whenever @xmath10 is in the closed convex hull of @xmath11 , @xmath12 and @xmath13 .
this problem is motivated by the hilbert transform along a curve @xmath14 defined by @xmath15 and the bilinear hilbert transform defined by @xmath16 among various curves , one simple model case is the parabola @xmath17 in the two dimensional plane .
this work was initiated by fabes and riviere @xcite in order to study the regularity of parabolic differential equations . in the last thirty years
, considerable work on this type of problems had been done . a nice survey on this type of operators
was written by stein and wainger @xcite .
for the curves on homogeneous nilpotent lie groups , the @xmath18 estimates were established by christ @xcite .
the work for the hilbert transform along more general curves with certain geometric conditions such as the `` flat '' case can be found in christ , duoandikoetxea and j. l. rubio de francia , and nagel , vance , wainger and weinberg s papers @xcite .
the general results were established recently in @xcite for the singular radon transforms and their maximal analogues over smooth submanifolds of @xmath19 with some curvature conditions .
+ in recent years there has been a very active trend of harmonic analysis using time - frequency analysis to deal with multi - linear operators . a breakthrough on the bilinear hilbert transform
was made by lacey and thiele @xcite . following lacey and thiele s work ,
the field of multi - linear operators has been actively developed , to the point that some of the most interesting open questions have a strong connection to some kind of non - abelian analysis .
for instance , the tri - linear hilbert transform @xmath20 has a hidden quadratic modulation symmetry which must be accounted for in any proposed method of analysis .
this non - abelian character is explicit in the work of b. kra and b. host @xcite who characterize the characteristic factor of the corresponding ergodic averages @xmath21 here , @xmath22 is a measure preserving system , @xmath23 is the sigma - field which describes the characteristic factor . in this case
, it arises from certain @xmath24-step nilpotent groups .
the limit above is in the sense of @xmath25-norm convergence , and holds for all bounded @xmath26 .
the ergodic analog of the bilinear hilbert transform along a parabola is the non - conventional bi - linear average @xmath27 where @xmath28 is the profinite factor , a subgroup of the maximal abelian factor of @xmath29 .
the proof of the characteristic factor result above , due to furstenberg @xcite , utilizes the characteristic factor for the three - term result .
we are indebted to m. lacey for bringing furstenberg s theorems to our attention . however , a notable fact is that our proof for the bilinear hilbert transform along a monimial curve does not have to go through the tri - linear hilbert transform .
the proof provided in this article heavily relies on the concept of `` quadratic '' uniformity and some kind of `` quadratic '' fourier analysis , initiated by gowers @xcite . and
perhaps this is a starting point to understand the tri - linear hilbert transform . +
another prominent theme is the relation of the bilinear hilbert transforms along curves and the multilinear oscillatory integrals .
the bilinear hilbert transforms along curves are closely associated to the multilinear oscillatory integrals of the following type .
@xmath30 where @xmath31 is a unit ball in @xmath32 , @xmath33 are vectors in @xmath32 , and the phase function @xmath34 satisfies a non - degenerate condition @xmath35 here @xmath36 s are unit vectors orthogonal to @xmath37 s respectively . for a polynomial phase @xmath34 with the non - degenerate condition ( [ nondegcond ] ) , it was proved in @xcite that @xmath38 holds for some positive number @xmath39 . for the particular vectors @xmath37 s and the non - degenerate phase @xmath34 encountered in our problem , an estimate similar to ( [ linftyest ] ) still holds
however , one of the main difficulties arises from the falsity of @xmath40 decay estimates for the trilinear form @xmath41 . in order to overcome this difficulty
, we end up introducing the `` quadratic '' uniformity , which plays a role of a `` bridge '' connecting two spaces @xmath40 and @xmath42 . +
the method used in this paper essentially works for those curves on nilpotent groups .
it is possible to extend theorem [ thm1 ] to the general setting of nilpotent lie groups .
but we will not pursue this in this article .
there are some related questions one can pose . besides the generalisation to the more general curves , it is natural to ask the corresponding problems in higher dimensional cases and/or in multi - linear cases .
for instance , in the tri - linear case , one can consider @xmath43 here @xmath44 are polynomials of @xmath45 .
the investigation of such problems will be discussed in subsequent papers .
+ * acknowledgement * the author would like to thank his wife , helen , and his son , justin , for being together through the hard times in the past two years . and
he is also very grateful to michael lacey for his constant support and encouragement .
0 let @xmath46 be a schwartz function supported in the union of two intervals @xmath47 $ ] and @xmath48 $ ] .
[ lemmaj ] let @xmath49 be a real polynomial with degree @xmath2 . and let @xmath50 .
suppose that the @xmath51-th order derivative of @xmath49 , @xmath52 , does not vanish .
let @xmath53 .
then @xmath54 is bounded from @xmath55 to @xmath9 for @xmath56 , @xmath57 and @xmath58 .
we may without loss of generality restrict @xmath59 , hence likewise the supports of @xmath5 , to fixed bounded intervals whose sizes depend on the coefficients of the polynomial @xmath49 .
this is possible because of the restriction @xmath60 in the integral .
let us restrict @xmath59 in a bounded interval @xmath61 .
it is obvious that @xmath54 is bounded uniformly from @xmath62 to @xmath42 and from @xmath63 to @xmath1 for @xmath64 and @xmath65 .
when @xmath66 in @xmath67 , then the boundedness from @xmath68 to @xmath1 can be obtained immediately by changing variable @xmath69 and @xmath70 since the jocobian @xmath71 . thus @xmath54 is bounded from @xmath68 to @xmath72 since @xmath59 is restricted to a bounded interval @xmath61 and then the lemma follows by interpolation .
when there is a real solution in @xmath73 to the equation @xmath74 , the trouble happens at a neighborhood of @xmath75 , where @xmath76 is the real solution to @xmath74 .
there are at most @xmath77 real solutions to the equation @xmath78 .
thus we only need to consider a small neighborhood containing only one real solution @xmath75 to @xmath74 .
let @xmath79 be a small neighborhood of @xmath75 which contains only one real solution to @xmath78 .
we should prove that @xmath80 for @xmath81 and @xmath82 with @xmath83 .
let @xmath84 be a suitable bump function supported in @xmath85 such that @xmath86 . to get ( [ estlem0 ] )
, it suffices to prove that there is a positive @xmath39 such that @xmath87 for all large @xmath88 , @xmath81 and @xmath82 with @xmath58 , since ( [ estlem0 ] ) follows by summing for all possible @xmath89 . by a translation argument
we need to show that @xmath90 for all large @xmath88 , @xmath81 and @xmath82 with @xmath58 , where @xmath91 is a polynomial of degree @xmath92 defined by @xmath93 .
it is clear that @xmath94 and @xmath95 . when @xmath96 , @xmath97 for some constant @xmath98 depending on the coefficients of @xmath49 .
let @xmath99 $ ] and @xmath100 be defined by @xmath101 \,\,{\rm for}\,\ , n=-1 , \cdots , \frac{(b_p - a_p)\cdot 2^{j}}{c_p}\,.\ ] ] notice that for a fixed @xmath102 , @xmath103 is in @xmath104 for some @xmath105 .
so we can restrict @xmath59 in one of @xmath100 s .
now let @xmath106 .
due to the restriction of @xmath59 , we only need to show that @xmath107 for all large @xmath89 , @xmath81 and @xmath82 with @xmath108 , where @xmath109 , @xmath110 and @xmath111 is independent of @xmath105 . by inserting absolute values throughout
we get @xmath112 maps @xmath8 to @xmath9 with a bound @xmath113 uniform in @xmath105 , whenever @xmath10 belongs to the closed convex hull of the points @xmath114 , @xmath115 and @xmath116 .
observe that @xmath117 since @xmath94 .
by @xmath118 and applying cauchy - schwarz inequality , we obtain for all @xmath88 large enough , @xmath119 hence an interpolation then yields a bound @xmath120 for all triples of reciprocal exponents within the convex hull of @xmath121 , @xmath122 , @xmath114 , @xmath115 and @xmath116 .
this finishes the proof of ( [ estlem3 ] ) .
therefore we complete the proof of lemma [ lemmaj ] notice that if @xmath49 is a monomial @xmath123 , then the lower bound for @xmath124 in lemma [ lemmaj ] can be improved to @xmath125 .
this is because @xmath126 has nonvanishing @xmath127 when @xmath128 .
we now give a counterexample to indicate that the lower bound @xmath129 for @xmath124 is sharp in lemma [ lemmaj ] .
[ counterexample ] let @xmath130 be integers such that @xmath2 and @xmath131 .
there is a real polynomial @xmath132 of degree @xmath133 whose @xmath51-th order derivative does not vanish such that @xmath134 is unbounded from @xmath8 to @xmath9 for all @xmath56 and @xmath135 with @xmath83 , where @xmath134 is the bilinear operator defined by @xmath136 .
let @xmath137 be a very large number .
we define @xmath138 by @xmath139 it is sufficient to prove that if @xmath134 is bounded from @xmath140 to @xmath9 for some @xmath56 and @xmath141 , then @xmath142 .
suppose there is a constant @xmath111 such that @xmath143 for all @xmath144 and @xmath145 .
let @xmath146 be a small positive number . and
let @xmath147}$ ] and @xmath148}$ ] .
let @xmath149 be the intersection point of the curves @xmath150 and @xmath151 in @xmath152-plane with @xmath153 , and let @xmath154 be the intersection point of the curves @xmath155 and @xmath156 in @xmath152-plane with @xmath153 .
let @xmath157 and @xmath158 .
then @xmath159 @xmath160 thus we have @xmath161 @xmath162 when @xmath137 is large and @xmath146 is small , any horizontal line between line @xmath163 and line @xmath164 has a line segment of length @xmath165 staying within the region bounded by curves @xmath156 , @xmath166 , @xmath151 and @xmath150 .
hence , we have @xmath167 by the boundedness of @xmath134 , we have @xmath168 by ( [ lhsbig ] ) we have @xmath169 since @xmath137 can be chosen to be a very large number and @xmath146 can be very small , ( [ sharpest ] ) implies @xmath170 , which completes the proof of lemma [ counterexample ] .
0 let @xmath171 be a standard bump function supported on @xmath48 $ ] . and let @xmath172 it is clear that @xmath46 is an odd function . to obtain the @xmath9 estimates for @xmath173 ,
it is sufficient to get @xmath9 estimates for @xmath174 defined by @xmath175 , where @xmath176 is @xmath177 let @xmath178 be a large positive number ( larger than @xmath179 ) . by lemma [ lemmaj ]
, we have that if @xmath180 , @xmath181 for all @xmath56 and @xmath83 , where the operator norm @xmath182 depends on the upper bound @xmath178 .
hence in the following we only need to consider the case when @xmath183 .
in fact we prove the following theorem .
[ thmlargej ] let @xmath176 be defined as in ( [ defoftgj ] ) .
then the bilinear operator @xmath184 is bounded from @xmath0 to @xmath1 .
clearly theorem [ thm1 ] follows by theorem [ thmlargej ] and lemma [ lemmaj ] .
the rest part of the article is devoted to a proof of theorem [ thmlargej ] .
+ we begin the proof of theorem [ thmlargej ] by constructing an appropriate decomposition of the operator @xmath185 .
this is done by an analysis of the bilinear symbol associated with the operator .
+ expressing @xmath176 in dual frequency variables , we have @xmath186 where the symbol @xmath187 is defined by @xmath188
first we introduce a resolution of the identity .
let @xmath189 be a schwarz function supported on @xmath190 such that @xmath191 if @xmath192 .
set @xmath193 to be a schwartz function satisfying @xmath194 then @xmath193 is a schwartz function such that @xmath195 is supported on @xmath196 and @xmath197 and for any @xmath198 , @xmath199 which is a bump function supported on @xmath200 . from ( [ defofphi ] )
, we can decompose @xmath176 into two parts : @xmath201 and @xmath202 , where @xmath203 is given by @xmath204 and @xmath202 is defined by @xmath205 define @xmath206 by @xmath207 clearly @xmath208 . in @xmath209
, the phase function @xmath210 does not have any critical point in a neighborhood of the support of @xmath46 , and therefore a very rapid decay can be obtained by integration by parts so that we can show that @xmath211 is essentially a finite sum of paraproducts ( see section [ paraproduct ] ) .
a critical point of the phase function may occur in @xmath212 and therefore the method of stationary phase must be brought to bear in this case , exploiting in particular the oscillatory term .
this case requires the most extensive analysis .
[ ht ] -plane for @xmath213 when
@xmath214,title="fig : " ] notice that there are only finitely many @xmath215 if @xmath216 is fixed in ( [ defoftj2 ] ) . without loss of generality , we can assume @xmath217 .
then in order to get the @xmath9 estimates for @xmath218 , it suffices to prove the @xmath9 boundedness of @xmath219 , where @xmath220 s are defined by @xmath221 it can be proved that @xmath222 is equal to @xmath223 , where @xmath224 is a paraproduct studied in theorem [ para0uniest ] .
this can be done by fourier series and the cancellation condition of @xmath46 and thus @xmath225 is essentially a paraproduct .
we omit the details for it since it is exactly same as those in section [ paraproduct ] for the case @xmath226 .
therefore , the most difficult term is @xmath227 . for this term
, we have the following theorem . [ thmoftm ]
let @xmath220 be a bilinear operator defined as in ( [ defoftm ] ) .
then there exists a constant @xmath111 such that @xmath228 holds for all @xmath229 .
a delicate analysis is required for proving this theorem .
we will prove it in subsection [ proofofthmtm ] .
theorem [ thmlargej ] follows from theorem [ thmoftm ] and the boundedness of @xmath226 .
the rest of the article is organized as follows . in section [ paraproduct ] ,
the @xmath9-boundedness will be established for @xmath226 .
some crucial bilinear restriction estimates will appear in section [ bfr ] and as a consequence theorem [ thmoftm ] follows .
sections [ sphasetri]-[proofprop2neg ] are devoted to a proof of the bilinear restriction estimates .
in this section we prove that @xmath211 is essentially a finite sum of certain paraproducts bounded from @xmath8 to @xmath9 .
first let us introduce the paraproduct encountered in our problem .
let @xmath230 , @xmath231 be positive integers and @xmath232 be integers .
@xmath233\ ] ] and @xmath234\,.\ ] ] let @xmath235 be a schwartz function whose fourier transform is a standard bump function supported on a small neighborhood of @xmath48 $ ] or @xmath47 $ ] , and @xmath236 be a schwartz function whose fourier transform is a standard bump function supported on @xmath237 $ ] and @xmath238 . for @xmath239 and @xmath240
, define @xmath241 by @xmath242 it is clear that @xmath243 is supported on @xmath244 . for locally integrable functions @xmath245 s ,
we define @xmath246 s by @xmath247 we now define a paraproduct to be @xmath248 for this paraproduct , we have the following uniform estimates .
[ para0uniest ] for any @xmath249 , @xmath250 with @xmath251 , there exists a constant @xmath111 independent of @xmath252 such that @xmath253 for all @xmath254 and @xmath255 .
the @xmath256 case can be handled by a telescoping argument .
the @xmath257 case is more complicated and it requires a time - frequency analysis .
a proof of theorem [ para0uniest ] can be found in @xcite .
the constant @xmath111 in theorem [ para0uniest ] may depend on @xmath231 .
it is easy to see that @xmath111 is @xmath258 .
it is possible to get a much better upper bound such as @xmath259 by tracking the constants carefully in the proof in @xcite .
but we do not need the sharp constant in this article .
the independence on @xmath260 is the most important issue here .
+ we now return to @xmath261 .
this sum can be written as @xmath262 , where @xmath263 is a bilinear operator defined by @xmath264 and @xmath265 is a bilinear operator given by @xmath266 .
we now prove that @xmath265 can be reduced to the paraproducts studied in @xcite .
indeed , if @xmath267 , then let @xmath268 be @xmath269 clearly , if @xmath270 , @xmath271 and @xmath272 thus if @xmath273 , then we have @xmath274 where @xmath275 is the fourier coefficient . from ( [ d1est ] ) and ( [ m1est ] ) , we have latexmath:[\[\label{fouriercoe1 } and notice that @xmath46 is an odd function , then we have @xmath277 where the fourier coefficient @xmath278 satisfies @xmath279 thus @xmath280 equals to a sum of paraproducts @xmath281 where @xmath282 is a function whose fourier transform is @xmath283 @xmath284 is a function whose fourier transform is @xmath285 actually , in the definition of @xmath282 and @xmath284 , @xmath195 should be a schwartz function supported in some neighborhood of @xmath195 and it is identically equal to @xmath286 on the support of @xmath195 .
we abuse the notions here .
but it does no harm to us since the propety of the function does not change significantly .
notice that the fourier transform of @xmath287 is a function supported in the interval @xmath288 $ ] .
we denote @xmath289 by @xmath290 .
thus by the definition of @xmath193 , @xmath291 where @xmath292 is a standard bump function supported in @xmath293 .
let @xmath294 $ ] .
then @xmath295 is supported in @xmath296 . hence due to the fast decay of the fourier coefficients ( [ fouriercoe1 ] ) and ( [ fconst1 ] ) , we actually run into two paraproducts in this case .
one of them is @xmath297 the other is @xmath298 the @xmath9 estimates of these paraproducts follow from theorem [ para0uniest ] .
in fact , for all @xmath299 and @xmath83 , applying theorem [ para0uniest ] , we obtain the @xmath300 estimates uniformly in @xmath301 with the operator norms @xmath302 and @xmath303 for @xmath304 and @xmath305 respectively .
the fast decay estimates of the fourier coefficients ( [ fouriercoe1 ] ) and ( [ fconst1 ] ) then allow us to conclude the desired @xmath9 boundedness of @xmath280 .
this case is similar to the case @xmath263 .
in fact , if @xmath306 , then let @xmath268 be @xmath307 clearly , if @xmath308 , @xmath309 and @xmath310 thus if @xmath311 , then we have @xmath312 where @xmath313 is the fourier coefficient . from ( [ d1estb12 ] ) and ( [ m1estb12 ] ) , we have latexmath:[\[\label{fouriercoe1b12 } and
notice that @xmath46 is an odd function , then we have @xmath315 where the fourier coefficient @xmath316 satisfies @xmath317 thus @xmath265 equals to a sum of paraproducts @xmath318 observe that the fourier transform of @xmath319 is a function supported in the interval @xmath320 $ ] .
we denote @xmath321 by @xmath322 .
thus by the definition of @xmath193 , @xmath323 where @xmath324 is a standard bump function supported in @xmath325 .
let @xmath326 $ ] .
then @xmath327 is supported in @xmath328 . hence due to the rapid decay of the fourier coefficients ( [ fouriercoe1b12 ] ) and ( [ fconst1b12 ] ) , we actually encounter two paraproducts in this case . one of them is @xmath329 the other is @xmath330 applying theorem [ para0uniest ] , for all @xmath299 and @xmath83 , we have the @xmath300 estimates uniformly in @xmath301 with the operator norms @xmath302 and @xmath303 for @xmath331 and @xmath332 respectively
. then the desired @xmath9 estimates of @xmath333 follow due to the fast decay of the fourier coefficients ( [ fouriercoe1b12 ] ) and ( [ fconst1b12 ] ) .
0 let @xmath334 .
we define a bilinear fourier restriction operator of @xmath5 by @xmath335 and @xmath336 where @xmath337 and @xmath338 are the fourier ( smooth ) restrictions of @xmath339 on the support of @xmath195 respectively .
more precisely , @xmath340 are given by @xmath341 @xmath342 by inserting absolute values throughout and applying cauchy - schwarz inequality , the boundedness of @xmath343 from @xmath344 to @xmath1 follows immediately .
moreover , since the fourier transform of @xmath5 are restricted on the support of @xmath195 , we actually can improve the estimate .
let us state the improved estimates by the following theorems , which are of independent interest .
[ thmbfr1 m ] let @xmath2 and @xmath343 be defined as in ( [ defofbfr ] ) and ( [ defofbfrjneg ] ) .
if @xmath345 , then there exists a constant @xmath111 independent of @xmath346 such that @xmath347 holds for all @xmath229 .
[ thmbfr1j ] let @xmath2 and @xmath343 be defined as in ( [ defofbfr ] ) and ( [ defofbfrjneg ] ) . if @xmath348 , then there exists a positive number @xmath349 and a constant @xmath111 independent of @xmath346 such that @xmath350 holds for all @xmath229 .
the positive number @xmath349 in theorem [ thmbfr1j ] can be chosen to be @xmath351 .
thmbfr1 m ] can be proved by a @xmath352 method .
however , the @xmath352 method fails when @xmath353 . to obtain theorem [ thmbfr1j ] ,
we will employ a method related to the uniformity of functions .
+ now we can see that theorem [ thmoftm ] is a consequence of theorem [ thmbfr1 m ] and theorem [ thmbfr1j ] .
define a bilinear operator @xmath354 to be @xmath355 let @xmath356 be defined by @xmath357 a rescaling argument , theorem [ thmbfr1 m ] and theorem [ thmbfr1j ] yield @xmath358 since @xmath359 , we obtain @xmath360 where @xmath361 @xmath362 clearly the right hand side of ( [ esttm221 ] ) is bounded by @xmath363
. therefore we finish the proof of theorem [ thmoftm ] .
+ we now start to make some reductions first for proving theorem [ thmbfr1 m ] and theorem [ thmbfr1j ] .
let @xmath364 be a nonnegative schwartz function such that @xmath365 is supported in @xmath366 $ ] and satisfies @xmath367 .
for any integer @xmath368 , let @xmath369 . and
for @xmath370 , let @xmath371\,.\ ] ] denote the characteristic function of the set @xmath372 by @xmath373 .
we define @xmath374 and @xmath375 here @xmath376 and @xmath377 can be considered as essentially @xmath378 .
clearly we have @xmath379 [ restrictx1 ] let @xmath235 be a schwartz function such that @xmath380 if @xmath381 and @xmath382 is supported on @xmath383 . and let @xmath384 and define @xmath385 by if @xmath386 , then @xmath387 if @xmath388 , then @xmath389
if there is a constant @xmath390 independent of @xmath51 such that @xmath391 holds for all @xmath229 , then for any positive number @xmath39 , @xmath392 where @xmath393 is a constant depending on @xmath39 only . without loss of generality , assume that @xmath394 .
and we only prove the case @xmath386 .
the case @xmath388 can be proved similarly and we omit the proof for this case . from ( [ sum1kn ] ) , we can express @xmath395 as @xmath396 here @xmath397 equals to @xmath398 where @xmath399 @xmath400 @xmath401 putting the absolute value throughout and utilizing the fast decay of @xmath376 , we estimate the sum of @xmath397 for all @xmath402 s with @xmath403 by @xmath404 for all positive integers @xmath105 . since @xmath405
when @xmath45 is in the support of @xmath46 , we dominate this sum by @xmath406 we now turn to sum @xmath407 for all @xmath408 and @xmath409 . observe that when @xmath346 are large , fourier transforms of @xmath410 and @xmath411 are supported in @xmath381 .
thus we have @xmath412 and then ( [ estofbjmn0 ] ) gives @xmath413 which is clearly bounded by @xmath414 combining ( [ k1k2large ] ) and ( [ k1k2small ] ) , we complete the proof .
let @xmath415 be measurable functions supported on @xmath416 .
define a trilinear form @xmath417 by @xmath418 by lemma [ restrictx1 ] , theorem [ thmbfr1 m ] and theorem [ thmbfr1j ] can be reduced to the following theorems respectively .
[ thmtriest1 m ] let @xmath2 and @xmath417 be defined as in ( [ defoflajmn00 ] ) . if @xmath419 , then there exists a constant @xmath111 independent of @xmath346 such that @xmath420 holds for all @xmath421 .
[ thmtriest1 ] let @xmath2 and @xmath417 be defined as in ( [ defoflajmn00 ] ) .
if @xmath348 , then there exist a positive number @xmath349 and a constant @xmath111 independent of @xmath346 such that @xmath422 holds for all @xmath423 and @xmath424 such that @xmath415 are supported on @xmath416 . a proof of theorem [ thmtriest1 m ] will be provided in section [ sphasetri ] and a proof of theorem [ thmtriest1 ] will be given in section [ triest000neg ] .
0 in section [ paraproduct ] , we see that fourier series can help us to reduce the problem to the paraproduct case when @xmath425 .
this method does not work for the case when @xmath426 .
this is because the critical points of the phase function may happen in a neighborhood of @xmath67 , say @xmath427 , which provides a stationary phase for the fourier integral @xmath206 .
this stationary phase gives a highly oscillatory factor in the integral .
we expect a suitable decay from the highly oscillatory factor . in this section we should prove theorem [ thmtriest1 m ] by utilizing essentially a @xmath352 method .
+ let @xmath428 . to prove theorem [ thmtriest1 m ] , it suffices to prove the following @xmath40 estimate for the trilinear form @xmath429 , @xmath430 holds for all @xmath421 . clearly @xmath429 can be expressed as if @xmath386 , @xmath431 and if @xmath388 , @xmath432 whenever @xmath433 , the second order derivative of the phase function @xmath434 is comparable to @xmath435 . we only need to focus on the worst situation when there is a critical point of the phase function in a small neighborhood of @xmath436 .
thus the method of stationary phase yields @xmath437 where @xmath438 is a constant depending only on @xmath92 . henceforth we reduce theorem [ thmtriest1 m ] to the following lemma .
[ triosc*222 ] let @xmath439 be defined by if @xmath386 then @xmath440 and if @xmath388 , then @xmath441 then there exists a positive constant @xmath111 such that @xmath442 holds for all @xmath421 . without loss of generality , we assume that @xmath195 is supported on @xmath48 $ ] ( or @xmath47 $ ] ) . and we only give a proof for the case @xmath386 since a similar argument yields the case @xmath388 . let @xmath443 be a phase function defined by @xmath444 and let @xmath445 and @xmath446 . changing variable @xmath447 and @xmath448 , we have that @xmath449 equals @xmath450 thus by cauchy - schwarz we dominate @xmath451 by @xmath452 where @xmath453 is defined by @xmath454 it is easy to see that @xmath455 equals to @xmath456 where @xmath457 @xmath458 changing variables @xmath459 and @xmath460 , we see that @xmath461 equals to @xmath462 where @xmath463 @xmath464 changing coordinates to @xmath465 , the inner integral becomes @xmath466 where @xmath467 is defined by @xmath468 when @xmath88 is large enough , the mean value theorem yields @xmath469 if @xmath470 . a well - known hrmander theorem on the non - degenerate phase @xcite gives that ( [ inner ] ) is estimated by @xmath471 hence by cauchy - schwarz inequality @xmath472 is bounded by @xmath473 for any @xmath474 .
by one more use of cauchy - schwarz inequality , @xmath472 is dominated by @xmath475 for any @xmath474 .
thus we have @xmath476 this completes the proof of proposition [ triosc*222 ] .
it is easy to see that @xmath477 fails for all @xmath478 .
indeed , let us only consider the case @xmath479 .
assume that ( [ la*jmest4221 ] ) holds for all @xmath479 .
let @xmath480 , then ( [ la*jmest4221 ] ) implies @xmath481 where @xmath482 simply taking @xmath483 , we obtain @xmath484 this clearly can not be true and hence we get a contradiction .
therefore , ( [ la*jmest4221 ] ) does not hold for all @xmath485 .
from this fact , we know that the @xmath352 method can not work for the case @xmath353 . in the following sections , we have to introduce a concept of uniformity and employ a `` quadratic '' fourier analysis .
0 we introduce a concept related to a notion of uniformity employed by gowers @xcite .
a similar uniformity was utilized in @xcite .
let @xmath486 $ ] , let @xmath487 be a collection of some real - valued measurable functions , and fix a bounded interval @xmath488 in @xmath6 .
[ defofunif ] a function @xmath489 is @xmath490-uniform in @xmath487 if @xmath491 for all @xmath492 .
otherwise , @xmath493 is said to be @xmath490-nonuniform in @xmath487 .
[ uniformitythm ] let @xmath178 be a bounded sub - linear functional from @xmath494 to @xmath495 , let @xmath496 be the set of all functions that are @xmath490-uniform in @xmath487 , and let @xmath497 then for all functions in @xmath494 , @xmath498 where @xmath499 clearly the complement @xmath500 is a set of all functions that are @xmath490-nonuniform in @xmath487 .
let us set @xmath501 clearly @xmath502 . in order to obtain ( [ lnonunif ] ) , it suffices to prove that if @xmath503 , then @xmath504 for any @xmath505 , there exists a function @xmath506 such that @xmath507 let @xmath508 be an inner product on @xmath494 defined by @xmath509 for all @xmath510 . since @xmath493 is @xmath490-nonuniform in @xmath487 , there exists a function @xmath511 in @xmath487 such that @xmath512 let @xmath513 such that @xmath514 and @xmath515
. then we can write @xmath493 as @xmath516 sub - linearity of @xmath178 and the triangle inequality then yield @xmath517 notice that @xmath518 if @xmath503 and @xmath519 then from ( [ suplf ] ) and ( [ lftogq ] ) , we have @xmath520 applying the elementary inequality @xmath521 if @xmath522 , we then get @xmath523 from ( [ nonunif1 ] ) , we have @xmath524 now let @xmath525 and we then obtain ( [ nonunif11 ] ) .
therefore we complete the proof .
0 without loss of generality , in the following sections we assume that @xmath526 is supported on @xmath527 for @xmath528 , where @xmath529 is either @xmath530 $ ] or @xmath531 $ ] .
let @xmath532 be a set of some functions defined by @xmath533 [ prop1 ] let @xmath534 be @xmath490-uniform in @xmath532 . and let @xmath386 and @xmath535 be defined as in ( [ defoflajmn00 ] ) .
then there exists a constant @xmath111 independent of @xmath536 such that @xmath537 holds for all @xmath538 and @xmath539 .
let @xmath540 and let @xmath541 be a bilinear operator defined by @xmath542 for all @xmath5 .
decompose @xmath417 into @xmath543 , where @xmath544 let @xmath545 be a fixed point in the interval @xmath546 . and set @xmath547 to be @xmath548 split @xmath549 into two terms : @xmath550 where @xmath551 is equal to @xmath552 and @xmath553 equals to @xmath554 for @xmath555 , let @xmath556 denote @xmath557 we now start to prove that @xmath558 the mean value theorem and the smoothness of @xmath235 yield that for @xmath559 , @xmath560 because @xmath405 when @xmath561 , @xmath551 can be written as @xmath562 where @xmath563 is an integer between @xmath564 and @xmath565 . putting absolute value throughout and applying ( [ estdifffjml1 ] ) plus cauchy - schwarz inequality , we then estimate @xmath566 by @xmath567 which clearly gives ( [ estb1jmnl1 ] ) by one more use of cauchy - schwarz inequality .
+ we now prove that @xmath568 from ( [ sumlk1111 ] ) , we get that @xmath569 equals to @xmath570 where @xmath571 is equal to @xmath572 and @xmath573 equals to @xmath574 cauchy - schwarz inequality yields that @xmath575 in order to obtain a similar estimate for @xmath576 , we change variables by @xmath577 and @xmath578 to express @xmath579 as @xmath580 where @xmath581 is the jocobian @xmath582 .
it is easy to see that the jocobian @xmath583 . as we did for @xmath584
, we dominate the previous integral by @xmath585 notice that @xmath586 whenever @xmath587 .
we then estimate @xmath588 ( [ estb1jmnl122 ] ) follows from ( [ estlajmnkl2 ] ) and ( [ estlajmnkl1 ] ) .
an interpolation of ( [ estb1jmnl1 ] ) and ( [ estb1jmnl122 ] ) then yields @xmath589 we now turn to prove that @xmath590 in dual frequency variables , @xmath591 can be expressed as @xmath592 where @xmath593 @xmath594 if @xmath595 is not in a small neighborhood of @xmath596 , then there is no critical point of the phase function @xmath597 occurring in a small neighborhood of @xmath598 .
integration by parts gives a rapid decay @xmath599 for @xmath600 .
thus in this case , we dominate @xmath601 by @xmath602 for any positive integer @xmath105 .
we now only need to consider the worst case when there is a critical point of the phse function @xmath597 in a small neighborhood of @xmath436 . in this case , @xmath595 must be in a small neighborhood of @xmath596 and the stationary phase method gives @xmath603 where @xmath438 is a constant depending on @xmath92 only .
thus the principle term of @xmath604 is @xmath605 where @xmath606 is a schwartz function supported on a small neighborhood of @xmath596 , and @xmath607 the key point is that the integral in the previous expression can be viewed as an inner product of @xmath608 and @xmath609 , where @xmath610 is a multiplier operator defined by @xmath611 here the multiplier @xmath612 is given by @xmath613 observe that @xmath614 is in @xmath532 for any @xmath615 and @xmath616 .
thus @xmath490-uniformity in @xmath532 of @xmath534 yields @xmath617 and henceforth we dominate @xmath618 by @xmath619 which clearly is bounded by @xmath620 now ( [ estlajmn(2 ) ] ) follows from ( [ nocriest ] ) and ( [ final222la2jmn ] ) . combining ( [ estlajmn(1 ) ] ) and ( [ estlajmn(2 ) ] ) , we finish the proof .
[ cor1 ] let @xmath535 be defined as in ( [ defoflajmn00 ] )
. then there exists a constant @xmath111 independent of @xmath621 such that @xmath622 holds for all @xmath623 which are @xmath490-uniform in @xmath532 , @xmath538 and @xmath624 . since there is a smooth restriction factor @xmath625 in the definition of @xmath626 , the right hand side of ( [ f1unifest ] )
can be sharpen to @xmath627 which is clearly bounded by @xmath628 [ prop2 ] let @xmath535 be defined as in ( [ defoflajmn00 ] ) .
then there exists a constant @xmath111 independent of @xmath621 such that @xmath629 holds for all @xmath630 , @xmath538 and @xmath624 , where @xmath631 is a positive constant defined in ( [ defofmathfrakd1 ] ) .
a proof of proposition [ prop2 ] will be provided in section [ proofprop2 ] .
0 let @xmath633 be a set of some functions defined by @xmath634 [ prop1neg ] let @xmath635 be @xmath490-uniform in @xmath636 . and
let @xmath388 and @xmath535 be defined as in ( [ defoflajmn00 ] ) .
then there exists a constant @xmath111 independent of @xmath536 such that @xmath637 holds for all @xmath638 and @xmath539 .
let @xmath540 and let @xmath541 be a bilinear operator defined by @xmath542 for all @xmath5 .
decompose @xmath417 into @xmath543 , where @xmath544 let @xmath545 be a fixed point in the interval @xmath546 . and set @xmath639 to be @xmath640 split @xmath549 into two terms : @xmath550 where @xmath551 is equal to @xmath641 and @xmath553 equals to @xmath642 for @xmath555 , let @xmath556 denote @xmath557 we now start to prove that @xmath643 the mean value theorem and the smoothness of @xmath235 yield that for @xmath559 , @xmath644 because @xmath405 when @xmath561 , @xmath551 can be written as @xmath645 where @xmath563 is an integer between @xmath564 and @xmath565 .
putting absolute value throughout and applying ( [ estdifffjml1neg ] ) plus cauchy - schwarz inequality , we then estimate @xmath566 by @xmath646 which clearly gives ( [ estb1jmnl1neg ] ) by one more use of cauchy - schwarz inequality .
+ we now prove that @xmath647 from ( [ sumlk1111neg ] ) , we get that @xmath569 equals to @xmath648 where @xmath571 is equal to @xmath649 and @xmath573 equals to @xmath650 cauchy - schwarz inequality yields that @xmath651 in order to obtain a similar estimate for @xmath576 , we change variables by @xmath652 and @xmath653 to express @xmath579 as @xmath654 where @xmath581 is the jocobian @xmath582 .
it is easy to see that the jocobian @xmath583 . as we did for @xmath584
, we dominate the previous integral by @xmath655 notice that @xmath656 whenever @xmath587 .
we then estimate @xmath657 ( [ estb1jmnl122neg ] ) follows from ( [ estlajmnkl2neg ] ) and ( [ estlajmnkl1neg ] ) .
an interpolation of ( [ estb1jmnl1neg ] ) and ( [ estb1jmnl122neg ] ) then yields @xmath658 we now turn to prove that if @xmath635 is @xmath490-uniform in @xmath636 , then @xmath659 in dual frequency variables , @xmath591 can be expressed as @xmath660 where @xmath661 @xmath662 if @xmath663 is not in a small neighborhood of @xmath596 , then there is no critical point of the phse function @xmath597 occurring in a small neighborhood of @xmath598 .
integration by parts gives a rapid decay @xmath599 for @xmath600 .
thus in this case , we dominate @xmath601 by @xmath664 for any positive integer @xmath105 .
we now only need to consider the worst case when there is a critical point of the phse function @xmath597 in a small neighborhood of @xmath436 . in this case
, @xmath663 must be in a small neighborhood of @xmath596 and the stationary phase method gives @xmath665 where @xmath438 is a constant depending on @xmath92 only .
thus the principle term of @xmath604 is @xmath666 where @xmath606 is a schwartz function supported on a small neighborhood of @xmath596 , and @xmath667 the key point is that the integral in the previous expression can be viewed as an inner product of @xmath608 and @xmath668 , where @xmath610 is a multiplier operator defined by @xmath669 here the multiplier @xmath612 is given by @xmath670 observe that @xmath671 is in @xmath636 for any @xmath615 and @xmath672 .
thus @xmath490-uniformity in @xmath636 of @xmath635 yields @xmath673 and henceforth we dominate @xmath618 by @xmath674 which clearly is bounded by @xmath675 now ( [ estlajmn(2)neg ] ) follows from ( [ nocriestneg ] ) and ( [ final222la2jmnneg ] ) . combining ( [ estlajmn(1)neg ] ) and ( [ estlajmn(2)neg ] ) , we finish the proof .
[ cor1neg ] let @xmath388 and @xmath535 be defined as in ( [ defoflajmn00 ] ) .
then there exists a constant @xmath111 independent of @xmath621 such that @xmath676 holds for all @xmath538 which are @xmath490-uniform in @xmath636 , @xmath623 and @xmath624 . since there is a smooth restriction factor @xmath677 in the definition of @xmath626 , the right hand side of ( [ f1unifestneg ] )
can be sharpen to @xmath678 which is clearly bounded by @xmath679 [ prop2neg ] let @xmath680 and @xmath535 be defined as in ( [ defoflajmn00 ] ) . then there exists a constant @xmath111 independent of @xmath621 such that @xmath681 holds for all @xmath682 , @xmath623 and @xmath624
. we will prove proposition [ prop2neg ] in section [ proofprop2neg ] .
we now are ready to provide a proof of theorem [ thmtriest1 ] .
+ corollary [ cor1 ] , proposition [ prop2 ] and theorem [ uniformitythm ] yield that @xmath683 is dominated by @xmath684 holds for all @xmath623 , @xmath538 and @xmath624 .
take @xmath490 to be @xmath685 , then we have @xmath686 this give a proof for the case @xmath386 .
for the case @xmath388 , applying corollary [ cor1neg ] , proposition [ prop2neg ] and theorem [ uniformitythm ] , we estimate @xmath687 by @xmath688 holds for all @xmath623 , @xmath538 and @xmath624
. now choose @xmath490 to be @xmath689 .
then we have @xmath690 which completes the proof of the case @xmath388 . therefore combining ( [ f1unifest221fff ] ) and ( [ f1unifest221ffnegf ] )
, we proved theorem [ thmtriest1 ] .
0 [ vdcorput2d ] let @xmath691 .
let @xmath692 and @xmath693 be fixed bounded intervals . and
let @xmath34 be a function from @xmath694 to @xmath695 satisfying @xmath696 for all @xmath697 .
assume an additional condition holds in the case @xmath698 , @xmath699 for all @xmath697 .
then there exists a constant depending on the length of @xmath692 and @xmath693 but independent of @xmath700 and the locations of @xmath692 and @xmath701 such that @xmath702 for all @xmath229 , where @xmath703 for any @xmath505 .
this lemma is related to a 2-dimensional van der corput lemma proved in @xcite .
the case @xmath704 was proved in @xcite . and a proof of the case @xmath698 can be found in @xcite .
the estimates on @xmath705 in ( [ defofmathfrakd1 ] ) are not sharp . with some additional convexity conditions on the phase function @xmath34 ,
@xmath706 might be improved to be @xmath707 ( see @xcite for some of such improvements ) .
but in this article we do not need to pursue the sharp estimates .
[ mixdevarphi ] let @xmath708 and @xmath34 be a function defined by @xmath709 define @xmath710 by @xmath711 then there exists a constant @xmath712 depending only on @xmath92 such that @xmath713 holds for all @xmath714 $ ] .
moreover , if @xmath214 , we have @xmath715 holds for all @xmath714 $ ] .
[ vdcorput2d111 ] let @xmath488 be a fixed interval of length @xmath286 . and
let @xmath717 be a bump function supported on @xmath718 $ ] ( or @xmath719 $ ] ) .
suppose that @xmath720 is a phase function defined by @xmath721 where @xmath722 are constants independent of @xmath723 such that @xmath724 .
let @xmath725 be a bilinear form defined by @xmath726 then we have @xmath727 holds for all @xmath728 and @xmath729 , where @xmath712 is a constant depending only on @xmath92 .
the bilinear form @xmath730 equals to @xmath731 , where @xmath732 is defined by @xmath733 by a change of variables , @xmath734 can be expressed as @xmath735 where @xmath736 @xmath737 @xmath738 changing coordinates @xmath739 by @xmath740 and @xmath741 , we write the inner double - integral in the previous integral as @xmath742 where @xmath743 is defined as in ( [ defofmqcjtau ] ) . from ( [ estmixd11 ] ) , ( [ estmixd12 ] ) and lemma [ vdcorput2d ] , we then estimate @xmath744 by @xmath745 which clearly is bounded by @xmath746 hence ( [ estcorputphidjm ] ) follows and therefore we complete the proof .
we now turn to the proof of proposition [ prop2 ] . for simplicity , we assume @xmath46 is supported on @xmath747 $ ] . for any function @xmath748 , we have @xmath749 where @xmath750 .
the stationary phase method yields that the principle part of ( [ q1inverse ] ) is @xmath751 where @xmath752 are constants depending only on @xmath92 .
thus to obtain proposition [ prop2 ] , it suffices to prove that there exists a constant @xmath111 such that @xmath753 holds for all @xmath754 , and @xmath755 , where @xmath756 is defined to be @xmath757 observe that @xmath382 is supported essentially in a bounded interval away from @xmath758 .
thus we can restrict the variable @xmath59 in a bounded interval @xmath759 whose length is comparable to @xmath760 and reduce the problem to estimate @xmath761 holds for an absolute constant @xmath111 and all @xmath762 , where @xmath763 is equal to @xmath764 here @xmath765 let @xmath488 be an interval of length @xmath286 .
a rescaling argument then reduces ( [ q1unif23estforsub ] ) to an estimate of a bilinear form @xmath766 associated to @xmath488 , that is , @xmath767 where @xmath768 is defined by @xmath769 notice that @xmath770 where @xmath771 are constants such that @xmath772 $ ] .
@xmath773 can be dropped by utilizing fourier series since @xmath382 is a schwartz function , because @xmath774 are restricted in bounded intervals respectively .
then ( [ q1unif23estforunit ] ) can be reduced to lemma [ vdcorput2d111 ] by a change of variable @xmath775 .
therefore we prove proposition [ prop2 ] .
+ [ phasetauneg ] let @xmath388 and @xmath776 $ ] . and
let @xmath777 be defined by @xmath778 suppose that @xmath478 and @xmath779 .
then @xmath780 holds whenever @xmath781 and @xmath782 , where @xmath529 is the interval @xmath783 $ ] or @xmath784 $ ] .
clearly @xmath785 which can be written as a sum @xmath786 where @xmath787 is @xmath788 and @xmath789 is equal to @xmath790 the mean value theorem yields that @xmath791 and @xmath792 where @xmath793 and @xmath794 is a point between @xmath795 and @xmath796 .
a simple computation gives @xmath797 now ( [ largederiofphi ] ) follows from ( [ smallphi2tau ] ) , ( [ deriofgtau ] ) , @xmath478 and @xmath779
. therefore we finish the proof .
[ vdcorputneg ] let @xmath798 be bump functions supported on @xmath692 and @xmath693 respectively , where @xmath529 is @xmath799 $ ] or @xmath800 $ ] .
suppose that @xmath801 , @xmath478 and @xmath720 is a phase function defined by @xmath802 where @xmath803 are constants independent of @xmath804 such that @xmath724 .
let @xmath805 be a bilinear form defined by @xmath806 then we have @xmath807 holds for all @xmath728 and @xmath729 , where @xmath712 is a constant depending only on @xmath92 .
the bilinear form @xmath808 equals to @xmath809 , where @xmath810 is defined by @xmath811 by a change of variables , we express @xmath812 as @xmath813 where @xmath814 @xmath815 @xmath816 @xmath817 changing coordinates @xmath739 by @xmath818 and @xmath741 , we write the inner double - integral in the previous integral as @xmath819 it is clear that @xmath820 has at most finite zeros for fixed @xmath821 .
thus , by lemma [ phasetauneg ] and van der corput lemma , we obtain @xmath822 holds for @xmath779 .
thus we estimate @xmath823 by @xmath824 which is dominated by @xmath825 henceforth , ( [ estcorputneg ] ) follows and we complete the proof .
we now turn to the proof of proposition [ prop2neg ] . for simplicity , we assume @xmath46 is supported on @xmath747 $ ] . for any function @xmath826
, we have @xmath827 where @xmath750 . by the stationary phase method
, we obtain that the principle part of ( [ q1inverseneg ] ) is @xmath828 where @xmath752 are constants depending only on @xmath92 .
thus to obtain proposition [ prop2 ] , it suffices to prove that there exists a constant @xmath111 such that @xmath829 holds for all @xmath830 , and @xmath755 , where @xmath831 equals to @xmath832 by a rescaling argument , it suffices to show that @xmath833 holds for all @xmath728 and @xmath729 , where @xmath834 is defined by @xmath835 here the constant @xmath803 satisfies @xmath836 and @xmath837 is a bump function supported on @xmath718 $ ] or @xmath719 $ ] .
clearly ( [ rescalneg0 ] ) is a consequence of lemma [ vdcorputneg ] .
therefore the proof of proposition [ prop2neg ] is completed . | we establish an @xmath0 to @xmath1 estimate for the bilinear hilbert transform along a curve defined by a monomial .
our proof is closely related to multilinear oscillatory integrals . |
the recent substantial evolution in the ellipticity of rich galaxy clusters was noted by mellot , chambers & miller ( 2001 ) . this result has been supported by plionis ( 2002 ) .
in particular , he argues that the ellipticity of aco clusters is correlated with their redshifts . using , among others , 309 aco clusters taken from apm , for which about 180 have determined redshifts , he claimed the existence of a significant ellipticity - redshift dependence .
our sample consists of 246 aco clusters .
all data come from 48 schmidt telescopes .
original plates were scanned at the rome and edinburgh astronomical observatories .
the majority of galaxy catalogues were obtained from dss using focas packages . for this
sample , visual verification of object classification was made .
it is well known that ellipticity depends also on the distance from cluster centre ( e.g. carter & metcalfe 1980 , flin 1984 , trevese et al .
1992 , struble & ftaclas 1994 ) .
therefore , we decided to check the investigated relation for various distances from cluster centres .
the radius changes from 0.5 to 1.5 mpc ( @xmath0 ) with a step of 0.25 mpc from the cluster centre .
thus , we have measured ellipticities at 5 distances from the cluster centre for each of the 246 clusters .
the ellipticities have been determined in a standard way , this is from the covariance matrix .
the redshifts are taken from literature .
for each cluster we have two parameters , the calculated ellipticity of the cluster and cluster redshift .
a straight line was fitted to each set of data .
the results of the fitting procedure for line @xmath1 are given in table 1 .
r _ in mpc is the radius of the area in which the ellipse has been calculated , _ a _ and _ b _ are fit parameters and _ r _ is the correlation coefficient .
we also calculated the mean value of the five ellipticities for various _
r _ radii .
the dependence of such mean ellipticities on redshift is given in fig.1 .
the last line of table 1 gives the relevant parameters .
it is clear that we do not observe any relation between ellipticities and redshifts .
this differs greatly from the plionis ( 2002 ) result .
however , the claimed dependence of ellipticity on redshift reported by him is very weak .
our result is affected by such factors as e.g. the adopted cluster centre , the way of smoothing the distribution and the applied method of ellipticity determination .
these factors are being carefully studied now .
the current result is consistent with mulchaey & zabludoff ( 1998 ) and zabludoff & mulchaey ( 1998 ) observation reporting the lack of recent significant dynamical evolution for poor groups and cd clusters .
moreover , the lack of the eccentricity evolution is in agreement with the latest hydrodynamic simulations of clusters ( floor et al .
2003a , 2003b ) .
1980 mnras , 191 , 325 1984 in clusters and groups of galaxies , eds .
mardirossian , f. , et al . ,
dordrecht , p. 163 2003a , apj , 591 , 741 2003b , astro - ph/0307539 2001 , apj , 559 , l75 1998 , apj , 496 , 73 2002 , apjl , 572 , 67 1994 , aj , 108 , 1 1992 , a&as , 94 , 237 1998 , apj , 496 , 39 | we analysed 246 scanned aco clusters and we do not find any dependence of galaxy cluster ellipticity on redshift . |
recently , there has been a great deal of interest @xcite in studying various collective phenomena in dense quantum plasmas which are ubiquitous in a wide - range of physical systems , including compact astrophysical objects ( e.g. the cores of white dwarfs @xcite and neutron stars @xcite ) , giant planetary cores ( e.g. the interior of the planet jupiter @xcite , high - energy density compressed plasmas created by powerful laser beams @xcite , as well as semiconductors @xcite and metallic nanostructures @xcite for high - technological applications , table - top quantum free - electron lasers ( fel ) @xcite for producing coherent tunable x - rays and gamma rays , and quantum diodes @xcite , etc .
quantum plasmas are usually composed of degenerate electron and nondegenerate ion fluids . since , in quantum plasmas the number density of degenerated electrons , @xmath0 , is rather high , the average inter - electron / ion distances @xmath1 can be comparable with the electron thermal de broglie wavelength , @xmath2 , which is much smaller than the landau length @xmath3 . here where @xmath4 is the magnitude of the electron charge , @xmath5 the boltzmann constant , @xmath6 the thermal electron temperature , @xmath7 the planck constant divided by @xmath8 , @xmath9 the rest mass of the electron , and @xmath10 the electron thermal speed . in quantum plasmas ,
the wavefunctions of degenerated fermionic ingredients ( electrons ) overlap giving rise to the fermi - dirac distribution function and density of states ( dos ) .
in fact , the importance of the quantum electron degeneracy was recognized eighty years ago by fowler @xcite and chandrasekhar @xcite in the study of non - relativistic and ultra - relativistic pressures for degenerate electrons that were required for setting up the hydrostatic stability of a gravitating star that eventually becomes a compact star called a white dwarf . clearly , in quantum plasmas , the equations of state ( eos ) significantly differ from classical plasmas where non - degenerate electrons obey the ordinary maxwell - boltzmann distribution function . in quantum plasmas , the electron and ion gamma factors ( which are defined as the ratio of the coulomb potential to the thermal energy of the plasma particles )
are given as @xmath11 and @xmath12 , respectively , being smaller and much larger than unity , implying a weak correlation for degenerate electrons and strong correlations for non - degenerate ion fluids in the quasineutral plasma regime .
the collective behavior of a quantum fermi plasmas with a statistical ensemble of degenerate electrons has been investigated in the pioneering papers of klimontovich and silin @xcite , and later by bohm and pines @xcite , who have used the quantum kinetic wigner and poisson equations to obtain the dispersion relation for quantum electron - plasma oscillations ( epos ) .
the frequency spectra of the latter can be shown to be , @xmath13 , where @xmath14 is the electron plasma frequency and @xmath15 is the fermi electron speed .
the frequency spectra of quantum epos have also been derived @xcite from the quantum electron fluid equations @xcite , composed of the electron continuity , non - relativistic electron momentum equation including the quantum statistical pressure and the quantum recoil effect , and poisson s equation .
on the other hand , in his classic experimental work , glenzer _ et al .
_ @xcite have conclusively demonstrated the quantum mechanical collective effect by measuring the scattering of electromagnetic waves off the quantum epos in a high - energy density ( warm dense matter ) laboratory plasma .
numerical analyses of poisson s equation and the nonlinear schrdinger equation for non - relativistic degenerate electron fluids revealed the formation of a dark soliton and two - dimensional quantized vortices @xcite , as well as nano - structures @xcite in quantum electron plasmas with immobile ions . on the other hand ,
inclusion of the ion dynamics leads to electrostatic oscillations ( esos ) supported by the restoring force due to the quantum statistical pressure , the quantum bohm force and the inertia of non - degenerate ions .
by using poisson s equation , the momentum equation for non - relativistic degenerate inertialess electrons and the classical hydrodynamic equations ( the continuity and momentum equations ) for non - correlated ions , hass _ et al _ @xcite derived a modified korteweg - de vries ( kdv ) equation for small amplitude nonlinear ion oscillations in an unmagnetized quantum plasma .
the kdv equation admits a localized solitary pulse as a stationary solution .
large amplitude localized ion oscillations in a collisionless quantum plasma were investigated by many authors @xcite ignoring the ion correlation effects . the latter and the ion fluid viscosity
have recently been incorporated in the small amplitude shock wave theory @xcite of ion oscillations in a quantum plasma . in this paper , we present a theory for fully nonlinear nonstationary es ion shock waves in a quantum viscous plasma with mildly coupled degenerate electrons and strongly correlated ion fluids for a wide - range of the chandrasekhar s relativity parameter and the plasma atomic number . for our purposes , we use the hydrodynamic force balance ( the electric force balancing the gradient of the degenerate electron pressure , gradient of the potentials involving the electron - exchange and electron - correlations , the gradient of the bohm potential ) for inertialess electrons and the generalized viscoelastic ion momentum ( gvim ) equation , together with the ion continuity equation , invoking the quasi - neutrality approximation .
the gvim equation contains the essential physics of the ion correlation decay rate , as well as of the ion fluid bulk and shear viscosities .
when the electric force is eliminated from the gvim equation by using the electric field , from the momentum equation for inertialess degenerate electron fluids , we obtain a pair of equations exhibiting the coupling between the ion fluid velocity and the ion number density .
the pair of equations are then numerically solved to study the dynamics of is ion shocks in planar as well as non - planar geometries .
our numerical analysis exhibits distinguished features for the formation of es shock structures in different ion mass - density ranges which might play a very important role in acceleration of charged particles in dense quantum plasmas , such as those in the cores of white dwarf stars and in compressed plasmas created by intense laser and electron beams .
we consider a fully - ionized quantum plasma composed of strongly correlated non - degenerate warm ion and mildly coupled degenerate electron fluids . since , the ion thermal energy is much smaller than the electron fermi - electron energy , we shall not consider quantum forces on the ion fluid .
the viscoelastic hydrodynamic ( vehd ) model for such an electron - ion plasma is governed by a set of equations @xcite @xmath16
\\ = \left [ { { \eta _ v}{\nabla ^2}{{\bf{u}}_i } + \left ( { \xi + \frac{{{\eta _ v}}}{3 } } \right)\nabla ( \nabla \cdot{{\bf{u}}_i } ) } \right ] , \hspace{3mm}\delta { \bf{u } } = { { \bf{u}}_i } - { { \bf{u}}_e}\\0 = - e{n_e}{\bf{e } } - \nabla { p_{tot } } + { m_e}{n_e}{\nu _ { ei}}\delta { \bf{u } } , \\ \end{array}\ ] ] where @xmath17 , @xmath18(@xmath19 ) , @xmath20 , and @xmath21 are the viscoelastic relaxation time , electron - ion ( ion - electron ) collision frequency , shear viscosity , and the ion charge state , respectively .
also , the total electron pressure , @xmath22 , consists of pressures due to the electron degeneracy , coulomb interaction , and electron - exchange interaction , respectively . the electron viscosity has been shown to become dominant only at much larger densities , such as in neutron - star crusts @xcite , hence , we will ignore it in the following analysis .
it is also known that in quantum plasmas the electron - electron collisions are rare due to the pauli blocking - mechanism @xcite .
therefore , it is convincing to only consider the ion - fluid viscosity effects in our model .
we simplify the model by supposing that degenerate electrons are inertialess and the plasma is quasi - neutral , viz . @xmath23 .
thus , we have @xmath24 , and obtain @xmath25 = \left [ { { \eta _ v}{\nabla ^2}{\bf{u } } + \left ( { \xi + \frac{{{\eta _ v}}}{3 } } \right)\nabla ( \nabla \cdot{\bf{u } } ) } \right ] , \\ \end{array}\ ] ] where , @xmath26 is the generalized plasma pressure .
moreover , it is convenient to express all involved parameters in terms of the chandrasekhar s relativity parameter , @xmath27 , where @xmath28 gr/@xmath29 and @xmath30 are the normalizing plasma mass - density and number of nucleons per electron with @xmath31 and @xmath21 being the atomic weight and number , and @xmath32 is the electron relativistic fermi momentum .
other quantities can be expressed in terms of the relativity parameter @xmath33 and the plasma atomic - number @xmath21 .
for instance , for the electron degeneracy pressure , we have @xcite @xmath34,\ ] ] while , for the coulomb interaction pressure in the spherical wigner - seitz cell approximation , we have @xcite @xmath35{r^4},\hspace{3mm}\alpha = \frac{{{e^2}}}{{\hbar c } } \simeq \frac{1}{{137}}.\ ] ] finally , for the electron exchange interaction @xcite , one may write @xmath36 } \right\ } , \end{array}\ ] ] where , @xmath37 . on the other hand ,
the ion - viscosity is @xcite @xmath38 [ figure1 ] , and the plasma ion fluid - viscosity , @xmath39 with respect to different mass - density ( relativistic - degeneracy parameter , @xmath40 ) and the atomic - number , @xmath21 .
the curve thickness in plots indicate the increase in the varied parameter in each plot.,title="fig : " ] where , @xmath41 is the coupling parameter with @xmath42 being the wigner - seitz spherical - cell radius containing only one electron on average .
variations of the parameters @xmath43 and @xmath44 are given in ref .
@xcite for various coupling parameter values .
according to dewitt @xcite , the strongly coupled plasma is in liquid form for the range of the plasma ion coupling of @xmath45 to which we focus our attention in the following analysis .
the plasma coupling factor , in terms of the relativity parameter for a typical white dwarf temperature of @xmath46 is approximately @xmath47 , valid for a wide ange of the plasma mass - density and atomic - number variations .
the viscoelastic relaxation - time is directly related to the plasma viscosity via the following relation @xcite @xmath48 where , @xmath49 is the free electron effective - mass .
the viscoelastic relaxation time is thus given as the following , with the fitting constants presented in ref .
@xcite @xmath50 furthermore , ignoring the ion pressure compared to that of the electrons , we may write the plasma generalized pressure in terms of a generalized effective potential @xcite @xmath51 with the effective plasma potential , defined as @xmath52 where , @xmath53 . the dimensionless set of vehd equations , eqs .
( [ scalar ] ) , can be obtained by using the following scalings @xmath54 where , the normalizing factors , @xmath55 , @xmath56 ( with @xmath57 and @xmath58 ) , and @xmath59 , denote the equilibrium plasma mass - density , characteristic ion plasma frequency and the plasma sound speed , respectively .
thus , the normalized equations , dropping the bar notation for simplicity , read @xmath60 = \eta\left [ { { \nabla ^2}{\bf{u } } + \frac{1}{3}\nabla ( \nabla \cdot{\bf{u } } ) } \right ] . \\
\end{array}\ ] ] where , @xmath61 and @xmath62 $ ] .
it should be noted that , at very high densities , such as in the neutron star crust region , the electron - electron collisions can be very important dominating the electron viscosity over that of the ions , the consideration of which is beyond the scope of the present investigation .
equations ( [ dimensionless ] ) can be considered as the most general model for an unmagnetized viscoelastic fully interacting quantum plasma , valid for a wide- range of the plasma mass density , atomic - number , and the ion - fluid coupling factor .
the quantum plasma parameters range of interest covering the planet interiors and white dwarfs are approximately , @xmath63 , @xmath64 and @xmath65 .
the quantum fluid model , following the closed equation - set relevant to different plasma geometries for the one - dimensional ( planar ( @xmath66 ) , cylindrical ( @xmath67 ) , and spherical ( @xmath68 ) ) coordinate system , can be written as @xmath69 = 0 , \\
\left [ { 1 + \tau ( \rho , z)\frac{d}{{dt } } } \right]\left [ { \rho ( r , t)\left ( { \frac{{du(r , t)}}{{dt } } + \frac{{\partial { \psi _ g}(\rho , z)}}{{\partial r } } } \right ) } \right ] \\ = \eta ( \rho , z)\left [ { \frac{1}{{{r^\nu}}}\frac{\partial } { { \partial r}}\left ( { { r^\nu}\frac{{\partial u(r , t)}}{{\partial r } } } \right ) + \frac{1}{3}\frac{\partial } { { \partial r}}\left ( { \frac{1}{{{r^\nu}}}\frac{{\partial { r^\nu}u(r , t)}}{{\partial r } } } \right ) } \right ] . \\ \end{array}\ ] ] in the following sections , we will evaluate the characteristics of the viscoelastic es nonlinear excitations in terms of fractional fermi - dirac plasma parameters . note that , in our model , we have ignored the bulk viscosity , @xmath70 , compared to that of shear , @xmath39 , due to negligible contributions from compressibility of high pressure plasma fluids .
figure 1 displays the variations of ( a ) the normalized viscoelastic relaxation time , @xmath71 , and ( b ) the normalized shear ion fluid - viscosity , @xmath39 with respect to the plasma degeneracy parameter , @xmath40 , for different plasma atomic - number , @xmath21 .
it is observed from fig .
1(a ) that for all atomic - number values the viscoelastic relaxation time increases with the increase of the plasma number density , related to the relativistic degeneracy parameter @xmath40 , and this increase is smaller for larger values of plasma atomic - number , @xmath21 .
figure 1(b ) , on the other hand , reveals that the plasma ion - viscosity first increases with the increase of the plasma number density up to a maximum value and then decreases again .
this behavior is observed for all plasma atomic - number value being more pronounced for plasmas with lower atomic numbers .
it is also remarked that the relative value of the ion fluid - viscosity parameter for a given plasma mass density is always lower for heavier element composed plasmas .
[ figure2 ] ) and the atomic - number @xmath21 . the curve thickness in plots indicates the increase in the varied parameter in each plot.,title="fig : " ] important information can be obtained considering the plasma response to linear perturbations .
this is accomplished by the fourier analysis of eq .
( [ simp ] ) , and by using the harmonic operators @xmath72 and @xmath73 .
accordingly , we have the following first order relations @xmath74 where @xmath75 from eqs .
( [ rel ] ) we readily obtain the generalized dispersion relation @xmath76 figure 2 depicts the linear dispersion of viscoelastic ion waves in a relativistically degenerate plasma and the variation of the phase - speed for different plasma mass density and atomic - numbers .
figure 2(a ) reveals that the linear wave dispersion is fundamentally governed by the plasma mass density . below a critical plasma density ,
a dispersion gap opens in the wave - number range , indicated for the thin curve in the plot .
as the plasma density increases above the critical value , for a fixed plasma composition , the dispersion gap disappears , but there still exists a sharp discontinuity in the linear wave speed . also , as the plasma mass - density increases further this discontinuity disappears .
such anomalous feature may be attributed to the interplay of strong plasma interactions and ion fluid couplings expected for a super - dense degenerate plasma .
figure 2(b ) shows the effect of the plasma composition on the linear wave dispersion .
it is evident that in the small and long - wavelength limits , the linear wave speed is not affected by the change to plasma composition , while this is converse in the medium - wavelength region due to the presence of the mentioned discontinuity in the plasma wave speed .
[ figure3 ] ) collisional quantum plasma for a given plasma number density ( @xmath77 ) and atomic - number ( @xmath78).,title="fig : " ] [ figure4 ] [ figure5 ] ) and the atomic - number @xmath21 .
each row in the figure indicatess the effect of the variation in atomic number on the shock height and its steepness and each column depicts the effect of the change in the ion mass density on the shock wave characteristic parameters.,title="fig : " ] [ figure6 ] ) , cylindrical ( @xmath67 ) , and spherical ( @xmath68 ) .
plot ( d ) reveals that the shock height increases as one goes from spherical to cylindrical and from cylindrical to planar geometry.,title="fig : " ] it is evident that the generalized set of vehd equations , eqs .
( [ simp ] ) , do not admit analytical arbitrary amplitude solutions .
therefore , at this stage , we aim at numerical investigation of arbitrarily large amplitude nonlinear waves which can have fundamental importance in superdense astrophysical environments . in order to evaluate the nonlinear blast - waves in the laboratory frame , we solve the time - dependent vehd equations compliant with appropriate periodic boundary conditions for different independent fractional plasma parameter values , namely , the relativistic degeneracy parameter @xmath40 ( representing the plasma mass density , @xmath79 ) , and the plasma composition parameter , @xmath21 . for the the planar geometry ( viz .
@xmath66 ) a gaussian pulses of the form @xmath80 with an initial speed profile of @xmath81 is placed at the origin in initial time .
the time evolution of the plasma mass - density perturbation is then rendered to future times based on the governing differential equation set , eq .
( [ simp ] ) , for given values of the plasma fractional parameters .
it is further assumed that the spatial variations in the plasma mass - density vanishes at the plasma boundaries in all future times . in our model
, we also incorporate the effects of time - dependence of the viscoelastic parameters due to the plasma mass - density variation , for completeness .
the results of simulation are depicted in figs . 3 and 4 for non - viscous ( @xmath82 ) and viscous ( @xmath83 ) quantum plasmas , respectively , for given initial plasma mass - density , temperature and composition relevant to typical white dwarf stars and cores of giant planets .
figure 3 indicates the time evolution of ion density wave into the train of density humps with the time elapse .
this is an example of ion wave breaking / steepening , arising from harmonic generation nonlinearity , where , the plasma dispersive effects are somewhat spreading the ion pulse , but not strong enough to allow the formation of a real solitary pulse .
these ion density pulses arise from balance between harmonic nonlinearity and the quantum statistical pressure and interaction effects in the limit of vanishing plasma viscosity and the corresponding relaxation time . for the case of strongly coupled quantum plasma the value of the plasma coupling factor , @xmath84 ,
is kept below @xmath85 in our simulations to ensure that the quantum plasma is in the liquid phase @xcite .
it is observed from fig .
4 that the time evolution of the gaussian density and velocity profiles develop into shock - like structures with pronounced shock wave front steepened by further elapse of the simulation time .
it is confirmed by fig .
5 that for all the considered set of plasma parameters the gaussian perturbation profile develops into shock - like structures after enough elapsed simulation time , say ( @xmath86 ) . comparing different plots in fig . 5
, we observe that an increase in both the plasma mass density and the atomic - number leads to an increase in the shock height and steepening of the shock front .
remember that the maximum value of the relativistic degeneracy parameter , @xmath87 used in this simulation , corresponds to high electron / ion number density , @xmath88 , expected for white dwarf stars .
the present simulation revealed that despite the significant effect of the temperature on the plasma coupling factor , @xmath84 , it has little effect on the viscosity parameters and the present study might be extended to the ( jupiter - like ) large planet cores and environments with high number densities and relatively low temperatures .
figure 6 depicts the evolution of ion - momentum distribution from an initial gaussian profile in different plasma geometries .
it is seen that in planar ( @xmath66 ) , cylindrical ( @xmath67 ) and spherical ( @xmath68 ) geometries , there exist large amplitude shock waves . in producing fig .
6(a - c ) , we have assumed an initial pulse shape as @xmath89 $ ] and @xmath90 $ ] in order to avoid singularities present at the origin for non - planar geometries . from fig .
6(d ) we observe that the amplitude of nonlinear ion shock structures are relatively higher in the planar geometry with respect to other geometries , for a given set of fixed plasma parameters .
different geometries than planar , like spherical and cylindrical , may be important in the study of physical mechanisms in stellar configurations and active galactic nuclei , for instance .
the present findings reveal the fundamental effects of the plasma mass density and its composition on the formation and characteristics of blast - like ion - waves in dense quantum plasmas .
furthermore , our study reveals that the characteristics of density perturbations and the formation of es ion shocks in the cores of white dwarf stars can be quite different from the existing shock - waves at the low density crust region .
in this paper , we have presented a nonlinear theory for arbitrary large amplitude electrostatic planar and non - planar shocks in collisional quantum plasmas composed of mildly coupled degenerate electron fluid and strongly correlated non - degenerate ions for a wide range of the plasma mass density and atomic number .
we have included the electrostatic and relevant pressure gradient , quantum forces and interactions on degenerate electron fluids , and have used the general vehd equations for non - degenerate ions to carry out laboratory - frame simulations in different geometries in order to deduce the complete picture of very large - amplitude electrostatic shock structures .
the latter may exist in white - dwarfs , the active galactic nuclei or many other high density astrophysical environments such as neutron stars , pulsars and magnetars .
it was found that that the increase / decrease in the ion charge state or atomic - number / mass - density would lead to stronger and steeper shock fronts in the relativistically degenerate viscous plasmas .
it was also shown that the linear dispersion in such quantum plasmas exhibits anomalous features attributed to the strong fermionic interactions and ion fluid couplings . in conclusion , we stress that the present investigation of the formation of arbitrary large amplitude es ion shocks will play a very important role in acceleration of electrons and ions in relativistically degenerate quantum plasmas , which are ubiquitous in high - energy density compressed plasmas produced by intense short wavelength laser beams , as well as in planetary systems ( e.g. the giant jupiter ) and in super dense compact astrophysical objects ( e.g. white dwarf stars and magnetars ) .
g. manfredi , fields inst .
commun . * 46 * , 263 ( 2005 ) .
p. k. shukla and b. eliasson , phys .
lett . * 96 * , 245001 ( 2006 ) .
g. brodin and m. marklund , phys .
e * 76 * , 055403(r ) ( 2007 ) .
g. brodin , m. marklund , and g. manfredi , phys .
lett . * 100 * , 175001 ( 2008 ) .
p. k. shukla and b. eliasson , phys .
usp . * 53 * , 51 ( 2010 ) .
p. k. shukla and b. eliasson , rev .
phys . * 83 * , 885 ( 2011 ) .
s. v. vladimirov and yu o. tyshetskiy , phys .
usp . * 54 * , 1243 ( 2011 ) .
f. haas , _ quantum plasmas : an hydrodyanmic approach _
( springer , new york , 2011 ) .
s. chandrasekhar , philos . mag . * 11 * , 992 ( 1931 ) .
s. chandrasekhar , mon . not .
soc . * 95 * , 207 ( 1935 ) .
m. chaichian , s. s. masood , c. montonen , a. prez martnez , and h. prez rojas , phys .
lett . * 84 * , 5261 ( 2000 ) . v. e. fortov , phys .
usp . * 52 * , 615 ( 2009 ) .
s. h. glenzer , o. l. landen , p. neumayer , r. w. lee , k. widmann , s. w. pollaine , phys .
lett . * 98 * , 065002 ( 2007 ) : s. h. glenzer and r. redmer , rev .
* 81 * , 1625 ( 2009 ) .
p. a. markowich , c. a. ringhofer , and c. schmeiser , _ semiconductor equations _
( springer - verlag , new york , 1990 ) . c. gardner , siam , j. appl
. math . * 54 * , 409(1994 ) .
n. crouseilles , p. a. hervieux , and g. manfredi , phys .
b * 78 * , 155412 ( 2008 ) . b. eliasson and p. k. shukla , phys
e * 85 * , 065401(r ) ( 2012 ) .
l. k. ang , t. j. t. kwan , and y. y. lau , phys .
* 91 * , 208303 ( 20030 ; l. k.
ang and p. zhang , _ ibid . _ * 98 * , 164802 ( 2007 ) .
r. h. fowler , mon . not .
. soc . * 87 * , 114 ( 1929 ) .
y. l. klimontovich and v. p. silin , dokl .
nauk ssr * 82 * , 361 ( 1952 ) ; zh .
* 23 * , 151 ( 1952 ) ; in _ plasma physics _ , ed .
j. e. drummond ( mcgraw hill , new york , 1961 ) , p. 35 .
d. bohm and d. pines , phys . rev . * 92 * , 609 ( 1953 ) .
g. manfredi and f. haas , phys .
b * 64 * , 075316 ( 2001 ) .
h. wilhelm , z. phys .
* 241 * , 1 ( 1971 ) .
d. shaikh and p. k. shukla , phys .
lett . * 99 * , 125002 ( 2007 ) .
f. haas , l. g. garcia , j. goedert , and g. manfredi , phys .
plasmas * 10 * , 3858 ( 2003 ) .
p. k. shukla and b. eliasson , j. plasma phys . * 75 * , 5(2008 ) .
b. eliasson and p. k. shukla , euro phys . lett . * 97 * , 15001 ( 2012 ) . m. akbari - moghanjoughi , astrophys .
space sci .
, * 322 * , 187 ( 2010 ) .
s. ghosh , n. chakarabarty , and p. k. shukla , phys .
plasmas * 19 * , 072123 ( 2012 ) .
n. itoh , y. kohyama , and h. takeuchi , apj * 317 * , 733 ( 1987 ) .
s. chandrasekhar , _ an introduction to the study of stellar structure _ , ( chicago university press , chicago , 1939 ) .
e. e. salpeter , astrophys .
j. * 134 * , 669 ( 1961 ) .
r. nandkumar and c. j. pethick , mon . not .
soc . * 209 * , 511 ( 1984 ) .
w. l. slattery , g. d. doolen , and h. e. dewitt , phys .
a. * 26 * , 2255 ( 1982 ) . m. akbari - moghanjoughi , phys .
plasmas * 19 * , 042701 ( 2012 ) . | we present a generalized nonlinear theory for large amplitude electrostatic ( es ) ion shocks in collisional quantum plasmas composed of mildly coupled degenerate electron fluids of arbitrary degeneracy and non - degenerate strongly correlated ion fluids with arbitrary atomic number . for our purposes ,
we use the inertialess electron momentum equation including the electrostatic , pressure gradient and relevant quantum forces , as well as a generalized viscoelastic momentum ( gvem ) equation for strongly correlated non - degenerate ions .
the ion continuity equation , in the quasi - neutral approximation , then closes our nonlinear system of equations .
when the electric field is eliminated from the gvem equation by using the inertialess electron momentum equation , we then obtain a generalized gvem and the ion continuity equations exhibiting nonlinear couplings between the ion number density and the ion fluid velocity . the pair of nonlinear equations is numerically solved to study the dynamics of arbitrary large amplitudes planar and non - planar es shocks arising from the balance between harmonic generation nonlinearities and the ion fluid viscosity for a wide range of the plasma mass - density and the ion atomic - number that are relevant for the cores of giant planets ( viz .
the jupiter ) and compact stars ( viz .
white dwarfs ) .
our numerical results reveal that the es shock density profiles strongly depend on the plasma number density and composition ( the atomic - number ) parameters .
furthermore , density perturbations propagate with mach numbers which significantly dependent on the studied plasma fractional parameters .
it is concluded that the dynamics of the es shocks in the super - dense degenerate plasma is quite different in the core of a white dwarf star from that in the lower density crust region . |
as we know , a charged particle walking through a region where only magnetic flux exists acquires a phase .
we call this aharonov - bohm effect ( abe ) .
since aharonov and bohm [ 1 ] proposed an experiment to assess the manifestations of the electromagnetic potential in the quantum mechanics in 1959 , some work [ 2 - 4 ] has been investigated by researchers , and they concluded that all observable phenomena are modulated by the magnetic flux through region , and that these phenomena with respect to @xmath0 demonstrate period with magnetic flux quantum @xmath1 . with the advance of lithography and semiconductor growth techniques , the nanoring - like devices [ 5 - 8 ] have been realized experimentally . due to unique character of the nanoring
, the self - assembled quantum ring has become the best ideal candidate to observe aharonov - bohm ( a - b ) oscillation [ 9 - 11 ] . in 1995 , chaplik [ 12 ] first put forward predication that the neutral exciton can exhibit a - b oscillator behavior in one - dimensional quantum ring . and
subsequently similar results were found in the literature [ 13 ] where a short - range interaction potential has been employed . over the last decade ,
some researchers had paid their attention to a - b effect of neutral exciton in two - dimensional quantum ring .
for instance , hu @xmath2 @xmath3 [ 9 ] theoretically studied the magnetic field effects on exciton in an @xmath4 nanoring , they found that the a - b effect of exciton only exists in a finite ( but small ) width nanoring in the presence of coulomb correlation .
ribeiro @xmath2 @xmath3 [ 14 ] experimentally presented a magneto - photoluminescence study of type - ii @xmath5 self - assembled quantum dot , which revealed that a - b type oscillations for neutral excitons when angular momentum of hole ground state ranges from @xmath6 to @xmath7 , @xmath8 , and @xmath9 .
dias da silva @xmath2 @xmath3 [ 15 ] reported impurity effects on the a - b optical signatures of neutral quantum ring magnetoexcitons .
their results , on the one hand , show that energy and oscillator strength of neutral excitons are strongly modulated by the magnetic field . on the other hand ,
scattering impurity enhances the photoluminescence ( pl ) intensity on otherwise `` dark '' magnetic field windows and nonzero pl emission appears for a wide magnetic field range even at zero temperature . recently ding @xmath2 @xmath3 [ 10 ]
has investigated the magneto - pl study of a single self - assembled semiconductor quantum ring which is realized by @xmath10 @xmath11 @xmath12 etching [ 16 ] and experimentally observed the a - b type oscillations in the photoluminescence energy . as mentioned above ,
studies about a - b effect are mainly focused on energy and oscillator strength of an exciton in a nanoring .
the research into a - b effect on nonlinear optical properties has been rarely reported so far .
understanding nonlinear optical properties of an exciton in a nanoring is of significance for designing optical devices .
hence , it is of interest to study the nonlinear optical properties of an exciton in a nanoring . in the present work
, we will focus on studying nonlinear optical absorption and nonlinear optical rectification of an exciton in a nanoring .
the paper is organized as follows : in section 2 we describe the model and theoretical framework , section 3 is dedicated to the results and discussions , and finally , our conclusions are given in section 4 .
here we are concerned with a nanoring with winternitz - smorodinsky confinement potential . in the presence of a magnetic field
applied along z - direction , hamiltonian of single article reads @xmath13 ^{2}+v(\textbf{r}),\ ] ] in eq.(1 ) , @xmath14 is effective mass , @xmath15 is the electron charge , @xmath16 is the speed of light in vacuum , @xmath17 is magnetic vector potential .
@xmath18 is given by reference [ 17 ] @xmath19 it should be noted that @xmath20 and @xmath21 are the confinement potential parameters which are connected with the outer radius @xmath22 and inner radius @xmath23 by the following relations [ 17 ] @xmath24 @xmath25 where for electron @xmath26 and for hole @xmath27 . when the magnetic vector potential @xmath17 is chosen as @xmath28 and applying the polar coordinates ,
the hamiltonian can be rewritten as @xmath29\nonumber \\ + i\frac{eb\hbar}{2\mu c}(\frac{\partial}{\partial \varphi}+i\frac{\phi}{\phi_{0}})-\frac{e^{2}b^{2}r^{2}}{8\mu c^{2}}+v(\textbf{r}),\ ] ] where @xmath30 is magnetic flux quantum . according to schr@xmath31dinger equation @xmath32
we can obtain wave function and energy spectrum of single as follows : @xmath33}{2^{m+1}n_{r}!(\gamma[m+1])^{2}\pi}}(\frac{r}{\lambda})^{m}e^{-\frac{r^{2}}{4\lambda^{2}}}f(-n_{r},m + 1;\frac{r^{2}}{2\lambda^{2}})e^{im\varphi},\ ] ] @xmath34 where @xmath35 in eq .
( 7 ) , @xmath36 is the confluent hypergeometric function . @xmath37
is the radial quantum number , @xmath38 is the magnetic quantum number . with @xmath39 and @xmath40 denoting the effective cyclotron frequency and effective magnetic length , respectively .
for an exciton in a nanoring , the hamiltonian can be written as @xmath41 here we assume that electron and hole are confined by the same potential . and the last term in right - hand of eq .
( 12 ) is assumed to be perturbation . according to perturbation method ,
we can obtain the correction from coulomb interaction to energy[17 ] .
the oscillator strength [ 18 ] is a very important physical quantity in the study of the optical properties which are related to the electronic dipole - allowed transitions . generally , the oscillator strength @xmath42 is defined as @xmath43 where @xmath44 denotes the energy difference between the initial state and the final state , and @xmath45 is the electric dipole moment of the transition from the @xmath46 state to the @xmath47 state .
based on the density matrix approach and the perturbation expansion method , the second - order nonlinear optical rectification coefficient is given by [ 19,20 ] @xmath48(\frac{t_{1}}{t_{2}}-1))}{\varepsilon_{0}[(e_{fi}-\hbar\omega)^{2}+(\frac{\hbar}{t_{2}})^{2}][(e_{fi}+\hbar\omega)^{2}+(\frac{\hbar}{t_{2}})^{2}]}.\ ] ] in the present work , we also performed calculation of the intensity - dependent resonant peaks of the nonlinear optical absorption and nonlinear optical rectification coefficient which are given by reference [ 21 ] @xmath49 and @xmath50 where @xmath51 is the vacuum permittivity , @xmath52 is the refractive index of semiconductor .
@xmath53 denotes the electron density in the nanoring . in addition , @xmath54 and @xmath55 are the longitudinal and the transverse relaxation times , respectively . in the calculation
, we set the relaxation time at @xmath56 , @xmath57 .
our calculations are performed for @xmath58 nanoring .
the parameters [ 17,22 ] chosen in this work are the followings : @xmath59 , @xmath60 where @xmath61 is the free electron mass .
@xmath62 , @xmath63 , @xmath64 . in fig
1 , we plotted transition energy as a function of inner radius @xmath23 , outer radius @xmath22 , respectively , for two different magnetic field @xmath65 , @xmath66 . from this figure
, we can find that transition energy has a continuous increase with respect to inner radius for a fixed outer radius , while that transition energy is reduced for the fixed inner radius as outer radius increases .
here we define width of a nanoring by the difference in size between outer radius and inner radius @xmath67 .
it should be noted that transition energy is not almost affected by the variations of inner or outer radius when @xmath68 is greater than @xmath69 .
also it can be easily seen that transition energy suddenly becomes very large when @xmath68 is very narrow ( around @xmath70 ) .
the physical origins , for these behaviors , can be interpreted in the following way . for wider ring ,
exciton moving in the ring can be considered to be two - dimensional , while for narrower ring , one can take exciton as one - dimensional .
two - dimensional nanoring evolves into one - dimensional when width of the nanoring continuously decreases .
in addition , in fig .
1 , the magnetic field has no influence on the transition energy of an exciton .
2 shows oscillator strength as a function of outer radius @xmath22 , inner radius @xmath23 , respectively , for two different magnetic field @xmath65 , @xmath66 .
we find from this figure that oscillator strength with respect to inner or outer radius has similar behaviors as shown in fig .
the reasons are that there is a greater probability of transition between the initial and the final state when exciton moves in a narrower ring . in fig .
3 , oscillator strength as a function of magnetic flux @xmath0 with @xmath71 and @xmath72 .
it is easily seen from this figure that oscillator strength decreases with increasing @xmath0 .
but the aharonov - bohm oscillations of exciton is not observed in the nanoring .
we can provide this phenomenon with the reason .
electron and hole are confined to the same potential in the nanoring and move over same path , leading to zero in electric dipole moment [ 23 ] . magnetic flux through nanoring
has an effect on wave function of electron and hole .
hence , the probability of transition between different states in a nanoring is dependent on magnetic flux . in fig .
4 , nonlinear optical absorption was presented as a function of inner radius @xmath23 , outer radius @xmath22 , respectively , for magnetic field @xmath65 .
we see that nonlinear optical absorption is not sensitive to the variations of inner or outer radius for the wider ring , while that nonlinear optical absorption instantly become very huge when @xmath68 is reduced to be quite narrow and can be taken as one - dimensional . and fig .
5 demonstrates that nonlinear optical absorption with respect to magnetic flux @xmath0 with three different widths of a ring .
it is clearly seen that effect of magnetic flux on the nonlinear absorption can be tuned by the width of a ring .
so we can tune the outer and inner radius to meet special needs in experiment , which is very significant in designing devices . in fig . 6 ,
nonlinear optical rectification was plotted as a function of photon energy with magnetic field for three different @xmath68 .
we note from this figure that the magnitude of resonant peak of nonlinear rectification increases as @xmath68 increases . also , the position of resonant peak of nonlinear optical rectification has a blue shift , which is modulated by width of a ring .
in addition , we also observe that the magnitudes of resonant peak are far less than results from literature [ 21 ] where parabolic confinement potential is employed to confine electron and hole . and
this indicates that the winternitz - smorodinsky potential is not realistic in crystal .
7 shows nonlinear optical rectification as a function of inner radius @xmath23 , outer radius @xmath22 , respectively , for magnetic field @xmath65 .
when outer radius is set at @xmath73 , nonlinear optical rectification is negative and decreases with inner radius ranging from @xmath74 to @xmath75 .
and then the nonlinear rectification monotonously increases with increasing inner radius .
in addition , we also see that the nonlinear optical rectification monotonously decreases with outer radius for a fixed @xmath23 . finally , we shows nonlinear optical rectification as a function of magnetic flux @xmath0 for @xmath76 and @xmath77 in fig .
8 . from this figure
, we can see that the nonlinear rectification increases with respect to @xmath0 . and
no a - b oscillation is observed in this figure .
we have investigated the effects of magnetic flux , size of quantum ring on transition energy , oscillator strength , nonlinear optical absorption and nonlinear optical rectification of exciton in a nanoring with winternitz - smorodinsky confinement potential .
our results show that transition energy , oscillator strength , nonlinear optical absorption and nonlinear optical rectification are strongly affected by width of a nanoring , and that one can control these properties of a nanoring by modulating the outer or inner radius to meet special needs in designing electro - optical devices .
in addition , it is found that the magnetic flux can tune transition energy , oscillator strength , nonlinear optical absorption and nonlinear optical rectification of a nanoring . in conclusion ,
the properties of an exciton in nanoring are strongly dependent on width of a ring and magnetic flux .
finally we hope our research can stimulate further studies in nonlinear optical properties of a nanoring .
this work is supported by national natural science foundation of china(under grant no .
11074055 ) . 0.5 cm 39 y.aharonov and d.bohm , phys . rev .
115 , 485 ( 1959 ) .
n. byers and c.n .
yang , phys .
lett . 7 , 46 ( 1961 ) . f. bloch , phys . rev
21 , 1241 ( 1968 ) .
m. b@xmath78ttiker , y. imry and r. landauer , phys .
96a , 365 ( 1983 ) .
g. timp , a.m. chang , j.e .
cunningham , t.y .
chang , p. mankiewich , r. behringer and r.e .
howard , phys .
58 , 2814 ( 1987 ) .
ford , t.j .
thornton , r. newbury , m. pepper , h. ahmed , c.t .
foxon , j.j .
harris and c. roberts , j. phys .
c 21 , l325 ( 1988 ) .
a. fuhrer , s. l@xmath78sher , t. ihn , t. heinzel , k. ensslin , w. wegsheider and m. bichler , nature ( london ) 413 , 822 ( 2001 ) ; microelectron .
63 , 47 ( 2002 ) .
yau , e.p.de poortere and m. shayegan , phys .
88 , 146801 ( 2002 ) .
hui hu , jia - lin zhu , dai - jun li and jia - jiong xiong , phys . rev .
b 63 , 195307 ( 2001 ) .
f. ding , n. akopian , b. li , u. perinetti , a. govorov , f.m .
peeters , c.c .
bof bufon , c , deneke , y. h. chen , a.rastelli , o.g .
schmidt and v.zwiller , phys . rev .
b 82 , 075309 ( 2010 ) .
r@xmath31mer and m.e .
raikh , arxiv : cond - mat/9906314v2 ( 2000 ) .
a. chaplik , pisma zh .
@xmath79ksp , teor .
62 , 855 ( 1995 ) [ jetp lett .
62 , 900 ( 1995 ) ] .
r@xmath31mer and m.e .
raikh , phys .
b 62 , 7045 ( 2000 ) .
e. ribeiro , a.o .
govorov , w. carvalho , jr . and g. medeiros - ribeiro , phys .
92 , 12 ( 2004 ) .
dias da silva , s.e .
u@xmath80oa and a.o .
govorov , phys . rev .
b 70 , 155318 ( 2004 ) .
f. ding , l. wang , s. kiravittaya , e. m@xmath78ller , a. rastelli and o.g .
schmidt , appl .
90 , 173104 ( 2007 ) .
a. k. atayan , e. m. kazaryan , a. v. meliksetyan , h. a. sarkisyan , journal of contemporary physics .
45 , 126 ( 2010 ) .
w.f . xie , phys .
e 43 , 49 ( 2010 ) .
s. baskoutas , e. paspalakis , a.f .
terzis , phys . rev .
b 74 , 153306 ( 2006 ) .
e. rosencher , ph .
bois , phys .
b 44 , 11315 ( 1991 ) .
m. zalu@xmath81ny , j. appl .
74 , 4716 ( 1993 ) .
w. f. xie , journal of luminescence 131 , 943 ( 2011 ) .
a. o. govorov , s. e. ulloa , k. karrai , and r. j. warburton , phys .
b 66 , 081309 ( 2002 ) .
1 . transition energy as a function of inner radius @xmath23 , outer radius @xmath22 , respectively , for two different magnetic field @xmath65 , @xmath66 . | we have performed calculations of nonlinear optical absorption and nonlinear optical rectification of an exciton in a nanoring in the presence of the magnetic flux .
our results show that one can control properties of nonlinear optical absorption and nonlinear optical rectification of a nanoring by tuning outer and inner radius .
in addition , we also find that nonlinear optical properties of a nanoring can be modulated by the magnetic flux through nanoring .
winternitz - smorodinsky , magnetic flux , absorption , rectification , exciton |
the bright lmxbs can be divided on basis of their correlated x - ray spectral and timing behavior in the z sources and the atoll sources , named after the tracks they produce in the x - ray color - color diagram ( cd ; hasinger & van der klis 1989 ) .
the globular cluster source 4u 182030 ( ngc 6624 ) was classified as an atoll source . before the launch of the _ rossi x - ray timing explorer _ ( _ rxte _ ) two types of qpos were observed in the z sources , which were not observed in the atoll sources ( although several indications for qpos or peaked noise were found ; see , e.g. , stella , white , & priedhorsky 1987a ; hasinger & van der klis 1989 ; yoshida et al .
1993 ) . on the top branches ( the so - called horizontal branches ) of the z sources
a 1560 hz qpo was detected ( the horizontal - branch oscillation or hbo ) for whose explanation the magnetospheric beat - frequency model was put forward ( alpar & shaham 1985 ; lamb et al .
1985 ) . that the hbo was seen only in z sources
was thought to be related to a higher magnetic field in the z sources compared to that in the atoll sources ( hasinger & van der klis 1989 ) .
in such an interpretation the hbo would be much weaker in the atoll sources , but they could be observable at some level .
another type of qpo was observed on the diagonal branches ( the normal branches ) of the z sources ( the normal - branch oscillation or nbo ) , when the mass accretion rate is near the eddington mass accretion rate ( e.g. , hasinger & van der klis 1989 ; penninx 1989 ; smale 1998 ) .
the nbos were therefore thought to be due to a process which was activated by such high accretion rates ( fortner et al .
1989 ; alpar et al .
1992 ) , and should be observable in the atoll sources if they would reach near eddington accretion rates . shortly after the launch of _ rxte _ in december 1995 a third type of qpo was discovered in both the z sources and the atoll sources with frequencies between 200 and 1200 hz ( the khz qpos ; see , e.g. , van der klis et al .
1996 ; strohmayer et al . 1996
; see smale , zhang , & white 1997 for the discovery of khz qpos in 4u 182030 ) . in both types of sources ,
the khz qpos were remarkably similar and were most likely caused by the same physical mechanism . in the atoll sources , besided the khz qpos , other qpos with frequencies between 20 and 70 hz were observed ( e.g. , strohmayer et al .
1996 ; ford et al . 1997 ; wijnands et al .
the properties of these qpos ( their frequencies and their relation to broad band noise just below the qpo frequency ; wijnands et al .
1998 ) suggested a link with the hbos of the z sources .
so far , similar qpos as the z source nbos have not been reported in the atoll sources . in this letter ,
we report the discovery of a qpo near 7 hz in the atoll source 4u 182030 at its highest observed mass accretion rate .
if this qpo is caused by the same physical mechanism as the nbos in the z sources , then the production mechanism for these qpos can also be actived when the luminosity is much less than the critical eddington luminosity .
c we used archival data obtained with the _ rxte _ satellite of 4u 182030 on 1996 may 1 09:0022:51 ut , and 1996 may 3 22:15 may 4 18:05 ut . a total of @xmath440 ks of data were obtained .
no x - ray bursts were observed .
the data were simultaneously collected with 16 s time resolution in 129 photon energy channels ( covering 260 kev ) , with 2 ms in 16 channels ( covering 218.2 kev ) , and with 16 @xmath5s in 16 channels ( 18.260 kev ) . during @xmath410% of the time , detector 4 and/or 5 ( out of five ) were off . for the color - color analysis we only used the data of the three detectors which were always on ; for the power spectral analysis we used all available data . in order to create the cd and the hardness - intensity diagram ( hid )
, we used as soft color the 3.56.4/2.13.5 kev count rate ratio and as hard color the 9.716.0/6.49.7 kev ratio .
as intensity we used the count rate in the energy range 2.116.0 kev .
the cd and hid of all data are shown in figure [ 4u1820 - 30_nbo_cd_hid ] . in the right hand side of both diagrams
what seems to be a banana branch is observed , with an additional patch of data that might be either the lower banana branch or part of an island state .
the power spectrum observed during this part ( see below ) indicates that the source was not in an island state but in the lower part of the banana branch , because the strong ( @xmath420% rms amplitude ) band - limited noise , characteristic of the island state , was not observed .
we conclude that during these observations the source was in the banana branch continuously .
the gap between the lower part of the banana branch with the rest is caused by the two day data gap between the 1996 may 1 and 1996 may 34 observations . during the may 1 observation the source was in the lower part of the banana branch and during
the may 34 observations the source moved up an down the banana branch .
c in order to study the timing behavior as a function of the position of 4u 182030 on the banana branch , we made power spectra of 16-s data segments of the 2 ms and 16 @xmath5s data combined .
this resulted in @xmath42500 power spectra between 1/16 and 256 hz .
the rms amplitude values quoted are those obtained for the energy range 260 kev , unless otherwise stated .
we selected the power spectra on basis of the position of the source in the cd by dividing the banana branch in the cd into six regions ( figure [ 4u1820 - 30_nbo_cd_hid ] _ left _ ) .
each region was analyzed separately .
the power spectra corresponding to each region were averaged and the dead - time modified poisson level ( zhang 1995 ; zhang et al .
1995 ) was subtracted .
the average power spectra of regions 2 and 3 were fitted with a fit function consisting of a power law ( representing the very - low frequency noise [ vlfn ] ) and an exponentially cutoff power law ( representing the high - frequency noise [ hfn ] ) . for region 1
a lorentzian was included in the fit function as well , to fit a qpo near 70 hz .
for the power spectra corresponding to the other three regions , we only used a power law for the vlfn and a lorentzian for the 7 hz qpo ( region 6 ) or the broad bump near 7 hz ( regions 4 and 5 ) .
the power spectra corresponding to each region are shown in figure [ 4u1820 - 30_nbo_cd_powerspectra ] . in the lower banana branch ( region 1 ) the atoll source vlfn and hfn components ( see hasinger & van der klis 1989 ) were observed .
a qpo at 71@xmath01 hz was also observed , with an rms amplitude of 2.2%@xmath00.2% and a fwhm of 21@xmath04 hz . there are also indications for a broad noise components at @xmath4100 hz , similar to what has found in other atoll sources ( ford , van der klis , & kaaret 1998 ; ford & van der klis 1998 ; wijnands & van der klis 1998a ) , however , the limited time resolution below 18.2 kev did not allow to investigate this in more detail . the vlfn was detected throughout the entire banana branch except in region 6 .
its rms amplitude ( integrated over 0.011 hz ) varied erratically between 2.7% and 7.9% , with a 95% confidence upper limit in region 6 of 3.4% .
the power law index varied erratically between 1.0 and 1.7 .
the hfn was observed in regions 13 .
its rms amplitude ( integrated over 1100 hz ) and cutoff frequency decreased from 6.3%@xmath00.1% and [email protected] hz in region 1 , to 3.1%@xmath00.1% and [email protected] hz in region 3 .
the power law index varied erratically between @xmath60.9 and @xmath61.4 . in regions 4 and 5 ,
broad peaked noise is present near 7 hz .
it is unclear whether the hfn evolved into this broad peaked noise component or was replaced by it . in region 6 ,
a very significant ( 9@xmath7 ) qpo is present near 7 hz , which might be related to the peaked noise in region 4 and 5 .
we fitted this peaked noise as well as the qpo with a lorentzian and found frequencies of [email protected] hz in region 4 and [email protected] hz in region 5 , consistent with the qpo frequency of [email protected] hz in region 6 . while the fwhm of the lorentzian decreased from [email protected] hz in region 4 to [email protected] hz in region 6 ,
the rms amplitude increased from 3.4%@xmath00.4% to 6.2%@xmath00.3% .
we investigated the presence of the 7 hz qpo in more detail and we found that it was only present during approximately 160 s of data during the 1996 may 4 observation .
the qpo could then be directly observed in the light curve ( figure [ 4u1820 - 30_nbo_lightcurve ] _ bottom panel _ ) . when measured in 32-s time intervals
the qpo frequency shifted erratically between 5.5 and 8 hz and had at certain times a fwhm of only [email protected] hz .
the frequency drifts of the qpo caused the fwhm to be @xmath42 hz in all qpo data combined .
we used these 160 s of data showing the qpo to study the energy dependence of its amplitude .
we fixed the frequency ( [email protected] hz ) and fwhm ( [email protected] hz ) to the values obtained by using the total energy range ( rms amplitude of 5.6%@xmath00.2% ; 11.8@xmath7 ) .
the rms amplitude in the energy bands 2.85.3 , 5.36.8 , 6.89.3 , 9.310.8 , and 10.818.2 kev , were 3.7%@xmath00.5% , 5.0%@xmath00.8% , 7.3%@xmath00.6% , 7.1%@xmath01.3% , and 6.9%@xmath00.9% , respectively .
the 95% confidence upper limits on the presence of the qpo outside the 160 s interval were between 1% and 1.5% ( total energy range ) .
we performed time lag measurements between the energy bands 2.86.8 and 6.818.2 kev .
the time lag of [email protected] ms is consistent with being zero .
we determined the x - ray flux during the time of the qpo by fitting the x - ray spectrum with an absorbed black body plus a power law .
we obtained a 225 kev x - ray flux of 1.1@xmath110@xmath8 erg s@xmath3 @xmath9 .
this corresponds to a 225 kev x - ray luminosity of 5.4@xmath110@xmath2 erg s@xmath3 , assuming a distance of 6.4 kpc ( vacca , lewin , & van paradijs 1986 ) .
the power law index was @xmath42.8 indicating that not much flux was present above 25 kev .
we have discovered a very significant @xmath47 hz qpo in the x - ray flux of the globular cluster lmxb and atoll source 4u 182030 .
this qpo was only observed during a small part of the data , when the source was in the uppermost part of the banana branch in the cd .
it is thought that the mass accretion rate increases monotonically along the banana branch from the lower to the upper banana ( hasinger & van der klis 1989 ) .
if true , then the 7 hz qpo was only detected during the highest observed mass accretion rate in this source .
note that , although the qpo was only detectable when the source was at the uppermost part of the banana branch , the count rate then was not the highest observed ( fig .
[ 4u1820 - 30_nbo_lightcurve ] ) .
it seems that in this source , similar to what has been observed in the atoll sources 4u 163653 ( van der klis 1990 ; prins & van der klis 1997 ) and 4u 160852 ( mndez et al . 1998 ) , the count rate does not have a one - to - one relation with the mass accretion rate .
owing to the small amount of data showing the qpo it is difficult to compare this qpo with qpos observed in other x - ray binaries .
similar frequency qpos have been observed in a wide range of sources .
for example , in black hole candidates ( bhcs ) qpos with frequencies around 7 hz were observed in several systems ( see van der klis 1995 for a review ) . however , recently it was shown ( wijnands & van der klis 1998b ) that at least some of these bhc qpos are most likely related to the qpos seen in atoll sources between 20 and 70 hz .
when 4u 182030 is on the lower part of the banana branch , thus at lower mass accretion rates , we detect a significant @xmath470 hz qpo ( fig .
[ 4u1820 - 30_nbo_cd_powerspectra ] ) , similar to what is observed in other atoll sources .
this indicates that the 7 hz qpo is most likely not related to the 7 hz bhc qpos , unless the 70 hz qpo decreased in frequency from 70 to 7 hz from the lower banana branch to the upper part , opposite to what would be expected ( e.g. , ford & van der klis 1998 ) .
in the x - ray dipper 4u 132362 a strong ( @xmath410% rms amplitude ) qpo was recently discovered near 1 hz ( jonker et al . 1998 ) .
the production of this qpo is thought to be related to the high ( @xmath1070 ) inclination of this system .
although the modulation of the x - ray flux at the 11-min orbital period ( stella , priedhorsky , & white 1987b ) suggested a high inclination for 4u 182030 , the strong modulation observed in the ultraviolet flux of this source indicates that the inclination most likely is between 35 and 50 ( anderson et al .
1997 ) , making a common production mechanism of the two types of qpos unlikely .
this conclusion is further strengthened by the differences between the qpos .
the strength of the 1 hz qpo in 4u 132362 only marginally increased with photon energy and it is observed almost continuously ( even in the dips and the x - ray bursts ) , both contrary to the case of our 7 hz qpo .
the most likely identification of the 7 hz qpo in 4u 182030 is with the 7 hz qpos observed in the z sources , when they are accreting near the eddington mass accretion rate ( see van der klis 1995 ) .
this identification can not be tested by comparing the time lags of the different qpos .
the non - detection of time lags in 4u 182030 is consistent with the non - detection in scorpius x-1 ( dieters et al .
1999 ) , but inconsistent with the large ( @xmath480 ms ) time lags reported in cygnus x-2 ( mitsuda & dotani 1989 ) .
however , besides the similar frequencies , also the power spectral evolution with increasing mass accretion rate is very similar between the z sources and 4u 182030 . in the z sources at low observed accretion rates , relatively strong ( 1020% ) band limited noise is present , which decreases in amplitude when the accretion rate is increasing . at a certain accretion rate , this noise is hardly detectable , but a new component arises near 7 hz , which evolves in a clear qpo .
the power spectral evolution of 4u 182030 is remarkably similar ( fig .
[ 4u1820 - 30_nbo_cd_powerspectra ] ) . seen in this light
, it is not surprising that the 7 hz qpo in 4u 182030 is only observed in the highest observed mass accretion rate .
the nbos in the z sources are observed when these sources accretion near the eddington accretion rate ( hasinger & van der klis 1989 ; see , e.g. , penninx 1989 and smale 1998 for the value of the accretion rate when the z sources are on their normal branches ) .
however , the x - ray luminosity in 4u 182030 is , assuming isotropic emission , at maximum @xmath440% of the eddington critical luminosity if the companion star has cosmic abundance . as the companion star is most likely a helium white dwarf ( stella et al .
1987a ) , the x - ray luminosity might be only @xmath420% of the ( helium ) eddington critical luminosity .
thus , barring a large anisotropy in the x - ray emission , the accretion rate in 4u 182030 is well below the eddington critical accretion rate and below the accretion rate of the z sources when they exhibited their 7 hz qpos .
this would mean that , if the 7 hz qpo in 4u 182030 is caused by the same physical mechanism as similar frequency qpos in the z sources , then the formation mechanism behind these qpos is already activated in 4u 182030 well below the critical eddington luminosity .
note added in manuscript : after we submitted this letter , we analyzed the pca data of 4u 182030 used by smale et al .
( 1997 ) and zhang et al .
( 1998 ) . during some of their observations
the 7 hz qpo was detected , again only at times when the source was in the uppermost part of the banana branch .
when the source was at slightly lower inferred mass accretion rate , broad peaked noise was present near 10 hz , which might be related to the qpo .
these observations will be discussed in a forthcoming paper .
this work was supported in part by the netherlands foundation for research in astronomy ( astron ) grant 781 - 76 - 017 .
this research has made use of data obtained through the high energy astrophysics science archive research center online service , provided by the nasa / goddard space flight center .
we thank jeroen homan for stimulating discussions and comments on an earlier version of this document . | we have discovered a [email protected] hz quasi - periodic oscillation ( qpo ) in the x - ray flux of the low - luminosity low - mass x - ray binary ( lmxb ) and atoll source 4u 182030 .
this qpo was only observable at the highest observed mass accretion rate , when the source was in the uppermost part of the banana branch , at a 225 kev luminosity of 5.4@xmath110@xmath2 erg s@xmath3 ( for a distance of 6.4 kpc ) .
the qpo had a fwhm of only [email protected] hz during small time intervals ( 32-s of data ) , and showed erratic shifts in the centroid frequency between 5.5 and 8 hz .
the rms amplitude over the energy range 260 kev was 5.6%@xmath00.2% .
the amplitude increased with photon energy from 3.7%@xmath00.5% between 2.8 and 5.3 kev to 7.3%@xmath00.6% between 6.8 and 9.3 kev , above which it remained approximately constant at @xmath47% .
the time lag of the qpo between 2.86.8 and 6.818.2 kev was consistent with being zero ( [email protected] ms ) .
the properties of the qpo ( i.e. , its frequency and its presence only at the highest observed mass accretion rate ) are similar to those of the 520 hz qpo observed in the highest luminosity lmxbs ( the z sources ) when they are accreting near the eddington mass accretion limit .
if this is indeed the same phenomenon , then models explaining the 520 hz qpo in the z sources , which require the near - eddington accretion rates , will not hold .
assuming isotropic emission , the 225 kev luminosity of 4u 182030 at the time of the 7 hz qpos is at maximum only 40% ( for a companion star with cosmic abundances ) , but most likely @xmath420% ( for a helium companion star ) of the eddington accretion limit . |
sterile neutrino oscillation models were proposed more than a decade ago as an explanation for the lsnd anomaly @xcite , an excess of events consistent with @xmath0 oscillations at high @xmath1 .
these models relate @xmath2 appearance ( @xmath3 ) with @xmath4 and @xmath2 disappearance ( @xmath5 and @xmath6 ) , motivating combined fits in all three oscillation channels . relatively early in the discussion of models ,
it was demonstrated @xcite that a three - active plus one - sterile ( 3 + 1 ) neutrino oscillation model could not reconcile the lsnd result with existing null results from other short - baseline ( sbl ) experiments , including karmen @xcite , nomad @xcite , bugey @xcite , chooz @xcite , ccfr84 @xcite , and cdhs @xcite , which had similar high @xmath1 sensitivity .
however , it was shown that a three - active plus two - sterile neutrino ( 3 + 2 ) oscillation scenario provided a better description of these data sets @xcite . in 2001 ,
the miniboone experiment began running with the goal to test the lsnd result using both neutrino and antineutrino beams .
this is a short - baseline appearance and disappearance experiment located at fermi national accelerator laboratory ( fermilab ) .
miniboone s first results , reported in 2007 , described a search for @xmath7 oscillations @xcite .
these results were incompatible with a simple two - neutrino oscillation interpretation of the lsnd signal and , within this model , miniboone excluded the lsnd result at the 98% cl .
however , this same analysis reported a 3.7@xmath8 excess of electron neutrino candidate events at low energies , between 300 - 475 mev , which remains unexplained .
reference @xcite included the miniboone first result in a global fit to all sbl experiments under the ( 3 + 1 ) and ( 3 + 2 ) oscillation scenarios .
the analysis built on an earlier study , which introduced the possibility of cp violation @xmath9 within ( 3 + 2 ) fits @xcite . including the first miniboone results into the global fit led to two observations in ref .
@xcite : 1 ) miniboone , lsnd , and the null appearance experiments ( karmen and nomad ) are compatible under a ( 3 + 2 ) sterile neutrino oscillation scenario with large cp violation .
2 ) there is severe tension between appearance and disappearance experiments , at a level of more than 3@xmath8 . in this paper
we will consider both observations in light of new appearance data .
also , we will show that the incompatibility between appearance and disappearance experiments arises mainly from two @xmath10 disappearance data sets : cdhs and atmospheric constraints .
motivated by three new results from miniboone , this paper re - examines the ( 3 + 1 ) and ( 3 + 2 ) global fits to the sbl data .
these new results are : 1 ) an updated @xmath11 result @xcite ; 2 ) first results for a @xmath12 search @xcite ; and 3 ) first @xmath7 results from the numi off - axis beam at miniboone @xcite .
we consider these new results in combination with seven sbl data sets .
these provide constraints on : @xmath10 disappearance ( from the ccfr84 and cdhs experiments ) , @xmath13 disappearance ( from the bugey and chooz experiments ) , @xmath7 oscillations ( from the nomad experiment ) , and @xmath12 oscillations ( from the lsnd and karmen experiments ) .
furthermore , we have taken into account atmospheric constraints based on the analysis of ref .
these constraints have been incorporated in our analysis following the method described in ref .
@xcite , and are included in fits to all sbl experiments , null sbl experiments , or as explicitly stated .
table [ tab : sblsummary ] summarizes all sbl data sets used in the fits presented in this paper .
.[tab : sblsummary ] short - baseline oscillation data sets considered in this paper , and oscillation channel that each data set constrains . [ cols="<,^",options="header " , ] table [ tab : tension ] summarizes the @xmath14-probability and pg results from ( 3 + 2 ) cp - violating fits .
the upper rows summarize @xmath14-probabilities and pg s from fits to all sbl experiments , as well as fits to appearance - only , disappearance - only , neutrino - only , antineutrino - only , neutrino appearance - only , and antineutrino appearance - only data sets .
appearance and disappearance data sets , as well as neutrino and antineutrino data sets , are incompatible with a pg of less than 0.1% . grouping sbl appearance - only data sets according to whether they are neutrino or antineutrino experiments yields significantly higher compatibilities98% among @xmath15 appearance experiments , and 74% among @xmath16 appearance experiments .
the compatibility between @xmath15 and @xmath16 appearance - only results is lower , at 6.8% but still acceptable . in the case where disappearance experiments are included in the comparison between neutrino and antineutrino fits ,
the compatibility among all @xmath16 sbl data sets remains considerably high , at 80% , as does the compatibility among all @xmath15 sbl data sets , at 43% .
however the compatibility between @xmath16 and @xmath15 results is only 0.06% .
the remaining rows of table [ tab : tension ] provide the @xmath14-probabilities of global fits under the same oscillation scenario where one experiment is excluded from the fit at a time ( as indicated by the `` - '' sign in the table ) .
the @xmath14 probabilities of all fits are acceptable , ranging between 29.2% for a fit excluding the bugey data set , and 72.2% for a fit excluding the lsnd data set . aside from lsnd ,
three experiments stand out as having the poorest compatibility when compared to a global fit with all other sbl data sets : 1 ) bnb - mb(@xmath15 ) , 2 ) cdhs and 3 ) atmospheric constraints ( atm ) .
these three experiments have been identified as the possible source of tension between appearance and disappearance experiments , or neutrino and antineutrino experiments .
the remaining combinations yield reasonably high compatibilities of at least 42% , with the exception of lsnd and bnb - mb(@xmath16 ) which are compatible with the remaining data sets at 2.9% and 13% , respectively . to further test the hypothesis that the tension between appearance and disappearance experiments is a result of the bnb - mb(@xmath15 ) and cdhs data sets and atmospheric constraints ,
the compatibility between appearance and disappearance experiments is re - evaluated several times .
each time , a different combination of these three experiments is excluded from the fits . the results are summarized in table [ tab : tension2 ] .
the compatibility among appearance and disappearance experiments with bnb - mb(@xmath15 ) , cdhs , and atmospheric constraints excluded from the fits is high , at 23.7% .
the bnb - mb(@xmath15 ) data set alone is not responsible for the disagreement between appearance and disappearance experiments , as suggested by the sixth row of the table [ tab : tension2 ] . even with bnb - mb(@xmath15 ) included in the fit , a compatibility of 2.9% can be obtained if cdhs and atmospheric constraints are excluded from the fit .
the same test can be performed between neutrino and antineutrino experiments .
the results are summarized in table [ tab : tension3 ] .
again , the compatibility between neutrino and antineutrino experiments is re - evaluated several times ; each time , a different combination of the bnb - mb(@xmath15 ) , cdhs , and atmospheric constraint data sets is excluded from the fits .
here , the compatibility among neutrino and antineutrino experiments with bnb - mb(@xmath15 ) , cdhs , and atmospheric constraints excluded from the fits is even higher , at 56% .
the bnb - mb(@xmath15 ) data set is just as responsible for the disagreement between neutrino and antineutrino experiments as the cdhs data set and atmospheric constraints alone .
the tension seems to be caused by all three experiments , as none of them independently excluded from the fit can accound for the increase in compatibility from @xmath171% ( or less ) to 56% .
it is possible that higher compatibility between bnb - mb(@xmath15 ) and all remaining sbl data sets may be achieved if the fits are to be repeated with the low energy region ( 200@xmath18475 mev ) excluded from the bnb - mb(@xmath15 ) data set . a global analysis with bnb - mb(@xmath15 ) ,
cdhs , and atmospheric constraints excluded from the fit yields a @xmath14-probability of 82% and @xmath1990% compatibility among all experiments .
we have re - examined global fits to sterile neutrino oscillation models , using new data from miniboone . those include the final miniboone neutrino mode results and the first , low statistics miniboone antineutrino results , as well as first results from the off - axis numi beam observed in the miniboone detector .
within a ( 3 + 1 ) cp- and cpt - conserving scenario , we have found that the data set collected by miniboone using the numi off - axis beam ( numi - mb ) currently provides very weak constraints to sterile neutrino fits , due to large systematic uncertainties .
updated numi results , expected soon , should have a greater impact on these fits . within the same oscillation framework , the miniboone antineutrino data set ( bnb - mb(@xmath16 ) )
is found in agreement with lsnd , yielding , in a combined analysis with lsnd and karmen under a ( 3 + 1 ) oscillation hypothesis , a @xmath14-probability of 29% , and best - fit parameters similar to those of lsnd .
updated miniboone antineutrino appearance results , with almost twice the current statistics , are expected in the near future .
the miniboone neutrino data set ( bnb - mb(@xmath15 ) ) , although suggestive of an excess that could be described by a ( 3 + 1 ) oscillation hypothesis with a @xmath14 probability of 35% , is found incompatible with the signals from the miniboone antineutrino and lsnd results . the remaining null appearance and disappearance experiments ( numi - mb(@xmath15 ) , karmen , nomad , bugey , chooz , cdhs , ccfr84 ) and atmospheric oscillation data impose strong constraints to the parameter space allowed by a combined ( 3 + 1 ) analysis of miniboone neutrino and antineutrino and lsnd data , excluding the 99% cl allowed region at 99% cl . however , the constraints from antineutrino disappearance experiments on the parameter space allowed by antineutrino appearance experiments ( bnb - mb(@xmath16 ) , lsnd , and karmen ) are weaker . in a ( 3 + 1 ) oscillation framework
, all antineutrino experiments yield a best - fit @xmath14-probability of 86% , and exclude the no - oscillations hypothesis at @xmath195.0@xmath8 .
the best - fit parameters are similar to those of lsnd , and correspond to a muon antineutrino disappearance amplitude of 0.35 , which may be addressed by upcoming results from miniboone and minos on muon antineutrino disappearance .
additionally , fits to all neutrino experiments yield a best - fit @xmath14-probability of 47% and exclude the null hypothesis at @xmath1990% cl .
furthermore , we find that with the addition of the new miniboone data sets , the ( 3 + 2 ) oscillation models provide a much better description of all sbl data sets compared to ( 3 + 1 ) models . in the case of ( 3 + 2 )
fits , cp violation allows for a significant improvement in @xmath14-probability for fits to only bnb - mb(@xmath15 ) + bnb - mb(@xmath16 ) + lsnd , and fits to only appearance experiments . in the case of global fits , however , the effect of cp violation is muted , as allowing for cp violation results in a relatively small improvement in the fit .
the @xmath14-probability for the best - fit ( 3 + 2 ) cpv hypothesis is 54% , compared to 52% for the cpc case .
the best - fit corresponds to large but not maximal cp violation ( @xmath20 ) .
the high incompatibility among appearance and disappearance data seen in the past @xcite in the case of ( 3 + 2 ) cp - violating fits still remains with the addition of the new miniboone results .
we have shown that the incompatibility is a result of the bnb - mb(@xmath15 ) and cdhs data sets and atmospheric constraints .
the compatibility between appearance and disappearance data with these three experiments excluded from the fits is significantly higher , at 24% .
neutrino and antineutrino results are also incompatible within a ( 3 + 2 ) cp - violating scenario , with a pg of less than 0.1% .
the compatibility improves to 6.8% in the case of comparing appearance - only neutrino versus antineutrino results .
overall , allowing for mixing with multiple sterile neutrino states and cp violation does not seem sufficient to allow incorporating all sbl experiments within a cpt - conserving , sterile neutrino oscillation framework .
it may be that there is an issue with one or more of the following data sets : lsnd , bnb - mb(@xmath15 ) , including the low - energy excess , cdhs , or atmospheric constraints ; alternatively , theories with cpt - violating oscillations or effective cpt violation @xcite may succeed in reconciling all short - baseline oscillation signatures , and should be explored .
99 c. athanassopoulos _ et al . _
[ lsnd collaboration ] , phys .
* 77 * , 3082 ( 1996 ) [ arxiv : nucl - ex/9605003 ] ; c. athanassopoulos _ et al . _
[ lsnd collaboration ] , phys .
c * 58 * , 2489 ( 1998 ) [ arxiv : nucl - ex/9706006 ] ; a.
aguilar _ et al . _
[ lsnd collaboration ] , phys .
d * 64 * , 112007 ( 2001 ) [ arxiv : hep - ex/0104049 ] .
o. l. g. peres and a. y. smirnov , nucl .
b * 599 * , 3 ( 2001 ) [ arxiv : hep - ph/0011054 ] .
a. strumia , phys .
b * 539 * , 91 ( 2002 ) [ arxiv : hep - ph/0201134 ] .
w. grimus and t. schwetz , eur .
j. c * 20 * , 1 ( 2001 ) [ arxiv : hep - ph/0102252 ] .
m. sorel , j. m. conrad and m. h. shaevitz , phys .
d * 70 * , 073004 ( 2004 ) [ arxiv : hep - ph/0305255 ] .
m. maltoni , t. schwetz , m. a. tortola and j. w. f. valle , nucl .
b * 643 * , 321 ( 2002 ) [ arxiv : hep - ph/0207157 ] .
b. armbruster _ et al .
_ [ karmen collaboration ] , phys .
d * 65 * , 112001 ( 2002 ) [ arxiv : hep - ex/0203021 ] .
p. astier _ et al .
_ [ nomad collaboration ] , phys .
b * 570 * , 19 ( 2003 ) [ arxiv : hep - ex/0306037 ] ; d. gibin , nucl .
* 66 * , 366 ( 1998 ) ; v. valuev [ nomad collaboration ] , http://www.slac.stanford.edu/spires/find/hep/www?irn=4920686[spires entry ] _
prepared for international europhysics conference on high - energy physics ( hep 2001 ) , budapest , hungary , 12 - 18 jul 2001 _ y. declais _ et al .
_ , nucl .
b * 434 * , 503 ( 1995 ) .
m. apollonio _ et al .
_ , [ arxiv : hep - ex/0301017 ] . i. e. stockdale _ et al .
_ , phys .
lett . * 52 * , 1384 ( 1984 ) .
f. dydak _ et al .
_ , phys .
b * 134 * , 281 ( 1984 ) .
a. a. aguilar - arevalo _ et al . _
[ the miniboone collaboration ] , phys .
lett . * 98 * , 231801 ( 2007 ) [ arxiv : hep - ex/0704.1500 ] . m. maltoni and t. schwetz , phys .
d * 76 * , 093005 ( 2007 ) [ arxiv : hep - ph/0705.0107 ]
. g. karagiorgi _ et al .
_ , phys .
d * 75 * , 013011 ( 2007 ) [ arxiv : hep - ph/0609177 ] .
a. a. aguilar - arevalo _ et al . _ [ miniboone collaboration ] , phys .
lett . * 102 * , 101802 ( 2009 ) [ arxiv : hep - ex/0812.2243 ] .
a. a. aguilar - arevalo _ et al . _ [ miniboone collaboration ] , [ arxiv : hep - ex/0904.1958 ] .
p. adamson _ et al .
_ , phys .
lett . * 102 * , 211801 ( 2009 ) [ arxiv : hep - ex/0809.2447 ] .
m. maltoni , t. schwetz , m. tortola and j. w. f. valle , new j. phys .
* 6 * , 122 ( 2004 ) [ arxiv : hep - ph/0405172 ] .
s. goswami and w. rodejohann , jhep * 0710 * , 073 ( 2007 ) [ arxiv:0706.1462 [ hep - ph ] ] .
a. melchiorri , o. mena , s. palomares - ruiz , s. pascoli , a. slosar and m. sorel , jcap * 0901 * , 036 ( 2009 ) [ arxiv:0810.5133 [ hep - ph ] ] .
a. d. dolgov and f. l. villante , nucl .
b * 679 * , 261 ( 2004 ) [ arxiv : hep - ph/0308083 ] .
a. a. aguilar - arevalo _ et al . _
[ miniboone collaboration ] , phys .
d * 79 * , 072002 ( 2009 ) [ arxiv : hep - ex/0806.1449 ] .
a. a. aguilar - arevalo _ et al . _
[ miniboone collaboration ] , nucl .
instrum .
a * 599 * , 28 ( 2009 ) [ arxiv : hep - ex/0806.4201 ] .
v. barger , y. b. dai , k. whisnant and b. l. young , phys .
d * 59 * , 113010 ( 1999 ) [ arxiv : hep - ph/9901388 ] .
b. kayser , [ arxiv : hep - ph/0211134 ] .
p. bremaud , _ markov chains : gibbs fields , monte carlo simulation , and queues _ ,
springer , new york , 1999 .
m. maltoni and t. schwetz , phys .
d * 68 * , 033020 ( 2003 ) [ arxiv : hep - ph/0304176 ] .
a. a. aguilar - arevalo _ et al . _ [ miniboone collaboration ] , phys .
lett . * 103 * , 061802 ( 2009 ) [ arxiv:0903.2465 [ hep - ex ] ] .
a. a. aguilar - arevalo _ et al . _ [ sciboone collaboration ] , arxiv : hep - ex/0601022 . k. b. m. mahn , private communication .
i. ambats _ et al . _
[ minos collaboration ] , fermilab - design-1998 - 02 .
minos , j. hartnell , ( 2009 ) , fnal seminar , 15 may 2009 .
+ url : http://theory.fnal.gov / jetp / talks / hartnell.pdf .
a. d. sakharov , pisma zh .
fiz . * 5 * , 32 ( 1967 ) [ jetp lett . * 5 * , 24 ( 1967 sopua,34,392 - 393.1991 ufnaa,161,61 - 64.1991 ) ] .
g. barenboim and j. d. lykken , phys .
b * 554 * , 73 ( 2003 ) [ arxiv : hep - ph/0210411 ] .
g. barenboim and j. d. lykken , arxiv:0908.2993 [ hep - ph ] .
s. hollenberg and h. pas , arxiv:0904.2167 [ hep - ph ] . | this paper examines sterile neutrino oscillation models in light of recently published results from the miniboone experiment .
the new miniboone data include the updated neutrino appearance results , including the low energy region , and the first antineutrino appearance results , as well as first results from the off - axis numi beam observed in the miniboone detector .
these new global fits also include data from lsnd , karmen , nomad , bugey , chooz , ccfr84 , and cdhs .
constraints from atmospheric oscillation data have been imposed .
we test the validity of the three - active plus one - sterile ( 3 + 1 ) and two - sterile ( 3 + 2 ) oscillation hypotheses , and we estimate the allowed range of fundamental neutrino oscillation parameters in each case . we assume cpt - invariance throughout .
however , in the case of ( 3 + 2 ) oscillations , cp violation is allowed . with the addition of the new miniboone data sets ,
there are clear incompatibilities between neutrino and antineutrino experiments under a ( 3 + 1 ) oscillation hypothesis . a better description of all short - baseline data over a ( 3 + 1 ) is provided by a ( 3 + 2 ) oscillation hypothesis with cp violation .
however , we still find large incompatibilities among appearance and disappearance experiments , consistent with previous analyses , as well as incompatibilities between neutrino and antineutrino experiments . aside from lsnd ,
the data sets responsible for this tension are the miniboone neutrino data set , cdhs , and the atmospheric constraints . on the other hand ,
fits to antineutrino - only data sets , including appearance and disappearance experiments , are found significantly more compatible , even within a ( 3 + 1 ) oscillation scenario . |
the extraordinary activity associated with quasars involves accretion onto a supermassive black hole ( smbh ) , with the uv / optical emission arising from a geometrically thin , optically thick cold accretion disk @xcite , and the x - ray continuum arising from a hot , optically thin corona that compton upscatters the disk uv photons ( e.g. , * ? ? ?
* ) . in highly accreting objects ,
like quasars ( @xmath13 , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the x - ray plasma geometry is expected to be that of a hot , possibly patchy , ionized ` skin ' that sandwiches the cold disk ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. however , the evidence for this is not conclusive , and relies on data from x - ray binaries and low-@xmath14 sources ( e.g. , see the dicussion by * ? ? ?
other geometries are possible , including an accretion disk that evaporates into a hot inner flow ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , or a combination of a hot inner flow and a corona that sandwiches the disk ( e.g. , * ? ? ?
* ; * ? ? ?
furthermore , radiation pressure can drive an outflow from the disk into the corona if the two are cospatial , thus altering the physics of the corona @xcite .
investigations of how quasar x - ray parameters depend on black hole mass , @xmath2 , and accretion rate relative to eddington , @xmath12 , offer important constraints on models of the disk / corona system .
there have been attempts to link the evolution of smbhs to analytic and semi - analytic models of structure formation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , where black holes grow by accreting gas funneled towards the center during a galaxy merger until feedback energy from the smbh expels gas and shuts off the accretion process ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . this ` self - regulated ' growth of black holes has recently been successfully applied in smoothed particle hydrodynamics simulations @xcite . within this framework ,
the agn or quasar phase occurs during the episode of significant accretion that follows the galaxy merger , persisting until feedback from the black hole ` blows ' the gas away ( e.g. , * ? ? ? * ) .
hydrodynamic calculations have shown that line pressure is more efficient than thermal pressure at driving an outflow @xcite , and therefore , the efficiency of agn feedback depends on the fraction of energy emitted through the uv / disk component as compared to the x - ray / corona component .
if the fraction of energy emitted in the uv as compared to the x - ray depends on @xmath2 or @xmath12 , then it follows that the efficiency of agn feedback will also depend on @xmath2 and @xmath12 .
this has important consequences for models of smbh growth , as the smbh may become more or less efficient at driving an outflow depending on its mass and accretion rate .
studies of the dependence of quasar x - ray / uv emission on black hole mass and accretion rate are therefore important as they allow us to constrain a @xmath2- or @xmath12-dependent feedback efficiency .
numerous previous studies have searched for a luminosity and redshift dependence of @xmath15 , the ratio of x - ray to uv / optical flux ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and @xmath1 , the x - ray spectra slope ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
most studies have found a correlation between @xmath0 and uv luminosity , @xmath16 , while the existence of a correlation between @xmath17 and @xmath14 is still a matter of debate ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in addition , studies of @xmath1 have produced mixed results .
some authors have claimed a correlation between @xmath1 and luminosity ( e.g. , * ? ? ?
* ; * ? ? ?
* ) or redshift ( e.g. , * ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ) , while , others find no evidence for a correlation between @xmath1 and @xmath16 or @xmath14 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . a correlation between @xmath1 and the @xmath18 of the h@xmath3 line has also been found ( e.g. , * ? ?
* ; * ? ? ?
* ) , suggesting a correlation between @xmath1 and black hole mass or eddington ratio ( e.g. , * ? ? ?
* ; * ? ? ?
recently , it has become possible to obtain estimates of @xmath2 for broad line agn by calibrating results from reverberation mapping @xcite for use on single - epoch spectra @xcite .
this has enabled some authors to confirm a correlation between @xmath1 and either @xmath2 or @xmath19 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , where the x - ray continuum hardens with increasing @xmath2 or softens with increasing @xmath19 .
in addition , previous work has also found evidence for quasars becoming more x - ray quiet as @xmath2 or @xmath19 increase @xcite ; however , studies involving the dependence of @xmath0 on @xmath2 or @xmath19 have remained rare compared to studies of @xmath1 .
it is important to note that the correlations inferred in previous work generally employ broad line mass estimates in combination with a constant bolometric correction .
therefore , most of the correlations found in previous work are , strictly speaking , between @xmath1 or @xmath0 and the estimates @xmath20 and @xmath21 , where @xmath22 . in this work ,
we investigate the dependence of @xmath0 and @xmath1 on black hole mass , optical / uv luminosity relative to eddington , and x - ray luminosity relative to eddington .
we combine the main sloan digital sky survey ( sdss ) sample of @xcite with the sample of @xcite , creating a sample of 318 radio - quiet quasars ( rqq ) with x - ray data from _ rosat _ and _ chandra _ , and optical spectra mostly from the sdss ; 153 of these sources have estimates of @xmath1 from _
chandra_. because the x - ray emission in radio - loud sources can have an additional component from the jet ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , we focus our analysis on the radio - quiet majority .
our sample has a detection fraction of @xmath23 and spans a broad range in black hole mass ( @xmath24 ) , redshift ( @xmath25 ) , and luminosity ( @xmath26@xmath7 \lesssim 10^{48}$ ] ) , enabling us to effectively look for trends regarding @xmath0 and @xmath1 .
the outline of this paper is as follows . in [ s - sample ]
we describe the construction of our sample , and in
[ s - specfits ] we describe the procedure we used to fit the optical continuum and emission lines . in [ s - mbhredd_est ] we describe how we obtain broad line mass estimates , our bolometric correction , and argue that a constant bolometric correction provides a poor estimate of the bolometric luminosity . in
[ s - alfox_bh ] we describe the results from a regression analysis of @xmath0 on @xmath27 and @xmath9 , and in [ s - gamx_bh ] we report evidence for a non - monotonic dependence of @xmath1 on either @xmath28 and @xmath9 . in [ s - discussion ] we discuss our results within the context of agn disk / corona models , and we discuss the implications for a dependence of quasar feedback efficiency on black hole mass or accretion rate . in [ s - summary ] we summarize our main results .
we adopt a cosmology based on the the wmap best - fit parameters ( @xmath29 , * ? ? ?
for ease of notation , we define @xmath30@xmath31@xmath32 and @xmath33 .
in this analysis we combine 169 rqqs from ( * ? ? ? * hereafter k07 ) with 149 rqqs from the main sdss sample of ( * ? ? ?
* hereafter s05 ) to create a sample of 318 rqqs . out of these 318 sources , 276 ( @xmath34 )
are detected in the x - rays .
the @xmath35 sources from the k07 sample were selected by cross - correlating the sdss dr3 quasar catalogue @xcite with the _ chandra _ public archive as of 2005 february 22 .
the @xmath36 sources from the k07 sample consist of targeted _ chandra _ rqqs taken from the literature @xcite , and new observations reported by k07 .
the sources taken from s05 were selected from the sdss to be contained within the inner @xmath37 of _ rosat _ pspc pointings with exposure times @xmath38 ksec .
the x - ray data for both samples are as reported by s05 and k07 . both the s05 and k07 samples consist only of radio - quiet quasars .
we focus our analysis on the radio - quiet majority because the radio - loud sources have an additional component of x - ray emission arising from the jet ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in addition , both s05 and k07 omitted bal qsos when possible .
it is necessary to remove the bal qsos because their high column density gives them the appearance of being x - ray weak ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , potentially biasing our analysis .
however , neither s05 nor k07 were able to remove the high - ionization bal quasars for @xmath39 , as their identification requires observations of the c iv line .
in addition , low - ionization bals can be identified at @xmath40 based on mg ii absorption .
@xcite found the fraction of bals in the sdss to be @xmath41 , and therefore we expect there to be @xmath42 bals in our sample at @xmath39 .
this number may be higher if one relaxes the definition of a bal quasar @xcite .
we exclude five sources from the s05 sample due to significant intrinsic narrow uv absorption or obvious host galaxy contamination : source sdss j103747.4 - 001643.9 ( @xmath43 ) had significant c iv absorption , and sources sdss j124520.7 - 002128.1 ( @xmath44 ) and sdss j103709.8 + 000235.2 ( @xmath45 ) had significant absorption in both c iv and l@xmath46 .
these three sources were omitted because the absorption prohibits obtaining an accurate line width measurement , necessary for broad line mass estimates , and to ensure that the x - ray emission under study is not effected by the absorption .
sources sdss j230440.6 - 082220.8 ( @xmath47 ) and sdss j023306.0 + 003856.4 ( @xmath48 ) have a significant host - galaxy component in their spectra .
in addition , we exclude source sdss j144340.8 + 585653.2 ( @xmath49 ) from the k07 sample because it has significant uv absorption .
we removed source sdss j142414.1 + 421400.1 ( @xmath50 ) from the k07 sample and source sdss j170441.4 + 604430.5 ( pg 1704 + 608 , @xmath51 ) from the s05 sample , as both sources are radio - loud .
we could not estimate black hole masses for sources sdss j083206.0 + 524359.3 ( @xmath52 ) , sdss j144231.7 + 011055.3 ( @xmath53 ) , and pc
0910 + 5625 ( @xmath54 ) .
all three of these sources are from the k07 sample .
the region containing the mg ii emission line for sdss j083206.0 + 524359.3 was missing from the sdss spectrum , the emission lines are too weak for sdss j144231.7 + 011055.3 , and an optical spectrum was not available for pc 0910 + 5625 .
optical spectra were obtained for most sources from the sdss .
we also obtained spectra for some of the high redshift quasars from @xcite , @xcite , and @xcite .
the values of @xmath16 and @xmath55 , for the k07 sources are taken from k07 .
we processed the optical spectra for the s05 sources in the same manner as for the k07 sources .
we do this for consistency and because s05 did not correct for quasar iron emission .
as described by k07 , we corrected the optical spectra for galactic absorption using the @xmath56 values taken from @xcite , as listed in the nasa / ipac extragalactic database ( ned ) , and the extinction curve of @xcite , assuming a value of @xmath57 .
we model the continuum as a power law of the form @xmath58 , and the fe emission as a scaled and broadened iron template extracted from i zw i. the optical iron template was extracted by @xcite , and the uv iron template was extracted by @xcite .
the continuum and iron emission were fit simultaneously using the levenberg - marquardt method for nonlinear @xmath59-minimization .
continuum flux densities were then estimated using the power law parameters .
we were not able to use a power - law fit to calculate @xmath16 for the @xmath60 sources , as the sdss spectral range for these sources does not contain the rest - frame uv continuum .
instead , we use the luminosity of the broad component of the h@xmath3 emission line , @xmath61 , as a proxy for @xmath16 . it is preferable to use the broad h@xmath3 emission line luminosity over , say , the optical continuum luminosity as a proxy for @xmath16 because the broad h@xmath3 emission line is not contaminated by emission from the host galaxy , and thus should provide an approximately unbiased estimate of @xmath16 .
host - galaxy contamination is likely negligible for all @xmath62 sources , as @xmath63 at @xmath64 for galaxies @xcite .
we fit a power - law relationship between @xmath16 and @xmath61 using the 44 sources at @xmath65 for which a measurement of both quantities if available .
we used the linear regression method of @xcite , which allows for measurement errors in both variables , and find @xmath66 the intrinsic scatter about this relationship is @xmath67 dex , implying a potential uncertainty in @xmath68 inferred from this relationship of the same magnitude .
there was no trend in the residuals with either @xmath14 or @xmath61 , implying that equation ( [ eq - luv_lhb ] ) should give unbiased estimates of @xmath68 for the @xmath69 sources .
values of @xmath16 were estimated using equation ( [ eq - luv_lhb ] ) for both the k07 and s05 @xmath69 sources , a total of 42 sources .
the distributions of @xmath16 and @xmath70 as a function of redshift are shown in figure [ f - lum_z_dist ] .
we calculate the ratio of optical to x - ray flux @xcite as @xmath71 where @xmath72 and @xmath73 are the rest - frame flux densities at 2 kev and @xmath64 , respectively .
if the flux density from @xmath64 to 2 kev is a simple power law , then @xmath0 is the spectral slope of this continuum , and thus @xmath0 may be thought of as a crude estimate of the shape of the ionizing continuum .
the parameter @xmath74 is an important parameter for model comparison , as it summarizes the amount of energy emitted in the x - ray region ( most likely a comptonized component ) , compared with that emitted in the optical - uv ( accretion disk component ) .
the distribution of @xmath0 as a function of @xmath16 and @xmath14 are also shown in figure [ f - lum_z_dist ] .
the error on @xmath0 is the result of measurement errors on the uv and x - ray flux , as well as error caused by quasar variability over the different epochs for the uv and x - ray observations . in general , the measurement errors on @xmath68 are negligible compared to the error on @xmath75 .
s05 estimates a error on the x - ray flux of @xmath76 dex , including both the contributions from measurement error and variability .
typical long - term x - ray variability for seyfert 1s is @xmath77@xmath78 with no obvious trend with luminosity @xcite .
the measurement errors in @xmath75 for the k07 sample are typically @xmath79 dex . assuming x - ray variability amplitudes of @xmath80
, this implies typical uncertainties in the x - ray luminosity of @xmath81 dex .
therefore , we estimate the uncertainty on @xmath0 to be @xmath82 for the k07 sources and @xmath83 for the s05 sources .
we extracted the h@xmath3 , mg ii , and c iv emission lines in order to use their widths in our black hole mass estimates ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these line were extracted by first subtracting the continuum and fe emission , interpolating over any narrow absorption features , and modelling all lines within the extraction region as a sum of gaussian functions .
any nearby lines were then subtracted , leaving only the broad emission line profile . in all cases the line profile extraction
was done interactively and every line fit was inspected visually . for h@xmath3 , we extracted the region within @xmath84 of @xmath85 , where we use the standard convention that negative velocities are blueward of a given wavelength .
the h@xmath3 profile was modeled as a sum of 23 gaussian functions .
the [ o iii ] @xmath86 and [ o iii ] @xmath87 lines were modeled as a sum of 12 gaussian functions , depending on the signal - to - noise of the lines .
a sum of two gaussian functions was used for the higher @xmath88 lines because the [ o iii ] line profiles are not exactly a gaussian function ; the individual gaussian components are not considered to be physically distinct components .
the widths of the narrow gaussian functions for h@xmath3 and [ o iii ] lines were fixed to be equal to eachother .
the [ o iii ] lines and the narrow component of the h@xmath3 line were then subtracted , leaving the broad component of h@xmath3 . for mg ii , we extracted the region within @xmath89 of @xmath90 .
there are no nearby non - iron emission lines that mg ii is blended with , so the extraction is trivial after removing the fe and continuum emission . for c iv , we extracted the region within @xmath91 and @xmath92 of @xmath93 .
the c iv line was modeled as a sum of 23 gaussian functions , and he ii @xmath94 and o iii ] @xmath95 were modeled as a sum of 12 gaussian functions each . after obtaining estimates of the he ii and o iii ] profiles , we subtracted these components .
we did not model the n iv ] @xmath96 emission line as this line is typically weak and lost in the c iv wings . in order to estimate @xmath2
, it is necessary to measure the @xmath18 of the emission lines . after extracting the line profiles
, we estimate the @xmath18 for the h@xmath3 , mg ii , and c iv emission lines by fitting them to a sum of 15 gaussian functions , enabling us to obtain a smooth representation of each line .
in contrast to our profile extraction technique , we choose the number of gaussian functions to minimize the bayesian information criterion ( @xmath97 , * ? ? ?
the @xmath97 is a common criterion to use for selecting the number of parameters in a model ( e.g. , see * ? ? ?
* ) ; the model that minimizes the @xmath97 is approximately the model that is most supported by the data . for gaussian errors ,
as assumed in this work , the bic is simply a modification to the standard @xmath59 statistic : @xmath98 where @xmath99 is the number of gaussian functions used , @xmath100 is the number of free parameters , and @xmath101 is the number of data points used in the fit . using the @xmath97 to ` fit ' the number of gaussian functions thus allows us more flexibility in obtaining a smooth representation of the line profile , as we are not choosing the number of gaussian functions arbitrarily .
once a smooth representation is obtained , we automatically measure the @xmath18 directly from the best fit line profile .
the standard errors on @xmath18 are estimated using a bootstrap method .
we simulated 100 ` observed ' emission lines by adding random gaussian noise to the best fit line profile with standard deviation equal to the noise level of the spectrum , including the propagated errors from the continuum and iron emission fitting .
we then fit each of the simulated emission lines , keeping the number of gaussian functions fixed at the number found from fitting the original profile , and measured the @xmath18 for each simulated line .
the standard error on @xmath18 was then estimated as the standard deviation of the @xmath18 values measured from the simulated line profiles .
recently , reverberation mapping studies of broad line agn ( e.g. , * ? ? ?
* ) established a correlation between the broad line region ( blr ) size , @xmath102 , and the continuum luminosity ( the @xmath102@xmath103 relationship , e.g. , * ? ? ?
* ; * ? ? ?
this has made it possible to estimate black hole virial mass @xmath104 for individual sources , where the blr velocity @xmath105 is estimated from the width of an emission line ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we choose the proportionality constant to give broad line mass estimates consistent with the @xmath2@xmath106 relationship @xcite , @xmath107 @xcite .
an estimate of the eddington luminosity can be computed as @xmath108 . in this work
we estimate @xmath2 from the h@xmath3 , mg ii , and c iv emission lines .
we use the relationship of @xcite to estimate @xmath2 from the h@xmath3 and c iv emission lines , and the datails of the mg ii calibration will be discussed in a forthcoming paper ( * ? ? ? * in preparation ) .
the calibration for the mg ii mass estimates was calculated to ensure that they are consistent with the mass estimates based on h@xmath3 and c iv .
we have 49 sources with both h@xmath3 and mg ii mass estimates , and 73 sources with both c iv and mg ii mass estimates .
both samples show consistent mass estimates between the different emission lines , within the intrinsic uncertainty in the broad line mass estimates ( @xmath109 dex ) .
we will denote the broad line mass estimates as @xmath110 , and @xmath111 .
it is important to distinguish between @xmath110 and @xmath2 , as @xmath112 is an estimate of @xmath2 derived from reverberation mapping , and thus in general @xmath113 .
the statistical uncertainty needs to be taken into account when analyzing correlations involving derived quantities like @xmath114 , as they can bias the results @xcite . the uncertainty in @xmath115 increases the formal statistical uncertainty in the broad line estimates of @xmath2 to @xmath116 dex .
our adopted formal uncertainty of @xmath116 dex is merely statistical , and additional systematic uncertainties in reverberation mapping may contribute @xcite . because our adopted uncertainty of @xmath116
describes the scatter in broad line mass estimates about the reverberation mapping estimates , as calibrated via the @xmath2@xmath106 relationship , the regression results found in this work should be understood as results that could have been obtained if we had reverberation - based @xmath2 for the sources in this work .
however , instead of reverberation - based mass estimates , we have broad line mass estimates with ` measurement error ' equal to @xmath117 dex with respect to the reverberation - based mass estimates , thus increasing the uncertainty from the regression analysis . for most sources , measurement errors on @xmath18 and @xmath118 did not significantly contribute to the uncertainty on @xmath2 .
if there were two emission lines in the same spectrum we averaged the two mass estimates , where the average was weighted by the uncertainties in the two estimates .
the distribution of @xmath2 as a function of @xmath14 for our sample is shown in figure [ f - mbh_vs_z ] .
the broad line masss estimates for the sources in our sample are reported in table [ t - sample ] .
a constant bolometric correction has been used in most previous studies involving the agn eddington ratio .
however , recent work by @xcite has suggested that bolometric corrections show a large spread with no obvious dependence on luminosity .
furthermore , these authors found evidence that the bolometric correction depends on the eddington ratio .
this implies that the error in the bolometric correction is correlated with eddington ratio , which therefore implies that the error in the estimated eddington ratio is correlated with the actual eddington ratio . an eddington ratio - dependent error in the bolometric correction my cause problems when using the estimated eddington ratios to infer correlations .
further difficulties with a constant bolometric correction are illustrated with figure [ f - alfox_redd ] . in [ f - alfox_redd ]
we plot @xmath0 as a function of @xmath119 and @xmath120 , where we estimate the eddington luminosity from the broad line mass estimates as @xmath121
. as with @xmath110 , we use the notation @xmath122 to emphasize that @xmath122 is an estimate of the true @xmath123 based on the broad line mass estimates , and therefore @xmath124 .
constant bolometric corrections are often applied to either the optical / uv or x - ray luminosity .
if a constant bolometric correction was valid for both @xmath16 and @xmath70 , then we would expect that @xmath125 .
however , while a correlation between @xmath17 and both @xmath119 and @xmath120 is apparent , they are of opposite sign . because the correlations are of opposite sign , it can not be true that both @xmath119 and @xmath120 are proportional to the eddington ratio , @xmath19 . because of the current significant uncertainty regarding rqq bolometric corrections , we take the conservative approach and merely compare @xmath0 and @xmath1 with @xmath10 and @xmath9 .
the estimated values of @xmath10 and @xmath126 for the sources in our sample are reported in table [ t - sample ] .
we can write @xmath127 and @xmath128 , where @xmath129 and @xmath130 are the inverses of the bolometric corrections for @xmath16 and @xmath70 , respectively .
the quantities @xmath129 and @xmath130 are proportional to the fraction of the bolometric luminosity emitted at 2500 and 2 kev .
then , correlations between either @xmath17 or @xmath1 and @xmath10 will result if @xmath17 or @xmath1 is correlated with @xmath129 , the eddington ratio , or both , and likewise for @xmath130 . while we do not use an estimate of the eddington ratio in our analysis , it is helpful to estimate the distribution of eddington ratios probed by our sample .
we assume the bolometric correction described in @xcite for the @xmath39 sources , and constant bolometric described in @xcite of @xmath131@xmath132 at @xmath133 . in figure [ f - mbh_vs_z ]
we also show the distribution of estimated eddington ratios as a function of @xmath14 . because the distribution of estimated @xmath19 is the true distribution of @xmath19 broadened by the distribution of errors in the estimates , our sample likely probes a smaller range in eddington ratio than that inferred from figure [ f - mbh_vs_z ] . therefore , at most our sample probes rqqs with eddington ratios @xmath134 .
we used our sample of 318 sources with estimates of @xmath2 to investigate the dependence of @xmath0 at a given black hole mass , @xmath135 and @xmath9 .
we use linear regression analysis in order to understand how @xmath0 varies with respect to these parameters .
we use the method of @xcite to estimate the regression parameters .
the method of @xcite accounts for measurement errors , non - detections , and intrinsic scatter .
in addition , @xcite adopts a bayesian approach , computing the posterior probability distribution of the parameters , given the observed data .
thus the uncertainties on the regression coefficients have a straight - forward interpretion , and do not rely on large - sample approximations .
many other methods , such as traditional maximum - likelihood , assume that the errors in the regression parameters follow a gaussian distribution , which is valid as the sample size approaches infinity .
however , this assumption is not necessarily valid for our finite sample size , especially in the presence of censoring ( i.e. , presence of upper / lower limits ) and significant measurement error .
the method of @xcite directly estimates the probability distribution of the regression parameters , and is therefore preferred .
we assess the simple 2-dimensional correlations between @xmath17 and @xmath136 , and @xmath9 , and compare with the @xmath0@xmath16 correlation .
the results from the regressions are @xmath137 where @xmath138 is the intrinsic dispersion in @xmath0 at a given @xmath139 or @xmath9 , @xmath140 is the linear correlation coefficient for @xmath0 and the respective independent variables , and the errors are quoted at the @xmath141 @xmath142 level .
all four relationships are significant , with rqqs becoming more x - ray quiet as @xmath143 or @xmath10 increases , and more x - ray loud as @xmath9 increases . because we have attempted to correct for the intrinsic statistical scatter in the broad line mass estimates , equations ( [ eq - alfox_luv])([eq - alfox_xedd ] )
refer to the intrinsic relationships involving @xmath2 , barring any systematic errors in reverberation mapping , and are not simply correlations between @xmath0 and the broad line mass estimates .
the estimated distributions of @xmath0 as a function of @xmath144 , and @xmath9 are shown in figure [ f - alfox_reg ] , along with the regression results .
the intrinsic dispersion in @xmath0 quantifies the magnitude of scatter in @xmath0 at a given @xmath145 or @xmath9 . because we have attempted to account for contribution to the scatter in @xmath0 resulting from measurement error and variability
, @xmath138 represents the dispersion in the real physical scatter in @xmath17 over the population of rqqs .
this ` residual ' scatter represents the amount of variation in @xmath0 that is not ` explained ' by variations in @xmath139 or @xmath126 , respectively . this intrinsic scatter in @xmath0 may be due to variations in accretion rate , viscosity , column density , and other quantities not included in our regression . in [ s - sample ]
we estimate that there are @xmath146 bal quasars in our sample at @xmath39 .
because these objects have the appearance of being x - ray weak , and because redshift is artificially correlated with luminosity and @xmath2 in a flux limited sample , we expect that the presence of unidentified bals at @xmath39 will produce an excess of x - ray weak objects at low @xmath16 and @xmath2 , thus flattening the inferred slopes .
inspection of the plot of @xmath0 and @xmath14 in figure [ f - lum_z_dist ] suggests an excess of x - ray weak objects at @xmath39 and @xmath147 , implying these objects are bal quasars .
we removed these 10 objects and refit the regressions .
omission of these objects resulted in a steepening of the slopes for the @xmath16 and @xmath2 regression , and a flattening of the slope for the @xmath10 regression .
in addition , the intrinsic dispersion in @xmath0 decreased for the @xmath16 and @xmath2 regressions , while it remained the same for the @xmath8 regression .
these changes were small ( @xmath148 ) and have no effect on our conclusions .
there was no difference in the results for the @xmath9 regression .
once can use equation ( [ eq - alfox ] ) to express the regression results ( eq . [ [ eq - alfox_luv]][[eq - alfox_xedd ] ] ) in the alternate form @xmath149 where the intrinsic dispersion in @xmath150 at a given @xmath139 and @xmath9 is @xmath151 and @xmath152 dex , respectively .
in contrast to equations ( [ eq - alfox_luv])([eq - alfox_xedd ] ) , we quote the @xmath153 ( @xmath154 ) uncertainties on the constants of proportionality , and the posterior standard deviations on the exponents . equations ( [ eq - lumrat_luv])([eq - lumrat_xredd ] ) may be more physically interpretable and allow easier comparison with models .
recent work has suggested a correlation between quasar x - ray spectral slope , @xmath157 and quasar eddington ratio as inferred from broad line mass estimates based on the h@xmath3 emission line ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the @xcite sample contains measurements of @xmath1 for 157 sources , and we were able to estimate black hole masses for 153 of them . in this section
we use these 153 rqqs to investigate the dependence of @xmath1 on @xmath28 and @xmath9 .
the distributions of @xmath1 as a function of estimated black hole mass , @xmath10 , and @xmath9 are shown in figure [ f - gamx_vs_bh ] for the entire sample , and in figure [ f - gamx_vs_bh_eline ]
seperately for each emission line .
while there does not appear to be a monotonic trend between @xmath1 and @xmath136 , or @xmath9 when using the entire sample , there is evidence for a trend between @xmath1 and these quantities when using the h@xmath3 line , and an opposite trend between @xmath1 and these quantities when using the c iv line .
we performed a linear regression of @xmath1 on @xmath158 and @xmath159
seperately for each emission line .
as before , we used the method of @xcite when performing the regression in order to correct for the intrinsic statistical uncertainty in the broad line estimates of @xmath2 .
the results for @xmath2 are @xmath160 the results for @xmath10 are @xmath161 and the results for @xmath9 are @xmath162 in these equations we have quoted the errors at @xmath141 @xmath142 confidence .
the probability distributions of the slope and intrinsic dispersion are shown in figure [ f - hbciv_reg ] .
the larger uncertainty in the results for the mg ii sample is likely caused by the more narrow range in @xmath163 and @xmath2 probed .
there are formally no significant linear correlations for the @xmath1@xmath2 relationship . however , there is a statistically significant difference between the h@xmath3 and c iv slopes , with @xmath164 of the posterior probability at @xmath165 , where @xmath166 denotes the @xmath1@xmath167 regression slope .
the probability distribution for the difference in slopes from the @xmath2 regression is shown in figure [ f - regdiff ] . the significant difference in the slope for the h@xmath3 and c
iv sample implies a nonlinear relationship between @xmath1 and @xmath168 , in spite of the fact that the h@xmath3 and c iv correlations themselves are not ` statistically significant ' .
results similar to the @xmath1@xmath169 regressions were found for the @xmath1@xmath10 and @xmath1@xmath170 regressions , but with opposite sign and higher statistical significance .
we performed monte carlo simulations as a consistency check on our inferred non - monotonicity of the @xmath1 relationships . while we have attempted to account for the significant statistical uncertainty on the broad line mass estimates , we employ these monte carlo simulations to ensure that the observed non - monotonic behavior is not a spurious result caused by the uncertainty on @xmath2 .
we performed @xmath171 simulations under two null hypotheses : ( 1 ) that @xmath1 is independent of @xmath10 , and ( 2 ) that @xmath1 depends linearly on @xmath172 .
for both cases we simulated black hole mass estimates derived from h@xmath3 and c iv seperately .
we first simulated ` true ' values of @xmath2 for each emission line from a normal distribution with means equal to the observed mean of the two respective subsamples , and variances equal to the difference between the observed variance of the subsamples and the average intrinsic variance in the broad line mass estimates . to simulate the uncertainty in the mass estimates , we added random gaussian errors to these ` true ' values of @xmath2 with standard deviation equal to the uncertainty in the mass estimates , @xmath109 dex . for the case where @xmath1 was assumed to be independent of @xmath10 , we
simulated values of @xmath1 from a normal distribution with mean equal to the sample mean of @xmath1 and variance equal to the difference between the observed variance in @xmath1 and the average of the variance in the measurement errors . for the case where @xmath1 was assumed to depend linearly on @xmath172 , we simulated values of @xmath1 according to our best fit relationship to the h@xmath3 subsample , given by equation ( [ eq - gamxreddu_hbeta ] ) . finally , for both cases we added random gaussian errors to the simulated values of @xmath1 by randomly reshuffling the dispersions in the measurement errors in @xmath1 . for each of the @xmath171 simulated samples
, we selected those samples that displayed a non - monotonic trend , i.e. , those sample where the slope for the h@xmath3-based regression had a different sign from the slope of the c iv - based regression . under the hypothesis that @xmath1 is independent of @xmath10 , only 3 of the @xmath171 simulated samples had both a non - monotonic trend and an absolute value of the difference in slopes between the h@xmath3 and c iv regression that were larger than that observed for our actual sample . under the hypothesis that @xmath1 depends linearly on @xmath173 , none of the @xmath171 simulated samples exhibited a non - monotonic trend
therefore , the observed non - monotonic trend in @xmath1 with eddington ratio is not a spurious result caused by the statistical uncertainty in the broad line mass estimates , in agreement with our bayesian regression results . in order to investigate whether the non - monotonicity in the dependence of @xmath1 on eddington ratio depends on @xmath2
, we performed a linear regression of @xmath1 simultaneously on @xmath173 and @xmath169 .
this also allows us to quantify whether the eddington ratio is the driver behind the @xmath1@xmath2 relationship .
in particular , the @xmath1@xmath2 relationship is weak compared to the eddington ratio relationships , and therefore it is reasonable to conclude that eddington ratio is the primary driver in these relationships .
we applied the multiple regression technique of @xcite
seperately to both the h@xmath3 and c iv subsamples .
the results are : @xmath174 here , we have quoted the errors at @xmath141 significance .
there is no statistically significant evidence that @xmath1 depends on @xmath2 at a given @xmath10 , and therefore we conclude that the primary driver in the @xmath1 relationships is eddington ratio .
however , this does not rule out the possibility that the non - monotonic trends with eddington ratio are the result of a discontinuous change in the slopes at a ` critical ' @xmath2 , as discussed in the next two sections .
the opposite correlations for h@xmath3 and c iv are intriguing but may represent problems with the broad line mass estimates . in particular , it is possible that the error in the broad line mass estimates is correlated with @xmath1 , but in opposite ways for h@xmath3 and c iv .
the most likely source of such a spurious correlation would be a correlation between @xmath1 and the scatter about the @xmath102@xmath103 relationship for h@xmath3 and c iv , respectively .
for example , if one were to systematically overestimate @xmath102 with increasing @xmath1 for the c iv emitting region , then one would infer a larger @xmath2 from c iv , and thus one would infer a spurious correlation between @xmath2 and @xmath1 .
however , the h@xmath3 line is only available at @xmath175 and the c iv line is only available at @xmath176 , and thus the change in sign for the @xmath1@xmath2 correlations could be due to different spectral components shifting into or out of the observable x - ray spectral region ( 0.37.0 kev ) . in figure [ f - mbh_vs_z2
] , we show the distribution of @xmath2 as a function of @xmath14 for the h@xmath3 and c iv samples .
as is clear from figure [ f - mbh_vs_z2 ] , the c iv line is probing sources with @xmath177 , while the h@xmath3 line is probing sources with @xmath178 .
therefore , the change in sign for the @xmath1 correlations could also be due to something more physically interesting , such as a change in the structure of the corona that occurs at some critical black hole mass or accretion rate . we can test if the change in sign for the @xmath1 correlations is the result of problems with the mass estimates for either h@xmath3 or c iv , or if it is the result of the differences in @xmath14 and @xmath2 probed by the two lines .
while the @xmath1@xmath2 relationship is weak compared to the eddington ratio dependencies , we can use the @xmath1@xmath2 relationship to test whether the non - monotonic dependency of @xmath1 on eddington ratio is a spurious result caused by systematic difference in the broad line mass estimates .
this is because the eddington ratios are inferred from the broad line mass estimates , and any systematic differences between the eddington ratio inferred from h@xmath3 as compared to c iv should also manifest themselves in the weaker @xmath1@xmath2 relationship .
this is true irregardless of whether the difference in slopes between the h@xmath3 and c iv @xmath1@xmath2 relationships is ` statistically significant ' or not .
we compiled five sources from the literature at @xmath179 with h@xmath3-based mass estimates of @xmath180 .
in addition , we compiled five more sources from the literature at @xmath181 and c iv - based mass estimates of @xmath182 . this ` test sample ' of 10 sources is listed in table [ t - test ] .
then , we test whether the h@xmath3 test sources are better described by the h@xmath3 regression or by the c iv regression , and likewise for the c iv test sources .
if the change in sign for the @xmath1 correlations is due to problems with the broad line mass estimates , then we would expect the h@xmath3-based mass estimates to be better described by the h@xmath3 regression .
however , if the change in sign is due to the difference in redshift and @xmath2 probed by the two regressions , then we would expect the h@xmath3 test sources to be better described by the c iv regression , as the h@xmath3 test sources are at high-@xmath14 and have high-@xmath2 .
a similar argument applies to the c iv test sources , since they are at low-@xmath14 and have low black hole masses .
figure [ f - test1 ] compares @xmath1 and @xmath2 for h@xmath3 and c iv for both the sources in our main sample and the test sources , as well as the best fit regression lines for the h@xmath3 and c iv samples , respectively .
the high-@xmath14 , high-@xmath2 h@xmath3 test sources appear to be better described by the high-@xmath14 , high-@xmath2 c iv - based regression , and likewise the low-@xmath14 , low-@xmath2 c iv test sources appear to be better described by the low-@xmath14 , low-@xmath2 h@xmath3-based regression .
we can quantify this result by calculating the probability that the h@xmath3 test sources ` belong ' to the h@xmath3-based regression , as compared to the probability that the h@xmath3 test sources ` belong ' to the c iv - based regression . assuming that the test sources are as equally likely to belong to either regression _ a priori _ ,
this ratio of probabilities is simply the ratio of the likelihood functions of the test sources for each regression relationship , where the likelihood functions are given by equation ( 24 ) in @xcite ; this ratio is called the ` bayes factor ' ( e.g. , * ? ? ?
in order to incorporate our uncertainty in the regression parameters , we use the value of the likelihood function averaged over the probability distribution of the regression parameters .
we find that the h@xmath3 test sources are @xmath183 times more likely to ` belong ' to the c iv - based regression , and that the c iv test sources are @xmath184 times more likely to ` belong ' to the h@xmath3-based regression .
because the test sources are independent , it follows that the test sources are @xmath185 times more likely to be described by the regression fit using the opposite emission line sample .
this is further evidence that @xmath1 and @xmath2 are not statistically independent ; if @xmath1 and @xmath2 were independent , then the test sources would not show a strong preference for either regression . based on this analysis , we conclude that the change in sign of the @xmath1 correlations is not due to problems associated with the use of the h@xmath3 and c iv emission lines , but rather due to the different range of @xmath14 and @xmath2 probed by the two subsamples . while it appears that @xmath1 has a nonmonotonic dependence on @xmath28 and @xmath9 , it is unclear as to whether the sign change in the correlations is dependent on @xmath14 or @xmath2 .
however , we can test this in the same manner as was used to test if the sign change is due to problems with the h@xmath3- or c iv - based mass estimates . in this case
, we need a sample of high-@xmath14 , low-@xmath2 sources in order to break the degeneracy between @xmath2 and @xmath14 .
if the sign change in the correlation is redshift dependent , then we would expect the test sources to be better described by the high-@xmath14 , high-@xmath2 c iv - based regression ; but , if the sign change is black hole mass dependent , then we would expect the test sources to be better described by the low-@xmath14 , low-@xmath2 h@xmath3-based regression .
our test sample consists of nine @xmath186 quasars from the cosmic evolution survey ( cosmos , * ? ? ?
* ) , with optical spectra from magellan @xcite and x - ray spectra from xmm - newton @xcite .
black hole masses for these objects ( * ? ? ? * in preparation ) were estimated from the mg ii and c iv emission lines in the same manner as above .
the objects are summarized in table [ t - cosmos ] , and their location in the @xmath2@xmath14 plane are shown in figure [ f - mbh_vs_z_cosmos ] .
the cosmos sources break the degeneracy between @xmath2 and @xmath14 present in our sdss sample , and are therefore adequate to test for a redshift dependence in the slope of the @xmath1@xmath2 relationship . in figure [ f - test2 ]
we compare the cosmos test sources with the h@xmath3- and c iv - based regressions .
as can be seen , the high-@xmath14 , low-@xmath2 cosmos sources are better described by the low-@xmath14 , low-@xmath2 regression .
we can quantify this in the same manner as described in
[ s - eline_test ] by averaging the likelihood function of the test sources over the posterior probability distribution .
we find that the cosmos test sources are @xmath187 times more likely to be better described by the h@xmath3-@xmath1 regression , and thus the sign change in the @xmath1 correlations is not due to the difference in redshifts probed by the h@xmath3 and c iv samples .
previous work has found evidence for a correlation between @xmath0 and both @xmath2 @xcite and @xmath188 @xcite , and for a correlation between @xmath1 and both @xmath2 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and @xmath188 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , in aggreement with the results found in this work .
however , our study differs from previous work in that we study a large sample of rqqs ( 318 sources with @xmath0 , 153 with @xmath1 ) over a broad range in black hole mass ( @xmath189 ) and redshift @xmath25 ; to date , this is the largest study of the dependence of the x - ray properties of rqqs on @xmath27 and @xmath9 .
in addition , this the first study of its kind to correct for the intrinsic statistical uncertainty in broad line mass estimates when quantifying the _ intrinsic _ trends between the x - ray emission and @xmath28 and @xmath126 .
currently , there are two main types of geometries being considered for the comptonizing corona .
the first of these is that of a ` slab'-type geometry , possibly patchy , that sandwiches the disk ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and the second is that of a hot spherical inner advection dominated flow ( e.g. , * ? ? ?
* ; * ? ? ?
* ) ; hybrids between the two geometries have also been considered ( e.g. , * ? ? ?
* ; * ? ? ?
there is a growing body of evidence that the advection dominated hot inner flow does not exist in objects with eddington ratios @xmath190 , as inferred from the existence of a relativistically broadened iron line ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , relativistically broadened reflection of ionized material @xcite , and by analogy with galactic black holes ( e.g. , * ? ? ?
* ; * ? ? ?
the range in eddington ratios probed by our study is at most @xmath134 , with a mean of @xmath191 .
therefore , the rqqs in our study are likely to have disks that extend approximately down to the last marginally stable orbit , and thus should only have the ` slab ' type geometries . in this work
we have found that rqqs become more x - ray quiet as @xmath2 increases , and confirmed the well - established relationship between @xmath0 and @xmath16 . because @xmath16 increases with @xmath2 and the accretion rate relative to eddington , @xmath12 ,
the well - known @xmath0@xmath16 correlation is likely driven by the @xmath0@xmath2 and @xmath17@xmath12 correlations . a correlation between @xmath0 and @xmath2
is expected even if the fraction of the bolometric luminosity emitted by the disk is independent of @xmath2 , as the effective temperature of the disk depends on @xmath2 . as @xmath2 increases , the effective temperature of the disk decreases , thus shifting the peak of the disk emission toward longer wavelengths .
because the flux density at @xmath64 lies redward of the peak in the disk sed over most of the range @xmath2 probed by our study , this shift in the disk sed toward longer wavelengths produces an increase in @xmath16 relative to @xmath70 .
we can use the standard thin disk solution to assess the evidence that the fraction of energy emitted by the corona depends on @xmath2 .
we assume a simple model where the spectrum for the disk emission is that expected for an extended thin accretion disk , and the spectrum for the corona emission is a simple power - law with exponential cutoffs at the low and high energy end . according to @xcite , the spectrum from a radially
extended thin accretion disk can be approximated as @xmath192 where @xmath193 is the normalization and @xmath194 is the cut - off frequency . in this work
, we choose the normalization to ensure that equation ( [ eq - thindisk ] ) integrates to unity , and therefore @xmath195 gives the shape of the disk emission . for a kerr black hole
, @xcite finds that the cut - off frequency is related to @xmath2 as @xmath196 we assume that the x - ray emission from the corona can be described by a simple power law with an exponential cutoff at the high and low end : @xmath197 here , @xmath198 is the corona spectrum normalization , @xmath199 is the high energy cutoff , and @xmath200 is the low energy cutoff .
we choose the low energy cutoff to be @xmath201 , and we choose the high energy cutoff to be @xmath202 ( e.g. , * ? ? ?
as with equation ( [ eq - thindisk ] ) , we choose the normalization in equation ( [ eq - corona ] ) to be equal to unity . denoting @xmath203 to be the fraction of bolometric luminosity emitted by the disk , our model rqq spectrum is then @xmath204 .
\label{eq - modelspec}\ ] ] we computed equation ( [ eq - modelspec ] ) assuming a value of @xmath205 and @xmath206 .
we chose the value of the @xmath207 because it is representative of the rqqs in our sample , and we chose the value @xmath206 because it gives values of @xmath0 typical of the rqqs in our sample .
we vary @xmath2 but keep @xmath203 and @xmath12 constant because we are interested in investigating whether there is evidence that assuming independence between @xmath2 and both @xmath203 and @xmath12 is inconsistent with our @xmath0 results .
we compute equation ( [ eq - modelspec ] ) for two forms of the dependence of @xmath1 on @xmath2 .
for the first model , we assume a constant value of @xmath208 . for the second model , we assume that @xmath1 depends on @xmath2 according to our best fit regression results , where @xmath1 depends on @xmath2 according to equation ( [ eq - gamxmbh_hbeta ] ) for @xmath209 , and @xmath1 depends on @xmath2 according to equation ( [ eq - gamxmbh_civ ] ) for @xmath210 .
we ignore the intrinsic dispersion in @xmath1 . in figure [ f - modelspec ]
we show the spectra computed from equation ( [ eq - modelspec ] ) for rqqs with @xmath211 and @xmath212 . the dependence of the location of the peak in the disk emission on @xmath2 is clearly illustrated . in figure [ f - alfox_model ]
we compare the @xmath17@xmath2 regression results with the dependence of @xmath17 on @xmath2 expected from equation ( [ eq - modelspec ] ) for both @xmath1@xmath2 models . under the thin disk approximation
, a correlation is expected between @xmath0 and @xmath2 , even if the fraction of bolometric luminosity emitted by the disk is independent of @xmath2 .
however , our data are inconsistent with the assumption that @xmath203 and @xmath2 are independent , given the thin disk approximation . under the assumption that @xmath203 and @xmath2 are independent , the @xmath0@xmath2 correlation is too flat , and a increase in the fraction of bolometric luminosity emitted by the disk with increasing @xmath2 is needed to match the steeper observed dependence of @xmath0 on @xmath2 . alternatively , if @xmath203 increases with increasing @xmath12 , as we argue in [ s - alfox_mdot ] , then a steeper @xmath0@xmath2 correlation also results if @xmath2 and @xmath12 are correlated . in this case , if @xmath203 increases with increasing @xmath12 , and if @xmath2 increases with increasing @xmath12 , then @xmath203 will also increase with increasing @xmath2 , thus producing a steeper observed @xmath0@xmath2 correlation . to the extant that equations ( [ eq - thindisk])([eq - modelspec ] ) accurately approximate the spectral shape of rqqs ,
our data imply that either the fraction of bolometric luminosity emitted by the disk increases with increasing @xmath2 , that @xmath2 and @xmath12 are correlated , or both .
some theoretical models have suggested that the fraction of bolometric luminosity emitted by the disk should depend on @xmath12 , but be relatively insensitive to @xmath2 ( e.g. , * ? ? ?
* ; * ? ? ?
therefore , while a significant dependence of @xmath203 on @xmath2 is not predicted by these disk / corona models , these models are still consistent with the interpretation that a @xmath2@xmath12 correlation is driving the steeper @xmath17@xmath2 correlation .
unfortunately , without accurate estimates of @xmath12 we are unable to distinguish between these two possibilities .
a shift in the peak of the disk sed with @xmath2 may also explain the dependence of @xmath0 on redshift observed by k07 .
k07 speculated that the observed hardening of @xmath0 with increasing @xmath14 at a given @xmath16 may be due to a correlation between @xmath0 and @xmath2 , manifested through a @xmath2@xmath14 correlation . at a given @xmath16 ,
an increase in @xmath2 will result in an increase in @xmath70 relative to @xmath16 , assuming that @xmath12 is not strongly correlated with @xmath2 .
this is because an increase in @xmath2 decreases the temperature of the disk , shifting the peak in the disk sed toward the red , and thus increasing the luminosity at 2500 .
however , since k07 investigated the dependence of @xmath0 on @xmath14 at a given @xmath16 , the luminosity at 2500 is held constant .
therefore , the overall disk emission must decrease in order to keep the luminosity at 2500 constant despite the increase in @xmath2 . as a result ,
an increase in @xmath2 at a given @xmath16 will result in an increase in the x - ray luminosity relative to the luminosity at 2500 . because @xmath2 and @xmath14 are correlated in our flux limited sample ( e.g. , see figure [ f - mbh_vs_z ] ) , an increase in @xmath14 will probe rqqs with higher @xmath2 . as a result , rqqs will become more x - ray loud with increasing @xmath14 , at a given 2500 luminosity .
consequently , deeper surveys that probe a greater range of @xmath2 should not see as strong of a correlation between @xmath2 and @xmath14 , thereby reducing the magnitude of a @xmath0@xmath14 correlation .
indeed , investigations based on samples that span a greater range in luminosity do not find evidence for a correlation between @xmath17 and @xmath14 ( e.g. , * ? ? ? * ; * ? ? ?
* ) , qualitatively consistent with our interpretation of a @xmath0@xmath14 correlation .
we have found that @xmath0 increases with increasing @xmath213 , and decreases with increasing @xmath9 .
the mere existence of these correlations is not particularly interesting , as we would expect that the ratio of optical / uv luminosity to x - ray luminosity would increase as the fraction of optical / uv luminosity relative to eddington increases , and vice versa for an increase in @xmath9 .
however , the relative magnitude of these dependencies carries some information regarding the dependence of @xmath0 on @xmath12 . a correlation between @xmath0 and
@xmath8 implies that @xmath214 increases as the quantity @xmath215 increases , where @xmath129 is the fraction of bolometric luminosity emitted at 2500 .
likewise , an anti - correlation between @xmath0 and @xmath9 implies that @xmath214 decreases as the quantity @xmath216 increases , where @xmath130 is the fraction of bolometric luminosity emitted at 2 kev .
if the fraction of the bolometric luminosity emitted by the disk , as compared to the corona , increases with increasing @xmath12 , then we would expect a strong increase in @xmath214 with the product @xmath217 , resulting from the dual dependency of @xmath214 on @xmath129 and @xmath12 .
furthermore , because the fraction of bolometric luminosity emitted by the disk should decrease with increasing @xmath130 , then , if the fraction of bolometric luminosity emitted by the disk increases with increasing @xmath12 , we would expect a weaker dependence of @xmath214 on the quantity @xmath216 .
this is because an increase in @xmath12 causes an increase in the disk emission relative to the corona emission , which will then work against the decrease in disk emission relative to the corona that results from an increase in @xmath130 .
the end result is a weaker dependence of @xmath214 on the product @xmath216 .
indeed , this is what we observe , where @xmath218 and @xmath219 .
therefore , we conclude that the disk emission relative to the corona emission increases with increasing @xmath12 .
this is in agreement with some models of corona with a slab geometry ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , where the @xmath0@xmath12 correlation arises due to a dependency of the size of the corona on @xmath12 . our result that @xmath0 is correlated with @xmath8 and anti - correlated with @xmath9 is inconsistent with a constant bolometric correction to both the optical / uv and x - ray luminosities
. instead , an increase in @xmath214 with increasing @xmath12 implies that the bolometric correction depends on @xmath12 . because we conclude that the fraction of bolometric luminosity emitted by the disk increases with increasing @xmath12 , this implies that the bolometric correction to the optical / uv luminosity decreases with increasing @xmath12 , while the bolometric correction to the x - ray luminosity increases with increasing @xmath12 .
the direction of this trend is consistent with the results of @xcite , who find that the bolometric correction to the x - ray luminosity increases with increasing @xmath188 .
similarly , we have found evidence that the fraction of bolometric luminosity emitted by the disk depends on @xmath2 , therefore implying that the bolometric correction also depends on @xmath2 .
even if the fraction of bolometric luminosity emitted by the disk is independent of @xmath2 , the bolometric correction will still depend on @xmath2 because the location of the peak in the disk emission will shift toward longer wavelengths as @xmath2 increases .
as @xmath2 varies , the luminosity at 2500 probes a different region of the quasi - blackbody disk emission , thereby producing a dependence of bolometric correction on @xmath2 . a significant amount of recent work suggests that radiative and mechanical feedback energy from agn plays an important part in galaxy and supermassive black hole coevolution ( e.g. , * ?
* ; * ? ? ?
* ; * ? ? ?
* ) . within the context of these models , a nuclear inflow of gas ,
possibly the result of a galaxy merger , feeds the smbh , thus igniting a quasar .
the smbh grows until feedback energy from the quasar is able to drive out the accreting gas , thus halting the accretion process .
hydrodynamic calculations of accretion flows have shown that the efficiency of the quasar in driving an outflow depends on the fraction of energy emitted through he uv / disk component as compared to the x - ray / corona component @xcite .
the disk component produces luminosity in the uv , which is responsible for driving an outflow via radiation pressure on lines , whereas the corona component produces luminosity in the x - rays , which is responsible for driving an outflow via thermal expansion .
calculations by @xcite have shown that radiation driving produces an outflow that carries more mass and energy than thermal driving .
if the efficiency of black hole feedback depends on the quasar sed , any dependence on @xmath2 and @xmath12 of the fraction of agn energy emitted in the uv as compared to the x - ray has important consequences for models of black hole growth . because we have found evidence that the fraction of bolometric luminosity emitted by the disk increases with increasing @xmath12 and @xmath2 , this implies that black holes becomes more efficient at driving an outflow with increasing @xmath12 and @xmath2 .
however , the @xmath0@xmath2 correlation may be due to the combination of both a correlation between @xmath2 and @xmath12 , and a dependence of the location peak in the disk sed on @xmath2 . if the fraction of energy emitted by the disk only depends weakly on @xmath2 , as some theoretical models have suggested ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , the fraction of energy emitted in the uv will still decrease with increasing @xmath2 becuase the peak of the disk emission will shift away from the uv . in this case , at a given @xmath12 we would expect that black holes will become less efficient at driving an outflow with increasing @xmath2 . in this work we have also found evidence that @xmath1 and @xmath220 and @xmath9 are not statistically independent .
moreover , the dependence of @xmath1 on black hole mass or eddington ratio appears to follow a non - monotonic form , although the @xmath1@xmath2 trend is weak compared to the dependency of @xmath1 on eddington ratio . for the @xmath1@xmath2 relationship , the x - ray continuum hardens with increasing black hole mass until @xmath221 , after which the x - ray continuum softens with increasing black hole mass .
the opposite is true for the @xmath1@xmath10 and @xmath1@xmath126 trends , and further work is needed to confirm this result .
previous studies have not seen this non - monotonic trend because they have only employed the h@xmath3 emission line , and therefore their samples have been dominated by low-@xmath14 , low-@xmath2 sources .
it is unlikely that the dependence of @xmath1 on @xmath27 or @xmath9 is due to redshifting of the observable spectra range . if this were the case , then as @xmath2 increases ,
so does @xmath14 due to selection effects , and thus we would observe a decrease in @xmath1 as the ` soft excess ' shifts out of the observed 0.37 kev spectral range , while the compton reflection component shifts into the observed spectral range
. however , there are lines of evidence that suggest that the @xmath1 correlation are not due to redshifting of the observable spectral region , and that at least some of the observed dependency of @xmath1 on @xmath27 and @xmath9 is real . first , in
[ s - ztest ] we tested whether a sample of nine @xmath222 test sources with @xmath223 were better described by a regression fit using the @xmath224 sources , or by a regression fit using the @xmath225 sources .
we found that the test sources were better fit using the regression of similar @xmath2 , and therefore that the difference in the @xmath1@xmath2 correlations primarily depends on @xmath2 .
second , similar trends at low redshift between @xmath1 and @xmath2 or @xmath19 have been seen in other studies that only analyze the hard x - ray spectral slope ( typically 212 kev , e.g. , * ? ? ?
* ; * ? ? ?
* ) , and thus these studies are not effected by the soft excess .
third , the compton reflection hump is unlikely to shift into the observabable spectral range until @xmath226 .
however , the contribution to the inferred @xmath1 from compton reflection at @xmath227 is likely weak , if not negligible , as our @xmath227 sources have @xmath228 and are highly luminous , and therefore are expected to only have weak reflection components @xcite .
there are two scenarios in which the non - monotonic behavior of @xmath1 with @xmath2 or eddington ratio may be artificially caused by selection .
we will focus on the eddington ratio dependency , as it is the strongest ; however , our argument also applies to @xmath2 .
first , the intrinsic dependency of @xmath1 on eddington ratio could be linear with increasing intrinsic scatter at high @xmath19 .
then , an inferred non - monotonic trend would occur if we were to systematically miss quasars with high @xmath188 and steep x - ray spectra .
alternatively , there could be no intrinsic dependency of @xmath1 on eddington ratio . in this case
, we would infer a non - monotonic trend if we were to systematically miss quasars with steep x - ray spectra at low and high @xmath19 , and quasars with flat x - ray spectra at moderate @xmath19 .
we do not consider it likely that the observed non - monotonic dependence of @xmath1 on eddington ratio is due solely to selection effects .
k07 describes the sample selection for sources with @xmath1 . with the exception of some of the @xmath229 quasars , all sources from k07
were selected by cross - correlating the sdss dr3 quasars with public _
chandra _ observations .
almost all sdss sources in k07 had serendipitious _
chandra _ observations , and therefore were selected without regard to their x - ray properties .
k07 estimated @xmath1 for all sources that were detected in x - rays at the level of @xmath230 or higher .
therefore , the only additional criterion beyond the sdss selection imposed by k07 is the requirement that the source had to be detected in x - ray , which was fulfilled by @xmath231 of the quasars ; the undetected sources were slightly more likely to be found at lower redshift , probably due to the presence of unidentified bal quasars . as a result ,
the k07 sample selection function is essentially equivalent to the sdss quasar selection function .
because the sdss selects quasars based on their optical colors , the most likely cause of selection effects is the optical color selection .
there is evidence that @xmath1 is correlated with the slope of the optical continuum , where the x - ray continuum flattens ( hardens ) as the optical continuum steepens ( softens ) ( * ? ? ?
* k07 ) . the sdss selection probability is lower for red sources @xcite , so we might expect to systematically miss sources with smaller @xmath1 .
however , for the two scenarios described above , this is opposite the trend needed to explain the @xmath1@xmath9 relationship , where we need to at least systematically miss sources with larger @xmath1 .
furthermore , the drop in sdss selection efficiency with optical spectral slope only occurs at @xmath232 @xcite , thus we would expect a redshift dependence for this selection effect . as we have argued above , and in [ s - ztest ] , the non - monotonic trends for @xmath1 can not be completely explained as the result of different redshift ranges being probed .
the dependence of @xmath1 on @xmath10 and @xmath9 is likely due to a dependence of @xmath1 on @xmath12 .
if these @xmath1 correlations were due to a dependence of @xmath1 on @xmath129 or @xmath130 , then we would expect opposite trends for @xmath8 and @xmath9 , as @xmath129 and @xmath130 should be anti - correlated .
the fact that the regression results for the @xmath1@xmath10 and @xmath1@xmath9 relationships are similar implies that @xmath1 depends on @xmath12 , and at most only weakly on @xmath129 or @xmath130 .
a non - monotonic dependence of @xmath1 on @xmath12 is predicted from the accreting corona model of @xcite , as well as a non - monotonic dependence of @xmath1 on the viscosity @xcite .
in addition , @xmath1 is expected to steepen with increasing optical depth ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
one could then speculate that the dependence of @xmath1 on @xmath2 or @xmath12 is due to a non - monotonic dependence of the corona optical depth on @xmath12 , which may indicate a change in the structure of the disk / corona system at @xmath233 or some critical @xmath12 .
recent work also suggests a non - monotonic dependence of the optical / uv spectral slope , @xmath234 , on @xmath12 @xcite . from this work
, it has been inferred that the optical / uv continuum becomes more red with increasing @xmath12 until @xmath235 , after which the optical / uv continuum becomes more blue with increasing @xmath12 .
assuming the bolometric corrections described in
[ s - eddrat ] , the turnover in the @xmath1@xmath19 relationship also occurs at @xmath235 .
@xcite suggested that the turnover in spectral slope at @xmath236 may be due to a change in accretion disk structure , where the inner part of the accretion disk becomes thicker due to increased radiation pressure @xcite .
@xcite performed a simple approximation to this ` slim disk ' solution and found that it is able to produce a non - monotonic trend between optical color and eddington ration .
if the inner disk structure changes at high @xmath12 , this change could alter the corona structure , producing the observed trend between @xmath1 and eddington ratio .
unfortunately , current models for corona geometry make a number of simplifying assumptions , and do not yet predict a specific relationship between @xmath237 and @xmath12 . ideally , full magneto - hydrodynamic simulations ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ) that include accretion disk winds ( e.g. , * ? ? ?
* ; * ? ? ?
* ) should be used to interpret the results found in this work . however
, mhd simulations have not advanced to the point where they predict the dependence of @xmath0 and @xmath1 on quasar fundamental parameters , but hopefully recent progress in analytical descriptions of the magneto - rotational instability @xcite will help to overcome some of the computational difficulties and facilitate further advancement .
in this work we have investigated the dependence of @xmath0 and @xmath1 on black hole mass and eddington ratio using a sample of 318 radio - quiet quasars with x - ray data from _ rosat _ @xcite and _ chandra _ @xcite , and optical data mostly from the sdss ; 153 of these sources have estimates of @xmath1 from _
chandra_. our sample spans a broad range in black hole mass ( @xmath189 ) , redshift ( @xmath25 ) , and luminosity ( @xmath238@xmath7 \lesssim 10^{48}$ ] ) . to date , this is the largest study of the dependence of rqq x - ray parameters on @xmath28 and @xmath9 .
our main results are summarized as follows : * we show that @xmath0 is correlated with @xmath10 and anti - correlated with @xmath9 .
this result is inconsistent with a constant bolometric correction being applicable to both the optical / uv luminosity and the x - ray luminosity .
this result , when taken in combination with recent work by @xcite , implies that constant bolometric corrections can be considerably unreliable and lead to biased results . instead , we argue that @xmath214 increases with increasing @xmath12 and increasing @xmath2 , therefore implying that the bolometric correction depends on @xmath12 and @xmath2 . *
we performed a linear regression of @xmath0 on luminosity , black hole mass , @xmath10 , and @xmath9 , and found significant evidence that @xmath0 depends on all four quantities : @xmath239 and @xmath240 .
the dependence of @xmath241 on @xmath16 may be due to the dual dependence of @xmath241 on @xmath2 and @xmath12 . because we have attempted to correct for the statistical uncertainties in @xmath0 and the broad line estimates of @xmath2
, these results refer to the _ intrinsic _ relationships involving @xmath0 and @xmath2 , and are not merely the relationships between @xmath241 and the broad line mass estimates , @xmath242 . *
a correlation between @xmath0 and @xmath2 is expected from the fact that the peak in the disk emission will shift to longer wavelengths as @xmath2 increases , even if the fraction of the bolometric luminosity emitted by the disk does not change with @xmath2 . using a simple model for rqq spectra
, we argue that the observed @xmath0@xmath2 correlation is steeper than that expected if both @xmath12 and the fraction of bolometric luminosity produced by the disk are independent of @xmath2 .
the observed @xmath0@xmath2 relationship therefore implies that either the fraction of bolometric luminosity emitted by the disk increases with increasing @xmath2 , that @xmath2 is correlated with @xmath12 , or both . * a correlation between @xmath0 and @xmath12
is predicted from several models of ` slab'-type corona .
we argue that the weaker dependence of @xmath0 on @xmath9 implies that @xmath214 increases with increasing @xmath12 .
considering that the efficiency of quasar feedback energy in driving an outflow may depend on the ratio of uv to x - ray luminosity , a correlation between @xmath0 and both @xmath2 and @xmath12 has important consequences for models of black hole growth .
in particular , if supermassive black holes become more x - ray quiet at higher @xmath12 , they will become more efficient at driving away their accreting gas , thus halting their growth . * because of a possible nonlinear dependence of @xmath1 on @xmath243 , or @xmath9 , we performed seperate regressions for the black hole mass estimates obtained from each emission line .
we confirmed the significant dependence of @xmath1 on @xmath10 and @xmath9 seen in previous studies as inferred from the broad line mass estimates based on the h@xmath3 line ; however , we also find evidence that the @xmath1 correlations change direction when including the c iv line .
in particular , for the h@xmath3 sample , the x - ray continuum hardens with increasing @xmath2 , while for the c iv sample , the x - ray continuum softens with increasing @xmath2 . similar but opposite trends are seen with respect to @xmath10 and @xmath244 , and we conclude that these relationships can be interpreted as resulting from a correlation between @xmath1 and @xmath12 .
results obtained from the mg ii line were too uncertain to interpret .
we analyzed two test samples to argue that this non - monotonic behavior is not due to the different redshifts probed by the two samples , or to problems with the estimates of @xmath2 derived from the two lines ; the different trends may be due to the difference in @xmath2 probed by the two samples .
a non - monotonic dependence of @xmath1 on @xmath2 and/or @xmath12 may imply a change in the disk / corona structure , although a non - monotonic dependence of @xmath1 on @xmath12 and the viscosity is predicted by some models of ` slab'-type coronal geometries .
we would like to thank the anonymous referee for a careful reading and comments that contributed to the improvement of this paper .
this research was funded in part by nsf grant ast03 - 07384 , nasa contract nas8 - 39073 , and chandra award number g05 - 6113x issued by the chandra x - ray observatory center .
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 .
abramowicz , m. a. , czerny , b. , lasota , j. p. , & szuszkiewicz , e. 1988 , , 332 , 646 akritas , m. g. , & bershady , m. a. 1996 , , 470 , 706 anderson , s. f. , et al .
2001 , , 122 , 503 avni , y. , & tananbaum , h. 1982 , , 262 , l17 ballantyne , d. r. , ross , r. r. , & fabian , a. c. 2001 , , 327 , 10 bechtold , j. , et al .
2003 , , 588 , 119 begelman , m. c. , & nath , b. b. 2005 , , 361 , 1387 bentz , m. c. , peterson , b. m. , pogge , r. w. , vestergaard , m. , & onken , c. a. 2006 , , 644 , 133 bian , w .- h .
2005 , chinese journal of astronomy and astrophysics supplement , 5 , 289 bianchi , s. , guainazzi , m. , matt , g. , & fonseca bonilla , n. 2007 , , 467 , l19 bisnovatyi - kogan , g. s. , & blinnikov , s. i. 1977 , , 59 , 111 boller , t. , brandt , w. n. , & fink , h. 1996 , , 305 , 53 bonning , e. w. , cheng , l. , shields , g. a. , salviander , s. , & gebhardt , k. 2007 , , 659 , 211 brandt , w. n. , mathur , s. , & elvis , m. 1997 , , 285 , l25 brandt , n. , & boller , t. 1998 , astronomische nachrichten , 319 , 163 brocksopp , c. , starling , r. l. c. , schady , p. , mason , k. o. , romero - colmenero , e. , & puchnarewicz , e. m. 2006 , , 366 , 953 bromley , j. m. , somerville , r. s. , & fabian , a. c. 2004 , , 350 , 456 brunner , h. , mueller , c. , friedrich , p. , doerrer , t. , staubert , r. , & riffert , h. 1997 , , 326 , 885 budavri , t. , et al .
2005 , , 619 , l31 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , , 345 , 245 collin , s. , kawaguchi , t. , peterson , b. m. , & vestergaard , m. 2006 , , 456 , 75 congdon , p. 2006
, bayesian statistical modelling ( 2nd . ed . ; west sussex : john wiley & sons ltd . ) constantin , a. , shields , j. c. , hamann , f. , foltz , c. b. , & chaffee , f. h. 2002 , , 565 , 50 czerny , b. , nikoajuk , m. , raska , a. , dumont , a .- m . , loska , z. , & zycki , p. t. 2003 , , 412 , 317 czerny , b. , witt , h. j. , & zycki , p. 1997
, the transparent universe , 382 , 397 dai , x. , chartas , g. , eracleous , m. , & garmire , g. p. 2004
, , 605 , 45 davis , s. w. , woo , j .- h . , & blaes , o. m. 2007 , , 668 , 682 dempster , a. , laird , n. , & rubin , d. 1977 , j. r. statist .
b. , 39 , 1 de villiers , j .-
p . , hawley , j. f. , & krolik , j. h. 2003 , , 599 , 1238 di matteo , t. , springel , v. , & hernquist , l. 2005 , , 433 , 604 elvis , m. , risaliti , g. , & zamorani , g. 2002 , , 565 , l75 elvis , m. , et al .
1994 , , 95 , 1 esin , a. a. , mcclintock , j. e. , & narayan , r. 1997 , , 489 , 865 fabian , a. c. 1999 , , 308 , l39 fabian , a. c. , et al .
2002 , , 335 , l1 fox , j. 1997 , applied regression analysis , linear models , and related methods ( thousand oaks : sage publications , inc . )
fuller , w. a. 1987 , measurement error models ( new york : john wiley & sons ) galeev , a. a. , rosner , r. , & vaiana , g. s. 1979 , , 229 , 318 gallagher , s. c. , brandt , w. n. , chartas , g. , & garmire , g. p. 2002
, , 567 , 37 gallagher , s. c. , richards , g. t. , hall , p. b. , brandt , w. n. , schneider , d. p. , & vanden berk , d. e. 2005 , , 129 , 567 gallagher , s. c. , brandt , w. n. , chartas , g. , priddey , r. , garmire , g. p. , & sambruna , r. m. 2006 , , 644 , 709 gebhardt , k. , et al .
2000 , , 539 , l13 gierliski , m. , & done , c. 2004 , , 349 , l7 gilli , r. , comastri , a. , & hasinger , g. 2007 , , 463 , 79 green , p. j. , aldcroft , t. l. , mathur , s. , wilkes , b. j. , & elvis , m. 2001 , , 558 , 109 grupe , d. , mathur , s. , wilkes , b. , & osmer , p. 2006
, , 131 , 55 grupe , d. , thomas , h .- c . , & beuermann , k. 2001 , , 367 , 470 haardt , f. , & maraschi , l. 1991 , , 380 , l51 haardt , f. , & maraschi , l. 1993 , , 413 , 507 hastie , t. , tibshirani , r. , & friedman , j. 2001 , the elements of statistical learning ( new york : springer - verlag ) hatziminaoglou , e. , mathez , g. , solanes , j .- m . , manrique , a. , & salvador - sol , e. 2003 , , 343 , 692 hopkins , p. f. , hernquist , l. , cox , t. j. , di matteo , t. , martini , p. , robertson , b. , & springel , v. 2005 , , 630 , 705 hopkins , p. f. , hernquist , l. , cox , t. j. , di matteo , t. , robertson , b. , & springel , v. 2006 , , 163 , 1 hopkins , p. f. , richards , g. t. , & hernquist , l. 2007 , , 654 , 731 janiuk , a. , & czerny , b. 2000 , new astronomy , 5 , 7 janiuk , a. , czerny , b. , & madejski , g. m. 2001 , , 557 , 408 just , d. w. , brandt , w. n. , shemmer , o. , steffen , a. t. , schneider , d. p. , chartas , g. , & garmire , g. p. 2007
, , 665 , 1004 kaspi , s. , maoz , d. , netzer , h. , peterson , b. m. , vestergaard , m. , & jannuzi , b. t. 2005 , , 629 , 61 kauffmann , g. , & haehnelt , m. 2000 , , 311 , 576 kelly , b. c. 2007 , , 665 , 1489 kelly , b. c. , & bechtold , j. 2007 , , 168 , 1 kelly , b. c. , bechtold , j. , siemiginowska , a. , aldcroft , t. , & sobolewska , m. 2007 , , 657 , 116 ( k07 ) kollmeier , j. a. , et al .
2006 , , 648 , 128 krolik , j. h. 2001 , , 551 , 72 krolik , j. h. , hawley , j. f. , & hirose , s. 2005 , , 622 , 1008 laor , a. , fiore , f. , elvis , m. , wilkes , b. j. , & mcdowell , j. c. 1997 , , 477 , 93 lee , j. c. , iwasawa , k. , houck , j. c. , fabian , a. c. , marshall , h. l. , & canizares , c. r. 2002 , , 570 , l47 liang , e. p. t. , & price , r. h. 1977 , , 218 , 247 liu , b. f. , mineshige , s. , & ohsuga , k. 2003 , , 587 , 571 lu , y. , & yu , q. 1999 , , 526 , l5 mainieri , v. 2007 , in press at apjs ( astro - ph/0612361 ) malkan , m. 1991 , iau colloq . 129 : the 6th institute dastrophysique de paris ( iap ) meeting : structure and emission properties of accretion disks , 165 markowitz , a. , edelson , r. , & vaughan , s. 2003 , , 598 , 935 mathur , s. 2000 , , 314 , l17 mcintosh , d. h. , rieke , m. j. , rix , h .- w . , foltz , c. b. , & weymann , r. j. 1999 , , 514 , 40 mclure , r. j. , & dunlop , j. s. 2004 , , 352 , 1390 mclure , r. j. , & jarvis , m. j. 2002 , , 337 , 109 merloni , a. , & fabian , a. c. 2002 , , 332 , 165 merritt , d. , & ferrarese , l. 2001 , , 547 , 140 mineo , t. , et al . 2000 , , 359 , 471 murray , n. , chiang , j. , grossman , s. a. , & voit , g. m. 1995 , , 451 , 498 nandra , k. , george , i. m. , mushotzky , r. f. , turner , t. j. , & yaqoob , t. 1997 , , 477 , 602 nayakshin , s. 2000 , , 534 , 718 nishihara , e. , yamashita , t. , yoshida , m. , watanabe , e. , okumura , s .-
i . , mori , a. , & iye , m. 1997 , , 488 , l27 nowak , m. a. , wilms , j. , & dove , j. b. 2002 , , 332 , 856 oneill , p. m. , nandra , k. , papadakis , i. e. , & turner , t. j. 2005 , , 358 , 1405 onken , c. a. , ferrarese , l. , merritt , d. , peterson , b. m. , pogge , r. w. , vestergaard , m. , & wandel , a. 2004 , , 615 , 645 page , k. l. , reeves , j. n. , obrien , p. t. , turner , m. j. l. , & worrall , d. m. 2004 , , 353 , 133 page , k. l. , turner , m. j. l. , reeves , j. n. , obrien , p. t. , & sembay , s. 2003 , , 338 , 1004 p ' eroux , c. , storrie - lombardi , l. j. , mcmahon , r. g. , irwin , m. , & hook , i. m. 2001 , , 121 , 1799 pessah , m. e. , chan , c .- k .
, & psaltis , d. 2006 , , 372 , 183 pessah , m. e. , chan , c .- k . , & psaltis , d. 2007 , in press at ( arxiv:0705.0352 ) peterson , b. m. , et al . 2004 , , 613 , 682 piconcelli , e. , jimenez - bailn , e. , guainazzi , m. , schartel , n. , rodrguez - pascual , p. m. , & santos - lle , m. 2005 , , 432 , 15 porquet , d. , reeves , j. n. , obrien , p. , & brinkmann , w. 2004 , , 422 , 85 poutanen , j. , krolik , j. h. , & ryde , f. 1997 , , 292 , l21 proga , d. 2005 , , 630 , l9 proga , d. 2007 , , 661 , 693 proga , d. , & kallman , t. r. 2004 , , 616 , 688 reeves , j. n. , & turner , m. j. l. 2000 , , 316 , 234 reeves , j. n. , turner , m. j. l. , ohashi , t. , & kii , t. 1997 , , 292 , 468 reichard , t. a. , et al .
2003 , , 126 , 2594 richards , g. t. , et al .
2006 , , 131 , 2766 risaliti , g. , & elvis , m. 2005 , , 629 , l17 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 schneider , d. p. , et al .
2005 , , 130 , 376 schwartz , g. 1979 , ann .
, 6 , 461 scoville , n. z. , et al .
2007 , in press at ( astro - ph/0612305 ) shakura , n. i. , & syunyaev , r. a. 1973 , , 24 , 337 shapiro , s. l. , lightman , a. p. , & eardley , d. m. 1976 , , 204 , 187 shemmer , o. , brandt , w. n. , netzer , h. , maiolino , r. , & kaspi , s. 2006 , , 646 , l29 silk , j. , & rees , m. j. 1998 , , 331 , l1 sobolewska , m. a. , siemiginowska , a. , & ycki , p. t. 2004a , , 608 , 80 sobolewska , m. a. , siemiginowska , a. , & ycki , p. t. 2004b , , 617 , 102 spergel , d. n. , et al .
2003 , , 148 , 175 springel , v. , di matteo , t. , & hernquist , l. 2005 , , 620 , l79 stapleton , d. c. & young , d. j. 1984 , econometrica , 52 , 737 steffen , a. t. , strateva , i. , brandt , w. n. , alexander , d. m. , koekemoer , a. m. , lehmer , b. d. , schneider , d. p. , & vignali , c. 2006 , , 131 , 2826 storrie - lombardi , l. j. , mcmahon , r. g. , irwin , m. j. , & hazard , c. 1996 , , 468 , 121 strateva , i. v. , brandt , w. n. , schneider , d. p. , vanden berk , d. g. , & vignali , c. 2005 , , 130 , 387 ( s05 ) tananbaum , h. , et al . 1979 , , 234 , l9 turner , n. j. 2004 , , 605 , l45 tremaine , s. , et al .
2002 , , 574 , 740 trump , j. r. , et al .
2006 , , 165 , 1 trump , j. , impey , c. , et al .
2007 , in press at ( astro - ph/0606016 ) trump , j. , impey , c. , et al .
2008 , in preparation uttley , p. , mchardy , i. m. , & papadakis , i. e. 2002 , , 332 , 231 vasudevan , r. v. , & fabian , a. c. 2007 , in press at , ( arxiv:0708.4308 ) v ' eron - cetty , m .-
p . , joly , m. , & v ' eron , p. 2004
, , 417 , 515 vestergaard , m. 2002 , , 571 , 733 vestergaard , m. 2004 , , 601 , 676 vestergaard , m. , et al .
2007 , in preparation vestergaard , m. , & peterson , b. m. 2006 , , 641 , 689 vestergaard , m. , & wilkes , b. j. 2001 , , 134 , 1 vignali , c. , brandt , w. n. , fan , x. , gunn , j. e. , kaspi , s. , schneider , d. p. , & strauss , m. a. 2001 , , 122 , 2143 vignali , c. , brandt , w. n. , schneider , d. p. , garmire , g. p. , & kaspi , s. 2003 , , 125 , 418 vignali , c. , brandt , w. n. , & schneider , d. p. 2003
, , 125 , 433 vignali , c. , brandt , w. n. , schneider , d. p. , & kaspi , s. 2005 , , 129 , 2519 vignali , c. , comastri , a. , cappi , m. , palumbo , g. g. c. , matsuoka , m. , & kubo , h. 1999 , , 516 , 582 wandel , a. 2000 , new astronomy review , 44 , 427 wandel , a. , peterson , b. m. , & malkan , m. a. 1999 , , 526 , 579 wang , j .- m . , watarai , k .- y . , & mineshige , s. 2004 , , 607 , l107 wilkes , b. j. , & elvis , m. 1987 , , 323 , 243 wilkes , b. j. , tananbaum , h. , worrall , d. m. , avni , y. , oey , m. s. , & flanagan , j. 1994 , , 92 , 53 woo , j. , & urry , c. m. 2002 , , 579 , 530 worrall , d. m. , tananbaum , h. , giommi , p. , & zamorani , g. 1987 , , 313 , 596 wyithe , j. s. b. , & loeb , a. 2003 , , 595 , 614 yuan , w. , siebert , j. , & brinkmann , w. 1998b , , 334 , 498 zamorani , g. , et al .
1981 , , 245 , 357 zdziarski , a. a. , lubinski , p. , & smith , d. a. 1999 , , 303 , l11 as a function of @xmath10 ( left ) and @xmath9 ( right ) .
the opposite dependence of @xmath0 on @xmath10 and @xmath9 suggests that at least one of these quantities is not proportional to @xmath19 . as such
, we do not employ bolometric corrections in this work , and instead compare directly with @xmath10 and @xmath9 .
[ f - alfox_redd],title="fig : " ] as a function of @xmath10 ( left ) and @xmath9 ( right ) .
the opposite dependence of @xmath0 on @xmath10 and @xmath9 suggests that at least one of these quantities is not proportional to @xmath19 . as such
, we do not employ bolometric corrections in this work , and instead compare directly with @xmath10 and @xmath9 .
[ f - alfox_redd],title="fig : " ] and @xmath245 and @xmath246 ( left ) and for the intrinsic dispersion in @xmath0 at a given @xmath139 or @xmath9 ( right ) .
the thick solid line denotes the probability distribution for the two quantities with respect to @xmath16 , the dashed line for the two quantities with respect to @xmath2 , the dashed - dotted line for the two quantites with respect to @xmath10 , and the thin solid line for the two quantities with respect to @xmath9 .
[ f - alfox_corr],title="fig : " ] and @xmath245 and @xmath246 ( left ) and for the intrinsic dispersion in @xmath0 at a given @xmath139 or @xmath9 ( right ) .
the thick solid line denotes the probability distribution for the two quantities with respect to @xmath16 , the dashed line for the two quantities with respect to @xmath2 , the dashed - dotted line for the two quantites with respect to @xmath10 , and the thin solid line for the two quantities with respect to @xmath9 .
[ f - alfox_corr],title="fig : " ] regressions of @xmath1 on @xmath168 .
while there is no significant evidence that either the h@xmath3 or c iv regression slope is different from zero , there _ is _ significant evidence that they are not the same , implying a nonmonotonic trend between @xmath1 and @xmath2 .
[ f - regdiff ] ] ) ([eq - modelspec ] ) , assuming @xmath208 ( left ) and a varying @xmath1 with @xmath2 ( right ) .
the spectra are computed for a rqq with @xmath247 ( thick solid line ) , @xmath248 ( dashed line ) , @xmath249 ( dashed - dotted line ) , and @xmath212 ( thin solid line ) . in all cases
we assume @xmath250 and that @xmath251 of the bolometric luminosity is emitted by the disk .
the vertical lines mark the locations of 2500 and 2 kev .
the dependence of the location of the peak in the disk emission on @xmath2 is apparent , producing a correlation between @xmath0 and @xmath2 even if the fraction of bolometric luminosity emitted by the disk is independent of @xmath2 .
[ f - modelspec],title="fig : " ] ) ([eq - modelspec ] ) , assuming @xmath208 ( left ) and a varying @xmath1 with @xmath2 ( right ) .
the spectra are computed for a rqq with @xmath247 ( thick solid line ) , @xmath248 ( dashed line ) , @xmath249 ( dashed - dotted line ) , and @xmath212 ( thin solid line ) . in all cases
we assume @xmath250 and that @xmath251 of the bolometric luminosity is emitted by the disk .
the vertical lines mark the locations of 2500 and 2 kev .
the dependence of the location of the peak in the disk emission on @xmath2 is apparent , producing a correlation between @xmath0 and @xmath2 even if the fraction of bolometric luminosity emitted by the disk is independent of @xmath2 .
[ f - modelspec],title="fig : " ] cccccc 00 02 30.7 & + 00 49 59.0 & 1.352 & @xmath252 & @xmath253 & @xmath254 + 00 06 54.1 & -00 15 33.4 & 1.725 & @xmath255 & @xmath256 & @xmath257 + 00 22 10.0 & + 00 16 29.3 & 0.574 & @xmath258 & @xmath259 & @xmath260 + 00 27 52.4 & + 00 26 15.7 & 0.205 & @xmath261 & @xmath262 & @xmath263 + 00 31 31.4 & + 00 34 20.2 & 1.735 & @xmath264 & @xmath265 & @xmath266 + 00 50 06.3 & -00 53 19.0 & 4.331 & @xmath267 & @xmath268 & @xmath269 + 00 57 01.1 & + 14 50 03.0 & 0.623 & @xmath270 & @xmath271 & @xmath272 + 00 59 22.8 & + 00 03 01.0 & 4.178 & @xmath273 & @xmath274 & @xmath275 + 01 06 19.2 & + 00 48 22.0 & 4.437 & @xmath276 & @xmath277 & @xmath278 + 01 13 05.7 & + 15 31 46.5 & 0.576 & @xmath279 & @xmath280 & @xmath281 + 01 13 09.1 & + 15 35 53.6 & 1.806 & @xmath282 & @xmath283 & @xmath284 + 01 15 37.7 & + 00 20 28.7 & 1.275 & @xmath285 & @xmath286 & @xmath287 + 01 26 02.2 & -00 19 24.1 & 1.765 & @xmath288 & @xmath289 & @xmath290 + lccccccccc pg 0026 + 129 & 00 29 13.7 & + 13 16 03.8 & 0.142 & c iv & 7.789 & 1.768 & 1 & 1.96 & 5 + fairall 9 & 01 23 45.7 & -58 48 21.8 & 0.046 & c iv & 7.760 & 0.265 & 1 & @xmath291 & 6 + pg 1202 + 281 & 12 04 42.2 & + 27 54 12.0 & 0.165 & c iv & 7.855 & 0.359 & 1 & @xmath292 & 7
+ pg 1211 + 143 & 12 14 17.7 & + 14 03 12.3 & 0.080 & c iv & 7.559 & 1.380 & 1 & @xmath293 & 8 + pg 1247 + 267 & 12 50 05.7 & + 26 31 07.7 & 2.038 & h@xmath3 & 10.41 & 0.379 & 2 & @xmath294 & 9 + q1346 - 036 & 13 48 44.1 & -03 53 25.0 & 2.370 & h@xmath3 & 9.946 & 0.609 & 3 & @xmath295 & 3 + mrk 478 & 14 42 07.5 & + 35 26 22.9 & 0.077 & c iv & 7.890 & 0.498 & 1 &
@xmath296 & 7 + pg 1630 + 377 & 16 32 01.1 & + 37 37 50.0 & 1.476 & h@xmath3 & 9.762 & 0.569 & 4 & @xmath297 & 10 + pg 1634 + 706 & 16 34 28.9 & + 70 31 33.0 & 1.334 & h@xmath3 & 10.27 & 0.734 & 4 & @xmath298 & 10 + he 2217 - 2818 & 22 20 06.8 & -28 03 23.9 & 2.414 & h@xmath3 & 10.12 & 0.807 & 3 & @xmath299 & 3 ccccccc 09 58 48.8 & + 02 34 42.3 & 1.551 & c iv & 8.276 & 0.131 & @xmath300 + 09 59 02.6 & + 02 25 11.8 & 1.105 & mg ii & 7.612 & 0.025 & @xmath301 + 09 59 49.4 & + 02 01 41.1 & 1.758 & c iv & 8.108 & 0.719 & @xmath302 + 10 00 50.0 & + 02 05 00.0 & 1.235 & mg ii & 7.692 & 0.501 & @xmath303 + 10 00 51.6 & + 02 12 15.8 & 1.829 & mg ii & 7.807 & 0.131 & @xmath304 + 10 00 58.9 & + 01 53 59.5 & 1.559 & c iv & 8.346 & 0.172 & @xmath305 + 10 02 19.6 & + 01 55 36.9 & 1.509 & c iv & 8.333 & 0.177 & @xmath306 + 10 02 34.4 & + 01 50 11.5 & 1.506 & c iv & 7.991 & 0.941 & @xmath307 + 10 02 43.9 & + 02 05 02.0 & 1.234 & mg ii & 7.817 & 0.303 & @xmath308 + | in this work we use a sample of 318 radio - quiet quasars ( rqq ) to investigate the dependence of the ratio of optical / uv flux to x - ray flux , @xmath0 , and the x - ray photon index , @xmath1 , on black hole mass , uv luminosity relative to eddington , and x - ray luminosity relative to eddington .
our sample is drawn from the literature , with x - ray data from _ rosat _ and _ chandra _ , and optical data mostly from the sdss ; 153 of these sources have estimates of @xmath1 from _
chandra_. we estimate @xmath2 using standard estimates derived from the h@xmath3 , mg ii , and c iv broad emission lines . our sample spans a broad range in black hole mass ( @xmath4 ) , redshift ( @xmath5 ) , and luminosity ( @xmath6@xmath7 \lesssim 10^{48}$ ] ) .
we find that @xmath0 increases with increasing @xmath2 and @xmath8 , and decreases with increasing @xmath9 .
in addition , we confirm the correlation seen in previous studies between @xmath1 and @xmath2 and both @xmath10 and @xmath9 ; however , we also find evidence that the dependence of @xmath1 of these quantities is not monotonic , changing sign at @xmath11 .
we argue that the @xmath0 correlations imply that the fraction of bolometric luminosity emitted by the accretion disk , as compared to the corona , increases with increasing accretion rate relative to eddington , @xmath12 .
in addition , we argue that the @xmath1 trends are caused by a dependence of x - ray spectral index on @xmath12 .
we discuss our results within the context of accretion models with comptonizing corona , and discuss the implications of the @xmath0 correlations for quasar feedback .
to date , this is the largest study of the dependence of rqq x - ray parameters on black hole mass and related quantities , and the first to attempt to correct for the large statistical uncertainty in the broad line mass estimates . |
the primary aim of ultra - relativistic heavy - ion collisions is to produce and study a state of high - density nuclear matter called the quark - gluon plasma ( qgp ) . in the search of this new form of matter , penetrating probes are essential in order to gain information from the early stage of the collisions .
the lifetime of the @xmath0 meson is about 1.3 fm / c , which is smaller than the life time of the system formed in @xmath6 collisions at such energy .
the @xmath0 measured via its hadronic decay channel ( branching ratio @xmath9 100@xmath5 ) can be used as a sensitive tool to examine the collision dynamics in the hadronic medium through its decay and regeneration . + elliptic flow , @xmath1 , is an observable which is thought to reflect conditions from early stage of the collisions [ 1 , 2 ] . in non - central heavy - ion collisions , the initial spatial anisotropy of the overlap region of the colliding nuclei is transformed into an anisotropy in momentum space through interactions among the produced particles .
systematic measurements of the @xmath1 of hadrons show that the @xmath1 scales with the number of constituent quarks in the intermediate @xmath10 region ( @xmath12 gev / c ) .
it has been proposed that the measurement of the @xmath1 of resonances can distinguish whether the resonances were produced from a hadronizing quark gluon plasma ( qgp - mechanism ) or in the hadronic final state via hadron - hadron rescattering ( hg mechanism ) [ 3 ] .
the results presented in this paper were obtained with the star detector [ 4 ] at the relativistic heavy ion collider ( rhic ) at brookhaven national laboratory ( bnl ) , usa .
the sub - detectors used in this analysis were the time projection chamber ( tpc ) , and the trigger detectors , namely zero degree calorimeter ( zdc ) .
the collision centrality was determined by charged hadron multiplicity measured in tpc within the pseudo - rapidity @xmath13 @xmath3 0.5 .
+ @xmath14 invariant mass distribution after background subtraction in @xmath6 collisions . ]
the @xmath0 yield in each @xmath10 bin was extracted from the invariant mass ( @xmath15 ) distribution of @xmath16 and @xmath17 candidates after subtraction of like - sign combinatorial background obtained from the geometric mean of the @xmath18 and @xmath19 invariant mass distributions in the same event .
the @xmath20 invariant mass distribution and the combinatorial background are normalized in the invariant mass range from 1.5 gev/@xmath21 to 2.5 gev/@xmath21 before subtraction .
the pions were identified through their @xmath22 energy loss in the star tpc [ 4 ] .
a typical @xmath8 invariant mass distributions after background subtraction for @xmath4@xmath5 centrality and @xmath23 gev / c in @xmath6 collisions at @xmath7 gev is shown in fig .
1 . the solid black line in fig . 1 is the sum of all the contributions in hadronic cocktail .
the @xmath24 was fit to a gaussian .
the @xmath25 shape was obtained from the hijing event generator [ 5 ] .
the @xmath26 , the @xmath27 , the @xmath28 ( 1270 ) and the @xmath29 were fit by relativistic breit - wigner functions [ 6 ] times the boltzmann factor which accounts for the phase space [ 7,8 ] in the hadronic cocktail . in the cocktail fit ,
the @xmath0 width was fixed at 160 mev/@xmath30 .
the @xmath29 mass and width were fixed at 630 mev/@xmath30 and 160 mev/@xmath30 , respectively .
the temperature in the phase space was taken to be 120 mev [ 8 ] . + two different techniques are being used to find out the @xmath1 of the @xmath0-meson .
one is @xmath1 vs. invariant mass method [ 9 ] and the other one is ( @xmath31 ) method [ 10 ] .
the invariant mass method involves calculating the @xmath1 of the same - event distribution as a function of @xmath15 and then fitting the resulting @xmath32 distribution using a multi parameters function : + @xmath33 where @xmath34 is the signal @xmath1 and @xmath35 is the background @xmath1 .
the signal @xmath34 contribution is coming from @xmath0 , @xmath29 , @xmath36 , @xmath37 , @xmath38 and @xmath28 , i.e. @xmath39 . the background @xmath35 is calculated from the @xmath40 and @xmath41 pairs @xmath1 .
the parameters @xmath42 and @xmath43 where @xmath44 is the sum of all the individual particles signal in the cocktail and @xmath45 is the background contribution from the @xmath18 and @xmath46 invariant mass distributions .
2 represents the @xmath1(total ) as a function of invariant mass for a particular @xmath10 bin ( @xmath47 gev / c ) .
the @xmath1(total ) is the total @xmath1 of all the @xmath8 combinations in the same event .
the signal @xmath1 is extracted by doing a fit using eq.(1 ) to the @xmath1(total ) as shown in fig .
counts as a function of @xmath31 .
the solid line is the result of the fitting function . ]
counts as a function of @xmath31 .
the solid line is the result of the fitting function . ] in order to compare the results obtained in the above method , we used another method to calculate the @xmath1 of @xmath0 which is called the standard ( @xmath31 ) method and described in the reference [ 10 ] .
3 shows the @xmath0-meson yield after background subtraction as a function of ( @xmath31 ) .
the observed @xmath1 is obtained from the distribution by fitting a function of the form @xmath48 $ ] .
the observed @xmath1 parameters were corrected for the event plane resolution to get the final @xmath1 values .
4 shows the corrected @xmath1 as a function of @xmath10 .
solid closed circles are the data points obtained from the invariant mass technique whereas the open closed circles are for the standard ( @xmath49 bin ) method .
it is clear from the figure that the results obtained in both the techniques are consistent within the statistical error .
5 represents the comparison of @xmath0 @xmath1 with @xmath24 and @xmath50 @xmath1 for the same centrality class , i.e. @xmath4@xmath5 , in @xmath6 collisions .
the solid circles are the data points for @xmath0 @xmath1 obtained from invariant mass method .
the open circles are the data points for @xmath24 and open squares are for @xmath50 . the @xmath24 and @xmath50 data points are taken from [ 11 ] .
it is clear from fig .
5 that the @xmath1 of @xmath0 is more close to the @xmath1 of @xmath24 than @xmath1 of @xmath50 in the region @xmath10 @xmath51 1.5 gev / c . in the low @xmath10 region
( @xmath52 gev / c ) @xmath0-meson seems to deviate from the usual mass ordering . + comparison of @xmath0 @xmath1 obtained in @xmath1 vs. invariant mass method and the standard event plane method .
only the statistical error bars were shown in the plot . ]
the number of constituent quark ( ncq ) scaling of @xmath1 of @xmath0-vector meson in the intermidiate @xmath10 is shown in fig .
the @xmath37 and @xmath53 @xmath1 are plotted in the same figure for comparison after scaling with @xmath54 and @xmath55 quarks , respectively .
this measurement shows that the @xmath0 @xmath1 scales with @xmath54 quarks in the intermediate @xmath10 range .
constituent quark number scaling of elliptic flow ( @xmath1 ) for the @xmath0-meson , the @xmath24 meson and the @xmath50 baryon in @xmath6 @xmath56 gev for @xmath4@xmath5 centrality .
the vertical error bars represent the statistical errors . ]
constituent quark number scaling of elliptic flow ( @xmath1 ) for the @xmath0-meson , the @xmath24 meson and the @xmath50 baryon in @xmath6 @xmath56 gev for @xmath4@xmath5 centrality .
the vertical error bars represent the statistical errors . ]
the @xmath0 @xmath1 is measured in @xmath6 collisions at 200 gev in two different methods and the results obtained are consistent within the statistical errors . from the number of constituent quark scaling of @xmath1 , it is clear that @xmath0-vector meson follows @xmath54 quarks in the intermediate @xmath10 range which implies most of the @xmath0s are formed from the quarks coalescence .
ollitrault , phys .
d * 46 * ( 1992 ) 229 .
h. sorge , phys .
* 82 * ( 1999 ) 2048 . c. nonaka et al .
c * 69 * ( 2004 ) 031902 .
ackermann et al . , nucl .
intr . meth . in phys .
res . a * 499 * ( 2003 ) 624 .
wang and m. gyulassy , phys .
d * 44 * ( 1991 ) 3501 ; compt .
phys . commun . 83 ( 1994 )
307 . c. adler et al .
* 89 * ( 2002 ) 272302 .
j. adams et al .
* 92 * ( 2004 ) 92301 .
shurayk and g.e .
brown , nucl .
phys . a * 717 * ( 2003 ) 322 .
n. borghini and j .- y .
ollitrault , phys .
c * 70 * ( 2004 ) 064905 .
a.m. poskanzer and s.a .
voloshin , phys .
c * 58 * ( 1998 ) 1671 .
b. i. abelev et al .
c * 77 * ( 2008 ) 54901 . | the first measurement of the @xmath0 - vector meson elliptic flow @xmath1 at mid - rapidity ( @xmath2 @xmath3 0.5 ) in @xmath4 @xmath5 centrality in @xmath6 collisions at @xmath7 gev from the star experiment at rhic is presented .
the study is through the @xmath8 hadronic decay channel of @xmath0 which has a branching ratio of @xmath9 100 % .
the analysis is being carried out in two different methods .
the @xmath1 results obtained in these methods are consistent .
number of constituent quark ( ncq ) scaling of @xmath1 of @xmath0 meson with respect to other hadrons at intermediate @xmath10 is observed .
the @xmath0 @xmath1 favors @xmath11 scaling , supporting the coalescence being the dominant mechanism of hadronization in the intermediate @xmath10 region at rhic . |
here we present an algorithm to compute projections of channels onto exponential families of fixed interactions .
the decomposition is geometrical , and it is based on the idea that , rather than joint distributions , the quantities we work with are channels , or conditionals ( or markov kernels , stochastic kernels , transition kernels , stochastic maps ) .
our algorithm can be considered a channel version of the iterative scaling of ( joint ) probability distributions , presented in @xcite .
exponential and mixture families ( of joints and of channels ) have a duality property , shown in section [ fam ] . by fixing some marginals
, one determines a mixture family . by fixing ( boltzmann - type ) interactions
, one determines an exponential family .
these two families intersect in a single point , which means that ( theorem [ dualmk ] ) _ there exists a unique element which has the desired marginals and the desired interactions_. as a consequence , theorem [ dualmk ] translates projections onto exponential families ( which are generally hard to compute ) to projections onto fixed - marginals mixture families ( which can be approximated by an iterative procedure ) .
section [ algo ] explains how this is done .
projections onto exponential families are becoming more and more important in the definition of measures of statistical interaction , complexity , synergy , and related quantities .
in particular , the algorithm can be used to compute decompositions of mutual information , as for example the ones defined in @xcite and @xcite , and it was indeed used to compute all the numerical examples in @xcite .
another application of the algorithm is explicit computations of complexity measure as treated in @xcite , @xcite , and @xcite .
examples of both applications can be found in section [ applic ] .
for all the technical details about the iterative scaling algorithm in its traditional version , we refer the interested reader to chapters 3 and 5 of @xcite .
all proofs can be found in the appendix .
we take the same definitions and notations as in @xcite , except that we let the output be multiple .
more precisely , we consider a set of @xmath0 input nodes @xmath1 , taking values in the sets @xmath2 , and a set of @xmath3 output nodes @xmath4 , taking values in the sets @xmath5 .
we write the input globally as @xmath6 , and the output globally as @xmath7 .
we denote by @xmath8 the set of real functions on @xmath9 , and by @xmath10 the set of probability measures on @xmath11 .
let @xmath12 and @xmath13 .
we call @xmath14 the space of functions which only depend on @xmath15 and @xmath16 : @xmath17 we can model the channel from @xmath11 to @xmath9 as a markov kernel @xmath18 , that assigns to each @xmath19 a probability measure on @xmath9 ( for a detailed treatment , see @xcite ) . here
we will consider only finite systems , so we can think of a channel simply as a transition matrix ( or stochastic matrix ) , whose rows sum to one .
@xmath20 the space of channels from @xmath11 to @xmath9 will be denoted by @xmath21 .
we will denote by @xmath11 and @xmath9 also the corresponding random variables , whenever this does not lead to confusion .
conditional probabilities define channels : if @xmath22 and the marginal @xmath23 is strictly positive , then @xmath24 is a well - defined channel .
viceversa , if @xmath25 , given @xmath26 we can form a well - defined joint probability : @xmath27 to extend the notion of divergence from probability distributions to channels , we need an `` input distribution '' : let @xmath26 , let @xmath28 . then : @xmath29 let @xmath30 be joint probability distributions on @xmath31 , and let @xmath32 be the kl - divergence . then the following `` chain rule '' holds : @xmath33
suppose we have a family @xmath34 of channels , and a channel @xmath18 that may not be in @xmath34 .
then we can define the `` distance '' between @xmath18 and @xmath34 in terms of @xmath35 .
let @xmath36 be an input distribution .
the divergence between a channel @xmath18 and a family of channels @xmath34 is given by : @xmath37 if the minimum is uniquely realized , we call the channel @xmath38 the _ ri - projection _ of @xmath18 on @xmath34 ( and simply `` an '' ri - projection if it is not unique )
. the families considered here are of two types , dual to each other : linear and exponential .
for both cases , we take the closures , so that the minima defined above always exist .
a _ mixture family _ of @xmath21 is a subset of @xmath21 defined by one or several affine equations , i.e. , the locus of the @xmath18 which satisfy a ( finite ) system of equations in the form : @xmath39 for some functions @xmath40 , and some constants @xmath41 .
[ [ example . ] ] example .
+ + + + + + + + consider a channel @xmath42 .
we can form the marginal : @xmath43 the channels @xmath44 such that @xmath45 form a mixture family , defined by the system of equations ( for all @xmath46 , @xmath47 ) : @xmath48 where the function @xmath49 is equal to 1 for @xmath50 , and zero for any other case .
more in general , let @xmath51 be a ( finite - dimensional ) linear subspace of @xmath52 , and let @xmath25 . then : @xmath53 is a mixture family , which we call _ generated by @xmath18 and @xmath51_. a ( closed ) _ exponential family _ of @xmath21 is ( the closure of ) a subset of @xmath21 of channels in the form : @xmath54 where @xmath55 satisfies affine constraints , @xmath18 is fixed , and : @xmath56 so that the channel is correctly normalized .
this is a sort of multiplicative equivalent of mixture families , as the exponent satisfies constraints similar to .
[ [ example.-1 ] ] example .
+ + + + + + + + let @xmath51 be a ( finite - dimensional ) linear subspace of @xmath52 , and let @xmath25 .
then the closure : @xmath57 is an exponential family , which again we call _ generated by @xmath18 and @xmath51_. this family is in some sense `` dual '' to the family in . the duality is expressed more precisely by the following result . [ dualmk ] let @xmath51 be a subspace of @xmath52 .
let @xmath26 be strictly positive .
let @xmath58 be a strictly positive `` reference '' channel .
let @xmath59 and @xmath60 . for @xmath61 ,
the following conditions are equivalent : 1 .
2 . @xmath63 , and @xmath64 .
3 . @xmath65 , and @xmath66 . in particular , @xmath67 is unique , and it is exactly @xmath68 .
geometrically , we are saying that @xmath69 , the ri - projection of @xmath18 on @xmath34 .
we call the mapping @xmath70 the _ ri - projection operator _ , and the mapping @xmath71 the _ i - projection operator _ these are the channel equivalent of the i - projections introduced in @xcite and generalized in @xcite .
the result is illustrated in figure [ fig : dualmk ] . .
the point @xmath67 at the intersection minimizes on @xmath34 the distance from @xmath18 , and minimizes on @xmath72 the distance from @xmath73 . ] as suggested by figure [ fig : dualmk ] , i- and ri - projections on exponential families satisfy a pythagoras - type equality .
for any @xmath74 , with @xmath34 exponential family : @xmath75 this statement follows directly from the analogous statement for probability distribution found in @xcite , after applying the chain rule .
the algorithm can be considered as a channel equivalent of the iterative scaling procedure for joint distributions , which can be found in chapter 5 of @xcite . translated into our language , that theorem says the following : [ jointit ] let @xmath76 be mixture families of joint distributions with nonempty intersection @xmath51 .
denote by @xmath77 the @xmath78-projection of a joint @xmath79 onto the family @xmath80 .
consider the sequence that starts at @xmath81 and is defined iteratively by : @xmath82 then @xmath83 converges , and the limit point is the @xmath78-projection of @xmath81 onto @xmath51 , i.e. if we call : @xmath84 then @xmath85 , and for any @xmath86 : @xmath87 our result depends on the theorem above , in the following way .
we define a marginal procedure for channels , which in general depends on the choice of an input distribution .
we define mixture families of channels with fixed marginals in a way compatible with the equivalent for joints .
we then define scalings of channels , and prove that they give the desired result at the joint level .
this makes it possible to translate the statement of theorem [ jointit ] to an analogous statement for channels , theorem [ convergence ] .
unless otherwise stated , all the input distributions here will be assumed strictly positive .
all our proofs can be found in the appendix .
consider an input distribution @xmath26 .
let @xmath88 ,
j\subseteq [ m ] , j\ne\emptyset$ ] .
we define the _ marginal operator _ for channels as : @xmath89 given the input @xmath36 .
[ jeq ] defined as above , @xmath90 is exactly the conditional probability for the marginal @xmath91 . in other words
, @xmath92 has marginal @xmath90 if and only if @xmath93 has marginal @xmath91 .
consider an input distribution @xmath26 .
let @xmath88 , j\subseteq [ m]$ ] , @xmath94 .
we define the mixture families @xmath95 as : @xmath96 where the @xmath97 are prescribed channel marginals .
analogously , let @xmath98 be a probability distribution in @xmath99 .
we define the mixture families : @xmath100 proposition [ jeq ] says that , for any @xmath25 , for any ( strictly positive ) @xmath26 , and for any @xmath88 , j\subseteq [ m]$ ] : @xmath101 [ presc ] @xmath95 is exactly the set @xmath102 of equation , where as @xmath51 we take the space @xmath14 of functions which only depend on the nodes in @xmath103 . just as in @xcite , the i - projections for single marginals can be obtained by scaling . for joint distributions
the scaling is done in this way : if @xmath104 is a `` prescribed '' marginal , then : @xmath105 will have the prescribed marginals , and even be the i - projection of @xmath36 on @xmath106 .
i.e. , @xmath107 , and : @xmath108 for the proof , see chapter 3 and section 5.1 of @xcite . in the case of channels ,
the scaling is instead done in two steps .
we define the ( unnormalized ) _ @xmath109-scaling _ as the operator @xmath110 , mapping @xmath18 to : @xmath111 we have that @xmath112 is _ not _ an element of @xmath113 , as in general it is not even in @xmath21 ( i.e. a correctly normalized channel ) .
however , at the joint level this corresponds exactly to the joint scaling : [ jlevel ] let @xmath26 , @xmath25 , and @xmath114 . then : @xmath115 this implies that @xmath116 is the @xmath78-projection of @xmath117 to the family @xmath118 .
we define the _ normalized @xmath109-scaling _ as the operator @xmath119 , mapping @xmath18 to : @xmath120 where : @xmath121 at the joint level , this corresponds to scaling of the input in the following way : [ inscale ] let @xmath26 , @xmath25 , and @xmath114 .
then : @xmath122}^{p } \operatorname{\sigma}_{ij}^{p \bar k } pk\;.\ ] ] this implies that @xmath123 is the @xmath78-projection of @xmath124 to the family with prescribed input @xmath23 . for brevity ,
let s call this family @xmath125}(p)$ ] .
now @xmath126 is an element of @xmath21 , but still _ not _ of @xmath113 .
however , if we iterate the operator @xmath127 , the resulting sequence will converge to the projector on @xmath113 .
more in general , the following result holds : [ convergence ] for @xmath128 , let @xmath129 $ ] be subsets of @xmath130 $ ] and @xmath131 $ ] be nonempty subsets of @xmath132 $ ] .
take an input distribution @xmath26 and a channel @xmath133 .
define the mixture families of prescribed marginals : @xmath134 and their intersection , which is also a mixture family ( nonempty , as it contains at least @xmath135 ) : @xmath136 choose a ( different ) channel @xmath137 and consider the sequence of normalized scalings starting at @xmath138 and defined iteratively by : @xmath139 then : * @xmath140 converges to a limit channel : @xmath141 * the limit @xmath142 is the @xmath78-projection of @xmath138 on @xmath72 , i.e. @xmath143 and : @xmath144 the proof can be found in the appendix . to apply the theorem [ convergence ] in our algorithm , we choose as initial channel @xmath138 exactly the reference channel @xmath73 of theorem [ dualmk ] , usually the uniform channel . as @xmath135 we take exactly the `` prescription channel '' @xmath18 of theorem [ dualmk ] , i.e. the channel which has the desired marginals .
the result of the iterative scaling will be the ri - projection of @xmath18 on the desired exponential family .
the algorithm presented here permits to compute the decompositions of mutual information between inputs and outputs in @xcite and @xcite .
we give here examples of computations of _ pairwise synergy _ as an ri - projection for channels , as described in @xcite .
it is not within the scope of this article to motivate this measure , we rather want to show how it can be computed . let @xmath18 be a channel from @xmath145 to @xmath9 .
let @xmath26 be a strictly positive input distribution .
we define in @xcite the synergy of @xmath18 as : @xmath146 where @xmath147 is the ( closure of the ) family of channels in the form : @xmath148 where : @xmath149 and : @xmath150 according to theorem [ dualmk ] , the ri - projection of @xmath18 on @xmath147 is the unique point @xmath67 of @xmath147 which has all the prescribed marginals : @xmath151 and can therefore be computed by iterative scaling , either of the joint distribution ( as it is traditionally done , see @xcite ) , or of the channels ( our algorithm ) . here
we present a comparison of the two algorithms , implemented similarly and in the same language ( mathematica ) .
the red dots represent our ( channel ) algorithm , and the blue dots represent the joint rescaling algorithm .
for the easiest channels ( see figure [ fig : xor ] ) , both algorithm converge instantly .
a more interesting example is a randomly generated channel ( figure [ fig : rand ] ) , in which both method need 5 - 10 iterations to get to the desired value .
however , the channel method is slightly faster .
the most interesting example is the synergy of the and gate , which should be zero according to the procedure @xcite . in that article
, we mistakenly wrote a different value , that here we would like to correct ( it is zero ) .
the convergence to zero is very slow , of the order of @xmath152 ( figure [ fig : and ] ) .
it is clearly again slightly faster for the channel method in terms of iterations . , the joint method ( blue ) proportionally to @xmath153 . ]
it has to be noted , however , that rescaling a channel requires more elementary operations than rescaling a joint distribution .
because of this , one single iteration with our method takes longer than with the joint method .
( as explained in section [ algo ] , a scaling for the channel corresponds to two scalings for the joint . ) in the end , despite the need of fewer iterations , the total computation time of a projection with our algorithm can be longer ( depending on the particular problem ) .
for example , again for the synergy of the and gate , we can plot the computation time as a function of the accuracy ( distance to actual value ) , down to @xmath154 .
the results are shown in figure [ fig : comp ] . to get to the same accuracy , though
, the channel approach used less iterations . in summary ,
our algorithm is better in terms of iteration complexity , but generally worse in terms of computing time .
iterative scaling can also be used to compute measures of complexity , as defined in @xcite , @xcite , and in section 6.9 of @xcite . for simplicity ,
consider two inputs @xmath155 , two outputs @xmath156 and a generic channel between them .
in general , any sort of interaction is possible , which in terms of graphical models ( see @xcite ) can be represented by diagrams such as those in figure [ fig : graphs1 ] .
.5 and @xmath157 are indeed correlated , but only indirectly , via the inputs .
b ) the graphical model corresponding to a non - complex system .
, title="fig : " ] [ fig : graph2 ] .45 and @xmath157 are indeed correlated , but only indirectly , via the inputs .
b ) the graphical model corresponding to a non - complex system .
, title="fig : " ] [ fig : gp1 ] any line in the graph indicates an interaction between the nodes .
in @xcite the outputs are assumed to be conditionally independent , i.e. they do not directly interact ( or , their interaction can be _ explained away _ by conditioning on the inputs ) .
in this case the graph looks like figure [ fig : graphs1]a , and the maginals to preserve are those of the family of pairs @xmath158 , @xmath159 with : @xmath160 , @xmath161 , @xmath162 , @xmath163 .
suppose now that @xmath156 correspond to @xmath155 at a later time .
in this case it is natural to assume that the system is not complex if @xmath164 does not depend ( directly ) on @xmath165 , and @xmath157 does not depend ( directly ) on @xmath166 .
intuitively , in this case `` the whole is exactly the sum of its parts '' . in terms of graphical models
, this means that our system is represented by figure [ fig : graphs1]b , meaning that the subsets of nodes in question are now only the ones given by @xmath167 , @xmath161 , @xmath168 , @xmath163 .
these channels ( or joints ) form an exponential family ( see @xcite ) which we call @xmath169 .
.5 a , with correlation between the outputs .
b ) the non - complex model of figure [ fig : graphs1]b , with correlation between the outputs . , title="fig : " ] [ fig : graph32 ] .45 a , with correlation between the outputs .
b ) the non - complex model of figure [ fig : graphs1]b , with correlation between the outputs . , title="fig : " ] [ fig : gp2 ] suppose now , though , that the outputs are not conditionally independent anymore , because of some `` noise '' ( see @xcite and @xcite ) .
this way the interaction structure would look like figure [ fig : graphs3 ] , i.e. the `` complete '' subset given by @xmath170 with @xmath171 and @xmath172 .
in particular , a non - complex but `` noisy '' system would be represented by figure [ fig : graphs3]b , and have subsets of nodes given by the pairs @xmath158 , @xmath173 with : @xmath167 , @xmath161 , @xmath168 , @xmath163 , @xmath174 , @xmath175 .
such channels form again an exponential family , which we call @xmath176 .
we would like now to have a measure of complexity for a channel ( or joint ) . in @xcite , the measure of complexity
is defined as the divergence from the family @xmath169 represented in figure [ fig : graphs1]b .
we will call such a measure @xmath177 . in case of noise , however , it is argued in @xcite and @xcite that the divergence should be computed from the family @xmath176 represented in [ fig : graphs3]b ( for example , as written in the cited papers , because such a complexity measure should be required to be upper bounded by the mutual information between @xmath11 and @xmath9 ) .
we will call such a measure @xmath178 .
both divergences can be computed with our algorithm . as an example
, we have considered the following channel : @xmath179 with : @xmath180 here @xmath181 represents a node of `` unknown input noise '' that adds correlation between the outputs ( of unknown form ) when if it is not observed .
we have chosen @xmath182 and @xmath183 , and a uniform input probability @xmath36 .
after marginalizing out @xmath181 ( obtaining then an element of the type of figure [ fig : graphs3]a ) , we can compute the two divergences : * @xmath184 . *
@xmath185 .
this could indicate that @xmath177 is incorporating part the correlation of the output nodes due to the `` noise '' , and therefore probably overestimating the complexity , at least in this case .
one could nevertheless also argue that @xmath178 can underestimate complexity , as we can see in the following `` control '' example .
consider the channel : @xmath186 with : @xmath187 which is represented by the graph in figure [ fig : graphs1]a . if the difference between @xmath177 and @xmath178 were just due to the noise , then for our new channel @xmath188 and @xmath189 should be equal .
this is not the case : * @xmath190 . * @xmath191 .
the divergences are still different .
this means that there is an element @xmath192 in @xmath176 , which does _ not _ lie in @xmath169 , for which : @xmath193 the difference is this time smaller , which could mean that noise still does play a role , but in rigor it is hard to say , since none of these quantities is linear , and divergences do not satisfy a triangular inequality .
we do not want to argue here in favor or against any of these measures .
we would rather like to point out that such considerations can be done mostly after explicit computations , which can be done with iterative scaling .
5 csiszr , i. and shields , p. c. . , 1(4):417528 , 2004 goodman , j. . : 916 , 2002 .
olbrich , e. , bertschinger , n. , and rauh , j. . , 17(5):35013517 , 2015 perrone , p. and ay , n. .
, 35(2 ) , 2016 . ay , n. . ,
17 , 24322458 , 2015 .
oizumi , m. , tsuchiya , n. , and amari , s. , 2015 .
amari , s. springer , 2016 .
kakihara , y. .
world scientific , 1999 .
csiszr , i. .
, 3:146158 , 1975 .
csiszr , i. and mat , f. .
, 49:14741490 , 2003 .
amari , s. .
, 47(5):17011709 , 2001 .
lauritzen , s. l. .
oxford , 1996 .
williams , p. l. and beer , r. d. . , 2010 .
amari , s. and nagaoka , h. . , 1982 .
amari , s. and nagaoka , h. .
oxford , 1993 .
@xmath194 : choose a basis @xmath195 of @xmath51 .
define the map @xmath196 , with : @xmath197 and : @xmath198 then : @xmath199 + \operatorname{\mathbb{e}}_p[\log z_\theta]\;.\ ] ] deriving ( where @xmath200 is w.r.t .
@xmath201 ) : @xmath202 + \operatorname{\mathbb{e}}_p\left [ \dfrac{\operatorname{\partial}_j z_\theta}{z_\theta } \right]\;.\ ] ] the term in the last brackets is equal to : @xmath203 so that now reads : @xmath204 + \operatorname{\mathbb{e}}_{pk_\theta}[f_j]\;.\ ] ] this quantity is equal to zero for every @xmath205 if and only if @xmath206 . now if @xmath207 is a minimizer , it satisfies , and so @xmath206 .
viceversa , suppose @xmath206 , so that it satisfies for every @xmath205 . to prove that it is a global minimizer , we look at the hessian : @xmath208 this is precisely the covariance matrix of the joint probability measure @xmath209 , which is positive definite .
@xmath210 : for every @xmath211 , we have : @xmath212\;. \end{aligned}\ ] ] if @xmath213 , then : @xmath214 = d_p(m||k ' ) + \operatorname{\mathbb{e}}_{pm}\left [ \log\dfrac{k'}{k_0 } \right]\;.\ ] ] by definition of @xmath34 , the logarithm in the last brackets belongs to @xmath51 , and since @xmath211 : @xmath215 = \operatorname{\mathbb{e}}_{pk}\left [ \log\dfrac{k'}{k_0 } \right ] = \operatorname{\mathbb{e}}_{pk'}\left [ \log\dfrac{k'}{k_0 } \right]\;.\ ] ] inserting in : @xmath216 = d_p(m||k ' ) + d_p(k'||k_0)\;.\ ] ] since @xmath217 , shows that @xmath67 is a minimizer . since @xmath218 is strictly convex in the first argument
, its minimizer is unique . for @xmath55 in @xmath14 :
@xmath220 = \sum_{x , y } p(x)\,k(x;y)\,f(x , y ) = \sum_{x_i , y_j } p(x_i)\,k(x_i;y_j)\,f(x_i;y_j)\;,\ ] ] and just as well : @xmath221 = \sum_{x , y } p(x)\,\bar{k}(x;y)\,f(x , y ) = \sum_{x_i , y_j } p(x_i)\,\bar{k}(x_i;y_j)\,f(x_i;y_j)\;.\ ] ] the definition in ( with strict positivity of @xmath36 ) requires exactly that : @xmath222=\operatorname{\mathbb{e}}_{p\bar{k}}[f]\ ] ] for every @xmath223 . using and
, the equality becomes : @xmath224 for every @xmath55 in @xmath14 , which means that @xmath225 .
take as initial distribution @xmath233 and form as in theorem [ jointit ] the sequence @xmath83 of @xmath78-projections . according to theorem [ jointit
] , this sequence converges to the @xmath78-projection of @xmath234 on @xmath51 . since @xmath235}(p)$ ] , this projection will have input marginal equal to @xmath23 , and so we can write it as @xmath236 for some uniquely defined channel @xmath142 .
we have , for @xmath237 : @xmath238 so in particular , for the subsequence of even - numbered terms also : @xmath239 this subsequence is defined iteratively by : @xmath240}^{p } \operatorname{\sigma}_{i_jj_j}^{p \bar k } q^{2(j-1)}\;.\ ] ] propositions [ jlevel ] and [ inscale ] imply then that : @xmath241 for every @xmath205 , where @xmath242 is the sequence defined in the statement of theorem [ convergence ]
. therefore this sequence converges : @xmath243 since @xmath244 for all @xmath245 , @xmath246 for all @xmath245 because of , which by definition means that @xmath247 .
moreover , @xmath248 is the @xmath78-projection @xmath79 of @xmath234 on @xmath51 , which means that : @xmath249 using the chain rule of the kl - divergence , we get : @xmath250 which means that @xmath142 is the @xmath78-projection of @xmath138 on @xmath72 . | here we define a procedure for evaluating kl - projections ( i- and ri - projections ) of channels
. these can be useful in the decomposition of mutual information between input and outputs , e.g. to quantify synergies and interactions of different orders , as well as information integration and other related measures of complexity .
the algorithm is a generalization of the standard iterative scaling algorithm , which we here extend from probability distributions to channels ( also known as transition kernels ) .
* keywords : * markov kernels , hierarchy , i - projections , divergences , interactions , iterative scaling , information geometry . |
for the solution of large stiff problems of the type@xmath0where @xmath1 arises from the discretization of unbounded sectorial operators and @xmath2 is a nonlinear function , in recent years much work has been done on the construction of exponential integrators that might represent a promising alternative to classical solvers ( see e.g. @xcite or @xcite for a comprehensive survey ) . as well known the computation of the matrix exponential or related functions of matrices is at the core of this kind of integrators .
the main idea is to damp the stiffness of the problem ( assumed to be contained in @xmath3 ) on these computations so that the integrator can be explicit . under the hypothesis
that the functions of matrices involved are exactly evaluated , the linear stability can be trivially achieved for both runge - kutta and multistep based exponential integrators and hence highly accurate and stable integrators can be constructed . on the other hand ,
the main problem with this class of integrators is just the efficient computation of such functions of matrices , so that , very few reliable codes have been written ( we remember the rosenbrock type exponential integrators presented in @xcite , @xcite , @xcite ) .
for this reason many authors are still doubtful about the potential of exponential integrators with respect to classical implicit solvers even for semilinear problem of type ( [ pr ] ) .
an exponential integrator requires at each time step the evaluation of a certain number ( depending on the accuracy ) of functions of matrices of the type @xmath4 , where @xmath5 being @xmath6 the time step .
actually this represents the general situation for the exponential time differencing methods , that is , the methods based on the variation - of - constants formula ; for lawson s type method ( also called integrating factor methods ) only the matrix exponential is involved .
we refer again to @xcite and the reference therein for a background . among the existing techniques for the computation of functions of matrices
( we quote here the recent book of higham @xcite for a survey ) , in this context the restricted - denominator ( rd ) rational arnoldi algorithm introduced independently in @xcite and @xcite for the computation of the matrix exponential seems to be an reliable approach .
it is based on the use of the so called rd rational forms , studied in @xcite for the exponential function , @xmath7where @xmath8 is a polynomial of degree @xmath9 we refer again to morno for the basic references about the properties and the use of such rational forms . while in the matrix case , the use of these approximants requires the solution of linear systems with the matrix @xmath10 , as shown in @xcite in the context of the solution of ( [ pr ] ) when @xmath3 is sectorial so typically sparse and well structured this linear algebra drawback can be almost completely overtaken organizing suitably the step - size control strategy and exploiting the properties of the rd arnoldi method concerning the choice of the parameter @xmath11 . in other words the number of linear systems to be solved can be drastically reduced with respect to the total number of computations of functions of matrices required by the integrator .
therefore the mesh independence property of the method , that leads to a very fast convergence with respect to a standard polynomial approach ( see again @xcite ) , can be fully exploited for the construction of competitive integrators .
a problem still open is that inside the integrator the rational arnoldi algorithm ( responsible for most of the computational cost ) have to be supported by a robust and sharp error estimator . in the self - adjoint case
the problem has been treated in @xcite where the author presents effective a - posteriori error estimates , even in absence of information on the location of the spectrum of @xmath3 .
anyway , in the general case , when ( pr ) arises for instance from the discretization of parabolic problems with advection terms and/or non - zero boundary conditions the numerical range of @xmath12 , that we denote by @xmath13 , may not reduce to a line segment . in this sense
the basic aim of this paper is to fill this gap providing error estimates for the non - symmetric case using as few as possible information about the location of @xmath13 .
it is necessary to keep in mind that a competitive code for ( [ pr ] ) should also be able to update @xmath3 ( interpreted as the jacobian of @xmath14 , @xcite , @xcite ) so that @xmath13 is may be not fixed during the integration , and so it is important to reduce as much as possible any pre - processing technique to estimate @xmath13 .
in particular assuming that @xmath15we shall provide a - posteriori error estimates for the rd arnoldi process using only information about the angle of the sector containing @xmath13 , angle that is typically independent of the sharpness of the discretization and hence computable working in small dimension .
the paper is organized as follows . in section 2
we present the basic idea of the rd rational arnoldi method and in section 3 we derive some first general error bounds based on the standard approaches . in section 4 , exploiting the relation between the derivatives of the function @xmath16 and the laguerre polynomials extended to the complex plane , we derive some a - posteriori error bounds .
the problem of defining reliable a - priori bounds is investigated in section 5 .
section 6 is devoted to the analysis of the generalized residual as error estimator , that can be used to obtain information about the choice of the parameter @xmath11 for the rational approximation . in section 7
we present some numerical examples arising from the discretization of a one - dimensional advection - diffusion model . in section 8
we provide some hints about the use of the rd rational arnoldi method inside an exponential integrator with the aim of reducing as much as possible the number of implicit computations of @xmath17 .
finally , in section 9 we furnish a deeper analysis concerning the fast rate of convergence of the method , that will provide further information about the optimal choice of the parameter @xmath11 .
in what follows we denote by @xmath18 the euclidean vector norm and its induced matrix norm . as already mentioned , the notation @xmath19 indicates the _ numerical range _ of @xmath3 , that is , @xmath20while the spectrum of @xmath3 is denoted by @xmath21 . the notation @xmath22 indicates the space of the algebraic polynomials of degree @xmath23 . given @xmath24 ,
let@xmath25be the unbounded sector of the left half complex plane , symmetric with respect to the real axis with vertex in @xmath26 and semiangle @xmath27 .
let moreover @xmath28 be the boundary of @xmath29 . throughout the paper
we assume that @xmath30 , the interior of @xmath31 .
accordingly , @xmath3 is a so - called sectorial operator ( see e.g. @xcite chap .
v , for a background ) . given a vector @xmath32 , with @xmath33 ,
consider the problem of computing @xmath34where @xmath35 is defined by ( [ defi ] ) .
the rd rational approach seeks for approximations to @xmath36 of the type @xmath37where @xmath38 is a suitable parameter .
turning to the matrix case , @xmath39 is approximated by elements of the krylov subspaces @xmath40 with respect to @xmath41 and the matrix @xmath42 defined by the transform @xmath43 in this sense the idea is to use a polynomial method to compute @xmath44 , where@xmath45is singular at @xmath26 .
for the construction of the subspaces @xmath46 we employ the classical arnoldi method .
as is well known it generates an orthonormal sequence@xmath47 , with @xmath48 , such that @xmath49 .
moreover , for every @xmath50 , @xmath51where @xmath52 $ ] , @xmath53 is upper hessenberg matrix with entries @xmath54 and @xmath55 is the @xmath56-th vector of the canonical basis of @xmath57 . the @xmath58-th rd - rational arnoldi approximation to @xmath59 is defined as ( see @xcite)@xmath60it can be seen that @xmath61where @xmath62 @xmath63 interpolates , in the hermite sense , the function @xmath64 in the eigenvalues of @xmath53 ( see saad2 ) . as mentioned in the introduction
this technique has been introduced independently in @xcite and @xcite .
anyway , the idea of using rational krylov approximations to matrix functions was originally introduced in @xcite .
more recently this approach has been extended to the case of multiple poles and is commonly referred to as rks ( rational krylov subspace ) approximation ( see @xcite , @xcite , @xcite ) .
before stating a general error bound for the method , we need to locate @xmath65 .
consider the function @xmath66 . denoting by @xmath67 the disk centered in @xmath68 with radius @xmath68 , let@xmath69its boundary , @xmath70 ,
is made by two circular arcs meeting with angle @xmath71 at @xmath26 and @xmath72 . regarding the field of values of @xmath42 , @xmath65 , we can state the following result that will be used frequently throughout the paper .
[ pint]if @xmath30 then @xmath73 obviously @xmath74 , so @xmath65 can not lie entirely outside @xmath75 .
now assume that there exists @xmath76 such that @xmath77 , that is , @xmath78 .
hence , there exists @xmath79 , @xmath80 , such that@xmath81defining @xmath82 we easily obtain@xmath83and hence@xmath84by ( [ rel1 ] ) we have@xmath85now let us define @xmath86 .
we have@xmath87and hence@xmath88that implies@xmath89now since @xmath90 and @xmath91 , by ( [ cmp ] ) it must be @xmath92 and @xmath93 so that @xmath94using this relation , by ( [ xn ] ) we finally have@xmath95that contradicts ( [ frel ] ) . since the field of values is connected the proof is complete . [ rem1]in order to provide information about the geometry of @xmath65 , it is worth referring to @xcite theorem 5.2 in which the author proves that if @xmath3 is an invertible matrix then@xmath96taking @xmath97 , we have that for small values of @xmath11@xmath98 going back to our method , the corresponding error @xmath99 can be expressed and bounded in many ways ( we quote here the recent papers @xcite and @xcite for a background on the error estimates for both polynomial and rational arnoldi approximation to matrix functions ) .
the following proposition states a general result .
[ stp]let @xmath100 be a compact such that @xmath101 and whose boundary @xmath102 is a rectifiable jordan curve . for every @xmath103@xmath104
using the properties of the arnoldi algorithm we know that for every @xmath105 , @xmath106hence from this identity it follows that , for @xmath107 @xmath108now since @xmath109 we can write ( [ fe ] ) in the dunford - taylor integral form@xmath110 dz.\]]collecting @xmath111 and using ( see @xcite)@xmath112we prove ( [ ge ] ) .
now since@xmath113where@xmath114(see @xcite ) , any bound for @xmath115 and any choice for @xmath116 and @xmath117 leads to a bound for @xmath118 .
this technique has been used for instance in @xcite and @xcite . in particular in @xcite the authors use the relation@xmath119and the inequality@xmath120 going back to our situation , the main problem is that if we simply assume that @xmath121 ( in other words @xmath13 arbitrarily large ) we have that @xmath122 as @xmath123 ( @xmath124 ) because we have to consider the singularity of @xmath125 at @xmath26 . therefore using a lower bound like ( [ lb ] ) ( but the situation remains true even for other approaches ( cf .
@xcite ) ) terms of the type @xmath126 would appear in ( [ ge ] ) .
in the exponential case ( @xmath127 ) this is not a problem because @xmath128 for @xmath129 , but for @xmath130 the situation changes completely since @xmath131for @xmath123 .
because of the difficulties just explained , our approach for deriving error bounds is not based on the use of the cauchy integral formula . exploiting the interpolatory nature of the standard arnoldi method , we notice , as pointed out also in @xcite , that the error can be expressed in the form @xmath132where ( cf . ( [ int]))@xmath133 in @xcite this relationship is used as the basis for the construction of restarted methods for the computation of matrix functions . we can state the following basic result that will be used throughout the paper and that allows to overcome the difficulties of working with formula ( [ ge ] ) . [ pg]let @xmath121 and let @xmath134 . then@xmath135where
@xmath136 . in the symmetric case
we can take @xmath137 . by @xcite we know that@xmath138and hence by ( [ try ] ) and ( [ erapr ] ) @xmath139now , by induction one proves that for @xmath140 @xmath141where @xmath142 and@xmath143putting ( [ fk ] ) in ( [ gkm ] ) we obtain@xmath144now , the polynomial @xmath145 interpolates in the hermite sense the function @xmath146 in the eigenvalues of @xmath53 and in @xmath147 .
henceforth @xmath148 is a divided difference that can be bounded using the hermite - genocchi formula ( see e.g. @xcite ) , so that @xmath149where @xmath150 denotes the convex hull of the point set given by @xmath151 , @xmath152 and @xmath72 .
since @xmath153 , and @xmath154 by proposition [ pint ] , the result follows .
by ( [ erb ] ) , in order to provide a - posteriori error estimates we just need to study the derivatives of the function @xmath155 . we need to introduce the generalized laguerre polynomials , defined by @xmath156we can state the following result .
let @xmath157 .
for @xmath107 @xmath158 first of all remember that @xmath159 . defining @xmath160 and using rodrigues formula for laguerre polynomials ( see @xcite p.101 ) we obtain@xmath161the result arises from the relation ( see @xcite p.240)@xmath162 before stating the main result we need to remember the following properties of the generalized laguerre polynomials , that can be found in @xcite pp . 785 - 786 . 1 .
@xmath163 2 .
@xmath164 3 .
@xmath165 given @xmath166 , let @xmath167 .
let moreover@xmath168then@xmath169 for @xmath170@xmath171using l1 with @xmath172 , @xmath173 , @xmath174 and @xmath175 , and then l2 with @xmath176 and @xmath177 , we have@xmath178since @xmath179formulas ( [ lk1 ] ) and ( [ lk2 ] ) are obtained applying l3 to @xmath180 and then to @xmath181 .
[ pro1]assume that @xmath121 , with @xmath182 .
then@xmath183where@xmath184and @xmath185 defined as in proposition [ pg ]
. for @xmath186@xmath187and@xmath188hence , using ( [ erb ] ) , ( [ ede ] ) and ( [ lk1 ] ) we obtain@xmath189since for @xmath190@xmath191(looking for the maximum with respect to @xmath192 ) , we immediately obtain ( fe1 ) . using again ( [ erb ] ) and ( [ ede ] ) but now with ( [ lk2 ] ) we arrive at the coarser bound ( [ fe2 ] ) . while formulas ( [ fe1 ] ) and ( [ fe2 ] ) theoretically hold for @xmath193 since @xmath194 for @xmath195 , it is necessary to point out that for @xmath196we may observe a rapid growth of the term@xmath197depending of course on the problem , so that the bounds may be useless .
this situation is caused by the bound ( [ lk1 ] ) that leads to the appearance of the term @xmath198 at the denominator .
working in inexact arithmetics the situation is even more difficult because of the loss of orthogonality of the vectors @xmath199 of the arnoldi algorithm and hence the accumulation of errors on the entries @xmath200 . for these reasons , in practice , formulas ( [ fe1 ] ) and ( [ fe2 ] ) should be used only for @xmath201 not much close to @xmath202 . for the exponential case ( @xmath127 ) we have@xmath203and hence by
( [ fe2])@xmath204 in the self - adjoint case @xmath205 we have @xmath206 and formula ( [ fe2 ] ) simplifies to@xmath207 the reason for which we consider two bounds in theorem [ pro1 ] is that the second one ( [ fe2 ] ) allows us to define suitably the parameter @xmath208 ( and then @xmath11 ) while the first one ( [ fe1 ] ) should be used whenever @xmath208 has been defined . indeed , assuming @xmath209 independent of @xmath11 and then of @xmath208 ( actually this is not true as we explain in section 9 ) by ( [ fe2 ] ) , looking for the minimum of @xmath210 we easily find that the optimal value for @xmath208 is given by@xmath211 the a - posteriori bounds provided by theorem [ pro1 ] depends substantially on the semiangle @xmath27 of the sector containing @xmath13 .
therefore , the most natural way to proceed is to compute the boundary of @xmath13 using the standard codes available in literature ( as for instance the matlab code ` fv.m ` by higham @xcite ) .
it is important to observe that @xmath27 is generally independent of the discretization so that one can work in smaller dimension . while the hypothesis @xmath121 of theorem [ pro1 ] is extremely general , the underlying assumption is that @xmath3 represents an arbitrary sharp discretization of an unbounded operator . on the other side ,
if it is known that @xmath13 is contained in a bounded sector then proposition [ stp ] can be used to derive sharper error estimates . in general
we may refer again to @xcite and the references therein for a background on the most used techniques based on the use of the integral representation of the error . anyway , here we want also to show how to adapt our approach in precence of more information on @xmath13 .
let @xmath212 be the disk centered at @xmath26 with radius @xmath213 , and assume that @xmath214 . using again ( [ erb ] ) and ( [ ede ] ) , we arrive at the bound@xmath215 in order to define a suitable value for @xmath208 , we just need to bound the laguerre polynomials as in ( [ lk2 ] ) , so that the optimal value is obtained ooking for the minimum of@xmath216a good approximation for this minimum is given by@xmath217that is obtained considering the bound@xmath218using this value of @xmath208 we can derive practical error bounds seeking for the maximum of the function @xmath219 ( cf .
( [ pri ] ) ) in the interval @xmath220 $ ] .
formula ( [ topt ] ) obviously requires to know the number of iterations that are necessary to achieve a certain accuracy . in this sense
we need to bound in some way @xmath221 . by ( [ try ] ) and since@xmath222for each monic polynomial @xmath223 of exact degree @xmath58 ( see @xcite p. 269 ) , a bound for @xmath221 can be stated using faber polynomials as explained in @xcite , that leads to@xmath224where @xmath225 is the logarithmic capacity of a compact@xmath226containing @xmath65 and where @xmath125 is analytic .
[ pap]let @xmath227 and assume that @xmath228 , with @xmath229
. then for @xmath230 @xmath231where@xmath232 since @xmath154 by proposition [ pint ] , let us consider the compact subset @xmath233 .
the associated conformal mapping @xmath234is given by@xmath235where @xmath236 .
the coefficient of the leading term of the laurent expansion ( [ le ] ) is the logarithmic capacity , so that by ( be ) we have@xmath237inserting this bound in ( [ fe2 ] ) we easily obtain for @xmath238@xmath239where the second inequality arises from the choice @xmath240 .
now , by the definition ( [ cj ] ) , it is rather easy to show that@xmath241so that@xmath242 ^{m}. \label{app}\]]since @xmath243 for @xmath244 , and since for each @xmath245@xmath246the proof is complete
. proposition [ pap ] shows the mesh - independence of the method for @xmath247 since the bound ( [ ap ] ) is independent of the discretization of the underlying sectorial operator . in section 9 this considaration
is extended to @xmath248 . by ( [ app ] ) and ( [ rot ] ) , in the self - adjoint case ( @xmath137 ) the bound ( [ ap ] ) reads@xmath249 it is worth noting that by ( [ fe ] ) for every @xmath103 we have that @xmath250where we assume that @xmath251 is compact , connected , with associated conformal mapping @xmath252 , and such that @xmath253 .
therefore , in principle , one could try to derive a - priori error bounds choosing suitably the polynomial sequence @xmath254 .
anyway , the classical results in complex polynomial approximation state that even taking @xmath255 as a sequence of polynomials that asymptotically behaves as the sequence of polynomial of best uniform approximation of @xmath125 on @xmath116 ( see e.g @xcite for a theoretical background and examples ) we have@xmath256 ^{1/m}\rightarrow \frac{1}{r}\text{\quad as } m\rightarrow \infty , \]]where @xmath213 is such that @xmath257 , since @xmath125 is singular at 0 ( _ maximal convergence _ property , see e.g @xcite chapter iv ) .
the main problem is that assuming @xmath3 to be unbounded , @xmath258 and consequently @xmath259 . for this reasons , in our opinion
the only reasonable approach to derive a - priori error bounds , is to define @xmath255 as a sequence of polynomials interpolating @xmath125 at point belonging to @xmath116 , and then to use the hermite - genocchi formula to bound the divided differences . using this formula and taking for instance @xmath117 as the sequence of interpolants at the zeros of faber polynomials
we just obtain the error bound given in proposition [ pap ] ( see @xcite ) .
by the integral representation of function of matrices and ( [ kn ] ) , we know that the error can be written as@xmath260dz
. \label{fot1}\]]in order to monitor the approximations during the computation we can consider the so - called generalized residual @xcite , defined as @xmath261which is obtained from ( [ fot1 ] ) by replacing the error@xmath262with the corresponding residual @xmath263using the fundamental relation ( [ i1 ] ) we have immediately@xmath264and inserting this relation in ( [ res ] ) we obtain @xmath265so that we may assume @xmath266 in order to show the reliability of this approximation let us consider the operator @xmath267 discretized with central differences in @xmath268 $ ] with uniform mesh @xmath269 , and dirichelet boundary conditions . for our examples , we consider the computation of @xmath4 for @xmath270 , with @xmath271 , comparing the exact error and the generalized residual .
we take @xmath272 , @xmath273 , and we consider the cases of @xmath274 and @xmath275 , whose corresponding sector semiangles are respectively @xmath276 and @xmath277 .
we define @xmath278 .
the results , collected in figure 1 , shows the accuracy of the approximation ( [ gr ] ) .
it is necessary to point out that the use of ( [ gr ] ) has the basic disadvantage that it requires the computation of @xmath279 , @xmath280 , and this is a computational drawback whenever a great amount of matrix functions evaluations are required to integrate a certain problem , even if @xmath281 can be considered much smaller than @xmath282 .
moreover , it frequently happens ( as in our experiments ) that the generalized residual tends to underestimate the error during the first iterations , and this can be particularly dangerous when computing @xmath283 with @xmath284 , as for instance in the case of the computation of the internal stages of an exponential runge - kutta method , in which @xmath285 . on the other side , exploiting the mesh independence of the method the generalized residual can be successfully used to estimate the optimal value for the parameter @xmath208 , that is @xmath286 . in other words , using a coarser discretization of the operator we look for the value of @xmath58 such that using the corresponding @xmath287 we obtain a certain tolerance in exactly @xmath58 iterations . for the experiments reported in figure 1 we considered the discretization of ( [ lu ] ) with only @xmath288 internal points , observing in both cases that @xmath289 for @xmath290 .
for this reason we have chosen @xmath291 .
for our numerical experiments we consider again the operator ( [ lu ] ) , discretized as in previous section .
we consider the computation of the functions @xmath4 , with @xmath41 as before and @xmath292 , for @xmath293 ( figure 2 ) and @xmath294 ( figure 3 ) . in all examples
we do not consider the symmetric case corresponding to @xmath295 ( already investigated in @xcite ) , but only the cases @xmath274 and @xmath275 .
as before , for the choice of @xmath208 we examined the behavior of the method for the coarser discretization of the same operator with only @xmath288 interior points , thus exploiting the mesh independence of the method .
the analysis suggested to take @xmath296 for all experiments with @xmath293 and @xmath278 for those with @xmath294 , thus independently of the function and @xmath297 , using the tolerance @xmath298 .
inside the arnoldi iterations the vectors @xmath299 , @xmath300 ( cf . section 2 ) , are computed via the lu factorization of @xmath301 .
the error bound is given by ( [ fe1 ] ) .
comparing figure 2 with figure 3 we can observe that the method tends to become slower reducing @xmath6 .
the reason is that for small values of @xmath6 , the rate of the decay of the singular values of @xmath42 becomes slower and this reduces the rate of the decay of @xmath221
. a deeper analysis of this behavior will be presented in section 9 .
employing the rd arnoldi method inside an exponential integrator requires some considerations .
first of all , in our opinion the method can be used only if the implicit computation of @xmath42 can be obtained with a sparse factorization technique .
the use of an inner - outer iteration can be too much expensive in this context .
indeed , the basic point is that organizing suitably the code one can heavily reduce the number of factorizations of @xmath303 ( see e.g @xcite ) , because the method seems to be really robust with respect to the choice of @xmath208 .
for this reason we want here to show what happens taking @xmath208 even quite far from the optimal one . for simplicity ( the situation is representative of what happens in general )
let us assume to work with exponential function and @xmath205 .
we assume moreover that the corresponding bound ( [ xp ] ) ( in which @xmath206 , @xmath304 ) is equal to a prescribed tolerance for a certain @xmath58 with the theoretical optimal choice @xmath305 .
we seek for the interval @xmath306 $ ] such that for @xmath307 the number of iterations necessary to achieve the same tolerance is at most @xmath308 ( @xmath309 ) . using ( [ xp ] ) and the approximation @xmath310 ( @xmath311 ) that is obtained forcing the equal sign in the a - priori bound ( [ bh ] ) , in figure 4 we can observe the result for @xmath312 . for each @xmath58
the corresponding extremal points @xmath313 and @xmath314 of the intervals @xmath315 and @xmath316 are plotted .
these points are obtained solving with respect to @xmath208 the equation ( cf .
( [ xp]))@xmath317for @xmath312 .
we point out that the results are even a bit conservative with respect to what happens in practice , and this is due to the approximation @xmath318 .
indeed larger intervals would be obtained taking @xmath319 as it occurs in practice . in
order prove the effectiveness of the above considerations let us consider again the operator ( [ lu ] ) with the usual discretization .
we consider the case @xmath274 , @xmath320 for @xmath273 . to define @xmath208
we consider again the discretization with @xmath288 interior points observing the generalized residual .
this leads us to define @xmath240 with @xmath321 . in figure 5
we consider the behavior of the method for @xmath208 , @xmath322 and @xmath323 .
the robustness of the method with respect to the choice of @xmath208 is maybe the most important aspect concerning its use inside an exponential integrator .
we want to give here some practical suggestions assuming to use a sparse factorization technique to solve the linear systems with @xmath324 , that , computationally , has to be considered the heaviest part of the method . 1 .
working in much smaller dimension compute @xmath27 and use the generalized residual to estimate the initial @xmath287 .
2 . for nonlinear problems , interpreting @xmath3 as the jacobian of the system ( @xcite , @xcite ) , it is necessary to introduce some strategies in order to reduce as much as possible the number of updates of @xmath3 during the integration , since each update would also require to update the factorization . as for exponential w - method ( see @xcite , @xcite ) , we suggest , whenever it is possible , to work with a time - lagged jacobian and hence to introduce the necessary order conditions in order to preserve the theoretical order .
3 . using a quasi - constant step - size strategy ( without jacobian update ) allows to keep the factorization of @xmath325 constant for a certain number of steps . whenever it is necessary to update the stepsize @xmath326 without changing the jacobian , if we want to keep the previous factorization of @xmath327 we just need to consider the ratio @xmath328 . if ( indicatively ) it is bigger than @xmath329 or smaller than @xmath330 ( cf . figure 4 and 5 ) , where @xmath287 arises from a previous analysis of the generalized residual , then we need to update the factorization ( cf . again
@xcite ) , otherwise we can keep the previous one . in this phase , however , one can even considers other strategies to define suitably the window of admissible values of @xmath208 around @xmath287 , taking into account of the local accuracy required by the integrator , the norm of @xmath41 , etc .
looking carefully at figure 5 we notice that while the analysis in smaller dimension suggested to take @xmath278 for reaching the desired tolerance in exactly @xmath331 iterations the method is unexpectedly a bit faster taking @xmath332 ( second picture ) .
the analysis was correct because in larger dimension the method actually achieves the tolerance in @xmath333 iterations ( first picture ) . in order to understand the reason of this behavior , we need to remember that the definition of @xmath334 given at the end of section 4 was based on the assumption that @xmath335 is independent of @xmath336 but this is not true . in what follows we try to provide a more accurate analysis studying the decay of @xmath221 .
we denote by @xmath337 , @xmath338 , the singular values of @xmath42 .
moreover we denote by @xmath339 , @xmath338 the eigenvalues of @xmath42 and assume that @xmath340 for @xmath338 .
we have the following result ( cf .
ne theorem 5.8.10 ) .
[ nev]assume that @xmath341 and@xmath342let @xmath343 .
then@xmath344where@xmath345 as already shown in section 4@xmath346for each monic polynomial @xmath223 of exact degree @xmath58 ( see @xcite p. 269 ) , so that theorem [ nev ] reveals that the rate of decay of @xmath335 is superlinear and depends on the @xmath347-summability of the singular values of @xmath42 .
we remark moreover that an almost equal bound has been obtained in @xcite studying the convergence of the smallest ritz value of the lanczos process for self - adjoint compact operators . in practice
, the use of ( [ bn ] ) requires the knowledge of @xmath347 and a bound for @xmath348 , that is , information about the singular values of the operator @xmath349 . as a model problem
we consider again the operator @xmath3 defined by ( lu ) with @xmath295 , whose eigenvalues are @xmath350 , @xmath338 , so that the eigenvalues of @xmath42 are given by @xmath351 . in this case
( [ psum ] ) holds for @xmath352 so that @xmath42 can be referred to as a _ trace class _ operator ( see again @xcite ) .
hence , taking for instance @xmath353 we have@xmath354and so@xmath355the bound ( [ pmh ] ) reveals that the rate of decay depends on the choice of @xmath11 and then on @xmath6 . for large values of @xmath11 ,
say @xmath356 , the bound ( [ tb ] ) can be heavily improved exploiting the properties of the @xmath357 function and the convergence is extremely fast .
the following proposition states a general superlinear bound that can be used when @xmath3 is an elliptic differential operator of the second order , so with singular values growing like @xmath358 .
the proof is straightforward since we just require to bound @xmath359 , and apply ( [ bn ] ) with @xmath353 .
let @xmath3 be an elliptic differential operator of the second order .
then there exists a constant @xmath360 such that@xmath361 this proposition can easily be generalized to operator of order @xmath362 , exploiting @xcite corollary 5.8.12 in which the author extends theorem [ nev ] for @xmath363 . anyway , this is beyond the purpose of this section . from a practical point of view ,
formula ( [ hh ] ) is almost useless since too much information on @xmath3 would be required . on the other side ,
it is fundamental to understand the dependence on @xmath11 .
setting as usual @xmath364 and putting the corresponding bound ( [ hh ] ) in theorem [ pro1 ] ( formula ( [ fe2 ] ) ) , we easily find that the theoretical optimal value for @xmath208 is obtained seeking for the minimum of@xmath365with respect to @xmath208 , that is,@xmath366this new value , less than @xmath367 , explains our considerations about figure 5 given at the beginning of this section .
we need to point out that since the choice of @xmath287 is independent of @xmath360 and @xmath6 , formula ( [ hh ] ) is quite coarse for small values of @xmath6 and not able to catch the fast decay of @xmath221 . in any case , if an estimate of @xmath360 is available an a - priori bound for the rd arnoldi method can be obtained taking@xmath368(cf .
( [ bh ] ) ) .
consequently we argue that@xmath369with @xmath287 close to @xmath370 for @xmath6 large and to @xmath367 for @xmath6 small .
in this paper we have tried to provide all the necessary information to employ the rd arnoldi method as a tool for solving parabolic problems with exponential integrators . the little number of codes available in literature , and consequently , the little number of comparisons with classical solvers is a source of skepticism about the practical usefulness of this kind of integrators . indeed , with respect to the most powerful classical methods for stiff problems , the computation of a large number of matrix functions , generally performed with a polynomial method , is still representing a drawback because of the computational cost .
the use of polynomial methods for these computations may even be considered inadequate whenever we assume to work with an arbitrarily sharp discretization of the operator , since this would result in a problem of polynomial approximation in arbitrarily large domains . for these reasons , the use of rational approximations as the one here presented ,
should be considered a valid alternative because of the fast rate of convergence and the mesh independence property , provided that we are able to exploit suitably the robustness of the method with respect to the choice of the poles , as explained in section 8 for our case | in this paper we investigate some practical aspects concerning the use of the restricted - denominator ( rd ) rational arnoldi method for the computation of the core functions of exponential integrators for parabolic problems .
we derive some useful a - posteriori bounds together with some hints for a suitable implementation inside the integrators .
numerical experiments arising from the discretization of sectorial operators are presented . |
though it is almost a century since the inception of quantum mechanics ( qm ) , its foundations and origin remain quite puzzling and continue to inspire intense inquiry . in this letter
, we attempt to illuminate the connection between the mathematical structure of qm and its physical characteristics by constructing a ` mutant ' quantum mechanical model which shares many , but not all of the mathematical features of canonical qm . by
investigating which characteristics of canonical qm survive the ` mutation ' and which ones do not , we hope to clarify the relation between the mathematical genotype and the physical phenotype . in canonical qm ,
the states of an @xmath1-level quantum system are described by vectors in the hilbert space @xmath2 . in the following ,
we introduce a ` mutation ' by replacing @xmath3 with @xmath4 @xcite , where @xmath5 is shorthand for the finite galois field @xmath0 , @xmath6 for some prime @xmath7 , and @xmath8 . for the case @xmath9
, we have @xmath10 . such replacements of the vector space have been considered previously , _
e.g. _ real qm in which @xmath3 is replaced by @xmath11 @xcite , and quaternionic qm in which it is replaced by @xmath12 @xcite .
however , the vector space @xmath13 , in contradistinction to @xmath14 , @xmath3 , or @xmath15 , lacks an inner product , normalizable states , and symmetric / hermitian operators .
nevertheless , we find that we can construct a perfectly ` quantum ' model on it , which predicts probabilities of physical measurements that can not be reproduced in any hidden variable theory .
what will not survive this ` mutation , ' however , are the super - classical correlations of canonical qm . in particular , we show that in our discrete qm , the clauser - horne - shimony - holt ( chsh ) @xcite version of bell s inequality @xcite is not violated .
before we proceed , we emphasize that our model is distinct from ` galois quantum systems ' discussed in the literature @xcite .
there , it is the phase space which is assumed to be @xmath16 , that is , the position and momentum of a particle take on values in @xmath5 . in our approach ,
it is the wave - functions that take on values in @xmath17 , while the outcomes of measurements take on values in @xmath18 .
our starting point is the following canonical expression for the probability of obtaining the outcome represented by the dual - vector @xmath19 when a measurement is performed on the state represented by the vector @xmath20 : @xmath21 here , @xmath22 is not normalized and the sum in the denominator runs over the duals of all the eigenstates of a hermitian operator which represents the observable in question .
however , for this expression to be interpretable as a probability , the necessary condition is that the dual - vectors in the sum span the entire dual vector space @xmath23 , and any reference to operators acting on @xmath3 is inessential .
the interpretation that the bracket @xmath24 is an inner product between two vectors also need not be imposed .
the probability depends only on the absolute values of the brackets @xmath25 . since we can multiply @xmath22 with any non - zero complex number without changing the probabilities defined via eq .
( [ pdef ] ) , we are compelled to identify vectors which differ by a non - zero multiplicative constant as representing the same physical state , endowing the state space with the complex projective geometry @xmath26 where each line going through the origin of @xmath27 is identified as a ` point . ' thus , to construct a ` mutant ' qm on @xmath13 , we represent states with vectors @xmath28 , and outcomes of measurements with dual - vectors @xmath29 .
observables are associated with a choice of basis of @xmath30 , each dual - vector in it representing a different outcome .
the bracket @xmath31 is converted into a non - negative real number @xmath25 via the absolute value function : @xmath32 here , underlined numbers and symbols represent elements of @xmath5 , to distinguish them from elements of @xmath18 or @xmath33 .
note that eq .
( [ abs ] ) is not to be interpreted as a condition imposed on @xmath31 ; all non - zero values of @xmath5 are mapped to one .
since @xmath34 is a cyclic multiplicative group , this assignment of ` absolute values ' is the only one consistent with the requirement that the map from @xmath5 to non - negative @xmath18 be product preserving , that is : @xmath35 . with these assignments , eq .
( [ pdef ] ) can be applied as it stands to calculate probabilities .
since the same absolute value is assigned to all non - zero brackets , all outcomes @xmath36 for which the bracket with the state @xmath22 is non - zero are given equal probabilistic weight .
the product preserving nature of the absolute value function guarantees that the probabilities of product observables on product states factorize in multi - particle systems : @xmath37 this property is crucial if we want to have isolated particle states , and is of course shared by canonical qm defined on @xmath38 .
note also that the multiplication of @xmath22 with a non - zero element of @xmath5 will not affect the probability .
thus , vectors that differ by non - zero multiplicative constants are identified as representing the same physical state , and the state space is endowed with the finite projective geometry @xcite @xmath39 where each ` line ' going through the origin of @xmath17 is identified as a ` point , ' in close analogy to the complex projective geometry of canonical qm .
to give a concrete example of our proposal , let us construct a 2-level system , analogous to spin , for which @xmath40 , and the state space is @xmath41 .
this geometry consists of @xmath42 ` points , ' which can be represented by the vectors @xmath43,\quad { | \,1\ , \rangle } = \left[\begin{array}{c } { \underline{0}}\\ { \underline{1}}\end{array}\right],\quad { | \,r\ , \rangle } = \left[\begin{array}{l } \underline{a}^{r-1 } \\ { \underline{1}}\end{array}\right],\ ] ] @xmath44 , where @xmath45 is the generator of the multiplicative group @xmath34 with @xmath46 .
the number @xmath42 results from the fact that of the @xmath47 non - zero vectors , every @xmath48 are equivalent , thus the number of inequivalent vectors are @xmath49 .
similarly , the @xmath42 inequivalent dual - vectors can be represented as : @xmath50\;,\cr { \langle \,\overline{1}\ , | } & = & \bigl[\,{\underline{1}}\;\;\phantom{-}{\underline{0}}\,\bigr]\;,\cr { \langle \,\overline{r}\ , | } & = & \bigl[\,{\underline{1}}\;-\!\underline{a}^{r-1}\,\bigr]\;,\qquad r=2,3,\cdots , q\;,\end{aligned}\ ] ] where the minus signs are dropped when the characteristic of @xmath5 is two . from these definitions , we find : @xmath51 and @xmath52 observables are associated with a choice of basis of @xmath30 : @xmath53 we assign the outcome @xmath54 to the first dual - vector of the pair , and the outcome @xmath55 to the second to make these observables spin - like .
this assignment implies @xmath56 .
the indices @xmath57 can be considered as indicating the direction of the ` spin , ' and the interchange of the indices as indicating a reversal of this direction . applying eq .
( [ pdef ] ) to this system , it is straightforward to show that @xmath58 and thus , @xmath59 so for each ` spin , ' there exist two ` eigenstates , ' one for @xmath54 ( ` spin ' up ) and another for @xmath55 ( ` spin ' down ) . for
all other states the two outcomes @xmath60 are equally probable . the states and observables `
rotate ' into each other under changes of bases . for the projective geometry @xmath41 ,
the group of all possible basis transformations constitute the projective group @xmath61 of order @xmath62 .
@xmath61 is formally a subgroup of @xmath63 , the group of all possible permutations of the @xmath42 states .
to show that our system is truly quantum , we use an argument analogous to those of greenberger , horne , shimony , and zeilinger @xcite , and of hardy @xcite for canonical qm .
let us construct a two ` spin ' system on the tensor product space @xmath64 .
the number of non - zero vectors in this space is @xmath65 , of which every @xmath48 are equivalent , so the number of inequivalent states is @xmath66 . of these , @xmath67 are product states , leaving @xmath68 that are entangled . as noted previously , eq .
( [ pdef ] ) applied to tensored spaces with the product preserving absolute value function eq .
( [ abs ] ) ensures that the expectation values of product observables factorize for product states , thereby rendering the distinction between product and entangled states meaningful .
the number of entangled states matches the order of the group @xmath61 , since arranging the 4 elements of an entangled state into a @xmath69 array gives rise to a non - singular matrix .
the entangled states fall into ` conjugacy ' classes , matching those of @xmath61 , that transform among themselves under @xmath61 ` rotations . '
the singlet state , corresponding to the conjugacy class of the unit element , can be expressed as @xmath70 for any two states @xmath71 and @xmath72 up to a multiplicative constant .
if the characteristic of @xmath5 is two , the minus sign is replaced by a plus sign .
products of the ` spin ' observables are defined as @xmath73 the four tensor products representing the outcomes @xmath74 , @xmath75 , @xmath76 , and @xmath77 , and the expectation value giving the correlation between the two ` spins . '
the probabilities of the four outcomes are particularly easy to calculate for the singlet state @xmath78 since @xcite @xmath79 thus @xmath80 and we obtain the probabilities and correlations listed in table .
[ probs ] .
.probabilities and expectation values of product observables in the singlet state @xmath78 .
the indices @xmath81 , @xmath82 , @xmath83 , and @xmath84 are distinct .
cases that can be obtained by flipping signs using @xmath85 are not shown . [ cols="^,^,^,^,^,^",options="header " , ] to demonstrate that these correlations can not be reproduced in any hidden variable theory , it suffices to look at the correlations between two observables that share an index .
for instance , consider the following two : @xmath86 first , from the first row of table [ probs ] we can discern that @xmath87 where we have added subscripts to distinguish between the two ` spins . '
this tells us that the pairs @xmath88 and @xmath89 are completely anti - correlated .
next , from the second row of table [ probs ] , we conclude : @xmath90 which means that if either one of the pairs @xmath91 and @xmath92 is @xmath54 , then its partner must be @xmath55 .
thus , the implications of either @xmath93 or @xmath94 would be : @xmath95 in either case , we can not classically have @xmath96 or @xmath97 , even though both configurations have quantum mechanical probabilities of @xmath98 .
thus , our ` mutant ' qm is truly ` quantum ' and its predictions do not allow any hidden variable mimic .
let us now look at what the clauser - horne - shimony - holt ( chsh ) bound @xcite would be in our ` mutant ' qm .
the chsh bound is the upper bound of the absolute value of the following combination of correlators : @xmath99 where @xmath100 and @xmath101 are two observables of particle 1 , and @xmath102 and @xmath103 are two observables of particle 2 .
all four observables are assumed to take on only the values @xmath60 upon measurement . for classical hidden variable theory ,
the bound on @xmath104 is 2 , while for canonical qm it is @xmath105 @xcite . to calculate this bound for our model
, it suffices to examine all possible correlators for the singlet state @xmath78 only .
this is because all @xmath62 entangled states can be transformed into @xmath78 via local @xmath61 rotations , that is , @xmath61 transformations on only one of the entangled particles .
we can also restrict the observables entering the correlator to those in which the indices are in increasing order , _
i.e. _ @xmath106 with @xmath107 , since @xmath108 these considerations simplify our task considerably , and we find that the absolute value of the chsh correlator is maximized for the cases @xmath109 thus , the chsh bound for our ` mutant ' qm is 2 .
in conclusion , we have constructed a ` mutant ' qm based on a linear vector space over the galois field @xmath110 .
we find that though it is fully ` quantum ' in the sense that no hidden variable theory can reproduce its predictions , the chsh bound of its correlations nevertheless has the ` classical ' value of 2 .
thus , our model provides an existence proof that ` quantum'-ness does not necessarily require the violation of the chsh bound .
the state space of our ` mutant ' qm is the finite projective space @xmath111 , in close analogy to the @xmath112 of canonical qm .
we recall that this complex projective space can be understood as the coset @xmath113 this structure incorporates unitary evolution described by the @xmath114 factor , thereby preserving the normalization of state vectors , with generic berry phases described by the @xmath115 factor , which characterize possible degenerate states , and the quantum mechanical @xmath116 phases @xcite .
the corresponding coset structure of @xmath111 is : @xmath117 where , @xmath118 is the general linear group on @xmath13 , @xmath119 its center consisting of @xmath120 unit matrices multiplied by a ` phase ' in @xmath34 , and @xmath121 is the affine linear group which keeps the direction of a vector in @xmath13 invariant .
the projective linear group we encountered earlier is itself the coset group @xmath122 thus , our discrete qm possesses analogs of the geometric structure of canonical qm , with @xmath118 generating evolution over finite time steps , and @xmath121 characterizing possible degeneracies in dynamical systems .
the extent that elements of this affine group and the center @xmath119 determines any super - selection sectors has yet to be explored @xcite .
the @xmath123 case of our model , constructed on @xmath124 , would be particularly simple .
it could , perhaps , be the simplest quantum theory imaginable and provide a setting to explore the most basic questions concerning the foundations of qm , as well as a platform to develop ideas relevant to quantum information and quantum computing @xcite .
a question we addressed in a previous publication @xcite was whether a super - quantum theory whose chsh bound exceeds the cirelson value of @xmath105 of canonical qm exists @xcite .
such super - quantum models are expected to go ` beyond ' canonical qm in one way or another .
the model discussed in this paper was discovered in the process of looking for such a model , though , of course , its correlations are sub - quantum instead .
an interesting question to ask is whether a ` super ' version of our discrete model can be constructed in which the chsh bound exceeds 2 .
we conjectured in ref . that a super - quantum theory may exist in the @xmath125 limit of canonical qm .
recalling that @xmath126 is effectively the curvature of @xmath112 , this complex projective space would degenerate to @xmath127 in the @xmath125 limit . extrapolating this intuition to our discrete qm
, whose geometry is described by @xmath128 , the super - quantum limit would correspond to degenerating this projective space into @xmath129 .
thus , the construction of a ` super ' version of discrete qm may require working in this space .
another interesting avenue , already alluded to above , would be to use our discrete qm to explore the structure of the conjectured general geometric quantum theory @xcite .
such a structure was argued to be relevant for quantum gravity . in particular , in that context , it was argued that non - linear grassmannian spaces are the natural generalization of complex projective spaces @xcite .
it view of our proposal , it would be interesting to explore discrete analogs of non - linear grassmannians by replacing complex spaces with @xmath5-spaces .
finally , since we have succeeded in constructing a qm on a space without an inner product , it can not be an essential gene necessary for the survival of a ` quantum ' theory .
the absence of an inner product allows the vector and dual - vector spaces to be distinct , and our construction gives separate physical meanings to the two : the dual - vectors represent possible outcomes of a measurement , while the vectors represent the latent possibilities of a state .
this removal of the required existence of an inner product potentially allows for the construction of quantum theories on spaces that have not heretofore been considered , _
e.g. _ banach spaces . we will explore these , and other related questions in upcoming publications @xcite .
we would like to thank sir anthony leggett and prof .
chia tze for helpful discussions .
zl , dm , and tt are supported in part by the u.s .
department of energy , grant de - fg05 - 92er40677 , task a. a similar proposal was made by b. schumacher and m. d. westmoreland in arxiv:1010.2929 [ quant - ph ] . in their work , probabilities were not defined .
our model would correspond to assigning equal probabilities to all ` possible effects ' in their model . for a pioneering discussion of quantum field theory over galois fields consult : y. nambu , `` field theory of galois fields , '' in _ quantum field theory and quantum statistics _
, vol . 1 ,
625 - 636 , eds .
i. a. batalin et al .
( iop publishing , 1987 ) . | we construct a discrete quantum mechanics using a vector space over the galois field @xmath0 .
we find that the correlations in our model do not violate the clauser - horne - shimony - holt ( chsh ) version of bell s inequality , despite the fact that the predictions of this discrete quantum mechanics can not be reproduced with any hidden variable theory . |
the hidden sector gaugino condensation has long been suggested toward the minimal supersymmetric standard model ( mssm ) via the gravity mediation of supersymmetry ( susy ) breaking @xcite .
the gravity mediation of the gaugino condensation introduces the squark mass splitting at the gravitino mass scale which has been naively estimated to be of order @xmath4 , leading to the hidden sector scale @xmath5 at @xmath6 gev . including extra dimensions , the kaluza - klein modes work @xmath7pro and con " toward a gauge coupling unification below the open - up scale of the extra dimensions @xcite . in this regard
, it is an interesting attempt to try to obtain a raised hidden sector scale @xmath5 together with the gravitino mass more suppressed by @xmath8 @xcite . in ref .
@xcite the gravitino mass is shown to be highly suppressed , @xmath9 , but there a specific form for the superpotential has been assumed .
thus , in the presence of the hidden sector gaugino condensation , it is an important question to ask , @xmath7what is really the gravitino mass ? " , and the answer can be applied to the gravity mediation and also to the gauge mediation of susy breaking ( gmsb ) @xcite . in this paper , we attempt to analyze a proper global susy breaking , which is then applied to obtain a gravitino mass in terms of the hidden sector scale .
we will obtain a gravitino mass suppressed by @xmath10 , but a smaller gravitino mass results from the strong dynamics of the hidden sector .
the hidden sector dynamics necessarily needs an information how the dilaton @xmath11 is stabilized . for the dilaton stabilization
, we can assume a ` race - track ' model @xcite , but here we will attempt to obtain the dilaton stabilization in the effective lagrangian approach @xcite in the presence of a dynamically generated susy breaking source .
the flipped - su(5 ) gut was introduced as another path of so(10 ) branching @xcite , and recently it has been applied to two dark matter components @xcite . in some @xmath12 compactifications of the heterotic string , it is possible to obtain three families in the flipped - su(5 ) gauge group in @xmath12 compactifications @xcite .
it is known that dynamical breaking of susy is possible in some chiral gauge models @xcite : @xmath13 motivated by this observation , we study one hidden su(5)@xmath0 family , @xmath1@xmath0 and @xmath2@xmath0 plus possible @xmath14 numbers of @xmath15@xmath0 and @xmath2@xmath0 . with one @xmath1@xmath0 and one @xmath2@xmath0
, we can not construct a composite superfield of the form , @xmath16 .
however , there exist two composite superfields constructed with the gluino and matter fields @xcite , @xmath17 where @xmath18 is the hidden sector gluino superfield , satisfying @xmath19 .
there is no more su(5)@xmath0 invariant independent chiral combination .
the addition of vector - like representations does not change the fate of susy breaking of the single family case since the witten index is not changed @xcite .
the pure susy gauge models lead to the effective superpotential @xmath20 below the confinement scale from which one can always obtain a susy condition @xcite . on the other hand ,
one family su(5)@xmath0 is classified as an @xmath21un - calculable model @xcite , but it has been argued that it would break susy @xcite . 0.5 cm in terms of the composite chiral fields @xmath22 and @xmath23 , two susy conditions can not be satisfied simultaneously and hence susy is broken @xcite . the key instanton diagram which dictates how one loop effect can be written
is shown as a spider diagram in fig .
[ fig : hsu5inst ] . assuming that we know the scale of the spider diagram , we can open up some lines of fig .
[ fig : hsu5inst ] with two gravitinos as shown in fig .
[ fig : gravitinomass ] . closing the gluino lines of fig .
[ fig : gravitinomass ] , we can estimate the contribution to the gravitino mass by fig .
[ fig : gravitinomass ] of order @xmath24 where @xmath25 is a dimension 1 chiral field and @xmath26 is its decay constant . the expression ( [ eq : gravmassint ] )
has the @xmath27 and the other factors are calculable in principle if the hidden sector dynamics is known . in sec .
[ sec : global ] , we show how the hidden sector scale @xmath5 appears in determining the susy breaking minimum . in sec .
[ sec : supergravity ] , supergravity effects are included and the gravitino mass is estimated . in the final section , sec . [ sec : comments ] , we comment how our observation can be used to obtain reasonable mssm models from an ultra - violet completed theory
. 0.5 cm
now , let us proceed to consider the one family su(5)@xmath0 model , with @xmath1@xmath0 and @xmath2@xmath0 . then
, the following su(5)@xmath0 singlet vector fields can be considered , @xmath28 these vector fields do not contribute to an effective potential but can give rise to couplings of the form @xmath29\nonumber\ ] ] where @xmath30 and @xmath31 are real functions of the su(5)@xmath0 singlet composite fields .
since we lack a method to treat this general form , we restrict to the known nonperturbative effects guided by the instanton interaction .
then , we can consider two su(5)@xmath0 singlet composite chiral fields considered in eqs .
( [ zdef ] ) and ( [ zpdef ] ) . due to the index theorem
, the model with a @xmath32 , a @xmath33 , plus @xmath14 copies of @xmath34 and @xmath33 breaks susy .
the quantum numbers of the global symmetries are as shown in table [ tab : sufive ] @xcite .
we defined the charges such that the u(1)@xmath35 is anomaly free , and there is only one anomalous u(1 ) : the u(1)@xmath36 .
the u(1)@xmath36 is broken by the su(5)@xmath0 instantons .
below the su(5)@xmath0 confinement scale , we can consider only two effective chiral fields as suggested in @xcite .
in addition , we must consider the global symmetries of the flavors @xmath34 and @xmath33 : @xmath37 .
this flavor symmetry @xmath38 must be realized below the confinement scale @xmath5 as @xcite : c|cccc & @xmath39 & @xmath40 & @xmath41 & @xmath42 + + @xmath32 & @xmath43&@xmath44 & @xmath45 & @xmath46 + @xmath33 & @xmath44&@xmath47 & @xmath48 & @xmath44 + @xmath34@xmath49 & @xmath44&@xmath50 & @xmath48 & @xmath44 + @xmath33@xmath51 & @xmath44&@xmath47 & @xmath48 & @xmath44 + @xmath22 & 0&@xmath52&2 & @xmath52 + @xmath23 & @xmath53&@xmath52 & @xmath54 & @xmath55 + @xmath56 & @xmath57&@xmath52&@xmath58 & @xmath57 + @xmath59 & @xmath60&@xmath52&@xmath61 & @xmath52 + @xmath62 & @xmath60&@xmath52&0 & @xmath52 * if the flavor symmetry remains unbroken , there appear @xmath14 singlets for the fundamental of @xmath63 and @xmath64 singlets for the fundamental of @xmath65 .
* if all or a part of the flavor symmetry is broken , there appear the corresponding composite goldstone boson multiplets . for @xmath66 , we can not find the matching number for the composite singlets .
for @xmath67 , we can find the matching number for the goldstone bosons , i.e. in the phases of @xmath68 the number of goldstone bosons in ( [ zpgold ] ) is @xmath64 , and the number of goldstone bosons in ( [ mgold ] ) is @xmath69 .
thus , the total number of goldstone bosons for the case of ( @xmath70 ) is @xmath71 .
this is the number resulting if @xmath38 is broken down to @xmath63 : @xmath72\rightarrow n_f^2 + 2n_f$ ] .
so we consider the composite fields considered in ( [ zdef ] ) and ( [ zpdef ] ) with @xmath73 .
the instanton interaction is a determinant , i.e. the flavor group singlet , and we consider all these fields appearing in the diagram , like @xmath59 for @xmath74 .
then , the effective superpotential is @xmath75.\label{eq : sufw}\ ] ] since there appear only one combination in ( [ eq : sufw ] ) , we can redefine @xmath59 for the effective interaction as considered in table [ tab : sufive ] .
but certainly there are more fields if @xmath74 , for which the stabilization of these extra ( @xmath76 ) fields must be taken into account also , which is out of scope of this paper . toward the directions @xmath22 and @xmath59 , for the dynamically generated effective superpotential respecting the global symmetries , we use eq .
( [ eq : sufw ] ) @xcite where @xmath77 is considered as a coupling .
it was shown that for @xmath78 , the susy conditions can not be satisfied and susy is dynamically broken @xcite .
let us define the following @xmath79 and @xmath25 such that their engineering dimensions are 1 , @xmath80 where @xmath81 and @xmath26 are scale parameters .
in principle , the decay constants @xmath81 and @xmath26 are determined by the hidden sector dynamics .
these are expected to be near the scale @xmath5 .
the effective superpotential is @xmath82\nonumber\\ & = f^2\phi\left[\log\left(\frac{f^{4 - 2n_f } f^ { ' 6 + 2n_f } \phi^{2-n_f } \phi'}{\lambda^{13-n_f}e^{\alpha+1 } } \right ) + 1\right]\nonumber\end{aligned}\ ] ] let us rescale @xmath5 such that @xmath83 and
hence we can use the following superpotential without loss of generality , @xmath84.\label{eq : normnf}\end{aligned}\ ] ] consider the case of @xmath78 which leads to a simple analysis , @xmath85 \label{eq : norm}\ ] ] @xmath86\\\frac{\partial w_{\rm su(5)}}{\partial \phi ' } = f^2\frac{\phi}{\phi'}\label{eq : fterms } \end{array}\ ] ] the susy conditions from the upper and lower equations of ( [ eq : fterms ] ) are @xmath87 and @xmath88 , respectively , which are mutually inconsistent .
note that @xmath79 and @xmath25 possess 4 real fields which are denoted as @xmath89 and @xmath90 , @xmath91 we can write @xmath92 as @xmath93 where @xmath94 the minimum conditions @xmath95 and @xmath96 give the same condition @xmath97 0.5 cm for given @xmath81 and @xmath26 , one can find a solution as depicted in fig .
[ fig : solution ] . without susy
, we expect that the condensation of two gluinos is much more stronger than that of hidden sector quark and anti - quark pairs . for the chiral representations , we have less information on the condensations .
furthermore , susy can change this view of condensations of fermions , and therefore we treat @xmath81 and @xmath26 as unknown free parameters near but somewhat below the scale @xmath5 .
then , the rhs of eq .
( [ eq : rhopxicond ] ) is expected to be negative , and a solution for @xmath98 is a possibility .
let that number be for example o(10 ) . then , @xmath99 .
the h - gluino condensation scale is @xmath100 which in this case is of order @xmath101 which gives about two orders smaller gaugino condensation scale compared to @xmath5 .
for example , for @xmath102 and @xmath103 we have @xmath104 and find an approximate solution @xmath105 or @xmath106 .
so , the gaugino condensation scale is of order @xmath107 the power @xmath108 of @xmath109 is small , and hence its effect is minor compared to the effect of the power of @xmath26 .
so , the gaugino condensation scale can be exponentially smaller than the hidden sector scale @xmath5 for @xmath110 , which is traced to the property of the high engineering dimension of the operator @xmath23 .
if supergravity is considered , the above dsb leads to a nonvanishing gravitino mass .
the gravitino mass is given in the supergravity lagrangian as @xcite , @xmath111 in our case , the susy breaking parameter @xmath112 is the f - terms , @xmath113 where @xmath114 is the superpotential modified khler function , @xmath115 , @xmath116 is the gauge kinetic function , and @xmath117 is the derivative of @xmath116 with respect to the chiral field @xmath118 . if the gauge kinetic function is a constant , the f - terms of @xmath22 and @xmath23 give the gravitino mass . since the effective lagrangian is arising from fig . [
fig : hsu5inst ] , the gravitino mass can be shown as fig . [
fig : gravitinomass ] .
let us now include the dilaton superfield @xmath11 in the gauge kinetic function .
motivated by the heterotic string , let us consider the following khler potential , @xmath119 from which the potential is calculated as @xmath120\ ] ] where @xmath121 and @xmath122 .
considering the result of sec .
[ sec : global ] , let us take @xmath123+c\nonumber\\ & = w_0-\frac{s}{4m_p}z + c \label{eq : wsugra}\end{aligned}\ ] ] where @xmath124 is @xmath125 of eq .
( [ eq : sufw ] ) with @xmath77 modified by the dilaton coupling , and a constant @xmath126 is added as a free parameter resulting from the u(1)@xmath35 breaking gravitational interaction .
the term @xmath127 in @xmath128 has been considered previously @xcite , but its consequence on the gravitino mass from one family su(5)@xmath0 model has never been presented before . since we are looking for a solution at a large value of @xmath11 , i.e. in the perturbative region , we can just read off the original gaugino coupling below the confinement scale as @xmath129 .
then , we must consider the following khler potential dependence , @xmath130 \nonumber\\ & d_\phi w = \frac{\partial w_0}{\partial\phi } -\frac{f^2 s}{4}+(w_0+c-\frac{f^2 s\phi}{4})\phi^ * \\ & d_{\phi ' } w=\frac{\partial w}{\partial\phi'}+ \frac{\partial k}{\partial\phi'}w= \frac{\partial w_0}{\partial\phi'}+(w_0+c-\frac{f^2 s\phi}{4})\phi'^ * \nonumber\end{aligned}\ ] ] so that we obtain @xmath131 where we set @xmath132 for a moment .
thus , the potential is expressed as @xmath133 \nonumber\\ & + \left[(\frac{w_0^*}{m_p^2}-\frac{f^2s^*\phi^*}{4m_p^3 } ) \left(\frac{\partial w_0}{\partial \phi}\phi+\frac{\partial w_0}{\partial \phi'}\phi'+\frac{1}{4m_p } f^2(s+s^*)\phi\right)+{\rm h.c.}\right]\big\ } \label{eq : vnoc}\end{aligned}\ ] ] where we omitted the terms depending on @xmath126 of eq .
( [ eq : wwithc ] ) .
this @xmath126 will be used only for adjusting the cosmological constant to zero , and hence for simplicity we will neglect its dependence of the dilaton stabilization .
let us choose the real fields as @xmath134 where the following @xmath135 ranges of the angles are assigned , @xmath136,\quad\delta=(-\pi,\pi],\quad \theta_{mi}=(-\pi,\pi].\ ] ] since the principal value of @xmath90 appears as polynomials in our expression , we choose our convenient range of @xmath90 ending at 0 .
then , we obtain the following potential @xmath137 -\frac{f^4\rho^2\sigma_{cr}}{2m_p^3 } + \frac{f^4\rho^2\sigma^2}{8m_p^4}\nonumber\\ & + \frac{2f^4\rho^2}{m_p^2}\big\ { ( -\log\xi+\log\epsilon+1-\frac{\sigma_{cr}}{4m_p } ) ( -\log\xi+\log\epsilon+1+\frac{\sigma_{cr}}{2m_p } ) \nonumber\\ & + \theta(\theta-\frac{\sigma\sin\theta_{mi}}{4m_p } ) \big\}\end{aligned}\ ] ] where @xmath92 is given in eq .
( [ eq : v ] ) . in applying the above equation ,
we take the positive parameters except the angles . representing @xmath138 in terms of dimensionless fields , @xmath139 and @xmath140
, we obtain @xmath141 + \frac{\tilde\sigma^2}{16}\nonumber\\ & \quad\quad\quad\quad\quad+h^2 \big[\tilde\sigma\cos\theta_{mi } ( -\log\xi+\log\epsilon+1)+ \frac{\tilde\sigma^2}{4 } \cos^2\theta_{mi}\big]\label{eq : finv}\\ & + { h^2}\left({h^2}+ h^{'2}\right)\cdot \big[(-\log\xi+\log\epsilon+1)^2 -\frac{\tilde\sigma}{2}\cos\theta_{mi}(-\log\xi+\log\epsilon+1)\nonumber\\ & \quad\quad\quad\quad\quad + \frac{\tilde\sigma^2}{16 } -\frac{\tilde\sigma}{2 } \theta \sin\theta_{mi}+ \theta^2\big]\big\}.\nonumber\end{aligned}\ ] ] note that in the limit of @xmath142 , i.e. @xmath143 , the dilaton is stabilized at @xmath144 in fig .
[ fig : dilastab ] we show the dilaton stabilization for @xmath145 .
at this dilaton stabilized point , the potential is @xmath146 $ ] .
so far we neglected @xmath126 of eq .
( [ eq : wsugra ] ) , and @xmath138 is tuned to 0 by an appropriate choice of @xmath126 . 0.5 cm in the limit @xmath142
, we obtain @xmath147 for a small @xmath148 .
then , the f - terms of @xmath149 and @xmath25 fields at the minimum of the supergravity potential are , viz
( [ eq : fterms ] ) , @xmath150 so that the gravitino mass from eq . ( [ eq : gravmass ] ) becomes @xmath151 where we used the limit @xmath152 and @xmath153 in the second line . toward a suppressed gravitino mass ,
we need both the ratio @xmath154 being sufficiently close to 1 and @xmath155 .
then , the gravitino mass is accordingly suppressed . from ( [ eq : finv ] ) , in this limit the phase stabilization is determined at @xmath156 and @xmath157 . at this vacuum ,
( [ eq : tunedgrm ] ) can give a reduced gravitino mass for @xmath158 .
thus , in principle , the gravitino mass is determined by the effective lagrangian approach , with which the dilaton is stabilized .
but due to our ignorance on the parameters in the confining sector , @xmath159 , and @xmath160 , we can not determine the gravitino mass exactly in the effective lagrangian approach . because of the difficulty in estimating the order of the absolute magnitude , one may resort to the original dynamical source due to the instanton diagram .
[ fig : hsu5inst ] has two more loops ( the hidden sector gluino quark loops ) compared to fig .
[ fig : gravitinomass ] ( the @xmath73 case ) . a naive estimation of these two loops would be @xmath161 times the hidden sector coupling times the relevant mass scale of the hidden sector . since fig .
[ fig : hsu5inst ] has 11 loops , the momenta going around one loop is averaged to @xmath162 .
so , we roughly estimate the missing two loops contribute @xmath163 where @xmath164 can be taken as 4 , the hidden sector coupling of o(1 ) and we use 10 instead of 11 ( considering fig .
[ fig : hsu5inst ] ) or 9 ( considering fig . [ fig : gravitinomass ] ) .
if there are @xmath14 pairs of @xmath165 and @xmath166 , we may divide @xmath5 by @xmath167 .
therefore , we estimate the diagram shown in fig . [
fig : gravitinomass ] as @xmath168 from which we estimate for @xmath169 @xmath170 for example , taking @xmath171 due to small @xmath172 , @xmath173 , and @xmath174 gev , eq .
( [ eq : massgravti ] ) gives @xmath175 gev for @xmath78 .
however , we have observed that the gravitino mass has one inverse power of @xmath8 and hence , correcting the above number by multiplying @xmath176 , we obtain @xmath177 gev .
therefore , hidden sector dynamics may lead to a very close @xmath178 and @xmath148 in eq .
( [ eq : tunedgrm ] ) .
in this paper , we estimated the gravitino mass in one hidden family su(5)@xmath0 models in terms of the hidden sector scale @xmath5 in the vacuum where the dilaton is stabilized and showed that it is possible to reduce the gravitino mass than the previous naive estimate .
so , starting with the hidden sector coupling much above @xmath179 gev , one can obtain a sufficiently small gravitino mass or the sub - tev mass splittings in the visible sector superfields by gravity mediation .
then , a tev order susy scale can be in principle calculated in the gmsb or / and anomaly mediation .
the reduced gravity mediation effect can be included if it is non - negligible . for an ultra - violet completion of this kind of dsb with one hidden family su(5)@xmath0 group , compactifications in @xmath12 orbifolds in the heterotic string and f - theory
are suitable .
for example , we find in @xmath12 orbifold compactifications that there frequently appear hidden sector one family su(5)@xmath0 groups . among these ,
choosing three visible sector families , the number of such models is drastically reduced . requiring other phenomenological constraints ,
this number is further reduced .
we presented two such models before @xcite .
if one starts from a universal gauge coupling at the gut unification point ( @xmath180 gev ) of the visible sector couplings , both of these models predict the hidden sector scale above @xmath3 gev , naively predicting the gravitino mass much above the tev scale .
of course , contributions from the kaluza - klein modes between the string scale and the gut scale may change @xcite this undesirable feature , but the models presented in @xcite and probably most models listed in @xcite work in the opposite direction because the @xmath181 function of the hidden sector gauge group is smaller than that of the visible sector gauge group .
the mechanism we discussed in this paper can remedy this dilemma .
namely , starting with a universal gauge coupling at the gut unification point even with a large hidden sector scale , presumably near @xmath182 gev , one can achieve a sufficiently small gravitino mass so that a tev order visible sector susy scale may result from the other sources such as from the gmsb @xcite . in this appendix , we present the mixing between the spin @xmath43 component and the spin @xmath183 component by the dsb .
we note that fig .
[ fig : gravitinomass ] is the essential one leading to the gravitino mass , including both the source of dsb of fig .
[ fig : hsu5inst ] and the u(1)@xmath35 invariance : the dsb source , the green and red blobs , give the u(1)@xmath35 invariance and the u(1)@xmath35 charges of two gravitino lines , and two gluino lines add up to zero .
[ fig : gravitinomass ] is the basis for calculating the gravitino mass , by relating the spin-@xmath43 component to the goldstino component spin-@xmath183 .
it is in parallel to the gauge boson mass by coupling the spin-1 component @xmath184 to the goldstone boson component spin-0 @xmath185 . in the flat space
, we note that for a particle with a nonzero helicity , the mass term changes the helicity @xmath186 because the massive particle can not move faster than the speed of light and one can lorentz - boost such that the helicity @xmath186 looks changed .
so , the coupling of two helicity states with @xmath187 is the mass of the nonzero helicity particle . for spontaneous breaking of u(1 ) gauge symmetry
, one starts with a coupling @xmath188 from @xmath189 where @xmath190 part is shown as a gauge invariant form under @xmath191 and @xmath192 . if @xmath193 , the u(1 ) gauge boson can not couple to a longitudinal component
. however , if the gauge symmetry is broken by a vacuum expectation value of @xmath194 , one can take a unitary gauge to represent @xmath194 as a goldstone boson dependent function @xmath195 .
then , the quantity inside the bracket @xmath196 can be gauge transformed to @xmath191 and @xmath192 .
the @xmath197 dependent function @xmath198 is identified as the dimensionless gauge function @xmath199 , and one obtains the @xmath184 ( transverse component ) to @xmath200 ( longitudinal component = goldstone boson ) coupling @xmath201 , and hence the gauge boson mass is @xmath202 .
we can view fig .
[ fig : gravitinomass ] as the susy invariant coupling respecting the global symmetries , and close the gluino lines to introduce another green blob .
namely , we integrate out strongly interacting fields and consider only light fields @xmath22 and @xmath203 , and hence only the second term of eq .
( [ eq : gravsusybr ] ) is considered . instead of fig .
[ fig : gravitinomass ] , along the above paragraph we try to obtain the @xmath204 ( transverse component ) to @xmath205 ( goldstino ) coupling to obtain the gravitino mass where @xmath206 is defined to carry dimension @xmath183 .
namely , @xmath207 where @xmath208 is the goldstino field .
in the zero vacuum energy , let us consider the magnetic moment type gravitino coupling to chiral fields to @xmath128 @xcite and the goldstino ( @xmath208 ) coupling @xmath209 where we can use ( [ eq : normnf ] ) for @xmath128 , @xmath114 is the goldstino superfield , and @xmath210 is an auxiliary field splitting scalar partner of @xmath114 from the massless @xmath208 .
so , we can assign the auxiliary field @xmath210 as @xmath211 so that the goldstino @xmath208 remains massless in the broken susy case .
@xmath208 has dimension 3/2 .
the first term of ( [ eq : gravgold ] ) is an interaction term and is not a gravitino mass term yet .
the susy transformations of ( [ eq : gravgold ] ) relate spin-@xmath43 components to spin-@xmath183 components .
firstly , the susy transformation of @xmath212 to @xmath213 generates @xmath214 because three components in @xmath212 in the @xmath215 gauge goes to one component in @xmath216 . note that @xmath216 has dimension 3/2 and the suffix @xmath183 denotes a two - component spinor
therefore , due to the goldstino definition as the f - term breaking of susy as given in eq .
( [ eq : fauxill ] ) , we interpret @xmath217 .
dimension @xmath218 goldstino field @xmath206 is the goldstino direction @xmath219 , taking into account the gravitational charge @xmath220 and the susy breaking scale @xmath221 .
dimension @xmath222 goldstino field @xmath223 is required to have the same dimension as the grassmann variable @xmath224 and hence @xmath225 .
the downward ( twice ) susy transformations of @xmath212 will lead to @xmath226 .
the upward ( twice ) susy transformations of a chiral field in @xmath128 will lead to the f - term of that field , i.e. @xmath227 for example .
thus , the upward susy transformation of @xmath228 will be @xmath229\nonumber\\ = \frac{f^2}{m_p^2}&=\frac{f^4}{m_p^2}[(-\log\xi+\log\epsilon)^2+\xi^2 ] \end{aligned}\ ] ] where we used eqs .
( [ eq : fterms ] ) . therefore , from downward and upward susy transformations inside the bracket of @xmath230
, we obtain @xmath231 , @xmath232 the susy breaking scale is @xmath233^{1/2}$ ] .
therefore , the coefficient of @xmath234 is @xmath235^{1/2}\label{eq : gravmsusytrans}\end{aligned}\ ] ] where we labeled the mass as ` mixing ' since we calculated it from the mixing of spin-@xmath43 and spin-@xmath183 components .
we can see that eq .
( [ eq : gravmsusytrans ] ) is basically the same as @xmath236 of eq .
( [ eq : tunedgrm ] ) .
note that the vacuum energy is given by @xmath237 where @xmath238 , and @xmath239 for the ds , flat and ads spaces , respectively .
namely , @xmath240 and the gravitino mass in the curved space has the same form in terms of @xmath221 , i.e. @xmath241 does not cancel @xmath242 exactly in the curved space .
# 1#2#3phys . rep .
* # 1 * ( # 3 ) # 2 # 1#2#3rev .
* # 1 * ( # 3 ) # 2 # 1#2#3annu .
nucl . part .
* # 1 * ( # 3 ) # 2 # 1#2#3nucl .
* b#1 * ( # 3 ) # 2 # 1#2#3phys . lett .
* b#1 * ( # 3 ) # 2 # 1#2#3phys .
* d#1 * , # 2 ( # 3 ) # 1#2#3phys .
# 1 * ( # 3 ) # 2 # 1#2#3jhep * # 1 * ( # 3 ) # 2 # 1#2#3jcap * # 1 * ( # 3 ) # 2 # 1#2#3z .
phys . * # 1 * ( # 3 ) # 2 # 1#2#3euro .
j. * # 1 * ( # 3 ) # 2 # 1#2#3int .
j. mod .
phys . * # 1 * ( # 3 ) # 2 # 1#2#3mod .
# 1 * ( # 3 ) # 2 # 1#2#3astrophys .
j. * # 1 * ( # 3 ) # 2 # 1#2#3nature * # 1 * ( # 3 ) # 2 # 1#2#3sov .
j. nucl .
# 1 * ( # 3 ) # 2 # 1#2#3astrophys .
j. * # 1 * ( # 3 ) # 2 # 1#2#3int .
j. mod .
phys . * # 1 * ( # 3 ) # 2 # 1#2#3astropart .
phys . * b#1 * , # 2 ( # 3 ) s. m. barr , ; j .-
derendinger , j. e. kim and d. v. nanopoulos , ; i. antoniadis , j. r. ellis , j. s. hagelin , and d. v. nanopoulos , ; ; k. dienes and a. faraggi , ; a. e. faraggi , r. s. garavuso and j. m. isidro , . | the hidden su(5)@xmath0 with one family , @xmath1@xmath0 and @xmath2@xmath0 , breaks supersymmetry dynamically . from the effective lagrangian approach
, we estimate the hidden sector gaugino candensation scale , the dilaton stabilization and the resulting gravitino mass . in some models
, this gravitino mass can be smaller than the previous naive estimate .
then , it is possible to raise the su(5)@xmath0 confining scale above @xmath3 gev . |
recent observations of a hourglass shaped excitation spectrum in cuprate superconductors point towards an underlying universal magnetic excitation spectrum for all of the cuprates@xcite . in insulating charge stripe ordered la@xmath7sr@xmath8coo@xmath4 an hourglass magnetic excitation spectrum has been observed , and explained within a charge stripe model with a weak inter - stripe spin interaction , relative to the intra - stripe spin interaction@xcite .
the magnetic excitations of la@xmath7sr@xmath8coo@xmath4 indicate that the underlying magnetism of the cuprates can originate from charge stripe fluctuations .
furthermore , the lack of an hour glass excitation spectrum in charge stripe ordered la@xmath2sr@xmath3nio@xmath4 is explained by the strong inter - stripe spin interaction of these materials .
the dispersion of the magnetic excitations of charge stripe ordered la@xmath0nio@xmath1 deviates from the theory that describes the magnetic excitations in la@xmath7sr@xmath8coo@xmath4 and la@xmath2sr@xmath3nio@xmath4 @xmath9 @xcite . in a limited energy range , @xmath1010 - 30mev , the centre of the magnetic excitations shifts inwards towards the two dimensional antiferromagnetic wavevector @xmath11 .
if the inter - stripe spin interaction was weak , the excitations would be expected to disperse towards @xmath12 , but the inter - stripe spin interaction of la@xmath0nio@xmath1 is strong@xcite . as the inter - stripe spin interaction
is mediated by the charge stripe electrons , it is important to ascertain what are the magnetic properties of the charge stripe electrons of la@xmath0nio@xmath1 . in la@xmath2sr@xmath3nio@xmath4
@xmath13 we have observed the charge stripe electrons have gapped quasi - one - dimensional antiferromagnetic spin correlations ( q-1d ) along the charge stripes@xcite .
a single crystals of la@xmath0nio@xmath1 ( @xmath14 ) was grown by the floating - zone technique@xcite .
the sample used here was also used in our previous neutron studies of @xmath14 @xcite , and is a rod of 8 mm diameter and 40 mm length .
thermo - gravimetric analysis ( tga ) of an as - grown crystal determined the oxygen excess to be @xmath15 .
data on the bulk magnetization of an as - grown crystal are published elsewhere@xcite .
neutron scattering measurements were performed on the triple - axis spectrometers in20 at the ill , and on puma at frm - ii .
the energies of the incident and scattered neutrons were selected by bragg reflection from crystal heusler arrays ( in20 ) , or arrays of pyrolytic graphite crystals ( puma ) .
the monochromators were horizontally focused on both instruments , and vertically focused on puma , to maximise neutron flux on the sample position .
the analyzers were horizontally focused on both instruments and vertically focused on puma .
data were collected with a fixed final neutron wavevector of 2.662@xmath16 .
a pyrolytic graphite filter was placed between the sample and analyzer to suppress higher - order harmonic scattering . on in20
polarized neutrons were employed , and the neutron spin polarization @xmath17 was maintained in a specified orientation with respect to the neutron scattering vector @xmath18 by an adjustable guide field of a few mt at the sample position .
a monitor is placed between the monochromator and sample position to determine the number of neutrons incident on the sample position .
we aligned the crystal so that the horizontal scattering plane was @xmath19 ( we refer here to the tetragonal unit cell of the space group @xmath20 with unit cell parameters @xmath21 and @xmath22 ) .
on in20 the sample was mounted in a standard orange cryostat , and on puma the sample was mounted in a closed cycle refrigerator .
we performed longitudinal polarized neutron scattering measurements on in20 . for these measurements we constrained the neutron polarization @xmath17 to be parallel to the scattering vector @xmath18 . in this
configuration scattering from electronic magnetic moments causes the neutron spin to flip , whereas scattering via non - magnetic processes does not . in figure
[ fig : elastic ] we show an elastic scan of the @xmath14 parallel to @xmath23 through the the magnetic bragg reflections at @xmath24 . in the spin flip ( sf ) data two sharp bragg reflections are observed at @xmath25 r.l.u . and @xmath26 r.l.u .
from the magnetic order . in the non - spin flip channel ( nsf ) weaker bragg reflections are observed due to the imperfect polarization of the neutron beam .
the ratio of the intensities of the nsf and sf bragg reflections gives a flipping ratio of 19 . in figure
[ fig : polinelastic](a ) we show scans along @xmath23 at @xmath27 mev of @xmath14 , for both the sf and nsf channels .
the data has been corrected for the imperfact neutron polariztion@xcite . in the nsf channel
no excitation peaks are observed , whereas in the sf channel a broad magnetic excitation peak centred on @xmath28 r.l.u . is observed . as the dispersion gradient of excitations from the ordered ni@xmath29 @xmath30 ordered spins is effectively infinite at 2.5mev
, they only account for the left hand side of the sf peak .
the right hand side of the excitation centred on @xmath31 is an unaccounted for antiferromagnetic excitation , that has no elastic component . at 2.5mev
we are below the out of plane anisotropy gap in @xmath14 , and the spins fluctuate around the ordered moment in the ni - o plane@xcite . for a wavevector along @xmath23 at 2.5mev
only the component of the spin fluctutaions perpendicular to this direction are probed .
if the spin excitations are observed at a wavector of the same length that is appproximately along @xmath32 , all the in plane spin fluctuations are probed and the intensity of the observed excitation will increase .
the q-1d in the @xmath5 was determined to preferentially fluctuate out of the ni - o plane@xcite .
if the q-1d is measured at a wavector appproximately along @xmath32 the intensity will decrease compared to a wavevector along @xmath33 .
figure [ fig : polinelastic ] ( b ) shows a polarized neutron scan of the magnetic excitations along @xmath34 which shows two peaks in the sf channel . in fig.[fig : polinelastic ] ( a ) and ( b ) the symmetry of scans reverses , so that the right hand side peak in ( b ) is from the ordered ni@xmath29 spins , which has increased in intensity as expected .
if the sf peak on the left of fig.[fig : polinelastic ] ( b ) was from q-1d excitations that preferential fluctuate out of the ni - o plane , the intensity of these excitations will be weaker in fig.[fig : polinelastic ] ( b ) than fig.[fig : polinelastic ] ( a ) , as observed . the polarized neutron scattering measurements of @xmath14 at 2.5mev show two magnetic excitations .
we have shown that one of these excitation modes is gapped with no elastic component , and has the same wavevector dependence for intensity as the q-1d in @xmath5 and @xmath35 .
after confirming the existence of a second magnetic excitation in @xmath14 , we studied the energy and temperature dependence of this magnetic excitation using unpolarized neutron scattering .
figure [ fig : eandt ] ( a ) we show three scans parallel to @xmath33 at @xmath36 of the @xmath14 at 4.2k . as we were able to collect significant neutron counts in the data , we clearly observe the lorentzian lineshape of the excitations , consistent with our previous results@xcite . at 2mev and @xmath37 a small side peak is clearly resolved from the excitations from the order moments .
the centring of this excitation agrees with the flexing of the q-1d observed in @xmath13@xcite . at 4mev
the excitations at @xmath38 and @xmath39 are observed to have reduced counts compared to 2mev , and by 6mev the excitations have broadened to become an extended tail to the excitation peaks from the ordered moments . when the q-1d of @xmath5 is scanned in an equivalent way to fig .
[ fig : eandt ] ( a ) , the q-1d is qualitatively observed to follow the same trend . unfortunately the fitting uncertainties for the data fits of the 4mev and 6mev data in fig .
[ fig : eandt ] ( a ) do not allow for a meaningful quantative comparison . in figure
[ fig : eandt ] ( b ) we show a scan parallel to @xmath33 at @xmath36 for 4mev of the @xmath14 , taken at 20k .
the scan at 20k is well described by a fit of two lorentzians on a sloping background , as indicated by the solid line of fig .
[ fig : eandt ] ( b ) . in the @xmath5 , on increasing temperature the q=1d remains observable to temperatures well in excess of 20k@xcite , unlike the scan at 20k of the @xmath14 .
the lack of observable q-1d at 20k for @xmath14 may be a result of the fit , with the sloping background and the lorentzian lineshape incorrectly accounting for the q-1d .
additional measurements in the @xmath40 scattering plane away from the excitations from the order ni@xmath29 spins , should resolve this issue .
we have presented experimental evidence of a second magnetic excitation mode in charge stripe ordered la@xmath0nio@xmath1 . a preferential polarization of the fluctuations out of the ni - o plane , and the energy dependence of the magnetic excitations is consistent with the gapped quasi - one - dimensional antiferromagnetic excitations of the charge stripe electrons ( q-1d ) in la@xmath2sr@xmath3nio@xmath4 @xmath13 .
the main components of the ground state magnetic excitations in la@xmath0nio@xmath1 , the out of plane anistropy gap , the inter- and intra - stripe interactions , and the q-1d observed in this work , are unchanged compared to la@xmath2sr@xmath3nio@xmath4 @xmath13@xcite . despite the lack of differences in the magnetic interactions of la@xmath0nio@xmath1 and la@xmath2sr@xmath3nio@xmath4 @xmath13 , the ground state magnetic excitations from the order ni@xmath29 spins in the energy range 10 - 30mev
have remarkable differences . in la@xmath0nio@xmath1
there is an inward dispersion of the magnetic excitations towards @xmath11@xcite , whereas in la@xmath2sr@xmath3nio@xmath4 @xmath13 there is a partial gapping of the spin excitations accompanied by a resonance like feature at the magnetic zone centre@xcite .
as the main magnetic interactions in these two materials are unaltered , we require further interactions to fully explain the magnetic excitation spectrums in the charge stripe ordered nickelates . as our previous studies for @xmath13 have indicated , we need to research beyond the magnetic interactions to understand these striking phenomena @xcite .
this research project has been supported by the european commission under the 7th framework programme through the research infrastructures action of the capacities programme , contract no : cp - csa_infra-2008 - 1.1.1 number 226507-nmi3 .
crystal growth was supported by the engineering and physical sciences research council of great britain .
10 hayden s. m. , mook h. a. , dai pengcheng , perring t. g. and doan f. , 2004 nature ( london ) * 429 * , 531 ; tranquada j. m. , woo h. , perring t. g. , goka h. , gu g. d. , xu g. , fujita m. and yamada k. , 2004 nature ( london ) * 429 * , 534 . a flipping ratio of f = 15 was used to correct the data , realistically reduced from the value of 19 observed in diffraction . correcting the data using @xmath41 results in insignificant differences to the data of fig .
[ fig : polinelastic ] . | the magnetic excitations of charge - stripe ordered la@xmath0nio@xmath1 where investigated using polarized- and unpolarized - neutron scattering to determine the magnetic excitations of the charge stripe electrons .
we observed a magnetic excitation mode consistent with the gapped quasi - one - dimensional antiferromagnetic correlations of the charge stripe electrons previously observed in la@xmath2sr@xmath3nio@xmath4 @xmath5 and @xmath6 . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.