article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
systematic deep photometric observations of the surroundings of the andromeda galaxy have revolutionized our knowledge of its satellite system . before 2004 , 12 dwarf galaxies were known to be m31 companions , including only 6 dwarf spheroidal galaxies , while 12 new dwarf galaxies have been discovered to inhabit this region of the local group sky over the last five years . a dedicated scan of the sloan digital sky survey ( sdss ) , targeting a region within @xmath4 of m31 s major axis , unveiled the presence of andromeda ix and x @xcite . in parallel , a contiguous mapping of the region within @xmath5 of andromeda , performed with the wide field camera on the isaac newton telescope , revealed the picture of a strikingly substructured inner stellar halo @xcite , and led to the discovery of andromeda xvii @xcite . mapping the outer regions of m31 s
halo was the goal of an extension to this survey .
performed with the megaprime / megacam @xmath6 camera on the canada - france - hawaii telescope ( cfht ) , it mapped a quarter of the m31 stellar halo , out to @xmath7 @xcite . confirming the clumpy nature of andromeda s stellar halo , the survey has also revealed the presence of numerous other dwarf galaxies : andromeda xi , xii , xiii , xv , xvi , xviii , xix and xx @xcite . at about the same time
, andromeda xiv was discovered serendipitously by @xcite , just outside of the edge of the megacam survey .
our understanding of these systems remains sparse , mainly due to the distance at which they reside , translating into difficult photometric observations that can not easily reach much deeper than the horizontal branch at the distance of m31 .
spectroscopic observations are also limited to the handful of bright red giant branch ( rgb ) stars that can be targeted in each system .
although progress is expected along these lines in the coming years , much can already be said of the generic properties of these new objects from the current survey data alone .
their absolute magnitude ranges from a very faint @xmath8 for and xii and and xx @xcite to a surprisingly bright @xmath9 for and xviii @xcite , patently showing the incompleteness of the m31 satellite luminosity function at the faint end , in regions that have so far only been surveyed with photographic plates .
given their relative faintness and significant sizes , these new systems are usually assumed to be dwarf spheroidal galaxies , in other words dwarf galaxies that are devoid of any significant amount of gas .
this is currently consistent with the analysis of hi surveys but the upper limits on their hi content remains relatively high ( @xmath10 ; @xcite ) .
therefore , it can not be entirely ruled out that some of these galaxies still contain non - negligeable amounts of gas .
building upon the previous megacam survey , we have initiated the pan - andromeda archaeological survey ( pandas ) , a large program using the cfht megacam imager to map the entire stellar halo of m31 and m33 out to distances of @xmath11 and @xmath5 respectively . here
, we report on the discovery of two new dwarf galaxies in the vicinity of the andromeda and triangulum galaxies based on the first year of pandas data .
these systems , dubbed andromeda xxi and xxii are sparse but unmistakable overdensities of stars that are also aligned along a rgb at the distance of m31 .
section 2 of this paper briefly summarizes the pandas data while section 3 presents the new systems and details their properties .
section 4 briefly discusses the new discoveries and section 5 concludes the paper . throughout this paper , the distance moduli of m31 and m33
are assumed to be @xmath12 and @xmath13 , or @xmath14 and @xmath15 respectively @xcite .
pandas builds upon our previous cfht / megacam surveys of m31 whose results are presented in @xcite , @xcite and @xcite .
we use a similar observational set - up and refer the reader to these papers for a description of the observing strategy , data reduction and data quality .
maps showing early results and the complete survey area to date can be found in @xcite and led to the discovery of and xxi and and xxii .
the survey consists of contiguous exposures , performed with the @xmath16 deg@xmath17 megaprime / megacam camera mounted on the cfht .
the camera is a mosaic of 36 @xmath18 ccds with a pixel size of 0.187arsec .
small gaps within the survey lead to a scientifically useable field - of - view of @xmath19 for each of the current 250pointings of the survey .
each field has been observed for 1350s in each of the megacam @xmath20 and @xmath21 filters , split into @xmath22s dithered subexposures .
good seeing ( @xmath23 ) ensures that the photometry reaches @xmath24 and @xmath25 with a signal - to - noise ratio of 10 and guarantees that the star / galaxy separation only degrades at magnitudes fainter than @xmath26 and @xmath27 .
data were preprocessed ( de - biased , flat - fielded and fringe corrected ) by the elixir system at cfht that also determines the photometric zero point of the observations , and then processed using a version of the cambridge astronomical survey unit ( casu ) photometry pipeline @xcite , adapted to cfht / megacam observations .
the pipeline registers and stacks the images , and also generates catalogues with object morphological classification before creating band - merged @xmath20 , @xmath21 products . in the following ,
dereddened magnitudes ( @xmath28 and @xmath29 ) have been determined from the @xcite @xmath30 extinction maps , using the following correction coefficients : @xmath31 and @xmath32 , listed in their table 6 .
two new dwarf galaxies were directly identified on matched - filter surface density maps of rgb candidate stars selected to have colors consistent with the average ( generally ) metal - poor locus of m31 dwarf spheroidals .
satellites in the very sparsely populated outer halo regions correspond to stellar overdensities that are striking enough for them to be easily spotted , even down to magnitudes of @xmath33 .
this was the case for and xii @xcite and and xx @xcite and also here with andromeda xxii .
an automated search is certainly warranted to better quantify the detection limits of the survey in uncovering new dwarf galaxies .
however , it is beyond the scope of this initial discovery paper and will be presented elsewhere .
map of dwarf galaxies located in the surroundings of m31 and m33 .
m31 is located at the center of the coordinate system and known satellites are represented by empty circles .
the two new discoveries , and xxi and and xxii , are represented by filled circles . the current extent of the pandas survey
is shown as dotted squares , with each square representing a single @xmath34 megacam field .
the ellipse represents the extent of m31 s hi disk .
constellation boundaries , taken from @xcite , are shown as thin lines . ]
the locations of the two new satellites are shown on the map of the pandas survey ( figure [ map ] ) . as with many previous detections ,
these two systems are noticeably close to the edge of the survey , suggesting that the survey limit of @xmath35 in projected distance from m31 does not represent the true extent of andromeda s satellite system ( see also figure 2 of @xcite ) . in the spirit of previous conventions ,
we dub these two satellites andromeda xxi and andromeda xxii ( and xxi and and xxii ; see the discussion below , in appendix [ naming_conventions ] ) .
both systems appear as overdensities of stars on the sky , as is visible in the left panels of figures [ map_andxxi ] and [ map_andxxii ] .
these stars are also aligned along a rgb that would be at , or close to , the distance of m31 or m33 .
the cmds within 2 half - light radii of the dwarfs ( determined in [ strparam ] below ) are shown in the middle panels of these figures and , when compared to the cmd of reference fields chosen in an annulus covering the same area at a distance of @xmath36 from the dwarfs centers ( right panels ) , indeed reveal an alignment of stars that follow the typical shape of a rgb . isolating these stars enhances the contrast of the overdensity of stars on the sky ( large symbols in the left panels ) . and
xxi is typical of the relatively bright dwarf galaxies that we have found before ( such as and xv or and xvi ) .
the overdensity of stars in the cmd region @xmath37 and @xmath38 corresponds to its horizontal branch but the degrading star / galaxy separation at these magnitudes , as well as the rapidly dropping completeness , makes it difficult to extract any accurate information from this region of the cmd . with only @xmath39stars visible along its rgb , and
xxii is one of the faintest m31 satellites found so far . , e. , dunkley , j. , nolta , m. r. , bennett , c. l. , gold , b. , hinshaw , g. , jarosik , n. , larson , d. , limon , m. , page , l. , spergel , d. n. , halpern , m. , hill , r. s. , kogut , a. , meyer , s. s. , tucker , g. s. , weiland , j. l. , wollack , e. , & wright , e. l. 2009 , , 180 , 330 , a. w. , irwin , m. j. , ibata , r. a. , dubinski , j. , widrow , l. m. , ct , p. , dotter , a. , navarro , j. f. , ferguson , a. m. n. , h. , p. t. , lewis , g. f. , babul , a. , barmby , p. , bienaym , o. , chapman , s. c. , cockcroft , r. , collins , m. l. m. , fardal , m. a. , harris , w. e. , huxor , a. , mackey , a. d. , pearrubia , j. , rich , r. m. , richer , h. , siebert , a. , tanvir , n. r. , valls - gabaud , d. , & venn , k. a. 2009 , , in press , h. r. , merrifield , m. r. , douglas , n. g. , kuijken , k. , romanowsky , a. j. , napolitano , n. r. , arnaboldi , m. , capaccioli , m. , freeman , k. c. , gerhard , o. , coccato , l. , carter , d. , evans , n. w. , wilkinson , m. i. , halliday , c. , & bridges , t. j. 2006 , , 369 , 120 , d. b. , kniazev , a. y. , bell , e. f. , martnez - delgado , d. , grebel , e. k. , rix , h .- w . , rockosi , c. m. , holtzman , j. a. , walterbos , r. a. m. , annis , j. , york , d. g. , ivezi , .
, brinkmann , j. , brewington , h. , harvanek , m. , hennessy , g. , kleinman , s. j. , krzesinski , j. , long , d. , newman , p. r. , nitta , a. , & snedden , s. a. 2004 , , 612 , l121 , d. b. , kniazev , a. y. , martnez - delgado , d. , bell , e. f. , rix , h .- w . , grebel , e. k. , holtzman , j. a. , walterbos , r. a. m. , rockosi , c. m. , york , d. g. , barentine , j. c. , brewington , h. , brinkmann , j. , harvanek , m. , kleinman , s. j. , krzesinski , j. , long , d. , neilsen , jr .
, e. h. , nitta , a. , & snedden , s. a. 2007 , , 659 , l21 | we present the discovery of two new dwarf galaxies , andromeda xxi and andromeda xxii , located in the surroundings of the andromeda and triangulum galaxies ( m31 and m33 ) .
these discoveries stem from the first year data of the pan - andromeda archaeological survey ( pandas ) , a photometric survey of the m31/m33 group conducted with the megaprime / megacam wide - field camera mounted on the canada - france - hawaii telescope .
both satellites appear as spatial overdensities of stars which , when plotted in a color - magnitude diagram , follow metal - poor , @xmath0 , red giant branches at the distance of m31/m33 . andromeda xxi is a moderately bright dwarf galaxy ( @xmath1 ) , albeit with low surface brightness , emphasizing again that many relatively luminous m31 satellites still remain to be discovered .
it is also a large satellite , with a half - light radius close to @xmath2 , making it the fourth largest local group dwarf spheroidal galaxy after the recently discovered andromeda xix , andromeda ii and sagittarius around the milky way , and supports the trend that m31 satellites are larger than their milky way counterparts .
andromeda xxii is much fainter ( @xmath3 ) and lies a lot closer in projection to m33 than it does to m31 ( 42 vs. 224 ) , suggesting that it could be the first triangulum satellite to be discovered .
although this is a very exciting possibility in the context of a past interaction of m33 with m31 and the fate of its satellite system , a confirmation will have to await a good distance estimate to confirm its physical proximity to m33 .
along with the dwarf galaxies found in previous surveys of the m31 surroundings , these two new satellites bring the number of dwarf spheroidal galaxies in this region to 20 . |
the life of an isolated star is univocally determined once the initial mass , chemical composition , distribution of the angular momentum and the possible presence of magnetic fields are fixed . for historical reasons
the so called `` classical '' models are , or at least were , computed neglecting the presence of rotation and magnetic fields .
the comparison between these `` classical '' models and their observational counterparts showed over the years many discrepancies that can not be explained by these models . among the others , the width of the upper ms and the color distribution of massive stars in the hr diagram ( * ? ? ? * and references therein ) , the relative proportions between o stars and the various wr subtypes ( * ? ? ? * and references therein ) ,
the surface enhancement of he and n in ms stars @xcite , the ratio of blue to red supergiants as a function of the metallicity @xcite , the peculiar chemical composition of many very metal poor stars ( * ? ? ?
* and references therin ) .
in order to improve the comparison between models and observed stars , uncertain physical phenomena like the extension of the convective regions and the efficiency of mass loss were varied more or less arbitrarily . in this way it was possible to improve one discrepancy or the other , but
no satisfactory general scenario has ever emerged from all these tests .
rotation in principle could be a good candidate to reduce all together these discrepancies since it may influence the mass loss rate as well as trigger flows of matter within the star . though the possible influence of rotation on the evolution of a star was already outlined even before the second world war @xcite , only in the last decade or so extended sets of massive star models computed taking into account rotation began to appear .
after the pioneering works by @xcite , @xcite , @xcite and @xcite , grids of rotating massive stars have been computed and discussed by @xcite and @xcite .
more than a decade ago we started a project aimed to study and progressively refine both the evolutionary properties of massive stars from the ms to the core collapse and the passage of the shock wave through the mantle in order to quantitatively determine the final explosive yields of a large number of nuclear species . in this paper
we discuss the inclusion of rotation in our stellar evolutionary code , the franec , and we present a first set of solar metallicity massive star models evolved up to the core collapse together to the following explosion and the final explosive yields . in a companion paper ( in advanced preparation ) we will provide a full database of tracks extending in metallicity between z=0 and z = solar . the paper is organized as follows : section 2 is devoted to the implementation of rotation in the franec code while the calibration of the rotationally induced instabilities is discussed in section 3 .
section 4 and 5 are devoted , respectively , to the h and he burning phases .
the advanced burnings and the explosive yields are discussed in sections 6 and 7 . a final summary and conclusions follows .
the latest version of the franec ( frascati raphson newton evolutionary code ) , major release 6 , now includes the effects of rotation on the evolution of a star .
here we describe only the main features of this new release with respect to the previous one , which is described in detail in @xcite . in principle
the inclusion of rotation in an evolutionary code would require at least a 2d scheme ( if not a 3d one ) .
@xcite and @xcite showed , however , that under few proper assumptions it is possible to simulate the mechanical and thermal distortions induced by rotation in a 1d code . in the following
we will not repeat the detailed derivation of the basic equations in presence of rotation , but we will simply summarize the basics of their derivation and address few comments we consider of interest to properly understand the differences among the various approaches followed by different groups .
we strongly suggest the reader to read carefully the works of @xcite and @xcite for a detailed derivation of the equations in presence of rotation .
the basic assumptions under which it is possible to simulate the average mechanical and thermal distortions induced by rotation in a 1d code , following the scheme proposed by @xcite , are : * cylindrical symmetry for the angular velocity @xmath1 ( needed in order to define equipotential surfaces and to keep the equations for the stellar structure in the same form as in spherical symmetry except for the introduction of two `` form factors '' that take into account the mechanical and thermal distortions induced by rotation ) ; * @xmath2 along the equipotentials and roche approximation ( needed in order to compute easily the shape of the equipotentials ) .
let us note , at this point , that this last assumption initially chosen for sake of simplicity , received additional support from the work of @xcite , who showed that the onset of a strong horizontal turbulence allows the growth of vigorous mixing that very probably keeps the chemical composition and the angular velocity constant along the isobars .
such an approximation is currently referred to as `` shellular rotation '' .
since one of the consequences of the first assumption ( cylindrical symmetry for the angular velocity @xmath1 ) implies that the equipotentials coincide with the isobars , let us therefore stress here that the @xcite scheme is based on the shellular rotation approximation even though it was not named in this way at that time .
as already mentioned above , the combination of cylindrical symmetry and shellular rotation necessarily implies a solid body rotation and this would limit the application of the @xcite scheme to differentially rotating stars .
a strong support to the work of @xcite came a few years later when @xcite showed that it was possible to maintain the equations of the stellar structure as modified by @xcite even dropping the requirement of cylindrical symmetry ( and hence the existence of a conservative field ) .
the price necessary to pay in this case is that the unique @xmath3 and t that can be associated to an isobar in the conservative case must now be replaced by appropriate averages on each isobar ( see * ? ? ? * for more details ) .
since the conservative and non conservative case differ basically for the meaning of the variables ( t and @xmath3 constant on an isobar in the conservative case and just average values in the non conservative case ) , from any practical point of view , the same scheme can be interpreted in either of the two ways .
in addition to the distortions induced by rotation on the structure of a star , it is very important to consider the physical phenomena that may lead to a significant redistribution of the angular momentum and/or chemical composition .
in addition to the more traditional dynamical instabilities ( determined on the basis of the scwharzchild and/or ledoux criteria ) , including semiconvection in h burning , induced semiconvection in he burning and the possible occurrence of some overshooting , we have also implemented two rotational instabilities : meridional circulation and shear instability .
once the instabilities to be included in the code have been identified or , actually , chosen , it is crucial to understand which is the equation that controls the mixing of both the angular momentum and the chemicals . as for the transport of the angular momentum , there are basically two different approaches .
the first one , more rigorous , takes into account the fact that the equation that controls the transport of the angular momentum is not a pure diffusive equation but it has an advective term , which means that the angular momentum is not necessarily spreaded out , flattened , but that it could be even steepened by the redistribution .
we implemented in the franec the same formulation adopted by @xcite and firstly derived by @xcite : @xmath4 where u represents the radial component of the velocity of the meridional circulation , @xmath5 the diffusion coefficient corresponding to the shear instability and all other quantities have their usual meaning .
we adopted the formulation for u given by @xcite ( their eq .
4.38 ) and the expression of @xmath5 proposed by @xcite and @xcite : @xmath6 where @xmath7 , @xmath8 and @xmath9 .
@xmath10 is the coefficient of horizontal turbulence given by @xmath11 @xcite , while @xmath12 is the thermal diffusivity @xcite .
alternative expressions for @xmath13 were proposed by @xcite and @xcite ( both leading to a larger efficiency of the horizontal turbulence ) , but we preferred to adopt the one proposed by @xcite in this ( our ) first paper on rotation for a better compatibility with the set of models published by @xcite .
for the same reason we did not adopt the alternative expression proposed by @xcite for the diffusion coefficient of the shear mixing . as for the transport of the chemicals , @xcite showed that in the asymptotic regime the transport of the chemical composition within a radiative region due to rotation can be described by a pure diffusive process . in this case the diffusion coefficient is given by @xmath14 where @xmath5 is given by equation [ dshear ] while @xmath15 is @xmath16 as suggested by @xcite .
since the numerical implementation of the advective - diffusive equation for the transport of the angular momentum is quite challenging and since the current computations show that in most cases the redistribution of angular momentum goes in the direction of flattening the @xmath1 profile , the formally correct equation is often substituted by a much simpler diffusion equation .
we implemented also this alternative approach in the code .
for the pure diffusion case we used a standard diffusion equation with the diffusion coefficients given by equations [ dshear ] , [ diffcoeff ] and [ diffcoeffmc ] .
the velocity @xmath17 that enters in equation [ diffcoeffmc ] now is not the more general one provided by @xcite ( their eq .
4.38 ) but the simpler expression provided by @xcite : @xmath18 when the pure diffusion approximation is adopted , the diffusion coefficients used for the mixing of the chemical composition are the same adopted for the transport of the angular momentum ( equations [ dshear ] , [ diffcoeff ] , [ diffcoeffmc ] and [ velmckipp ] ) .
the set of equations formed by the four ones describing the physical structure of the star plus the `` n '' ones describing the total chemical evolution ( i.e. the local burning due to the nuclear reactions plus the various kinds of mixing : convection , semiconvection and rotationally induced mixing ) are coupled together and solved simultaneously by means of a relaxation technique .
the temporal evolution of the angular momentum together to the determination of the profile of the velocity of the meridional circulation is solved separately , again by means of a relaxation technique .
mass loss has been included following the prescriptions of @xcite for the blue supergiant phase ( @xmath19 ) , @xcite for the red supergiant phase ( @xmath20 ) and @xcite for the wolf - rayet phase .
the criteria adopted to define the different wr subclasses are the same adopted in @xcite .
the enhancement of the mass loss due to the formation of dust during the red supergiant phase has been included following the prescriptions of @xcite .
mass loss is enhanced , in rotating models , following the prescription of @xcite .
the nuclear network includes 163 isotopes ( from h to @xmath21 ) and 448 reactions for h and he burning , and 282 isotopes ( from h to @xmath22 ) and 2928 reactions for the more advanced nuclear burning stages up to the procollapse phase . in total 293 isotopes and about 3000 processes were explicitly included in the various nuclear burning stages .
the database of the cross sections adopted in this first paper on rotating stellar models is the same adopted in @xcite .
the heavy element solar mixture is the one provided by @xcite .
the initial values of the global metallicity @xmath23 and of the he abundance @xmath24 have been set to @xmath25 and @xmath26 , respectively .
these values have been obtained by requiring that a @xmath27 , computed by taking into account the atomic diffusion , reproduces the main properties of the present sun ( radius , luminosity , rotation rate and helium abundance ) after @xmath28 ( see * ? ? ? * for more details ) .
all models are started on the hayashi track when the central temperature is of the order of @xmath29 k. the angular velocity is raised above zero when the stellar model reaches the main sequence : at this point @xmath1 is raised slowly and kept constant throughout the star until the chosen surface equatorial velocity is reached .
less than 1% of h is burnt in this phase .
the diffusion coefficients discussed so far are intrinsically uncertain and must be regarded as an estimate of the efficiency of the transport of both the angular momentum and the chemicals . for this reason we decided to consider two free parameters , namely , @xmath30 and @xmath31 , to be calibrated somehow .
similar parameters have already been introduced by , e.g. , @xcite , @xcite and @xcite . the first free parameter , @xmath30 , simply multiplies the total diffusion coefficient defined in equation [ diffcoeff ] and adopted to transport the chemicals , i.e. : @xmath32 this means that @xmath30 directly controls the speed at which the chemical composition is mixed in the various zones inside a star without affecting the transport of the angular momentum .
the second free parameter , @xmath31 , multiplies the gradient of molecular weight ( @xmath33 ) and regulates the influence of this quantity on the mixing of both the angular momentum and the chemical composition .
this parameter is necessary because the inclusion of the gradient of the molecular weight at its face value strongly inhibits the transport @xcite . in the present set of models ,
computed by adopting the advection - diffusion scheme , we chose a conservative approach by fixing @xmath34 and by calibrating @xmath31 so that stars of solar metallicity in the mass range 15 - 20 @xmath35 , and settling on the main sequence with an initial equatorial velocity of 300 km / s , increase the initial surface nitrogen abundance by roughly a factor of three at core h depletion ( see , e.g. , * ? ? ?
the @xmath31 value that gives such an increase is 0.03 .
figure [ checka03b ] shows the run of the surface n abundance as a function of the central h mass fraction for three models of 20 @xmath35 , all computed with @xmath34 but for three different values of @xmath31 , namely , @xmath36 .
the largest nitrogen increase is obviously obtained for the case @xmath37 and is of the order of 5 .
how already stated above , we implemented also a pure diffusive scheme to treat the transport of the angular momentum .
also this scheme obviously requires a calibration of the two parameters @xmath30 and @xmath31 . in this case
it is not possible to use @xmath34 because the mixing is now much more efficient . since there is more than one combination of @xmath30 and @xmath31 values that may provide the same surface n enhancement , we explored some alternatives by calibrating the value of @xmath30 for three values of @xmath31 , namely @xmath38 ( the same values shown in figure [ checka03b ] ) .
the values of the @xmath30 parameter that provide the requested surface nitrogen enrichment in each one of the three 20 @xmath0models are @xmath39 , respectively , and the corresponding surface n enhancements are shown in figure [ checka04b ] . obviously the larger the @xmath40 barrier the larger the @xmath30 value needed to obtain the same mixing efficiency .
the black dashed line in figure [ checka04b ] refers to our `` standard '' case , i.e. the model computed in the advective - diffusive framework with @xmath34 and @xmath41 .
it is worth noting that a surface n enhancement of the order of 3 at core h depletion , for solar metallicity rotating stars in the mass range 15 - 20 @xmath0with @xmath42 , is comparable to what is obtained also by other groups . without having the aim of discussing their calibrations ,
let us simply note that , for example , @xcite obtain @xmath43 for a 20 @xmath0star with @xmath44 , while @xcite find @xmath45 for a 20 @xmath0star with @xmath46 . a last point we want to stress here concerns the fact that the surface n enhancement ( @xmath47n ) at core h depletion depends significantly also on the initial abundances of carbon and nitrogen .
the reason is straightforward , since the @xmath47n comes form the conversion of c into n ( cn cycle ) that occurs in the deep interior of the star : the smaller the initial n abundance the larger the @xmath47n , and also the larger the initial c abundance the larger the @xmath47n as well .
such an occurrence has been well shown quantitatively by @xcite .
as for the size of the convective regions , we did not attempt any calibration but we included an overshooting of 0.2 @xmath48 for compatibility with our previous set of models @xcite .
the core h burning phase is by far the longer lasting evolutionary phase of a star .
therefore it is either the phase during which a star can be more probably observed and also the one during which the secularly unstable rotationally mixing phenomena have more time to operate .
how we have already stated in the previous sections , the influence of rotation on the evolution of a star during core h and core he burning phases has been already discussed many times .
hence here we will only discuss a few aspects we consider of interest for the reader .
let us firstly show how the various effects of rotation sum up to determine the position of a rotating star on the main sequence and affect its core h burning phase .
of course , the influence of rotation will depend first of all on the amount of angular momentum injected in the star at the beginning : all models computed for this test have an initial surface equatorial velocity of 300 km / s , which corresponds ( in these specific models ! ) to an @xmath49 , where @xmath50 is the classical critical angular velocity and , in the framework of the roche model , is given by @xmath51 , where @xmath52 is the polar radius of the star when the surface rotates with the critical velocity .
this velocity is already larger than the bulk of observational data available up to now .
figure [ checka02b ] shows five tracks : the first one ( black solid line ) refers to a standard 20 @xmath0computed without rotation while the second ( red dotted line ) refers to the evolution of the same stellar model in which only the @xmath53 and @xmath54 factors have been taken into account .
this means that only the mechanical and thermal distortion of the structure is considered without any kind of rotationally induced mixing : it goes without saying that this automatically implies local conservation of the angular momentum .
the red dotted line is shifted towards slightly lower luminosities and surface temperatures as a consequence of the centrifugal force that sums to the pressure gradient in sustaining the star ( i.e. , the effective gravity is reduced and the star is more expanded ) .
the third track ( blue dashed line ) refers to a model computed like the red one but including also the transport of the angular momentum by either the meridional circulation and the shear instability .
the inclusion of the transport of the angular momentum increases progressively the shift of the model towards lower effective temperatures because it raises the centrifugal force in the mantle of the star ( hence lowers even more the effective gravity ) .
the green dash - dot line refers to a model computed by adding also the mixing of the chemical composition due to the meridional circulation : in this case the star shifts slightly blueward during the core h burning phase because the partial mixing of the matter throughout the whole star smears out the gradient of molecular weight and makes the star more compact on average ( see below ) . note that this mixing increases the size of the h depleted zone ( in some sense it acts similarly to the overshooting ) and this is clearly shown by the much higher luminosity at which the overall contraction occurs .
the last sequence ( violet dash - dot - dot line ) refers to the complete model in which also the shear instability contributes to the mixing of the matter .
since the net result is simply a more robust mixing of the matter inside the star , this stellar model shows a more pronounced blueward shift and a higher luminosity at the end of the central h burning . a basic comparison between the h burning phase of two different sets of models ( in this specific case rotating versus non - rotating models ) naturally implies a comparison of the paths in the hr diagram , of the internal structures , of the lifetimes and of the total amount of mass lost .
also possible differences in the surface chemical composition are extremely relevant first of all because the surface of a star is our basic observational counterpart . in the case of rotating models also the variation of the surface velocity with time must be discussed because this is another quantity that can be observed .
the evolutionary path of rotating ( with an initial equatorial velocity of @xmath55 ) and non rotating models in the hr diagram is shown in figure [ hrtot ] while the main evolutionary properties of the two sets of models are reported in tables 1 and 2 .
the various columns report , for each mass :
1 ) the evolutionary phase , 2 ) the lifetime ( in yr ) , 3 ) the mass size of the convective core ( in solar masses ) , 4 ) and 5 ) the average effective temperature and luminosity , 6 ) , 7 ) and 8) the total mass , the he core mass and the co core mass ( in solar masses ) , 9 ) the equatorial velocity ( in km / s ) , 10 ) the surface angular velocity ( in @xmath56 ) , 11 ) the ratio of the surface angular velocity with respect to the critical angular velocity , 12 ) the total angular momentum ( in @xmath57 ) , 13 ) , 14 ) and 15 ) the surface abundances of h , he and n in mass fraction , 16 ) and 17 ) the n / c and n / o surface ratios . according to what was discussed above , all rotating models settle on a ms position slightly less luminous and redder than their respective non rotating models .
the path of the models in the hr diagram will depend on the balance between the outward transport of the angular momentum ( that tends to push the models towards lower effective temperatures ) and the transport of the chemicals ( that , viceversa , tends to push the track blueward because of the smearing out of the molecular weight gradient ) .
the transport of the angular momentum prevails in the lower mass stars hence the tracks of these models remain roughly parallel to those of the non rotating ones . on the contrary ,
the transport of the chemicals prevails in the higher mass stars so that the path of these models tends to shorten and even cross that of their respective non rotating counterparts .
it is therefore clear that the larger the efficiency of mixing of the chemicals the smaller the minimum `` crossing mass '' , i.e. the mass for which the rotating and non rotating tracks intersect during the h burning phase .
an inspection of the sets of models published by @xcite and @xcite shows that in both bases such minimum `` crossing mass '' is much lower than in our case .
this clearly points towards a much more efficient mixing in their models .
this is well confirmed by figure [ confazoto ] ( see below ) that shows a comparison of the surface n enrichment among the @xcite , @xcite and our models : the @xcite model shows a very efficient mixing since the beginning that prolongs for the whole h burning phase while the @xcite one shows an efficient mixing for more the first 40% of the h burning phase or so .
our model , viceversa , show a rapid growth of the surface n abundance only towards the end of the central h burning phase .
it is not easy to understand where these differences come from because each set of models assumes different coefficients for the mixing of the chemicals : we adopt the diffusion coefficient proposed by @xcite , that includes the contribution of the horizontal turbulence as given by @xcite , while @xcite adopt the coefficient provided by @xcite that does not include the horizontal turbulence .
@xcite adopt a even different approach @xcite and moreover take into account also the transport of the angular momentum by magnetic fields ( but not the transport of the chemicals ) .
the very strong dependence of the mass loss rate ( in ms ) on the luminosity ( i.e. on the initial mass ) naturally divides massive stars in two subgroups : a first group formed by those stars that do not lose a significant amount of mass during core h burning ( @xmath58 ) and those whose ms evolution is dominated by mass loss . during core h burning ,
rotationally induced mixing drives a continuous slow ingestion of fresh fuel into the convective core , as well as a slow mixing of the freshly synthesized h burning products into the envelope of the star .
such a mixing is clearly visible by comparing both the h and the @xmath59 profiles , in rotating and non rotating models , at different stages during core h burning ( figures [ confstru1 ] and [ confstru2 ] ) . in particular ,
it is readily evident in figures [ confstru1 ] and [ confstru2 ] , that the chemical profiles are much shallower in rotating models than in the non rotating ones .
this is a clear effect of a slow diffusive mixing between the h convective core and the radiative h - rich envelope . such a mixing is driven by both the meridional circulation , that dominates at the base of the radiative envelope , and by the shear instability that is more efficient in the outer layers ( see right panels of figures [ confstru1 ] and [ confstru2 ] )
the main effect of this slow ingestion of fresh fuel within the h convective core is that of slightly increasing the h burning lifetime of rotating models with respect to the non rotating ones .
the differences , however , do not exceed the 15% at the lower end of the mass interval ( 13 @xmath0 ) and flatten out at a 10% level above 60 @xmath0or so ( see figure [ times ] ) .
another interesting information provided by figures [ confstru1 ] and [ confstru2 ] , and in particular from the internal profiles of the two `` form factors '' @xmath53 and @xmath60 , is that the departure from the spherical symmetry ( i.e. where @xmath53 significantly decreases below 1.0 ) is negligible in the more internal zones while it progressively increases moving outward in mass .
the different paths followed by rotating and non rotating models in the hr diagram together to the slight differences in the lifetime naturally lead to different mass loss histories and hence to different total masses at core h exhaustion .
figure [ masshburn ] shows such a comparison . since rotating models have lower effective temperatures , higher luminosities and longer lifetimes compared to their corresponding non rotating ones , and since mass loss scales directly with the luminosity and inversely with the effective temperature , rotating models are expected to lose more mass than their respective non rotating counterparts .
moreover , mass loss in rotating models is enhanced due to the effect of the centrifugal force ( see section [ franec ] ) .
figure [ masshburn ] shows that this expectation is fulfilled for masses @xmath61 and for the 120 @xmath0model while it is completely reversed for both the 60 @xmath0and the 80 @xmath0models . in order to understand such an `` unexpected '' result
, it must be remembered that the mass loss rate provided by @xcite has two threshold temperatures that mark dramatic changes in the mass loss rate .
this means that if a star crosses or not one of these so called `` bi - stability jumps '' , its mass loss rate may or may not change dramatically .
both the rotating 60 @xmath0and 80 @xmath0models evolve to higher effective temperatures compared to their non rotating counterparts and simply do not cross such threshold temperatures that mark the strong enhancement of the mass loss rate and therefore they do not lose as much mass as their respective non rotating models .
an important property of the models , worthwhile to be addressed , is the surface chemical composition at core h depletion .
figure [ azotofinal ] shows the ratio between the final ( at core h exhaustion ) and the initial surface @xmath59 mass fraction as a function of the initial mass for rotating and non rotating models .
no physical phenomenon can drive a change of the surface chemical composition during core h burning in classical non rotating models .
in fact non rotating models with initial masses @xmath61 do not show any increase of the @xmath62/@xmath63 ratio .
non rotating models with initial masses larger than 40 @xmath0 , on the contrary , show a quite large @xmath62/@xmath63 ratio , which is almost independent of the initial mass .
the reason for such a @xmath59 enhancement is that mass loss becomes so efficient in these models that it peels off the outer envelope down to within the initial size of the convective core , in this way the products of the h burning may show up at the surface .
since the @xmath59 equilibrium abundance , resulting from the cno processing , has a very mild dependence on the temperature , all these stars show roughly the same @xmath62/@xmath63 value at the end of core h burning .
rotating models with initial masses lower than 40 @xmath0behave differently because in this case rotationally induced mixing is able to bring some products of the h burning ( see figures [ confstru1 ] and [ confstru2 ] ) so that a sizable @xmath59 enhancement may be obtained in this case . as the initial mass increases , the mass loss increases as well and hence a stronger @xmath59 enhancement is expected .
this is clearly visible in figure [ azotofinal ] although the @xmath59 enhancement is not as extreme as in the non rotating models .
the reason is that stars in the mass interval 60 @xmath0to 80 @xmath0lose less mass than their non rotating counterparts ( see figure [ masshburn ] ) and therefore are not able to expose to the surface zones previously located within the convective core . since figure [ azotofinal ] provides the surface chemical composition at core h depletion , it is not much useful to compare the expected surface properties of rotating models and the observed abundances in o - type and b - type stars .
in fact the probability to see a star in a given evolutionary phase is proportional to the time spent in that phase .
therefore a much more meaningful information , for this purposes , is the temporal evolution of the surface @xmath59 mass fraction during core h burning .
figure [ nitrotemp ] shows the surface @xmath59 mass fraction normalized to the initial value ( top panel ) , the equatorial velocity @xmath64 ( middle panel ) and the @xmath65 ratio ( bottom panel ) as a function of the h burning lifetime .
this figure shows that the @xmath59 surface enrichment does not occur at the beginning but , on the contrary , in an advanced stage of core h burning .
there is also a dependence on the initial mass .
the smaller the mass the later the surface enrichment : in the 13 - 15 @xmath0models the enrichment occurs during the latest 20% or so of core h burning . as the initial mass increases , the beginning of the surface enrichment extends over a progressively larger fraction of the h - burning lifetime . in the extreme case of the 120 @xmath0model ,
the surface enrichment starts just after the first 20% of the core h - burning .
the surface velocity remains roughly constant during most of the h burning , a small slowdown ( 20% at most ) occurring only in the more massive models .
a comparison between the temporal evolution of the surface @xmath59 mass fraction and of the rotational velocity implies that in a steady state situation in which all stars rotate initially at 300 km / s , one should expect the following : ( 1 ) the majority of the stars ( for any reasonable imf , the lower mass stars are the most numerous ones ) share a similar velocity , i.e. the initial one , and does not show any substantial surface @xmath59 enrichment ; ( 2 ) the small fraction of stars showing a sizable @xmath59 enhancement should span a range of surface velocities between @xmath66 and @xmath67 km / s , because the surface @xmath59 enrichment occurs during the late stages of core h burning which is characterized by the largest variation ( decreasing ) of the surface velocity . since the distribution of stars in the `` ( @xmath59/@xmath68@xmath69 -@xmath70 '' diagram ( where @xmath70 is the projected rotational velocity , `` i '' being the angle between the rotational axis and the line of sight ) plays a pivotal role in the comparison between the predictions provided by rotating models and the observed properties of the rotating stars , it is worth mentioning here that the temporal evolution of the surface abundance of @xmath59 depends on the basic scheme adopted to perform the rotationally induced mixing . to be clearer , in the previous section we showed in figure [ checka04b ] a plot of the run of the surface @xmath59 versus central h abundance for four different evolutionary tracks of a 20 @xmath0 .
the black line refers to our standard run ( therefore computed within the advection - diffusion scheme discussed in section [ franec ] ) while the red , blue and green lines refer to three evolutions computed in the pure diffusion scheme but different choices for the free parameters @xmath30 and @xmath31 ( see section [ calibration ] ) . note that , though all four tracks reach a similar final overabundance of @xmath59 at the end of core h burning , the change of the surface abundance with time ( or , equivalently , the central h mass fraction ) depends on the specific run .
for example , while in the green run ( pure diffusion with @xmath71 and @xmath72 ) the surface abundance begins to increase since almost the beginning of core h burning , in the red case ( pure diffusion with @xmath73 and @xmath74 ) the surface abundance remains at its initial value for most of the time , raising at the required value just before the end of the h burning .
our standard track ( black line ) shows an intermediate behavior .
of course these differences will affect the expected distribution of stars in the ( @xmath59/@xmath68@xmath69-@xmath75 diagram .
given the importance of such a diagram , we compare the evolutionary properties of our 20 @xmath0rotating model with those of similar models presently available in literature , namely , the one by @xcite , @xcite and @xcite ( see caption for more details on the models ) .
the upper left panel of figure [ confazoto ] compares the path of these models in the hr diagram : the four models show a quite similar core h burning phase , all starting basically from the same location and diverging slightly during the evolution probably because of small differences in the extension of the convective core and in the efficiency of both the angular momentum transport and the rotationally induced mixing .
also the temporal evolution of the surface equatorial velocity is not much different ( upper right panel , here the velocity is plotted as a function of the central h mass fraction ) , the largest discrepancy occurring in the @xcite model : their model is the only one to show a systematic ( even if not strong ) drop of the surface velocity during core h burning . the temporal evolution of the surface n abundance ( normalized to the initial one ) is shown in the lower left panel . both our model and the @xcite one show a very mild increase during most of the core h burning lifetime and a more consistent increase once the models enter the bi - stability jump .
viceversa , the @xcite model quickly reaches an asymptotic value that remains roughly constant during most of the core h burning lifetime .
note , however , that these models are significantly different from the others since they include the effects of magnetic fields on the transport of the angular momentum ( but not on the transport of the chemicals ) . the model taken from the @xcite database shows an even different behavior , since it steeply increases all along the core h burning phase .
the lower right panel shows the path of the four models in the ( @xmath59/@xmath68@xmath69-@xmath76 diagram with superimposed 7 dots along each track , marking the 20% , 40% , 60% , 80% , 90% , 95% and 100% of the core h burning lifetime ( @xmath77 ) , respectively . with the aid of these time - slices we can easily compare the predictions of the four different models .
our track ( black line ) predicts that a 20 @xmath0star remains @xmath78 of its @xmath77 at basically its initial rotational velocity , with a surface n overabundance below a factor of two .
only the last @xmath79 of @xmath77 is spent by this model with a relatively lower surface velocity ( of the order of 150 - 200 km / s ) and a higher n overabundance ( between a factor of 2 and 3 ) .
the model of @xcite shows a similar behavior but with the difference that the typical rotational velocity during the last @xmath79 of its @xmath77 is of the order of 50 - 100 km / s . the model provided by @xcite predicts a quite large spread of n overabundances at roughly constant equatorial velocity ( @xmath80 , i.e. , the initial one ) for about 90% of @xmath77 and then a mild spread of the rotational velocities ( 150 - 250 km / s ) at roughly constant n overabundance for the remaining @xmath81 of its @xmath77 .
the behavior of the @xcite model is even more extreme in the sense that it predicts a wide spread of n overabundances but basically at constant equatorial velocity all along the core h burning phase .
these different behaviors of the four models will obviously reflect in a different distribution of the stars in the @xmath59/@xmath68@xmath69-@xmath76 diagram .
this issue will be discussed in a forthcoming paper .
another property of rotating models which is worth discussing is the efficiency of the angular momentum transport during core h burning .
figure [ deltamangh ] shows , for three selected models ( i.e. , 15 , 60 and 120 @xmath0 ) , the logarithm of the ratio between the final and initial specific angular momentum , as a function of the mass coordinate , at the end of core h burning .
inspection of this figure reveals that the angular momentum is transported from the innermost zones towards the outer ones and that it is eventually lost by stellar wind . as a result of these combined processes ,
rotating models lose roughly between @xmath82 and @xmath83 of their initial angular momentum during core h burning . before closing this section
let us briefly mention that the more massive stars may lose such a large fraction of their initial mass during core h burning , that they become wolf - rayet ( wr ) ( see * ? ? ?
* for the definition of the various wr stages ) even before leaving the main sequence .
the minimum mass to become a wr already during core h burning is 120 @xmath0 , for a set of models with a main sequence equatorial rotational velocity of 300 km / s , while it drops to 80 @xmath0 in the non rotating case .
the rotating 120 @xmath0enters the wnl stage when the central h mass fraction is @xmath84 while the non rotating 80 and 120 @xmath0models enter the wnl stage at @xmath85 and @xmath86 , respectively .
it goes without saying that such a limiting mass would vary in the range 80 - 120@xmath0as the initial rotational velocity spans the range 0 - 300 km / s .
this means that it is not straightforward to define unambiguously such ( actually any - see below ) a limiting mass in a realistic case in which a generation of stars is expected to show a flat ( i.e. constant ) initial distribution of rotational velocities ( hereinafter idrov ) .
note that in principle the idrov is two - dimensional since the initial rotational velocity may depend either on the mass but also vary from star to star of the _ same _ mass .
we ll come back to this point below .
the first step that ferries a star from the end of core h burning to the central he ignition is the contraction and heating of the core , up to the typical he burning temperatures ( @xmath87 ) , coupled to an expansion of the h - rich mantle . during this phase the star evolves from a blue supergiant ( bsg ) - high temperature - configuration , to a red supergiant ( rsg ) - low temperature - structure .
the transition from a bsg to a rsg configuration is mainly controlled by the h - rich envelope .
if the h - rich envelope is thick `` enough '' , the star keeps a rsg structure while , on the contrary , as soon as the mass of the h - rich envelope decreases below a critical threshold value the star turns to a bsg configuration . by the way
let us remind the reader that we are discussing solar metallicity stars ; at lower metallicity a bsg structure may be retained even in presence of massive h - rich mantles . in the present set of models
only the two ( rotating and non rotating ) 120 @xmath0fail to reach a rsg configuration ( figure [ hrtot ] ) because of the very thin h - rich envelope ; all other models , rotating or not , become rsg at the very beginning of core he burning .
the timescale of the transition from a bsg to a rsg configuration after core h depletion can not be robustly determined on the basis of first principles .
the reason is that it depends on the efficiency of the mixing in the region of variable h left by the receding convective core during core h burning ( figure [ confstru2 ] ) that in turn is still highly uncertain .
in fact while the strict adoption of the schwarzschild would imply that this region is unstable , the stabilizing effect of the @xmath40 gradient would keep these layers stable ( ledoux criterion ) .
this zone is often referred to as the `` semiconvective '' region .
if the schwarzschild criterion for convection is adopted , these layers are mixed on a dynamical timescale and the redward excursion occurs on a nuclear timescale .
this would imply that the region between the ms and the red giant branch ( rgb ) in the hr diagram should be well populated . on the contrary
, the adoption of the ledoux criterion prevents a mixing on a dynamical timescale and in this case the redward evolution occurs on the much faster kelvin - helmoltz timescale . in this case one would expect very few stars ( basically a gap ) in the hr diagram between the ms and the rgb .
unfortunately , the ledoux criterion is not that robust because the gradient of mean molecular weight ( @xmath88 ) , that stabilizes the zone against convection , could be ( at least at small scales ) destroyed by any kind of stochastic turbulence , favoring the dynamical mixing that , in turn , would reduce even more @xmath88 .
this is clearly an unstable situation .
the physics of this phenomenon has been studied by many authors by means of linear - stability analyses .
for example @xcite showed that infinitesimal perturbations would grow in this region on a thermal diffusion timescale , due to heat dissipation processes , with the net result of mixing eventually the whole region .
unfortunately , the timescale over which these perturbations become finite as well as the real time - scale of the mixing process is still unknown .
hence the only guidance we have to treat these layers comes from the observations which clearly show a shortage in the number of stars between the ms and the rgb ( * ? ? ? * and references therein ) .
this implies that at most a small efficiency of the mixing is allowed by the observations . in the present models we treat the semiconvective region following the same formalism introduced by @xcite and we fix the free parameter @xmath89 to the largest value that allows all the models to become rsg at the very beginning of core he burning : in our models it amounts to @xmath90 .
figure [ mhecore ] summarizes some selected properties of the models at core h exhaustion and shows that the he core masses are almost independent of rotation , the only exceptions being the 120 @xmath0since the rotating star lose much more mass during the ms than its non rotating counterpart ( see previous section ) entering the wolf - rayet phase when the central h abundance is still relatively high : the very efficient mass loss associated to the wr phase has therefore much more time to operate . as already discussed above , the position of a model in the hr diagram in core he burning depends on the mass size of the h - rich envelope ( defined as the amount of mass above the he core mass ) .
figure [ mhecore ] ( dotted lines ) also shows that rotating and not rotating models reach the end of the core h burning with substantially different h - rich mantles .
the consequence of this difference on the evolutionary path of the models in the hr diagram during core he burning is shown in figure [ red_blue ] .
in particular , all models ( rotating and not ) with @xmath91 spend all their core he burning as rsgs since in both cases the mass of the mantle remains above the critical value needed for a blueward evolution .
on the contrary , all models with mass @xmath92 evolve to the blue at the very beginning of the he burning , regardless of the mass size of the envelope , because of the huge mass loss typical of this mass interval that is able to eject all the mantle anyway .
the largest differences between rotating and non rotating models manifest in the mass interval between the 20 and the 40 @xmath0 .
the reason is that stars in this mass interval spend part of their central he burning lifetime as rsg and part as bsg .
as already discussed above the transition occurs once the mass size of the envelope has greatly reduced and since rotating models reach the he ignition with smaller envelope masses , they will lose their h rich envelope earlier than their non rotating counterparts ( that have more massive h - rich envelopes ) . hence the ratio between the time spent as rgb and the total time spent in central he burning is systematically smaller in presence of rotation .
also the minimum mass that experiences a blueward evolution ( which means the smallest mass that becomes wr , see below ) reduces in presence of rotation .
figure [ red_blue ] shows the @xmath93 ratio as a function of the initial mass for both sets of models . in the specific case of stars rotating initially at 300 km /
s the minimum mass that becomes wr is smaller than 20 @xmath0(to be compared to a limiting mass of the order of 25 @xmath0obtained in the non rotating case ) .
of course since not all stars are born with the same rotational velocity , the concept of a _
unique _ limiting mass loses of meaning in presence of rotation and should be replaced by the idea that a _ spread _ of limiting masses exists due to the variety of possible initial rotational velocities .
let us eventually note that non rotating models in the interval 20 - 35 @xmath0experience a blue loop during core he burning , turning back towards a rsg / ysg ( ysg stands for yellow supergiant ) configuration at core he depletion .
let us now turn to the internal properties of these models starting with the non rotating models with mass @xmath94 .
these models are characterized by a he convective core that either advances or remains stable in mass .
if the schwarzschild criterion were adopted , the internal chemical composition of these stars , at core he depletion , would be characterized by a sharp chemical discontinuity marking the maximum extension of the he convective core .
behind such a discontinuity , the chemical composition would be homogeneous and dominated by @xmath95 and @xmath96 , i.e. , the main ashes of he burning .
outside the chemical discontinuity , the chemical composition would be dominated by @xmath97 and @xmath59 , i.e. the main products of core h burning , and would be homogeneous as a result of the convective mixing during core h burning .
the adoption of the ledoux criterium , viceversa , creates a small region of variable chemical composition at the border of the convective core in which @xmath95 is already enhanced with respect to the pre he burning abundance but the @xmath59 has not yet been destroyed by @xmath98 captures .
this special region remains untouched up to the end of the central he burning and is of great importance because its chemical composition resembles the one observed on the surface of a quite rare subset of wr stars , namely the wnc stars ( @xmath99 ) ( upper left panel of figure [ 1560endhe ] ) . in these models , however , mass loss is not strong enough to expose these layers to the surface , hence none of them will show up as wnc stars .
non rotating stars with mass larger than 40 @xmath0 , on the contrary , experience a strong mass loss after core h burning so that they completely lose the h rich envelope and a fraction of the he core during core he burning .
such an occurrence deeply affects the evolution of these stars .
the upper right panel of figure [ 1560endhe ] shows the internal chemical composition of a non rotating 60 @xmath0star , taken as representative of the high mass models .
the he profile looks quite different from the one shown by the 15 @xmath0 . in this case ,
in fact , the continuous shrinking of the he convective core caused by the progressive reduction of the he core mass induced by the strong mass loss ( note that a similar effect could be obtained in the case of a roche lobe overflow ) leads to the formation of a shallow he profile . by the way ,
this is a clear demonstration of the influence of the he core mass size on the evolution of the internal regions : the smaller the ( he core ) mass the smaller the size of the convective core .
let s now turn to the rotating models ( lower panels in figure [ 1560endhe ] ) .
in the 15 @xmath0case , the he profile shows now a pronounced slope within the whole he core .
this is not the consequence of a reduction of the he core ( that is not eroded by mass loss in this case ) , but the result of the work done by the rotationally induced mixing that partially spreads the products of the he burning within the he core and also slightly within the tail of the h burning shell .
this slow ingestion of fuel ( he in this case ) within the convective core has the visible consequence of lowering significantly the @xmath100 ratio and of increasing the mass of the co core at the end of core he burning ( see * ? ? ?
* for a detailed discussion of the influence of the mixing on the @xmath100 ratio at core he exhaustion ) , while the contemporaneous penetration of fresh @xmath101 into the tail of the h burning shell activates the production of a locally consistent amount of primary @xmath102 .
induced rotational mixing operates also in more massive models at the beginning of core he burning and produces the same effects discussed for the 15 @xmath0 , but then mass loss begins to reduce progressively the he core so that the chemical composition present within the he core at core he exhaustion results from the combined effect of the rotationally induced mixing ( operating during the early stages of core he burning ) and the progressive reduction of the he core due to mass loss ( working at later stages of core he burning ) .
tables 3 and 4 show the time spent by all the present models in each specific wr stage .
the various columns show : 1 ) the initial mass , 2 ) the time spent by the model as o - type star ( in yr ) , 3 ) the time spent globally as wr star ( in yr ) , 4 ) the time spent as wnl ( in yr ) , 5 ) the central abundance , between brackets , when the star enters the wnl phase , 6 ) the time spent as wne ( in yr ) , 7 ) the central abundance , between brackets , when the star enters the wne phase , 8) the time spent as wnc ( in yr ) , 9 ) the central abundance , between brackets , when the star enters the wnc phase , 10 ) the time spent as wco ( in yr ) and 11 ) the central abundance , between brackets , when the star enters the wco phase .
inspection of these tables shows that the minimum mass for the formation of a wr star is of the order of @xmath103 if @xmath55 , value that raises to @xmath104 in the non rotating case .
it is interesting to note how strongly the wne and the wnc stages are influenced by rotation when the initial rotational velocity is @xmath55 .
while only a minor fraction of the region confined between the convective core and the border of the he core is enriched in @xmath101 in the non rotating case ( see above ) , the rotationally induced mixing spreads out fresh @xmath101 produced in the core within the whole he core ( raising the c / n ratio above 0.1 ) so that no he rich layer resembles any more the typical composition of a wne star but that of a wnc one ( figure [ comp60he ] ) . more specifically , all rotating models with mass @xmath105 change directly from a wnl to a wnc stage ( with the only exception of the @xmath106 ) and fully skip the wne stage . on the contrary , both the wne and wnc stages
are populated in absence of rotation , being 40 @xmath0and 60 @xmath0the minimum masses that become wne and wnc respectively .
figure [ wrconf ] shows a comparison between the expected and the observed numbers of wolf - rayet stars with respect to the o - type stars ( wr / o ) and the wr - subtypes / wr ratios for the two sets of models . by the way
let us remind that this comparison is obtained by making the nonphysical assumption that the idrov is flat and that the rotational velocity is 300 km / s .
note that the assumed idrov is bidimentionally flat in the sense that we assume either that the initial surface equatorial velocity is independent on the mass but also that all stars of the same mass born with the same initial velocity ( there is no evident reason for this ) .
for this reason a more realistic prediction of the number ratios among the various wr - subtypes and o - type stars should be based on synthetic stellar population synthesis in which an idrov as anchored as possible to the real world is taken into account . keeping in mind this general warning , figure [ wrconf ]
shows a comparison between the expected ratios under the assumption of a steady state distribution of stars of solar metallicity and a salpeter imf ( @xmath107 ) with @xmath108 .
the observed wr / wr - subtypes ratios have been taken from the compilation of @xcite the grey band in figure [ wrconf ] marks a region corresponding to an uncertainty of a factor of 2 .
the integrated wn / wr and wco / wr ratios are basically unaffected by rotation while both the wr / o and the wnc / wr ratios predicted by the present rotating models are a factor of 3 to 4 higher than in the non rotating case .
a comparison between our theoretical predictions and those obtained by @xcite shows an excellent agreement for the wr / o , wn / wr , and wco / wr ratios ( figure [ wrconf ] ) .
viceversa , the wnc / wr ratio differs by a factor of @xmath109 . such a difference should come out from a different efficiency of the rotational mixing in the region between the he convective core and the he core , and from the presence and extension of the semiconvective region at the border of the he convective core ( see above ) . the comparison between the observed and the predicted ratios is quite satisfactory in both the rotating and non rotating cases , at least within a factor of 2 .
the only significant discrepancies occur for the wr / o ratio , in the non rotating case , and for the wnc / wr ratio , in the rotating case .
the large underestimated wr / o ratio , obtained in the non rotating models , means that the minimum mass for wr star formation is too large , a reduction of this limiting mass should significantly improve the fit .
as far as the rotating models are concerned , the large overestimated wnc / wr ratio is connected to the very efficient rotationally induced mixing during core he burning that , as discussed above , produces an extended zone between the outer edge of the he convective core and the tail of the h burning shell , where both @xmath101 and @xmath102 are quite abundant .
this efficient mixing is also responsible for the lack of wne stars predicted in this case . as a final comment ,
it is interesting to note that , in spite of previous claims that rotation is necessary in order to predict a non zero population of wnc stars @xcite , our non rotating models predict a number of wnc stars that is compatible with the observed number .
let us now turn to the evolution of the angular momentum and of the equatorial surface velocity during core he burning .
figure [ deltamanghe ] shows the logarithm of the ratio between the final ( at core he exhaustion ) and the initial ( at core h exhaustion ) specific angular momentum for three selected models , namely , 15 ( upper panel ) , 60 ( middle panel ) and 120 @xmath0(lower panel ) .
inspection of figure [ deltamanghe ] reveals that these models lose a consistent fraction of their angular momentum also in he burning ; models of lower mass lose roughly 10 - 15% of the angular momentum present in the core , while the more massive models roughly halves the amount of angular momentum located in the interior .
models that burn he as rsgs ( 13 and 15 @xmath0 ) maintain a very small rotational velocity because of the very extended radii .
viceversa , those becoming bsgs experience a substantial increase of the rotational velocity due to the contraction and the consequent speed up of the outer layers . however , all these models converge towards similar structures at core he exhaustion , the final surface velocity being of the order of 100 km / s .
the only exception is the 120 @xmath0 : this model experiences such a huge mass loss ( and angular momentum ) during core h burning that the surface equatorial velocity at core he exhaustion drops to roughly 50 km / s .
after the core he depletion , he burning shifts in a shell and eventually a convective zone above the he burning shell forms . such a he convective shell develops either in a region of variable he ( i.e. in all rotating models and in the non rotating ones with @xmath110 ) , or in a region where he is flat ( i.e. the non rotating models with @xmath111 ) . in the first case
the he convective shell turns out to be hotter then in the second one @xcite ; such a different behavior will have profound and interesting implications on the production of some specific isotopes ( see below ) . before closing this section
let us eventually point out that the he convective shell destroys most of the primary @xmath102 produced by the induced rotational mixing so that no primary n is ejected in the interstellar medium .
the physical parameters driving the evolutionary properties of a massive star after the core he exhaustion are ( 1 ) the co core mass ( that takes the role of the total mass ) , ( 2 ) the @xmath101 mass fraction profile left by core he burning and , if the star rotates , ( 3 ) both the total and the interior distribution of the angular momentum .
these basic properties of the models at core he exhaustion , as a function of the initial mass , are shown in figures [ mcohe0 ] , [ ccen ] and [ omegahe0 ] .
we have already shown in the previous section ( figure [ 1560endhe ] ) that an important consequence of the rotationally induced mixing during core he burning is a lower @xmath101 abundance and a larger co core mass at core he exhaustion , compared to the non rotating case ( this is also readily evident from figures [ mcohe0 ] , [ ccen ] ) .
these `` effects '' , however , do not affect all the models in the same way but , on the contrary , they reach their maximum for the lowest masses and reduce progressively as the initial mass increases .
the reason is partly due to the fact that we adopted an initial equatorial surface velocity equal for all the models , independent on the initial mass . since this choice implies that also the initial @xmath112 reduces as the initial mass increases , the effects of rotation become progressively less important as well .
moreover , it must also be reminded that the evolutionary timescales decrease with the initial mass , hence the larger the mass the faster the evolution and hence the smaller the timescale over which the rotationally induced secular instabilities may operate .
the black lines in figure [ omegahe0 ] show the trend of the angular velocity as a function of the interior mass , for four selected rotating models at core he exhaustion .
the basically flat profile shown by all the models within their co core descends from the fact that the co core basically corresponds to the he convective core and that we impose a solid body rotation in the convective regions .
note the sharp reduction of the angular velocity in the envelope of the 15 @xmath0model which is the only one , among the four , that is a rsg at this stage . in order to address the role of rotation on the advanced evolutionary phases we show in figure [ omegahe0 ] the internal variation of both the angular velocity and the form factor ( @xmath53 ) at two selected evolutionary stages after core he depletion , i.e. , at core si exhaustion ( dotted lines ) and at the presupernova stage ( dashed lines ) .
it is quite evident that , in spite of the spin up of the more internal zones due to the progressive contraction of the core , the structural deformations induced by rotation start to be sizable only after core si exhaustion , i.e. , the form factor ( red lines ) decreases substantially below 1 only beyond this phase .
this means that all the advanced evolutionary phases ( at least from core he exhaustion to core si depletion ) are not much affected by rotation and hence that the final structural differences between rotating and non rotating models depend , almost exclusively , on the differences in the co core mass and in the central @xmath101 mass fraction at core he depletion .
let us briefly remind that the larger is the co core - and/or the lower is the @xmath101 mass fraction at core he exhaustion - the faster is the advancing of the c shell and the more compact is the core of the star . according to these general rules , for the same initial mass , rotating models behave , in some sense , like more massive stars and hence end their life with more compact structures .
figures [ kip020r ] and [ kip020n ] show , as an example , a comparison between the kippenhahn diagrams of a rotating and a non rotating 20 @xmath0model , respectively . while the non rotating model forms a convective core during core c burning , followed by three convective shell episodes in the further evolutionary phases ( typical behavior of the less massive stars ) , the rotating model burns c radiatively in the core and presents only two c convective shell episodes in the further evolution ( typical of a more massive star )
. the larger compactness of rotating models at the presupernova stage is shown in figure [ mara4 ] , where the final mass - radius ( m - r ) relations of a subset of rotating models are compared to their respective non rotating counterparts .
it is important to remind , at this point , that the m - r relation plays a crucial role not only in the evolution of the innermost zones of the exploding mantle during the following explosion of the star ( determining , e.g. , the amount of falling back material ) , but also in determining the final explosive yields .
the reason is that the volumes within which the various explosive burnings occur do not depend on the physical structure crossed by the shock wave and depend very modestly on the energy of the explosion .
therefore , the m - r relation determines essentially the amount of mass present in each one of these volumes and hence the amount of mass processed by the various explosive burnings ( see * ? ? ? * for a detailed discussion of this topic ) .
for this reason we show in figure [ mara4 ] also the radii ( vertical black dotted lines ) identifying the zones undergoing the various explosive burnings for an explosion energy of 1 foe ( 1 foe = @xmath113 erg ) , for each selected presupernova model : from the left to the right the 4 vertical lines mark the distances at which the peak temperature of the shock wave drops to 5 gk ( 3700 km , limit of the complete explosive si burning ) , 4 gk ( 5000 km , limit of the incomplete explosive si burning ) , 3.3 gk ( 6400 km , limit of the explosive o burning ) and 1.9 gk ( 13400 km , limit of the ne / c explosive burnings ) . another important quantity playing a crucial role in determining the isotope distribution of the explosive yields is the electron fraction ( @xmath114 ) profile at the presupernova stage .
this quantity is mostly important for all the explosive burnings where a full- or a quasi- nuclear statistical equilibrium is achieved , i.e. , complete and incomplete explosive si burning and explosive o burning @xcite . while rotation leads to more compact final structures , it is worth noting that the @xmath114 profile is not significantly affected by rotation , the largest ( anyway small ) difference occurring for the 15 @xmath0 in the zone exposed to the explosive si burning ( see figure [ mara4 ] ) .
we will discuss the impact of these similarities and differences between rotating and non rotating models on the explosive yields in the next section .
the modest role played directly by rotation during the advanced burning phases becomes much more relevant beyond the core si burning .
figure [ omegahe0 ] clearly shows that @xmath53 decreases significantly below 1 within the fe core after the core si depletion in both the 15 and 30 @xmath0(upper panels ) . by the way
, let us remind again that our choice of adopting a constant initial surface equatorial velocity for all masses implies that the smaller the mass the larger the initial @xmath112 and hence the impact of rotation on the evolution of the stars .
the large drop of @xmath53 implies a decrease of the effective gravity that in turn determines a slow down of the contraction .
the cores of these rotating models therefore move towards a much more degenerate configuration compared to their non rotating counterparts .
figure [ tcroc ] shows in fact that , while the evolutionary paths of the two sets of models in the @xmath115 plane remain quite similar up to the central si depletion ( the rotating models evolving at slightly larger entropies because of the systematically larger co core masses ) , the rotating models definitely turn towards much higher densities and lower temperatures than their non rotating counterparts . in order to really proof that this behavior is due to rotation ,
we recomputed the evolution of the 30 @xmath0from the core si exhaustion onward , by imposing that @xmath53 and @xmath54 could not decrease below 0.9 .
the resulting track , reported in figure [ tcroc ] as long dashed green line , shows that in this case the contraction of the core is not hindered anymore by rotation .
this result is extremely interesting and leads to a possible speculation about the final evolution of fast rotating massive stars .
if rotation were able to prevent the collapse of the core for a not negligible amount of time , the fe core could grow much more than in the non rotating case before the onset of the collapse .
such an occurrence would hinder the passage of the shock wave because of the huge energy losses the shock wave undergoes inside the fe core and , since the temporal delay would be controlled by the timescale of the outward leakage of angular momentum , a variety of different explosions could in principle occur .
it would be very important to investigate in greater details the possible consequences of the stronger degeneracy we found to occur in fast rotating stellar models and we plan to address this issue in a forthcoming dedicated paper .
figure [ deltamangadv ] shows the variation of the specific angular momentum between the end of core he burning and the presupernova stage .
the _ shark teeth _
visible in the figure mark the convective regions where the rotational velocity is assumed to be flattened out ( see section 2 ) and hence the outward transport of the angular momentum maximized . in the radiative zone , viceversa , the specific angular momentum is , at a large extent , locally preserved because the speed up of the evolution due to the enormous neutrino energy losses largely inhibits both the secular shear and the meridional circulation . as a consequence ,
since no convective region crosses the border of the co core , the total angular momentum locked in the co core will remain basically constant up to the onset of the core collapse .
the limiting masses that mark the passage from one kind of core collapse supernova to another are collected in tables 5 and 6 . for each progenitor mass ( column 1 ) , the amount of h and he present in the envelope at the time of the collapse are shown in columns 2 and 3 for rotating and non rotating models , respectively .
the `` kind '' of star at the onset of the collapse is shown in column 4 , i.e. whether the star is a rsg , ysg , or a wr , while the expected core collapse supernova type is shown in column 5 .
note that , in order to classify the various core collapse supernovae we adopted the following h and he limiting masses ( see also * ? ? ?
* ) : ( 1 ) @xmath116 ( the minimum h mass for a type iip sn ) ; ( 2 ) @xmath117 ( the minimum h mass for a type iib sn ) ; ( 3 ) @xmath118 ( the minimum he mass for a type ib sn )
. inspection of the two mentioned tables show that in both cases , i.e. with and without rotation , the maximum mass for the explosion of a sniip is @xmath119 . on the contrary ,
the minimum mass for the explosion of a snib is @xmath120 in the non rotating set while it lowers to @xmath121 if rotation with @xmath122 is taken into account .
no progenitor star can produce a type ic sn lightcurve because of the rather high mass of he present in the envelope at the time of the explosion ; note however that @xcite question the upper limiting value of @xmath123 to have a type ic supernova explosion since they find that cases in which the he lines ( in particular the he i 10830 line ) are not excited even in presence of a quite large he mass fraction . before closing this section
let us note that the limiting masses recently obtained by the geneva group @xcite are in good agreement with the present findings .
the explosion of the mantle of each model was followed by means of a hydrodynamic code developed by us that solves the fully compressible reactive hydrodynamic equations using the piecewise parabolic method ( ppm ) of @xcite in the lagrangean form .
each explosion was started by means of a kinetic bomb , i.e. by imparting instantaneously an initial velocity @xmath124 to a mass coordinate of @xmath125 , i.e. well within the iron core @xcite .
since these explosions are not obtained from _ first principles _ , the initial velocity must be tuned in some way . in general , @xmath124 can be calibrated in order to obtain a given final kinetic energy of the ejecta or a given amount of mass ejected ( i.e. a given amount of radioactive @xmath126 ) . in the present paper
we choose , for each model , the minimum initial velocity which provides the ejection of the whole mantle above the fe core .
this second case leaves the freedom to choose the mass cut ( i.e. , the mass coordinate which separates the ejecta from the compact remnant ) a posteriori , e.g. by requiring the ejection of a specific amount of @xmath126 . tables [ tabyields_rot ] and [ tabyields_norot ] show , for each exploded rotating and non rotating model , the fe core mass ( in solar masses ) , the kinetic energy of the ejecta resulting from the full ejection of the mass above the fe core ( in foes , i.e. in units of @xmath113 erg ) , the mass cut corresponding to the ejection of 0.1 @xmath0of @xmath126 , the ejected mass of each nuclear species , in solar masses , corresponding to the ejection of 0.1 @xmath0of @xmath126 .
the evolution of the chemical composition during the explosion is followed up to a time of @xmath127 , so that all the unstable isotopes with a very short half - life have time to decay in this time interval .
let us eventually mention that , although the yields provided in tables 7 and 8 have been obtained for a specific choice of the mass cut , the full set of cumulative isotopic yields as a function of the mass cut down to the fe core mass are freely available at the web site http://www.iasf-roma.inaf.it/orfeo/public{_}html[http://www.iasf-roma.inaf.it/orfeo/public\{_}html ] .
the comparison between the yields produced by rotating and non rotating models is not straightforward because , as already mentioned above , each single explosion must be tuned and we do not know if the same tuning should be applied to both sets of models . to be clearer , if we tune the explosion of the non rotating 15 @xmath0model in such a way that 0.1 @xmath0of @xmath126 is ejected , which kind of tuning should be used for the corresponding rotating model to make the comparison meaningful ? and which is the role of the initial rotational velocity , i.e. , how the tuning of the explosion should be changed with the initial rotational velocity ?
we do not have an answer and hence we will simply assume that both rotating and non rotating models eject the same amount of @xmath126 .
we are aware of the fact that probably this choice is not the best one , but this is the only approach we find viable at the moment .
figure [ pfni56 ] shows the comparison between the production factors for all the computed models , but the 120 @xmath0 , obtained by assuming all the models to eject 0.1 @xmath35 of @xmath126 .
the eight panels show that the overall differences are not large
. however some interesting odds are present between the two sets of yields . in general
, rotating models tend to overproduce , on average , all the elements , the ones with @xmath128 preferentially the intermediate mass nuclei while the ones with @xmath129 the weak s - process component .
rotating stars with mass larger than 40 @xmath0 , on the contrary , do not show any significant overproduction because the amount of angular momentum injected in these models is not enough to induce sizable effects on the final yields .
a similar result is obtained by comparing the yields averaged over a salpeter initial mass function @xmath107 ( with @xmath108 ) . in this case , however , the differences are less pronounced than in the single mass comparisons ( figure [ pfimf ] ) , although the depletion of li and overproduction of both f and s - process elements obtained in the rotating models is still sizable . as a final comment ,
let us point out that , as already discussed in the previous section , the rotating models are more compact than the non rotating ones , so that the binding energy of their mantle is systematically larger than the one of the corresponding non rotating counterparts . as a consequence , if we would assume for all models ( rotating and not ) the same final kinetic energy of the ejecta , e.g. 1 foe ( @xmath130 ) , we would expect the rotating models to leave more massive remnants and to pollute less efficiently the interstellar medium .
such a possible scenario , however , could be reversed if at least part of the rotational energy could feed the outgoing shock wave : in this case one could have exactly the opposite result .
in this paper we presented a first set of solar metallicity massive star models computed by taking into account the effects of rotation .
the mechanical and thermal distortions induced by rotation have been included in our stellar evolutionary code ( franec ) following the scheme proposed by @xcite and @xcite that can be considered a general approach @xcite .
we also included two rotation driven instabilities , namely the meridional circulation and the shear , following the schemes proposed either by @xcite ( and references therein ) and @xcite ( and references therein ) .
our results are similar to those obtained by other authors in the sense that the basic effects of rotation are similar , for similar initial conditions .
they can be summarized as follows .
rotating stars ( not too close to the break out velocity ) spend most of their core h burning lifetime at lower luminosities and effective temperatures than their non rotating counterparts ; the chemical mixing triggered by both the meridional circulation and the shear leads to larger h convective cores and also to a modification of the surface chemical composition .
since the average luminosity during core h burning is lower and the amount of fuel ( h ) is larger , the core h burning lifetime of the rotating models is longer than that of their non rotating counterparts .
the amount of mass lost by the rotating stars is larger in most cases , with the noticeable exception of the mass interval between 60 and 80 @xmath0 .
the reason for a such a non monotonic effect of rotation on the mass loss is that the mass loss rate provided by @xcite for ob stars is highly discontinuous in the proximity of two critical temperatures ( bi - stability jumps ) , so that even minor changes in the path in the hr diagram may drastically change the amount of mass lost .
two important predictions are the temporal variations of the surface equatorial velocity and of the surface chemical composition ( mainly n and he ) because they have clear observational counterparts . a large program on the vlt with the flames instrument ( pi .
smartt ) , focused on a survey of ob - type stars in the galaxy and the magellanic clouds , started a few years ago .
as part of this large program , a series of papers came out over the years presenting a large database of stars for which a quite detailed surface chemical abundances and projected rotational velocities ( @xmath131 ) are now available @xcite .
the comparison between theoretical predictions and these sample of stars in the `` so called '' hunter diagram ( i.e. @xmath132 versus @xmath131 ) showed a significant discrepancy .
we will not enter in this discussion here because most of the differences concern the magellanic cloud samples and here we presented solar metallicity stars . however , let us stress again that the present inclusion of rotation in a ( any ) evolutionary code is still very phenomenological and hence that the invocation of new exotic unknown phenomena could be untimely .
rotation affects also the he burning phase , both directly and indirectly .
on one side the rotating models , on the average , end the core h burning phase with smaller h rich mantles ( but quite similar he core masses ) than their corresponding non rotating models .
this difference affects _ per se _ the core he burning because a ) increases the time spent as bsg with respect to that spent as rsg , b ) lowers the minimum mass that becomes a wr star , c ) changes the initial final mass relation , d ) changes the relative frequencies of the various core collapse supernova sub - types . on the other side , rotation triggers a partial mixing of matter between the he convective core and the surrounding he core ( basically fresh @xmath133 is brought inside and fresh @xmath101 outside ) and such an occurrence has the double consequence of lowering the amount of @xmath101 left by core he burning and of increasing the co core mass .
moreover , the mixing of freshly synthesized @xmath101 from the he convective core to the more external he- and n - rich zones increases the c / n number ratio within the he core favoring the entrance in the wr - wnc stage when these zones are exposed to the surface . in the present set of rotating models ( @xmath122 ) ,
such a rotationally induced mixing is efficient enough that the number ratio c / n raises above 0.1 in the whole he core , therefore all the stars losing their h - rich envelope show up directly as wr - wnc skipping completely the wne phase . during core he burning
the outward transport of the angular momentum is efficient enough to reduce its internal initial ( at core he ignition ) value by about a fifty per cent . on the contrary , the structural distortions due to rotation during this phase are quite modest , i.e. , both @xmath53 and @xmath54 never decrease significantly below 1 .
the evolution beyond core he exhaustion is characterized by the local conservation of the angular momentum because neither the meridional circulation nor the shear have time to operate any more .
angular momentum is redistributed only in the convective regions where the flattening of the angular velocity is assumed to occur on a dynamical timescale . from core he exhaustion to core si depletion , both @xmath53 and @xmath54 never decrease significantly below 1 , hence rotation does not play a considerable direct role during these stages . the only influence of rotation during these evolutionary phases is indirect , through the increase of the co core mass and the decrease of the @xmath101 mass fraction at core he exhaustion .
both these occurrences have the effect of simulating the behavior of a more massive star , i.e. , the mass - radius relation of any given rotating model at the presupernova stage is substantially steeper than the one of the corresponding non rotating counterpart .
since a steeper mass - radius relation means a larger gravitational binding energy , this result may have interesting consequences for the dynamics of the iron core collapse and fallback .
the influence of rotation on the final yields is difficult to assess because it is not clear how rotating and non rotating exploded models must be compared .
if the comparison is made for the same final kinetic energy of the ejecta , we expect rotating models to leave more massive remnants and to pollute less than their non rotating counterparts ( because of their larger compactness ) .
viceversa , if the comparison is made by assuming the ejection of the same amount of @xmath126 , the chemical composition of the ejecta is quite similar with few differences worth being mentioned : ( 1 ) rotating models with mass in the range 20 - 40 @xmath0tend to overproduce by a factor of @xmath134 all the weak component s - process elements ( cu to sr ) because of their prolonged core he burning phase ; ( 2 ) the same models also show a large overproduction of f , although its production factor remains substantially lower than that of the o ; ( 3 ) the intermediate mass elements ( s to ca ) are slightly enhanced together to o in the lower mass ( @xmath128 ) rotating models .
let us note that the amount of angular momentum locked in the core would lead ( if fully preserved ) to a rotation velocity of the pulsar much larger ( by one order of magnitude or so ) than observed in the present available sample of millisecond pulsars @xcite .
this is a well known problem @xcite .
the presently preferred possible candidate for an additional efficient transport of the angular momentum is the magnetic field : its role would be that of forcing a solid body rotation and hence maximize the efficiency of the outward transport of the angular momentum .
though the magnetic field could certainly play a pivotal role in this respect , our poor knowledge of the details of the collapse of a massive star and the following formation of the remnant leaves room for the possibility that a significant fraction of the angular momentum could be lost just during the collapse or the first phase of formation of the remnant from which a pulsar would emerge .
the strong influence rotation has on the amount of mass lost by a star during its life significantly affects the predicted limiting masses for the formation of the various wr subtypes as well as for the production of the various core collapse supernova types . in a simple ( even if unrealistic ) scenario in which all massive stars are born with the same surface equatorial velocity
, one could easily redefine new limiting masses .
for example , in a generation of massive stars with @xmath122 ( our present models ) , the minimum mass that would show up as a type ib supernova would be of the order of 20 @xmath0 . the corresponding value for a generation of non rotating stars would be of the order of 30 @xmath0 .
also the various limiting masses for the wr subtypes would change as described in the previous sections .
actually we think that this approach is in principle misleading because the main characteristics of rotation is that it varies from one star to another and hence it must be seen as a new `` space parameter '' . in other words rotation
would simply spread all these limiting masses over a range of possible values .
for example , a fast rotating 20 @xmath0may explode as a type ib supernova while another one , born with a smaller amount of angular momentum , may explode as a type ii - l supernova . the same could hold for the initial mass - wr subtypes relation . if the distribution of the angular momentum among the different masses would be proved to follow a general rule , only in that case it would be possible to determine `` average '' new limiting masses . at present we feel safe to conclude that the limiting masses we obtained with and without rotation define the range of possible values for initial rotational velocities lower than 300 km / s . finally , we find that the maximum mass exploding as sniip is @xmath119 , independent of the inclusion of rotation
. the substantial reduction of this quantity with respect to the value obtained up to now ( @xmath135 ) is due to the inclusion , in the present models , of the dust driven mass loss during the rsg stage @xcite which causes the star to lose a lot of mass during core he burning and to eventually enter the wr stage well before the explosion .
this limiting value is in remarkable good agreement with the quite recent determinations of progenitor masses for sniip based on observed archival images provided by @xcite ( and references therein ) which report a value of the order of @xmath136 .
lcccccccccccccccc h & @xmath137 & @xmath138 & @xmath139 & @xmath140 & @xmath141 & @xmath142 & @xmath143 & @xmath144 & @xmath145 & @xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & @xmath151 & @xmath152 + he & @xmath153 & @xmath154 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath159 & @xmath160 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath167 & @xmath168 + c & @xmath169 & @xmath170 & @xmath171 & @xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath146 & @xmath180 & @xmath181 & @xmath182 & @xmath183 + ne & @xmath184 & @xmath185 & @xmath171 & @xmath172 & @xmath173 & @xmath139 & @xmath175 & @xmath176 & @xmath177 & @xmath186 & @xmath179 & @xmath146 & @xmath180 & @xmath181 & @xmath182 & @xmath183 + o & @xmath187 & @xmath188 & @xmath171 & @xmath172 & @xmath173 & @xmath139 & @xmath175 & @xmath176 & @xmath177 & @xmath186 & @xmath179 & @xmath146 & @xmath180 & @xmath181 & @xmath182 & @xmath183 + si & @xmath189 & @xmath190 & @xmath171 & @xmath172 & @xmath173 & @xmath139 & @xmath191 & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath146 & @xmath180 & @xmath181 & @xmath182 & @xmath183 + psn & @xmath192 & & @xmath171 & @xmath172 & @xmath173 & @xmath139 & @xmath191 & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath146 & @xmath180 & @xmath181 & @xmath182 & @xmath183 + h & @xmath193 & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath143 & @xmath199 & @xmath200 & @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath205 & @xmath206 & @xmath207 + he & @xmath208 & @xmath209 & @xmath210 & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 + c & @xmath224 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath138 & @xmath229 & @xmath230 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 + ne & @xmath239 & @xmath240 & @xmath226 & @xmath227 & @xmath228 & @xmath138 & @xmath229 & @xmath241 & @xmath242 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 + o & @xmath243 & @xmath244 & @xmath226 & @xmath227 & @xmath228 & @xmath138 & @xmath229 & @xmath241 & @xmath242 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 + si & @xmath245 & @xmath246 & @xmath226 & @xmath227 & @xmath228 & @xmath138 & @xmath247 & @xmath241 & @xmath242 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 + psn & @xmath248 & & @xmath226 & @xmath227 & @xmath228 & @xmath138 & @xmath247 & @xmath241 & @xmath242 & @xmath249 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 + h & @xmath250 & @xmath251 & @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath143 & @xmath256 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & @xmath261 & @xmath262 & @xmath263 & @xmath264 + he & @xmath265 & @xmath266 & @xmath267 & @xmath268 & @xmath269 & @xmath143 & @xmath253 & @xmath270 & @xmath271 & @xmath202 & @xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 + c & @xmath278 & @xmath143 & @xmath267 & @xmath279 & @xmath280 & @xmath143 & @xmath281 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & @xmath273 & @xmath286 & @xmath287 & @xmath288 & @xmath289 + ne & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath280 & @xmath143 & @xmath281 & @xmath294 & @xmath295 & @xmath296 & @xmath285 & @xmath273 & @xmath286 & @xmath287 & @xmath288 & @xmath289 + o & @xmath264 & @xmath297 & @xmath292 & @xmath298 & @xmath280 & @xmath143 & @xmath281 & @xmath299 & @xmath300 & @xmath301 & @xmath285 & @xmath273 & @xmath286 & @xmath287 & @xmath288 & @xmath289 + si & @xmath302 & @xmath303 & @xmath292 & @xmath293 & @xmath280 & @xmath143 & @xmath281 & @xmath304 & @xmath305 & @xmath296 & @xmath285 & @xmath273 & @xmath286 & @xmath287 & @xmath288 & @xmath289 + psn & @xmath306 & & @xmath292 & @xmath293 & @xmath280 & @xmath143 & @xmath281 & @xmath307 & @xmath308 & @xmath296 & @xmath285 & @xmath273 & @xmath286 & @xmath287 & @xmath288 & @xmath289 + h & @xmath309 & @xmath310 & @xmath311 & @xmath312 & @xmath313 & @xmath314 & @xmath143 & @xmath315 & @xmath316 & @xmath317 & @xmath276 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 + he & @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath143 & @xmath328 & @xmath199 & @xmath329 & @xmath330 & @xmath331 & @xmath273 & @xmath332 & @xmath333 & @xmath334 & @xmath335 + c & @xmath336 & @xmath143 & @xmath337 & @xmath338 & @xmath339 & @xmath143 & @xmath340 & @xmath341 & @xmath342 & @xmath343 & @xmath344 & @xmath273 & @xmath345 & @xmath346 & @xmath347 & @xmath348 + ne & @xmath349 & @xmath350 & @xmath337 & @xmath338 & @xmath339 & @xmath143 & @xmath340 & @xmath341 & @xmath351 & @xmath352 & @xmath344 & @xmath273 & @xmath345 & @xmath346 & @xmath347 & @xmath348 + o & @xmath353 & @xmath354 & @xmath355 & @xmath356 & @xmath339 & @xmath143 & @xmath357 & @xmath358 & @xmath359 & @xmath360 & @xmath344 & @xmath273 & @xmath345 & @xmath346 & @xmath347 & @xmath348 + si & @xmath361 & @xmath362 & @xmath363 & @xmath356 & @xmath339 & @xmath143 & @xmath364 & @xmath365 & @xmath366 & @xmath367 & @xmath344 & @xmath273 & @xmath345 & @xmath346 & @xmath347 & @xmath348 + psn & @xmath368 & & @xmath369 & @xmath370 & @xmath339 & @xmath143 & @xmath364 & @xmath371 & @xmath372 & @xmath373 & @xmath344 & @xmath273 & @xmath345 & @xmath346 & @xmath347 & @xmath348 + h & @xmath374 & @xmath375 & @xmath376 & @xmath377 & @xmath378 & @xmath379 & @xmath143 & @xmath380 & @xmath381 & @xmath382 & @xmath383 & @xmath384 & @xmath385 & @xmath386 & @xmath387 & @xmath388 + he & @xmath389 & @xmath390 & @xmath391 & @xmath392 & @xmath141 & @xmath143 & @xmath393 & @xmath394 & @xmath395 & @xmath352 & @xmath396 & @xmath273 & @xmath345 & @xmath397 & @xmath398 & @xmath399 + c & @xmath400 & @xmath143 & @xmath312 & @xmath401 & @xmath402 & @xmath143 & @xmath403 & @xmath404 & @xmath405 & @xmath373 & @xmath406 & @xmath273 & @xmath407 & @xmath408 & @xmath409 & @xmath410 + ne & @xmath411 & @xmath412 & @xmath413 & @xmath414 & @xmath402 & @xmath143 & @xmath415 & @xmath416 & @xmath417 & @xmath418 & @xmath406 & @xmath273 & @xmath407 & @xmath408 & @xmath409 & @xmath410 + o & @xmath419 & @xmath354 & @xmath413 & @xmath414 & @xmath402 & @xmath143 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & @xmath406 & @xmath273 & @xmath407 & @xmath408 & @xmath409 & @xmath410 + si & @xmath424 & @xmath425 & @xmath413 & @xmath414 & @xmath402 & @xmath143 & @xmath420 & @xmath426 & @xmath427 & @xmath423 & @xmath406 & @xmath273 & @xmath407 & @xmath408 & @xmath409 & @xmath410 + psn & @xmath150 & & @xmath413 & @xmath414 & @xmath402 & @xmath143 & @xmath420 & @xmath428 & @xmath429 & @xmath430 & @xmath431 & @xmath273 & @xmath407 & @xmath408 & @xmath409 & @xmath410 + h & @xmath432 & @xmath433 & @xmath434 & @xmath435 & @xmath436 & @xmath437 & @xmath143 & @xmath438 & @xmath439 & @xmath440 & @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath445 & @xmath446 + he & @xmath447 & @xmath448 & @xmath337 & @xmath401 & @xmath449 & @xmath143 & @xmath450 & @xmath451 & @xmath452 & @xmath430 & @xmath453 & @xmath273 & @xmath454 & @xmath455 & @xmath456 & @xmath457 + c & @xmath458 & @xmath143 & @xmath293 & @xmath459 & @xmath460 & @xmath143 & @xmath461 & @xmath462 & @xmath463 & @xmath464 & @xmath465 & @xmath273 & @xmath466 & @xmath467 & @xmath468 & @xmath469 + ne & @xmath470 & @xmath471 & @xmath293 & @xmath459 & @xmath460 & @xmath143 & @xmath461 & @xmath472 & @xmath473 & @xmath176 & @xmath474 & @xmath273 & @xmath466 & @xmath467 & @xmath468 & @xmath469 + o & @xmath475 & @xmath476 & @xmath293 & @xmath459 & @xmath460 & @xmath143 & @xmath461 & @xmath477 & @xmath478 & @xmath464 & @xmath474 & @xmath273 & @xmath466 & @xmath467 & @xmath468 & @xmath469 + si & @xmath479 & @xmath480 & @xmath279 & @xmath459 & @xmath460 & @xmath143 & @xmath461 & @xmath481 & @xmath482 & @xmath483 & @xmath474 & @xmath273 & @xmath466 & @xmath467 & @xmath468 & @xmath469 + psn & @xmath484 & & @xmath279 & @xmath459 & @xmath460 & @xmath143 & @xmath461 & @xmath485 & @xmath486 & @xmath483 & @xmath474 & @xmath273 & @xmath466 & @xmath467 & @xmath468 & @xmath469 + h & @xmath487 & @xmath488 & @xmath489 & @xmath490 & @xmath491 & @xmath492 & @xmath143 & @xmath493 & @xmath494 & @xmath495 & @xmath496 & @xmath497 & @xmath498 & @xmath499 & @xmath500 & @xmath501 + he & @xmath502 & @xmath503 & @xmath355 & @xmath504 & @xmath505 & @xmath143 & @xmath506 & @xmath507 & @xmath508 & @xmath509 & @xmath510 & @xmath273 & @xmath511 & @xmath512 & @xmath513 & @xmath514 + c & @xmath515 & @xmath143 & @xmath516 & @xmath517 & @xmath505 & @xmath143 & @xmath449 & @xmath518 & @xmath519 & @xmath520 & @xmath521 & @xmath273 & @xmath160 & @xmath522 & @xmath523 & @xmath524 + ne & @xmath525 & @xmath526 & @xmath516 & @xmath517 & @xmath505 & @xmath143 & @xmath527 & @xmath528 & @xmath529 & @xmath530 & @xmath521 & @xmath273 & @xmath160 & @xmath522 & @xmath523 & @xmath524 + o & @xmath531 & @xmath532 & @xmath516 & @xmath517 & @xmath505 & @xmath143 & @xmath527 & @xmath533 & @xmath534 & @xmath530 & @xmath535 & @xmath273 & @xmath160 & @xmath522 & @xmath523 & @xmath524 + si & @xmath536 & @xmath537 & @xmath516 & @xmath517 & @xmath505 & @xmath143 & @xmath527 & @xmath538 & @xmath539 & @xmath540 & @xmath535 & @xmath273 & @xmath160 & @xmath522 & @xmath523 & @xmath524 + psn & @xmath541 & & @xmath516 & @xmath517 & @xmath505 & @xmath143 & @xmath527 & @xmath533 & @xmath542 & @xmath530 & @xmath535 & @xmath273 & @xmath160 & @xmath522 & @xmath523 & @xmath524 + h & @xmath543 & @xmath544 & @xmath545 & @xmath546 & @xmath547 & @xmath548 & @xmath143 & @xmath549 & @xmath550 & @xmath241 & @xmath551 & @xmath498 & @xmath552 & @xmath553 & @xmath554 & @xmath555 + he & @xmath556 & @xmath557 & @xmath363 & @xmath490 & @xmath558 & @xmath143 & @xmath559 & @xmath358 & @xmath560 & @xmath561 & @xmath296 & @xmath273 & @xmath562 & @xmath563 & @xmath564 & @xmath565 + c & @xmath566 & @xmath143 & @xmath138 & @xmath567 & @xmath568 & @xmath143 & @xmath569 & @xmath570 & @xmath571 & @xmath373 & @xmath147 & @xmath273 & @xmath572 & @xmath573 & @xmath574 & @xmath575 + ne & @xmath576 & @xmath143 & @xmath577 & @xmath578 & @xmath568 & @xmath143 & @xmath569 & @xmath579 & @xmath580 & @xmath581 & @xmath147 & @xmath273 & @xmath572 & @xmath573 & @xmath574 & @xmath575 + o & @xmath582 & @xmath583 & @xmath577 & @xmath578 & @xmath568 & @xmath143 & @xmath569 & @xmath485 & @xmath584 & @xmath581 & @xmath585 & @xmath273 & @xmath572 & @xmath573 & @xmath574 & @xmath575 + si & @xmath586 & @xmath587 & @xmath577 & @xmath578 & @xmath568 & @xmath143 & @xmath569 & @xmath579 & @xmath588 & @xmath581 & @xmath585 & @xmath273 & @xmath572 & @xmath573 & @xmath574 & @xmath575 + psn & @xmath589 & & @xmath577 & @xmath578 & @xmath568 & @xmath143 & @xmath569 & @xmath462 & @xmath590 & @xmath418 & @xmath585 & @xmath273 & @xmath572 & @xmath573 & @xmath574 & @xmath575 + h & @xmath591 & @xmath592 & @xmath593 & @xmath594 & @xmath595 & @xmath596 & @xmath143 & @xmath597 & @xmath598 & @xmath599 & @xmath600 & @xmath468 & @xmath601 & @xmath602 & @xmath603 & @xmath604 + he & @xmath605 & @xmath606 & @xmath363 & @xmath490 & @xmath558 & @xmath143 & @xmath607 & @xmath608 & @xmath609 & @xmath610 & @xmath611 & @xmath273 & @xmath612 & @xmath613 & @xmath614 & @xmath615 + c & @xmath616 & @xmath143 & @xmath138 & @xmath578 & @xmath568 & @xmath143 & @xmath617 & @xmath618 & @xmath619 & @xmath620 & @xmath621 & @xmath273 & @xmath622 & @xmath623 & @xmath624 & @xmath625 + ne & @xmath626 & @xmath143 & @xmath577 & @xmath627 & @xmath568 & @xmath143 & @xmath617 & @xmath628 & @xmath629 & @xmath630 & @xmath621 & @xmath273 & @xmath622 & @xmath623 & @xmath624 & @xmath625 + o & @xmath631 & @xmath532 & @xmath577 & @xmath627 & @xmath568 & @xmath143 & @xmath617 & @xmath632 & @xmath351 & @xmath633 & @xmath621 & @xmath273 & @xmath622 & @xmath623 & @xmath624 & @xmath625 + si & @xmath634 & @xmath635 & @xmath577 & @xmath627 & @xmath568 & @xmath143 & @xmath617 & @xmath636 & @xmath395 & @xmath631 & @xmath621 & @xmath273 & @xmath622 & @xmath623 & @xmath624 & @xmath625 + psn & @xmath637 & & @xmath577 & @xmath594 & @xmath568 & @xmath143 & @xmath617 & @xmath638 & @xmath639 & @xmath640 & @xmath621 & @xmath273 & @xmath622 & @xmath623 & @xmath624 & @xmath625 lcccccccccccc h & @xmath641 & @xmath356 & @xmath642 & @xmath643 & @xmath644 & @xmath645 & @xmath143 & @xmath646 & @xmath647 & @xmath648 & @xmath649 & @xmath650 + he & @xmath651 & @xmath652 & @xmath653 & @xmath654 & @xmath655 & @xmath656 & @xmath657 & @xmath658 & @xmath659 & @xmath660 & @xmath661 & @xmath662 + c & @xmath663 & @xmath664 & @xmath665 & @xmath666 & @xmath667 & @xmath668 & @xmath669 & @xmath658 & @xmath659 & @xmath660 & @xmath661 & @xmath662 + ne & @xmath670 & @xmath671 & @xmath665 & @xmath666 & @xmath667 & @xmath668 & @xmath672 & @xmath658 & @xmath659 & @xmath660 & @xmath661 & @xmath662 + o & @xmath673 & @xmath674 & @xmath665 & @xmath666 & @xmath667 & @xmath668 & @xmath672 & @xmath658 & @xmath659 & @xmath660 & @xmath661 & @xmath662 + si & @xmath675 & @xmath676 & @xmath665 & @xmath666 & @xmath667 & @xmath668 & @xmath672 & @xmath658 & @xmath659 & @xmath660 & @xmath661 & @xmath662 + psn & @xmath677 & & @xmath665 & @xmath666 & @xmath667 & @xmath668 & @xmath678 & @xmath658 & @xmath659 & @xmath660 & @xmath661 & @xmath662 + h & @xmath679 & @xmath680 & @xmath311 & @xmath681 & @xmath682 & @xmath683 & @xmath143 & @xmath646 & @xmath647 & @xmath648 & @xmath649 & @xmath650 + he & @xmath684 & @xmath685 & @xmath665 & @xmath686 & @xmath687 & @xmath688 & @xmath689 & @xmath690 & @xmath691 & @xmath692 & @xmath693 & @xmath694 + c & @xmath695 & @xmath696 & @xmath155 & @xmath697 & @xmath698 & @xmath688 & @xmath699 & @xmath690 & @xmath691 & @xmath692 & @xmath693 & @xmath694 + ne & @xmath700 & @xmath701 & @xmath155 & @xmath697 & @xmath698 & @xmath688 & @xmath702 & @xmath690 & @xmath691 & @xmath692 & @xmath693 & @xmath694 + o & @xmath703 & @xmath704 & @xmath155 & @xmath697 & @xmath698 & @xmath688 & @xmath702 & @xmath690 & @xmath691 & @xmath692 & @xmath693 & @xmath694 + si & @xmath261 & @xmath587 & @xmath155 & @xmath697 & @xmath698 & @xmath688 & @xmath702 & @xmath690 & @xmath691 & @xmath692 & @xmath693 & @xmath694 + psn & @xmath523 & & @xmath155 & @xmath697 & @xmath698 & @xmath688 & @xmath702 & @xmath690 & @xmath691 & @xmath692 & @xmath693 & @xmath694 + h & @xmath705 & @xmath706 & @xmath376 & @xmath707 & @xmath708 & @xmath709 & @xmath143 & @xmath646 & @xmath647 & @xmath648 & @xmath649 & @xmath650 + he & @xmath710 & @xmath711 & @xmath711 & @xmath712 & @xmath713 & @xmath714 & @xmath715 & @xmath716 & @xmath717 & @xmath718 & @xmath719 & @xmath720 + c & @xmath721 & @xmath471 & @xmath722 & @xmath355 & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727 & @xmath728 & @xmath729 & @xmath730 + ne & @xmath731 & @xmath732 & @xmath722 & @xmath355 & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727 & @xmath728 & @xmath729 & @xmath730 + o & @xmath733 & @xmath734 & @xmath722 & @xmath355 & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727 & @xmath728 & @xmath729 & @xmath730 + si & @xmath735 & @xmath635 & @xmath722 & @xmath355 & @xmath723 & @xmath724 & @xmath736 & @xmath726 & @xmath727 & @xmath728 & @xmath729 & @xmath730 + psn & @xmath737 & & @xmath722 & @xmath355 & @xmath723 & @xmath724 & @xmath736 & @xmath726 & @xmath727 & @xmath728 & @xmath729 & @xmath730 + h & @xmath738 & @xmath739 & @xmath740 & @xmath741 & @xmath742 & @xmath743 & @xmath143 & @xmath646 & @xmath647 & @xmath648 & @xmath649 & @xmath650 + he & @xmath744 & @xmath745 & @xmath746 & @xmath213 & @xmath747 & @xmath748 & @xmath312 & @xmath749 & @xmath750 & @xmath751 & @xmath752 & @xmath753 + c & @xmath754 & @xmath755 & @xmath756 & @xmath757 & @xmath758 & @xmath748 & @xmath759 & @xmath760 & @xmath761 & @xmath751 & @xmath762 & @xmath763 + ne & @xmath764 & @xmath765 & @xmath756 & @xmath757 & @xmath758 & @xmath748 & @xmath370 & @xmath760 & @xmath761 & @xmath751 & @xmath762 & @xmath763 + o & @xmath766 & @xmath676 & @xmath756 & @xmath757 & @xmath758 & @xmath767 & @xmath370 & @xmath760 & @xmath761 & @xmath751 & @xmath762 & @xmath763 + si & @xmath768 & @xmath635 & @xmath756 & @xmath757 & @xmath758 & @xmath748 & @xmath370 & @xmath760 & @xmath761 & @xmath751 & @xmath762 & @xmath763 + psn & @xmath769 & & @xmath756 & @xmath757 & @xmath758 & @xmath748 & @xmath370 & @xmath760 & @xmath761 & @xmath751 & @xmath762 & @xmath763 + h & @xmath770 & @xmath771 & @xmath376 & @xmath213 & @xmath772 & @xmath706 & @xmath143 & @xmath646 & @xmath647 & @xmath648 & @xmath649 & @xmath650 + he & @xmath773 & @xmath328 & @xmath774 & @xmath281 & @xmath644 & @xmath775 & @xmath776 & @xmath319 & @xmath318 & @xmath777 & @xmath778 & @xmath779 + c & @xmath780 & @xmath781 & @xmath782 & @xmath783 & @xmath559 & @xmath784 & @xmath785 & @xmath204 & @xmath203 & @xmath786 & @xmath787 & @xmath788 + ne & @xmath789 & @xmath765 & @xmath736 & @xmath783 & @xmath559 & @xmath784 & @xmath790 & @xmath204 & @xmath203 & @xmath786 & @xmath787 & @xmath788 + o & @xmath791 & @xmath792 & @xmath793 & @xmath783 & @xmath559 & @xmath784 & @xmath790 & @xmath204 & @xmath203 & @xmath786 & @xmath787 & @xmath788 + si & @xmath794 & @xmath795 & @xmath796 & @xmath783 & @xmath559 & @xmath784 & @xmath790 & @xmath204 & @xmath203 & @xmath786 & @xmath787 & @xmath788 + psn & @xmath797 & & @xmath798 & @xmath783 & @xmath559 & @xmath784 & @xmath790 & @xmath204 & @xmath203 & @xmath786 & @xmath787 & @xmath788 + h & @xmath799 & @xmath800 & @xmath311 & @xmath709 & @xmath801 & @xmath802 & @xmath143 & @xmath646 & @xmath647 & @xmath648 & @xmath649 & @xmath650 + he & @xmath803 & @xmath804 & @xmath712 & @xmath805 & @xmath806 & @xmath143 & @xmath807 & @xmath273 & @xmath808 & @xmath809 & @xmath810 & @xmath811 + c & @xmath812 & @xmath143 & @xmath267 & @xmath813 & @xmath814 & @xmath143 & @xmath815 & @xmath273 & @xmath808 & @xmath809 & @xmath816 & @xmath817 + ne & @xmath818 & @xmath471 & @xmath292 & @xmath255 & @xmath814 & @xmath143 & @xmath815 & @xmath273 & @xmath808 & @xmath809 & @xmath816 & @xmath817 + o & @xmath819 & @xmath820 & @xmath292 & @xmath255 & @xmath814 & @xmath143 & @xmath815 & @xmath273 & @xmath808 & @xmath809 & @xmath816 & @xmath817 + si & @xmath821 & @xmath795 & @xmath292 & @xmath255 & @xmath814 & @xmath143 & @xmath815 & @xmath273 & @xmath808 & @xmath809 & @xmath816 & @xmath817 + psn & @xmath822 & & @xmath292 & @xmath255 & @xmath814 & @xmath143 & @xmath815 & @xmath273 & @xmath808 & @xmath809 & @xmath816 & @xmath817 + h & @xmath823 & @xmath824 & @xmath825 & @xmath517 & @xmath826 & @xmath827 & @xmath143 & @xmath828 & @xmath829 & @xmath751 & @xmath830 & @xmath831 + he & @xmath832 & @xmath503 & @xmath355 & @xmath459 & @xmath833 & @xmath143 & @xmath834 & @xmath273 & @xmath835 & @xmath836 & @xmath837 & @xmath838 + c & @xmath839 & @xmath143 & @xmath293 & @xmath840 & @xmath841 & @xmath143 & @xmath842 & @xmath273 & @xmath843 & @xmath844 & @xmath845 & @xmath846 + ne & @xmath847 & @xmath848 & @xmath298 & @xmath849 & @xmath841 & @xmath143 & @xmath842 & @xmath273 & @xmath843 & @xmath844 & @xmath845 & @xmath846 + o & @xmath850 & @xmath851 & @xmath298 & @xmath852 & @xmath841 & @xmath143 & @xmath853 & @xmath273 & @xmath843 & @xmath844 & @xmath845 & @xmath846 + si & @xmath536 & @xmath587 & @xmath298 & @xmath852 & @xmath841 & @xmath143 & @xmath853 & @xmath273 & @xmath843 & @xmath844 & @xmath845 & @xmath846 + psn & @xmath854 & & @xmath855 & @xmath852 & @xmath841 & @xmath143 & @xmath853 & @xmath273 & @xmath843 & @xmath844 & @xmath845 & @xmath846 + h & @xmath856 & @xmath857 & @xmath858 & @xmath546 & @xmath859 & @xmath860 & @xmath143 & @xmath861 & @xmath862 & @xmath751 & @xmath863 & @xmath864 + he & @xmath865 & @xmath866 & @xmath363 & @xmath490 & @xmath867 & @xmath143 & @xmath868 & @xmath273 & @xmath869 & @xmath870 & @xmath871 & @xmath872 + c & @xmath873 & @xmath874 & @xmath516 & @xmath567 & @xmath558 & @xmath143 & @xmath617 & @xmath273 & @xmath875 & @xmath876 & @xmath877 & @xmath878 + ne & @xmath879 & @xmath143 & @xmath138 & @xmath578 & @xmath558 & @xmath143 & @xmath617 & @xmath273 & @xmath875 & @xmath876 & @xmath877 & @xmath878 + o & @xmath880 & @xmath881 & @xmath138 & @xmath627 & @xmath558 & @xmath143 & @xmath617 & @xmath273 & @xmath875 & @xmath876 & @xmath877 & @xmath878 + si & @xmath882 & @xmath883 & @xmath138 & @xmath627 & @xmath558 & @xmath143 & @xmath617 & @xmath273 & @xmath875 & @xmath876 & @xmath877 & @xmath878 + psn & @xmath884 & & @xmath577 & @xmath627 & @xmath558 & @xmath143 & @xmath617 & @xmath273 & @xmath875 & @xmath876 & @xmath877 & @xmath878 + h & @xmath885 & @xmath886 & @xmath887 & @xmath888 & @xmath889 & @xmath890 & @xmath143 & @xmath891 & @xmath892 & @xmath893 & @xmath894 & @xmath895 + he & @xmath896 & @xmath897 & @xmath369 & @xmath898 & @xmath899 & @xmath143 & @xmath900 & @xmath273 & @xmath901 & @xmath902 & @xmath903 & @xmath904 + c & @xmath905 & @xmath143 & @xmath516 & @xmath906 & @xmath907 & @xmath143 & @xmath908 & @xmath273 & @xmath909 & @xmath910 & @xmath911 & @xmath912 + ne & @xmath913 & @xmath143 & @xmath516 & @xmath914 & @xmath907 & @xmath143 & @xmath908 & @xmath273 & @xmath909 & @xmath910 & @xmath911 & @xmath912 + o & @xmath915 & @xmath916 & @xmath516 & @xmath906 & @xmath907 & @xmath143 & @xmath908 & @xmath273 & @xmath909 & @xmath910 & @xmath911 & @xmath912 + si & @xmath917 & @xmath918 & @xmath516 & @xmath906 & @xmath907 & @xmath143 & @xmath908 & @xmath273 & @xmath909 & @xmath910 & @xmath911 & @xmath912 + psn & @xmath919 & & @xmath516 & @xmath906 & @xmath907 & @xmath143 & @xmath908 & @xmath273 & @xmath909 & @xmath910 & @xmath911 & @xmath912 ccccccccccc 13 & & & & & & & & & & + 15 & @xmath920 & & & & & & & & & + 20 & @xmath921 & @xmath922 & @xmath923 & he(0.29 ) & & & @xmath924 & he(0.02 ) & & + 25 & @xmath925 & @xmath926 & @xmath927 & he(0.12 ) & & & @xmath928 & he(0.01 ) & & + 30 & @xmath929 & @xmath930 & @xmath931 & he(0.25 ) & & & @xmath932 & he(0.01 ) & & + 40 & @xmath933 & @xmath934 & @xmath935 & he(0.74 ) & & & @xmath936 & he(0.17 ) & @xmath937 & he(0.10 ) + 60 & @xmath938 & @xmath939 & @xmath926 & he(0.95 ) & & & @xmath940 & he(0.48 ) & @xmath941 & he(0.45 ) + 80 & @xmath942 & @xmath943 & @xmath944 & he(0.99 ) & & & @xmath945 & he(0.70 ) & @xmath935 & he(0.65 ) + 120 & @xmath946 & @xmath947 & @xmath948 & h(0.15 ) & @xmath949 & he(0.74 ) & @xmath950 & he(0.72 ) & @xmath951 & he(0.58 ) ccccccccccc 13 & & & & & & & & & & + 15 & @xmath952 & & & & & & & & & + 20 & @xmath953 & & & & & & & & & + 25 & @xmath954 & @xmath955 & @xmath955 & adv . & & & & & & + 30 & @xmath956 & @xmath957 & @xmath957 & he(0.12 ) & & & & & & + 40 & @xmath958 & @xmath959 & @xmath960 & he(0.30 ) & @xmath961 & he(0.02 ) & & & & + 60 & @xmath962 & @xmath963 & @xmath964 & he(0.97 ) & @xmath965 & he(0.67 ) & @xmath966 & he(0.20 ) & @xmath967 & he(0.16 ) + 80 & @xmath968 & @xmath969 & @xmath970 & h(0.01 ) & @xmath971 & he(0.74 ) & @xmath972 & he(0.38 ) & @xmath973 & he(0.32 ) + 120 & @xmath974 & @xmath975 & @xmath976 & h(0.08 ) & @xmath977 & he(0.81 ) & @xmath978 & he(0.55 ) & @xmath979 & he(0.54 ) cccccc 13 & 3.18 & 2.77 & rsg & sniip + 15 & 1.80 & 2.18 & rsg & sniip + 20 & 0.00 & 0.74 & wnc & snib + 25 & 0.00 & 0.76 & wnc & snib + 30 & 0.00 & 0.78 & wnc & snib + 40 & 0.00 & 0.57 & wco & snib + 60 & 0.00 & 0.45 & wco & snib + 80 & 0.00 & 0.45 & wco & snib + 120 & 0.00 & 0.63 & wco & snib cccccc 13 & 5.46 & 4.49 & rsg & sniip + 15 & 5.76 & 4.91 & rsg & sniip + 20 & 0.26 & 3.83 & rsg & sniib + 25 & 0.16 & 4.45 & ysg & sniib + 30 & 0.03 & 4.33 & ysg & snib + 40 & 0.00 & 4.20 & wne & snib + 60 & 0.00 & 0.92 & wco & snib + 80 & 0.00 & 0.80 & wco & snib + 120 & 0.00 & 0.54 & wco & snib lccccccccc @xmath980 & 1.73 & 1.85 & 1.83 & 2.03 & 2.14 & 2.00 & 2.27 & 2.17 & 1.80 + @xmath981 & 1.00 & 1.00 & 1.17 & 1.91 & 2.05 & 2.21 & 7.39 & 6.55 & 2.91 + @xmath982 & 1.84 & 2.02 & 1.92 & 2.42 & 2.37 & 2.33 & 2.87 & 3.11 & 2.68 + @xmath983 & @xmath984 & @xmath985 & @xmath986 & @xmath987 & @xmath988 & @xmath989 & @xmath990 & @xmath991 & @xmath992 + @xmath993 & @xmath994 & @xmath995 & @xmath996 & @xmath997 & @xmath998 & @xmath999 & @xmath1000 & @xmath1001 & @xmath1002 + @xmath1003 & @xmath1004 & @xmath1005 & @xmath1006 & @xmath1007 & @xmath1008 & @xmath1009 & @xmath1010 & @xmath1011 & @xmath1012 + @xmath1013 & @xmath1014 & @xmath1015 & @xmath1016 & @xmath1017 & @xmath1018 & @xmath1019 & @xmath1020 & @xmath1021 & @xmath1022 + @xmath1023 & @xmath1024 & @xmath1025 & @xmath1026 & @xmath1027 & @xmath1028 & @xmath1029 & @xmath1030 & @xmath1031 & @xmath1032 + @xmath1033 & @xmath1034 & @xmath1035 & @xmath1036 & @xmath1037 & @xmath1038 & @xmath1039 & @xmath1040 & @xmath1041 & @xmath1042 + @xmath1043 & @xmath1044 & @xmath1045 & @xmath1046 & @xmath1047 & @xmath1048 & @xmath1049 & @xmath1050 & @xmath1051 & @xmath1052 + @xmath1053 & @xmath1054 & @xmath1055 & @xmath1056 & @xmath1057 & @xmath1058 & @xmath1059 & @xmath1060 & @xmath1061 & @xmath1062 + @xmath1063 & @xmath1064 & @xmath1065 & @xmath1066 & @xmath1067 & @xmath1068 & @xmath1069 & @xmath1070 & @xmath1071 & @xmath1072 + @xmath1073 & @xmath1074 & @xmath1075 & @xmath1076 & @xmath1077 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath1081 & @xmath1082 + @xmath1083 & @xmath1084 & @xmath1085 & @xmath1086 & @xmath1087 & @xmath1088 & @xmath1089 & @xmath1090 & @xmath1091 & @xmath1092 + @xmath1093 & @xmath1094 & @xmath1095 & @xmath1096 & @xmath1097 & @xmath1098 & @xmath1099 & @xmath1100 & @xmath1101 & @xmath1102 + @xmath1103 & @xmath1104 & @xmath1105 & @xmath1106 & @xmath1107 & @xmath1108 & @xmath1109 & @xmath1110 & @xmath1111 & @xmath1112 + @xmath1113 & @xmath1114 & @xmath1115 & @xmath1116 & @xmath1117 & @xmath1118 & @xmath1119 & @xmath1120 & @xmath1121 & @xmath1122 + @xmath1123 & @xmath1124 & @xmath1125 & @xmath1126 & @xmath1127 & @xmath1128 & @xmath1129 & @xmath1130 & @xmath1131 & @xmath1132 + @xmath1133 & @xmath1134 & @xmath1135 & @xmath1136 & @xmath1137 & @xmath1138 & @xmath1139 & @xmath1140 & @xmath1065 & @xmath1141 + @xmath1142 & @xmath1143 & @xmath1144 & @xmath1145 & @xmath1146 & @xmath1147 & @xmath1148 & @xmath1149 & @xmath1150 & @xmath1151 + @xmath1152 & @xmath1153 & @xmath1154 & @xmath1155 & @xmath1156 & @xmath1157 & @xmath1158 & @xmath1159 & @xmath1160 & @xmath1161 + @xmath1162 & @xmath1163 & @xmath1164 & @xmath1165 & @xmath1166 & @xmath1167 & @xmath1168 & @xmath1169 & @xmath1170 & @xmath1171 + @xmath1172 & @xmath1173 & @xmath1174 & @xmath1175 & @xmath1176 & @xmath1177 & @xmath1178 & @xmath1092 & @xmath1179 & @xmath1180 + @xmath1181 & @xmath1182 & @xmath1183 & @xmath1184 & @xmath1185 & @xmath1186 & @xmath1187 & @xmath1188 & @xmath1189 & @xmath1190 + @xmath1191 & @xmath1192 & @xmath1193 & @xmath1194 & @xmath1195 & @xmath1193 & @xmath1196 & @xmath1197 & @xmath1198 & @xmath1199 + @xmath1200 & @xmath1201 & @xmath1202 & @xmath1203 & @xmath1204 & @xmath1205 & @xmath1206 & @xmath1207 & @xmath1208 & @xmath1209 + @xmath1210 & @xmath1211 & @xmath1212 & @xmath1213 & @xmath1214 & @xmath1215 & @xmath1216 & @xmath1217 & @xmath1218 & @xmath1219 + @xmath1220 & @xmath1221 & @xmath1222 & @xmath1223 & @xmath1224 & @xmath999 & @xmath1225 & @xmath1226 & @xmath1227 & @xmath1228 + @xmath1229 & @xmath1230 & @xmath1231 & @xmath1232 & @xmath1233 & @xmath1234 & @xmath1235 & @xmath1236 & @xmath1237 & @xmath1238 + @xmath1239 & @xmath1240 & @xmath1241 & @xmath1242 & @xmath1243 & @xmath1244 & @xmath1245 & @xmath1125 & @xmath1246 & @xmath1247 + @xmath1248 & @xmath1249 & @xmath1250 & @xmath1251 & @xmath1252 & @xmath1253 & @xmath1254 & @xmath1255 & @xmath1256 & @xmath1257 + @xmath1258 & @xmath1259 & @xmath1260 & @xmath1261 & @xmath1262 & @xmath1263 & @xmath1264 & @xmath1265 & @xmath1266 & @xmath1267 + @xmath1268 & @xmath1269 & @xmath1270 & @xmath1271 & @xmath1272 & @xmath1273 & @xmath1274 & @xmath1275 & @xmath1276 & @xmath1277 + @xmath1278 & @xmath1279 & @xmath1280 & @xmath1281 & @xmath1282 & @xmath1283 & @xmath1284 & @xmath1285 & @xmath1286 & @xmath1287 + @xmath1288 & @xmath1289 & @xmath1290 & @xmath1291 & @xmath1292 & @xmath1293 & @xmath1294 & @xmath1295 & @xmath1296 & @xmath1297 + @xmath1298 & @xmath1299 & @xmath1300 & @xmath1301 & @xmath1302 & @xmath1303 & @xmath1304 & @xmath1305 & @xmath1306 & @xmath1307 + @xmath1308 & @xmath1309 & @xmath1310 & @xmath1311 & @xmath1312 & @xmath1313 & @xmath1314 & @xmath1315 & @xmath1316 & @xmath1317 + @xmath1318 & @xmath1319 & @xmath1271 & @xmath1320 & @xmath1321 & @xmath1322 & @xmath1323 & @xmath1324 & @xmath1325 & @xmath1326 + @xmath1327 & @xmath1328 & @xmath1329 & @xmath1330 & @xmath1331 & @xmath1332 & @xmath1333 & @xmath1334 & @xmath1335 & @xmath1336 + @xmath1337 & @xmath1338 & @xmath1339 & @xmath1340 & @xmath1341 & @xmath1342 & @xmath1343 & @xmath1344 & @xmath1345 & @xmath1346 + @xmath1347 & @xmath1348 & @xmath1349 & @xmath1350 & @xmath1351 & @xmath1352 & @xmath1353 & @xmath1354 & @xmath1355 & @xmath1356 + @xmath1357 & @xmath1358 & @xmath1359 & @xmath1360 & @xmath1361 & @xmath1362 & @xmath1363 & @xmath1364 & @xmath1365 & @xmath1366 + @xmath1367 & @xmath1368 & @xmath1369 & @xmath1370 & @xmath1371 & @xmath1372 & @xmath1286 & @xmath1373 & @xmath1374 & @xmath1374 + @xmath1375 & @xmath1376 & @xmath1377 & @xmath1378 & @xmath1379 & @xmath1380 & @xmath1381 & @xmath1382 & @xmath1383 & @xmath1384 + @xmath1385 & @xmath1386 & @xmath1387 & @xmath1388 & @xmath1389 & @xmath1390 & @xmath1391 & @xmath1392 & @xmath1393 & @xmath1394 + @xmath1395 & @xmath1396 & @xmath1397 & @xmath1398 & @xmath1399 & @xmath1400 & @xmath1401 & @xmath1402 & @xmath1403 & @xmath1404 + @xmath1405 & @xmath1406 & @xmath1407 & @xmath1408 & @xmath1409 & @xmath1410 & @xmath1411 & @xmath1412 & @xmath1413 & @xmath1414 + @xmath1415 & @xmath1416 & @xmath1417 & @xmath1418 & @xmath1419 & @xmath1420 & @xmath1421 & @xmath1422 & @xmath1423 & @xmath1424 + @xmath1425 & @xmath1426 & @xmath1427 & @xmath1428 & @xmath1429 & @xmath1430 & @xmath1431 & @xmath1432 & @xmath1433 & @xmath1434 + @xmath1435 & @xmath1436 & @xmath1437 & @xmath1438 & @xmath1439 & @xmath1440 & @xmath1441 & @xmath1442 & @xmath1443 & @xmath1444 + @xmath1445 & @xmath1446 & @xmath1447 & @xmath1448 & @xmath1449 & @xmath1361 & @xmath1450 & @xmath1451 & @xmath1452 & @xmath1453 + @xmath1454 & @xmath1455 & @xmath1456 & @xmath1457 & @xmath1458 & @xmath1459 & @xmath1276 & @xmath1460 & @xmath1461 & @xmath1462 + @xmath1463 & @xmath1464 & @xmath1465 & @xmath1466 & @xmath1467 & @xmath1468 & @xmath1469 & @xmath1470 & @xmath1471 & @xmath1472 + @xmath1473 & @xmath1474 & @xmath1475 & @xmath1476 & @xmath1477 & @xmath1478 & @xmath1479 & @xmath1480 & @xmath1481 & @xmath1482 + @xmath1483 & @xmath1484 & @xmath1485 & @xmath1212 & @xmath1486 & @xmath1487 & @xmath1488 & @xmath1489 & @xmath1490 & @xmath1491 + @xmath1492 & @xmath1493 & @xmath1494 & @xmath1495 & @xmath1496 & @xmath1497 & @xmath1498 & @xmath1499 & @xmath1500 & @xmath1501 + @xmath1502 & @xmath1503 & @xmath1504 & @xmath1505 & @xmath1506 & @xmath1507 & @xmath1508 & @xmath1509 & @xmath1510 & @xmath1511 + @xmath1512 & @xmath1513 & @xmath1514 & @xmath1515 & @xmath1516 & @xmath1517 & @xmath1518 & @xmath1519 & @xmath1520 & @xmath1521 + @xmath1522 & @xmath1523 & @xmath1524 & @xmath1525 & @xmath1468 & @xmath1526 & @xmath1527 & @xmath1528 & @xmath1529 & @xmath1530 + @xmath1531 & @xmath1532 & @xmath1533 & @xmath1534 & @xmath1535 & @xmath1536 & @xmath1537 & @xmath1538 & @xmath1539 & @xmath1540 + @xmath1541 & @xmath1542 & @xmath1543 & @xmath1544 & @xmath1545 & @xmath1546 & @xmath1547 & @xmath1548 & @xmath1549 & @xmath1550 + @xmath1551 & @xmath1552 & @xmath1553 & @xmath1554 & @xmath1555 & @xmath1556 & @xmath1557 & @xmath1558 & @xmath1559 & @xmath1560 + @xmath1561 & @xmath1562 & @xmath1563 & @xmath1564 & @xmath1565 & @xmath1566 & @xmath1567 & @xmath1568 & @xmath1569 & @xmath1570 + @xmath1571 & @xmath1572 & @xmath1573 & @xmath1574 & @xmath1575 & @xmath1576 & @xmath1577 & @xmath1575 & @xmath1578 & @xmath1578 + @xmath1579 & @xmath1580 & @xmath1513 & @xmath1581 & @xmath1582 & @xmath1583 & @xmath1331 & @xmath1584 & @xmath1585 & @xmath1586 + @xmath1587 & @xmath1588 & @xmath1589 & @xmath1590 & @xmath1591 & @xmath1592 & @xmath1593 & @xmath1594 & @xmath1595 & @xmath1596 + @xmath1597 & @xmath1598 & @xmath1599 & @xmath1600 & @xmath1601 & @xmath1602 & @xmath1603 & @xmath1604 & @xmath1605 & @xmath1606 + @xmath1607 & @xmath1608 & @xmath1609 & @xmath1610 & @xmath1611 & @xmath1612 & @xmath1613 & @xmath1614 & @xmath1615 & @xmath1616 + @xmath1617 & @xmath1618 & @xmath1619 & @xmath1620 & @xmath1621 & @xmath1622 & @xmath1623 & @xmath1624 & @xmath1625 & @xmath1626 + @xmath1627 & @xmath1628 & @xmath1629 & @xmath1630 & @xmath1631 & @xmath1632 & @xmath1633 & @xmath1634 & @xmath1114 & @xmath1635 + @xmath1636 & @xmath1637 & @xmath1638 & @xmath1639 & @xmath1640 & @xmath1583 & @xmath1641 & @xmath1642 & @xmath1643 & @xmath1644 + @xmath1645 & @xmath1646 & @xmath1647 & @xmath1648 & @xmath1649 & @xmath1650 & @xmath1651 & @xmath1652 & @xmath1653 & @xmath1654 + @xmath1655 & @xmath1656 & @xmath1657 & @xmath1658 & @xmath1659 & @xmath1660 & @xmath1661 & @xmath1662 & @xmath1663 & @xmath1664 + @xmath1665 & @xmath1666 & @xmath1667 & @xmath1668 & @xmath1669 & @xmath1670 & @xmath1671 & @xmath1672 & @xmath1673 & @xmath1674 + @xmath1675 & @xmath1676 & @xmath1677 & @xmath1401 & @xmath1678 & @xmath1679 & @xmath1680 & @xmath1197 & @xmath1681 & @xmath1682 + @xmath1683 & @xmath1684 & @xmath1685 & @xmath1686 & @xmath1687 & @xmath1688 & @xmath1689 & @xmath1690 & @xmath1691 & @xmath1692 + @xmath1693 & @xmath1694 & @xmath1695 & @xmath1696 & @xmath1697 & @xmath1698 & @xmath1676 & @xmath1699 & @xmath1595 & @xmath1700 + @xmath1701 & @xmath1702 & @xmath1703 & @xmath1704 & @xmath1705 & @xmath1706 & @xmath1707 & @xmath1708 & @xmath1709 & @xmath1710 + @xmath1711 & @xmath1712 & @xmath1713 & @xmath1714 & @xmath1715 & @xmath1716 & @xmath1717 & @xmath1718 & @xmath1719 & @xmath1720 + @xmath1721 & @xmath1722 & @xmath1723 & @xmath1724 & @xmath1725 & @xmath1726 & @xmath1727 & @xmath1728 & @xmath1729 & @xmath1730 + @xmath1731 & @xmath1732 & @xmath1733 & @xmath1734 & @xmath1735 & @xmath1736 & @xmath1600 & @xmath1134 & @xmath1497 & @xmath1737 + @xmath1738 & @xmath1739 & @xmath1740 & @xmath1741 & @xmath1244 & @xmath1742 & @xmath1743 & @xmath1744 & @xmath1745 & @xmath1746 + @xmath1747 & @xmath1748 & @xmath1749 & @xmath1706 & @xmath1750 & @xmath1751 & @xmath1752 & @xmath1753 & @xmath1754 & @xmath1755 + @xmath1756 & @xmath1757 & @xmath1758 & @xmath1759 & @xmath1760 & @xmath1761 & @xmath1159 & @xmath1762 & @xmath1247 & @xmath1763 + @xmath1764 & @xmath1765 & @xmath1766 & @xmath1767 & @xmath1768 & @xmath1769 & @xmath1770 & @xmath1771 & @xmath1772 & @xmath1773 + @xmath1774 & @xmath1775 & @xmath1369 & @xmath1776 & @xmath1777 & @xmath1669 & @xmath1778 & @xmath1779 & @xmath1780 & @xmath1781 + @xmath1782 & @xmath1783 & @xmath1784 & @xmath1785 & @xmath1786 & @xmath1787 & @xmath1788 & @xmath1789 & @xmath1790 & @xmath1199 + @xmath1791 & @xmath1792 & @xmath1793 & @xmath1562 & @xmath1206 & @xmath1794 & @xmath1195 & @xmath1193 & @xmath1795 & @xmath1796 + @xmath1797 & @xmath1798 & @xmath1799 & @xmath1754 & @xmath1800 & @xmath1801 & @xmath1802 & @xmath1803 & @xmath1804 & @xmath1805 + @xmath1806 & @xmath1807 & @xmath1808 & @xmath1809 & @xmath1810 & @xmath1811 & @xmath1812 & @xmath1813 & @xmath1814 & @xmath1815 + @xmath1816 & @xmath1817 & @xmath1818 & @xmath1819 & @xmath1820 & @xmath1821 & @xmath1822 & @xmath1823 & @xmath1824 & @xmath1825 + @xmath1826 & @xmath1827 & @xmath1828 & @xmath1829 & @xmath1830 & @xmath1831 & @xmath1832 & @xmath1833 & @xmath1834 & @xmath1698 + @xmath1835 & @xmath1836 & @xmath1837 & @xmath1838 & @xmath1839 & @xmath1840 & @xmath1841 & @xmath1842 & @xmath1843 & @xmath1844 + @xmath1845 & @xmath1846 & @xmath1847 & @xmath1848 & @xmath1849 & @xmath1850 & @xmath1851 & @xmath1852 & @xmath1698 & @xmath1853 + @xmath1854 & @xmath1855 & @xmath1856 & @xmath1857 & @xmath1858 & @xmath1859 & @xmath1860 & @xmath1861 & @xmath1862 & @xmath1863 + @xmath1864 & @xmath1865 & @xmath1866 & @xmath1867 & @xmath1868 & @xmath1869 & @xmath1870 & @xmath1871 & @xmath1872 & @xmath1873 + @xmath1874 & @xmath1875 & @xmath1876 & @xmath1877 & @xmath1878 & @xmath1879 & @xmath1880 & @xmath1881 & @xmath1882 & @xmath1883 + @xmath1884 & @xmath1885 & @xmath1886 & @xmath1887 & @xmath1888 & @xmath1889 & @xmath1890 & @xmath1891 & @xmath1892 & @xmath1893 + @xmath1894 & @xmath1895 & @xmath1896 & @xmath1897 & @xmath1898 & @xmath1899 & @xmath1900 & @xmath1901 & @xmath1902 & @xmath1903 + @xmath1904 & @xmath1905 & @xmath1906 & @xmath1907 & @xmath1908 & @xmath1314 & @xmath1909 & @xmath1910 & @xmath1911 & @xmath1912 + @xmath1913 & @xmath1914 & @xmath1915 & @xmath1916 & @xmath1917 & @xmath1918 & @xmath1919 & @xmath1920 & @xmath1921 & @xmath1922 + @xmath1923 & @xmath1924 & @xmath1925 & @xmath1926 & @xmath1927 & @xmath1928 & @xmath1929 & @xmath1930 & @xmath1931 & @xmath1932 + @xmath1933 & @xmath1934 & @xmath1935 & @xmath1936 & @xmath1937 & @xmath1938 & @xmath1939 & @xmath1940 & @xmath1941 & @xmath1942 + @xmath1943 & @xmath1944 & @xmath1945 & @xmath1946 & @xmath1947 & @xmath1948 & @xmath1949 & @xmath1950 & @xmath1951 & @xmath1952 + @xmath1953 & @xmath1954 & @xmath1126 & @xmath1853 & @xmath1955 & @xmath1956 & @xmath1957 & @xmath1583 & @xmath1958 & @xmath1959 + @xmath1960 & @xmath1961 & @xmath1962 & @xmath1963 & @xmath1964 & @xmath1965 & @xmath1966 & @xmath1967 & @xmath1968 & @xmath1837 + @xmath1969 & @xmath1970 & @xmath1971 & @xmath1972 & @xmath1973 & @xmath1974 & @xmath1975 & @xmath1976 & @xmath1977 & @xmath1978 + @xmath1979 & @xmath1582 & @xmath1515 & @xmath1980 & @xmath1981 & @xmath1519 & @xmath1982 & @xmath1160 & @xmath1983 & @xmath1984 + @xmath1985 & @xmath1986 & @xmath1987 & @xmath1988 & @xmath1989 & @xmath1990 & @xmath1991 & @xmath1992 & @xmath1993 & @xmath1994 + @xmath1995 & @xmath1996 & @xmath1997 & @xmath1998 & @xmath1999 & @xmath1676 & @xmath2000 & @xmath1129 & @xmath2001 & @xmath2002 + @xmath2003 & @xmath2004 & @xmath2005 & @xmath1126 & @xmath2006 & @xmath2007 & @xmath2008 & @xmath2009 & @xmath2010 & @xmath2011 + @xmath2012 & @xmath2013 & @xmath2014 & @xmath2015 & @xmath2016 & @xmath2017 & @xmath2018 & @xmath2019 & @xmath2020 & @xmath2021 + @xmath2022 & @xmath2023 & @xmath2024 & @xmath2025 & @xmath2026 & @xmath2027 & @xmath1079 & @xmath2028 & @xmath2029 & @xmath2030 + @xmath2031 & @xmath2032 & @xmath2033 & @xmath2034 & @xmath2035 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath2036 & @xmath1082 + @xmath2037 & @xmath2038 & @xmath2039 & @xmath2040 & @xmath2041 & @xmath2042 & @xmath2043 & @xmath2044 & @xmath2045 & @xmath2046 + @xmath2047 & @xmath2048 & @xmath2049 & @xmath2050 & @xmath2051 & @xmath1482 & @xmath2052 & @xmath2053 & @xmath2054 & @xmath2055 + @xmath2056 & @xmath2057 & @xmath2058 & @xmath1126 & @xmath1676 & @xmath2059 & @xmath2060 & @xmath1517 & @xmath2061 & @xmath2062 + @xmath2063 & @xmath2064 & @xmath2065 & @xmath2066 & @xmath2067 & @xmath2068 & @xmath2069 & @xmath2070 & @xmath2071 & @xmath2072 + @xmath2073 & @xmath1850 & @xmath2074 & @xmath2075 & @xmath2076 & @xmath2077 & @xmath1821 & @xmath1910 & @xmath2078 & @xmath2079 + @xmath2080 & @xmath2081 & @xmath1345 & @xmath2082 & @xmath2083 & @xmath2084 & @xmath2085 & @xmath2086 & @xmath2087 & @xmath2088 + @xmath2089 & @xmath2090 & @xmath1977 & @xmath1401 & @xmath2091 & @xmath2092 & @xmath2093 & @xmath2094 & @xmath2095 & @xmath2096 + @xmath2097 & @xmath2098 & @xmath2099 & @xmath2100 & @xmath2101 & @xmath2102 & @xmath2103 & @xmath2104 & @xmath2105 & @xmath2106 + @xmath2107 & @xmath2108 & @xmath2109 & @xmath2110 & @xmath2096 & @xmath2111 & @xmath2112 & @xmath2113 & @xmath2114 & @xmath2115 + @xmath2116 & @xmath1153 & @xmath1643 & @xmath2117 & @xmath2118 & @xmath2119 & @xmath2120 & @xmath2121 & @xmath2122 & @xmath2123 + @xmath2124 & @xmath2125 & @xmath1203 & @xmath1244 & @xmath2126 & @xmath1795 & @xmath2127 & @xmath2128 & @xmath2129 & @xmath2130 + @xmath2131 & @xmath2132 & @xmath1642 & @xmath1160 & @xmath2133 & @xmath2134 & @xmath2121 & @xmath2135 & @xmath2136 & @xmath2137 + @xmath2138 & @xmath1698 & @xmath1761 & @xmath2139 & @xmath2140 & @xmath2141 & @xmath2142 & @xmath2143 & @xmath2144 & @xmath2145 + @xmath2146 & @xmath2147 & @xmath2148 & @xmath2149 & @xmath2150 & @xmath2151 & @xmath2152 & @xmath2153 & @xmath2154 & @xmath2155 + @xmath2156 & @xmath2157 & @xmath1075 & @xmath1076 & @xmath2158 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath1081 & @xmath1082 + @xmath2159 & @xmath2160 & @xmath2161 & @xmath2162 & @xmath2163 & @xmath2164 & @xmath2165 & @xmath2166 & @xmath2167 & @xmath2168 + @xmath2169 & @xmath1651 & @xmath2170 & @xmath2171 & @xmath2172 & @xmath2173 & @xmath2174 & @xmath2175 & @xmath2176 & @xmath2177 + @xmath2178 & @xmath2179 & @xmath1208 & @xmath2180 & @xmath2181 & @xmath1402 & @xmath2182 & @xmath2183 & @xmath2184 & @xmath2185 + @xmath2186 & @xmath2187 & @xmath1235 & @xmath2188 & @xmath2189 & @xmath2190 & @xmath2191 & @xmath2192 & @xmath2193 & @xmath1470 + @xmath2194 & @xmath1978 & @xmath2195 & @xmath2137 & @xmath2196 & @xmath2180 & @xmath2197 & @xmath2198 & @xmath2199 & @xmath2200 + @xmath2201 & @xmath2202 & @xmath1237 & @xmath2203 & @xmath2204 & @xmath2205 & @xmath2206 & @xmath1925 & @xmath1464 & @xmath2207 + @xmath2208 & @xmath2209 & @xmath2210 & @xmath2211 & @xmath2212 & @xmath2213 & @xmath2214 & @xmath2215 & @xmath2216 & @xmath2217 + @xmath2218 & @xmath1565 & @xmath2219 & @xmath2220 & @xmath2221 & @xmath2222 & @xmath2223 & @xmath2224 & @xmath2225 & @xmath2226 + @xmath2227 & @xmath2148 & @xmath2228 & @xmath2229 & @xmath2230 & @xmath2231 & @xmath2232 & @xmath2233 & @xmath2234 & @xmath2235 + @xmath2236 & @xmath2237 & @xmath2238 & @xmath2239 & @xmath2240 & @xmath1453 & @xmath2241 & @xmath2242 & @xmath2243 & @xmath2244 + @xmath2245 & @xmath2246 & @xmath2247 & @xmath2248 & @xmath2249 & @xmath2183 & @xmath2250 & @xmath2251 & @xmath2252 & @xmath2253 + @xmath2254 & @xmath2255 & @xmath2256 & @xmath1408 & @xmath2257 & @xmath2258 & @xmath2259 & @xmath1285 & @xmath2260 & @xmath2261 + @xmath2262 & @xmath2263 & @xmath2264 & @xmath2265 & @xmath1565 & @xmath1685 & @xmath1271 & @xmath2266 & @xmath2267 & @xmath2268 + @xmath2269 & @xmath2270 & @xmath1503 & @xmath2271 & @xmath2272 & @xmath2273 & @xmath2274 & @xmath2275 & @xmath2276 & @xmath2277 + @xmath2278 & @xmath2279 & @xmath2280 & @xmath1199 & @xmath2281 & @xmath2282 & @xmath2283 & @xmath2284 & @xmath2285 & @xmath2286 + @xmath2287 & @xmath2288 & @xmath2289 & @xmath2290 & @xmath1125 & @xmath2291 & @xmath2094 & @xmath2292 & @xmath2293 & @xmath2294 + @xmath2295 & @xmath2296 & @xmath2297 & @xmath2298 & @xmath2299 & @xmath2300 & @xmath2301 & @xmath2302 & @xmath2303 & @xmath2304 + @xmath2305 & @xmath1074 & @xmath1075 & @xmath1076 & @xmath1077 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath1081 & @xmath1082 + @xmath2306 & @xmath2307 & @xmath2308 & @xmath2309 & @xmath2219 & @xmath2310 & @xmath2200 & @xmath2311 & @xmath2312 & @xmath2313 + @xmath2314 & @xmath2315 & @xmath2316 & @xmath2317 & @xmath2318 & @xmath2319 & @xmath2320 & @xmath2321 & @xmath1981 & @xmath1594 + @xmath2322 & @xmath2323 & @xmath1648 & @xmath2324 & @xmath2325 & @xmath2326 & @xmath2327 & @xmath2328 & @xmath1790 & @xmath1682 + @xmath2329 & @xmath2330 & @xmath2331 & @xmath2332 & @xmath1526 & @xmath2333 & @xmath2334 & @xmath2335 & @xmath1642 & @xmath1910 + @xmath2336 & @xmath2326 & @xmath2337 & @xmath2338 & @xmath2339 & @xmath1565 & @xmath2340 & @xmath2341 & @xmath2342 & @xmath2343 + @xmath2344 & @xmath1331 & @xmath2345 & @xmath2007 & @xmath2346 & @xmath2347 & @xmath2093 & @xmath1203 & @xmath2348 & @xmath1206 + @xmath2349 & @xmath2350 & @xmath2351 & @xmath2352 & @xmath2353 & @xmath2354 & @xmath2351 & @xmath2355 & @xmath2356 & @xmath2357 + @xmath126 & @xmath2358 & @xmath2359 & @xmath2360 & @xmath2360 & @xmath2358 & @xmath2361 & @xmath2362 & @xmath2361 & @xmath2363 + @xmath2364 & @xmath2365 & @xmath2366 & @xmath2312 & @xmath1446 & @xmath2221 & @xmath2367 & @xmath1021 & @xmath1635 & @xmath2368 + @xmath2369 & @xmath2370 & @xmath2371 & @xmath2372 & @xmath2373 & @xmath2374 & @xmath2375 & @xmath2376 & @xmath2377 & @xmath2378 + @xmath2379 & @xmath2195 & @xmath2380 & @xmath2381 & @xmath2382 & @xmath2383 & @xmath1790 & @xmath2384 & @xmath1992 & @xmath1199 + @xmath2385 & @xmath2386 & @xmath2219 & @xmath2387 & @xmath2388 & @xmath2389 & @xmath2390 & @xmath2391 & @xmath2392 & @xmath2393 + @xmath2394 & @xmath2395 & @xmath1399 & @xmath2396 & @xmath2397 & @xmath2398 & @xmath2399 & @xmath2400 & @xmath2401 & @xmath2402 + @xmath2403 & @xmath2091 & @xmath2180 & @xmath2404 & @xmath2405 & @xmath2222 & @xmath2406 & @xmath2407 & @xmath2408 & @xmath2409 + @xmath2410 & @xmath1981 & @xmath2411 & @xmath1243 & @xmath1400 & @xmath2412 & @xmath2413 & @xmath2414 & @xmath2415 & @xmath2416 + @xmath2417 & @xmath2418 & @xmath2419 & @xmath2420 & @xmath2421 & @xmath2422 & @xmath1631 & @xmath2423 & @xmath2424 & @xmath2425 + @xmath2426 & @xmath2427 & @xmath2428 & @xmath2429 & @xmath1237 & @xmath2430 & @xmath1967 & @xmath2431 & @xmath2432 & @xmath2433 + @xmath2434 & @xmath1074 & @xmath1075 & @xmath1076 & @xmath1077 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath1081 & @xmath1082 + @xmath2435 & @xmath2436 & @xmath2437 & @xmath2438 & @xmath2439 & @xmath2440 & @xmath2441 & @xmath2442 & @xmath2443 & @xmath2444 + @xmath2445 & @xmath2446 & @xmath2447 & @xmath2448 & @xmath2449 & @xmath2450 & @xmath2451 & @xmath2452 & @xmath2453 & @xmath2454 + @xmath2455 & @xmath2456 & @xmath2457 & @xmath2458 & @xmath2459 & @xmath2460 & @xmath2461 & @xmath2462 & @xmath2463 & @xmath2464 + @xmath2465 & @xmath2466 & @xmath2467 & @xmath1910 & @xmath2468 & @xmath2469 & @xmath2470 & @xmath2471 & @xmath2472 & @xmath2473 + @xmath2474 & @xmath2475 & @xmath2476 & @xmath2477 & @xmath2478 & @xmath1827 & @xmath2479 & @xmath2480 & @xmath2481 & @xmath2482 + @xmath2483 & @xmath2484 & @xmath2485 & @xmath2486 & @xmath2487 & @xmath2488 & @xmath2489 & @xmath2490 & @xmath2246 & @xmath2491 + @xmath2492 & @xmath2493 & @xmath1748 & @xmath2494 & @xmath2495 & @xmath2496 & @xmath2497 & @xmath1755 & @xmath2498 & @xmath1315 + @xmath2499 & @xmath1981 & @xmath2500 & @xmath2501 & @xmath2502 & @xmath2503 & @xmath2504 & @xmath2505 & @xmath2506 & @xmath2507 + @xmath2508 & @xmath2509 & @xmath2510 & @xmath2511 & @xmath2512 & @xmath2513 & @xmath2514 & @xmath2515 & @xmath2516 & @xmath2517 + @xmath2518 & @xmath2519 & @xmath2520 & @xmath2521 & @xmath2522 & @xmath2523 & @xmath2524 & @xmath2525 & @xmath2526 & @xmath2527 + @xmath2528 & @xmath2529 & @xmath2530 & @xmath2531 & @xmath2532 & @xmath2533 & @xmath2534 & @xmath2535 & @xmath2536 & @xmath2537 + @xmath2538 & @xmath2115 & @xmath2112 & @xmath1402 & @xmath1740 & @xmath2539 & @xmath2540 & @xmath2541 & @xmath2542 & @xmath2543 + @xmath2544 & @xmath2545 & @xmath2546 & @xmath2547 & @xmath2548 & @xmath2549 & @xmath2550 & @xmath2551 & @xmath2552 & @xmath2553 + @xmath2554 & @xmath2555 & @xmath2556 & @xmath2557 & @xmath2558 & @xmath2559 & @xmath2560 & @xmath2324 & @xmath2561 & @xmath2562 + @xmath2563 & @xmath1513 & @xmath2564 & @xmath2565 & @xmath2076 & @xmath1158 & @xmath2566 & @xmath1520 & @xmath2567 & @xmath2418 + @xmath2568 & @xmath2569 & @xmath1911 & @xmath2570 & @xmath2136 & @xmath2571 & @xmath1244 & @xmath1994 & @xmath2507 & @xmath2113 + @xmath2572 & @xmath2573 & @xmath2574 & @xmath2575 & @xmath2576 & @xmath2577 & @xmath2578 & @xmath2579 & @xmath2580 & @xmath2581 + @xmath2582 & @xmath2583 & @xmath1593 & @xmath2584 & @xmath2585 & @xmath2586 & @xmath2587 & @xmath2588 & @xmath2589 & @xmath2590 + @xmath2591 & @xmath2592 & @xmath2593 & @xmath2594 & @xmath2595 & @xmath2596 & @xmath2597 & @xmath1887 & @xmath2598 & @xmath2457 + @xmath2599 & @xmath2600 & @xmath2601 & @xmath2602 & @xmath2603 & @xmath2496 & @xmath2604 & @xmath2605 & @xmath2606 & @xmath2607 + @xmath2608 & @xmath2609 & @xmath2610 & @xmath1214 & @xmath2611 & @xmath2612 & @xmath2613 & @xmath2614 & @xmath2615 & @xmath2616 + @xmath2617 & @xmath1074 & @xmath1075 & @xmath1076 & @xmath1077 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath1081 & @xmath1082 + @xmath2618 & @xmath2619 & @xmath2620 & @xmath2621 & @xmath2622 & @xmath2623 & @xmath2624 & @xmath2625 & @xmath2626 & @xmath2627 + @xmath2628 & @xmath2629 & @xmath2630 & @xmath2631 & @xmath2632 & @xmath2633 & @xmath2634 & @xmath2635 & @xmath2636 & @xmath2637 + @xmath2638 & @xmath2639 & @xmath2640 & @xmath2150 & @xmath2641 & @xmath2642 & @xmath2643 & @xmath2644 & @xmath2645 & @xmath2646 + @xmath2647 & @xmath2648 & @xmath2649 & @xmath2650 & @xmath2651 & @xmath2652 & @xmath2653 & @xmath2654 & @xmath2541 & @xmath2655 + @xmath2656 & @xmath2657 & @xmath2658 & @xmath1900 & @xmath2659 & @xmath2660 & @xmath2661 & @xmath2662 & @xmath2663 & @xmath1842 + @xmath2664 & @xmath2665 & @xmath2666 & @xmath2667 & @xmath2668 & @xmath2669 & @xmath2670 & @xmath2671 & @xmath2672 & @xmath2673 + @xmath2674 & @xmath2675 & @xmath2676 & @xmath2677 & @xmath1821 & @xmath2678 & @xmath2679 & @xmath2007 & @xmath2680 & @xmath2681 + @xmath2682 & @xmath2683 & @xmath2684 & @xmath2685 & @xmath2686 & @xmath2687 & @xmath2688 & @xmath2689 & @xmath2690 & @xmath2691 + @xmath2692 & @xmath2693 & @xmath2694 & @xmath2695 & @xmath2696 & @xmath1642 & @xmath1315 & @xmath2697 & @xmath2576 & @xmath1594 + @xmath2698 & @xmath2699 & @xmath2700 & @xmath2701 & @xmath2702 & @xmath2072 & @xmath2703 & @xmath2704 & @xmath1917 & @xmath2705 + @xmath2706 & @xmath2707 & @xmath2708 & @xmath2709 & @xmath2710 & @xmath2711 & @xmath2712 & @xmath2713 & @xmath2714 & @xmath2715 + @xmath2716 & @xmath2717 & @xmath2718 & @xmath2719 & @xmath2720 & @xmath2721 & @xmath2722 & @xmath2723 & @xmath2724 & @xmath2725 + @xmath2726 & @xmath2727 & @xmath2728 & @xmath2729 & @xmath2730 & @xmath2731 & @xmath2732 & @xmath2733 & @xmath2734 & @xmath2735 + @xmath2736 & @xmath2737 & @xmath2738 & @xmath2739 & @xmath2740 & @xmath2741 & @xmath2742 & @xmath2743 & @xmath2744 & @xmath2745 + @xmath2746 & @xmath2747 & @xmath2748 & @xmath2749 & @xmath2750 & @xmath2751 & @xmath2752 & @xmath2753 & @xmath2754 & @xmath2755 + @xmath2756 & @xmath2757 & @xmath2758 & @xmath2759 & @xmath2760 & @xmath2761 & @xmath2762 & @xmath2763 & @xmath2751 & @xmath1898 + @xmath2764 & @xmath2765 & @xmath2766 & @xmath2767 & @xmath1593 & @xmath2768 & @xmath2769 & @xmath2770 & @xmath1801 & @xmath2771 + @xmath2772 & @xmath2773 & @xmath2774 & @xmath2775 & @xmath1525 & @xmath2776 & @xmath2777 & @xmath2778 & @xmath1591 & @xmath2779 + @xmath2780 & @xmath2781 & @xmath2782 & @xmath2783 & @xmath2784 & @xmath2785 & @xmath2786 & @xmath2787 & @xmath2770 & @xmath2788 + @xmath2789 & @xmath2790 & @xmath2791 & @xmath2193 & @xmath2792 & @xmath1981 & @xmath1755 & @xmath2793 & @xmath2794 & @xmath2795 + @xmath2796 & @xmath2694 & @xmath2797 & @xmath2798 & @xmath2799 & @xmath2800 & @xmath2801 & @xmath2139 & @xmath2802 & @xmath2557 + @xmath2803 & @xmath2804 & @xmath2805 & @xmath2806 & @xmath2807 & @xmath2808 & @xmath2809 & @xmath2810 & @xmath2808 & @xmath2811 + @xmath2812 & @xmath2813 & @xmath2814 & @xmath2815 & @xmath1931 & @xmath2816 & @xmath1472 & @xmath2817 & @xmath1582 & @xmath1757 + @xmath2818 & @xmath2819 & @xmath2820 & @xmath2821 & @xmath2822 & @xmath2823 & @xmath2824 & @xmath2752 & @xmath2825 & @xmath2826 + @xmath2827 & @xmath2828 & @xmath2829 & @xmath2830 & @xmath2831 & @xmath2832 & @xmath2833 & @xmath2834 & @xmath2835 & @xmath2836 + @xmath2837 & @xmath2838 & @xmath2839 & @xmath2840 & @xmath2841 & @xmath2842 & @xmath2843 & @xmath2844 & @xmath1720 & @xmath2845 + @xmath2846 & @xmath2847 & @xmath2848 & @xmath2826 & @xmath2849 & @xmath2850 & @xmath2753 & @xmath2851 & @xmath2852 & @xmath2853 + @xmath2854 & @xmath2855 & @xmath2856 & @xmath2857 & @xmath2858 & @xmath2859 & @xmath2860 & @xmath2861 & @xmath2862 & @xmath2751 + @xmath2863 & @xmath2433 & @xmath2864 & @xmath2865 & @xmath2866 & @xmath2867 & @xmath2868 & @xmath2869 & @xmath2870 & @xmath2871 + @xmath2872 & @xmath2873 & @xmath2874 & @xmath2064 & @xmath2875 & @xmath1917 & @xmath2876 & @xmath2877 & @xmath2878 & @xmath2879 + @xmath2880 & @xmath2881 & @xmath2882 & @xmath1942 & @xmath2883 & @xmath2884 & @xmath2885 & @xmath2886 & @xmath2887 & @xmath2888 + @xmath2889 & @xmath2890 & @xmath2891 & @xmath2892 & @xmath2893 & @xmath2894 & @xmath1968 & @xmath2895 & @xmath2896 & @xmath2476 + @xmath2897 & @xmath2898 & @xmath2899 & @xmath2900 & @xmath2901 & @xmath2902 & @xmath2903 & @xmath2904 & @xmath2905 & @xmath2906 + @xmath2907 & @xmath2908 & @xmath2909 & @xmath2910 & @xmath2911 & @xmath2912 & @xmath2913 & @xmath2914 & @xmath2915 & @xmath2916 + @xmath2917 & @xmath2918 & @xmath2919 & @xmath2920 & @xmath2921 & @xmath2574 & @xmath2922 & @xmath2923 & @xmath2924 & @xmath2925 + @xmath2926 & @xmath2927 & @xmath2928 & @xmath1704 & @xmath2929 & @xmath2930 & @xmath1126 & @xmath2931 & @xmath2932 & @xmath2933 + @xmath2934 & @xmath2935 & @xmath2936 & @xmath2937 & @xmath2938 & @xmath1757 & @xmath2939 & @xmath2574 & @xmath2940 & @xmath2941 + @xmath2942 & @xmath2943 & @xmath2944 & @xmath2945 & @xmath2946 & @xmath2947 & @xmath2948 & @xmath1126 & @xmath2288 & @xmath1755 + @xmath2949 & @xmath2950 & @xmath2951 & @xmath2952 & @xmath2953 & @xmath2954 & @xmath2955 & @xmath2956 & @xmath2957 & @xmath2958 + @xmath2959 & @xmath2919 & @xmath2960 & @xmath2961 & @xmath2962 & @xmath1928 & @xmath2963 & @xmath2964 & @xmath2965 & @xmath2966 + @xmath2967 & @xmath2968 & @xmath2969 & @xmath2970 & @xmath2971 & @xmath2972 & @xmath2973 & @xmath2974 & @xmath2975 & @xmath2976 + @xmath2977 & @xmath2978 & @xmath2979 & @xmath2980 & @xmath1302 & @xmath2981 & @xmath2982 & @xmath2983 & @xmath2984 & @xmath2985 + @xmath2986 & @xmath2987 & @xmath2988 & @xmath2989 & @xmath2990 & @xmath1141 & @xmath2991 & @xmath2992 & @xmath2993 & @xmath2994 + @xmath2995 & @xmath2996 & @xmath2997 & @xmath2998 & @xmath2999 & @xmath3000 & @xmath3001 & @xmath3002 & @xmath3003 & @xmath3004 + @xmath3005 & @xmath3006 & @xmath3007 & @xmath3008 & @xmath3009 & @xmath3010 & @xmath3011 & @xmath3012 & @xmath3013 & @xmath3010 + @xmath3014 & @xmath3015 & @xmath3016 & @xmath2065 & @xmath3017 & @xmath3018 & @xmath3019 & @xmath3020 & @xmath3021 & @xmath1818 + @xmath3022 & @xmath3023 & @xmath3024 & @xmath3025 & @xmath3026 & @xmath2088 & @xmath3027 & @xmath3028 & @xmath3029 & @xmath3030 + @xmath3031 & @xmath3032 & @xmath3033 & @xmath3034 & @xmath2703 & @xmath2797 & @xmath3035 & @xmath1847 & @xmath1471 & @xmath1906 + @xmath3036 & @xmath3037 & @xmath3038 & @xmath3039 & @xmath3040 & @xmath1941 & @xmath3041 & @xmath3042 & @xmath3043 & @xmath3044 + @xmath3045 & @xmath3046 & @xmath3047 & @xmath3048 & @xmath3049 & @xmath3050 & @xmath1101 & @xmath3051 & @xmath3052 & @xmath3053 + @xmath3054 & @xmath3055 & @xmath3056 & @xmath3057 & @xmath3058 & @xmath3059 & @xmath2810 & @xmath3060 & @xmath3061 & @xmath3062 + @xmath3063 & @xmath3064 & @xmath3065 & @xmath3066 & @xmath3067 & @xmath3068 & @xmath3069 & @xmath3070 & @xmath3071 & @xmath3072 + @xmath3073 & @xmath3074 & @xmath1937 & @xmath2202 & @xmath3075 & @xmath3076 & @xmath3077 & @xmath3078 & @xmath1467 & @xmath1231 + @xmath3079 & @xmath3080 & @xmath3081 & @xmath2904 & @xmath3082 & @xmath3083 & @xmath3084 & @xmath1965 & @xmath3085 & @xmath3086 + @xmath3087 & @xmath3088 & @xmath3089 & @xmath3090 & @xmath3091 & @xmath3092 & @xmath3093 & @xmath3094 & @xmath3095 & @xmath2871 + @xmath3096 & @xmath3097 & @xmath2207 & @xmath3098 & @xmath3099 & @xmath3100 & @xmath3101 & @xmath1906 & @xmath2817 & @xmath3102 + @xmath3103 & @xmath3104 & @xmath3105 & @xmath3106 & @xmath3102 & @xmath3107 & @xmath3108 & @xmath3109 & @xmath1698 & @xmath3110 + @xmath3111 & @xmath3112 & @xmath3062 & @xmath3113 & @xmath3114 & @xmath3106 & @xmath2797 & @xmath1705 & @xmath1917 & @xmath1930 + @xmath3115 & @xmath3116 & @xmath1844 & @xmath3117 & @xmath3118 & @xmath3119 & @xmath3120 & @xmath1638 & @xmath1831 & @xmath3121 + @xmath3122 & @xmath3123 & @xmath3124 & @xmath3125 & @xmath3126 & @xmath3127 & @xmath3128 & @xmath3129 & @xmath3130 & @xmath3131 + @xmath3132 & @xmath3133 & @xmath3134 & @xmath3135 & @xmath3136 & @xmath3137 & @xmath3138 & @xmath3139 & @xmath3140 & @xmath3141 + @xmath3142 & @xmath3143 & @xmath3144 & @xmath3145 & @xmath3146 & @xmath3147 & @xmath3148 & @xmath3149 & @xmath3150 & @xmath3151 + @xmath3152 & @xmath3153 & @xmath3154 & @xmath3155 & @xmath3156 & @xmath3157 & @xmath3158 & @xmath3159 & @xmath3160 & @xmath3161 + @xmath3162 & @xmath3163 & @xmath2643 & @xmath3164 & @xmath3165 & @xmath3166 & @xmath3167 & @xmath3168 & @xmath3169 & @xmath3170 + @xmath3171 & @xmath3172 & @xmath2840 & @xmath3173 & @xmath3174 & @xmath3175 & @xmath3176 & @xmath3177 & @xmath3178 & @xmath3178 + @xmath3179 & @xmath3180 & @xmath3181 & @xmath3182 & @xmath3183 & @xmath3184 & @xmath3185 & @xmath3186 & @xmath3187 & @xmath3188 + @xmath3189 & @xmath3041 & @xmath1827 & @xmath3190 & @xmath3191 & @xmath3192 & @xmath3193 & @xmath3194 & @xmath2966 & @xmath2652 + @xmath3195 & @xmath3196 & @xmath3197 & @xmath1839 & @xmath1964 & @xmath2961 & @xmath3198 & @xmath3199 & @xmath3200 & @xmath3201 + @xmath3202 & @xmath3203 & @xmath3204 & @xmath3205 & @xmath3206 & @xmath3207 & @xmath2564 & @xmath3208 & @xmath3209 & @xmath2766 + @xmath3210 & @xmath3211 & @xmath3212 & @xmath3213 & @xmath3214 & @xmath3215 & @xmath3216 & @xmath3217 & @xmath3218 & @xmath3219 + @xmath3220 & @xmath3221 & @xmath3222 & @xmath2901 & @xmath3223 & @xmath3224 & @xmath3225 & @xmath1934 & @xmath3226 & @xmath3227 + @xmath3228 & @xmath3229 & @xmath3230 & @xmath3231 & @xmath2821 & @xmath3232 & @xmath3233 & @xmath3234 & @xmath3235 & @xmath3236 + @xmath3237 & @xmath1837 & @xmath1938 & @xmath2908 & @xmath3238 & @xmath3239 & @xmath1818 & @xmath1930 & @xmath3240 & @xmath1130 + @xmath3241 & @xmath3242 & @xmath2164 & @xmath3243 & @xmath3244 & @xmath3245 & @xmath3246 & @xmath3247 & @xmath3248 & @xmath1704 + @xmath3249 & @xmath3250 & @xmath3251 & @xmath3240 & @xmath3252 & @xmath3253 & @xmath3254 & @xmath3255 & @xmath1583 & @xmath2335 + @xmath3256 & @xmath3257 & @xmath3258 & @xmath3259 & @xmath3260 & @xmath3261 & @xmath3262 & @xmath3263 & @xmath3264 & @xmath3265 + @xmath3266 & @xmath3267 & @xmath1934 & @xmath3268 & @xmath1838 & @xmath3269 & @xmath1842 & @xmath3270 & @xmath3083 & @xmath2163 + @xmath3271 & @xmath3272 & @xmath3273 & @xmath3274 & @xmath3275 & @xmath3276 & @xmath3277 & @xmath3278 & @xmath3279 & @xmath3280 + @xmath3281 & @xmath3282 & @xmath3283 & @xmath3284 & @xmath3285 & @xmath3286 & @xmath3287 & @xmath3288 & @xmath3289 & @xmath3290 + @xmath3291 & @xmath3292 & @xmath3293 & @xmath3294 & @xmath3295 & @xmath3296 & @xmath3297 & @xmath3298 & @xmath3299 & @xmath3300 + @xmath3301 & @xmath3302 & @xmath3303 & @xmath3304 & @xmath3305 & @xmath3306 & @xmath3307 & @xmath3308 & @xmath3309 & @xmath3310 + @xmath3311 & @xmath3312 & @xmath3313 & @xmath3314 & @xmath3315 & @xmath3316 & @xmath3317 & @xmath2044 & @xmath2043 & @xmath1657 + @xmath3318 & @xmath3319 & @xmath2066 & @xmath1766 & @xmath3320 & @xmath1748 & @xmath2765 & @xmath3321 & @xmath3322 & @xmath2940 + @xmath3323 & @xmath3324 & @xmath3325 & @xmath3326 & @xmath3327 & @xmath2935 & @xmath1887 & @xmath3328 & @xmath3329 & @xmath3330 + @xmath3331 & @xmath3332 & @xmath3333 & @xmath3334 & @xmath3335 & @xmath3336 & @xmath3050 & @xmath3337 & @xmath3338 & @xmath3339 + @xmath3340 & @xmath1768 & @xmath3341 & @xmath3342 & @xmath3343 & @xmath3344 & @xmath1467 & @xmath2885 & @xmath3345 & @xmath3346 + @xmath3347 & @xmath1839 & @xmath3348 & @xmath3349 & @xmath3269 & @xmath3350 & @xmath3351 & @xmath3352 & @xmath2814 & @xmath2815 + @xmath3353 & @xmath3354 & @xmath3355 & @xmath3356 & @xmath3357 & @xmath3358 & @xmath3359 & @xmath3191 & @xmath3360 & @xmath3344 + @xmath3361 & @xmath3362 & @xmath3363 & @xmath1885 & @xmath3364 & @xmath3365 & @xmath3083 & @xmath2202 & @xmath3366 & @xmath3203 + @xmath3367 & @xmath3368 & @xmath3369 & @xmath3370 & @xmath3371 & @xmath3372 & @xmath3373 & @xmath3374 & @xmath3375 & @xmath3376 + @xmath3377 & @xmath3378 & @xmath3379 & @xmath3380 & @xmath3381 & @xmath2100 & @xmath2850 & @xmath3382 & @xmath3383 & @xmath3384 + @xmath3385 & @xmath3386 & @xmath3387 & @xmath3388 & @xmath3389 & @xmath3390 & @xmath3391 & @xmath3392 & @xmath3393 & @xmath3394 + @xmath3395 & @xmath3396 & @xmath1657 & @xmath3397 & @xmath3398 & @xmath3399 & @xmath3400 & @xmath3401 & @xmath2045 & @xmath1664 + @xmath3402 & @xmath3403 & @xmath3404 & @xmath3405 & @xmath3406 & @xmath1945 & @xmath3272 & @xmath3407 & @xmath3408 & @xmath3409 + @xmath3410 & @xmath3411 & @xmath3412 & @xmath3413 & @xmath3414 & @xmath3313 & @xmath1057 & @xmath3415 & @xmath3416 & @xmath3417 + @xmath3418 & @xmath2751 & @xmath3419 & @xmath3420 & @xmath3421 & @xmath3422 & @xmath3423 & @xmath3424 & @xmath2202 & @xmath1482 + @xmath3425 & @xmath3426 & @xmath3427 & @xmath3428 & @xmath3429 & @xmath3181 & @xmath3430 & @xmath3431 & @xmath3432 & @xmath3433 + @xmath3434 & @xmath3435 & @xmath3436 & @xmath3437 & @xmath3438 & @xmath2806 & @xmath3439 & @xmath3440 & @xmath3441 & @xmath3442 + @xmath3443 & @xmath3444 & @xmath3445 & @xmath3446 & @xmath3447 & @xmath3448 & @xmath3449 & @xmath3450 & @xmath3309 & @xmath3451 + @xmath3452 & @xmath2512 & @xmath3453 & @xmath3454 & @xmath3455 & @xmath3456 & @xmath3457 & @xmath2993 & @xmath3458 & @xmath3459 + @xmath3460 & @xmath3461 & @xmath3462 & @xmath1099 & @xmath1934 & @xmath3463 & @xmath3464 & @xmath3465 & @xmath3466 & @xmath3082 + @xmath3467 & @xmath3186 & @xmath3468 & @xmath3469 & @xmath3470 & @xmath3471 & @xmath3472 & @xmath3473 & @xmath3474 & @xmath3475 + @xmath3476 & @xmath3477 & @xmath3478 & @xmath3479 & @xmath3480 & @xmath3480 & @xmath3059 & @xmath3481 & @xmath3482 & @xmath3483 + @xmath3484 & @xmath3485 & @xmath3486 & @xmath2811 & @xmath2862 & @xmath2472 & @xmath3487 & @xmath3488 & @xmath3489 & @xmath1939 + @xmath3490 & @xmath3491 & @xmath3492 & @xmath3493 & @xmath3494 & @xmath3390 & @xmath3495 & @xmath3338 & @xmath3496 & @xmath3341 + @xmath3497 & @xmath2472 & @xmath3498 & @xmath3499 & @xmath1840 & @xmath3371 & @xmath3500 & @xmath3501 & @xmath3502 & @xmath3017 + @xmath3503 & @xmath3463 & @xmath3504 & @xmath3505 & @xmath3506 & @xmath3507 & @xmath3508 & @xmath3509 & @xmath3501 & @xmath3510 + lccccccccc @xmath980 & 1.14 & 1.40 & 1.40 & 1.48 & 1.56 & 1.58 & 1.48 & 1.75 & 1.92 + @xmath981 & 1.00 & 1.00 & 1.00 & 1.15 & 1.63 & 1.77 & 4.22 & 4.32 & 5.48 + @xmath982 & 1.38 & 1.42 & 1.57 & 1.65 & 1.79 & 2.08 & 2.21 & 2.43 & 2.81 + @xmath983 & @xmath3511 & @xmath3512 & @xmath3513 & @xmath1188 & @xmath3514 & @xmath3515 & @xmath3516 & @xmath3517 & @xmath3518 + @xmath993 & @xmath1919 & @xmath3519 & @xmath3520 & @xmath3521 & @xmath3522 & @xmath3523 & @xmath3524 & @xmath3525 & @xmath1794 + @xmath1003 & @xmath1863 & @xmath3526 & @xmath3527 & @xmath1377 & @xmath3528 & @xmath3529 & @xmath3530 & @xmath3531 & @xmath3532 + @xmath1013 & @xmath3533 & @xmath2422 & @xmath3534 & @xmath3535 & @xmath3536 & @xmath3537 & @xmath1021 & @xmath3538 & @xmath3539 + @xmath1023 & @xmath3540 & @xmath3541 & @xmath3542 & @xmath3543 & @xmath1188 & @xmath3544 & @xmath3545 & @xmath3546 & @xmath3547 + @xmath1033 & @xmath3548 & @xmath3549 & @xmath1044 & @xmath3550 & @xmath3551 & @xmath3552 & @xmath2976 & @xmath3553 & @xmath3554 + @xmath1043 & @xmath3236 & @xmath3325 & @xmath3555 & @xmath3556 & @xmath3557 & @xmath3558 & @xmath3559 & @xmath3196 & @xmath1901 + @xmath1053 & @xmath3560 & @xmath3561 & @xmath3562 & @xmath3563 & @xmath3564 & @xmath3565 & @xmath3566 & @xmath3313 & @xmath3302 + @xmath1063 & @xmath1946 & @xmath3567 & @xmath2996 & @xmath3568 & @xmath3569 & @xmath3570 & @xmath3571 & @xmath2831 & @xmath3572 + @xmath1073 & @xmath3573 & @xmath1079 & @xmath3574 & @xmath3575 & @xmath3576 & @xmath3577 & @xmath3578 & @xmath3579 & @xmath3580 + @xmath1083 & @xmath3186 & @xmath3581 & @xmath3582 & @xmath3583 & @xmath3584 & @xmath3585 & @xmath3586 & @xmath3587 & @xmath3588 + @xmath1093 & @xmath3057 & @xmath1888 & @xmath2900 & @xmath3080 & @xmath2860 & @xmath3589 & @xmath3328 & @xmath3590 & @xmath3591 + @xmath1103 & @xmath3592 & @xmath3593 & @xmath3594 & @xmath3595 & @xmath3596 & @xmath3597 & @xmath3598 & @xmath3599 & @xmath3600 + @xmath1113 & @xmath3601 & @xmath3602 & @xmath3603 & @xmath3604 & @xmath3605 & @xmath1271 & @xmath3606 & @xmath2387 & @xmath2391 + @xmath1123 & @xmath2887 & @xmath3607 & @xmath3608 & @xmath2704 & @xmath2966 & @xmath1313 & @xmath3609 & @xmath1581 & @xmath2432 + @xmath1133 & @xmath3610 & @xmath1037 & @xmath3611 & @xmath3297 & @xmath3612 & @xmath3613 & @xmath3614 & @xmath3615 & @xmath3616 + @xmath1142 & @xmath3617 & @xmath3618 & @xmath3619 & @xmath3620 & @xmath3621 & @xmath3622 & @xmath3623 & @xmath3624 & @xmath3625 + @xmath1152 & @xmath3626 & @xmath3627 & @xmath3628 & @xmath3629 & @xmath3630 & @xmath2470 & @xmath3631 & @xmath3632 & @xmath3633 + @xmath1162 & @xmath3634 & @xmath3635 & @xmath3636 & @xmath3637 & @xmath3638 & @xmath3639 & @xmath3640 & @xmath3641 & @xmath3642 + @xmath1172 & @xmath3643 & @xmath3644 & @xmath3645 & @xmath3646 & @xmath3647 & @xmath3648 & @xmath3649 & @xmath3650 & @xmath2088 + @xmath1181 & @xmath1557 & @xmath3651 & @xmath3652 & @xmath3653 & @xmath3654 & @xmath3655 & @xmath3656 & @xmath3657 & @xmath3658 + @xmath1191 & @xmath3659 & @xmath3660 & @xmath1973 & @xmath3661 & @xmath3662 & @xmath1398 & @xmath3663 & @xmath3664 & @xmath3665 + @xmath1200 & @xmath1456 & @xmath3666 & @xmath3667 & @xmath3668 & @xmath3669 & @xmath3670 & @xmath2110 & @xmath3671 & @xmath3672 + @xmath1210 & @xmath3673 & @xmath3674 & @xmath3675 & @xmath3676 & @xmath3677 & @xmath3678 & @xmath3679 & @xmath3680 & @xmath3681 + @xmath1220 & @xmath3682 & @xmath3683 & @xmath3684 & @xmath3685 & @xmath3686 & @xmath3687 & @xmath3688 & @xmath3689 & @xmath3690 + @xmath1229 & @xmath3691 & @xmath3588 & @xmath3488 & @xmath3491 & @xmath3692 & @xmath3693 & @xmath3694 & @xmath1765 & @xmath1232 + @xmath1239 & @xmath1918 & @xmath3695 & @xmath2871 & @xmath3696 & @xmath3697 & @xmath1520 & @xmath3630 & @xmath1317 & @xmath3698 + @xmath1248 & @xmath3699 & @xmath3700 & @xmath3701 & @xmath3702 & @xmath3703 & @xmath3704 & @xmath3705 & @xmath3706 & @xmath3707 + @xmath1258 & @xmath3708 & @xmath3709 & @xmath3710 & @xmath3711 & @xmath3712 & @xmath3713 & @xmath3714 & @xmath3715 & @xmath3716 + @xmath1268 & @xmath2325 & @xmath3717 & @xmath3718 & @xmath3719 & @xmath2407 & @xmath3720 & @xmath3721 & @xmath3722 & @xmath3723 + @xmath1278 & @xmath3724 & @xmath3725 & @xmath3726 & @xmath3727 & @xmath3728 & @xmath3729 & @xmath1363 & @xmath3730 & @xmath3731 + @xmath1288 & @xmath3732 & @xmath3733 & @xmath3734 & @xmath3735 & @xmath3736 & @xmath3737 & @xmath3738 & @xmath3739 & @xmath3740 + @xmath1298 & @xmath3741 & @xmath3742 & @xmath3743 & @xmath3744 & @xmath3745 & @xmath3746 & @xmath2015 & @xmath3747 & @xmath3748 + @xmath1308 & @xmath1840 & @xmath1231 & @xmath1513 & @xmath3749 & @xmath3750 & @xmath2567 & @xmath1396 & @xmath3751 & @xmath1315 + @xmath1318 & @xmath1021 & @xmath3752 & @xmath3753 & @xmath3754 & @xmath3755 & @xmath3756 & @xmath3757 & @xmath3758 & @xmath3759 + @xmath1327 & @xmath3760 & @xmath3761 & @xmath3762 & @xmath3035 & @xmath3763 & @xmath1957 & @xmath3764 & @xmath3632 & @xmath3765 + @xmath1337 & @xmath3766 & @xmath3767 & @xmath3768 & @xmath3769 & @xmath3770 & @xmath3771 & @xmath3772 & @xmath3773 & @xmath3774 + @xmath1347 & @xmath1460 & @xmath1552 & @xmath3775 & @xmath3776 & @xmath1437 & @xmath3777 & @xmath3778 & @xmath3779 & @xmath3780 + @xmath1357 & @xmath2388 & @xmath3781 & @xmath3782 & @xmath3783 & @xmath3784 & @xmath3785 & @xmath3786 & @xmath3787 & @xmath3788 + @xmath1367 & @xmath3789 & @xmath3790 & @xmath3791 & @xmath3792 & @xmath3793 & @xmath3794 & @xmath3795 & @xmath3796 & @xmath3797 + @xmath1375 & @xmath3798 & @xmath3799 & @xmath3800 & @xmath3801 & @xmath3802 & @xmath3803 & @xmath3804 & @xmath3805 & @xmath3806 + @xmath1385 & @xmath3807 & @xmath3808 & @xmath3809 & @xmath3810 & @xmath3811 & @xmath3812 & @xmath3813 & @xmath3814 & @xmath3815 + @xmath1395 & @xmath1153 & @xmath3816 & @xmath3817 & @xmath3818 & @xmath3819 & @xmath3820 & @xmath3821 & @xmath3822 & @xmath3823 + @xmath1405 & @xmath3824 & @xmath3825 & @xmath1320 & @xmath2240 & @xmath3826 & @xmath3827 & @xmath3828 & @xmath3829 & @xmath3830 + @xmath1415 & @xmath3831 & @xmath3832 & @xmath3833 & @xmath3834 & @xmath3835 & @xmath3836 & @xmath3837 & @xmath3312 & @xmath3838 + @xmath1425 & @xmath3839 & @xmath3840 & @xmath3841 & @xmath3842 & @xmath3843 & @xmath2151 & @xmath3844 & @xmath3845 & @xmath3846 + @xmath1435 & @xmath3847 & @xmath3848 & @xmath3849 & @xmath3850 & @xmath3851 & @xmath3852 & @xmath3853 & @xmath3854 & @xmath3855 + @xmath1445 & @xmath1565 & @xmath3856 & @xmath3857 & @xmath3858 & @xmath3859 & @xmath3860 & @xmath3723 & @xmath3861 & @xmath3862 + @xmath1454 & @xmath1271 & @xmath3863 & @xmath2377 & @xmath3864 & @xmath3865 & @xmath3866 & @xmath3867 & @xmath3868 & @xmath3869 + @xmath1463 & @xmath3348 & @xmath3870 & @xmath3871 & @xmath3872 & @xmath3873 & @xmath3874 & @xmath3875 & @xmath3876 & @xmath3877 + @xmath1473 & @xmath3878 & @xmath3879 & @xmath3880 & @xmath3881 & @xmath3329 & @xmath1840 & @xmath3882 & @xmath3883 & @xmath3884 + @xmath1483 & @xmath3885 & @xmath3886 & @xmath3887 & @xmath3888 & @xmath3138 & @xmath3889 & @xmath3890 & @xmath3891 & @xmath3892 + @xmath1492 & @xmath3893 & @xmath1035 & @xmath3894 & @xmath3895 & @xmath3896 & @xmath3897 & @xmath3838 & @xmath3447 & @xmath3898 + @xmath1502 & @xmath3899 & @xmath1632 & @xmath3900 & @xmath1115 & @xmath3824 & @xmath3901 & @xmath3902 & @xmath1201 & @xmath3903 + @xmath1512 & @xmath2894 & @xmath3904 & @xmath2886 & @xmath2966 & @xmath1848 & @xmath3905 & @xmath3906 & @xmath1912 & @xmath3907 + @xmath1522 & @xmath3908 & @xmath3909 & @xmath3910 & @xmath3911 & @xmath3091 & @xmath3912 & @xmath3520 & @xmath2784 & @xmath2498 + @xmath1531 & @xmath3913 & @xmath3914 & @xmath3915 & @xmath3916 & @xmath3917 & @xmath3918 & @xmath3919 & @xmath3920 & @xmath3921 + @xmath1541 & @xmath3922 & @xmath3923 & @xmath3143 & @xmath3924 & @xmath3925 & @xmath3926 & @xmath3927 & @xmath3928 & @xmath3929 + @xmath1551 & @xmath3930 & @xmath3931 & @xmath3932 & @xmath3933 & @xmath2358 & @xmath3934 & @xmath3935 & @xmath3936 & @xmath3937 + @xmath1561 & @xmath1206 & @xmath3938 & @xmath3939 & @xmath3940 & @xmath3602 & @xmath3941 & @xmath3942 & @xmath3943 & @xmath3944 + @xmath1571 & @xmath3945 & @xmath3946 & @xmath3947 & @xmath3948 & @xmath3949 & @xmath3950 & @xmath3951 & @xmath3952 & @xmath3953 + @xmath1579 & @xmath3041 & @xmath3114 & @xmath2351 & @xmath2965 & @xmath3954 & @xmath1590 & @xmath3955 & @xmath3956 & @xmath3957 + @xmath1587 & @xmath3246 & @xmath3207 & @xmath1831 & @xmath3958 & @xmath3959 & @xmath3960 & @xmath1158 & @xmath1784 & @xmath3961 + @xmath1597 & @xmath3962 & @xmath3963 & @xmath3964 & @xmath3965 & @xmath3966 & @xmath3967 & @xmath3968 & @xmath3969 & @xmath2151 + @xmath1607 & @xmath3970 & @xmath3971 & @xmath3972 & @xmath3973 & @xmath1534 & @xmath3974 & @xmath3975 & @xmath3976 & @xmath3977 + @xmath1617 & @xmath3978 & @xmath3979 & @xmath3980 & @xmath3981 & @xmath3982 & @xmath3983 & @xmath3984 & @xmath3985 & @xmath3986 + @xmath1627 & @xmath1977 & @xmath1811 & @xmath2180 & @xmath3987 & @xmath3988 & @xmath3989 & @xmath3990 & @xmath3991 & @xmath3992 + @xmath1636 & @xmath3993 & @xmath3994 & @xmath3995 & @xmath3120 & @xmath3996 & @xmath1331 & @xmath1970 & @xmath3997 & @xmath3998 + @xmath1645 & @xmath3999 & @xmath4000 & @xmath4001 & @xmath2490 & @xmath2396 & @xmath4002 & @xmath4003 & @xmath4004 & @xmath4005 + @xmath1655 & @xmath4006 & @xmath4007 & @xmath4008 & @xmath4009 & @xmath1946 & @xmath4010 & @xmath4011 & @xmath2548 & @xmath4012 + @xmath1665 & @xmath4013 & @xmath4014 & @xmath4015 & @xmath4016 & @xmath4017 & @xmath2258 & @xmath4018 & @xmath1326 & @xmath4019 + @xmath1675 & @xmath4020 & @xmath1751 & @xmath4021 & @xmath1333 & @xmath4022 & @xmath4023 & @xmath4024 & @xmath4025 & @xmath2137 + @xmath1683 & @xmath2398 & @xmath4026 & @xmath4027 & @xmath4028 & @xmath4029 & @xmath3948 & @xmath4030 & @xmath3781 & @xmath4031 + @xmath1693 & @xmath1827 & @xmath4032 & @xmath4033 & @xmath4034 & @xmath4035 & @xmath4036 & @xmath4037 & @xmath4038 & @xmath2469 + @xmath1701 & @xmath4039 & @xmath4040 & @xmath4041 & @xmath2885 & @xmath2694 & @xmath2676 & @xmath4042 & @xmath4043 & @xmath4044 + @xmath1711 & @xmath4045 & @xmath4046 & @xmath4047 & @xmath4048 & @xmath4049 & @xmath4050 & @xmath4051 & @xmath3038 & @xmath4052 + @xmath1721 & @xmath4053 & @xmath4054 & @xmath4055 & @xmath4056 & @xmath4057 & @xmath4058 & @xmath4059 & @xmath4060 & @xmath4061 + @xmath1731 & @xmath4062 & @xmath4063 & @xmath4064 & @xmath4065 & @xmath4066 & @xmath4067 & @xmath4068 & @xmath4069 & @xmath4070 + @xmath1738 & @xmath4071 & @xmath4072 & @xmath1978 & @xmath1811 & @xmath1203 & @xmath2282 & @xmath4073 & @xmath4074 & @xmath4075 + @xmath1747 & @xmath4076 & @xmath4077 & @xmath2354 & @xmath4078 & @xmath4079 & @xmath4080 & @xmath4081 & @xmath4082 & @xmath4083 + @xmath1756 & @xmath4084 & @xmath4085 & @xmath4086 & @xmath3627 & @xmath2947 & @xmath1699 & @xmath4087 & @xmath4088 & @xmath1787 + @xmath1764 & @xmath2822 & @xmath4089 & @xmath1100 & @xmath4090 & @xmath4091 & @xmath1843 & @xmath4092 & @xmath4093 & @xmath2675 + @xmath1774 & @xmath4094 & @xmath4095 & @xmath4096 & @xmath4097 & @xmath4098 & @xmath2240 & @xmath4099 & @xmath4100 & @xmath4101 + @xmath1782 & @xmath1819 & @xmath4102 & @xmath3957 & @xmath4103 & @xmath4104 & @xmath1763 & @xmath4105 & @xmath4106 & @xmath4107 + @xmath1791 & @xmath1595 & @xmath1592 & @xmath4108 & @xmath4109 & @xmath2136 & @xmath4107 & @xmath3717 & @xmath4110 & @xmath4111 + @xmath1797 & @xmath3954 & @xmath4112 & @xmath4113 & @xmath2697 & @xmath4083 & @xmath4114 & @xmath4115 & @xmath4116 & @xmath2380 + @xmath1806 & @xmath2469 & @xmath4117 & @xmath4118 & @xmath4119 & @xmath4120 & @xmath2490 & @xmath4121 & @xmath3670 & @xmath4122 + @xmath1816 & @xmath4123 & @xmath4124 & @xmath4125 & @xmath3077 & @xmath4126 & @xmath4127 & @xmath1821 & @xmath4128 & @xmath2078 + @xmath1826 & @xmath2755 & @xmath4129 & @xmath3882 & @xmath1464 & @xmath4130 & @xmath1233 & @xmath4131 & @xmath3960 & @xmath2345 + @xmath1835 & @xmath1383 & @xmath4132 & @xmath4133 & @xmath4134 & @xmath4135 & @xmath4136 & @xmath4137 & @xmath4138 & @xmath4139 + @xmath1845 & @xmath2193 & @xmath4033 & @xmath3749 & @xmath4140 & @xmath2356 & @xmath4141 & @xmath4142 & @xmath1583 & @xmath2498 + @xmath1854 & @xmath4143 & @xmath4144 & @xmath4145 & @xmath4146 & @xmath4147 & @xmath4148 & @xmath4149 & @xmath4150 & @xmath4151 + @xmath1864 & @xmath4152 & @xmath4153 & @xmath4154 & @xmath4155 & @xmath4156 & @xmath4157 & @xmath4158 & @xmath4159 & @xmath4160 + @xmath1874 & @xmath4161 & @xmath4162 & @xmath4163 & @xmath4164 & @xmath4165 & @xmath4166 & @xmath4167 & @xmath4168 & @xmath4169 + @xmath1884 & @xmath4170 & @xmath4171 & @xmath3420 & @xmath4172 & @xmath4173 & @xmath4174 & @xmath4175 & @xmath2753 & @xmath4176 + @xmath1894 & @xmath4177 & @xmath4178 & @xmath4179 & @xmath4180 & @xmath4181 & @xmath4182 & @xmath4183 & @xmath4184 & @xmath4185 + @xmath1904 & @xmath4186 & @xmath2351 & @xmath4187 & @xmath4188 & @xmath1529 & @xmath4189 & @xmath1958 & @xmath1677 & @xmath4190 + @xmath1913 & @xmath3589 & @xmath4191 & @xmath3041 & @xmath3243 & @xmath4192 & @xmath2494 & @xmath4193 & @xmath4194 & @xmath4195 + @xmath1923 & @xmath4134 & @xmath4196 & @xmath3420 & @xmath2850 & @xmath3270 & @xmath1235 & @xmath4197 & @xmath2887 & @xmath4198 + @xmath1933 & @xmath4199 & @xmath4200 & @xmath4201 & @xmath2860 & @xmath3420 & @xmath4202 & @xmath3339 & @xmath1827 & @xmath2908 + @xmath1943 & @xmath4203 & @xmath4204 & @xmath4205 & @xmath4206 & @xmath4207 & @xmath4208 & @xmath4209 & @xmath4210 & @xmath4211 + @xmath1953 & @xmath4212 & @xmath4213 & @xmath3957 & @xmath1753 & @xmath4214 & @xmath4215 & @xmath2947 & @xmath2948 & @xmath2411 + @xmath1960 & @xmath4216 & @xmath4217 & @xmath4218 & @xmath4219 & @xmath1232 & @xmath4220 & @xmath4221 & @xmath2163 & @xmath4076 + @xmath1969 & @xmath4222 & @xmath4223 & @xmath4224 & @xmath4225 & @xmath4226 & @xmath4227 & @xmath4228 & @xmath1790 & @xmath4229 + @xmath1979 & @xmath4230 & @xmath4231 & @xmath4232 & @xmath2057 & @xmath1698 & @xmath4233 & @xmath2467 & @xmath2133 & @xmath4234 + @xmath1985 & @xmath1762 & @xmath4235 & @xmath4236 & @xmath4237 & @xmath4238 & @xmath4239 & @xmath1977 & @xmath4240 & @xmath1244 + @xmath1995 & @xmath4241 & @xmath4242 & @xmath3763 & @xmath4243 & @xmath1155 & @xmath2345 & @xmath4244 & @xmath4245 & @xmath3633 + @xmath2003 & @xmath4246 & @xmath2925 & @xmath4247 & @xmath3108 & @xmath4248 & @xmath1911 & @xmath4249 & @xmath4250 & @xmath4251 + @xmath2012 & @xmath4252 & @xmath4253 & @xmath4254 & @xmath4255 & @xmath4256 & @xmath2610 & @xmath4257 & @xmath4258 & @xmath4259 + @xmath2022 & @xmath4260 & @xmath4261 & @xmath4262 & @xmath4263 & @xmath4264 & @xmath4265 & @xmath4266 & @xmath4267 & @xmath4268 + @xmath2031 & @xmath4269 & @xmath1079 & @xmath4270 & @xmath3575 & @xmath4271 & @xmath3577 & @xmath3578 & @xmath3579 & @xmath4272 + @xmath2037 & @xmath4273 & @xmath4274 & @xmath4275 & @xmath4276 & @xmath4277 & @xmath3451 & @xmath1069 & @xmath2046 & @xmath4278 + @xmath2047 & @xmath4279 & @xmath3052 & @xmath4280 & @xmath4281 & @xmath2066 & @xmath4282 & @xmath4283 & @xmath4284 & @xmath4285 + @xmath2056 & @xmath4286 & @xmath4287 & @xmath4288 & @xmath4289 & @xmath3956 & @xmath1807 & @xmath4290 & @xmath4291 & @xmath4292 + @xmath2063 & @xmath3492 & @xmath4293 & @xmath4294 & @xmath4295 & @xmath3116 & @xmath3374 & @xmath4296 & @xmath4297 & @xmath3994 + @xmath2073 & @xmath4298 & @xmath4299 & @xmath1707 & @xmath4300 & @xmath4301 & @xmath2335 & @xmath1519 & @xmath1971 & @xmath4302 + @xmath2080 & @xmath4303 & @xmath4304 & @xmath4305 & @xmath1733 & @xmath4306 & @xmath4307 & @xmath4308 & @xmath1040 & @xmath4309 + @xmath2089 & @xmath4310 & @xmath4311 & @xmath4312 & @xmath1245 & @xmath2381 & @xmath4313 & @xmath1015 & @xmath4314 & @xmath4315 + @xmath2097 & @xmath4316 & @xmath4317 & @xmath4318 & @xmath4319 & @xmath4320 & @xmath4321 & @xmath4322 & @xmath4323 & @xmath4324 + @xmath2107 & @xmath4325 & @xmath4326 & @xmath2195 & @xmath1992 & @xmath4327 & @xmath2282 & @xmath4328 & @xmath4329 & @xmath4330 + @xmath2116 & @xmath4331 & @xmath3764 & @xmath1986 & @xmath4332 & @xmath4333 & @xmath4334 & @xmath4335 & @xmath4336 & @xmath4106 + @xmath2124 & @xmath4107 & @xmath4229 & @xmath4337 & @xmath4338 & @xmath1563 & @xmath4339 & @xmath4340 & @xmath2129 & @xmath1022 + @xmath2131 & @xmath1700 & @xmath1517 & @xmath4291 & @xmath1801 & @xmath4341 & @xmath4342 & @xmath4343 & @xmath2380 & @xmath1975 + @xmath2138 & @xmath1707 & @xmath4344 & @xmath1956 & @xmath4022 & @xmath1128 & @xmath4345 & @xmath4106 & @xmath4346 & @xmath4347 + @xmath2146 & @xmath4348 & @xmath4349 & @xmath4350 & @xmath4351 & @xmath4352 & @xmath4353 & @xmath4354 & @xmath4355 & @xmath4356 + @xmath2156 & @xmath3573 & @xmath1079 & @xmath3574 & @xmath3575 & @xmath3576 & @xmath3577 & @xmath3578 & @xmath3579 & @xmath3580 + @xmath2159 & @xmath4357 & @xmath1238 & @xmath4358 & @xmath2701 & @xmath3349 & @xmath2166 & @xmath3084 & @xmath1479 & @xmath4359 + @xmath2169 & @xmath2562 & @xmath3670 & @xmath4360 & @xmath4361 & @xmath4362 & @xmath4363 & @xmath3604 & @xmath3992 & @xmath4364 + @xmath2178 & @xmath1681 & @xmath3938 & @xmath4229 & @xmath2561 & @xmath4365 & @xmath4366 & @xmath4367 & @xmath4368 & @xmath4369 + @xmath2186 & @xmath4370 & @xmath4371 & @xmath4372 & @xmath4373 & @xmath4374 & @xmath4375 & @xmath4376 & @xmath1470 & @xmath2945 + @xmath2194 & @xmath4377 & @xmath4378 & @xmath4379 & @xmath4380 & @xmath2182 & @xmath4381 & @xmath4382 & @xmath4383 & @xmath4384 + @xmath2201 & @xmath4385 & @xmath4386 & @xmath4387 & @xmath4388 & @xmath4389 & @xmath4390 & @xmath4391 & @xmath3113 & @xmath3507 + @xmath2208 & @xmath4392 & @xmath4393 & @xmath4394 & @xmath4395 & @xmath4396 & @xmath4397 & @xmath4398 & @xmath4399 & @xmath4400 + @xmath2218 & @xmath4401 & @xmath4402 & @xmath4384 & @xmath1567 & @xmath4403 & @xmath4404 & @xmath1121 & @xmath1509 & @xmath4405 + @xmath2227 & @xmath4406 & @xmath4407 & @xmath4408 & @xmath4409 & @xmath4410 & @xmath4411 & @xmath4412 & @xmath3456 & @xmath4413 + @xmath2236 & @xmath4414 & @xmath3723 & @xmath4415 & @xmath4416 & @xmath4417 & @xmath4418 & @xmath4419 & @xmath4420 & @xmath4421 + @xmath2245 & @xmath1992 & @xmath2586 & @xmath4422 & @xmath1629 & @xmath1193 & @xmath4329 & @xmath4423 & @xmath2285 & @xmath4424 + @xmath2254 & @xmath4425 & @xmath4426 & @xmath4427 & @xmath4428 & @xmath4429 & @xmath3708 & @xmath3728 & @xmath2260 & @xmath4430 + @xmath2262 & @xmath4381 & @xmath4431 & @xmath4432 & @xmath1565 & @xmath4433 & @xmath4434 & @xmath4435 & @xmath4436 & @xmath4437 + @xmath2269 & @xmath2136 & @xmath1193 & @xmath4438 & @xmath3992 & @xmath4439 & @xmath4440 & @xmath4441 & @xmath4442 & @xmath4443 + @xmath2278 & @xmath2074 & @xmath4444 & @xmath1982 & @xmath2469 & @xmath4445 & @xmath4446 & @xmath4447 & @xmath2129 & @xmath4448 + @xmath2287 & @xmath4449 & @xmath4450 & @xmath4451 & @xmath4452 & @xmath3697 & @xmath4453 & @xmath1193 & @xmath4454 & @xmath1019 + @xmath2295 & @xmath4455 & @xmath4456 & @xmath4457 & @xmath4458 & @xmath4459 & @xmath4460 & @xmath4461 & @xmath4462 & @xmath4463 + @xmath2305 & @xmath3573 & @xmath1079 & @xmath3574 & @xmath3575 & @xmath3576 & @xmath3577 & @xmath3578 & @xmath3579 & @xmath3580 + @xmath2306 & @xmath4074 & @xmath4464 & @xmath4465 & @xmath4466 & @xmath4467 & @xmath1572 & @xmath2367 & @xmath1406 & @xmath4468 + @xmath2314 & @xmath4469 & @xmath1751 & @xmath4470 & @xmath4471 & @xmath4472 & @xmath1586 & @xmath3960 & @xmath3906 & @xmath4117 + @xmath2322 & @xmath4473 & @xmath1192 & @xmath1628 & @xmath4346 & @xmath4474 & @xmath4377 & @xmath4475 & @xmath2381 & @xmath2145 + @xmath2329 & @xmath4476 & @xmath4112 & @xmath4477 & @xmath2648 & @xmath4478 & @xmath4479 & @xmath2799 & @xmath1642 & @xmath1700 + @xmath2336 & @xmath4480 & @xmath4346 & @xmath4481 & @xmath4482 & @xmath4483 & @xmath3719 & @xmath4484 & @xmath4485 & @xmath4486 + @xmath2344 & @xmath4476 & @xmath1906 & @xmath4487 & @xmath4488 & @xmath4489 & @xmath4490 & @xmath1972 & @xmath2091 & @xmath2145 + @xmath2349 & @xmath3498 & @xmath3382 & @xmath4491 & @xmath1466 & @xmath4492 & @xmath4493 & @xmath3954 & @xmath2766 & @xmath4494 + @xmath126 & @xmath2360 & @xmath2359 & @xmath2358 & @xmath2363 & @xmath2363 & @xmath4495 & @xmath2361 & @xmath2363 & @xmath4430 + @xmath2364 & @xmath4496 & @xmath4497 & @xmath4498 & @xmath4499 & @xmath4439 & @xmath4029 & @xmath4500 & @xmath2177 & @xmath1022 + @xmath2369 & @xmath4501 & @xmath4502 & @xmath1511 & @xmath1448 & @xmath1273 & @xmath4503 & @xmath1576 & @xmath4504 & @xmath4505 + @xmath2379 & @xmath1994 & @xmath4107 & @xmath3663 & @xmath4025 & @xmath1977 & @xmath2559 & @xmath4506 & @xmath4507 & @xmath4475 + @xmath2385 & @xmath4508 & @xmath4509 & @xmath4510 & @xmath2311 & @xmath4511 & @xmath1568 & @xmath4434 & @xmath4512 & @xmath4513 + @xmath2394 & @xmath4514 & @xmath4228 & @xmath1199 & @xmath2421 & @xmath4515 & @xmath4516 & @xmath4517 & @xmath4518 & @xmath1505 + @xmath2403 & @xmath4519 & @xmath4520 & @xmath2561 & @xmath4521 & @xmath4522 & @xmath2177 & @xmath4511 & @xmath4497 & @xmath4405 + @xmath2410 & @xmath4523 & @xmath4524 & @xmath4525 & @xmath4119 & @xmath4506 & @xmath1680 & @xmath1650 & @xmath2183 & @xmath4521 + @xmath2417 & @xmath1583 & @xmath1595 & @xmath1127 & @xmath1398 & @xmath4106 & @xmath2246 & @xmath4526 & @xmath4527 & @xmath4528 + @xmath2426 & @xmath4529 & @xmath4530 & @xmath4531 & @xmath4318 & @xmath2849 & @xmath4532 & @xmath3319 & @xmath2813 & @xmath2961 + @xmath2434 & @xmath3573 & @xmath1079 & @xmath3574 & @xmath3575 & @xmath3576 & @xmath3577 & @xmath3578 & @xmath3579 & @xmath3580 + @xmath2435 & @xmath4533 & @xmath4534 & @xmath4535 & @xmath4536 & @xmath4537 & @xmath4538 & @xmath4539 & @xmath4540 & @xmath4541 + @xmath2445 & @xmath4542 & @xmath4543 & @xmath4544 & @xmath4545 & @xmath4546 & @xmath4547 & @xmath4548 & @xmath4549 & @xmath4550 + @xmath2455 & @xmath2458 & @xmath4551 & @xmath4552 & @xmath4553 & @xmath4554 & @xmath4555 & @xmath4556 & @xmath4557 & @xmath4558 + @xmath2465 & @xmath4237 & @xmath2678 & @xmath4559 & @xmath1986 & @xmath2485 & @xmath4560 & @xmath4319 & @xmath2099 & @xmath4561 + @xmath2474 & @xmath2885 & @xmath4562 & @xmath4563 & @xmath4564 & @xmath4565 & @xmath4566 & @xmath4567 & @xmath4568 & @xmath4569 + @xmath2483 & @xmath4570 & @xmath4571 & @xmath1754 & @xmath4572 & @xmath4342 & @xmath4573 & @xmath4574 & @xmath2179 & @xmath4575 + @xmath2492 & @xmath4576 & @xmath2187 & @xmath4577 & @xmath4578 & @xmath3100 & @xmath2798 & @xmath1759 & @xmath4579 & @xmath2467 + @xmath2499 & @xmath4035 & @xmath2871 & @xmath1954 & @xmath4580 & @xmath3817 & @xmath2560 & @xmath2326 & @xmath4581 & @xmath4582 + @xmath2508 & @xmath4583 & @xmath4151 & @xmath4584 & @xmath4585 & @xmath4586 & @xmath4587 & @xmath4588 & @xmath4589 & @xmath4590 + @xmath2518 & @xmath4591 & @xmath4592 & @xmath4593 & @xmath3840 & @xmath4594 & @xmath4595 & @xmath4596 & @xmath4597 & @xmath4598 + @xmath2528 & @xmath4599 & @xmath4600 & @xmath4601 & @xmath4602 & @xmath4603 & @xmath4604 & @xmath4605 & @xmath4606 & @xmath4607 + @xmath2538 & @xmath1685 & @xmath4608 & @xmath4609 & @xmath4073 & @xmath2180 & @xmath4610 & @xmath2856 & @xmath4611 & @xmath4612 + @xmath2544 & @xmath4613 & @xmath4614 & @xmath4615 & @xmath4616 & @xmath4617 & @xmath4618 & @xmath4619 & @xmath4620 & @xmath4621 + @xmath2554 & @xmath4622 & @xmath2576 & @xmath1986 & @xmath2786 & @xmath4623 & @xmath4624 & @xmath2347 & @xmath1192 & @xmath2506 + @xmath2563 & @xmath4625 & @xmath4626 & @xmath3345 & @xmath3119 & @xmath4627 & @xmath4082 & @xmath2132 & @xmath1982 & @xmath1982 + @xmath2568 & @xmath3960 & @xmath4628 & @xmath1824 & @xmath4629 & @xmath4630 & @xmath4025 & @xmath4631 & @xmath2507 & @xmath2263 + @xmath2572 & @xmath3246 & @xmath3954 & @xmath4632 & @xmath4633 & @xmath3960 & @xmath2466 & @xmath4634 & @xmath4635 & @xmath4636 + @xmath2582 & @xmath1750 & @xmath4637 & @xmath1954 & @xmath1643 & @xmath4638 & @xmath4574 & @xmath1740 & @xmath4639 & @xmath4640 + @xmath2591 & @xmath1068 & @xmath4641 & @xmath4642 & @xmath4643 & @xmath4644 & @xmath4645 & @xmath4646 & @xmath2750 & @xmath4647 + @xmath2599 & @xmath4648 & @xmath4649 & @xmath3501 & @xmath4650 & @xmath4651 & @xmath2193 & @xmath4652 & @xmath2074 & @xmath4653 + @xmath2608 & @xmath4654 & @xmath4655 & @xmath4656 & @xmath4461 & @xmath4657 & @xmath4658 & @xmath4659 & @xmath4660 & @xmath4661 + @xmath2617 & @xmath3573 & @xmath1079 & @xmath3574 & @xmath3575 & @xmath3576 & @xmath3577 & @xmath3578 & @xmath3579 & @xmath3580 + @xmath2618 & @xmath4662 & @xmath4663 & @xmath3985 & @xmath4664 & @xmath4665 & @xmath4666 & @xmath4667 & @xmath4668 & @xmath4669 + @xmath2628 & @xmath4670 & @xmath4671 & @xmath4672 & @xmath4673 & @xmath4674 & @xmath4675 & @xmath4676 & @xmath4677 & @xmath1251 + @xmath2638 & @xmath1134 & @xmath4678 & @xmath4679 & @xmath4680 & @xmath4681 & @xmath4682 & @xmath4683 & @xmath4684 & @xmath4685 + @xmath2647 & @xmath1641 & @xmath4686 & @xmath4687 & @xmath2925 & @xmath4688 & @xmath3870 & @xmath1718 & @xmath2821 & @xmath3325 + @xmath2656 & @xmath4279 & @xmath4689 & @xmath4690 & @xmath1936 & @xmath3393 & @xmath4691 & @xmath4692 & @xmath4648 & @xmath3116 + @xmath2664 & @xmath4693 & @xmath3415 & @xmath2041 & @xmath4694 & @xmath2668 & @xmath4695 & @xmath4696 & @xmath2672 & @xmath4697 + @xmath2674 & @xmath2054 & @xmath4698 & @xmath1932 & @xmath4699 & @xmath4700 & @xmath4082 & @xmath1333 & @xmath4701 & @xmath1763 + @xmath2682 & @xmath4702 & @xmath4703 & @xmath2643 & @xmath4356 & @xmath4704 & @xmath4705 & @xmath4706 & @xmath4707 & @xmath4708 + @xmath2692 & @xmath4709 & @xmath4710 & @xmath4711 & @xmath4712 & @xmath1758 & @xmath4713 & @xmath4714 & @xmath1821 & @xmath4715 + @xmath2698 & @xmath4716 & @xmath4717 & @xmath4718 & @xmath4719 & @xmath3394 & @xmath4720 & @xmath1844 & @xmath1469 & @xmath3208 + @xmath2706 & @xmath4721 & @xmath4722 & @xmath4723 & @xmath4724 & @xmath4725 & @xmath4726 & @xmath4727 & @xmath4728 & @xmath4729 + @xmath2716 & @xmath4730 & @xmath4731 & @xmath1728 & @xmath4732 & @xmath4733 & @xmath4734 & @xmath4735 & @xmath4736 & @xmath4737 + @xmath2726 & @xmath4738 & @xmath1748 & @xmath4739 & @xmath2877 & @xmath2864 & @xmath4740 & @xmath4741 & @xmath4742 & @xmath2971 + @xmath2736 & @xmath4743 & @xmath4744 & @xmath1345 & @xmath4745 & @xmath4746 & @xmath4747 & @xmath4748 & @xmath4749 & @xmath4750 + @xmath2746 & @xmath3492 & @xmath4751 & @xmath4752 & @xmath2098 & @xmath4753 & @xmath1961 & @xmath4754 & @xmath4755 & @xmath3589 + @xmath2756 & @xmath1716 & @xmath4756 & @xmath4644 & @xmath4757 & @xmath4758 & @xmath4759 & @xmath2811 & @xmath4760 & @xmath4761 + @xmath2764 & @xmath4762 & @xmath1748 & @xmath2330 & @xmath2074 & @xmath4763 & @xmath1958 & @xmath1596 & @xmath4701 & @xmath4764 + @xmath2772 & @xmath1099 & @xmath4136 & @xmath1475 & @xmath4281 & @xmath4491 & @xmath4765 & @xmath4766 & @xmath2868 & @xmath4767 + @xmath2780 & @xmath4768 & @xmath4769 & @xmath2330 & @xmath4770 & @xmath4771 & @xmath3628 & @xmath1956 & @xmath2787 & @xmath4772 + @xmath2789 & @xmath2477 & @xmath1941 & @xmath4773 & @xmath1467 & @xmath2888 & @xmath4774 & @xmath4775 & @xmath4776 & @xmath2345 + @xmath2796 & @xmath1467 & @xmath4777 & @xmath1637 & @xmath1919 & @xmath2868 & @xmath4043 & @xmath1595 & @xmath1242 & @xmath4764 + @xmath2803 & @xmath4778 & @xmath4779 & @xmath4780 & @xmath4781 & @xmath4782 & @xmath4783 & @xmath2874 & @xmath2819 & @xmath3059 + @xmath2812 & @xmath4784 & @xmath4785 & @xmath4786 & @xmath4787 & @xmath3090 & @xmath4788 & @xmath2494 & @xmath2676 & @xmath4789 + @xmath2818 & @xmath4790 & @xmath1058 & @xmath3276 & @xmath4791 & @xmath4792 & @xmath4793 & @xmath4794 & @xmath4171 & @xmath2098 + @xmath2827 & @xmath4795 & @xmath4796 & @xmath1499 & @xmath4797 & @xmath4798 & @xmath4799 & @xmath4800 & @xmath4801 & @xmath4802 + @xmath2837 & @xmath4803 & @xmath4618 & @xmath4804 & @xmath4805 & @xmath4806 & @xmath4807 & @xmath4808 & @xmath4809 & @xmath4810 + @xmath2846 & @xmath4811 & @xmath3557 & @xmath4812 & @xmath4813 & @xmath1101 & @xmath2936 & @xmath3382 & @xmath4814 & @xmath4815 + @xmath2854 & @xmath4800 & @xmath4816 & @xmath4817 & @xmath4818 & @xmath4819 & @xmath4820 & @xmath4646 & @xmath4821 & @xmath4822 + @xmath2863 & @xmath2166 & @xmath4565 & @xmath3760 & @xmath4823 & @xmath3247 & @xmath2353 & @xmath1589 & @xmath2565 & @xmath4824 + @xmath2872 & @xmath3037 & @xmath2654 & @xmath4825 & @xmath4826 & @xmath3419 & @xmath4827 & @xmath4828 & @xmath4829 & @xmath2055 + @xmath2880 & @xmath4830 & @xmath4831 & @xmath4832 & @xmath4202 & @xmath3336 & @xmath3198 & @xmath4833 & @xmath4834 & @xmath4835 + @xmath2889 & @xmath4561 & @xmath4836 & @xmath4837 & @xmath4838 & @xmath3496 & @xmath4491 & @xmath4839 & @xmath2919 & @xmath3086 + @xmath2897 & @xmath3235 & @xmath4177 & @xmath4840 & @xmath4841 & @xmath4842 & @xmath4843 & @xmath4136 & @xmath2456 & @xmath4844 + @xmath2907 & @xmath1925 & @xmath2477 & @xmath4845 & @xmath3911 & @xmath4578 & @xmath1130 & @xmath4846 & @xmath4847 & @xmath4848 + @xmath2917 & @xmath3366 & @xmath4576 & @xmath4849 & @xmath4850 & @xmath4851 & @xmath2071 & @xmath4476 & @xmath1589 & @xmath3695 + @xmath2926 & @xmath4852 & @xmath4853 & @xmath2189 & @xmath4854 & @xmath4855 & @xmath4856 & @xmath4653 & @xmath4857 & @xmath3110 + @xmath2934 & @xmath4858 & @xmath4859 & @xmath4860 & @xmath4861 & @xmath2066 & @xmath4768 & @xmath1917 & @xmath2766 & @xmath4862 + @xmath2942 & @xmath3104 & @xmath4863 & @xmath3118 & @xmath1472 & @xmath4246 & @xmath4864 & @xmath4247 & @xmath2484 & @xmath1315 + @xmath2949 & @xmath4865 & @xmath4866 & @xmath4867 & @xmath4868 & @xmath4869 & @xmath4870 & @xmath4871 & @xmath4872 & @xmath4873 + @xmath2959 & @xmath4874 & @xmath4875 & @xmath4876 & @xmath2790 & @xmath1466 & @xmath4877 & @xmath4878 & @xmath4879 & @xmath2965 + @xmath2967 & @xmath4880 & @xmath4881 & @xmath4882 & @xmath4883 & @xmath4884 & @xmath4885 & @xmath1302 & @xmath4886 & @xmath4887 + @xmath2977 & @xmath4888 & @xmath4889 & @xmath4890 & @xmath4891 & @xmath4892 & @xmath4893 & @xmath4894 & @xmath2980 & @xmath4895 + @xmath2986 & @xmath4896 & @xmath4897 & @xmath4898 & @xmath4899 & @xmath3312 & @xmath4900 & @xmath1424 & @xmath1660 & @xmath4901 + @xmath2995 & @xmath4902 & @xmath4903 & @xmath3065 & @xmath4904 & @xmath4905 & @xmath4906 & @xmath4907 & @xmath4908 & @xmath4909 + @xmath3005 & @xmath4910 & @xmath4911 & @xmath4912 & @xmath1390 & @xmath4913 & @xmath4914 & @xmath4915 & @xmath4916 & @xmath4917 + @xmath3014 & @xmath3366 & @xmath1968 & @xmath4918 & @xmath3088 & @xmath3510 & @xmath4919 & @xmath4920 & @xmath4878 & @xmath3091 + @xmath3022 & @xmath4921 & @xmath4922 & @xmath1299 & @xmath4923 & @xmath4924 & @xmath4925 & @xmath4926 & @xmath4927 & @xmath4928 + @xmath3031 & @xmath4929 & @xmath3319 & @xmath4784 & @xmath4373 & @xmath4930 & @xmath4931 & @xmath4932 & @xmath1472 & @xmath4198 + @xmath3036 & @xmath1045 & @xmath4933 & @xmath4934 & @xmath4935 & @xmath4936 & @xmath2849 & @xmath4937 & @xmath4938 & @xmath4939 + @xmath3045 & @xmath1951 & @xmath4940 & @xmath1719 & @xmath4941 & @xmath4942 & @xmath4943 & @xmath4944 & @xmath4945 & @xmath4946 + @xmath3054 & @xmath4947 & @xmath4948 & @xmath4949 & @xmath4950 & @xmath4180 & @xmath4951 & @xmath3335 & @xmath3487 & @xmath4952 + @xmath3063 & @xmath4953 & @xmath4954 & @xmath3276 & @xmath3468 & @xmath4955 & @xmath4956 & @xmath3470 & @xmath3070 & @xmath4957 + @xmath3073 & @xmath2936 & @xmath4958 & @xmath4959 & @xmath4960 & @xmath3190 & @xmath3017 & @xmath3265 & @xmath4961 & @xmath2564 + @xmath3079 & @xmath4962 & @xmath4963 & @xmath3080 & @xmath4964 & @xmath4965 & @xmath4966 & @xmath4967 & @xmath3357 & @xmath4968 + @xmath3087 & @xmath1965 & @xmath3243 & @xmath4969 & @xmath4970 & @xmath2944 & @xmath1930 & @xmath3876 & @xmath1750 & @xmath4971 + @xmath3096 & @xmath2663 & @xmath4972 & @xmath2476 & @xmath4973 & @xmath4974 & @xmath4975 & @xmath2965 & @xmath1328 & @xmath4976 + @xmath3103 & @xmath4977 & @xmath4978 & @xmath2765 & @xmath2564 & @xmath2911 & @xmath2332 & @xmath4979 & @xmath1753 & @xmath2132 + @xmath3111 & @xmath4980 & @xmath2904 & @xmath2891 & @xmath1968 & @xmath4981 & @xmath3075 & @xmath4982 & @xmath1695 & @xmath1470 + @xmath3115 & @xmath4983 & @xmath2893 & @xmath4984 & @xmath2938 & @xmath4985 & @xmath4769 & @xmath4560 & @xmath4986 & @xmath4987 + @xmath3122 & @xmath4988 & @xmath3798 & @xmath4989 & @xmath4990 & @xmath4991 & @xmath4992 & @xmath3895 & @xmath4993 & @xmath4994 + @xmath3132 & @xmath4995 & @xmath4996 & @xmath4997 & @xmath4998 & @xmath4999 & @xmath5000 & @xmath5001 & @xmath5002 & @xmath5003 + @xmath3142 & @xmath5004 & @xmath5005 & @xmath5006 & @xmath5007 & @xmath5008 & @xmath5009 & @xmath5010 & @xmath5011 & @xmath5012 + @xmath3152 & @xmath5013 & @xmath5014 & @xmath5015 & @xmath5016 & @xmath5017 & @xmath5018 & @xmath5019 & @xmath5020 & @xmath5021 + @xmath3162 & @xmath2955 & @xmath5022 & @xmath5023 & @xmath5024 & @xmath5025 & @xmath3026 & @xmath1384 & @xmath3170 & @xmath5026 + @xmath3171 & @xmath5027 & @xmath5028 & @xmath5029 & @xmath5030 & @xmath2841 & @xmath5031 & @xmath5032 & @xmath5033 & @xmath5034 + @xmath3179 & @xmath5035 & @xmath5036 & @xmath4708 & @xmath5037 & @xmath5038 & @xmath5039 & @xmath5040 & @xmath4012 & @xmath5041 + @xmath3189 & @xmath1925 & @xmath3341 & @xmath5042 & @xmath1844 & @xmath5043 & @xmath5044 & @xmath5045 & @xmath4476 & @xmath2816 + @xmath3195 & @xmath5046 & @xmath5047 & @xmath3333 & @xmath5048 & @xmath3058 & @xmath5049 & @xmath4295 & @xmath5050 & @xmath5051 + @xmath3202 & @xmath5052 & @xmath4281 & @xmath1925 & @xmath1236 & @xmath5053 & @xmath5054 & @xmath1469 & @xmath3876 & @xmath2817 + @xmath3210 & @xmath5055 & @xmath5056 & @xmath5057 & @xmath5058 & @xmath5059 & @xmath5060 & @xmath5061 & @xmath5062 & @xmath5063 + @xmath3220 & @xmath5064 & @xmath5065 & @xmath5066 & @xmath5067 & @xmath5068 & @xmath5069 & @xmath4690 & @xmath4836 & @xmath5070 + @xmath3228 & @xmath5071 & @xmath5072 & @xmath5073 & @xmath1087 & @xmath2804 & @xmath5074 & @xmath3068 & @xmath3177 & @xmath5075 + @xmath3237 & @xmath5076 & @xmath5077 & @xmath5078 & @xmath3201 & @xmath5079 & @xmath1694 & @xmath4032 & @xmath1932 & @xmath2653 + @xmath3241 & @xmath4183 & @xmath5080 & @xmath3336 & @xmath3085 & @xmath3358 & @xmath5081 & @xmath5082 & @xmath2675 & @xmath1471 + @xmath3249 & @xmath5083 & @xmath4931 & @xmath2192 & @xmath3106 & @xmath5084 & @xmath2912 & @xmath2867 & @xmath1698 & @xmath1315 + @xmath3256 & @xmath5033 & @xmath5085 & @xmath4951 & @xmath5086 & @xmath5087 & @xmath1925 & @xmath5088 & @xmath5089 & @xmath1927 + @xmath3266 & @xmath5090 & @xmath5091 & @xmath5092 & @xmath5093 & @xmath5094 & @xmath3227 & @xmath5095 & @xmath1237 & @xmath3097 + @xmath3271 & @xmath5096 & @xmath5097 & @xmath5098 & @xmath5099 & @xmath5100 & @xmath5101 & @xmath5102 & @xmath5103 & @xmath5104 + @xmath3281 & @xmath5105 & @xmath5106 & @xmath5107 & @xmath5108 & @xmath5109 & @xmath5110 & @xmath4151 & @xmath5111 & @xmath5112 + @xmath3291 & @xmath5113 & @xmath5114 & @xmath3804 & @xmath5115 & @xmath5116 & @xmath5117 & @xmath5118 & @xmath5119 & @xmath5120 + @xmath3301 & @xmath5121 & @xmath5122 & @xmath5123 & @xmath5124 & @xmath2686 & @xmath5125 & @xmath5126 & @xmath5127 & @xmath5128 + @xmath3311 & @xmath5129 & @xmath5130 & @xmath5131 & @xmath5132 & @xmath5133 & @xmath5134 & @xmath5135 & @xmath5136 & @xmath5137 + @xmath3318 & @xmath2163 & @xmath1914 & @xmath1236 & @xmath4973 & @xmath4773 & @xmath5138 & @xmath4854 & @xmath5139 & @xmath5140 + @xmath3323 & @xmath3172 & @xmath4780 & @xmath5141 & @xmath5142 & @xmath5143 & @xmath4813 & @xmath5144 & @xmath1891 & @xmath2936 + @xmath3331 & @xmath4909 & @xmath5145 & @xmath5146 & @xmath5147 & @xmath5148 & @xmath5149 & @xmath5150 & @xmath3370 & @xmath5151 + @xmath3340 & @xmath3371 & @xmath5152 & @xmath5153 & @xmath5154 & @xmath4040 & @xmath5155 & @xmath5156 & @xmath3750 & @xmath5157 + @xmath3347 & @xmath5158 & @xmath5159 & @xmath5160 & @xmath5161 & @xmath3203 & @xmath2730 & @xmath3509 & @xmath5162 & @xmath5054 + @xmath3353 & @xmath4322 & @xmath5163 & @xmath3365 & @xmath2163 & @xmath1938 & @xmath5164 & @xmath5165 & @xmath5166 & @xmath5167 + @xmath3361 & @xmath5168 & @xmath5169 & @xmath5170 & @xmath4319 & @xmath5171 & @xmath5172 & @xmath5173 & @xmath5174 & @xmath4875 + @xmath3367 & @xmath5175 & @xmath5176 & @xmath5177 & @xmath1765 & @xmath4295 & @xmath2207 & @xmath5178 & @xmath5179 & @xmath5180 + @xmath3377 & @xmath5181 & @xmath5182 & @xmath4200 & @xmath4717 & @xmath5183 & @xmath4173 & @xmath5158 & @xmath5184 & @xmath5185 + @xmath3385 & @xmath5186 & @xmath2655 & @xmath4793 & @xmath5187 & @xmath5188 & @xmath4171 & @xmath5189 & @xmath5190 & @xmath3082 + @xmath3395 & @xmath5191 & @xmath5192 & @xmath5193 & @xmath5194 & @xmath5195 & @xmath5135 & @xmath1383 & @xmath5196 & @xmath5197 + @xmath3402 & @xmath5198 & @xmath5199 & @xmath5200 & @xmath5201 & @xmath5202 & @xmath4048 & @xmath3406 & @xmath5203 & @xmath5204 + @xmath3410 & @xmath5205 & @xmath5206 & @xmath3455 & @xmath5036 & @xmath2992 & @xmath5207 & @xmath5208 & @xmath5209 & @xmath5210 + @xmath3418 & @xmath1885 & @xmath4181 & @xmath5211 & @xmath4946 & @xmath3383 & @xmath5212 & @xmath1903 & @xmath5213 & @xmath4918 + @xmath3425 & @xmath5214 & @xmath5215 & @xmath3895 & @xmath5216 & @xmath5217 & @xmath1045 & @xmath5218 & @xmath5219 & @xmath1501 + @xmath3434 & @xmath5220 & @xmath1071 & @xmath5221 & @xmath5222 & @xmath5223 & @xmath3441 & @xmath5224 & @xmath5225 & @xmath5226 + @xmath3443 & @xmath5227 & @xmath5228 & @xmath5229 & @xmath5230 & @xmath5231 & @xmath4012 & @xmath2830 & @xmath5232 & @xmath5233 + @xmath3452 & @xmath5234 & @xmath5235 & @xmath5236 & @xmath5237 & @xmath4204 & @xmath5238 & @xmath3303 & @xmath5239 & @xmath3428 + @xmath3460 & @xmath3485 & @xmath5240 & @xmath2459 & @xmath2826 & @xmath4173 & @xmath2100 & @xmath5241 & @xmath5242 & @xmath5243 + @xmath3467 & @xmath5244 & @xmath1423 & @xmath1070 & @xmath5073 & @xmath5245 & @xmath5246 & @xmath5247 & @xmath5248 & @xmath2541 + @xmath3476 & @xmath5249 & @xmath5250 & @xmath1099 & @xmath4951 & @xmath2862 & @xmath1899 & @xmath4175 & @xmath5251 & @xmath5252 + @xmath3484 & @xmath5253 & @xmath1099 & @xmath2657 & @xmath2099 & @xmath4181 & @xmath5254 & @xmath2890 & @xmath5255 & @xmath5256 + @xmath3490 & @xmath5144 & @xmath3492 & @xmath5257 & @xmath5258 & @xmath5259 & @xmath5260 & @xmath5261 & @xmath2918 & @xmath5262 + @xmath3497 & @xmath5263 & @xmath5264 & @xmath3589 & @xmath2102 & @xmath5265 & @xmath1767 & @xmath5266 & @xmath1235 & @xmath5267 + @xmath3503 & @xmath2471 & @xmath3487 & @xmath4754 & @xmath5268 & @xmath3366 & @xmath3341 & @xmath5269 & @xmath4968 & @xmath5270 + asplund , m. , grevesse , n. , sauval , a. j. & scott , p. 2009
, , 47 , 481 brott , i. , hunter , i. , anders , p. , & langer , n. 2008 , in first stars iii , eds . b. w. oshea & a. heger , aip conf .
ser . , 990 , 273 brott , i. , de mink , s. e. , cantiello , m. , langer , n. , de koter , a. , evans , c. j. , hunter , i. , trundle , c. & vink , j. s. 2011 , , 530 , 115 chaboyer , b. & zahn , j .- p .
1992 , , 253 , 173 chieffi , a. , limongi , m. , & straniero , o. 2000 , in `` the evolution of the milky way '' , eds . f. matteucci and f. giovannelli , kluwer academic publisher , p.417 colella , p. & woodward , p. r. 1984 , j. comput .
phys . , 54 , 174 decressin , t. , meynet , g. , charbonnel , c. , prantzos , n. , & ekstrm , s. 2007 , , 464 , 1029 dessart , l. , hillier , d.j . , livne , e. , yoon , s.c . , woosley , s. , waldman , r. & langer , n . 2011 , , 414 , 2985 dunstall , p. r. , brott , i. , dufton , p. l. , et al . 2011 , , 536 , a6 de jager , c. , nieuwenhuijzen , h. , & van der hucht , k. a. 1988 , , 72 , 259 eddington , a. s. 1929 , , 90 , 54 ekstrm , s. , meynet , g. , chiappini , c. , hirschi , r. , & maeder , a. 2008 , , 489 , 685 ekstrm , s. , georgy , c. , eggenberger , p. , et al .
2012 , , 537 , 146 endal , a. s. , & sofia , s. 1976 , , 210 , 184 georgy , c. , ekstrm , s. , meynet , g. , et al .
2012 , , 542 , a29 gratton , l. 1945 , , 17 , 5 hachinger , s. , mazzali , p. a. , taubenberger , s. , et al .
2012 , , 422 , 70 heger , a. , & langer , n. 2000 , , 544 , 1016 heger , a. , langer , n. , & woosley , s. e. 2000 , , 528 , 368 heger , a. , woosley , s. e. , & spruit , h. c. 2005 , , 626 , 350 hirschi , r. , meynet , g. , & maeder , a. 2004 , , 425 , 649 hunter , i. , dufton , p. l. , smartt , s. j. , et al .
2007 , , 466 , 277 hunter , i. , lennon , d. j. , dufton , p. l. , et al .
2008 , , 479 , 541 hunter , i. , brott , i. , langer , n. , et al .
2009 , , 496 , 841 imbriani , g. , limongi , m. , gialanella , l. , terrasi , f. , straniero , o. & chieffi , a. 2001 , , 558 , 903 kaspi , v.m . ,
manchester , r.n . , siegman , b. , johnston , s. & lyne , a. g. 1994 , , 422 , l83 kato s. 1966 , , 18 , 374 kippenhahn , r. , & thomas , h .- c .
1970 , iau colloq . 4 : stellar rotation , 20 kippenhahn , r. , & weigert 1990 , `` book - review - stellar structure and evolution '' , , 80 , 5 langer , n. 1989 , , 220 , 135 langer , n. 1991 , a&a , 252 , 669 langer , n. , & maeder , a. 1995 , , 295 , 685 levesque , e.m . , massey ,
p. , olsen , k.a.g .
, pletz , b. , josselin , e. , maeder , a. & meynet , g. 2005 , , 628 , 973 limongi , m. & chieffi , a. 2003 , , 592 , 404 limongi , m. & chieffi , a. 2006 , , 647 , 483 limongi , m. , & chieffi , a. 2012 , , 199 , 38 maeder , a. 1995 , , 299 , 84 maeder , a. 1997 , , 321 134 maeder , a. 2003 , , 399 263 maeder and zahn 1998 , 334 , 1000 maeder , a. & meynet , g. 1994 , , 287 , 803 maeder , a. & meynet , g. 1996 , , 313 , 140 marshall , f. e. , gotthelf , e. v. , zhang , w . , middleditch , j. , & wang , q. d. 1998 , , 499 , l179 massey , p. 2003
, , 41 , 15 mathis , s. , palacios , a. & zahn , j .- p .
2004 , , 425 , 243 meynet & maeder 1997 , , 321 , 465 meynet & maeder 2000 , , 361 , 101 meynet , g. & maeder , a. 2003 , , 404 , 975 meynet , g. & maeder , a. 2005 , , 429 , 581 mokiem , m. r. , de koter , a. , evans , c. j. , et al .
2006 , , 456 , 1131 mokiem , m. r. , de koter , a. , evans , c. j. , et al .
2007 , , 465 , 1003 muslimov , a. & page , d. 1996 , , 458 , 347 nugis , t. , & lamers , h. j. g. l. m. 2000 , , 360 , 227 palacios , talon , a.s . , charbonnel1 , c. & forestini , m. 2003 , , 399 , 603 pinsonneault , m. h. , kawaler , s. d. , sofia , s. , & demarque , p. 1989
, , 338 , 424 smartt , s. j. 2009 , , 47 , 63 straniero .
o. , chieffi , a. & limongi , m. 1997 , , 490 , 425 talon , s. , zahn , j.p . , maeder , a. & meynet , g. 1997 , , 322 , 209 talon , s. & zahn , j.p .
1997 , 317 , 749 trundle , c. , dufton , p. l. , hunter , i. , et al .
2007 , , 471 , 625 van der hucht , k.a . 2006 , , 458 , 453 van loon , j. t. , cioni , m .-
l. , zijlstra , a. a. , & loup , c. 2005 , , 438 , 273 vink , j. s. , de koter , a. , & lamers , h. j. g. l. m. 2000 , , 362 , 295 vink , j. s. , de koter , a. , & lamers , h. j. g. l. m. 2001 , , 369 , 574 von zeipel , h. 1924 , , 84 , 665 woosley , s.e . , & heger , a. 2006 , , 637 , 914 yoon , s .- c . , & langer , n. 2005 , , 443 , 643 yoon , s .- c . , langer , n. , & norman , c. 2006 , , 460 , 199 zahn , j. p. , 1974
, in `` stellar instabilities and evolution '' , iau symp 59 , ed .
p. ledoux et al . , reidel publ .
co. , p. 1185
zahn , j. p. , 1992 , 265 , 115 | we present the first set of a new generation of models of massive stars of solar composition extending between 13 and 120 @xmath0 , computed with and without the effects of rotation .
we included two instabilities induced by rotation , namely the meridional circulation and the shear instability .
we implemented two alternative schemes to treat the transport of the angular momentum : the advection - diffusion formalism and the simpler purely diffusive one .
the full evolution from the pre main sequence up to the presupernova stage is followed in detail with a very extended nuclear network .
the explosive yields are provided for a variety of possible mass cut and are available at the website http://www.iasf-roma.inaf.it/orfeo/public{_}html[http://www.iasf-roma.inaf.it/orfeo/public\{_}html ] .
we find that both the he and the co core masses are larger than those of their non rotating counterparts .
also the c abundance left by the he burning is lower than in the non rotating case , especially for stars of initial mass 13 - 25 @xmath0 , and this affects the final mass - radius relation , basically the final binding energy , at the presupernova stage .
the elemental yields produced by a generation of stars rotating initially at 300 km / s do not change substantially with respect to those produced by a generation of non rotating massive stars , the main differences being a slight overproduction of the weak _ s_-component and a larger production of f. since rotation also affects the mass loss rate , either directly and indirectly , we find substantial differences in the lifetimes as o - type and wr - subtypes between rotating and non rotating models .
the maximum mass exploding as type iip supernova ranges between 15 and 20@xmath0 in both sets of models ( this value depending basically on the larger mass loss rates in the red super giant phase due to the inclusion of the dust driven wind ) .
this limiting value is in remarkable good agreement with current estimates . |
the discovery of the standard model ( sm)-like higgs at the cern large hadron collider ( lhc ) provides the last link towards validating the sm @xcite .
however , much work still remains in establishing the nature of the higgs boson , as well as studying possible extensions of the sm higgs sector .
a complete characterization of all possible decay modes of the sm - like higgs is crucial . while the discovery of the higgs relied on conventional search channels @xmath0 , @xmath1 , @xmath2 , @xmath3 and @xmath4 with event rates compatible with sm predictions , the limit on the branching ratio to exotic states is rather weak @xcite .
if the higgs decay branching fractions indeed show a departure from the sm , it would become very important to get a handle on possible exotic decays as they will help narrow down the various possibilities beyond the sm .
a particularly interesting example is the exotic decay of the observed sm - like higgs @xmath17 with a light cp - odd higgs @xmath8 .
many beyond the sm scenarios are constructed by extending the higgs sector - well known examples are the minimal supersymmetric standard model ( mssm ) @xcite , next to minimal supersymmetric standard model ( nmssm ) @xcite and two - higgs doublet models ( 2hdm)@xcite .
in addition to the sm - like higgs boson in these models , there are other cp - even higgses , cp - odd higgses , as well as charged ones .
other than the decay of those extra higgses into the sm - final states @xmath0 , @xmath1 , @xmath2 , @xmath3 and @xmath4 , which have been the focus of the current higgs searches , the decay of heavy higgses into light higgses , or higgs plus gauge boson could also be sizable . in this study , we explore the decay of @xmath6 or @xmath5 , with @xmath18 and @xmath8 referring to the cp - even and cp - odd higgs respectively . in generic 2hdm or nmssm ,
both decays @xmath19 and @xmath20 could appear with large branching fractions @xcite .
@xmath21 could also have a sizable cross - section in the low @xmath22 region of the mssm with the light cp - even @xmath23 being sm - like @xcite . in this work
, we study the process @xmath10 for @xmath11 , @xmath12 , exploring the exclusion bounds as well as discovery reach at the lhc for various combinations of @xmath9 . for simplicity
, we use @xmath5 to refer to both types of decays . since we do not make use of angular correlations , the results obtained for @xmath5 apply to @xmath6 as well with the values of @xmath24 and @xmath25 switched .
details of our analysis can be found in ref .
a similar analysis of @xmath26 in the 2hdm has also been carried out in ref .
@xcite .
the paper is organized as follows . in sec .
[ sec : analysis ] , we briefly present our analysis describing the efficiencies of the cuts employed for both the signal and dominant sm backgrounds . in sec .
[ sec : results ] , we present the 95% c.l .
exclusion as well as 5@xmath15 discovery limits for @xmath27 at the 14 tev lhc with 100 , 300 and 1000 @xmath13 integrated luminosity . to make connection with existing models
, we also compare these exclusion and discovery limits to the case of the nmssm .
we conclude in sec .
[ sec : conclusions ] .
we study the dominant gluon fusion higgs production @xmath28 with @xmath29 and @xmath30 for @xmath31 . the dominant sm backgrounds for @xmath32 final states are @xmath33 with leptonic @xmath34 decay , @xmath35 with leptonically decaying top quarks , @xmath36 , and @xmath37 @xcite .
we have ignored the subdominant backgrounds from @xmath38 , @xmath2 , @xmath39 , @xmath40 , multijet qcd background , @xmath41 , @xmath42 as well as @xmath43 .
those backgrounds either have small production cross sections , or can be sufficiently suppressed by the cuts imposed .
we use madgraph 5/madevent v1.5.11 @xcite to generate our signal and background events .
these events are passed to pythia v2.1.21 @xcite to simulate initial and final state radiation , showering and hadronization .
the events are further passed through delphes 3.07 @xcite with the snowmass combined lhc detector card @xcite to simulate detector effects . for the signal process
, we generated event samples at 14 tev lhc for @xmath28 with the daughter particle mass fixed at @xmath44200 , 125 , and 50 gev while varying the parent particle mass @xmath24 in the range 150 @xmath16 600 gev .
we applied the following cuts to identify the signal from the backgrounds background . however , including pile - up effects introduces @xmath45 in the signal events , which renders the cut inefficient .
we thank meenakshi narain and john stupak for pointing this out to us . ] : * * two leptons , two tagged @xmath46 s : * we require exactly two identified isolated leptons and two tagged bottom jets using the snowmass delphes detector card : @xmath47 this implies that the leptons and bottom jets are in the central detector region @xmath48 and satisfy a minimum transverse momentum cut of @xmath49 gev and @xmath50 15 gev . for jet reconstruction , the anti-@xmath51 jet algorithm with @xmath52=0.5 is used . *
* lepton trigger : * we require minimum transverse momenta for the leptons to pass the trigger : @xmath53 * * dilepton mass @xmath54 : * we require the dilepton mass to be in the @xmath34-mass window : @xmath55 * * @xmath56 vs. @xmath57 : * we require the dijet mass @xmath56 to be close to the daughter - higgs mass @xmath25 and the mass @xmath58 to be close to the parent - higgs mass @xmath59 .
the two masses are correlated , i.e. , if we underestimate @xmath56 we also underestimate @xmath57 . to take this into account
we apply a two - dimensional cut : @xmath60 where @xmath61 is the width for the dijet mass window .
note that the slightly shifted reconstructed higgs mass @xmath56 ( 0.95 @xmath25 instead of @xmath25 ) is due to the reconstruction of the @xmath46-jet with a small size of @xmath52=0.5 .
the second condition describes two lines going through the points @xmath62 with slope @xmath63 .
we choose a width for the @xmath64 peak of @xmath65 where @xmath66 is the width of a sm higgs with mass @xmath24 .
this accounts for both small higgs masses for which the width of the peak is caused by detector effects and large higgs masses for which the physical width dominates .
the effectiveness of this cut is shown in fig .
[ fig:2d_mbb_mllbb ] for @xmath67 gev and @xmath68 gev , with two horizontal lines indicating the @xmath56 range and two slanted lines indicating the @xmath58 range as given in eq .
( [ cut_5 ] ) .
+ -axis ) of @xmath56 vs. @xmath64 for the signal ( left ) , and the dominating @xmath69 ( central ) and @xmath35 ( right ) backgrounds for @xmath67 gev and @xmath68 gev .
two horizontal lines indicate the @xmath56 range and two slanted lines indicate the @xmath58 range , as given in eq .
( [ cut_5 ] ) .
, title="fig : " ] -axis ) of @xmath56 vs. @xmath64 for the signal ( left ) , and the dominating @xmath69 ( central ) and @xmath35 ( right ) backgrounds for @xmath67 gev and @xmath68 gev .
two horizontal lines indicate the @xmath56 range and two slanted lines indicate the @xmath58 range , as given in eq .
( [ cut_5 ] ) .
, title="fig : " ] -axis ) of @xmath56 vs. @xmath64 for the signal ( left ) , and the dominating @xmath69 ( central ) and @xmath35 ( right ) backgrounds for @xmath67 gev and @xmath68 gev .
two horizontal lines indicate the @xmath56 range and two slanted lines indicate the @xmath58 range , as given in eq .
( [ cut_5 ] ) .
, title="fig : " ] * * transverse momentum : * we require the sum of the transverse momenta of the bottom jets and the sum of the transverse momenta of the bottom jets and leptons to be large : @xmath70 .signal and background cross sections with cuts for the signal benchmark point @xmath59 = 300 gev and @xmath25 = 125 gev at the 14 tev lhc .
we have chosen a nominal value for @xmath27 of 100 fb to illustrate the cut efficiencies for the signal process .
the last column of @xmath71 is shown for an integrated luminosity of @xmath72 . [ cols="<,>,>,>,>,>,>,>",options="header " , ] in table [ tab : cuts ] , we show the signal and background cross sections with cuts for signal benchmark point of @xmath59 = 300 gev and @xmath25 = 125 gev at the 14 tev lhc .
we have chosen a nominal value for @xmath27 of 100 fb to illustrate the cut efficiencies for the signal process .
the last column of @xmath71 is shown for an integrated luminosity of @xmath72 .
in fig . [ fig : cslimit ] , we show the 95% c.l .
exclusion ( left panel ) and 5@xmath15 discovery ( right panel ) reach of @xmath27 at the 14 tev lhc with 300 @xmath13 integrated luminosity as a function of @xmath24 for @xmath44 50 , 125 and 200 gev . also shown
are the reach of 100 @xmath13 ( dashed lines ) and 1000 @xmath13 ( dash - dotted lines ) for the case of @xmath73 gev , as well as the reach assuming 10% systematic error for the backgrounds ( purple lines ) .
the exclusion / discovery limits were calculated using the theta - auto program @xcite . for 300 @xmath13 integrated luminosity with @xmath73 gev , the 95% c.l .
limit is about 26 ( 2.2 ) fb for @xmath74 gev , while the 5 @xmath15 discovery limit is about 65 ( 5.6 ) fb .
the limits typically get better for large @xmath24 and smaller @xmath25 , except for very large @xmath24 with very small @xmath25 , when the two @xmath46 jets are highly collimated and thus do not pass the two @xmath46 jets requirement .
typical range of @xmath75 in the nmssm @xcite are plotted as scattered dots for the daughter higgs mass being @xmath76 gev , which quickly drops for higgs mass around 350 gev due to the opening of the @xmath35 channel .
@xmath5 , or @xmath6 , with @xmath77 final states , therefore , could be a useful channel for heavy higgs searches , especially for higgs mass below 350 gev . in the generic type - ii 2hdm ,
@xmath78 could even compete with the @xmath35 channel for higgs mass above 350 gev @xcite .
reach ( right panel ) in @xmath27 at the 14 tev lhc with 300 @xmath13 integrated luminosity as a function of the parent particle mass @xmath24 for various daughter particle mass @xmath44 50 , 125 and 200 gev . also shown are the reach of 100 @xmath13 ( dashed lines ) and 1000 @xmath13 ( dash - dotted lines ) for the case of @xmath73 gev , as well as the reach assuming 10% systematic error for the backgrounds ( purple lines ) .
scattering dots are the possible @xmath75 range in the nmssm for the daughter higgs mass being @xmath76 gev.,title="fig : " ] reach ( right panel ) in @xmath27 at the 14 tev lhc with 300 @xmath13 integrated luminosity as a function of the parent particle mass @xmath24 for various daughter particle mass @xmath44 50 , 125 and 200 gev . also shown
are the reach of 100 @xmath13 ( dashed lines ) and 1000 @xmath13 ( dash - dotted lines ) for the case of @xmath73 gev , as well as the reach assuming 10% systematic error for the backgrounds ( purple lines ) .
scattering dots are the possible @xmath75 range in the nmssm for the daughter higgs mass being @xmath76 gev.,title="fig : " ]
while the conventional searches for the higgs boson focus on its decays to sm final states : @xmath0 , @xmath1 , @xmath2 , @xmath3 and @xmath4 , in this study , we explored the decay @xmath6 or @xmath5 with @xmath79 final state and presented model independent limits on the 95% c.l .
exclusion and @xmath80 discovery of the heavy higgs in this channel . for 14 tev lhc with 300 @xmath13 integrated luminosity , the 95% c.l .
limits on @xmath75 vary between 30 fb to a few fb for the parent heavy higgs mass in the range of 200 gev to 600 gev , while the limit for 5@xmath15 discovery is about 2@xmath163 times larger . comparing with the specific case of nmssm
, the @xmath81 channel can be useful for higgs masses all the way up to 350 gev , when the @xmath35 channel opens up .
the reach could be even better in the generic 2hdm .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 717 * , 70 ( 2012 ) [ arxiv:1206.2443 [ hep - ex ] ]
. g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 718 * , 369 ( 2012 ) [ arxiv:1207.0210 [ hep - ex ] ] .
[ cms collaboration ] , cms - pas - hig-11 - 017 .
[ cms collaboration ] , cms - pas - hig-13 - 012 .
[ atlas collaboration ] , atlas - conf-2013 - 034 .
[ cms collaboration ] , cms - pas - hig-13 - 005 .
h. p. nilles , phys .
* 110 * , 1 ( 1984 ) .
h. e. haber and g. l. kane , phys .
* 117 * , 75 ( 1985 ) .
r. barbieri , riv .
nuovo cim . * 11n4 * , 1 ( 1988 ) . j. r. ellis , j. f. gunion , h. e. haber , l. roszkowski and f. zwirner , phys . rev .
d * 39 * , 844 ( 1989 ) .
m. drees , int .
j. mod .
phys . a * 4 * , 3635 ( 1989 ) .
g. c. branco , p. m. ferreira , l. lavoura , m. n. rebelo , m. sher and j. p. silva , phys .
rept . * 516 * , 1 ( 2012 ) [ arxiv:1106.0034 [ hep - ph ] ] .
haber , g.l .
kane and t. sterling , nucl .
* b161 * , 493 ( 1979 ) .
f. febres cordero , l. reina and d. wackeroth , phys .
d * 80 * , 034015 ( 2009 ) [ arxiv:0906.1923 [ hep - ph ] ] .
n. kidonakis , arxiv:1205.3453 [ hep - ph ] .
j. m. campbell , r. k. ellis and c. williams , jhep * 1107 * , 018 ( 2011 ) [ arxiv:1105.0020 [ hep - ph ] ] .
s. dittmaier _ et al . _
[ lhc higgs cross section working group collaboration ] , arxiv:1101.0593 [ hep - ph ] .
s. ovyn , x. rouby and v. lemaitre , arxiv:0903.2225 [ hep - ph ] .
a. avetisyan _ et .
_ , snowmass energy frontier simulations for hadron colliders " .
ott / theta / theta - auto/ | while the conventional searches for the higgs boson focus on its decays to standard model ( sm ) particles : @xmath0 , @xmath1 , @xmath2 , @xmath3 and @xmath4 , in this study , we explore the decays @xmath5 or @xmath6 .
such decays could appear in the exotic decay of the observed sm - like higgs @xmath7 with a light @xmath8 , or for extra higgses in beyond the sm new physics scenarios with an extended higgs sector .
we study the exclusion bounds as well as discovery reach at the lhc for various combinations of @xmath9 for the process : @xmath10 for @xmath11 , @xmath12 .
we found that for 14 tev lhc with 300 @xmath13 integrated luminosity , the 95% c.l .
limits on @xmath14 vary between about 30 fb to a few fb for the parent heavy higgs mass in the range 200 gev to 600 gev , while the limits for 5@xmath15 discovery are about 2@xmath163 times larger . comparing with the specific case of next to minimal supersymmetric standard model
, we find that this channel could be useful for heavy higgs searches , especially for higgs masses below 350 gev . |
deep multi - wavelength lookback surveys carried out over the past five years have improved our understanding of galaxy evolution in the young universe tremendously . at least over the past 10 gyr of lookback time , a correlation between the rate of star formation and the amount of assembled stellar mass in star - forming galaxies ( sfgs ) , dubbed the main sequence of star formation ( ms ) , has been observed ( noeske et al .
2007 ; elbaz et al . 2007 ; daddi et al . 2007 ) .
while the scatter around the ms is not observed to evolve strongly with redshift , the zeropoint does , in the sense that high - redshift sfgs form stars at a higher rate than similar mass galaxies today . complementing the observational results on the locus of sfgs in sfr - mass space , direct measurements of the co molecular line emission from galaxies residing on the ms have revealed molecular gas mass fractions as high as 35 ( 45 ) % at @xmath6 ( tacconi et al .
2010 ) . linking surface densities of gas and star formation , genzel et al .
( 2010 ) recently argued that high - redshift sfgs follow the same star formation law as local sfgs , unlike their merging counterparts that have shorter depletion timescales and follow a kennicutt - schmidt relation with a higher zeropoint ( see also daddi et al .
2010 ) . from the observed zeropoint evolution of the ms , and tightness of the relation
, a picture has emerged in which the bulk of sfgs are forming stars gradually over timescales that are long ( one to two orders of magnitude larger ) relative to their dynamical times ( genzel et al . 2010 ; wuyts et al . 2011 ) .
this is only possible by a continuous replenishment of their gas reservoirs .
cold , filamentary streams of gas have been proposed on the basis of cosmological simulations as an efficient mechanism to penetrate the galaxies surrounding halos , and deposit new ( or recycled ) fuel for star formation ( keres et al .
2005 , 2009 ; dekel & birnboim 2006 ; dekel et al .
evidence from kinematics strongly supports such a continuous ( i.e. , non - major merger ) triggering and maintenance of star formation for the bulk of sfgs ( genzel et al . 2008 ; frster schreiber et al .
2009 ; shapiro et al . 2008 ; frster schreiber et al . 2011a ) .
in such a scenario , a secular mode of star formation dominates the cosmic sfr history , and the scatter around the ms can , aside from a contribution by measurement uncertainties , be attributed to burstiness induced by occasional ( minor ) merger events . while the excess star formation due to mergers may only amount to 5 - 10% ( hopkins et al .
2010 ; rodighiero et al .
2011 ) , they may play a crucial role in shaping the diversity of galaxies observed both at low and high redshift . the merger paradigm ( sanders et al .
1988 ; hopkins et al . 2006 )
links a wide variety of galaxy types in an evolutionary sequence in which the collision between sfgs creates tidal torques that channel gas to the center where it triggers a nuclear starburst ( e.g. , barnes & hernquist 1991 , 1996 ) . when the starburst is accompanied by accretion onto the central supermassive black hole(s ) ( smbhs ) , this produces agn emission , and naturally accounts for the origin of a correlation between smbh and bulge mass ( di matteo et al .
finally , a combination of gas exhaustion , supernova and agn feedback leads to a rapid decline in the sfr , leaving a post - starburst , red and dead galaxy as remnant ( wuyts et al . 2010 ; snyder et al .
2011 ) .
indeed , such classes of galaxies offset from the ms have also been identified observationally out to @xmath5 and even beyond : towards the high - sfr end with for instance bright sub - millimeter galaxies ( smgs , smail et al .
1997 ; barger et al . 1998
; hughes et al . 1998 ; see blain et al .
2002 for a review ) , and towards the low - sfr end ( daddi et al . 2005 ; kriek et al . 2006 ; van dokkum et al .
observations of their structural properties , albeit often limited to small samples , could lend credence to their merger connection .
several of the smgs imaged in co show two - component morphologies and/or highly disturbed velocity fields , implying a major merger nature ( tacconi et al .
massive ( @xmath7 ) quiescent galaxies at high redshift are increasingly more compact relative to similar mass galaxies today ( @xmath8 ) ( e.g. , van der wel et al . 2008 ; van dokkum et al .
2008 , but see mancini et al .
2010 for a contrasting view ) , in agreement with expectations from binary merger scenarios in which the gas content of the progenitors increases with redshift ( khochfar & silk 2006 ; hopkins et al .
2009 ; wuyts et al .
2010 ) . due to the highly dissipational formation process , wuyts et al .
( 2010 ) predict a larger degree of rotation in the remnants than in lower redshift early - type galaxies .
observational evidence for high - redshift compact quiescent disks ( mcgrath et al . 2008 ; van der wel et al . 2011 ) could possibly reflect this , or alternatively may signal other quenching mechanisms at play . with the high - resolution optical ( acs ) and
near - infrared ( wfc3 ) cameras onboard hst , detailed structural measurements can now be obtained for ever larger samples , bringing statistical significance also for the more extreme populations .
in particular , the ongoing cosmic assembly near - infrared deep extragalactic legacy survey ( candels , grogin et al . 2011 ; koekemoer et al .
2011 ) is collecting deep rest - frame optical ( and rest - uv ) space - based imaging over unprecedented areas .
at the same time , the pacs evolutionary probe ( pep , lutz et al .
2011 ) opens a window on the far - infrared emission of large samples of sfgs , and allows the calibration of other sfr indicators for use in fields that lack such ( deep ) far - ir data ( wuyts et al .
2011 ) . in this paper
, we exploit high - resolution imaging and a cross - calibrated ladder of sfr indicators to study the structure of galaxies and their mode of star formation , as function of their position in the sfr - mass diagram .
we discuss how properties such as size , surface brightness profile shape , sfr surface density , and @xmath9 ratios vary along and across the ms , and how these trends evolve with redshift ( from @xmath10 till today ) . to this end , we assembled samples of unprecedented size , with 639924 galaxies at @xmath2 , 132328 galaxies at @xmath3 , and 35649 galaxies at @xmath4 .
the data quality and sample size allow us to extend previous studies of the relation between structure and stellar populations in the nearby ( kauffmann et al .
2003 ; schiminovich et al .
2007 ; bell 2008 ) and intermediate - redshift ( scarlata et al . 2007 ; maier et al . 2009 ) universe , and to build on the ground - based explorations of these early epochs ( franx et al .
2008 ; williams et al .
2010 ) .
the paper is structured as follows .
we present an overview of the fields and data used , the methods to compute derived products , and the final sample in section [ obs_method_sample.sec ] .
readers who wish to skip directly to our results , should turn to sections [ structure.sec ] and [ mode.sec ] , where we analyze galaxy structure and the mode of star formation in the sfr - mass diagram respectively .
section [ discussion.sec ] discusses the observational results in a physical context .
finally , we summarize our conclusions in section [ summary.sec ] . throughout this paper , we quote magnitudes in the ab system , assume a chabrier ( 2003 ) inital mass function ( imf ) , and adopt the following cosmological parameters : @xmath11 .
in our analysis , we divide the galaxies in three redshift bins : @xmath12 , @xmath13 , and @xmath14 , which we will in short refer to as @xmath2 , @xmath3 , and @xmath4 .
the @xmath2 sample is extracted from the sloan digital sky survey ( sdss ) , as detailed in section [ sdss_galex.sec ] .
the higher redshift samples are extracted from four deep fields that host some of the richest multi - wavelength data sets available to date : cosmos , uds , goods - south , and goods - north .
we describe the data available for them in sections [ cosmos.sec ] to [ goodsn.sec ] , and summarize the methods applied to compute derived properties such as photometric redshifts , stellar masses , sfrs and structural parameters in section [ derived_prop.sec ] .
sdss offers a robust local universe anchor for our study of galaxy properties in the sfr - mass diagram , and its evolution with redshift .
we use the spectroscopic sdss sample ( dr7 , abazajian et al .
2009 ) in overlap between the mpa - jhu and ny - vagc value - added galaxy catalogs .
sfrs and stellar masses were taken from the mpa - jhu compilation .
briefly , the stellar masses were estimated by fitting bruzual & charlot ( 2003 , hereafter bc03 ) stellar population synthesis models with a wide range of star formation histories ( sfhs ) to the broad - band photometry ( salim et al .
2007 ) . unless explicitly stated otherwise , we use total , dust- and aperture - corrected sfrs .
sfrs within the sdss spectroscopic fiber were primarily based on h@xmath15 , using the calibration of brinchmann et al .
( 2004 ) , which includes an extinction correction based on the balmer decrement .
aperture corrections to total use fits to the photometry of the outer regions of the galaxies , as described by salim et al .
we use the sdss ( dr7 ) - galex ( gr5 ) matched catalog from bianchi et al .
( 2011 ) to compute the unobscured part of the star formation @xmath16 .
sizes and profile shapes were measured by the ny - vagc team by means of sersic fits to the galaxy surface brightness profiles ( blanton et al .
2003 , 2005 ) . by default , we adopt the structural parameters measured in the @xmath17 band , but we will discuss the dependence on wavelength where appropriate .
we make use of public multi - wavelength photometry from ilbert et al .
( 2009 ) and gabasch et al .
( 2008 ) over an effective area of 1.48 deg@xmath18 in the cosmos hst field ( scoville et al . 2007 ;
koekemoer et al .
our working catalog is cut at @xmath19 to guarantee sufficient signal - to - noise to allow for a reliable determination of photometric redshifts and other derived properties .
a total of 36 medium and broad bands sample the sed from galex to irac wavelengths . at infrared wavelengths ,
pacs imaging from pep , extracted using 24 @xmath20 m sources as prior , reaches depths of 5.7 and 11.9 mjy ( 3@xmath21 ) at 100 @xmath20 m and 160 @xmath20 m respectively .
mips 24 @xmath20 m imaging in cosmos ( sanders et al .
2007 ; le floch et al .
2009 ) is used down to 60 @xmath20jy .
an x - ray catalog based on xmm - newton observations is released by cappelluti et al .
high - resolution @xmath22-band imaging with hst / acs ( @xmath23 ) spanning the entire field , and probing the rest - frame @xmath24 and @xmath25 band at @xmath3 and @xmath4 respectively , forms the foundation for structural measurements in cosmos . a deep wfc3 @xmath26-selected catalog ( @xmath27 , 5@xmath21 ) with consistent photometry in 16 bands from @xmath24 to 8 @xmath20 m was constructed on the basis of candels , seds ( pi g. fazio ) , spuds ( pi j. dunlop ) , and ancillary ground - based data in the @xmath28 arcmin@xmath18 candels - covered uds area ( guo et al . in prep )
. mips 24 @xmath20 m imaging to a depth of 30 @xmath20jy ( 3@xmath21 ) was carried out as part of the spuds survey .
ueda et al . ( 2008 ) presented the x - ray source catalog of the subaru / xmm - newton deep survey , in which the candels / uds area is embedded .
we performed the structural measurements on the wfc3 @xmath26 mosaic , which with 4/3 orbits per pointing forms part of candels - wide .
the @xmath26 band probes the rest - frame @xmath29 and @xmath30 band at @xmath3 and @xmath4 respectively .
following similar procedures as for uds , guo et al .
( in prep ) constructed a deep wfc3 @xmath26-selected catalog ( @xmath31 , 5@xmath21 ) for the early release science ( ers , pi oconnell ) and candels - deep area in goods - south .
the catalog contains consistent photometry in 14 passbands from @xmath32 to 8 @xmath20 m .
spanning the same wavelength range , we use fireworks ( @xmath33 , wuyts et al . 2008 ) as supporting multi - wavelength catalog to exploit candels - wide data in the bottom 20% of the 148 sq .
arcmin goods - south field .
as part of the pep survey ( lutz et al .
2011 ) , pacs photometry was obtained to a 5@xmath21 depth of 1.8 mjy , 1.9 mjy and 3.3 mjy at 70 @xmath20 m , 100 @xmath20 m and 160 @xmath20 m respectively , using the position of 24 @xmath20 m sources as prior .
the 24 @xmath20 m imaging itself ( 30 @xmath20jy , 5@xmath21 ) was taken from magnelli et al .
where appropriate , we select x - ray sources from the chandra 2 ms source catalog by luo et al .
we performed the structural measurements on wfc3 @xmath26-band imaging drizzled to a @xmath34 scale , obtained as part of the ers and candels programs .
the wfc3/ir observations of the ers were obtained in f098 m , f125w and f160w for a total depth of 2 orbits per pointing in each filter ( windhorst et al .
2011 ) and the mosaics used here were produced using multidrizzle ( koekemoer et al .
2002 ) following the approach outlined in koekemoer et al .
the 4-epoch candels - deep area is similar in depth , while candels - wide received half the integration time .
we derived photometric redshifts and stellar masses for galaxies in goods - north based on a @xmath35-selected catalog with photometry in 16 bands from galex to irac wavelengths ( berta et al .
2010 ) . additional longer wavelength information at 24 @xmath20 m ( 30 @xmath20jy , 5@xmath21 ) , 100 @xmath20 m ( 5.1 mjy , 5@xmath21 ) , and 160 @xmath20 m ( 8.7 mjy , 5@xmath21 ) was acquired as part of the goods and pep surveys .
where appropriate , we select x - ray sources from the chandra 2 ms source catalog by alexander et al .
structural measurements in goods - north are based on the goods acs @xmath36-band imaging ( giavalisco et al . 2004 ) , corresponding to the rest - frame @xmath24 and @xmath37 band at @xmath3 and @xmath4 respectively .. for all deep fields , we use identical procedures to derive photometric redshifts , stellar masses , sfrs , and structural parameters , hence optimizing the consistency of the combined data set .
we determined photometric redshifts by fitting a superposition of 6 galaxy templates to the observed broad - band seds using the photometric redshift code eazy ( brammer et al .
2008 ) . for the 6955 out of 132328 ( 5.3% ) of galaxies at @xmath13 , and 497 out of 35649 ( 1.4% ) of galaxies at @xmath38 that have a spectroscopic redshift from one of the many spectroscopic campaigns in the above four fields ( see , e.g. , vanzella et al . 2008 ; lilly et al .
2009 ; barger et al .
2008 ) , we adopt the @xmath39 as redshift in our analysis . comparing the photometric redshifts @xmath40 and spectroscopic redshifts @xmath39 of spectroscopically confirmed galaxies , we find a median and scatter ( normalized median absolute deviation ) in @xmath41 of ( -0.001 ; 0.015 ) in the @xmath3 and ( -0.007 ; 0.052 ) in the @xmath4 redshift bin .
naturally , these uncertainties only apply to galaxies with similarly bright magnitudes as represented in the spectroscopic sample based on which the scatter was computed .
e.g. , the public zcosmos - bright spectroscopic survey ( lilly et al . 2009 ) reaches down to @xmath42 . for this spectroscopic sample in cosmos ,
we measure a scatter in @xmath41 of 0.010 . as detailed in appendix b.3 , the estimated uncertainty increases to @xmath43 for cosmos galaxies with @xmath44 , which account for 40% of the galaxies in that field , and 30% of our entire sample .
the estimated uncertainty is consistent with earlier work by ilbert et al .
for the other deep fields as well , the increase in the photometric redshift uncertainty with respect to the spectroscopic sample amounts to a factor of @xmath45 for the faintest galaxies that enter our analysis . an extensive analysis in appendix
b.3 shows that our results are robust against photometric redshift uncertainties , even when accounting for the fact that the @xmath40 quality of the entire sample is likely lower than that of the ( typically brighter ) spectroscopically confirmed subsample .
we used fast ( kriek et al .
2009a ) to fit bc03 models with exponentially declining sfhs to the @xmath46 m broad - band seds of the galaxies . following the recipe presented by wuyts et al .
( 2011 ) , we impose a minimum e - folding time of 300 myr .
the time since the onset of star formation was allowed to vary between 50 myr and the age of the universe at the observed redshift .
our choice of sfhs guarantees consistency with most of the literature , and in addition has the advantage ( as opposed to , e.g. , only exponentially increasing sfhs ) that it is also applicable to lower redshift galaxies and high - redshift quiescent galaxies , that clearly formed stars at a higher rate in their past than at the epoch of observation .
maraston et al .
( 2010 ) showed that adopting exponentially increasing sfhs for star - forming galaxies at @xmath4 leads to negligible changes in the estimated stellar masses , unless also a stringent constraint on formation redshift is imposed , in which case the estimated stellar mass of these galaxies can increase by a few 0.1 dex .
solar metallicities were assumed throughout , and dust attenuation was modeled using the calzetti et al .
( 2000 ) reddening law , with visual extinctions in the range @xmath47 . for a detailed description of the dependence of
estimated stellar masses on the assumed metallicity and attenuation law , we refer the reader to wuyts et al .
we revisit the assumption of solar metallicity in section [ model_iruv.sec ] in the context of a model for ir / uv ratios that involves a metallicity scaling , and conclude that it has a negligible impact on our results . as for the sdss galaxies , we assumed a universal chabrier ( 2003 ) imf for all galaxies in the lookback surveys . exploiting deep pacs photometry in the goods - south field , wuyts et al .
( 2011 ) established a continuity across sfr indicators , ranging from uv + fir , over uv + mir , to sed modeled sfrs .
we adopt this cross - calibrated ladder of sfr indicators to obtain an estimate of the sfr for each galaxy in our sample individually . briefly ,
if the galaxy is detected in an ir band , we compute the total sfr by summing the unobscured and re - emitted emission from young stars , following kennicutt ( 1998 ) : @xmath48 = 1.09 \times 10^{-10}\ ( l_{ir } + 3.3 l_{2800})/l_{\sun } \label{sfruvir.eq}\ ] ] where @xmath49 was computed with eazy from the best - fitting sed . here , the total ir luminosity @xmath50 is derived monochromatically from the longest wavelength ir band that is significantly ( @xmath51 ) detected : either pacs 160 , 100 , or 70 @xmath20 m , or mips 24 @xmath20 m . we use the mid- to far - infrared sed template by wuyts et al .
( 2008 ) for the conversion to @xmath52 .
this conversion leads to a consistency between 24 @xmath20 m and pacs - derived @xmath52 , unlike locally calibrated template sets that have been proven to overestimate the @xmath52 of @xmath53 galaxies based on their rest - frame 8 @xmath20 m emission ( nordon et al . 2010 ; elbaz et al .
2010 ; wuyts et al .
2011 ) . for galaxies that lack an ir detection
, we adopt the sfr associated to the best - fit stellar population synthesis model from section [ sedmodeling.sec ] . for the low- to intermediate sfr regime , where @xmath54
is only attainable in the deepest of our four fields ( goods - south ) , wuyts et al .
( 2011 ) demonstrated that sed modeled sfrs are consistent with those obtained from equation [ sfruvir.eq ] .
as for the stellar masses , the values of @xmath55 , obtained without imposing explicit constraints on the formation redshift , would not change significantly when assuming exponentially increasing rather than decreasing sfhs ( maraston et al .
2010 ) . in this paper , we limit our analysis of galaxy structure to a parametric approach , focussing on two basic properties of the surface brightness distribution : its extent and cuspiness . the former is expressed as the circularized effective radius @xmath56 corresponding to the best - fit sersic profile . the latter is quantified by the sersic index @xmath57 , with @xmath57 restricted to the range @xmath58 .
we adopt the structural measurements performed on the longest wavelength high - resolution imaging available for each field ( wfc3 for goods - s and uds , acs for goods - n and cosmos , see section [ fields_data.sec ] ) .
the galapagos package ( barden et al .
2009 ) or equivalent procedures were used to pre - determine the sky background and the neighboring sources that need to be simultaneously fit or masked , and to automate fitting sersic profiles to the 2d surface brightness distributions with galfit ( c. y. peng et al .
galfit takes into account convolution by the point spread function ( psf ) .
since the longest wavelength at which high - resolution imaging is presently available varies between our four deep fields , our study lacks the potential to consistently probe the rest - frame optical regime , and fully control morphological k - corrections , if present .
however , since we are interested not only in galaxies on the main sequence , but also in those rare systems that form the high - sfr tail , we prefer to sacrifice this aspect of our analysis , rather than area . as a sanity check
, we compared sizes and sersic indices measured in the @xmath36 and @xmath26 imaging of goods - s ( see appendix b.2 ) .
we find the median acs - wfc3 deviations to be small ( -0.01 in @xmath59 and -0.10 in @xmath57 at @xmath38 ) relative to the trends that will be discussed in this paper , suggesting that any biases from morphological k - corrections have only a limited impact on our conclusions .
bond et al .
( 2011 ) arrive at a similar conclusion comparing rest - frame optical and rest - frame uv size measurements of @xmath4 galaxies .
moreover , we also verified that the results obtained for each of the four fields individually are consistent with those presented for the combined data set in this paper , albeit with more noise due to the smaller number statistics ( see appendix b.1 ) . while beyond the scope of this paper , the wavelength dependence of structural properties can potentially reveal interesting clues on the physical processes shaping galaxies in the young universe ( see , e.g. , guo et al .
2011 ; szomoru et al .
2011 ; and cassata et al .
2011 , who find evidence for mild morphological k - corrections such that the centers of galaxies in their samples tend to be somewhat redder ) .
for an in depth discussion of resolved stellar populations inferred from spatial variations in color , we defer the reader to wuyts et al .
( in prep ) . our four deep fields are not uniform in depth .
consequently , they have different completeness limits in the sfr - mass diagram . since we are mainly interested in individual galaxy characteristics rather than abundances ( number or mass densities ) as function of position along or across the ms , we refrain from applying any incompleteness corrections to the observed populations .
under the premise that , at any given redshift , galaxies form a two - parameter family described by their mass and sfr , this approach works well .
i.e. , a median galaxy property can be computed reliably , and unbiased by any completeness issues , based on the objects observed in a given bin of sfr - mass space .
our final sample comprises 639924 galaxies at @xmath12 , 132328 galaxies at @xmath13 , and 35649 galaxies at @xmath38 .
the relative breakdown in galaxies of different masses is determined by the depth of the observations , and the stellar mass function at the respective redshifts . above @xmath60 our sample
counts 532131 , 31127 , and 8895 galaxies at @xmath2 , @xmath3 , and @xmath4 respectively . above @xmath61
, the numbers drop to 147922 , 2767 , and 1059 galaxies at @xmath2 , @xmath3 , and @xmath4 respectively .
an overview of the sample size per field is provided in table [ fields.tab ] .
we start by analyzing the surface brightness profile shape as a function of position in the sfr - mass diagram in figure [ n.fig ] .
the three panels show from left to right the @xmath2 , @xmath3 , and @xmath4 bins respectively . instead of indicating the relative abundance of galaxies in different regions of the diagram
, we use the color - coding to mark the median value of the sersic index @xmath57 of all galaxies in each @xmath62 $ ] bin . for displaying purposes ,
we restrict the range of the colorbar to @xmath63 , and assign the same color as @xmath64 and @xmath65 to bins with median @xmath66 or median @xmath67 respectively .
the fractions @xmath68 of galaxies lying outside these bounds amounts to ( 0.09 ; 0.24 ) , ( 0.41 ; 0.14 ) , and ( 0.41 ; 0.16 ) at @xmath2 , @xmath3 , and @xmath4 respectively .
the fraction of @xmath62 $ ] bins with median n outside this range is small : ( 0.02 ; 0.11 ) at @xmath2 , ( 0.14 ; 0.15 ) at @xmath3 , and ( 0.11 ; 0.11 ) at @xmath4 .
the resulting diagrams present a remarkably smooth variation in the typical galaxy profile shape across the diagram .
moreover , despite the loss of information on number densities , the so - called main sequence of star formation is immediately apparent , and its presence persists out to the highest observed redshifts .
this structural main sequence consists of galaxies with near - exponential profiles ( @xmath69 ) and shows a similar behavior as the conventional number main sequence as identified on the basis of number densities in the sfr - mass diagram ( e.g. , noeske et al .
2007 ; elbaz et al . 2007
; daddi et al . 2007 ) .
namely , an upward shift of the zeropoint is observed with increasing lookback time . at each epoch , the ms in figure [ n.fig ] is well approximated by a slope of unity ( _ white line _ ) .
the sfr at which the median n reaches a minimum in a mass slice around @xmath70 roughly coincides with the mode of the @xmath71 distribution in that mass slice , but depending on the fitting method and sample definition used to weed out quiescent galaxies , a somewhat shallower slope than unity may be measured for the number main sequence at the massive end ( see , e.g. , rodighiero et al .
2010 ) . below the structural ms ,
a cloud of galaxies with cuspy , near - de vaucouleurs ( @xmath72 ) profiles is visible .
this population of massive quiescent galaxies is present at all observed epochs .
our first and foremost conclusion from figure [ n.fig ] is therefore that a correlation between the structure and stellar populations of galaxies is already in place since @xmath5 . in the local universe
, the presence of such a correlation has been described in depth by , e.g. , kauffmann et al .
( 2003 ) and brinchmann et al .
scarlata et al . (
2007 ) demonstrated that photometrically selected massive quiescent galaxies are already dynamically relaxed ( i.e. , have the morphological appearance of early - type galaxies ) at @xmath3 .
also out to @xmath3 , maier et al . ( 2009 ) exploited the spectroscopic zcosmos survey to reveal a correlation between the specific sfr of galaxies and the sersic index of their surface brightness profiles . at larger lookback times ,
our results are in agreement with previous findings , based on smaller samples , reporting a similarly early emergence of the hubble sequence on the basis of stellar surface mass density , galaxy size and visual appearance ( franx et al .
2008 ; toft et al .
2009 ; kriek et al .
we find that , across cosmic time , the typical sersic index of galaxies is not optimally described as a function of their absolute sfr , or even their absolute specific sfr , but rather as a function of their position relative to the ms at the epoch of observation .
the correspondence between mass , sfr and structure , as quantified by the sersic index , is equivalent to the hubble sequence .
based on samples of unprecedented size , we see that such a sequence already existed at @xmath4 ; bulge - dominated morphologies go hand in hand with a more quiescent nature .
first indications of such a correlation between sersic index and specific sfr out to high redshift were recently reported by szomoru et al .
( 2011 ) based on 27 galaxies at @xmath3 and 16 galaxies at @xmath4 .
in detail , an additional interesting aspect is revealed by the @xmath57 distribution in sfr - mass space .
the surface brightness profile shape does not vary monotonically across the main sequence , but instead shows a reversal in the high - sfr tail of the distribution , such that the upper envelope of the ms is composed of galaxies whose typical cuspiness is intermediate between those of normal ms galaxies and the quiescent population below the ms .
this is illustrated further in figure [ dev.fig ] , where we plot a cut through of the @xmath73 mass slice .
the top panels show the histogram of deviations from the ms measured along the sfr axis : @xmath74 .
the top left panel illustrates the well - known bimodality between star - forming and passive galaxies in the local universe ( e.g. , kauffmann et al . 2003 ) .
a similar bimodality , albeit with a reduced amplitude of the quiescent peak , is visible at @xmath3 ( and becomes more pronounced if we were to limit ourselves to higher masses only ) .
while both populations are also found in our highest redshift bin , the quiescent fraction is further reduced .
incompleteness in our shallowest field , as well as uncertainties in photometric redshifts and estimated sfrs inhibit the identification of a true bimodality in the @xmath4 histogram .
the bottom panels of figure [ dev.fig ] show the median dependence of @xmath57 on @xmath74 ( _ black curve _ ) , as well as errors on the determination of the median ( _ black error bars _ ) .
with grey polygons , we mark the central 68th percentile .
clearly , a variety of profile shapes is measured , even at a given @xmath75 , and particularly in the @xmath3 and @xmath4 bins .
this stems partly from measurement uncertainties , that dominate in the low surface brightness regime , but in addition likely means that in detail galaxies are more complex than a simple two - parameter ( @xmath62 $ ] ) family .
interestingly , the scatter seems to be minimized for galaxies that reside on the ms .
frster schreiber et al . ( 2011 ) present a detailed analysis of large and clumpy ms galaxies at @xmath4 , that exhibit shallow surface brightness profiles ( @xmath66 ) .
such a population is also present in our sample ( see figure [ dev.fig ] and the above quoted @xmath76 ) .
focussing on the overall trend , we find the results at @xmath3 and @xmath4 to be very similar to those established at @xmath2 .
namely , galaxies on the ms resemble exponential disks , high - sfr outliers have a centrally enhanced surface brightness profile , and once galaxies end up below the ms , they quickly reach a plateau of @xmath72 .
we discuss the physical implications of this empirical relation in section [ quenching.sec ] .
we verified that , when excluding x - ray sources from our analysis , the same median dependence of the surface brightness profile shape on the location of galaxies in the sfr - mass diagram remains present .
this is the case even for the fields with the deepest x - ray data .
the same robustness also applies to the other galaxy properties discussed later in this paper .
our test indicates that the observed trends ( particularly the enhanced sersic indices of the high - sfr outliers at a given mass ) is not merely due to a nuclear , non - stellar point source .
when focussing on the x - ray detected sources alone , we find that they occupy the mass regime above @xmath77 , and span a wide range in sfrs , from the quiescent class to on and above the ms . within a given bin ,
their surface brightness profiles differ in the sense that they have higher sersic indices and smaller sizes than normal galaxies of the same @xmath62 $ ] .
point source contamination would drive the structural parameters in the observed direction , but extensive simulations ( see , e.g. , simmons & urry 2008 ) are required to constrain whether in addition there is any evidence for a different stellar structure of the hosts . before moving to the next structural property , we have a closer look at the @xmath2 bin in figure [ dev.fig ] , where we use structural measurements in five passbands to investigate how the observed trend depends on wavelength .
the dotted colored curves in the bottom left panel of figure [ dev.fig ] illustrate the median relation derived from sersic fits in the sdss @xmath78 , @xmath79 , @xmath80 , and @xmath81 bands .
together with the default @xmath17-band profile fits ( _ solid black curve _ ) , a trend emerges in which the overall relation between @xmath57 and @xmath75 is remarkably robust against morphological k - corrections .
the largest difference is observed for galaxies that lie on the ms ( @xmath82 ) , such that a cuspier profile is inferred from redder waveband imaging .
this comes as no surprise , as it is well known that many normal sfgs in the nearby universe are composed of a ( red , @xmath72 ) bulge and ( blue , @xmath69 ) disk component ( sargent et al .
2007 ) . ] . at low and high values of @xmath75 , the sersic index levels of at similar values ( @xmath83 and @xmath84 respectively ) , irrespective of wavelength .
this suggests that the light concentration in the centers of these galaxies corresponds to a mass concentration , and is not due ( only ) to the large luminosity of a young stellar population .
having established the existence of a hubble sequence out to @xmath5 , we now turn to the zeroth order structural measurement of size , and address how it varies along and across the ms at each of our observed epochs .
figure [ re.fig ] is similar in nature to figure [ n.fig ] , but the color - coding is now used to mark differences in the spatial extent of the surface brightness distribution ( @xmath85 ) . as in figure
[ n.fig ] , a qualitatively similar behavior is observed at all redshifts .
overall , more massive galaxies tend to be larger .
however , the mass - size relation is not fundamental in the sense that at a given mass substantial variations in size occur that are not random , but instead correlate with the offset from the ms ( see also the appendix , figure [ dev_app.fig ] ) .
this translates to different mass - size relations followed by passive and star - forming galaxies , as reported previously by shen et al .
( 2003 ) based on sdss , but also by williams et al .
( 2010 ) and szomoru et al .
( 2011 ) who extended the analysis for massive galaxies out to @xmath86 , based on ground - based @xmath87-band imaging in uds and wfc3 @xmath26-band imaging in the hubble ultra deep field respectively .
consistent with previous results by , e.g. , franx et al .
( 2008 ) and toft et al . (
2009 ) , high - redshift quiescent galaxies are more compact than the bulk of sfgs at the same epoch . also at @xmath2 , galaxies below the ms ( at least at @xmath88 ) tend to be smaller than those on the ms .
the high - sfr tail of the @xmath2 distribution differs from normal ms galaxies in the sense that they tend to have higher surface mass densities ( i.e. , are more compact for a given mass ) .
as we march up in redshift , our measurements are consistent with a similar behavior across all lookback times considered . comparing the panels for the three redshift bins , it is immediately apparent that galaxies grow over time .
we quantified the growth by fitting a @xmath89 relation to the galaxies in each @xmath62 $ ] bin .
it is important to note that this reflects the size evolution of the galaxy population as a whole , and not per se that of any given galaxy individually , as galaxies will move from one @xmath62 $ ] bin to another as they build up stellar mass , and new systems will be added over time .
we find @xmath15 to be varying across the sfr - mass diagram , from @xmath90 below @xmath91 to @xmath92 above @xmath93 . here , the range indicates the normalized median deviation over the different @xmath62 $ ] bins .
splitting the latter class into galaxies above and below the ms at @xmath2 ( _ white line in the left - hand panel of figure [ re.fig]_ ) , we obtain @xmath94 and @xmath95 respectively . in other words ,
the rate of size evolution correlates with the specific star formation rate , reaching the lowest values in the ( especially massive ) quiescent population .
having established the variation of galaxy structure along and across the ms , we now do the same for two parameters that characterize the mode of star formation : its surface density ( section [ surfdens.sec ] ) , and its breakdown into unobscured and dust - enshrouded star formation ( section [ obscuration.sec ] ) .
a discussion of the physical implications of the observed relationships follows in section [ model_iruv.sec ] .
figure [ surf.fig ] presents the dependence of the sfr surface density @xmath96 on the position of galaxies in the sfr - mass diagram . by construction
, this figure has the same information content as figure [ re.fig ] , but its interpretation in terms of mode of star formation is more straightforward .
galaxies that form stars more actively are not just upscaled versions of lower sfr systems .
their ism conditions differ in the sense that also normalized by area they form more stars per unit time .
lines of constant @xmath0 run diagonally , implying a tighter correlation with specific sfr than with sfr or stellar mass separately .
schiminovich et al .
( 2007 ) discussed this trend extensively for galaxies in the nearby universe . at higher redshifts ,
this is consistent with recent findings based on rest - frame uv size measurements by elbaz et al .
we confirm that the same trend is seen when rest - frame optical size measurements are used instead .
we contrast @xmath0 to @xmath97 in figure [ surfssfr.fig ] , where we plot the 25th and 75th percentile contours ( _ solid and dashed lines _ ) for each of the considered redshift bins . over 10 gyr of lookback time ,
the bulk of galaxies follow a similar , linear relation between specific sfr and surface density of star formation .
as we probe higher redshifts , the upper end of this relation becomes more densily populated .
this trend can naturally be explained by increased gas mass fractions in the young universe ( e.g. , tacconi et al .
2010 ) . in equation [ sfrsurfdens.eq ]
, we adopted the same size measurement as discussed in section [ structural_params.sec ] and [ size.sec ] to represent the radius within which half the star formation takes place .
i.e. , we adopted the best - fit @xmath85 of a single sersic profile fit . at @xmath2
, we used the @xmath17 band by default .
for the higher redshift samples , we used the longest wavelength high - resolution image available ( @xmath26 in the case of uds and goods - s , @xmath36 and @xmath22 for goods - n and cosmos respectively ) . in principle , radial variations in sfh and/or dust attenuation could bias the measurement of a half - sfr radius .
e.g. , in the case of a superposition of a star - forming disk and a red and dead bulge , the total half - light radius would contain less than half the star formation ( particularly when measured in the red ) .
in contrast , in the case of a galaxy where old and young stars trace the same distribution , but the dust column is centrally enhanced , the observed @xmath85 would contain more than half of the star formation ( particularly when measured in the blue ) .
we take two simple approaches to investigate the impact of these potential biases on our results .
first , we repeated our analysis at @xmath2 adopting each of the @xmath98 sdss bands to determine @xmath85 and based thereupon @xmath0 .
we find that this induces a minor change only , in the sense that , for galaxies below the main sequence , @xmath0 increases by a few 0.1 dex from the bluest to the reddest band ( see the left - hand panel of figure [ dev_app.fig ] in the appendix ) .
overall , the pattern of @xmath0 in sfr - mass space does not depend appreciably on the adopted photometric band at @xmath2 .
second , we stress that at @xmath3 as well as at @xmath4 , the dependence of @xmath0 on location in the sfr - mass diagram remains the same ( although with somewhat enhanced noise due to the smaller number statistics ) when considering the fields with rest - frame optical and rest - frame uv imaging separately ( see appendix b.1 ) .
finally , we consider the ratio of re - emitted to unobscured star formation . what we denote as @xmath1 is for ir - detected galaxies computed as the ratio of the ir and uv contribution to the total @xmath54 ( see equation [ sfruvir.eq ] ) :
@xmath99 . for those galaxies in the deep lookback surveys that lack an ir detection , and for which we adopted the best - fit sfr from stellar population modeling
, we derive @xmath1 as @xmath100 .
here , @xmath101 is the extinction at 2800 as inferred from the sed modeling .
this approach has the advantage of always producing positive values , unlike adopting @xmath102 as definition , which can reach negative values when @xmath103 .
the latter situation can occur for quiescent galaxies with low ratios of present to past - averaged sfr , in which case the rest - frame 2800 emission contains a substantial contribution from older stars . for the @xmath2 sample , total sfrs ( based on dust- and aperture - corrected @xmath104 measurements , see section [ sdss_galex.sec ] ) are available from the mpa - jhu release , and we computed @xmath16 as for the higher redshift galaxies , where we interpolated @xmath105 from the galex nuv and sdss @xmath78-band photometry .
the derived @xmath106 at @xmath2 are well defined for the star - forming population , but ill - constrained for the passive population for the above reasons .
we therefore choose to exclude the passive @xmath2 galaxies from our analysis .
the resulting sfr - mass diagrams , color - coded by @xmath1 , are presented in figure [ iruvobs.fig ] .
again , the qualitative appearance of the diagrams looks similar over the wide range of redshifts considered . at a given mass , a proportionally larger fraction of the star formation activity
is revealed by reprocessed ir emission as we consider more actively star - forming systems . a positive correlation between bolometric luminosity and obscuration
was presented previously for galaxies in the nearby universe ( heckman et al .
1998 ) and at @xmath4 ( reddy et al . 2010 ) , but these authors did not trace the trend with sfr while specifically controlling for stellar mass . likewise ,
an increase in @xmath1 is observed as we move along the ms from the low- to high - mass end .
such a trend was not seen as clearly in @xmath0 ( figure [ surf.fig ] ) , a finding that we will interpret as due to metallicity variations across the diagram in section [ model_iruv.sec ] .
finally , objects forming the upper rim of the galaxy distribution , also at lower masses , show enhanced ir emission .
the latter trend is most prominent in the @xmath3 bin , but a hint of the same feature is seen at both lower and higher redshift as well .
the observation that a precursor of the hubble sequence , with disk - like sfgs and cuspier quiescent systems , was already in place 10 gyr ago ( see section [ profile.sec ] ) has important implications . to name one , a mechanism for galaxies to transition from the star - forming to quiescent phase must be universally present across cosmic time , at least out to @xmath5 .
it is during this epoch around @xmath4 that the cosmic sfr density is observed to reach its peak ( e.g. , hopkins & beacom 2006 ) , and that large accretion rates onto galaxy halos are expected from cosmological hydrodynamical simulations ( keres et al . 2005 , 2009 ; dekel & birnboim 2006 ; dekel et al .
we can therefore exclude gas exhaustion as sole driver of the transition from the active to passive phase . instead
, a quenching mechanism must be at play .
moreover , the observed relation between structure and stellar populations implies a causal connection between the quenching and the morphological transition . as early as by toomre & toomre ( 1972 ) ,
mergers have been invoked as a potential mechanism to transition from late to early hubble types .
in such a scenario , much of the gas content of the progenitors is consumed by star formation in the nucleus , building up a central cusp .
violent relaxation distributes the stars that were already formed prior to final coalescence in a near - de vaucouleurs ( @xmath65 ) distribution .
feedback from short - lived massive stars that explode as supernovae , as well as energy that is released from accreting material onto the smbh(s ) and that couples to the surrounding ism drives out the remaining gas and heats up the galaxy halo preventing new infall ( e.g. , hopkins et al .
2006 , and references therein ) . in this light , the reversal of the decreasing trend of @xmath57 with @xmath74 ( see figure [ dev.fig ] ) may signal that we are observing the galaxies at the high - sfr tail of the distribution during the peak of their sfh , in the process of building up a central concentration of stellar mass , much like that typically observed in the quiescent population .
this would also explain their increased sfr surface densities and small sizes relative to normal sfgs on the ms .
while galaxies above and below the ms by definition form the opposite extremes in terms of sfr , they may thus share a more similar nature in terms of structural shape . according to the merger evolutionary sequence
, this would place them as nuclear starbursts in between ms galaxies and quiescent remnants .
a crude duty cycle argument reveals that , if all quiescent galaxies at @xmath13 with @xmath107 and @xmath108 underwent a starburst of @xmath109 prior to quenching , the timescale of that burst must have been of order 110 myr to account for the observed numbers of such bursts : @xmath110 here , @xmath111 represents the number density of objects with @xmath112 , and @xmath113 equals the number of newly quenched ( @xmath108 ) systems since @xmath114 ( which is most of them , given the relatively low number density at @xmath115 ) .
the time interval probed in the @xmath3 bin spans @xmath116 gyr . for the @xmath4 bin , similar arguments lead to a value of @xmath117 myr . to determine these timescales , we restricted ourselves to the deepest fields , goods - south and uds , to avoid any incompleteness effects .
our wfc3 @xmath26-selected samples in these fields do not suffer from incompleteness above @xmath118 , even for maximally old , high @xmath119 systems in our highest redshift bin ( see appendix b.1 ) . increasing the number statistics by pushing down to @xmath120 , where we are still highly complete
, we obtain timescales of 170 myr and 120 myr at @xmath3 and @xmath4 respectively .
the derived timescales should be considered as crude upper limits .
first , uncertainties in sfr estimates may lead to a contamination in the high - sfr tail by galaxies with intrinsically lower sfr .
second , in the case that not all excursions to high sfrs are followed by quenching , the inferred timescale of bursts associated with a quenching event will be shortened further .
overall , the timescales obtained are not inconsistent with those expected from merger - induced star formation on the basis of hydrodynamical simulations ( mihos & hernquist 1996 ; cox et al .
2008 ) , although there may be room for other evolutionary paths to reach the quiescent phase as well .
our observations of the structural ms and the quiescent population located below in figure [ n.fig ] reveal two aspects that any galaxy formation and quenching model should account for .
first , we find spheroid - like ( @xmath121 ) quiescent galaxies to be present over a wide range of stellar masses ( 1.5 - 2 dex ) at all redshifts , down to at least @xmath118 .
second , galaxies with a large spread in star formation activities co - exist at the same stellar mass over a mass range of an order of magnitude ( roughly @xmath122 ) .
if all galaxies at a given stellar mass are central galaxies and reside in an equally massive dark matter halo , this would imply that quenching is not a simple step function of halo mass . carrying out a detailed abundance matching analysis between galaxies at @xmath123 and their halo hosts , conroy & wechsler ( 2009 )
also do not find evidence for a sharp characteristic halo mass at which star formation truncates .
if on the other hand halo quenching is responsible for the transition from active to passive systems , this would imply that quiescent galaxies live in higher mass halos ( i.e. , have a smaller baryon fraction ) than sfgs of similar stellar mass .
finally , a third quenching mechanism that can contribute to the build - up of the quiescent population is environmental ( satellite ) quenching . studying the role of environment and its impact on the morphological properties of ( part of ) the quenched population , is beyond the scope of this paper .
we do however note that y. j. peng et al .
( 2010 ) argue this physical process only becomes relevant at late times ( @xmath124 ) and low masses ( @xmath125 ) .
based on an empirical model that successfully reproduces the stellar mass function of star - forming and passive galaxies in sdss , y. j. peng et al .
( 2010 ) attribute the quenching at larger lookback times to a combination of a mass - dependent component ( dominating at the high - mass end ) , and a mass - independent merger quenching component .
together , these processes are claimed to be responsible for the double schechter mass function of passive galaxies .
it is important to point out that their mass quenching is probabilistic in nature , rather than a step function , with a rate proportional to the sfr , and an underlying physical origin that may well be associated with agn or star formation feedback rather than a halo - scale phenomenon . in sections [ surfdens.sec ] and [ obscuration.sec ] , we presented observational results on how the surface density and obscuration of star formation depends on the position of galaxies in the sfr - mass diagram . here
, we investigate whether those two observations can be understood in the context of one self - consistent physical model that relates the amount of reprocessed emission to the column of obscuring material , and indirectly to @xmath0 .
the aim of such a model is to explain qualitatively the difference in shape of iso-@xmath0 and iso-@xmath1 regions in the sfr - mass plane ( figure [ surf.fig ] versus figure [ iruvobs.fig ] ) , and to reproduce quantitatively the observed @xmath1 ratio starting from @xmath0 , with the help of a few fundamental , and where possible empirically calibrated , laws .
our model involves three steps : a conversion from sfr surface density to gas surface density , a conversion from gas column to dust column , and a geometry - dependent translation from dust column to ir - over - uv ratio ( see top panel of figure [ iruvcomp.fig ] ) . the first step boils down to an application of the kennicutt - schmidt ( ks ) star formation law ( kennicutt 1998 ) .
recently , genzel et al . ( 2010 ) made the case for a ks law of the form @xmath126 ) = & 1.17 \log ( \sigma_{mol\ gas } [ m_{\sun } pc^{-2 } ] ) \nonumber \\ & - 3.48 \label{ks.eq}\end{aligned}\ ] ] on which both nearby and high - redshift main sequence sfgs lie .
major mergers follow a relation of similar slope according to these authors , but may be offset in normalization , so that a higher level of star formation is obtained for a given reservoir of molecular gas . for simplicity ,
we assume equation [ ks.eq ] holds for all galaxies in our sample , at low and high redshift .
we apply equation [ ks.eq ] to our observed @xmath0 , account for a 36% mass correction for helium to obtain the @xmath127 surface mass density , and add a fiducial plateau of @xmath128 surface mass density @xmath129 ( krumholz , mckee & tumlinson 2009 ) to obtain the total hydrogen column . in most cases ,
the contribution of hi to the total gas column is negligible .
the second step of our model accounts for the fact that not all galaxies have the same metallicity .
this is relevant , because the limited number of cases in which metallicities as well as dust - to - gas ratios have directly been observed ( milky way , small and large magellanic cloud ) , illustrate that it is not the dust - to - gas ratio , but rather the dust - to - metal ratio that remains relatively constant between galaxies .
we adopt a gas - to - dust ratio equal to that of the milky way @xmath130 ( bohlin et al . 1978 ) , with a linear scaling accounting for gas - phase metallicities deviating from solar : @xmath131 even in the era of multi - object near - infrared spectrographs , spectral metallicity tracers of high - redshift galaxies , such as @xmath132/h\alpha$ ] are notoriously difficult to obtain for large samples of galaxies , let alone for the sample size considered here .
we therefore adopt an empirically calibrated functional form to compute the metallicity on the basis of known stellar population properties . rather than assuming an evolution of the mass - metallicity relation with redshift @xmath133 ( e.g. , erb et al .
2006 ; maiolino et al .
2008 ; mannucci et al .
2009 ; buschkamp et al . in prep ) , we parametrize the metallicity as a non - evolving function of position in the sfr - mass diagram : @xmath134 . using the sdss spectral database , mannucci et al .
( 2010 ) established such a fundamental surface in sfr - mass - metallicity space .
furthermore , these authors find that galaxies out to @xmath5 lie on the same surface as do nearby systems .
the observed evolution of the mass - metallicity relation stems in this picture merely from the fact that galaxies with higher sfrs are more abundantly present at high redshift .
that higher sfr systems at a given mass have lower metallicities is interpreted as evidence for recent gas infall that simultaneously fuels the formation of new stars , and dilutes the metal - content of the interstellar medium . even in the local universe ,
measurements of the gas - phase metallicity ( and therefore the normalization of the mass - metallicity relationship ) vary by factors of @xmath135 depending on the tracer and calibration used ( kewley et al .
given this systematic uncertainty , we choose to renormalize the mannucci et al .
( 2010 ) relation so that it reaches a plateau of solar metallicity ( @xmath136 , asplund et al .
2004 ) at the high - mass end , so that for galaxies with milky way properties ( @xmath137 ) , the milky way dust - to - gas ratio applies : @xmath138 where @xmath139 . for @xmath140 , the smooth polynomial form from mannucci et al .
( 2010 , equation 4 ) is applied , downscaled by the same factor .
the scaling of the column with @xmath134 accounts at least in part for the difference in shape of the @xmath0 and @xmath1 diagrams ( figure [ surf.fig ] and [ iruvobs.fig ] respectively ) . while along the ms little evolution in @xmath0 is observed , gas - phase metallicities are expected to increase as we march up the ms .
this leads to an increase in the dust column , and consequently of the reprocessed fraction of the star formation towards the high - mass end of the ms , as observed in figure [ iruvobs.fig ] .
in other words , our observations do not probe metallicities directly , but are consistent with the presence of a mass - metallicity and/or mass - sfr - metallicity relation . )
is in principle inconsistent with the gas - phase metallicity scalings applied in our model . as a sanity check
, we verified that all conclusions drawn in this paper remain valid if we were to adopt a stellar metallicity equal to the gas - phase metallicity according to mannucci et al .
( 2010 ) : @xmath141 .
the effect of such a scaling is to slightly skew the sfr - mass diagram , such that generally lower mass galaxies ( and at a given mass particularly those with higher sfrs ) shift to masses that are lower by 0 to -0.2 dex . at metallicities of a half solar , the stellar masses decrease by a few percent in the median , while the median decrease amounts to -0.1 dex at a quarter of solar . ] using the calzetti et al .
( 2000 ) attenuation law in combination with equation [ nh_av.eq ] , we now compute the optical depth @xmath142 at the rest - frame 2800 wavelength where we defined @xmath16 : @xmath143 the third and final step of our model relates the optical depth to the ratio of reprocessed over unobscured star formation . for the geometry of a uniform foreground screen , i.e. , where all stars are hiding behind the total dust column , the intrinsic uv emission from young stars is reduced by a factor @xmath144 .
we quickly dismissed such a geometry , as it is physically implausible , and leads to inferred @xmath1 that exceed the observed values by several orders of magnitude .
clearly , uv emission from young stars is observed , and assuming a more realistic mixture of dust and stars in our model can account for this .
we choose to adopt the simplest possible model , in which the emitting sources and the dust are homogeneously mixed .
in such a configuration , the intrinsic emission is attenuated by a factor @xmath145 ( mcleod et al .
1993 ; frster schreiber et al .
rewritten to a functional form for @xmath1 , this becomes : @xmath146 we present the comparison of our model results to the observed values of @xmath1 in figure [ iruvcomp.fig ] .
solid and dashed contours enclose 25% and 75% of the galaxy distribution respectively . in the @xmath2 bin ,
passive galaxies , for which our @xmath1 is ill - defined ( see section [ obscuration.sec ] ) were excluded from the analysis . for nearby galaxies , and even at @xmath3 , we find a surprisingly good agreement between model and observations .
a slight bias is present at @xmath3 , in the sense that the inferred @xmath1 for the bulk of the galaxies exceeds the observed values by a factor of @xmath147 .
this systematic offset increases further toward higher redshift , amounting to an order of magnitude at @xmath4 .
we speculate briefly on the physical origin of this trend .
a refined model should account for an increased amount of uv emission escaping unhindered from high - redshift galaxies .
this situation is reminiscent of adelberger & steidel ( 2000 ) , who found that for galaxies of a given sfr , at high redshift a larger proportion of that star formation is emitted in the uv .
daddi et al .
( 2007 ) also discussed that @xmath4 galaxies are more `` transparent '' to uv radiation than local galaxies with similar levels of star - forming activity .
the offsets found in the @xmath3 and @xmath4 panels of figure [ iruvcomp.fig ] could signal shortcomings in any of the three steps of our simple model .
since the first two steps ( star formation law , and metallicity scaling ) are empirically motivated , and the third ( geometry ) is not , it makes sense to address the latter first .
given that star formation , dust production , and the physical processes ( such as winds ) that determine the distribution of the ism are intimately connected , the idealized model of a homogeneous mixture of dust and ( young ) stars will likely break down at some level .
if high - redshift galaxies are more patchy than their local counterparts , this would naturally explain why a relatively larger fraction of the intrinsic emission can escape the galaxies unattenuated .
in contrast to galaxies today , the uv and ir emission of high - redshift systems would in that case more often be associated with physically distinct regions within a galaxy .
if such a patchy configuration would indeed be more representative for @xmath4 galaxies , we would expect this to be revealed in terms of rich internal color variations in multi - wavelength high - resolution imaging .
such internal color variations , and their interpretation in terms of spatially resolved stellar populations will be the focus of a forthcoming paper ( wuyts et al . in prep ) .
first steps in this direction have been taken through visual classification ( cameron et al .
2010 ) and quantitative internal color dispersion methods ( bond et al . 2011 ) to compare acs and wfc3 morphologies of distant galaxies in the ers and hudf regions . alternatively , the results of our model comparison may signal deviations from the adopted star formation law ( equation [ ks.eq ] ) . a tuning of the ks law to enforce an agreement between the modeled and observed
@xmath1 would require a higher normalization and/or a steeper slope than 1.17 , so that a given @xmath0 corresponds to a smaller column of obscuring material .
genzel et al .
( 2010 ) demonstrate that for a subsample of galaxies ( those identified as mergers ) a higher normalization is indeed appropriate . depending on their contribution to the overall galaxy population
, this may lead to a somewhat steeper slope of the effective ks relation fitted to the combined sample of normal sfgs and mergers .
the change in slope required to bring our model prediction in agreement with the observations ( keeping the zeropoint fixed ) is substantial : a slope of 1.35 at @xmath3 and 1.75 at @xmath4 .
ks slopes of this magnitude have been reported in the literature ( e.g. , bouch et al .
2007 ) , but in combination with a lower zeropoint of the relation .
if instead we keep the slope fixed to 1.17 , a zeropoint increase to -3.2 at @xmath3 and -2.5 at @xmath4 would be required to obtain agreement between model and observations .
a large fraction of @xmath4 galaxies for which gas surface densities have been measured do lie on a relation with slope 1.17 and zeropoint -2.44 , but that is because due to sensitivity limitations these samples have often been biased to high - sfr outliers .
recent measurements for @xmath4 galaxies on the main sequence ( i.e. , representative for the bulk of the @xmath4 population ) are consistent with the local relation for normal sfgs ( genzel et al .
2010 ; daddi et al .
2010 ) , and fitted by themselves give a zeropoint of -3.11 ( and slope of 1.17 ) .
the required change therefore seems to be in tension with the observational constraints at @xmath4 available to date .
finally , a lower dust column for a given @xmath0 could also be obtained if our metallicities of high - redshift galaxies are overestimated .
if this is the case , this would imply that gas - phase metallicity is determined by three parameters ( @xmath148 ) rather than a non - evolving dependence on sfr and stellar mass .
we stress that current observations out to @xmath5 do not require such an evolution with redshift of the sfr - mass - metallicity plane , but the high - redshift samples presently show a large scatter ( mannucci et al .
moreover , while we are still facing large systematic uncertainties in metallicity measurements of even local galaxies , we already took the liberty to rescale the mannucci et al .
( 2010 ) relation to reach a plateau of solar metallicity at the high - mass end , which is among the lower end of calibrations presented in the literature ( kewley et al .
vice versa , if we were to adopt the mannucci et al .
( 2010 ) relation without rescaling ( i.e. , reaching a plateau of @xmath149 , or 2.6 times solar ) , the model prediction of @xmath1 would exceed the observed values by a factor 5 , 10 , and 60 at @xmath2 , @xmath3 , and @xmath4 .
we conclude that a simple model , combining a mixed dust - star geometry with an empirical star formation law and metallicity scaling , is successful in describing the obscuration and surface density of star formation in a self - consistent picture . towards higher redshifts , an adjustment of this model
is desired , allowing more of the uv emission to escape .
this may be due to the fact that galaxies at higher lookback times feature more patchy geometries .
we use statistically significant samples of unprecedented size ( @xmath150 galaxies at @xmath2 , @xmath151 at @xmath3 , and @xmath152 at @xmath4 ) to analyze how the structure and mode of star formation of galaxies at different lookback times depends on their position in the sfr - mass diagram . to this end
, we combine 4 deep lookback surveys ( cosmos , uds , goods - n , goods - s ) with high - resolution imaging and rich multi - wavelength data sets reaching from the optical and near - infrared to mid- and far - infrared wavelengths . value - added sdss catalogs are complemented with galex data to investigate equivalent galaxy properties in the nearby universe .
we find remarkable similarities over 10 gyr of lookback time .
e.g. , the relation between surface brightness profile shape on the one hand and the position of a galaxy with respect to the main sequence of star formation on the other hand is strikingly similar at all observed epochs . phrased differently , we find evidence for a correlation between galaxy structure and stellar populations ( i.e. , a hubble sequence ) to be in place already three billion years after the big bang .
the fact that a structurally distinct quiescent population below the main sequence is already present since @xmath5 implies that the quenching mechanism responsible for the shutdown of star formation must be universally present already since that epoch , and must be causally connected to the morphological transition as well ( or vice versa , see martig et al .
2009 who analyze a morphological quenching scenario ) .
low- and high - redshift galaxies also seem to follow a qualitatively similar behavior in terms of star formation surface density and its dependence on the specific sfr ( see also elbaz et al .
2011 ) , and the variation in obscuration ( parametrized by @xmath1 ) along and across the ms .
they differ in the sense that the high end of the relation between @xmath0 and @xmath97 is more abundantly populated at larger lookback times .
another aspect in which low- and high - redshift galaxies differ is their typical size .
measured at fixed @xmath62 $ ] , galaxies on the ms today experienced a slower growth over time ( @xmath153 ) than the quiescent population ( @xmath154 ) .
these growth rates are consistent with earlier work based on ground - based imaging ( trujillo et al .
2006 ; franx et al .
2008 ; toft et al .
2009 ) and recent wfc3 measurements by nagy et al .
( 2011 ) . as star - forming galaxies in our sample that satisfy the same selection criteria
. their steeper growth rate ( @xmath155 within @xmath156 ) stems from the fact that they do not anchor the size evolution to @xmath157 , i.e. , they effectively use a different parametrization . ]
a simple mass - size relation is inadequate to describe the size variations of the entire galaxy population .
at all epochs , star - forming galaxies on the ms are the largest among galaxies of a given mass .
the profile shape of galaxies , quantified by the sersic index @xmath57 , is not a monotonic function of the level of star formation . instead , at a given stellar mass a reversal towards higher @xmath57
is observed in the high - sfr tail of the distribution .
both our low- and high - redshift samples show evidence of such a relation .
this trend may hint at an evolutionary connection between the high - sfr outliers , which are in the process of building up a nuclear concentration of mass , and the quiescent galaxy population , which is characterized by cuspy surface brightness profiles , as proposed by the merger paradigm .
the variation of @xmath0 across the sfr - mass diagram shows a different behavior than that of @xmath1 . while iso-@xmath0 regions run diagonal , following approximately lines of constant specific sfr ( see also schiminovich et al .
2007 for @xmath2 ) , the @xmath1 ratio increases as we move up along the ms ( i.e. , at fixed specific sfr ) .
this strongly suggests variations in metallicity along the main sequence , as expected from studies of the mass - metallicity ( erb et al .
2006 ; buschkamp et al . in prep ) and sfr - mass - metallicity ( mannucci et al . 2010 ) relation .
an increase in gas - phase metallicity towards the high - mass end of the ms would cause the obscuring column to increase , even if the gas column remains the same . using a simple model involving the observed kennicutt - schmidt star formation law ( genzel et al .
2010 ) and an empirically calibrated metallicity scaling ( mannucci et al . 2010 ) in combination with an assumed dust - star geometry , we compare the observed @xmath1 ratios to those inferred from the observed @xmath0 . our model is successful at qualitatively relating the two properties , and explaining why their iso - contours in the sfr - mass plane are of different shape . for a metallicity scaling that flattens at solar values at the high - mass end , we find a good quantitative correspondence at @xmath2 and @xmath3 when assuming a homogeneous mixture of dust and stars . at higher redshifts ,
progressively more patchy dust - star geometries appear to be required in order to have sufficient uv emission escaping to match the observed @xmath1 .
this work is based on observations taken by the candels multi - cycle treasury program with the nasa / esa hst , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
funding for the sloan digital sky survey ( sdss ) has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , princeton university , the united states naval observatory , and the university of washington .
the galaxy evolution explorer ( galex ) is a nasa small explorer .
the mission was developed in cooperation with the centre national detudes spatiales of france and the korean ministry of science and technology .
similar to figure [ dev.fig ] in section [ profile.sec ] , we here present the dependence of the effective radius @xmath85 , surface density of star formation @xmath0 , and obscuration @xmath1 of galaxies with @xmath158 as a function of deviation from the ms ( figure [ dev_app.fig ] ) .
the scatter at a given @xmath74 ( _ grey polygons _ ) stems from a combination of uncertainties in the derived parameters and intrinsic variations within the galaxy population of a given specific sfr .
especially the relations reflecting the mode of star formation ( @xmath0 and @xmath1 ) with respect to the ms are remarkably tight .
we follow two approaches to address the robustness of the trends in sfr - mass space described in this paper . first
, we compare empirically the results for individual fields ( b.1 ) and for data sets of different depth and wavelength in a given field ( b.2 ) .
second , we use monte carlo simulations ( b.3 ) to propagate different sources of uncertainty ( photometry , redshifts , sfrs , structural parameters , ... ) and assess their impact on the median properties of the overall galaxy population .
while specific measurements on individual objects may sometimes be subject to significant uncertainties , the conclusions drawn in this paper are much less affected as they are based on median properties over large numbers of galaxies , and the level of systematic biases is small .
lrrrrrr[t ] cosmos & 1.480 & @xmath22 & 27.2 & 25.0 & 106080 & 21430 + uds & 0.056 & @xmath26 & 26.7 & 26.7 & 10443 & 6796 + goods - s & 0.041 & @xmath26 & 27.0 & 27.0 & 7008 & 3973 + goods - n & 0.042 & @xmath36 & 27.6 & 26.8 & 8797 & 3450 + [ fields.tab ] our results at intermediate and high redshift are based on 4 lookback surveys , each with a wealth of multi - wavelength data .
they complement each other in area and depth , and differ in some cases in the selection band of the multi - wavelength catalog as well as the resolution and waveband of available hst imaging ( see table [ fields.tab ] ) .
a comparison of the results obtained for each of the fields individually therefore serves as an empirical robustness check that encompasses field - to - field variations , the impact of imaging depth on morphological measurements and completeness within a given @xmath62 $ ] bin , resolution , and morphological k - corrections , if present . in figure
[ field_per_field.fig ] , we demonstrate that the median behavior of @xmath57 , @xmath85 , @xmath0 and @xmath1 looks strikingly similar for our samples in cosmos , uds , goods - s and goods - n , despite inhomogeneities between the data sets . in cases where the patterns look somewhat less smooth ,
this is to be expected from the smaller number statistics . from the similarity of the @xmath0 and @xmath1 diagrams
, it follows that also our simple model and interpretation in terms of patchier dust geometries at high redshift holds for each of the fields separately .
the region of sfr - mass space covered by each of the fields in figure [ field_per_field.fig ] , reflects differences in depth .
we expand on this in figure [ completeness.fig ] , where the color - coding illustrates the estimated completeness of our sample . for the cosmos and goods - n fields ,
the completeness in each @xmath62 $ ] bin was defined empirically as the fraction of uds and goods - s galaxies that would enter the @xmath80- and @xmath81-band limited samples for those fields respectively . for the deep wfc3 @xmath159-band selected samples in uds and goods - s , we estimated the completeness indirectly . for a given @xmath62 $ ] bin , we dimmed the seds of galaxies with similar specific sfr , but higher sfr and mass ( i.e. , located in a highly complete region of the diagram ) .
as completeness estimate , we adopted the fraction of galaxies downscaled to the @xmath62 $ ] bin of interest that would still be sufficiently bright to enter the sample .
the latter is best interpreted as an upper bound on the completeness . as extinction
is known to scale with the star formation activity ( see , e.g. , the increased obscuration as one moves along the ms toward the high - mass end in figure [ iruvobs.fig ] ) , our simple approach of downscaling the sed by the same factor as the sfr and mass , without altering the sed shape , likely leads to a completeness estimate that is too conservative .
figure [ completeness.fig ] illustrates that the completeness level of our cosmos sample is lower than that of the samples extracted from the other fields .
as is evident from figure [ field_per_field.fig ] , cosmos therefore gets zero weight at the lowest sfrs and masses probed in the @xmath4 bin . in @xmath62
$ ] bins that do contain cosmos galaxies , two tests indicate that the somewhat lower completeness level does not lead to significant biases in the median morphology or mode of star formation .
first , the trends in cosmos do not differ from those observed in the other fields ( figure [ field_per_field.fig ] ) .
second , we applied the same @xmath80-band limit adopted for cosmos to the highly complete wfc3-selected catalog in uds .
we verified that this does not change the median morphological and mode of star formation parameters in @xmath62 $ ] bins that still contain sources after imposing the @xmath80-band cut .
while boosting confidence in the robustness of our results , the field - to - field comparison in b.1 did not discriminate between different aspects of inhomogeneity between the data sets . using the goods - s field as a test case
, we now isolate the effect of depth , and the effect of waveband on which the surface brightness profile fitting is performed .
the top panels of figure [ depth_wave.fig ] illustrate how the morphological parameters change when measured on the presently available full depth mosaic of the candels - deep region , and on a version of the mosaic built with only the first half of the exposures ( 0.35 mag shallower , similar in depth to the candels - wide data used for uds ) .
unsurprisingly , the scatter in @xmath160 is observed to increase toward fainter magnitudes . however , down to the faintest magnitudes we find no evidence for a systematic bias in the size measurement .
this implies that population studies which describe the median size , such as presented in this paper , can safely probe to fainter levels than would be considered reliable on an individual object level .
hussler et al . (
2007 ) arrive at the same qualitative conclusion when inserting simulated @xmath64 and @xmath65 galaxies in mock acs images , and recovering the morphological parameters using galfit . for similar tests aimed specifically at the candels data sets ,
we defer the reader to van der wel et al .
( in prep ) .
it is relevant to note that our conclusion may not necessarily hold for other methods to determine @xmath85 than the galfit code used in our analysis ( see also hussler et al . 2007 ; sargent et al . 2007 ) .
the difference @xmath161 between the sersic index measured on the 2- and 4-epoch mosaics also exhibits an increased scatter toward fainter magnitudes , but again the median @xmath161 remains consistent with zero . in detail , however , a trend is noted such that shallow profile shapes are on average recovered correctly down to the faintest magnitudes , while the sersic index of intrinsically cuspy systems shows a systematic deviation at the faint end . of order -2 or lower can by definition not be reached for disk - dominated systems because of the allowed range @xmath58 . ] in relative terms , the underestimate for faint @xmath67 galaxies corresponds to a median @xmath162 . in sfr - mass space , these galaxies are predominantly located at @xmath163 and @xmath164 . if they were the dominant population in that area of sfr - mass space , they could potentially bias the median @xmath57 to low values .
our deepest , 4 epoch goods - s data suggests however that they only account for less than 10% of the galaxy population in the low mass and low sfr regime .
the bottom panels of figure [ depth_wave.fig ] compare measurements of @xmath57 and @xmath85 carried out on the wfc3 @xmath26 mosaic , and on the acs @xmath36 mosaic .
we plot the deviation ( negative where the @xmath36 measurement yields a smaller value ) as a function of redshift to facilitate the interpretation in terms of morphological k - corrections in the rest - frame .
variations in the extent of the @xmath36- and @xmath26-band surface brightness profiles are small compared to the amplitude of the size variation across the sfr - mass diagram .
median offsets are limited to the few percent level for @xmath85 , and of order @xmath165% for the sersic index ( such that the longer wavelength surface brightness profiles tend to be cuspier ) .
overall , the systematic deviations in @xmath57 are small compared to the amplitude of trends with profile shape across the sfr - mass plane , although we caution that the random scatter in @xmath57 due to morphological k - corrections becomes substantial for systems with @xmath67 in the @xmath26 band .
we ran extensive monte carlo ( mc ) simulations to estimate the effects of photometric and photometric redshift uncertainties on our analysis .
first , we selected a representative subset of galaxies in each of the four deep fields . we sample four redshift intervals ( @xmath166 , @xmath167 , @xmath168 , and @xmath169 ) , drawing for each one three galaxies from each @xmath62 $ ] bin , spanning the lower , middle and upper tertile in signal - to - noise . for the resulting subsample of 7480 galaxies , we then compiled one hundred mock photometric catalogs .
the flux - points in each optical - to - ir band were drawn randomly from a gaussian with the measured flux as the mean and its error as the standard deviation .
next , we subject each mock catalog to the same fitting procedures to derive photometric redshifts , stellar masses , and sfrs , and also propagate the newly obtained redshifts into the conversion from angular to physical half - light radii .
together , the 100 mock realizations provide us with an uncertainty distribution on ( and often correlated between ) each of the parameters that enter our analysis . in order to apply the mc results , we associate each galaxy from our full sample with a galaxy in the mc subsample that lies in the same field , and occupies the same redshift , sfr , mass and signal - to - noise bin .
the deviations with respect to the actual measurements of redshift , mass , sfr , and @xmath85 are drawn from one of the 100 mock realizations of the associated mc target , and applied to the galaxy of interest .
the resulting mock population is then again divided in @xmath62 $ ] bins and we compute the median of the property of interest ( @xmath57 , @xmath85 , @xmath0 , @xmath1 ) , exactly as we did for our analysis of the real measurements . in principle , the deviations in derived parameters ( which unlike the photometry are not assumed or forced to be symmetric ) should be applied to error - free measurements .
since the latter are not available , our test serves as a conservative estimate of how important photometric errors are in our analysis . upon visual inspection of the mc - ed equivalents of figures [ n.fig ] to [ iruvcomp.fig ]
, we find all the trends described in this paper to be robust against photometric uncertainties .
however , the @xmath40 uncertainty distribution as inferred from running eazy on the 100 perturbed photometric catalogs , is likely too narrow .
we find evidence for this by running a similar set of 100 simulations on the full spectroscopically confirmed sample .
the spectroscopic redshift lies within the formal 1@xmath21 confidence interval derived from the mc simulation in only about 30% of the cases , much less than the expectation value of 68% were photometric errors the only source of uncertainty in deriving the redshift . in a second iteration , we therefore base deviations in @xmath40 on the empirical comparison between photometric redshifts and the available spectroscopic redshifts .
photometric and photometric redshift uncertainties are then propagated to the other parameters as before . here , it is important to account for the fact that the spectroscopic sample is not randomly drawn from the overall galaxy sample , but instead biased towards bright magnitudes .
following scarlata et al .
( 2007 ) , we quantified the photometric redshift uncertainty at a given magnitude semi - empirically by dimming the seds of spectroscopically confirmed galaxies to that magnitude ( perturbing the photometry accordingly ) , and deriving the resulting @xmath41 distribution . as expected , this yields a broader distribution , at the faintest levels by up to a factor of 5 compared to the normalized median absolute deviation of the undimmed spectroscopic sample .
however , even at the faintest levels , no significant systematic offsets in @xmath40 are revealed .
figure [ mc.fig ] presents the mc - ed equivalents of figures [ n.fig ] to [ iruvobs.fig ] , where photometric errors and semi - empirical photometric redshift uncertainties have been applied and propagated in the analysis .
we considered the distribution in @xmath170 , where @xmath171 stands for the median structural ( @xmath57 , @xmath85 ) or mode of star formation ( @xmath0 , @xmath1 ) property in a given @xmath62 $ ] cell .
systematic deviations from zero are limited to @xmath172 dex . at @xmath13 ,
the scatter of the @xmath173 distribution is modest , amounting to 0.04 , 0.04 , 0.07 , and 0.12 dex in @xmath57 , @xmath85 , @xmath0 and @xmath1 respectively . at @xmath38 , the scatter in @xmath173 doubles .
we conclude that , when following this more conservative approach , we still find that the patterns in sfr - mass space described in this paper are robust against photometric and photometric redshift uncertainties . in a third and final iteration
, we now also include random uncertainties in mass , @xmath174 , @xmath85 , and @xmath57 that do not stem from photometric or photometric redshift uncertainties . to this end
, we adopt fiducial errors of 0.2 dex in each of these parameters .
this crudely reflects our ability to recover the intrinsic surface brightness profile shape ( b.2 ) , as well as imperfections with which observables are translated to physical parameters ( see , e.g. , wuyts et al . 2009 , 2011a ) .
the scatter in @xmath173 is boosted by @xmath175 compared to the mc simulation that only accounted for photometric and photometric redshift uncertainties .
nevertheless , the overall conclusions drawn in this paper remain valid .
abazajian , k. n. , et al .
2009 , apjs , 182 , 543 adelberger , k. l. , & steidel , c. c. 2000 , apj , 544 , 218 alexander , d. m. , et al .
2003 , aj , 126 , 539 asplund , m. , grevesse , n. , sauval , a. j. , allende prieto , c. , & kiselman , d. 2004 , a&a , 417 , 751 barger , a. j. , cowie , l. l. , sanders , d. b. , fulton , e. , taniguchi , y. , sato , y. , kawara , k. , & okuda , h. 1998 , nature , 394 , 248 barger , a. j. , cowie , l. l. , & wang , w .- h .
2008 , apj , 689 , 687 barnes , j. e. , & hernquist , l. 1991 , apj , 370 , 65 barnes , j. e. , & hernquist , l. 1996 , apj , 471 , 115 bell , e. f. 2008 , apj , 682 , 355 berta , s. , et al .
2010 , a&a , 518 , 30 bianchi , l. , efremova , b. , herald , j. , girardi , l. , zabot , a. , marigo , p. , & martin , c. 2011 , mnras , 411 , 2770 blain , a. w. , smail , i. , ivison , r. j. , kneib , j .- p .
, & frayer , d. t. 2002 , phr , 369 , 111 birnboim , y. , & dekel , a. 2003 , mnras , 345 , 349 blanton , m. r. , et al .
2003 , apj , 594 , 186 blanton , m. r. , et al .
2005 , aj , 129 , 2562 bohlin , r. c. , savage , b. d. , & drake , j. f. 1978 , apj , 224 , 132 bond , n. a. , gawiser , e. , & koekemoer , a. m. 2011 , apj , 729 , 48 bouch , n. , et al .
2007 , apj , 671 , 303 brammer , g. b. , van dokkum , p. g. , & coppi , p. 2008 , apj , 686 , 1503 brinchmann , j. , charlot , s. , white , s. d. m. , tremonti , c. , kauffmann , g. , heckman , t. , & brinkmann , j. 2004 , mnras , 351 , 1151 bruzual , g , & charlot , s. 2003 , mnras , 344 , 1000 calzetti , d. , armus , l. , bohlin , r. c. , kinney , a. l. , koornneef , j. & storchi - bergmann , t. 2000 , apj , 533 , 682 cameron , e. , carollo , c. m. , oesch , p. a. , bouwens , r. j. , illingworth , g. d. , trenti , m. , labb , i. , & magee , d. 2010 , submitted to apj ( arxiv1007.2422 ) cappelluti , n. , et al .
2009 , a&a , 497 , 635 cassata , p. , et al .
2011 , apj , in press ( arxiv1106.4308 ) chabrier , g. 2003 , pasp , 115 , 763 conroy , c. , & wechsler , r. h. 2009 , apj , 696 , 620 cox , t. j. , jonsson , p. , somerville , r. s. , primack , j. r. , & dekel , a. 2008 , mnras , 384 , 386 daddi , e. , et al .
2005 , apj , 626 , 680 daddi , e. , et al . 2007a , apj , 670 156 daddi , e. , et al .
2010 , apj , 714 , 118 dekel , a. , & birnboim , y. 2006 , mnras , 368 , 2 dekel , a. , et al .
2009 , nature , 457 , 451 di matteo , t. , springel , v. , & hernquist , l. 2005 , nature , 433 , 604 elbaz , d. , et al .
2007 , a&a , 468 , 33 elbaz , d. , et al .
2010 , a&a , 518 , 29 elbaz , d. , et al .
2011 , a&a , 533 , 119 erb , d. k. , shapley , a. e. , pettini , m. , steidel , c. c. , reddy , n. a. , & adelberger , k. l. 2006 , apj , 644 , 813 frster schreiber , n. m. , genzel , r. , lutz , d. , kunze , d. , & sternberg , a. 2001 , apj , 552 , 544 frster schreiber , n. m. , et al .
2009 , apj , 706 , 1364 frster schreiber , n. m. , et al .
2011 , apj , 731 , 65 franx , m. , van dokkum , p. g. , frster schreiber , n. m. , wuyts , s. , labb , i. , & toft , s. 2008 , apj , 688 , 770 gabasch , a. , goranova , y. , hopp , u. , noll , s. , & pannella , m. 2008 , mnras , 383 , 1319 genzel , r. , et al .
2008 , apj , 687 , 59 genzel , r. , et al .
2010 , mnras , 407 , 2091 giavalisco , m. , et al .
2004 , apj , 600 , 93 grogin , n. a. , et al .
2011 , submitted to apjs ( arxiv1105.3753 ) guo , y. , et al .
2011 , apj , 735 , 18 hussler , b. , et al .
2007 , apjs , 172 , 615 heckman , t. m. , robert , c. , leitherer , c. , garnett , d. r. , & van der rydt , f. 1998 , apj , 503 , 646 hopkins , a. m. , & beacom , j. f. 2006 , apj , 651 , 142 hopkins , p. f. , hernquist , l. , cox , t. j. , di matteo , t. , robertson , b. , & springel , v. 2006 , apjs , 163 , 1 hopkins , p. f. , hernquist , l. , cox , t. j. , keres , d. , & wuyts , s. 2009 , apj , 691 , 1424 hopkins , p. f. , younger , j. d. , hayward , c. c. , narayanan , d. , & hernquist , l. 2010 , mnras , 402 , 1693 hughes , d. h. , et al .
1998 , nature , 394 , 241 ilbert , o. , et al .
2009 , apj , 690 , 1236 kauffmann , g. , et al .
2003 , mnras , 341 , 54 kennicutt , r. c. 1998 , ara&a , 36 , 189 keres , d. , katz , n. , weinberg , d. h. , & dav , r. 2005 , mnras , 363 , 2 keres , d. , katz , n. , fardal , m. , dav , r. , & weinberg , d. h. 2009 , mnras , 395 , 160 kewley , l. j. , & ellison , s. l. 2008 , apj , 681 , 1183 khochfar , s. , & silk , j. 2006 , apj , 648 , 21 koekemoer , a. m. , fruchter , a. s. , hook , r. n. , & hack , w. 2002 , hst calibration workshop ( eds .
s. arribas , a. m. koekemoer , b. whitmore ; stsci : baltimore ) , 337 koekemoer , a. m. , et al .
2007 , apjs , 172 , 196 koekemoer , a. m. , et al .
2011 , submitted to apjs ( arxiv1105.3754 ) kriek , m. , et al .
2006 , apj , 649 , 71 kriek , m. , van dokkum , p. g. , labb , i. , franx , m. , illingworth , g. d. , marchesini , d. , & quadri , r. f. 2009a , apj , 700 , 221 kriek , m. , van dokkum , p. g. , franx , m. , illingworth , g. d. , & magee , d. k. 2009b , apj , 705 , 71 krumholz , m. r. , mckee , c. f. , & tumlinson , j. 2009 , apj , 693 , 216 le floch , e. , et al .
2009 , apj , 703 , 222 lilly , s. j. , et al .
2009 , apjs , 184 , 218 luo , b. , et al .
2008 , apjs , 179 , 19 lutz , d. , et al .
2011 , a&a , 532 , 90 magnelli , b. , elbaz , d. , chary , r. r. , dickinson , m. , le borgne , d. , frayer , d. t. , & willmer , c. n. a. 2009 , a&a , 496 , 57 maier , c. , et al .
2009 , apj , 694 , 1099 maiolino , r. , et al .
2008 , a&a , 488 , 463 mannucci , f. , et al .
2009 , mnras , 398 , 1915 mannucci , f. , cresci , g. , maiolino , r. , marconi , a. , & gnerucci , a. 2010 , mnras , 408 , 2115 maraston , c. , pforr , j. , renzini , a. , daddi , e. , dickinson , m. , cimatti , a. , & tonini , c. 2010 , mnras , 407 , 830 martig , m. , bournaud , f. , teyssier , r. , & dekel , a. 2009 , apj , 707 , 250 mcgrath , e. j. , stockton , a. , canalizo , g. , iye , m. , & maihara , t. 2008 , apj , 682 , 303 mcleod , k. k. , rieke , g. h. , rieke , m. j. , & kelly , d. m. 1993 , apj , 412 , 111 mihos , j. c. , & hernquist , l. 1996 , apj , 464 , 641 nagy , s. r. , law , d. r. , shapley , a. e. , & steidel , c. c. 2011 , apjl , 735 , 19 noeske , k. g. , et al .
2007 , apj , 660 , 43 nordon , r. , et al .
2010 , a&a , 518 , 24 peng , c. y. , ho , l. c. , impey , c. d. , & rix , h .- w .
2010 , aj , 139 , 2097 peng , y. j. , et al .
2010 , apj , 721 , 193 reddy , n. a. , erb , d. k. , pettini , m. , steidel , c. c. , shapley , a. e. 2010 , apj , 712 , 1070 rodighiero , g. , et al .
2010 , a&a , 518 , 25 rodighiero , g. , et al .
2011 , apjl , 739 , 40 salim , s. , et al .
2007 , apjs , 173 , 267 sanders , d. b. , soifer , b. t. , elias , j. h. , madore , b. f. , matthews , k. , neugebauer , g. , & scoville , n. z. 1988 , apj , 325 , 74 sanders , d. b. , et al .
2007 , apjs , 172 , 86 sargent , m. t. , et al .
2007 , apjs , 172 , 434 scarlata , c. , et al .
2007 , apjs , 172 , 494 schiminovich , d. , et al .
2007 , apjs , 173 , 315 scoville , n. z. , et al .
2007 , apjs , 172 , 1 shapiro , k. l. , et al .
2008 , apj , 682 , 231 shen , s. , mo , h. j. , white , s. d. m. , blanton , m. r. , kauffmann , g. , voges , w. , brinkmann , j. , & csabai , i. 2003 , mnras , 343 , 978 simmons , b. d. , & urry , c. m. 2008 , apj , 683 , 644 smail , i. , ivison , r. j. , & blain , a. w. 1997 , apj , 490 , 5 snyder , g. f. , cox , t. j. , hayward , c. c. , hernquist , l. , & jonsson , p. 2011 , submitted to apj ( arxiv1102.3689 ) szomoru , d. , franx , m. , bouwens , r. j. , van dokkum , p. g. , labb , i. , illingworth , g. d. , & trenti , m. 2011 , apj , 735 , 22 tacconi , l. j. , et al . 2008 , apj , 680 , 246 tacconi , l. j. , et al . 2010 , nature , 463 , 781 toft , s. , franx , m. , van dokkum , p. , frster schreiber , n. m. , labb , i. , wuyts , s. , & marchesini , d. 2009 , apj , 705 , 255 toomre , a. , & toomre , j. 1972 , apj , 178 , 623 van der wel , a. , holden , b. p. , zirm , a. w. , franx , m. , rettura , a. , illingworth , g. d. , & ford , h. c. 2008 , apj , 688 , 48 trujillo , i. , et al .
2006 , apj , 650 , 18 ueda , y. , et al .
2008 , apjs , 179 , 124 van der wel , a. , et al .
2011 , apj , 730 , 38 van dokkum , p. g. , et al .
2008 , apj , 677 , 5 van dokkum , p. g. , et al .
2011 , submitted to apjl ( arxiv1108.6060 ) vanzella , e. , et al .
2008 , a&a , 478 , 83 williams , r. j. , quadri , r. f. , franx , m. , van dokkum , p. g. , toft , s. , kriek , m. , & labb , i. 2010 , apj , 713 , 738 windhorst , r. a. , et al .
2011 , apjs , 193 , 27 wuyts , s. , et al .
2007 , apj , 655 , 51 wuyts , s. , labb , i. , frster schreiber , n. m. , franx , m. , rudnick , g. , brammer , g. b. , & van dokkum , p. g. 2008 , apj , 682 , 985 wuyts , s. , franx , m. , cox , t. j. , hernquist , l. , hopkins , p. f. , robertson , b. e. , & van dokkum , p. g. 2009 , apj , 696 , 348 wuyts , s. , cox , t. j. , hayward , c. c. , franx , m. , hernquist , l. , hopkins , p. f. , jonsson , p. , & van dokkum , p. g. 2010 , apj , 722 , 1666 wuyts , s. , et al .
2011 , apj , 738 , 106 | we analyze the dependence of galaxy structure ( size and sersic index ) and mode of star formation ( @xmath0 and @xmath1 ) on the position of galaxies in the sfr versus mass diagram .
our sample comprises roughly 640000 galaxies at @xmath2 , 130000 galaxies at @xmath3 , and 36000 galaxies at @xmath4 .
structural measurements for all but the @xmath2 galaxies are based on hst imaging , and sfrs are derived using a herschel - calibrated ladder of sfr indicators .
we find that a correlation between the structure and stellar population of galaxies ( i.e. , a hubble sequence ) is already in place since at least @xmath5 .
at all epochs , typical star - forming galaxies on the main sequence are well approximated by exponential disks , while the profiles of quiescent galaxies are better described by de vaucouleurs profiles . in the upper envelope of the main sequence
, the relation between the sfr and sersic index reverses , suggesting a rapid build - up of the central mass concentration in these starbursting outliers .
we observe quiescent , moderately and highly star - forming systems to co - exist over an order of magnitude or more in stellar mass . at each mass and redshift
, galaxies on the main sequence have the largest size .
the rate of size growth correlates with specific sfr , and so does @xmath0 at each redshift .
a simple model using an empirically determined sf law and metallicity scaling , in combination with an assumed geometry for dust and stars is able to relate the observed @xmath0 and @xmath1 , provided a more patchy dust geometry is assumed for high - redshift galaxies . |
a comprehensive physics programme is employed by the h1 and zeus experiments at the hera @xmath0 collider .
together with measuring the structure of the proton , the deep inelastic collisions ( dis ) produced at hera , at a centre of mass up to 318 gev , provide an ideal environment to study rare processes , set constraints on the standard model ( sm ) and search for new particles and physics beyond the standard model ( bsm ) .
events containing a high transverse momentum ( @xmath1 ) isolated electron or muon and large missing transverse momentum have been observed at hera @xcite .
the main sm contribution to such a topology comes from the production of real @xmath2 bosons @xmath3 with subsequent leptonic decay @xmath4 .
an excess of hera i ( 19942000 ) data events compared to the sm prediction was reported by the h1 collaboration in @xcite , which was not confirmed by the zeus collaboration , although using a slightly different analysis approach @xcite .
the h1 analysis has been updated to include new @xmath5 data from the ongoing hera ii phase ( since 2003 ) , resulting in a total analysed luminosity of 279 pb@xmath6 @xcite .
a total of 40 events are observed in the data , compared to a sm prediction of 34.3 @xmath7 4.8 . at large values of hadronic transverse momentum @xmath8 an excess of data events
is observed over the sm expectation , as can be seen in figure [ fig : isolep ] .
additionally , the observed h1 excess is only present in the @xmath9 data , where for @xmath10 25 gev a total of 15 data events are observed compared to a sm prediction of 4.6 @xmath7 0.8 .
a re - analysis of the zeus electron channel has been performed @xcite , using @xmath9 data from 1999 - 2000 in addition to hera ii @xmath9 data from 2003 - 04 ( total luminosity 106 pb@xmath6 ) , leading to similar backgrounds as in the h1 analysis .
however , the zeus analysis does not confirm the h1 excess , where only one electron candidate is observed in the region @xmath10 25 gev .
coupling compared to limits derived by the cdf @xcite and the l3 @xcite experiments . ]
coupling compared to limits derived by the cdf @xcite and the l3 @xcite experiments . ]
both experiments have also searched for the non standard production of single top quarks at hera i via the flavour changing neutral current ( fcnc ) process , where the analysis strategy is an extension of the isolated lepton analysis @xcite .
some of the h1 events at large @xmath8 show kinematics compatible with the single top hypothosis , but no signal can be claimed .
figure [ fig : singletop ] shows limits derived at 95% confidence level ( cl ) by both experiments on the anomalous fcnc @xmath11 coupling .
it can be seen that hera has the best sensitivity to the @xmath11 coupling in the region where @xmath12 is small .
electons in the di electron ( left ) and tri electron ( right ) h1 analyses compared to the sm expectation , which is dominated by pair production @xcite.,title="fig : " ] electons in the di electron ( left ) and tri electron ( right ) h1 analyses compared to the sm expectation , which is dominated by pair production @xcite.,title="fig : " ] searches for multi
electron production at high transverse momentum have been previously carried out by the h1 @xcite and zeus experiments @xcite , using the hera i data sample .
h1 has additionally studied the production of high @xmath1 muon pairs @xcite . the main sm process for multi lepton production in @xmath0 collisions is photon photon interactions @xmath13 , where quasi real photons radiated from the incoming electron and proton interact to produce a pair of leptons .
high mass events ( @xmath14 100 gev ) are observed in the in the di electron sample of both experiments , and additionally in the h1 tri electron sample , regions where the sm expectation is low .
h1 has recently updated the analysis with the new hera ii data , now exploiting a total luminosity of 209 pb@xmath6 @xcite .
the analysis examines @xmath15 , @xmath16 , @xmath15 , @xmath17 and @xmath18 topologies , searching for events with at least two high @xmath1 electrons or muons .
figure [ fig : multilep ] shows the invariant mass distribution of the h1 di electron ( left ) and tri electron ( right ) samples .
the data are found to be in good overall agreement with the sm , although interesting events are seen at high masses ; in the electron di electron and tri electron channels in the hera i data and more recently in the electron muon channels in the hera ii data . the production in @xmath0 collisions of a doubly charged higgs boson @xmath19 could be a source of events containing multiple high @xmath1 leptons and the observed high mass events in the h1 analysis
have been investigated in this context @xcite .
only one @xmath20 event satisfies the additional selection criteria and the hera limits on the @xmath19 coupling to @xmath20 are not competitive to those set by the opal experiment @xcite .
however , limits derived from the hera data in the @xmath15 decay channel of the @xmath19 extend to higher masses beyond the reach of the previous searches performed by the cdf @xcite and lep @xcite experiments and new constraints were obtained for the @xmath19 coupling to @xmath21 .
a model independent general search for deviations from the sm has been performed by h1 using a hera i data sample corresponding to an integrated luminosity of 117 pb@xmath6 @xcite .
all high @xmath1 final state configurations involving electrons ( @xmath22 ) , muons ( @xmath23 ) , jets ( @xmath24 ) , photons ( @xmath25 ) or neutrinos ( @xmath26 ) are considered .
all final state configurations containing at least two such objects with @xmath27 20 gev in the central region of the detector are investigated and classified into exclusive event classes , @xmath22@xmath24 , @xmath23@xmath24@xmath26 , @xmath24@xmath24@xmath24 and so on .
data events are found in 22 such event classes and good agreement is observed between data and the sm expectation in most event classes , as can be seen in figure [ fig : general ] .
a non biased statistical method is employed to search for deviations of the data with respect to the sm .
a good agreement is found in all channels , except in the @xmath23@xmath24@xmath26 event class , where 4 data events are observed compared to a sm expectation of 0.8 @xmath7 0.2 as previously reported in @xcite .
additionally , in the @xmath22@xmath24@xmath24@xmath24@xmath24 event class 1 event is observed in the data compared to a sm prediction of 0.026 @xmath7 0.011 .
the @xmath0 collsions at hera provide a unique possibility to investigate the formation of a new particle coupling to a lepton
quark pair . in the buchmller , rckl and wyler ( brw ) classification of such states , termed `` leptoquarks '' , 7 scalar and 7 vector particles
are proposed , all of which can couple to an @xmath28 pair and 4 of which can also couple to both @xmath28 and @xmath29 @xcite .
a search for first generation leptoquark production , @xmath30 or @xmath31 has been performed on the hera i data by the h1 @xcite and zeus @xcite collaborations by looking for deviations in the mass spectra of neutral current ( nc ) and charged current ( cc ) dis interactions , the major sm background .
no evidence of leptoquark production is observed , and limits on the 14 leptoquark couplings are derived as function of mass , as shown for example from h1 for @xmath32 scalar leptoquarks in figure [ fig : h1lq ] . in the h1 ( zeus )
analysis , leptoquark masses up to 325 ( 386 ) gev are ruled out at 95% cl for a coupling @xmath33 of electromagnetic strength ( @xmath33 = @xmath34 0.3 ) .
a search for lepton flavour violating ( lfv ) processes mediated by leptoquark exchange has also been performed by h1 @xcite and zeus @xcite , investigating the interactions @xmath35 and @xmath36 .
no evidence for lfv is observed and , similarly to the above analyses , limits are derived for the appropriate leptoquark couplings as a function of mass .
an example is shown in figure [ fig : zeuslfvlq ] from the zeus analysis of the @xmath36 channel using 113 pb@xmath6 of @xmath9 data , where the zeus results improve the existing limits from rare @xmath37 , @xmath38 or @xmath39 decays ( see references given in @xcite ) . as a function of leptoquark mass for the vector leptoquark @xmath40 in the @xmath37 decay channel @xcite . ] as a function of leptoquark mass for the vector leptoquark @xmath40 in the @xmath37 decay channel @xcite . ]
h1 has previously performed a search for squarks ( @xmath41 ) , the scalar supersymmeteric ( susy ) partners of quarks , in models with r parity violation ( ) @xcite
. the search rules out squarks of all flavours with masses up to 275 gev for a coupling of electromagnetic strength in the minimal supersymmetric standard model ( mssm ) framework . a similar analysis recently performed by the zeus collaboration set comparable limits in a search for stop quark ( @xmath42 ) production in mssm susy @xcite . as a function of the stop mass , obtained from the analysis of each bosonic stop decay channel .
right : exclusion limits at the 95% cl in the ( @xmath43 , @xmath44 ) plane for @xmath45 0.3 @xcite.,title="fig : " ] as a function of the stop mass , obtained from the analysis of each bosonic stop decay channel .
right : exclusion limits at the 95% cl in the ( @xmath43 , @xmath44 ) plane for @xmath45 0.3 @xcite.,title="fig : " ] in the afforementioned analyses all squarks are assumed to be degenerate .
a complementary analysis , based on the assumption @xmath46 has been performed by the h1 collaboration using @xmath9 data corresponding to an integrated luminosity of 106 pb@xmath6 . in this case , the stop produced in fusion via the @xmath47 coupling decays bosonically @xmath48 @xcite .
the decay of the sbottom quark @xmath49 and leptonic and hadronic @xmath2 decays are considered . the decay @xmath50 is also examined for completeness .
the bosonic stop decay leads to three different final state topologies depending on the decay of the @xmath2 boson ; a jet , a lepton ( electron or muon ) decay into @xmath51 , where @xmath52 hadrons @xmath53 is not investgated in this analysis . ] and missing transverse momentum ( @xmath54 channel and @xmath55 channel ) or , if the @xmath2 decays to jets , three jets and missing transverse momentum ( @xmath56 channel ) . a slight excess of events compared to the sm prediction is observed in the @xmath55 channel , confirming the results of the dedicated h1 analysis to such a topology @xcite . assuming the presence of a stop mass @xmath43 decaying bosonically , the observed event yields are used to determine the allowed range for a stop production cross section @xmath57 , as illustrated in figure [ fig : bosonicstop ] ( left ) .
it can be seen that the excess in the @xmath55 channel is not supported by the other channels , and hence no evidence for stop production is observed .
the results from the different channels are combined to derive constraints on stop quarks decaying bosonically in the mssm .
the resulting limits projected on the ( @xmath43 , @xmath44 ) plane for @xmath45 0.3 are shown in figure [ fig : bosonicstop ] ( right ) , where for @xmath58 100 gev stop masses up to 275 gev are ruled out at 95% cl . , @xmath59 ) plane from h1 for various values of the coupling @xmath60 ( @xmath61 1,2 ) @xcite . ] , @xmath59 ) plane from h1 for various values of the coupling @xmath60 ( @xmath61 1,2 ) @xcite . ]
if squarks are very heavy , susy phenomena can manifest in @xmath0 collisions by gaugino production via @xmath62channel slepton exchange . a search for gauginos
is performed by the zeus collaboration , where the interaction proceeds via @xmath62channel exchange of a selectron and production of a neutralino @xmath63 @xcite .
the neutralino subsequently cascade decays into two quarks and an electron , positron or neutrino and hence the main sm background arises from di jets in nc and cc dis .
no deviation from the sm is observed , and exclusion limits in the ( @xmath64 , @xmath23 ) plane for gaugino production are derived at 95% cl , as shown in figure [ fig : gaugino ] .
these limits extend previous limits set by the lep experiments @xcite .
in other , so called gauge mediated supersymmetry breaking ( gmsb ) models , the lightest susy particle ( lsp ) is the gravitino @xmath65 and the next to lightest susy particle ( nlsp ) can be the lightest neutralino @xmath63 , which decays to the stable gravitino and a photon .
this specific decay channel of the neutralino is investigated by the h1 @xcite and zeus @xcite collaborations as a search for light gravitino production , that is resonant single neutralino production @xmath63 via @xmath62channel selectron exchange , @xmath66 , where it is assumed the @xmath63 is the nlsp and the subsequent decay @xmath67 occurs with an unobservably small lifetime .
the experimental signature is thus a photon , a jet from the struck quark and missing transverse momentum due to the gravitino and hence the main sm background is due to radiative cc dis .
no significant deviation from the sm is observed by either h1 or zeus and the results are used to derive constraints on gmsb models for different values of the coupling @xmath60 for fixed values of the relevant susy parameters .
figure [ fig : gravitino ] displays excluded regions in the ( @xmath68 , @xmath59 ) plane derived from @xmath9 hera data from h1 for various values of @xmath60 . for small mass differences between the neutralino and selectron , neutralino masses up to 112
gev are ruled out at 95% cl for an coupling @xmath69 1.0 .
the results presented in this section are the first constraints from hera on susy models independent of the squark sector .
many searches for new physics have been performed at hera by the h1 and zeus collaborations , some including data from the new hera ii phase .
interesting events containing isolated leptons and missing @xmath1 and multiple high @xmath1 leptons at high masses have been observed by h1 .
no evidence is seen by either collaboration in searches for a large selection of exotic particles and competitive limits have been derived on such bsm scenarios .
the continued data taking in the hera ii phase by both h1 and zeus experiments will hopefully clarify the high @xmath1 lepton events seen by h1 and further the search for new physics at hera .
h1 collab . , contributed paper to _
eps05 _ , lisbon 2005 , * abstract 637 * ; + updated preliminary results for the _ desy prc _
2005 ; available at : + http://www-h1.desy.de/psfiles/confpap/eps2005/h1prelim-05-164_prc_nov05.ps delphi collab . , j. abdallah _ et .
* b552 * ( 2003 ) 127 ; + l3 collab . , p. achard _ et .
lett . * b576 * ( 2003 ) 18 ; + opal collab . , g. abbiendi _ et .
* b526 * ( 2002 ) 221 , delphi collab .
, j. abdallah _ et .
j. * c36 * ( 2004 ) 1 ; + delphi collab .
, j. abdallah _ et .
al . _ , eur .
j. * c37 * ( 2004 ) 129 ; + aleph collab .
, a. heister _ et .
j. * c31 * ( 2003 ) 1 . | the latest results from the h1 and zeus collaborations are presented on rare standard model processes and searches for physics beyond the standard model .
intriguing events containing high transverse momentum leptons are observed by h1 and many competitive limits are set by both collaborations on the production of new particles in a variety of supersymmetry scenarios . |
gravitational microlensing was proposed by @xcite as a method to determine if the dark halo of the milky way was comprised of objects such as jupiters , brown dwarfs , or white dwarfs in the planetary to stellar mass range .
this proposed program has been carried out by both the macho and eros collaborations with remarkable success .
objects ranging from @xmath1 to @xmath2 are excluding from dominating the mass of the milky way s dark halo @xcite .
this rules out what were formerly the leading astrophysical dark matter candidates , and leaves exotic particle dark matter as the clear leading candidate to comprise the dark matter halo of the milky way .
however , macho collaboration has reported a microlensing optical depth toward the central regions of the large magellanic cloud ( lmc ) that is substantially in excess of the expected microlensing optical depth from known stellar populations @xcite .
this implies that @xmath3% of the milky way s dark halo could be in the form of massive compact halo objects or machos .
the timescales of these microlensing events suggest that these machos could be low mass stars or white dwarfs .
however , the eros collaboration has found a substantially lower microlensing optical depth @xcite with a survey that focused on the outer regions of the lmc . when interpreted in terms of lensing by a halo population ,
the macho and eros results are in contradiction .
a number of ways to resolve this apparent contradiction have been considered .
one possibility is an error by one of the groups . within the macho team ,
concern initially focused on the event detection methods and the detection efficiency methods employed by eros . unlike macho @xcite
, eros did not add simulated stars to images from their survey in order to estimate their detection efficiency , and the eros survey failed to find a number the microlensing events detected by macho .
however , simulated image tests by eros show that any systematic error in their efficiency determination is not nearly large enough to explain the difference in the microlensing optical depth measurements toward the lmc @xcite .
furthermore , eros has also shown that the lmc microlensing events that they did not detect probably are properly accounted for by their detection efficiency calculation
. there has also been some concern that the macho sample might be contaminated by non - microlensing variability .
this was suggested by @xcite , but their methods were criticized by @xcite and @xcite .
however , it was later discovered by eros @xcite that one of the macho microlensing candidates ( lmc-23 ) had a repeat variation .
this could , in principle , be caused by a binary source or lens , but neither bump on the light curve provides a very good fit to a microlensing light curve , so it is most likely that this event , lmc-23 , is a variable star .
the possibility of variable star contamination was considered in some detail by @xcite , who found that improved difference imaging photometry of macho data as well as ctio follow - up data generally confirmed the microlensing interpretation of the macho events .
they also showed that the @xcite classified lmc-23 and some background supernovae as microlensing events , while classifying events with extremely good microlensing light curve fits as likely variables .
@xcite developed a method to correct the macho lmc optical depth measurement for variable star contamination , and showed that this correction only modestly reduced the macho lmc microlensing optical depth measurement . since experimental error by either macho or eros seems unlikely , we are led to consider astrophysical solutions to this apparent discrepancy .
since the lmc microlensing rate is too small to be consistent with a dark halo comprised of only machos , there is no reason to assume that the lens objects trace the density distribution of the dark halo .
however , the distinction between different potential microlensing populations is complicated by the fact that the crucial observable in microlensing , the event duration , typically admits degeneracy in the three fundamental microlensing parameters : the mass , distance and velocity of the lens .
this makes it difficult to distinguish between the two principal geometric arrangements which may explain large magellanic cloud ( lmc ) microlensing : a ) mw - lensing , in which the lens is part of the milky way ( mw ) and b ) self - lensing , in which the lens is part of the lmc . in self - lensing
, the lens may belong to the disk+bar or spheroid of the lmc , while the source star may come from either of these components or some sort of background population to the lmc .
most efforts to distinguish between mw - lensing and self - lensing focus on modeling the lmc self - lensing contribution to the optical depth and comparing this to the observed optical depth .
since the observed optical depth is substantially higher than the modelled self - lensing contribution for standard models of the lmc , we have inferred that there must be a substantial mw - lensing signal @xcite , although lmc self - lensing is certainly a possibility if we allow non - standard lmc models . in this work , we take a different approach in which we discriminate between mw lensing and self - lensing by locating the source stars and comparing their location with predictions for various self - lensing and mw - lensing geometries .
we consider three types of lmc self - lensing : lmc disk+bar self - lensing , lmc spheroid self - lensing and background lensing .
each type of self - lensing has a distinct geometry ( location of source star and lens ) , described in detail in [ lmcmodels ] .
the most important distinction is that the source star populations are different for each self - lensing geometry . in particular , in each self - lensing geometry a different fraction , @xmath4 , of the source stars will lie behind the lmc disk .
the remaining fraction @xmath5 of the source stars lie in the lmc disk+bar . therefore , if we determine the location of the observed microlensing source stars we may compare the observed value of @xmath4 with the prediction for each model of lmc self - lensing . in this way we may eliminate some self - lensing geometries as possible explanations for the lmc microlensing signal .
we emphasize , that as defined in this paper , @xmath4 implies the fraction of source stars which lie behind the entire lmc disk , not behind the mid - plane of the disk .
if all self - lensing geometries which satisfy other external constraints are eliminated and the observed value of @xmath4 remains consistent with its prediction in the mw - lensing geometry , we may conclude that the most likely explanation for the microlensing signal is lensing by massive compact halo objects ( machos ) in the mw .
we determine the fraction of macho source stars drawn from a background population lying behind the lmc disk , @xmath4 , using the suggestion of @xcite , @xcite , and @xcite .
these works point out that source stars from a background population should be preferentially fainter and redder than the average population of the lmc because they will suffer from the extinction of the lmc disk , and are displaced along the line of sight .
@xcite present a model for this background population with an additional mean reddening relative to the average population of the lmc , @xmath6 mag , and a displacement from the lmc of @xmath7 kpc .
the displacement results in an increase of the distance modulus by @xmath8 mag .
we note that , in reality , even disk+bar self - lensing would have a source star distribution which is redder than for mw - lensing . in mw - lensing
, the source stars will be distributed evenly throughout the lmc disk . in the disk+bar self - lensing disk
there will be more source stars on the far side of the lmc disk ( e.g. , behind the mid - plane ) than on the near side of the lmc disk .
the fraction of far - side source stars depends on your model for the lmc disk and the reddening effect depends sensitively on your assumptions about the distribution of dust in the lmc ( e.g. , the dust scale height ) .
this level of extra reddening is a much more subtle effect , and much more model dependent .
we do not address this situation here , instead we only consider the reddening effects of source stars lying behind the entire lmc disk .
we test for possible extra reddening in our source stars relative to the average population of the lmc by comparing a color - magnitude diagram ( cmd ) of the source stars with the cmd of all nearby lmc stars .
the cmds are created using wide field planetary camera 2 ( wfpc2 ) hubble space telescope ( hst ) photometry of the 13 microlensing source stars selected using cut a of @xcite and their surrounding fields .
the source stars are identified by deriving accurate centroids in the ground - based macho images using difference image analysis ( dia ) and then transforming the dia coordinates to the wfpc2 frame .
we compare the source star color magnitude diagram ( cmd ) with a 2-d kolmogorov - smirnov ( ks ) test to model source star cmds in which varying fractions of the source stars are located in and behind the lmc . in [ lmcmodels ]
we describe the four models of lmc microlensing . in [ hstobs ]
we construct a cmd of the average lmc population by combining the cmds of thirteen wfpc2 fields centered on past macho microlensing events in the outer lmc bar . in [ identify ] we describe the identification of the microlensing source stars in these fields by difference image analysis , and find that there are no background galaxies in the vicinity of the source stars . in
[ blend_fits ] we show that the source star brightness implied by the microlensing fits matches the source star observed in the hst images for most events and discuss the reasons why the fit magnitudes occasionally fail to match the observed magnitudes . in
[ modelpops ] we describe the construction of model source star cmds with varying fractions of background source stars , @xmath4 . in [ stats ]
we determine the likelihoods that the observed microlensing source stars were drawn from each of the model source star cmds by using kolmogorov - smirnov ( ks ) tests to compare the observed and model distributions .
we discuss the results of the ks test in
[ ssdiscuss ] , in the context of four models of microlensing : a ) mw halo - lensing , b ) lmc disk+bar self - lensing , c ) lmc spheroid self - lensing , and d ) background lensing .
we also consider other constraints on these microlensing models , such as the microlensing results from the eros project and events which have lens distance determinations .
finally , in [ sssum ] , we summarize our results .
the most straightforward interpretation of the lmc microlensing signal is that the source stars in the disk+bar of the lmc are being lensed by machos in the milky way halo . in this work
we refer to this geometry as mw - lensing .
since all the source stars are located in the lmc disk+bar , the expected value of @xmath4 is @xmath9 .
the other models we consider are all variants on lmc self - lensing , meaning that the lenses themselves also belong to some component of the lmc . early works by @xcite and @xcite considered lensing by the lmc in slightly different contexts .
@xcite considered lensing by the dark halo of the lmc as well as known stellar populations and showed that the lmc dark halo should have a microlensing optical depth that is slightly larger than the optical depth measured by macho .
he also argued that lensing by lmc stars could be important .
@xcite argued that the entire lmc microlensing signal could be explained by ordinary stellar lmc populations , so that no dark matter lensing was needed . in this work
, we refer to this geometry as lmc disk+bar self - lensing .
both these early works suggested that lmc disk+bar self - lensing could account for a substantial fraction of the observed optical depth .
this claim has since been disputed by several other groups @xcite who show that when considering only disk stars the lmc self - lensing optical depth is far too low to account for the observed optical depth .
@xcite show that allowing for contributions to the lens and source populations from the lmc bar does not substantially increase the lmc self - lensing optical depth .
@xcite find a low lmc self - lensing optical depth for a flared lmc disk .
@xcite also show that much of the disagreement in models of the lmc disk+bar self - lensing optical depth results from disagreement about the fundamental parameters of the lmc , such as the total disk mass and inclination angle . within their region of allowed parameters
@xcite also make a strong case that lmc disk+bar self - lensing makes a small contribution to the observed optical depth , at most @xmath1025% ( using the macho result , @xmath11 , corrected for variable star contamination @xcite ) .
@xcite show that the spacial distribution of microlensing events indicates that a large fraction of them are not caused by the lmc disk+bar . in the lmc disk+bar self - lensing model ,
the source stars will be preferentially drawn from the far side of the lmc disk ( e.g. , beyond the midplane of the disk ) , but are still part of the lmc disk and are not drawn from a distinct entity behind the entire lmc disk .
therefore , this model also gives @xmath12 .
although the upper limit on the optical depth expected from lmc disk+bar self - lensing is strongly constrained to be substantially below the observed optical depth , the self - lensing optical depth may be raised much higher if one allows for lenses in an lmc stellar halo population as first shown by @xcite .
the principal problem surrounding an lmc stellar halo contribution is that tracers of old populations in the lmc seem to indicate that this population has too little mass to contribute more than @xmath13% of the optical depth measured by macho .
however , another possibility has arisen from the results of @xcite who claims that lmc microlensing may be caused by a _
non - virialized _ stellar halo or `` shroud '' . in this geometry
the lens may lie in the lmc disk+bar or the foreground part of the shroud , while the source stars lie in the lmc disk+bar or the background part of the shroud .
the term shroud was introduced by @xcite and is meant to imply an lmc population which is like a halo in that it is spatially not part of the lmc disk , but unlike a halo in that it is non - virialized and thus may have a relatively low velocity dispersion . such a population is suggested by the simulations of @xcite , who finds that the lmc s dynamical interaction with the mw may torque the lmc disk in such a way that the lmc disk is thickened and a spheroid component is populated without isotropizing the stellar orbits and thereby leaving disklike kinematics intact . however , observational work on rr lyrae by @xcite , which does not rely on the specific kinematics of the spheroid , limits the total mass of any type of halo ( virialized or non - virialized ) to perhaps 5% of the mass of the lmc , too small to contribute more than @xmath105% of the observed optical depth .
a more recent revisiting of this argument by @xcite finds room to increase this optical depth contribution to at most 20% , still only a small fraction of the total .
the @xcite models have been challenged on various grounds by @xcite . in order for an lmc shroud to account for the total optical depth it must have a mass comparable to that of the lmc disk+bar @xcite .
even if we accept the existence of such a massive shroud , the microlensing implications are somewhat in dispute .
@xcite finds an lmc self - lensing optical depth comparable to the observed optical depth .
however , this estimate is reduced by a factor of 3 by @xcite who repeat the @xcite microlensing analysis using lower values for the disk total mass and inclination angle , and a proper weighting over all observed macho fields . in this work
we refer to both lmc halo and shroud self - lensing as lmc spheroid self - lensing . in spheroid
self - lensing there are four event geometries : a ) background spheroid source and disk lens , b ) disk source and foreground spheroid lens , c ) disk source and disk lens and , d ) background spheroid source and foreground spheroid lens .
we might naively expect the former two types dominate the number of expected events , and so if we were to ignore the contribution from the latter two types we would conclude that spheroid lensing would imply @xmath14
. however , in order to produce the entire observed optical depth , the spheroid must be so massive that it is no longer self - consistent to ignore the final term .
calculations performed in the formalism of @xcite suggest instead that events with a background spheroid source and a foreground spheroid lens become an important contributor and increase the expected fraction of background source stars to @xmath15 .
however , this estimate assumes that the spheroid has the same stellar population as the disk+bar .
if the spheroid population is much older , it could lack the main sequence a and f stars that serve as sources for 75% of the lmc microlensing events . in this case
, a much smaller value of @xmath4 would be appropriate .
yet another self - lensing geometry was introduced by @xcite who suggests that the observed events are due to `` background '' self - lensing in which the source stars are located in some background population , displaced at some distance behind the lmc . a veritable plethora of lenses for this population is then supplied by the disk+bar of the lmc .
a background population has the advantage of being nearly impossible to confirm or reject observationally , as there are nearly no limits on its size or content ( provided of course , it is small enough to `` hide '' behind the lmc ) . since all the source stars belong to the background population , @xmath16 .
each of the above lmc microlensing geometries is depicted pictorially in figure [ modelpictures ] .
a final possibility is foreground self - lensing .
this is not self - lensing in the classical sense as in this case the lenses are not drawn from the lmc itself , but rather from some kinematically distinct foreground population , such as an intervening dwarf galaxy .
@xcite claim a detection of a population of stars from such an entity .
however , @xcite claim that this population is a morphological feature of the lmc red clump , while others show that such a population consistent with other observational constraints could not produce a substantial microlensing signal @xcite .
@xcite argue that may be possible to evade these constraints by tuning the imf of this putative foreground population .
but the _ a priori _ probability for such a large mass of stars to exist in the foreground of the lmc is already quite low , so this explanation seems rather unlikely .
we do not consider this model any further .
observations were made under proposals go-05901 , go-7306 and go-8676 ( principal investigator : kem cook ) with the wfpc2 on the hst between may 1997 and july 2000 through the f555w ( @xmath17 ) and f814w ( @xmath18 ) filters .
the planetary camera ( pc ) was centered on the location of past macho microlensing events .
the microlensing events , positions , and exposure times are listed in table 1 .
multiple exposures of a field were combined using a sigma - clipping algorithm to remove deviant pixels , usually cosmic rays .
the pc has a pixel size of 0046 which easily resolves the great majority of stars in our frames .
most stars are also resolved in the wide field ( wf ) fields which have a pixel size of 01 .
instrumental magnitudes were calculated from aperture photometry using daophot ii @xcite with a radius of 025 and centroids derived from point - spread function ( psf ) fitting photometry .
aperture corrections to 05 were performed individually for each frame .
we choose this aperture radius as this is how the @xcite calibrations are defined .
we correct for the wfpc2 charge transfer effect using the equations from instrument science report wfpc2 97 - 08 .
we also make the minimal corrections for contaminants which adhere to the cold ccd window according to table 28.2 of the hst data handbook .
we transform our instrumental magnitudes to landolt @xmath17 and @xmath18 using the calibrations from @xcite .
we create a composite lmc cmd by combining the pc and wf photometry for all of our fields except the field of lmc-1 . in the case of lmc-1
the @xmath17 and @xmath18 observations were taken at different roll angles and there is little area of overlap except in the pc frame .
we therefore include the pc field from lmc-1 but not the wf fields .
the composite hst cmd is shown in figure [ cmd ] .
a ground - based macho image has a pixel size of 06 and a seeing of at least 14 . in a typically crowded region of the outer lmc bar ,
a macho seeing disk will contain @xmath1011 stars of @xmath19 .
this means that faint `` stars '' in ground - based macho photometry are usually not single stars at all , but rather blended composite objects made up of several fainter stars . henceforth , we distinguish between these two words carefully , using _
object _ to denote a collection of stars blended into one seeing disk , and _
star _ to denote a single star , resolved in an hst image or through difference image analysis .
the characteristics of the macho object that was lensed tell us little about the actual lensed star .
however , with the microlensing object centroid from the macho images we can hope to identify the microlensing source star in the corresponding hst frame .
a direct coordinate transformation from the macho frame to the hst frame often places the baseline macho object centroid in the middle of a group of faint hst stars with no single star clearly identified . to resolve this ambiguity
we have used dia .
this technique is described in detail by @xcite and @xcite , but we review the main points here . dia is an image subtraction technique designed to provide accurate photometry and centroids of variable stars in crowded fields .
the basic idea is to subtract a flux and seeing matched high s / n reference image from each from each program image , leaving a differenced image containing only the variable components .
applied to microlensing , we subtract baseline images from images taken at the peak of the microlensing light curve , leaving a differenced image containing only the flux from the microlensing source star and not the rest of the object .
we typically observe a centroid shift between the macho object s location and the residual lensed flux in the difference image .
if the centroid from the differenced image is transformed to the hst frame we find that it usually clearly identifies the hst microlensed source star .
this process is illustrated in figures [ macho ] and [ lmc4 ] . in figures
[ dia1 ] and [ dia2 ] we present wfpc2 finding charts identifying the macho object centroid , the centroid in the dia image and the centroid of the star in the wfpc2 image closest to the dia centroid . this technique allows us to unambiguously identify the 13 microlensed source stars which pass the cut a microlensing tests in @xcite . in table 2
we present the @xmath17 , @xmath20 , and @xmath18 magnitudes of our source stars from the hst data .
the errors presented here include the formal photon counting errors returned by daophot ii as well as an additional @xmath21 mag uncertainty due to aperture corrections .
we note that the cut a microlensing events do not include lmc-9 ( which appeared in a similar analysis by @xcite ) .
although lmc-9 is certainly a valid microlensing event , it is a binary microlensing event @xcite and therefore its non - standard light curve does not pass the stringent cut a requirements . although we list its photometric information in table 2 , it is not used in our statistical analysis .
our identification of lmc-5 revealed it to be the rather rare case of a somewhat blended hst star @xcite . although there are two stars evident , at an aperture of 025 the flux of one star was contaminated by that of its neighbor . therefore , in this case , we perform psf fitting photometry using psfs kindly provided by peter stetson .
the errors presented in table 2 for lmc-5 are those returned by the profile fitting routine allstar .
the dia centroid falls 2 pixels closer to the centroid of star one than star two , clearly preferring star one as the source star .
furthermore , as predicted by @xcite and @xcite , star two is a rather red object which is very faint in the v band .
fits to the macho light curve presented by @xcite suggest lensed flux fractions in the v and r bands of 1.00 and 0.46 respectively , confirming the dia choice of the much bluer star as the lensed source star .
in addition to these hst / wfpc2 observations @xcite , a second epoch of hst observations using the advanced camera for surveys ( acs ) were obtained @xcite , and these data combined with a microlensing parallax analysis @xcite resulted in the first stellar mass measurement of a single star besides the sun @xcite .
the lmc-5 lens star has a mass of @xmath22 .
an important source of background for microlensing events in these crowded fields is due to distant galaxies behind the lmc .
this was first noticed in @xcite , where event lmc-11 was seen to be very close to an obvious background spiral galaxy .
this and the asymmetric light curve indicated that the variability was almost certainly due to a background supernova and not microlensing .
background supernovae were considered in some detail in @xcite , where a detailed supernova rejection procedure was developed .
events were rejected if 1 .
a supernova type ia model provided a better fit than microlensing , or 2 .
a background galaxy was visible in the proximity of the event .
this procedure is expected to work because the type ii supernovae with light curves that do nt resemble type ia light curves are generally associated with galaxies that are closer and have more young stars , so that they should be visible in the macho images .
this procedure seems to remove all the supernovae contamination from the macho sample , but there are a few events , such as lmc-10 and 22 , for which host galaxy is so compact that it looks like a star in the macho images .
thus , an important check of our background rejection procedure is that all of the events should not be associated with background galaxies .
a careful inspection of the hst images ( see figs . [ dia1 ] and [ dia2 ] ) reveals no extended sources in the vicinity of the lens events , so the efficiency of our supernovae rejection method is confirmed .
as discussed above , in a ground - based macho image each microlensed `` star '' is often a blended composite image of several fainter stars . only one of the stars which make up the composite object
is microlensed . in this section ,
we have described one method for identifying the microlensed star in the wfpc2 image .
however , there exists another method for estimating the percentage of the total object flux which was actually microlensed .
the ground - based object flux may separated into lensed and unlensed components by performing a microlensing blend fit to the ground - based lightcurve .
this procedure is described in detail by @xcite and @xcite . in table 2 , we calibrate the lensed flux for each event as presented in table 5 of @xcite . the calibrated lensed flux is an estimate of the source star ( not object ) magnitude , derived entirely from the ground - based photometry .
thus , we would expect @xmath23 and @xmath24 .
we show a plot of the difference @xmath25 in figure [ blend ] .
we plot the difference versus the event duration , @xmath26 and the reduced chi - squared of the microlensing blend fit ( see table 5 of @xcite ) .
the blend fits work reasonably well for all events except lmc-7 , lmc-13 , lmc-21 and lmc-23 .
the case of lmc-13 is considered by @xcite , who show that the blend fit is consistent with the source identified in the hst frames when follow - up ctio photometry and improved dia photometry of the macho survey data is used in the fits . for lmc-21
, the blend fits indicate a source that is @xmath27 mag fainter than the corresponding source in the hst frames , but the @xmath28 colors of the blend fit sources and the hst image are equal to within their errors .
thus , this event might be explained if the source has a binary companion of almost equal brightness at a separation larger than a few au . for lmc-7 , the blend fit source also has a color that is similar to the color seen in the hst images , but in this case the blend fit source is brighter .
so , this can not be explained by a source companion .
the light curve is symmetric and achromatic , with a peak magnification of 6.9 , but it also has significant deviations from the standard , single lens microlensing light curve .
this could be caused by a non - standard microlensing effect , such as a weak binary lens , microlensing parallax @xcite or xallarap @xcite , which is the result of orbital motion of the source ( i.e.parallax reversed ) .
the hst star for this event does appear to be slightly redder than the blend fit source .
if this does not reflect photometry problems ( as was the case with lmc-13 ) , it might indicate a very red foreground milky way disk lens as was the case for events lmc-5 @xcite and lmc-20 @xcite .
( lmc-20 is a lower s / n microlensing event that did not pass all the sample a cuts . )
if this excess red flux is really from the source , then there is a very good chance that the light curve will show a parallax deviation .
lmc-23 is the one source that shows a large color offset between the blend fits and the hst photometry , with the hst blend fits indicating a source about 0.3 mag bluer than the hst frame .
this event also has significant light curve deviations that can probably not be explained by a reasonable microlensing model @xcite .
in fact , data from the eros @xcite and ogle ( udalski , 2004 , private communication ) groups show an additional brightening that occurred after the macho survey stopped taking data .
it is possible , but unlikely , for binary source or binary lens model to produce two light curve bumps separated by a year or more @xcite , but it is virtually impossible to also ascribe the poor microlensing fit to non - standard microlensing .
the only reasonable conclusion is that this event is not microlensing , but is actually due to a variable star .
however , this is the only macho lmc event that has been confirmed to be something besides a microlensing event , and a careful analysis of all the macho lmc events indicates that contamination of the macho sample by non - microlensing variability leads to only a small correction to the microlensing optical depth @xcite .
we consider source stars drawn from two distinct populations : the disk+bar of the lmc and a background population located behind the lmc disk .
source stars from the disk+bar of the lmc are well represented by our composite hst cmd shown in figure [ cmd ] .
source stars drawn from the background population will be on average redder than the disk+bar by an amount equal to the mean extinction through the lmc and may also be displaced behind the lmc by some distance .
the mean internal extinction of the lmc has been determined using ubv and ubvi photometry of young , hot ( o a type ) stars in @xcite and @xcite .
@xcite find a mean internal reddening due to dust in the lmc disk of @xmath29 and @xcite find @xmath30 .
a more recent study @xcite suggests that the reddening of the lmc may be population dependent .
@xcite find @xmath31 for hot , young stars ( @xmath32 k ) , but a much lower internal reddening for older , cooler stars ( @xmath33 ) of @xmath34 .
the lower reddening for older stars may be due to the difference in scale heights of an old population and a young population relative to the scale height of the dust ( i.e. , the old population may have a larger scale height ) combined with the presence of low - extinction `` holes '' in the dust layer . in this interpretation
, the large scale height of old stars implies that about half the cool stars live in front of the lmc dust plane and are only extincted by galactic foreground dust .
in contrast , the small scale height of the young stars implies that the majority live within the dust plane and suffer from both foreground and internal extinction .
of the stars lying behind the lmc midplane , some fraction are found in `` holes '' and are again only extincted by the galactic foreground dust . according to @xcite
such a model works well in predicting both the hot star and cold star extinctions in the lmc .
if the above interpretation of population dependent reddening is correct , then the mean reddening for the young hot stars is an appropriate measure of the total internal reddening of the lmc disk .
however , the concept of a patchy lmc extinction is problematic for our kolmogorov - smirnov test in which we assume that the reddening of all of our fields is approximately the same .
empirically , we find that the reddening of our fields is approximately the same .
if we shift our cmds up and down the reddening axis until they seem to line up perfectly ( by eye ) , we find that the maximum shift is within @xmath35 of the mean . the high reddenings for young , hot stars may also be due to environmental effects , e.g. , the young stars may still live very close to the dust in which they were formed . in this case
, the young star reddening may significantly overestimate the true mean internal reddening of the lmc disk . in this work
, we adopt the same value as used in @xcite , @xmath30 .
this is the lowest mean value determined using young stars , but is significantly higher than the @xcite value for old stars .
we note , that if the true value is significantly lower than @xmath6 , the statistical significance of our kolmogorov - smirnov test would decrease substantially .
the distance to the background population of @xmath36 from @xcite is very loosely derived by the requirement that the background population be at least transiently gravitationally bound to the lmc .
we consider three different displacement distances from @xmath37 , @xmath38 ( @xmath39 kpc ) and @xmath40 ( @xmath41 kpc ) .
we have no constraints on the location , size or content of a possible background population , except that it must be small enough and similar enough to the lmc stellar population to have avoided direct detection . for each model source star population
we begin with the composite hst cmd and then shift some fraction , @xmath4 , of the stars into the background .
that is , we redden and displace a fraction @xmath4 of the cmd by amounts @xmath42 @xmath43 the coefficient in the conversion from @xmath44 to @xmath45 is drawn from table 6 of @xcite and @xmath46 mag .
we use the landolt coefficients from @xcite since the @xcite coefficients calibrate the wfpc2 magnitudes to this system .
each cmd now contains a fraction @xmath47 source stars from the lmc disk+bar and a fraction @xmath4 of source stars from the background population .
the cmd now represents the distribution of source stars down to @xmath48 .
however , not all possible microlensing events are detected in the macho images .
to create a cmd representing a population of source stars which produce detectable microlensing events we must convolve the hst cmd with the macho detection efficiency .
the macho efficiency pipeline is extensively described by @xcite and @xcite and the detection efficiency as a function of stellar magnitude , @xmath49 , and einstein ring crossing time has been calculated .
we average this function over the event durations of the thirteen events from @xcite and present the macho detection efficiency as a function of @xmath49 in figure [ eff ] .
we convolve this function with our hst cmds to produce the final model source star cmds . in figure
[ model ] , we illustrate two model source star populations , both with @xmath50 , one with @xmath51 ( all source stars in the lmc disk+bar ) and one with @xmath52 and @xmath50 ( all source stars in a background population displaced by 7.5 kpc ) . we overplot the observed microlensing source stars of table 2 as large red stars .
this procedure is based on several assumptions .
first , we assume that our thirteen hst fields collectively well represent the stellar population of the lmc disk .
this assumption has two parts , the first being that an observation at a random line of sight in the lmc bar is dominated by stars in the lmc disk and the second that the stellar population across the lmc is fairly constant .
the first part holds so long as the surface density of the background population is much smaller than that of the lmc itself .
if this were not the case , this population would have been directly detected .
the second part has been confirmed by many lmc population studies including @xcite , @xcite , and @xcite , as well as our own comparison of individual cmds and luminosity functions .
in addition , as mentioned above , we find that all of our fields have reddenings within about @xmath35 of their mean value .
we now determine which model source star population is most consistent with the observed cmd of microlensed source stars by performing a two - dimensional kolmogorov - smirnov test . in the familiar one dimensional case ,
a ks test of two samples with number of points @xmath53 and @xmath54 returns a distance statistic @xmath55 , defined to be the maximum distance between the cumulative probability functions at any ordinate . associated with @xmath55 is a corresponding probability @xmath56 that _ if _ two random samples of size @xmath53 and @xmath54 are _ drawn from the same distribution _ a worse value of @xmath55 will result .
this is equivalent to saying that we can exclude the hypothesis that the two samples are drawn from the same distribution at a confidence level of @xmath57 . if @xmath58 , then this is also equivalent to excluding at a @xmath57 confidence level the hypothesis that sample 1 is drawn from sample 2 .
the concept of a cumulative distribution is not defined in more than one dimension .
however , it has been shown that a good substitute in two dimensions is the integrated probability in each of four right - angled quadrants surrounding a given point ( fasano & franceschini 1987 ; peacock 1983 ) .
that is , consider a point @xmath59 in distribution @xmath60 and draw two perpendicular lines through this point ( one line perpendicular to the @xmath61 axis and the other perpendicular to the @xmath62 axis ) which divide the space into four quadrants .
the integrated probability of each quadrant for each distribution is the fraction of data from the distribution which lies in that quadrant .
the two dimensional ks statistic @xmath55 is now taken to be the maximum difference ( ranging both over all data points and all quadrants ) between the integrated probability of distribution @xmath60 and @xmath63 . see press et al.(1992 ) for further details .
the distance statistic @xmath55 and the corresponding @xmath56 are subject to the same interpretation as in the one dimensional case .
we show the resulting @xmath56 plotted versus @xmath4 in figure [ results ] .
the results are shown separately for @xmath64 and @xmath65 in red circles , green triangles , and blue squares , respectively .
the error bars indicate the scatter about the mean value for 20 simulations of each model .
because the creation of the efficiency convolved cmd is a weighted random draw from the hst cmd , the model population created in each simulation differs slightly .
this in turn leads to small differences in the ks statistics .
we find that the 2-d ks - test probability is highest at a fraction of source stars behind the lmc @xmath66 with very little dependence on the value of the displacement , @xmath67 .
the significant preference for low @xmath4 arises mostly from reddening / color . in a strict statistical sense
, the ks test can only be used to reject or confirm the hypothesis that two datasets have a common parent distribution . as a general guideline ,
a probability of less than 10% suggests that the two distributions are significantly different ( press et al.1992 ) .
it is not strictly allowable to compare relative probability , e.g. , a ks test probability of 80% does not mean that it is twice as likely that two distributions are identical than if they had returned a ks test probability of 40% .
however , we still find it valuable to show the shape of our ks test probability function to demonstrate its insensitivity to @xmath67 .
[ 2.6 ] we now compare the observational results for @xmath4 ( figure 10 ) with the expected value for our four models of microlensing : a ) background lensing ( @xmath68 ) , b ) lmc spheroid self - lensing ( @xmath69 ) , c ) lmc disk+bar self - lensing ( @xmath70 ) , and d ) mw lensing ( @xmath70 ) .
we rule out a model in which the source stars all belong to some background population at a confidence level of @xmath71% ( e.g. , the ks test probability for @xmath52 is @xmath72 ) .
of course , one event of this sample , lmc-5 , is known to be caused by a lens in the milky way disk , but @xmath73 and @xmath74 are excluded by @xmath75% and @xmath76% , respectively .
so , background source models would seem to be excluded .
but , this conclusion might be weakened if we had allowed for patchy extinction within the lmc .
we can rule out spheroid self - lensing models where the spheroid has the same luminosity function as the disk+bar ( @xmath77 ) at the statistically marginal confidence level of @xmath78% .
but old spheroid self - lensing models without a and f stars in the spheroid have @xmath79 and are not significantly disfavored by this analysis .
the relatively low number density of lmc halo rr lyrae @xcite suggests that the lmc spheroid microlensing optical depth should be quite low , but there are several reasons not to dismiss the lmc spheroid self - lensing models .
first , as @xcite emphasized , the lmc is not an isolated galaxy , and the interactions of the lmc with the milky way and the small magellanic could ( smc ) significantly complicate attempts to properly model its structure .
thus , the lack of an lmc model that provides a sufficient microlensing optical depth while fitting the other observational constraints may just reflect our failure to model the mw + lmc + smc system .
the allowed region of the @xmath4 plot ( @xmath80% ) is consistent with the expected location of the source stars in both the mw - lensing and lmc disk+bar self - lensing geometries .
however , detailed modelling of lmc disk+bar self - lensing suggests that it contributes only 25% of the observed optical depth @xcite .
so , the test presented here seems most consistent with having the majority of events located in the mw halo , with some contribution from the lmc disk+bar and mw disk
. there are , however , additional reasons that favor the lmc spheroid self - lensing model . the strongest point in favor of lmc
spheroid self - lensing is the apparent spacial distribution of the lmc microlensing optical depth .
while the spacial distribution of the macho events does not match expectations from lmc disk+bar self - lensing @xcite , lmc spheroid self - lensing should produce a much more uniform distribution of events across the central regions of the lmc where the macho survey had high sensitivity ( within @xmath81 of the lmc - bar center ) .
however , the eros survey has very low sensitivity to microlensing in this central part of the lmc , as most of their sensitivity comes from fields that are @xmath82 from the center of the lmc - bar . since lmc spheroid self - lensing should provide a much lower optical depth in these outer fields , it would seem that lmc spheroid self - lensing might provide a reasonable explanation for both the macho and eros results .
some of the handful of magellanic cloud microlensing events with distance estimates also provide some clues to the locations of the other lenses .
lmc-5 and lmc-20 ( from the b sample of @xcite ) have both been confirmed to be mw disk lenses @xcite , but they have been detected by methods that can only detect lenses that are close to us , so they do nt really inform us about the location of other events .
the situation is similar for the binary source event , lmc-14 , which has been shown to be caused by a lens in the lmc disk+bar or spheroid @xcite .
the binary source effects that lead to this conclusion are only detectable for lenses that reside in the lmc , some self - lensing events are expected for any reasonable model of the lmc . on the other hand
, only a minority of self - lensing events should show this effect , so one might argue that such an event should be accompanied by a number of other self - lensing events that do nt show this effect . but this argument does nt have much statistical significance since it is based on only a single event .
it is more informative to consider events that can reveal distance information no matter where the lens resides .
there are two types of such events : binary lens events with resolved caustic crossings , and events with space - based microlensing parallax measurements .
event lmc-9 is a binary lens event with a caustic crossing that is marginally resolved by only two measurements @xcite .
the long caustic crossing time favors a lmc disk+bar lens , but due to the poor sampling of the caustic crossing alternative interpretations are possible .
event macho-98-smc-1 was announced by the macho alert system @xcite prior to the second caustic crossing , which enabled the eros , mps , ogle and planet collaborations to join macho in monitoring the event .
this allowed the second caustic crossing to be well resolved , and the observed caustic crossing timescale indicated that the lens resided in the smc @xcite .
however , the smc is known to be much more extended along the line - of - sight than the lmc , so its self - lensing optical depth is expected to be much larger .
so , the fact that this lens is likely to be in the smc does not imply that most lmc sources have lenses that are also in the lmc .
the situation with the only event ( ogle-2005-smc-1 ) with a space - based microlensing parallax measurement @xcite is somewhat more interesting because the analysis favors a mw halo lens over a lens in the smc by a 20:1 likelihood ratio .
however , if the lens is in the mw halo , then it is almost certainly more massive than the lenses for the lmc microlensing events found by macho .
so , it does nt really support the mw halo model for the lmc lenses , either .
however , it does seem plausible that this event could be explained by a smc - spheroid lensing population that is larger than predicted by standard smc models .
we have presented hst follow - up observations of 13 lmc microlensing events presented in @xcite including the 12 genuine microlensing events passing the cuts for for selection criteria a. these hst data support our conclusions , based on light curve modeling , that these are genuine microlensing events , and they also strengthen the conclusion that the macho lmc microlensing sample does not have significant contamination due to non - microlensing variability .
we have also compared the colors and magnitudes of our source stars to the predictions of various models to determine which models of lmc microlensing might be consistent with the observed events . at present ,
the strength of this analysis is severely limited by the number of microlensing events .
we are currently able to exclude the most extreme model ( @xmath68 ) , but with more events , it should be possible to more accurately determine @xmath4 and judge intermediate models such as lmc spheroid or shroud self - lensing . ongoing microlensing search projects ( ogle - iii , moa - ii , supermacho ) may supply a sufficient sample of events in the next few years .
the technique outlined in this paper should prove a powerful method for locating the sources , and , subsequently , the lenses with these future datasets . at present , there exist no theoretical models that adequately explain the microlensing optical depth seen towards the lmc .
lmc spheroid or shroud models seem most likely to fit the spacial distribution of microlensing events seen in the combined macho plus eros data sets . the work of can , ajd , khc , pp .
mg , sck , and slm work was performed under the auspices of the u.s .
department of energy by lawrence livermore national laboratory in part under contract w-7405-eng-48 and in part under contract de - ac52 - 07na27344 .
d.p.b . was supported by grant ast-0708890 from the nsf .
is supported by the basal cata , by fondap center for astrophysics 15010003 , and by the milky way millennium center .
afonso , c. , et al .
2000 , , 532 , 340 alcock , c. , et al .
1995 , , 454 , l125 alcock , c. , et al .
1996 , , 463 , l67 alcock , c. , et al .
1997 , , 486 , 697 alcock , c. et al.1998 , , 499 , l9 alcock , c. , et al . 1999a , , 518 , 44 alcock , c. , et al .
1999b , , 521 , 602 alcock , c. , et al .
2000a , , 541 , 270 alcock , c. , et al . 2000b , , 542 , 281 alcock , c. , et al .
2001a , , 550 , l169 alcock , c. , et al . 2001b , , 552 , 259 alcock , c. , et al .
2001c , , 552 , 582 .
alcock , c. , et al . 2001d , , 136 , 439 alcock , c. , et al .
2001e , , 414 , 617 alves , d.r .
2000 , in `` galaxy disks and disk galaxies , '' eds . s.funes & e. corsini , 537 alves , d.r . & nelson , c.a .
2000 , , 542 , 789 beaulieu , j. , & sackett , p. 1998
, , 116 , 209 belokurov , v. , evans , n. w. , & le du , y. 2004 , , 352 , 233 bennett , d.p .
1998 , , 493 , l79 bennett , d. p. 2005
, , 633 , 906 bennett , d. p. , becker , a. c. , & tomaney , a. 2005 , , 631 , 301 bennett , d. p. , et al .
1996 , nuc .
, 51 , 152 calchi novati , s. , de luca , f. , jetzer , p. , & scarpetta , g. 2006 , , 459 , 407 dong , s. , et al . 2007 , , 664 , 862 drake , a. j. , cook , k. h. , & keller , s. c. 2004 , , 607 , l29 evans , n.w .
& kerins , e. 2000 , , 529 , 917 fasano , g. , & franceschini , a. 1987 , , 225 , 155 gates , e. i. , & gyuk , g. 2001 , , 547 , 786 geha , m. , et al . 1998 , , 115 , 1045 gould , a. 1992 , , 392 , 442 gould , a. 1995 , , 441 , 77 gould , a. 1998 , , 499 , 728 gould , a. 2004 , , 606 , 319 gould , a. , bennett , d. p. , & alves , d. r. 2004 , , 614 , 404 gould , a. , bahcall , j.n . , & flynn , c. 1997 , , 482 , 913 graff , d.s . ,
gould , a.p . ,
suntzeff , n.b . ,
schommer , r.a . , & hardy , e. 2000 , , 540 , 211 griest , k. , & hu , w. 1992 , , 397 , 362 griest , k. , & thomas , c. l. 2005 , , 359 , 464 gyuk , g. , dalal , n. , & griest , k. 2000 , , 535 , 90 harris , j. , zaritsky , d. , & thompson , i. 1997 , , 114 , 1933 holtzman , j.a . , burrows , c.j . ,
casertano , s. , hester , j.j . ,
trauger , j.t . , watson , a.m. , & worthey , g. 1995 , , 107 , 1065 oestricher , m.o . , & schmidt - kaler , t. 1996 , a&as , 117 , 303 kallivayalil , n. , patten , b. m. , marengo , m. , alcock , c. , werner , m. w. , & fazio , g. g. 2006 , , 652 , l97 kinman , t.d . , et al .
1991 , , 103 , 1279 kuijken , k. & garcia - ruiz , i. 2001 , in `` galaxy disks and disk galaxies , '' eds . s. funes & e. corsini .
, 401 lasserre , t. , et al .
2000 , , 355 , l39 mancini , l. , calchi novati , s. , jetzer , p. , & scarpetta , g. 2004 , , 427 , 61 minniti , d. , borissova , j. , rejkuba , m. , alves , d. r. , cook , k. h. , & freeman , k. c. 2003 , science , 301 , 1508 olsen , k.a.g . , 1999 , , 117 , 2244 paczyski , b. 1986 , , 304 , 1 peacock , j.a . 1983 , , 202 , 615 poindexter , s. , afonso , c. , bennett , d. p. , glicenstein , j .- f . , gould , a. , szymaski , m. k. , & udalski , a. 2005 , , 633 , 914 press , w.h . ,
teukolsky , s.a . , vetterling , w.t . , &
flannery , b.p .
1992 , numerical recipes in c , second edition ( cambridge : cambridge univ . press ) sahu , k.s .
1994 , , 370 , 275 schlegel , d.j . ,
finkbeiner , d.p . , &
davis , m. 1998 , , 500 , 525 stetson , p.b .
1987 , , 99 , 191 stetson , p.b .
1991 , data analysis workshop iii ( garching : eso ) , 187 tisserand , p. , et al .
2007 , , 469 , 387 tomaney , a.b . , &
crotts , a.p.s .
1996 , , 112 , 2872 weinberg , m. 2000 , , 532 , 922 wu , x. 1994 , , 435 , 66 zaritsky , d. 1999 , , 118 , 2824 zaritsky , d. , & lin , d.n.c .
1997 , , 114 , 2546 zaritsky , d. , shectman , s. a. , thompson , i. , harris , j. , & lin , d. n. c. 1999 , , 117 , 2268 zhao , h. 1999 , , 527 , 167 zhao , h. 2000 , , 530 , 299 zhao , h. , graff , d.s . , & guhathakurta , p. 2000 , , 532 , l37 lllcccl lmc-1 & 05:14:44.53 & @xmath8368:47:56.31 & @xmath84 & @xmath85 & @xmath86 & 1997/12/16 + lmc-4 & 05:17:14.59 & @xmath8370:46:55.28 & @xmath87 & @xmath88 & @xmath89 & 1997/12/12 + lmc-5 & 05:16:41.22 & @xmath8370:29:21.22 & @xmath84 & @xmath90 & @xmath91 & 1999/05/13 + lmc-6 & 05:26:13.77 & @xmath8370:21:16.37 & @xmath84 & @xmath90 & @xmath91 & 1999/08/26 + lmc-7 & 05:04:03.61 & @xmath8369:33:21.00 & @xmath84 & @xmath90 & @xmath91 & 1999/04/12 + lmc-8 & 05:25:09.65 & @xmath8369:47:53.82 & @xmath84 & @xmath90 & @xmath91 & 1999/03/12 + lmc-9 & 05:20:20.52 & @xmath8369:15:13.84 & @xmath84 & @xmath90 & @xmath91 & 1999/04/13 + lmc-13 & 05:24:02.83 & @xmath8368:49:14.77 & @xmath87 & @xmath88 & @xmath87 & 2000/07/28 + lmc-14 & 05:34:44.54 & @xmath8370:25:10.09 & @xmath92 & @xmath90 & @xmath93 & 1997/05/13 + lmc-15 & 05:05:45.86 & @xmath8369:43:54.02 & @xmath87 & @xmath88 & @xmath87 & 2000/07/17 + lmc-18 & 05:45:20.87 & @xmath8371:09:15.20 & @xmath87 & @xmath88 & @xmath87 & 2000/07/21 + lmc-20 & 04:54:18.81 & @xmath8370:02:21.39 & @xmath87 & @xmath88 & @xmath87 & 2000/07/29 + lmc-21 & 04:57:13.82 & @xmath8369:27:50.57 & @xmath87 & @xmath88 & @xmath87 & 2000/07/26 + lmc-23 & 05:06:16.85 & @xmath8370:58:49.98 & @xmath87 & @xmath88 & @xmath87 & 2000/07/18 + lmc-25 & 05:02:15.86 & @xmath8368:00:55.10 & @xmath87 & @xmath88 & @xmath87 & 2000/07/14 + lrrrrrr lmc-1 & @xmath94 & @xmath95 & @xmath96 & @xmath97 & @xmath98 & 1.510 + lmc-4 & @xmath99 & @xmath100 & @xmath101 & @xmath102 & @xmath103 & 1.560 + lmc-5 & @xmath104 & @xmath105 & @xmath106 & @xmath107 & @xmath108 & 0.940 + lmc-6 & @xmath109 & @xmath110 & @xmath111 & @xmath112 & @xmath113 & 0.850 + lmc-7 & @xmath114 & @xmath115 & @xmath116 & @xmath117 & @xmath101 & 1.410 + lmc-8 & @xmath118 & @xmath119 & @xmath120 & @xmath121 & @xmath122 & 3.280 + lmc-9@xmath123 & @xmath124 & @xmath125 & @xmath126 & @xmath127 & @xmath128 & 1.490 + lmc-13 & @xmath129 & @xmath130 & @xmath131 & @xmath132 & @xmath133 & 1.210 + lmc-14 & @xmath134 & @xmath135 & @xmath136 & @xmath137 & @xmath138 & 0.810 + lmc-15 & @xmath139 & @xmath140 & @xmath141 & @xmath142 & @xmath143 & 0.940 + lmc-18 & @xmath134 & @xmath144 & @xmath145 & @xmath146 & @xmath147 & 1.290 + lmc-21 & @xmath148 & @xmath149 & @xmath150 & @xmath151 & @xmath152 & 2.580 + lmc-23 & @xmath153 & @xmath154 & @xmath155 & @xmath156 & @xmath157 & 1.480 + lmc-25 & @xmath158 & @xmath96 & @xmath159 & @xmath160 & @xmath161 & 0.960 + | we present hubble space telescope ( hst ) wfpc2 photometry of 13 microlensed source stars from the 5.7 year large magellanic cloud ( lmc ) survey conducted by the macho project .
the microlensing source stars are identified by deriving accurate centroids in the ground - based macho images using difference image analysis ( dia ) and then transforming the dia coordinates to the hst frame .
none of these sources is coincident with a background galaxy , which rules out the possibility that the macho lmc microlensing sample is contaminated with misidentified supernovae or agn in galaxies behind the lmc .
this supports the conclusion that the macho lmc microlensing sample has only a small amount of contamination due to non - microlensing forms of variability .
we compare the wfpc2 source star magnitudes with the lensed flux predictions derived from microlensing fits to the light curve data . in most cases
the source star brightness is accurately predicted .
finally , we develop a statistic which constrains the location of the large magellanic cloud ( lmc ) microlensing source stars with respect to the distributions of stars and dust in the lmc and compare this to the predictions of various models of lmc microlensing .
this test excludes at @xmath0% confidence level models where more than 80% of the source stars lie behind the lmc .
exotic models that attempt to explain the excess lmc microlensing optical depth seen by macho with a population of background sources are disfavored or excluded by this test .
models in which most of the lenses reside in a halo or spheroid distribution associated with either the milky way or the lmc are consistent which these data , but lmc halo or spheroid models are favored by the combined macho and eros microlensing results . |
understanding the status and evolution of dwarf galaxies is of crucial importance in current astrophysics and cosmology .
they are important ingredients in common scenarios of galaxy formation , either as building blocks or as left - overs of the formation process ( silk 1987 ) .
in addition , dwarf galaxies forming at redshift around and below 1 have been suggested to be responsible for the excess of faint blue galaxies seen in deep photometric surveys ( babul @xmath9 rees 1992 ; babul @xmath9 ferguson 1996 ) .
dwarf galaxies come in three main classes : ( i ) dwarf spheroidals , which contain population ii stars and have low surface brightness ( e.g. , wirth & gallagher 1984 ) ; ( ii ) dwarf irregulars , with various levels of star - formation activity and a high gas mass ( hunter & gallagher 1986 ) ; ( iii ) starbursting dwarfs , including blue compact dwarfs ( thuan 1991 ) , h ii galaxies ( terlevich 1991 ) , and blue amorphous galaxies ( sandage & brucato 1979 ) . the boundary between dwarf and giant galaxies is usually taken to be around @xmath10 mag ( tammann 1994 ) .
an intriguing property of dwarf galaxies is the pronounced dichotomy between the two main classes of dwarf spheroidals and dwarf irregulars : most dwarfs are either gas - poor spheroids or gas - rich disk - like systems , with very few transition objects ( van den bergh 1977 ) .
the relationship between the two classes is still poorly understood : attempts to unify the complex zoo of dwarf galaxies have not been fully successful ( binggeli 1994 ) .
some effort has been made to understand the interrelationship between individual classes in terms of an evolutionary sequence ( e.g. , gallagher , hunter , & tutukov 1984 ) .
starbursting dwarfs can play a key role : a powerful starburst might strip a gas - rich dwarf irregular of its gas and transform it into a gas - poor dwarf - elliptical ( kormendy 1985 ; marlowe 1997 ) .
understanding the star - formation histories in starbursting dwarf galaxies is a prerequisite for understanding evolutionary connections between different object classes ( e.g. , gallagher 1996 ) since during cosmologically brief periods of time their star formation increases dramatically .
this has profound implications for the interstellar medium .
stellar winds and supernovae inject energy and may cause a ` blow - out ' , initiating a galactic superwind ( heckman 1995 )
. numerous examples of galactic superwinds are known from observations ( e.g. , heckman , armus , & miley 1990 ) and their cosmological implications have been discussed ( e.g. , de young & heckman 1994 ) .
however , attempts to relate the superwind properties and the star - formation histories in individual galaxies are quite rare and do not go beyond a qualitative level ( heckman 1990 ; leitherer , robert , & drissen 1992 ) .
yet establishing such a relation is crucial for theoretical models of the superwind hydrodynamics ( suchkov 1994 ; tenorio - tagle & muoz - tuon 1997 ) .
ngc 1569 (= ugc 3056 = arp 210 = vii zw 16 = iras 4260 + 6444 ) has been indicated by gallagher , hunter , @xmath9 tutukov ( 1984 ) as an outstanding object in their sample of active star forming galaxies .
indeed , it is a prime candidate for a detailed study of the star - formation history in a starburst galaxy since at @xmath11 mpc it is the closest starburst galaxy known ( israel 1988a ) . adopting the literature reddening @xmath12 = 0.56 ( israel 1988a ) , photometry of individual stars is feasible with hst down to @xmath13 .
this corresponds to stars with mass of about 3 m@xmath6 in their core - helium burning phase , and consequently to a look - back time of @xmath00.4 gyr ( e.g. , schaller 1992 ) .
the properties of ngc 1569 are typical of a dwarf galaxy . with a distance modulus of @xmath14 = 26.71 ,
its total absolute b magnitude is @xmath15 ( israel 1988a ) , which is intermediate between the two magellanic clouds .
its total mass and hydrogen content are estimated @xmath16 m@xmath6 and @xmath17 m@xmath6 , respectively ( israel 1988a ) .
numerous studies of the chemical composition exist .
the published range of oxygen abundances is very narrow : @xmath18 ( hunter , gallagher , & rautenkranz 1982 ) , 8.37 ( calzetti , kinney , & storchi - bergmann 1994 ) , 8.26 ( devost , roy , & drissen 1997 ) , 8.29 ( gonzlez - delgado et al . 1997 ) , and 8.19 ( kobulnicky & skillman 1997 ; martin 1997 ) .
searches for chemical composition gradients were done by devost et al . and
kobulnicky & skillman .
no significant evidence for chemical inhomogeneities was found .
if [ o / fe ] = 0.0 is assumed , the average metallicity of ngc 1569 is @xmath19 , with about 0.2 dex uncertainty .
thus , ngc 1569 is a gas - rich system with smc - like composition in a relatively early stage of its chemical evolution .
the most detailed study of the warm ( @xmath010@xmath20 k ) gas of ngc 1569 was done by waller ( 1991 ) , who found evidence for a link between the star - formation history and the morphology and kinematics of extended gas structures .
the same connection is suggested by hot ( @xmath010@xmath21 k ) gas observed with the rosat and asca satellites ( heckman 1995 ; della ceca 1996 ) .
the extended x - ray emission is consistent with a starburst driven galactic superwind which could in principle lead to a large - scale disruption of the interstellar medium .
evidence for large numbers of supernovae was also found from the non - thermal radio spectrum ( israel 1988b ) .
one of the most spectacular properties of ngc 1569 are its ` super star clusters ' .
these high - density star clusters were first detected and discussed by arp & sandage ( 1985 ) and melnick & moles , & terlevich ( 1985 ) and later studied by oconnell , gallagher , & hunter ( 1994 ) , ho & filippenko ( 1996 ) , de marchi ( 1997 ) , and gonzlez - delgado ( 1997 ) .
the super star clusters reflect a recent starburst in which at least @xmath22 m@xmath6of gas was transformed into stars within about 1 pc .
super star clusters are similar to young galactic globular clusters observed shortly after their formation ( meurer 1995 ) .
pre - costar hst studies of the field population were done by oconnell ( 1994 ) and vallenari & bomans ( 1996 ) ( hereinafter vb ) . according to the latter authors , ngc
1569 experienced a strong starburst until a few myr ago , when its activity subsided .
the repaired hst offers the possibility of significant improvement over these previous studies .
we have therefore embarked on an extensive spectroscopic and photometric hst study in an attempt to confront the observational data with the predictions of theoretical models , and derive detailed informations on the star formation ( sf ) history in ngc 1569 . in a first paper ( de marchi 1997 ) we discussed the photometry of the sscs , while in this paper we present the results for the resolved stars in the same field , which are interpreted in terms of the recent sf history via theoretical simulations .
the data and their analysis are described in section 2 .
the derived color - magnitude diagram and luminosity function are in section 3 .
our theoretical models are described in section 4 .
they are compared with the observations in section 5 . in section 6
we discuss our results and their implications for the evolution of dwarf galaxies , and we present in section 7 our general conclusions .
the observations of ngc1569 described in this paper were obtained with wfpc2 on board the hst on 1996 january 10 ( ut ) .
ngc1569 was centered in the field of view of the planetary camera ( pc ) yielding an effective plate scale of @xmath23pixel@xmath7 and a @xmath24 field of view .
the plate scale and field of view of the adjacent wide field chips ( wf ) are @xmath25pixel@xmath7 and @xmath26 , respectively .
images were obtained in three broadband filters , f380w , f439w and f555w , whose photometric properties are similar yet not identical to the johnson @xmath27 filters , and are described in detail in the wfpc2 handbook ( biretta 1996 ) .
since the photometric quality of the f380w images turned out to be poor , it was not possible to measure stars fainter than @xmath024 mag with photometric errors smaller than @xmath00.2 mag .
we have then concentrated our analysis on the f439w and f555w data only , and used the f380w images to check for spurious detections ( see below ) .
a set of 16 deep exposures was collected for each of these filters , all of duration 630 s in f439w and 450 s in f555w .
images were taken at 4 different positions , dithered by a few pixel in order to improve the sampling of the final combined images , to recover faint stars with reliability , and to eliminate high frequency noise .
the whole dataset has been processed using the standard stsci pipeline procedure .
after correction for geometrical distortion , images taken with the same filter have been registered and combined to remove cosmic ray hits .
total exposure times of the final combined images are @xmath010,000 s and @xmath07200 s in f439w and f555w , respectively . for each filter ,
a set of short exposures was also obtained to provide photometric measurements of stars which would saturate in the deep images .
exposures of 20 s , 40 s , and 60 s duration were obtained for the f555w , f439w , and f380w filters , respectively , at two different pointings and are described in de marchi ( 1997 ) .
many bright star clusters and several thousand stars are clearly visible in the deep wf+pc images .
the automated star detection routine daofind was applied to the data , by setting the detection threshold at @xmath28 above the local background level .
we carefully examined by eye each individual object detected by daofind and discarded saturated stars , a number of features ( psf tendrils , noise spikes , etc . )
that daofind had misinterpreted as stars , as well as a dozen extended objects ( star clusters ) whose full width at half maximum ( fwhm ) exceeded that of typical stellar objects in pc frames ( fwhm @xmath29pixel ) by a factor of two or more . in this way
, a total of @xmath07000 unsaturated stars were detected in the field covered by the pc .
the severe crowding of our images , particularly towards the center of the galaxy , led us to prefer the use of the psf - fitting daophot photometry package over the simpler , but more limited , aperture photometry techniques .
a sample psf was built by using three relatively isolated and moderately bright stars ( @xmath2 @xmath2 21.5 ) located on opposite sides with respect to the body of the galaxy .
instrumental magnitudes measured in this way were converted into the wfpc2 synthetic system vegamag as indicated in holtzman ( 1995 ) .
interested readers are referred to that paper for further details on the characteristics of the wfpc2 photometric system . in this paper , and
refer to the wfpc2 synthetic system .
although these magnitudes are very similar to johnson @xmath30 and @xmath31 magnitudes , no attempt has been made to convert our magnitudes into the standard johnson system .
the main source of uncertainty in our photometry is crowding , particularly in the central regions of ngc1569 , where the average star - to - star distance is on the order of @xmath04 pixels . the photometric error estimated by daophot ( ) , and
confirmed by artificial star tests ( see below ) , is smaller than @xmath32mag in both bands down to @xmath2 @xmath2 24 , and it grows to @xmath00.3 mag at magnitude 26 in both bands .
the behavior of the photometric error with magnitude is shown in fig .
[ fig - sigma ] .
we carried out a series of artificial star tests in order to estimate the fraction of stars lost to crowding as a function of their brightness .
we ran 10 trials in each of the two filters , by adding each time a number of artificial stars not exceeding @xmath33 of the total number of objects in the frame , to avoid overcrowding .
the magnitudes of the artificial stars were distributed according to the empirical luminosity function .
recursive daofind runs , with the same parameters as in the reduction of the scientific data , allowed us to assess the fraction of added stars that were recovered and the magnitudes at which these objects were detected . since the artificial star routine ( addstar ) places objects at random over the area of the frame , these tests provide an average value of the photometric incompleteness in our images .
artificial star tests were run on both filter frames , but the f439w image , being less deep , in practice defines the overall completeness limit .
all artificial stars not recovered in the f555w frames were also missing in the f439w ones .
this is partly due to the intrinsic color of the stars , and partly to the lower efficiency of the instrumental setup ( wfpc2 + f439w ) in the bluer band .
since the assumed color distribution of the artificial stars follows the mean locus observed in the color - magnitude diagram ( cmd ) , the incompleteness factors derived in the two filters are not independent and , as such , the overall incompleteness fraction is set by the larger of the two values at each magnitude . for each artificially added star we compared the input magnitude with
that at which the star was recovered , and found no systematic difference .
in addition , the standard deviation of the recovered values is in excellent agreement with the @xmath34 uncertainty of the photometric reduction at any magnitude level .
our simulations show that at @xmath2 25.0 on average about 30% of the artificial stars added to the pc chip are not detected by the automated routine , but this number goes down to only 8% at @xmath2 23.5 . at magnitude levels
brighter than @xmath2 23 all artificial stars are detected , except when they fall on top of one of the bright super star clusters .
our photometry is , therefore , @xmath0100% complete for @xmath35 23 .
table [ tab - compl ] shows the completeness factors as a function of the f439w magnitude , as determined by the artificial star tests .
these are the values adopted in our theoretical simulations .
the artificial star experiments provide us with a statistical estimate of the number of stars lost to crowding .
there is , however , another issue that must be considered when comparing the observations with theoretical simulations : blending affects the derived magnitude of detected objects .
the extent to which the images of two stars can merge with each other , thereby resembling a single object with photometric properties different from those of the parent stars , is a function of their angular separation and of their magnitude difference .
if the projected angular distance between two stars is smaller than the psf - fwhm , they will be detected as a single object by the photometric routine , and the number of stars lost because of this is included in the incompleteness factors . concerning the photometric error , we select objects with @xmath35 0.1mag and take 0.1mag as an estimate for the photometric accuracy of the data .
thus , we are essentially interested in those cases in which the uncertainty introduced by blending on the determined magnitudes is larger than this value .
this occurs when the blended stars differ in magnitude by less than @xmath02 mag .
our tests show , however , that even in this case the sharp psf of the pc camera considerably limits the effects of blending , in spite of the high density of stellar objects in our frames .
the results of the artificial star experiments , as well as statistical considerations , indicate that the probability of blending for a pair of stars differing less than 2mag in flux amounts to @xmath36 for objects brighter than @xmath2 25 .
therefore , it is unlikely that blending introduces a large photometric uncertainty .
blending of objects with a large magnitude difference is much more frequent . in this case , the fainter component is virtually lost with a small effect on the brighter one , and this is accounted for in the simulations by correcting for photometric error and incompleteness .
two of the wf frames have also been reduced , the third having been discarded because of the presence of a highly saturated star .
we found @xmath01000 objects with @xmath35 0.2mag in both the f555w and f439w filters .
since the general characteristics of the stellar distributions in the cmds derived for the wf and for the pc frames are very similar , we proceeded with a quantitative analysis only for the cmd derived from the pc frames .
[ fig - cmdo ] shows the cmd in the hst f439w and f555w bands resulting from the analysis of the region of ngc 1569 contained in the pc field of view . as discussed by de marchi ( 1997 )
, this region contains three super star clusters ( sscs ) , two of which form a close pair which is barely resolved .
only integrated photometry is feasible for the sscs , which are not resolved into stars .
the resulting @xmath37 and @xmath38 colors are indicative of young ages , but is not possible to date the clusters precisely .
the diagram in panel a ) shows the 2809 objects measured in both the f555w and f439w bands with a photometric error @xmath35 0.2 , while the diagram in panel b ) shows the subsample of 801 objects with @xmath35 0.1 . by comparing the two panels it is apparent that the major features of the galaxy stellar population are well delineated in both cases .
clearly , the diagram with the tighter selection criterion contains a lower number of stars because it reaches a brighter magnitude limit , and because it rejects most of the objects with extreme colors . in the cmd shown in panel
a ) there are some very blue objects , populating the bottom left corner .
taking into account the average extinction affecting our field , objects appearing with ( ) @xmath3 0.5 have intrinsic colors bluer than @xmath2 1 .
however , neither galactic nor extragalactic objects can have colors that blue : both white dwarfs with magnitudes in the relevant range , and the bluest qsos have intrinsic @xmath39 .
we have thus examined in detail all the 72 objects with ( ) @xmath40 and photometric error @xmath35 0.2mag in both bands .
they are all concentrated in the region of the pc field dominated by the main body of ngc 1569 , suggesting that these extremely blue objects originate from phenomena associated with the galaxy ( either physical or spurious ) , and not from background or foreground effects . a detailed examination in all of the three photometric bands revealed that most of these objects are spurious detections , due to noise peaks falling on the wings of stellar images , or to the blend of 2 stars .
when selecting only the objects with errors @xmath35 0.1mag no point source with extremely blue color is found . since the shape and position of the bulk of the blue plume and of the red tail
look unaffected by the selection criterion , we will adopt the diagram in panel b ) as reference observational cmd for the stellar population in the field of ngc 1569 .
the general morphology of the cmd of ngc 1569 is similar to that of local group irregulars observed from the ground ( see , e.g. , freedman 1988 ; tosi 1991 , hereinafter tgmf ; greggio 1993 , hereinafter gmtf ; marconi 1995 , hereafter mtgf ; gallart , aparicio , @xmath9 vlchez 1996b ) , with a quite scattered distribution and a prominent concentration of stars in the blue plume .
the cmd shows a smaller fraction of red stars , but this is mostly due to the selection effect of the magnitude limit for object detection in the f439w frame .
indeed , the @xmath41 vs. @xmath42 diagram of ngc 1569 obtained vb shows a significant population of red ( mostly intermediate age ) stars .
the median color of the blue plume ( ) @xmath2 0.45 is quite red as a consequence of high foreground extinction . by correcting this color for the average foreground reddening @xmath43 as estimated from uv observations by israel ( 1988a ) , an intrinsic ( )
@xmath44 = 0.1 is derived .
this is equal to or slightly redder than the median @xmath45 color of the blue plumes in other nearby irregulars ( see , e.g. , mtgf and gallart 1996b for ngc 6822 ; tgmf for sextans b ; bresolin 1993 and gmtf for ngc 3109 ) .
however , since gonzlez - delgado ( 1997 ) have derived @xmath46 for the super star clusters in ngc 1569 ( using uv , optical and near - ir spectra ) , it may well be that this galaxy is affected by internal differential reddening .
thus , the intrinsic colors of the blue plume stars could be bluer than the previous estimate . given the low galactic latitude @xmath47 = + 11@xmath48 of ngc 1569 , the observed cmd is likely to be populated by a significant number of foreground stars .
despite our original request , the time allocated on hst was not sufficient to take exposures of adjacent fields and measure directly the galaxy contamination .
since the cmd derived from the wf images looks very similar to that derived from the pc frames , the galaxy extends substantially over the wf frames , which can not then be safely used to discriminate between members and foreground / background objects .
however , using ground based observations , vb derive a very low density of external field objects in the direction of ngc 1569 .
thus our cmd should be almost unaffected by foreground / background contamination .
the full dots in fig .
[ fig - fl ] represent the luminosity function ( lf ) of the 801 selected stars of ngc 1569 , whereas the open circles correspond to the stars with 0 @xmath1 ( ) @xmath1 0.5 , which in the following will be referred to as the _ reference blue plume_. this subsample has been selected simply for the purpose of comparing our results with those relative to other galaxies and/or other authors , and does not represent either the whole blue plume or the pure main sequence ( ms ) stars . in the past ,
the lf of the blue plume was usually adopted for estimating the mass function , under the assumption that it is essentially populated by ms stars .
however , as already noticed for other irregulars ( e.g. , freedman 1988 ; tgmf and gmtf ; gallart 1996b ) , the comparison of the observational cmd with theoretical stellar models demonstrates that the blue plume is populated also by a significant fraction of stars at the hot edge of the blue loop evolutionary phase .
this represents a serious problem for authors deriving the initial mass function ( imf ) from the luminosity functions of the blue plume assuming it contained only ms stars .
the ms post - ms blend is apparent also in the @xmath41 vs. @xmath42 diagrams described by vb . as discussed there and in the next section , it is virtually impossible to discriminate ms from post - ms stars at the brightest magnitudes , where the morphology of the theoretical evolutionary tracks of massive stars in the cmd leads to the merging of the blue edges of the loops into the region populated by ms stars .
our ( ) selected _ reference blue plume _ is therefore just a compromise between the requirement of removing as much as possible the evolved stars and the need of keeping the brightest ms objects .
the slopes of the lfs are usually derived with least squares linear fitting to their complete portions . however , as shown in gmtf and mtgf , the large statistical fluctuations due to the small number of sampled stars , especially at the brightest magnitudes , may significantly alter the results of a method requiring data binning . to overcome this problem ,
we have applied a maximum likelihood fitting to evaluate the slope of the lf in the _ reference blue plume_. down to = 23 , where our sample is complete ( see previous section ) , the slope turns out to be @xmath49 log @xmath50 .
this value is higher than @xmath51 derived by vb , a result of the different methods : for instance , a linear fit gives a flatter slope of @xmath52 for our data set .
furthermore , the fit of vb is not restricted to the data portion where completeness is reached .
our slope @xmath53 is consistent with the average @xmath54 derived by freedman ( 1985 ) and hoessel ( 1986 ) from a large sample of irregulars .
it appears , however , slightly steeper than the slopes derived by us in a homogeneous way ( see tgmf ; gmtf ; mtgf ) for four nearby dwarf irregulars , whose average value is @xmath55 .
the observed cmd and the ( total ) lf described in the previous section have been compared to corresponding theoretical simulations , based on stellar evolutionary tracks .
this method has proven to be a powerful tool to investigate the star formation history of galaxies with young stellar populations ( see tosi 1994 ; greggio 1994 ) .
synthetic cmds are constructed via monte carlo extractions of ( mass , age ) pairs , according to a prescribed imf and a star - formation rate ( sfr ) law .
the extracted object is placed on the theoretical h
r diagram ( hrd ) by interpolating on evolutionary tracks , and its luminosity and effective temperature are converted into magnitudes using tables for bolometric corrections and colors . a photometric error and incompleteness test ,
both as measured on the observational frames , is finally applied , and the synthetic star is placed on the cmd .
once the number of objects populating the synthetic cmd equals that of the observed one , the procedure is stopped , yielding the quantitative level of the sfr consistent with the observational data , for the prescribed imf and shape of the sf law .
this procedure automatically accounts for the stochastic nature of the sf process , small number statistics , photometric errors , and incompleteness factors affecting the real data .
in addition , it easily allows us to explore the effects of using different sets of stellar tracks .
we have considered several sets of stellar tracks ( table [ tab - mod ] ) in order to compare their effect on the derived sf history , and to check their adequacy in reproducing the main features of the observed cmd .
the interpolation algorithm was carefully chosen to reproduce the main characteristics of the evolutionary sequences . due to numerical effects ,
the short lived , post - ms phases may be underrepresented in the monte carlo simulations .
thus , the tracks have been trimmed in subphases , and for each extracted object the evolutionary subphase is first determined .
interpolation between adjacent tracks is then performed within the appropriate subphase .
this procedure ensures that the main features of the synthetic diagrams , ( e.g. , the blue - to - red ratio , the ratio between the ms and post - ms objects , the number of w - r stars , etc . )
are consistent with the adopted set of tracks .
the luminosity and effective temperature of the synthetic objects are converted into magnitudes and colors by interpolating on tables constructed from the collection of model atmospheres by lejeune , cuisinier , @xmath9 buser ( 1997 ) , which cover ( although non - uniformly ) the effective temperature and gravity ranges : 2000 k @xmath35 @xmath35 50,000 k , 1.02 @xmath35 log @xmath56 @xmath35 5.5 , and are characterized by several values of the metallicity ( [ fe / h ] ) , with solar elemental ratios .
it should be noticed here that the tracks used for the simulations do not reach very low gravities , or temperatures lower than @xmath03600 k. therefore , the model atmospheres used for our synthetic diagrams are essentially those of kurucz ( 1992 ) . the sets at metallicities [ fe / h ] = 0.5 and [ fe / h ] = 1 have been selected for our application . the conversion tables have been prepared as follows . for each model atmosphere in the grid ,
a theoretical bolometric correction has been computed : @xmath57 where @xmath58 is the @xmath31-filter transmission function in the vegamag system .
similar tables for the @xmath59 and @xmath59 colors ( in the vegamag system ) have been produced by processing the model atmospheres through the calcphot routine in the synphot package .
for the objects populating our synthetic hrd , the following relation applies @xmath60 where @xmath61 is a calibration constant . in order to ensure that the theoretical magnitudes are in the _ same photometric system as the observational cmd _
, @xmath61 has been fixed by enforcing that for the solar model atmosphere ( = 5777 k , log @xmath56= 4.4 , [ fe / h ] = 0 . ) , we obtain @xmath62 = 4.7785 if we adopt @xmath63 erg s@xmath7 . this is the value in the vegamag system found by synphot when _ observing _ a star with the same flux distribution , the solar radius , at a distance of 10 pc . in this system ,
the solar model and the model atmosphere representing vega ( = 9400 k , log @xmath56 = 3.9 , [ fe / h ] = 0.5 ) have ( ) = 0.717 and 0.0066 , respectively .
for each synthetic object , characterized by ( mass , log @xmath64 , log ) , the theoretical bolometric correction and colors are determined through interpolation in these tables .
@xmath62 is then derived from eq .
( [ eq : mag ] ) , and henceforth also the hst magnitudes in the vegamag system .
the synthetic cm diagrams ( cmd ) and lfs depend on the input parameters in the simulation : imf , history of the sfr , distance , reddening and theoretical stellar evolutionary tracks . adopting a value for the distance modulus and reddening , the superposition of the evolutionary tracks to the observed cmd yields the range of ages and masses sampled by our data . as an example , in fig .
[ fig - cmdth001 ] we show the geneva tracks with z=0.001 trasformed to the observational plane , having adopted the canonical values @xmath65 ,
@xmath66 from the literature .
comparing fig .
[ fig - cmdth001 ] with fig .
[ fig - cmdo ] , panel a ) , one derives that the bluest stars in the blue plume are ms objects with masses larger than @xmath0 9 , and then ages younger than @xmath0 30 myr .
the wide color distribution shows the presence of evolved stars , with masses larger than @xmath0 4 , and then ages younger than @xmath0 0.15 gyr .
thus , sf has been active in our field over @xmath0 0.15 gyr .
although the smoothness of the star distribution in the cmd qualitatively suggest a continuous kind of sf , this issue can only be quantitatively assessed computing the simulations , which take into account the smearing effect of photometric errors and the incompleteness at faint magnitudes . in our procedure , then , we first try the simplest possible case , i.e. a constant sf over the look - back time , and then vary the sfr shape and imf slope to reproduce the data distributions . when an acceptable representation of the observation is reached , we perturb the parameters to test the stability of the solution . the comparison of the theoretical simulations with the observations is performed quantitatively on the lf , and mostly qualitatively on the appearance of the cmd .
a detailed quantitative comparison of the color distribution of stars ( like e.g. , tolstoy @xmath9 saha 1996 ) would not add much information , due to the smearing effect of photometric errors , and , more importantly , to the intrinsic uncertainty affecting the effective temperatures of models of massive stars in the post - ms phases ( renzini 1992 ) .
small variations of the theoretical result in large differences of the predicted distribution of objects throughout the cmd , due to the high sensitivity of colors and bolometric corrections to the effective temperature in the relevant range .
because of systematic uncertainties , there is often no significant difference between the best simulation and other similar solutions .
rather , our aim is to isolate the range of imf and sfr parameters which are consistent with the data .
we have computed a few hundred synthetic cmds varying the input parameters mentioned above . in the next subsections
we discuss our results and the influence of the input parameters on the conclusions .
we start with adopting a distance of 2.2 mpc , and present our preferred solutions for various sets of evolutionary models .
second we discuss how different sf histories influence the results .
finally we consider the effect of an uncertainty in the distance .
the reddening is varied , within the observational estimate , in such a way that the average color of the synthetic blue plume agrees with the observations .
the sf history derived from the analysis of an observational cmd depends on the set of tracks adopted in the simulation ( greggio 1994 ) .
we have therefore considered several sets of stellar tracks ( table [ tab - mod ] ) .
although the metallicity of ngc 1569 is z @xmath0 0.004 we have computed simulations based on the geneva tracks with several values of @xmath67 .
this is motivated by the fact that the overall appearence of the cmd is sensitive to the morphology of the blue loops , which in turn critically depends on z ( ritossa 1996 ) .
we have also explored the difference in the derived interpretation when adopting the padova set of tracks at z = 0.004 .
the main difference between the padova and the geneva sequences concerns the treatment of convective overshooting : both sets include core overshoot , but to a lesser extent in the geneva models .
in addition , the padova tracks also include overshooting from the base of the convective envelope .
it is worth recalling some general features of the evolutionary sequences : \i ) all sets are characterized by extended loops for masses above 4 m@xmath6 with the maximum extension for the geneva @xmath67 = 0.001 and padova @xmath67 = 0.004 sets .
\ii ) among the geneva tracks , the fraction of evolutionary lifetime spent in the blue portion of the cmd during the post - ms phase is largest for the @xmath67 = 0.001 tracks ( actually close to 100@xmath68 for stars more massive than @xmath04 m@xmath6 ) , and it decreases with increasing metallicity .
the padova set at @xmath67 = 0.004 is characterized by a value of this ratio which is intermediate between the geneva @xmath67 = 0.001 and @xmath67 = 0.004 sets .
\iii ) the ratio between post - ms and ms evolutionary lifetimes for stars more massive than @xmath08 m@xmath6 is @xmath00.1 for the geneva tracks .
the same ratio is considerably smaller for the padova tracks , reaching a minimum value of @xmath00.06 for the 12 m@xmath6 track .
these characteristics crucially determine the main differences in the simulations constructed with the various sets of tracks . adopting the geneva tracks with @xmath67 = 0.008 and
@xmath67 = 0.004 we were not able to reach a satisfactory agreement between the synthetic diagrams and the observations , regardless of the choice of the imf and sf parameters .
as an example , in fig .
[ fig - synth1fl ] we show two of the best cases obtained with @xmath67 = 0.008 ( panel a ) and @xmath67 = 0.004 ( panel b ) .
the corresponding lfs are shown in panel c ) , with the solid line referring to the simulation displayed in panel a ) .
the adopted reddening values are @xmath69 and 0.56 for panels a ) and b ) respectively , which compensate for the different metallicities of the tracks , and are both consistent with israel s ( 1988a ) estimate .
the cmd in panel a ) has been computed with two partially overlapping sf episodes , so that the sfr peaks around @xmath00.05 gyr ago .
the adopted imf exponent is @xmath70 ( in these units the salpeter s exponent is 2.35 ) .
the simulation in panel b ) adopts a mildly decreasing sfr , and a salpeter imf . in both simulations
the synthetic blue plume extends up to too bright magnitudes and the bright portion of the lf is overpopulated over a wide magnitude range .
these characteristics appear in all the simulations with this sets because they are related to the shape of the tracks on the cmd .
infact , due to the relatively short color extension of the blue loops of models with mass around 5 ( see fig .
[ fig - cmdth008 ] for the @xmath71 set ) , the faint portion of the blue plume has to be mainly populated by ms stars .
this is a consequence of the photometric incompleteness in the f439w frames , which depletes the number of evolved stars when their colors are too red .
thus , with these tracks a large number of massive stars is required , whose post - ms progeny shows up as bright evolved supergiants . in the specific case of the @xmath67 = 0.008 tracks , at magnitudes brighter than @xmath2 21.5
we find post - ms objects with masses between 15 and 25 m@xmath6 , which are the evolved progeny of fainter ms objects populating the blue plume in the range @xmath2 25 22.5 .
thus , the excess of bright objects can not be suppressed without substantially underpopulating the faint portion of the blue plume , and could only be avoided by invoking such a delicate fine - tuning of the sf behavior , to appear very contrived .
similar arguments apply to the @xmath67 = 0.004 set , in which the extension of the blue loops of the 5 m@xmath6 and 7 m@xmath6 tracks is so short , that the herztsprung gap appears at the faint magnitudes , in spite of the large photometric error .
the cmds in fig .
[ fig - synth1fl ] were computed with the low mass loss option in the geneva models .
switching to the high mass - loss sets does not improve the comparison between the simulations and the data .
for @xmath67 = 0.008 , the higher mass loss only affects the location of the core helium burning phase for the more massive models . as a consequence ,
the excess objects are still present , yet shifted to intermediate and red colors . for @xmath67 = 0.004 , the adoption of a high mass loss rate would only affect the evolution of the most massive stars ( @xmath72 m@xmath6 ) , spanning a mass range which is almost not populated in our cmd . as can be seen in fig .
[ fig - cmdth001 ] , the loops for the geneva tracks with @xmath67 = 0.001 extend to very blue colors , and a substantial fraction of the faint blue plume can be populated by relatively low mass stars in their core helium burning phase .
unlike the previous cases , we find satisfactory agreement with the observed cmd and lf with various combinations of the imf slope and sfr , provided that the activity ceased @xmath07 myr ago ( corresponding to the evolutionary lifetime of a @xmath2 30 m@xmath6 star ) to avoid an excess of bright stars .
[ fig - synth2fl ] shows our best models for a constant sfr ( panel a ) and an exponentially declining sfr with an e - folding time of 0.6 gyr ( panel b ) .
notice that this latter case implies a variation of the sfr of only @xmath0 20@xmath68 over our short look - back time .
the corresponding lfs are displayed in panel c ) .
a reddening of e(b - v)=0.56 has been adopted for these simulations .
it can be seen that both cases give a very good fit to the observed lf .
we adopted a steeper imf for the constant sf case , in order to avoid too many bright stars .
the synthetic cmds also give a fair representation of the data , except for a slight paucity of stars at intermediate and red colors . in order to see whether the data allow for large variations of the sf activity during the last 0.15 gyr we have computed simulations with two separate sf episodes .
[ fig - synth3fl ] shows our best models obtained by assuming that the sfr in the most recent episode is higher ( panel a ) , and lower ( panel b ) with respect to the previous episode . in both cases
the sfrs in the two episodes differ by a a factor of @xmath0 1.2 .
again , we achieve a fair representation of the data when appropriately tuning the imf slope with the sf history . in all acceptable simulations , however ,
the sfr in the last 0.15 gyr has proceeded approximately at the same level , and the slope of the imf is close to salpeter s value or slightly steeper .
our procedure corresponds to counting and weighing the stars which appear in the observational cmd , and determine their ages .
the derived level of the sfr is then relatively insensitive to the combination of imf and sf history adopted in the simulations . indeed , our models in fig .
[ fig - synth2fl ] and fig .
[ fig - synth3fl ] show agreement with the observations , in spite of the different prescriptions . yet , the average values of the sfrs over the last 0.15 gyr , in stars more massive than 4 , are 0.07 and 0.09 m@xmath6 yr@xmath7for the simulations in fig .
[ fig - synth2fl ] ( panels a and b respectively ) , and 0.08 and 0.11 for the simulations in fig .
[ fig - synth3fl ] ( panels a and b respectively ) .
this applies only to the mass range directly observed . when extrapolating the sfr for stars with masses ranging from 0.1 to 120 m@xmath6 , the average rates become @xmath0 3 m@xmath6 yr@xmath7for @xmath70 and @xmath00.5 m@xmath6 yr@xmath7 , for @xmath73 .
unlike the geneva tracks with the same metallicity , the padova tracks with @xmath67 = 0.004 are characterized by well extended loops for masses larger than @xmath05 m@xmath6 whose blue edges almost merge into the ms . as a consequence
, the faint blue plume is populated by ms and helium burning stars . moreover , in the range 9 25 m@xmath6
the ratio between post - ms and ms lifetime is significantly smaller than in the geneva tracks , implying a lower ratio between bright evolved supergiants and their ms progenitors . for these reasons the synthetic diagrams with this set of tracks are fairly different from those obtained with
the geneva tracks with the same metallicity .
we are able to obtain a fair representation of the data , provided that the activity stopped @xmath74 myr ago to avoid an excess of bright stars . by adopting the spectroscopic value for the reddening , @xmath12=0.56 , the color of the synthetic blue plume appears somewhat redder than the observed one .
we have therefore computed cases with both @xmath66 and 0.46 .
although assuming a constant sfr we obtain acceptable simulations , the best fitting models are characterized by two distinct episodes of sf , the most recent one starting @xmath0 30 myr ago .
we find acceptable solutions with sfrs both increasing and decreasing with time , in combination with different values of the imf slope and adopted reddening .
however , all the solutions indicate that the sfr has not varied by a large amount within the last 0.15 gyr , and that the imf slopes are slightly steeper than salpeter s value , and close to that derived for solar neighborhood stars ( e.g. , scalo 1986 ) . as an example , fig .
[ fig - synth4fl ] shows the synthetic cmds obtained with two bursts , the most recent one occurring at a lower rate in both panels a ) and b ) .
the theoretical and observed cmds and lfs are in good agreement , perhaps with a slight excess of red supergiants in the models . in these models
the ratio between the sfr in the older and most recent episode is 1.5 ( panel a ) and 1.2 ( panel b ) .
thus , also with this set of tracks the data are consistent with an almost constant sfr over the last 0.15 gyr ( within 20@xmath68 to 50@xmath68 ) .
the average level of the sfr in the simulations in fig . [ fig - synth4fl ] are 0.1 , 0.07 m@xmath6 yr@xmath7 in panels a and b respectively , in stars younger than 0.15 gyr and more massive than 4 m@xmath6 .
when extrapolating the imf down to a lower mass cut - off of 0.1 m@xmath6 we obtain an average sfr of is @xmath01 m@xmath6 yr@xmath7if @xmath75 , or @xmath03 m@xmath6 yr@xmath7 if @xmath70 .
this holds when adopting @xmath76 . for the higher reddening
the derived rates are @xmath0 1.3 times larger . given the smooth distribution of stars in the cmd , and the shape of the lf , large variations of the average sfr within the last @xmath0 0.15 gyr are not consistent with the data , as we have tested numerically . in this section
we rather concentrate on the epoch of the recent stop in the sf , and on the possibility of short quiescence periods .
[ fig - synth10fl ] shows the sensitivity of the results on the adopted age for the recent stop in the sf in the case of the geneva tracks with @xmath67 = 0.001 . in the bottom panel
the sf has ceased 5 myr ago , while in the top panel it is still ongoing .
a steep imf has been adopted for these simulations , in order to minimize the number of massive stars . despite that
, an ongoing sf predicts an excess of bright objects , while a 5 myr stop is consistent with the data .
[ fig - synth6fl ] panel b ) shows the case of a constant and currently ongoing sf having adopted the padova @xmath67 = 0.004 models . in spite of the steep exponent of the imf , the cmd turns out to be overpopulated at the bright end , over a range of @xmath01.5 mag . with this set of tracks , adopting @xmath77=3 , we still find an excess of bright stars stopping the sf 7 myr ago .
it is worth noticing that the age of the stop is constrained by the bright portion of the lf , over a range of @xmath0 1 mag or more .
we now turn to consider the possibility of discrete episodes of sf , separated by quiescent phases . in general , this behavior has a definite signature on the cmd , i.e. the lack of objects around the isochrones with ages corresponding to the quiescence period , which shows up in the evolved portion of the diagram as an almost horizontal gap . due to the effect of photometric errors , however , this signature can go undetected on the observed cmd .
a quiescent period lasting from 35 to 30 myr ago has been assumed in the simulations shown in fig .
[ fig - synth4fl ] , in order to improve the appearance of the synthetic cmds .
the simulations with a constant sfr are characterized by a clump of blue supergiants at @xmath2 22 ( see fig .
[ fig - synth6fl ] , panel b ) populated by objects with masses in the range @xmath09 to 12 m@xmath6 .
the blue loops of these sequences happen to converge to the same location in the cmd , providing an excess of objects in this region .
isochrones with ages of 35 to 30 myr are populated by post - ms objects with masses in the range @xmath0 9 to 10 m@xmath6 and the assumed quiescent period makes this clump of blue sgs less prominent .
it should be noticed that a slightly different shape of the loops of these sequences would not produce this feature , making the quiescent period unnecessary .
we also find that quiescent periods longer than @xmath010 myr should be excluded , since they show up in the synthetic cmd with a very evident horizontal gap , which is absolutely absent in the observed diagram .
this is illustrated in the top panel of fig .
[ fig - synth6fl ] , where we show a simulation in which no sf has occurred between 40 and 30 myr ago . the simulations in fig .
[ fig - synth3fl ] , based on the geneva tracks with @xmath67 = 0.001 , have been computed assuming two contiguous sf episodes .
also for these tracks we have tested the possibility of short interburst periods .
again , we find that quiescent periods longer than @xmath010 myr in the recent past ( less than 0.1 gyr ago ) would be evident in the observational cmd , and are then excluded by the smoothness of the data .
the distance to ngc 1569 is relatively uncertain .
the most widely quoted values in the literature are between 2.2 and 2.5 mpc ( arp & sandage 1985 ; israel 1988a ; oconnell 1994 ) .
they are based on the properties of the still resolved brightest individual supergiants .
however , distances of up to 4 mpc are possible since ngc 1569 is a probable member of the ic 341/maffei 1 & 2 group , whose distance is uncertain .
if ngc 1569 were at 4 mpc , it would be quite extraordinary in several respects .
apart from overluminous individual supergiants it would harbor the most luminous resolved super star clusters known in the universe .
the brightest cluster has @xmath78 to @xmath79 already for the shorter distance ( oconnell 1994 ; de marchi 1997 ) . at 4 mpc
, this cluster would have appeared at @xmath80 if observed at peak brightness 8 myr ago !
a larger distance would make the properties of the ring nebula in the outskirts quite unusual as well .
drissen & roy ( 1994 ) noted that its size and luminosity exceeds those of similar ring nebulae in m33 by up to an order of magnitude .
this apparent discrepancy would be even more amplified for the larger distance . while none of these arguments by itself provides convincing evidence for or against one or the other distance , taken together they all point toward the shorter distance .
however , a larger distance would imply younger ages for the stars in our observed cmd , as well as larger levels of the recent sfr . to test the effect of the uncertainty in the distance to ngc 1569 we have computed simulations assuming a distance of 4 mpc .
the observed cmd now samples stars younger than @xmath0 50 myr .
two of our best cases are shown in fig .
[ fig - synth8fl ] , for the padova tracks with @xmath81 .
the simulations assume a still ongoing sf ( bottom panel ) and a recent stop , at 5 myr ( top panel ) , and again the slope of the imf has been chosen to minimize the number of massive stars .
it can be noticed that , even adopting a large distance , a stop in the recent sf activity is preferred , not to overpopulate the bright end of the lf . the level of the sfr in these simulations is higher by a factor of @xmath0 2.8 with respect to what derived with the shorter distance .
in the previous section we have shown that the geneva tracks with metallicities @xmath67 = 0.008 and 0.004 do not allow us to reproduce the main features of the observed cmd and lf , regardless of the adopted combination of imf and sfr .
on the contrary , we find a fair agreement between the simulations and the data when adopting either the geneva tracks with @xmath67 = 0.001 , or the padova tracks with @xmath67 = 0.004 , the latter value being consistent with the metallicity derived observationally .
these results , however , are mostly based on the shape of the blue loops , which are intrinsically uncertain . therefore , we can neither conclude that there is an indication for a lower metallicity in this galaxy , nor that the input parameters used in the padova tracks are to be preferred to those in the geneva tracks , as our analysis does not allow us to constrain the shape of the tracks down to this level of detail .
rather , we maintain that both sets of tracks have the combination of input physics and metallicity parameter that yields a good representation of our data .
the color distribution of the evolved stars can not be used to determine the metallicity of the system , other than to derive a general ranking .
we can derive several consistent indications on the sf history in the field of ngc 1569 , regardless the adopted sequences : \i ) the sf activity stopped 510 myr ago .
\ii ) during the last 0.10.15 gyr the sfr has been approximately constant
. if quiescent periods have occurred , they did not last longer than @xmath010 myr .
\iii ) the best agreement is generally reached with imf slopes steeper than salpeter s , but the latter would not be inconsistent with the data .
\iv ) the average sfr in our field is approximately 3 , 1 , and 0.5 m@xmath6 yr@xmath7 for imf slopes of 3 , 2.6 , and 2.35 , respectively .
these values refer to stars with masses ranging from 0.1 to 120 m@xmath6 but it should be emphasized that our data only constrain the star formation of objects more massive than @xmath04.5 m@xmath6 . the dependence of these derived values on the adopted set of tracks , reddening and shape of the sf law is found to be small .
\v ) assuming that the minimum mass for a type ii sn explosion is 8 m@xmath6 the derived sfr implies that in the last 0.1 gyr in our region about @xmath82 type ii events have occurred , for an imf slope ranging from 3 to 2.35 .
it may seem that enhancing the recent sfr while steepening the imf would lead to the same result , and to some extent this is true : enhancing the recent sf would populate the cmd with young , and therefore massive , stars , thereby counterbalancing the effect of a steepening of the imf ( see fig .
[ fig - synth2fl ] and fig .
[ fig - synth3fl ] ) .
however , we find that acceptable solutions have parameters in a relatively small range . in other words , when the shape of the sfr is radically different from the one giving a fair solution , it is difficult to obtain acceptable simulations by varying the imf slope ( and vice versa ) . indeed
, the two parameters do not act in precisely the same way , as strengthening the recent sf with respect to the past enhances the population along the younger isochrones by a constant factor .
conversely , varying the slope of the imf affects the distribution of stars along all the isochrones .
thus , for example , when the imf slope is too flat , the lf of the synthetic population is so skewed towards the bright end that only very complicated sf histories may fit the data . as can be seen from figs .
[ fig - cmdth008 ] and [ fig - cmdth001 ] , the shape of the tracks , and therefore of the isochrones as well , is such that in a given magnitude bin we find stars of different generations and different masses , so that it is not possible to disentangle the effects of sfr and imf .
the number of stars falling within suitably chosen boxes in the cmd ( e.g. , bertelli 1992 ) may allow this task .
however , it is very hard to define these suitable boxes in this case , due to the shape of the tracks for massive models in the observed cmd and to the photometric errors .
this cautionary note applies in particular to the splitting of the population of the blue plume into the ms and post - ms contributions . for these reasons we need theoretical simulations , rather than isochrone superposition , to interpret the data
moreover , as shown by our experiments , the shape of the blue loops heavily affects the synthetic cmd and lf so that the final solution depends on the adopted set of tracks .
the flexibility of the simulation code allows us to easily explore this parameter space as well .
it is also worth noting that the sf stop that we infer in the field of ngc 1569 is not particularly sensitive to the adopted upper mass cut off for the imf .
for example , when the mass distribution is truncated at @xmath040 m@xmath6 the synthetic diagrams obtained with an ongoing sf are still overpopulated at the brightest magnitudes . indeed , lowering the mass cut off does not appreciably alter neither the slope of the lf , nor the overall ratio between the objects populating the fainter and brighter portions of the lf .
conversely , when stopping the sf activity a few myr ago , the whole ms population is reduced with respect to the evolved one .
our quoted values for the sfr refer to a single - slope imf with stars distributed from 0.1 to 120 m@xmath6 , and their dependence on the lower mass cut - off is of course very high .
for example , the mass in stars within 0.1 and 1 m@xmath6 is @xmath00.55 , 0.75 , and 0.9 of the total for @xmath73 , 2.6 , and 3 , respectively .
furthermore , these values depend on the imf slope at the low mass end . in a recent study by gould , bahcall , @xmath9 flynn ( 1997 ) ,
the imf for m dwarfs in the galactic disk can be fitted by a two - slopes imf , with a break at @xmath83 m@xmath6 and with a slope @xmath84 ( in our usual units ) in the low mass range . when adopting these prescriptions for the low mass end of the imf , and varying the slope of the mass distribution only over the range @xmath85 m@xmath6 , our average sfrs become @xmath00.8 , 0.5 , and 0.3 m@xmath6 yr@xmath7 for slopes of 3 , 2.6 , and 2.35 , respectively .
the only previous study primarily dedicated to the resolved stellar population in the field of ngc 1569 is that by vb . by using pre - refurbishment wfpc images ,
these authors derive the cmd in the f555w and f785lp filters , roughly equivalent to the @xmath31 and @xmath41 bands .
the method used for the interpretation of the data relies , like ours , on simulations of cmds based on evolutionary tracks , but without including the effect of photometric errors and incompleteness .
an additional difference concerns the use of the lf as a tool to discriminate between possible models .
vb conclude that ngc 1569 experienced a global burst of sf from 0.1 gyr to 4 myr ago , and that a previous episode of sf had taken place in the galaxy , at a substantially lower rate .
our observations are not deep enough in the f439w band to detect the agb stars that reveal this older episode in the @xmath41 vs @xmath42 diagram . as for the recent sf history ,
our results are roughly consistent with their finding .
we prefer an earlier stop of the sf activity , since , when adopting an sf rate ongoing up to 4 myr ago , the synthetic lf tends to be overpopulated at the bright end , especially for the padova tracks with z=0.004 , which is the same set adopted by vb .
vb could not establish quantitatively this effect without the comparison of the theoretical and observed lfs .
in addition , we believe that the photometric incompleteness prevents the determination of the epoch at which the burst started , since stars older than [email protected] gyr would go mostly or totally undetected in our images .
finally , the level of the sfr of the recent burst that we derive using the same set of tracks and imf is consistent with theirs , judging from the estimate of the recent snii activity .
to summarize , our results are in general agreement with vb , but we believe that our conclusions are more robust , due to both the superior quality of the refurbished hst data , and to the method employed for the interpretation .
ngc 1569 has been suggested to be in a post - starburst phase ( israel 1988a ) .
the definition of the starburst and the post - starburst phase is a matter of choice .
here we define the starburst phase as being characterized by ongoing sf . with this definition
the field population of ngc 1569 clearly is in a post - starburst phase since we find that the cmd requires no sf activity over the last few myr . on the other hand , in the same region
there are many hii regions and three sscs , whose age is likely to be younger than 10 myr ( gonzlez - delgado 1997 ) .
therefore the stellar population in ngc 1569 is made of two components : the resolved stars with ages older than @xmath0 10 myr , and the stars in the sscs , which formed more recently .
the resolved stellar component is in a post - starburst stage , while the most recent star formation is taking place in the sscs .
this finding may suggest a picture of discrete sf events mainly generating sscs , which dissolve to create the field stellar population in a certain time interval .
the @xmath010 myr quiescence period would correspond to a characteristic time step of the discrete sf activity .
on the other hand , the compactness of the sscs suggests that they are gravitationally bound systems , and that they are true young globular clusters ( ho @xmath9 filippenko 1996 ) .
in addition , the age of the brightest component of ssc - a has been estimated to be as old as 8 myr ( gonzlez - delgado 1997 ) , and still presents a half - light radius of @xmath01.6 pc ( de marchi 1997 ) . therefore the field population is likely to be the relic of a major episode of sf which involved the whole observed region , and which ended several myr ago .
the sf activity in the field is most likely responsible for the outflow of the interstellar medium analyzed by waller ( 1991 ) , who derived a kinematic age of 10 to 60 myr for the h@xmath77 filaments around ngc 1569 .
this sf activity , however , has not depleted the gas reservoir , and more recent sf is currently occurring in the hii regions and in the sscs .
this may suggest that only the densest regions are able to collapse to form stars following the burst of sf and its consequent energy input in the ism .
the sfr in dwarf irregulars has been found to be characterized by long activity periods , separated by short interburst phases ( _ gasping sf _ , e.g. , tosi 1994 ) . in this respect ,
the field stellar population in ngc 1569 exhibits similar properties , with a sf occurring at an approximately constant level over @xmath00.15 gyr , with quiescence periods shorter than @xmath010 myr .
a quantitative comparison is instructive . among the different possible ways in which the strength of the star formation activity can be characterized , we consider the sfr per unit surface ( see hunter
@xmath9 gallagher 1986 ) . at a distance of 2.2 mpc ,
the observed pc field surveys an area of approximately 0.14 kpc@xmath86 , so that our derived recent sf rates are about 20 , 7 , and 4 m@xmath6 yr@xmath7 kpc@xmath87 , for imf slopes of 3 , 2.6 , and 2.35 , respectively ( again having extrapolated the imf down to 0.1 m@xmath6 ) .
it is worth recalling that the typical sfr quoted for the solar neighborhood ( e.g. timmes , woosley , @xmath9 weaver 1995 ) is @xmath01000 times lower than this value .
the comparison of the sfr in ngc 1569 with those derived for other dwarf irregulars is not straightforward since there is a non - trivial dependence on the method used to measure the sfr .
even when considering the sfr derived from the analysis of the cmd via theoretical simulations , there may be systematic differences due to the adopted set of stellar tracks ( greggio 1994 ) . finally , the sfr depends on the adopted imf , both on the slope and on the mass cut - offs , the latter being not always clearly specified in the literature . with this caveat in mind ,
the recent sfr derived via theoretical simulations for dwarf irregular galaxies in the local group range from approximately @xmath88 m@xmath6 yr@xmath7 kpc@xmath87 to @xmath89 m@xmath6 yr@xmath7 kpc@xmath87 ( mtgf ; tolstoy 1996 ; gallart 1996a ) .
when allowing for the uncertainties introduced by the different methods used , we can conclude that the recent sfr in ngc 1569 is @xmath0 100 times larger than in the most active local dwarf irregular .
ngc 1569 fully deserves its classification as a ( post)-starburst galaxy , as opposed to normal , star - forming galaxies .
could it be that strong sf bursts took place in dwarf irregulars in the past ? the chemical properties , and the gas content of the most active of them are not consistent with the extrapolation of the currently observed activity over a long time .
chemical evolution models show that several bursts of sf may have occurred in these galaxies , but separated by long quiescence periods .
strong bursts of sf could perhaps be accommodated but the duty cycle of this activity has to be low ( matteucci @xmath9 tosi 1985 ; pilyugin 1994 ; marconi , matteucci , @xmath9 tosi 1994 ) .
more direct information on the past sf activity occurred in dwarf irregulars has been derived for some local group objects by analyzing the stellar population along the giant branches ( gallart 1996a ; aparicio , gallart , @xmath9 bertelli 1997a , b ) .
although the sfrs derived in this way are less detailed and reliable than those deduced from the analysis of the ms and core he - burning stars , there is no hint that these galaxies experienced a much stronger sfr in the past than recently . on the other hand , the analysis of the giant branches does not permit a fair time resolution , and the past activity in these galaxies could also have been characterized by recurrent strong bursts , whose average rate is reflected by the bright portion of the cmd . the cmd derived by vb exhibits a well populated agb , which definitely argues for previous sf in ngc 1569 .
vb concluded that a sf episode took place at a significantly lower rate than the most recent one between 1.5 and 0.15 gyr ago .
however , the values of these parameters are uncertain due to observational and theoretical limitations .
nevertheless , the mere presence of agb stars gives evidence for a previous sf activity at intermediate ages .
ages as old as @xmath010 gyr can be probed only with nicmos observations looking for the presence of stars at the tip of the first red giant branch in the infrared . in the meantime ,
indirect information on the past sf history can be derived from chemical evolution arguments .
a recent determination of the oxygen , nitrogen and helium abundances in the interstellar medium of ngc 1569 was done by kobulnicky & skillman ( 1997 ) . comparing their values to the predictions of chemical evolution models for bursting dwarfs (
marconi , matteucci , @xmath9 tosi 1994 ) we find that several bursts of sf may have happened in this galaxy , provided that strong galactic winds have occurred .
the products of snii explosions should be preferentially lost in these winds , thus keeping the overall metallicity low .
to summarize , the comparison of the sf history in ngc 1569 and that of dwarf irregular galaxies in the local group is inconclusive .
while the general characteristics of the sf appear very similar , the level of the current sfr in ngc 1569 seems significantly larger than that of the dwarf irregulars in the local group over their whole lifetime .
this suggests that ngc 1569 is distinctively more active than the local group dwarf irregulars .
an evolutionary link between dwarf spheroidals and dwarf irregulars has been suggested several times ( e.g. , kormendy 1985 ; binggeli 1994 ) .
strong bursts of sf are known to have occurred in dwarf spheroidals in the past ( e.g. , smecker - hane et al .
unfortunately , to our knowledge a quantitative determination of the level and duration of these bursts has not yet been attempted .
the evidence that ngc 1569 is currently experiencing a galactic outflow seems well documented , an effect which would turn the galaxy into an inconspicuous , perhaps forever fading object . on the other hand ,
the ultimate fate of the expanding gas is rather uncertain , due to the complications of hydrodynamic modelling . a stronger case could perhaps be made by comparing the sfr in the bursts which occurred in the dwarf spheroidals , to what we measure for ngc 1569 .
we now consider the possible relevance of galaxies like ngc 1569 to the interpretation of faint galaxy counts . to do this ,
we first discuss the possible evolutionary path for the sfr in this galaxy .
the direct census of the stellar population in ngc 1569 suggests that a strong burst of sf has occurred at a rate of @xmath00.5 m@xmath6 yr@xmath7 , with a duration of @xmath00.15 gyr .
this level of the sfr can not have been sustained over a long period of time .
the total mass of the galaxy as estimated by israel ( 1988a ) is @xmath03.3 @xmath90 m@xmath6 . considering a sfr of 0.5 m@xmath6 yr@xmath7 as representative of the whole galaxy ( i.e. neglecting the activity in the regions external to our pc field , and considering a turn - over in the imf at the low mass end ) it only takes @xmath00.7 gyr to cycle the entire mass into stars .
the gas exhaustion timescale is longer than this because of the mass return from stellar winds and supernovae .
however , this corresponds to a small correction since 1 gyr after its birth a single age stellar population has returned only @xmath00.2 of its total mass ( for a salpeter imf ) .
thus it seems very unlikely that such a high sfr could have been supported throughout a lifetime of several gyr , even allowing for the occurrence of galactic winds in the past , which would imply a larger initial mass .
furthermore , the metallicity resulting from such extreme activity would inevitably be much larger than derived from the spectra of hii regions .
though the above estimate for the gas exhaustion timescale is sensitive to the adopted lower mass cut - off , the metal production depends on the mass range that we directly observe .
we thus suggest that , in spite of the rather constant sfr characterizing the more recent epochs , the average past sfr in ngc 1569 should have been much lower , perhaps with short intense bursts coupled with long quiescence periods . considering the energy input from the recent burst of sf , vb suggest that at least part of the gas currently present in ngc 1569 will be lost from the galaxy .
we confirm this expectation , since our snii rate is consistent with what found by vb .
however , the burst giving birth to the stars in our observed field did not completely exhaust the gas reservoir , since the sscs and hii regions are younger than the field .
it could be that , while the burst of sf was able to initiate a general outflow of the ism , the highest density clumps succeeded in collapsing to form stars , possibly stimulated by the sn explosions .
the ultimate fate of the outflowing gas depends on several uncertain factors , including dissipation .
in addition , hydrodynamic models for bursting dwarfs suggest that , once the energy input from stellar winds and supernovae has ceased , the interstellar medium in the outer disk of the galaxy falls back to the center , where it could power a successive burst of sf ( dercole @xmath9 brighenti 1998 ) . to summarize , the recent burst could be the strongest sf episode in this galaxy over its entire lifetime , though the future evolution is rather uncertain . among the various models proposed in the literature to explain the excess population of faint blue galaxies , we consider those by gronwall @xmath9 koo ( 1995 ) and babul @xmath9 rees ( 1992 ) . in the model proposed by gronwall @xmath9 koo ( 1995 )
, the blue galaxy population does not evolve .
the luminosity evolution of ngc 1569 is likely to be characterized by on and off phases , corresponding to the recurrent bursts , with long interburst periods .
therefore , if ngc 1569 were a typical object in the class of galaxies mostly contributing to the excess blue objects , a constant luminosity evolution of the population could be accomplished only with an appropriate number evolution of the galaxies of this kind not a very attractive scenario .
in other words , it seems unlikely that ngc 1569 represents the local counterpart of the galaxies mostly contributing to the faint blue counts in the gronwall @xmath9 koo ( 1995 ) model . according to babul @xmath9 rees ( 1992 ) , the excess blue objects are bursting dwarf galaxies ( at redshifts close to @xmath01 ) which , after the burst , become gas - free because of sn explosions .
later evolution is bound to passive fading .
thus , ngc 1569 can not be regarded as a possible descendant of the bursting dwarfs , since it is bursting now .
still , its properties may be relevant to the babul @xmath9 rees model . in a quantitative study of this models , babul @xmath9 ferguson ( 1996 )
show that the current constraints on the faint galaxy population can be met if the sfr in the bursting dwarfs is large , so that the object is bright enough to contribute to the counts at the appropriate redshift .
our derived sfr for the most recent burst in ngc 1569 seems close to the rate of @xmath01 m@xmath6 yr@xmath7 required by their model .
however , babul @xmath9 ferguson adopt a burst duration of 10 myr , one order of magnitude lower than what found here for ngc 1569 .
the success of their model depends on the combination of the duration of the burst , efficiency of star formation , and on the decay rate of the dwarf galaxy formation probability .
it is likely that longer burst durations can be accommodated when varying the other parameters . the model by babul @xmath9 ferguson also requires substantial dimming of the bursting dwarfs after the major burst .
this is needed in order to avoid an excess of objects at brighter magnitudes , where the counts are dominated by the evolution of bigger galaxies . as emphasized before
, we do not know if ngc 1569 will definitely lose its gas .
the fact that ngc 1569 has experienced at least one previous burst and that the most recent one has had a relatively long duration , raises doubts on this hypothesis .
nevertheless , the case of ngc 1569 demonstrates that dwarf galaxies are capable to sustain bursts of sf of the magnitude required by this model .
in this paper we have presented hst wfpc2 f555w and f439w photometry of resolved stars in the dwarf irregular galaxy ngc 1569 .
about 2800 stars were measured with photometric error smaller than = 0.2mag out of a total of @xmath07000 objects detected .
the corresponding cmd extends down to , @xmath2 26 , and is complete for @xmath3 23 .
it samples stars more massive than @xmath04 m@xmath6 allowing us to derive the sf history over the last 0.15 gyr .
the interpretation of the data has been performed via theoretical simulations , based on stellar evolutionary tracks .
we have considered several sets of stellar tracks from the geneva and padova groups , and were able to reach a satisfactory representation of the data adopting either the geneva tracks with @xmath67 = 0.001 , or the padova tracks with @xmath67 = 0.004 .
since the applicability of these tracks is related to the shape of the blue loops , which is sensitive to details in the input parameters used in the computations , we consider it premature to dismiss alternative sets of evolutionary tracks .
the two sets provide a similar picture of the recent sf history in the field of ngc 1569 .
the galaxy has experienced a general burst of sf , of duration @xmath91 gyr , at an approximately constant rate . the burst seems to have stopped @xmath05 10 myr ago but sf continues in the hii regions and in the sscs .
if this general burst consists of successive episodes , these must have occurred at a similar rate , and be separated by short quiescent periods .
qualitatively , this behavior looks similar to that inferred for dwarf irregulars in the local group .
quantitatively , though , the level of the sfr in this recent burst in ngc 1569 is approximately 2 orders of magnitude higher .
we find that the salpeter imf is consistent with the observed cmd and lf , but our best simulations are characterized by slightly steeper slopes ( 2.6 3 ) .
this seems to disagree with observations of other starbursts which generally indicate a salpeter slope ( leitherer 1997 ) .
the present study , however , refers to the field population of stars less massive than @xmath030 m@xmath6 , whereas most starburst imfs are derived for the nuclear burst and for dense ob clusters in the mass range above @xmath020 m@xmath6 . since most of the sf activity in the ngc 1569 field has subsided about 7
10 myr ago , we have no direct information on the imf of those stars with lifetimes less than this value .
the corresponding minimum masses are around 30 m@xmath6 .
the field - star imf in our galaxy and in the magellanic clouds is steeper ( biased against massive stars ) than the cluster imf ( massey , johnson , @xmath9 degioia - eastwood 1995a ; massey 1995b ) , with the cluster value being close to salpeter . this could be due to a richness effect , where the most massive stars are not observed simply because of their small expected numbers .
ngc 1569 may be another case where the most massive stars with masses above @xmath030 m@xmath6 are preferentially formed in clusters , but we emphasize the different mass range sampled by our and that of lmc / smc studies .
our preferred imf is quite similar to that derived for field stars of comparable mass in our galaxy .
this is in agreement with the notion of a virtually constant imf , at least as far as the shape is concerned , and does not support the expectations of flatter slopes in starbursting regions ( e.g. , padoan , nordlund , @xmath9 jones 1997 ) , or in low metallicity environments , as sometimes invoked to better reproduce the properties of elliptical galaxies ( e.g. , vazdekis 1996 ) . in the past
, the sf in ngc 1569 is likely to have proceeded at a substantially lower average rate than in the recent burst .
this follows from the estimate of the gas exhaustion timescale , and from the relatively low metallicity of the ism in this galaxy .
this leaves the possibilities of either an approximately constant sfr , or short strong bursts , with long interburst periods .
detailed chemical evolution models are needed to explore quantitatively the possible evolutionary paths for ngc 1569 , and this will be the subject of a forthcoming paper .
the recent burst of sf is certainly not the first episode in ngc 1569 ( see also vb ) but may be the last .
the velocity field of the ionized gas ( tomita , ohta , @xmath9 saito 1994 ) and energetics arguments support this possibility ( heckman 1995 ) .
if this was the case , in the future ngc 1569 may turn into a dwarf spheroidal galaxy .
in addition , its luminosity evolution would be characterized by passive fading , which seems required by the babul @xmath9 ferguson ( 1996 ) bursting dwarfs model to explain the excess faint galaxy counts .
however , it is also possible that the outflowing gas falls back onto the galaxy , triggering a successive burst of sf , an option which has to be explored with the computation of detailed hydrodynamic modelling for bursting dwarf galaxies ( dercole @xmath9 brighenti 1998 ) . in spite of these uncertainties
, our results shows that dwarf galaxies are capable of sustaining bursts of sf at large enough rates to be relevant for the interpretation of the faint galaxy counts .
this work has been partially supported by asi through grant 9770070as .
support for this work was also provided by nasa grant go-06111.01 - 94a from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555
. l.g . and m.t .
gratefully acknowledge the hospitality of stsci .
aparicio , a. , gallart , c. , & bertelli , g. , 1997a , , 114 , 669 . 1997b , , 114 , 680 arp , h. , & sandage , a. 1985 , , 90 , 1163 babul , a. , & ferguson , h. c. 1996 , , 458 , 100 babul , a. , & rees , m. j. 1992 , , 255 , 346 bertelli , p. , mateo , m. , chiosi , c. , & bressan , a. 1992 , , 388 , 400 binggeli , b. 1994 , in dwarf galaxies , ed .
g. meylan & p. prugniel ( garching : eso ) , 13 biretta , j. 1996 , wfpc2 instrument handbook ( baltimore : stsci ) bresolin , f. , capaccioli , m. , & piotto , g. 1993 , , 105 , 1779 calzetti , d. , kinney , a. , & storchi - bergmann , t. 1994 , , 429 , 582 charbonnel , c. , meynet , a. , maeder , a. , schaller , g. , & schaerer , d. 1993 , a&as , 101 , 415 dercole , a. , & brighenti , f. 1998 , in preparation devost , d. , roy , j .-
r . , & drissen , l. 1997 , , 482 , 765 della ceca , r. , griffiths , r. e. , heckman , t. m. , & mackenty , j. w. 1996 , , 469 , 662 de marchi , g. , clampin , m. , greggio , l. , leitherer , c. , nota , a. , & tosi , m. 1997 , , 479 , l27 de young , d. , & heckman , t. m. 1994 , , 431 , 589 drissen , l. , & roy , j .- r .
1994 , , 106 , 974 fagotto , f. , bressan , a. , bertelli , g. , & chiosi , c. 1994 , a&as , 105 , 29 freedman , w. l. 1985 , , 299 , 74 . 1988 , , 96 , 1248 gallagher iii , j. s. 1996 , in from stars to galaxies : the impact of stellar physics on galaxy evolution , ed . c. leitherer , u. fritze - v .
alvensleben , & j. huchra ( san francisco : asp ) , 315 gallagher iii , j. s. , & hunter , d. a. 1984 , araa , 22 , 37 gallagher iii , j. s. , hunter , d. a. , & tutukov , a. v. 1984 , , 284 , 544 gallart , c. , aparicio , a. , bertelli , g. , & chiosi , c. 1996a , , 112 , 1950 gallart , c. , aparicio , a. , & vlchez , j. m. 1996b , , 112 , 1928 gonzlez - delgado , r. m. , leitherer , c. , heckman , t. m. , & cervio , m. 1997 , , 483 , 705 gould , a. , bahcall , j. n. , & flynn , c. 1997 , , 482 , 913 greggio , l. 1994 , in the local group : comparative and global properties , ed .
a. layden , r. c. smith , & j. storm ( garching : eso ) , 72 greggio , l. , marconi , g. , tosi , m. , & focardi , p. 1993
, , 105 , 894 ( gmtf ) gronwall , c. , & koo , d. c. 1995 , , 440 , l1 heckman , t. m. 1995 , in the interplay between massive star formation , the ism , and galaxy evolution , ed .
d. kunth , b. guiderdoni , m. heydari - malayeri , & t. x. thuan ( gif - sur - yvette : editions frontieres ) , 159 heckman , t. m. , armus , l. , & miley , g. k. 1990 , apjs , 74 , 833 heckman , t. m. , dahlem , m. , lehnert , m. d. , fabbiano , g. , gilmore , d. , & waller , w. h. 1995 , apj , 448 , 98 ho , l. c. , & filippenko , a. v. 1996 , , 466 , l83 hoessel , j. g. 1986 , in luminous stars and associations in galaxies , ed . c. w. h. de loore , a. j. willis , & p. laskarides ( dordrecht : reidel ) , 439 holtzman , j. , burrows , c. , casertano , s. , hester , j. , trauger , j. , watson , a. , & worthey , g. 1995 , , 107 , 1065 hunter , d. a. , & gallagher iii , j. s. 1986 , , 98 , 5 hunter , d. a , gallagher , j. s. , & rautenkranz , d. 1982 , , 49 , 53 israel , f. p. 1988a
, , 194 , 24 . 1988b , , 198 , 109 johnson , h. l. 1966 , , 4 , 193 kobulnicky , h. a. , & skillman , e. d. 1997 , , 489 , 636 kormendy , j. 1985 , , 295 , 73 kurucz , r. l. 1992 , in the stellar populations of galaxies , ed . b. barbuy & a. renzini ( dordrecht : kluwer ) , 255 leitherer , c. 1997 , in the stellar initial mass function , ed . g. gilmore ( san francisco : asp ) , in press leitherer , c. , robert , c. , & drissen , l. 1992 , apj , 401 , 596 lejeune , t. , cuisinier , f. , & buser , e. 1997 , a&as , 125 , 229 marconi , g. , matteucci , f. , & tosi , m. 1994 , , 270 , 35 marconi , g. , tosi , m. , greggio , l. , & focardi , p. 1995
, , 109 , 173 ( mtgf ) marlowe , a. t. 1997 , ph.d .
thesis ( baltimore : the johns hopkins university ) martin , c. l. 1997 , , 491 , 561 massey , p. , johnson , k. e. , & degioia - eastwood , k. 1995a , , 454 , 151 massey , p. , lang , c. c. , degioia - eastwood , k. , & garmany , c. d. 1995b , , 438 , 188 matteucci , f. , & tosi , m. 1985 , , 217 , 391 melnick , j. , moles , m. , & terlevich , r. 1985 , , 149 , l24 meurer , g. 1995 , nature , 375 , 742 meynet , a. , maeder , a. , schaller , g. , schaerer , d. , & charbonnel , c. 1994 , a&as , 103 , 97 oconnell , r. w. , gallagher , j. s. , & hunter , d. a. 1994 , apj , 433 , 65 padoan , p. , nordlund , a. p. , & jones , b. j. t. 1997 , , 288 , 145 pilyugin , l. s. 1994 , , 287 , 387 renzini , a. , greggio , l. , ritossa , c. , & ferrario , l. 1992 , , 400 , 280 ritossa , c. 1996 , , 281 , 970 sandage , a. , & brucato , r. 1979 , aj , 84 , 42 scalo , j. m. 1986 , fund .
phys . , 11 , 1 schaerer , d. , meynet , a. , maeder , a. , & schaller , g. 1993 , a&as , 98 , 523 schaller , g. , schaerer , d. , meynet , a. , & maeder , a. 1992 , a&as , 96 , 269 silk , j. 1987 , in high redshift and primeval galaxies , ed .
j. bergeron , d. kunth , b. rocca - volmerange , & j. tran than van ( gif - sur - yvette : editions frontieres ) , 301 smecker - hane , t. a. , stetson , p. b. , hesser , j. e. , & lehnert , m. d. 1994 , , 108 , 507 suchkov , a. a. , balsara , d. s. , heckman , t. m. , & leitherer , c. 1994 , apj , 430 , 511 tammann , g. 1994 , in dwarf galaxies , ed . g. meylan & p. prugniel , ( garching : eso ) , 3 tenorio - tagle , g. , & muoz - tun , c. 1997 , apj , 478 , 134 terlevich , r. , melnick , j. , masegosa , j. , moles , m. , & copetti , m. v. f. 1991 , a&as , 91 , 285 thuan , t. x. 1991 , in massive stars in starbursts , ed . c. leitherer , n. walborn , t. heckman , & c. norman ( cambridge : cup ) , 183 timmes , f. x. , woosley , s. e. , & weaver , t. a. 1995 , , 98 , 617 tomita , a. , ohta , k. , & saito , m. 1994 , , 295 , 73 tolstoy , e. 1996 , , 462 , 684 tolstoy , e. , & saha , a. 1996 , , 462 , 672 tosi , m. 1994 , in dwarf galaxies , ed . g. meylan & p. prugniel ( garching : eso ) , 465 tosi , m. , greggio , l. , marconi , g. , & focardi , p. 1991
, , 102 , 951 ( tgmf ) vallenari , a. , & bomans , d. j. 1996 , , 313 , 713 ( vb ) van den bergh , s. 1977 , in the evolution of galaxies and stellar populations , ed . b. m. tinsley & r. b. larson ( new haven : yale ) , 19 vazdekis , a. , casuso , e. , peletier , r. f. , @xmath9 beckman , j. e. 1996 , , 106 , 307 walborn , n. r. 1991 , in massive stars in starbursts , ed . c. leitherer , n. walborn , t. m. heckman , & c. a. norman ( cambridge : cup ) , 145 waller , w. 1991 , , 370 , 144 wirth , a. , & gallagher iii , j. s. 1984 , apj , 282 , 85 | we present wfpc2hst photometry of the resolved stellar population in the post - starburst galaxy ngc 1569 .
the color - magnitude diagram ( cmd ) derived in the f439w and f555w photometric bands contains @xmath02800 stars with photometric error @xmath1 0.2 mag down to , @xmath2 26 , and is complete for @xmath3 23 . adopting the literature distance modulus and reddening , our cmd samples stars more massive than @xmath04 , allowing us to study the star formation ( sf ) history over the last @xmath00.15 gyr .
the data are interpreted using theoretical simulations based on stellar evolutionary models .
the synthetic diagrams include photometric errors and incompleteness factors .
testing various sets of tracks , we find that the ability of the models to reproduce the observed features in the cmd is strictly related to the shape of the blue loops of the sequences with masses around 5 .
the field of ngc 1569 has experienced a global burst of star formation of duration @xmath40.1 gyr , ending @xmath0 5@xmath510 myr ago . during the burst , the sf rate was approximately constant , and , if quiescent periods occurred , they lasted less than @xmath010 myr
. the level of the sf rate was very high : for a single - slope initial mass function ( imf ) ranging from 0.1 to 120 we find values of 3 , 1 , and 0.5 m@xmath6 yr@xmath7 for @xmath8 , 2.6 , and 2.35 ( salpeter ) , respectively .
when scaled for the surveyed area , these rates are approximately 100 times larger than found in the most active dwarf irregulars in the local group .
the data are consistent with a salpeter imf , though our best models indicate slightly steeper exponents .
we discuss the implications of our results in the general context of the evolution of dwarf galaxies . |
recently , there are remarkable developments in experiments for the nucleon electromagnetic form factors : + 1 ) for the space - like momentum , the ratio of the electric and magnetic form factors of proton , @xmath0 and @xmath1 respectively , was shown to be a decreasing function of the squared momentum transfer @xmath2 and the experimental results imply that the proton electric form factor vanishes for @xmath3 @xcite-@xcite .
+ 2 ) for the neutron magnetic form factor , @xmath4 , very accurate experimental data were obtained and it approximately satisfies @xmath5 , with @xmath6 being represented in terms of gev / c , for fairly wide range of squared momentum transfer @xmath7 ( gev / c)@xmath8 @xcite , @xcite ( clas collaboration ) .
+ 3 ) for the time - like momentum the ratio @xmath9 was obtained @xcite , ( babar collaboration ) , while previously the data of form factors had been analyzed under the assumption @xmath10 or @xmath11 .
+ asymptotically , the experimental data of nucleon magnetic factors decrease more rapidly than the dipole formula for large @xmath2 and the decrease has been understood as a realization of perturbative qcd @xcite , the behavior of which can be formulated in terms of the dispersion theory with appropriate conditions on the absorptive parts ; we assumed unsbtracted dispersion relations for the charge and magnetic moment form factors .
to realize the asymptotic form of qcd we imposed the superconvergence conditions .
as the data for the time - like momentum have become accurate , it is necessary to investigate the form factors for the space - like and time - like momentums systematically . for this purpose
the dispersion theory is effective .
the dispersion theoretical calculations performed so far , the value of @xmath4 turned out to be larger than the above mentioned new experimental data for @xmath7 ( gev / c)@xmath8 ( see ref .
it is of vital importance to investigate if it is possible to realize the experimental data simply by the adjustment of parameters or by the refinement of absorptive parts in the dispersion relation .
it is the purpose of this paper to analyze experimental data of nucleon form factors by the dispersion theory , with the qcd constraints imposed , taking account of the above mentioned new experimental results .
organization of the paper is given as follows : in sec .
2 we explain the superconvergent dispersion relation and give conditions which are used in this paper .
we summarize the absorptive parts , which are broken up into three parts : low , intermediate and asymptotic momentum regions . for each momentum region
the imaginary parts are given .
the asymptotic part is expressed as an expansion in terms of the analytically regularized running coupling constant in the renormalization group for qcd . in sec .
3 we remark on the numerical analysis . in sec .
4 numerical results are summarized .
the final section is devoted to general discussions .
we assume the unsubtracted dispersion relations for the charge and magnetic moment form factors , @xmath12 and @xmath13 , respectively , with @xmath14 denoting the isospin state @xmath15 .
that is , @xmath16 where the threshold is @xmath17 . here
@xmath18 is the pion mass being taken as the average of the neutral and charged pion masses .
we impose conditions on @xmath19 to realize the qcd conditions .
experimental data imply that the magnetic form factors of nucleon decrease more rapidly than the dipole formula for large squared momentum transfer .
the decrease agrees with the prediction of perturbative qcd , where magnetic form factors of nucleon decrease for @xmath20 as @xmath21 where @xmath22 is the running coupling constant of qcd and @xmath23 with @xmath24 being the number of flavor .
@xmath25 is the qcd scale parameter having the dimension of momentum . to realize the qcd predictions we impose the following conditions on the charge and magnetic moment form factors : @xmath26 , \nonumber \\
f_2(q^2)&\to & { \rm const}/[q^4 ( \ln q^2/{\lambda}^2)^{\gamma } ] , \label{qcd12}\end{aligned}\ ] ] for @xmath27 with @xmath28 .
we briefly summarize the asymptotic theorems which are used to incorporate the constraints of qcd @xcite , where the proof is given in ref .
let @xmath29 satisfy the dispersion relation ( [ unsubtracted ] ) , and @xmath30 is given as @xmath31^{\gamma+1 } } + o\left(\frac{1}{[\ln(t'/\lambda^2)]^{\gamma+2}}\right)\ ] ] for @xmath32 with @xmath33 .
then @xmath29 becomes @xmath34^{\gamma+1}}\\ \nonumber & \to & \frac{c}{\pi\gamma\ln(|t|/\lambda^2)]^{\gamma } } \label{disp1}\end{aligned}\ ] ] for @xmath35 .
generally , when @xmath36 satisfies @xmath37^{\gamma+1 } } + o\left(\frac{1}{[\ln(t'/\lambda^2)]^{\gamma+2}}\right ) \label{asymp2}\ ] ] for @xmath38 and the superconvergence conditions @xmath39 @xmath29 given by ( [ unsubtracted ] ) approaches for @xmath40 to the following formula : @xmath41 which can be proved by using ( [ disp1 ] ) and ( [ asym4 ] ) together with the identity @xmath42 indeed , by using ( [ asym4 ] ) we have @xmath43 which leads to ( [ asym5 ] ) as @xmath44 satisfies ( [ asymp2 ] ) . to obtain the asymptotic formulas ( [ qcd12 ] ) , therefore , we impose the superconvergence conditions on the imaginary part of form factors , @xmath45 ( @xmath46 ; @xmath14 denotes isospin ) in the unsubtracted dispersion relation ( [ unsubtracted ] ) : + @xmath47where @xmath48 satisfies the asymptotic conditions for @xmath49 @xmath50^{\gamma+1 } \quad ( i=1,2 ) .
\label{asymptotic}\ ] ] in addition to the conditions ( [ sup1 ] ) and ( [ sup2 ] ) we impose the normalization conditions at @xmath51 : @xmath52where @xmath53 is the anomalous magnetic moments of nucleons with the isospin @xmath14 . let us discuss the imaginary parts of nucleon form factors , which are broken up into three parts : the low momentum , the intermediate , and the asymptotic regions .
the imaginary parts of the charge and magnetic moment form factors , @xmath54 , are given in terms of two pion contribution as follows : @xmath55 & = & \frac{m}{2}\frac{(t-4\mu^2)}{4m^2-t } \left(\frac{t-4\mu^2}{t}\right)^{1/2}\\% \nonumber \\ & \times & { \rm re}\big[m^{*}(t ) \big\{f_{+}^{(-)1}(t ) -\frac{t}{4m^2}\frac{m}{\sqrt 2}f_{-}^{(-)1}(t)\big\}\big],% \\\end{aligned}\]]@xmath56 & = & \frac{m}{2}\frac{(t-4\mu^2)}{(4m^2-t ) } \left(\frac{t-4\mu^2}{t}\right)^{1/2 } \label{imh}\\% \nonumber\\ & \times & { \rm re}\big[m^{*}\big\{\frac{m}{\sqrt 2}f_{-}^{(-)1}(t ) -f_{+}^{(-)1}(t)\big\}\big ] , \nonumber\end{aligned}\]]where @xmath57 are helicty amplitudes for @xmath58 , @xmath59 is the pion form factor and @xmath18 is the pion mass .
the superscript @xmath60 denotes the iso - vector part . for the helicity amplitudes we use the numerical values given by hhler and schopper @xcite and parameterize @xmath59 according to them . @xmath61^{-1},\ ] ] where @xmath62 and @xmath63 are the @xmath64 meson mass and width respectively and @xmath65 the imaginary parts thus obtained are denoted as @xmath66 hereafter .
it must be remarked that the @xmath64 meson contribution is included in the helicity amplitudes of ref .
the uncorrelated kaon pair is neglected here as the effect was estimated to be small @xcite .
the intermediate states @xmath67 are approximated by the addition of the breit - wigner terms , with the imaginary part parameterized as follow : @xmath68 where @xmath69 here @xmath70 and @xmath71 are the mass and width of resonance , respectively , the threshold @xmath72 is @xmath17 and @xmath73 is treated as an adjustable parameter .
g is introduced to cut - off the breit - wigner formula .
we write the intermediate part as the summation of resonances @xmath74 where @xmath14 is the isospin and @xmath75 is the labeling of resonances ( see table i ) . here
the suffix @xmath76 denotes @xmath46 , corresponding to the charge and magnetic moment form factors @xmath77 and @xmath78 ( @xmath79 = n or p ) .
the same formulas for @xmath80 are used for @xmath81 and @xmath82 .
we express the form factors as power series in the running coupling constant of qcd , @xmath22 . to calculate the absorptive part
, it is necessary to perform analytic continuation to the time - like momentum .
here we give only the necessary procedure for the analytic continuation of the running coupling constant to the time - like momentum by using the analytic regularization @xcite @xcite , as the formulation is given in ref .
@xcite .
let @xmath83 be the running coupling constant in the renormalization group calculated by the perturbative qcd as the function of the squared momentum @xmath2 for the space - like momentum .
we use the three loop approximation for @xmath83 , which is expressed in the pad form .
@xmath84^{-1}. \label{pqcd}\ ] ] @xmath85 is the qcd scale parameter , and @xmath86 are expressed in terms of the @xmath87 function of qcd , @xmath88 where @xmath89 with @xmath24 being the number of flavor .
we perform the analytic continuation of the squared momentum to the time - like region , @xmath90 , by the replacement in ( [ pqcd ] ) @xmath91 then @xmath92 becomes complex and is expressed as follows : @xmath93where @xmath94 and @xmath95 are given as @xmath96 \nonumber \\ & & + \frac{a_3\ln(s/\lambda^2)}{\ln^2(s/\lambda^2)+\pi^2 } , % \\ 2010 - 01 - 29 check\end{aligned}\]]@xmath97 \nonumber \\ & & -\frac{\pi a_3}{\ln\{\ln^2(s/\lambda^2)+\pi^2\}},\end{aligned}\]]with @xmath98 the running coupling constant is given by the dispersion integral both for the space - like and the time - like momentum @xmath99 with @xmath100 @xmath101 represented by ( [ regular ] ) is called analytically regularized running coupling constant as it has no singular point for @xmath102 .
the regularization eliminates the ghost pole of @xmath83 , given by ( [ pqcd ] ) , appearing at @xmath103 where @xmath104 for the number of flavor @xmath105 .
calculating ( [ regular ] ) , we find that @xmath101 is approximately given by the simple formula with the ghost pole subtracted @xmath106 where the residue @xmath107 is @xmath108 we use ( [ regular-1 ] ) as the regularized coupling constant ; for the time - like momentum we replace @xmath109 in ( [ regular-1 ] ) as was mentioned before .
the qcd parts , @xmath110 ( @xmath111 ; @xmath14 = 0,1 ) for the squared time - like momentum , are written as follows : @xmath112 where @xmath113 s are given as expansion in terms of the running coupling constant @xmath114 for the time - like squared momentum @xmath90 .
we multiply by the function @xmath115 in ( [ qcd1 ] ) to assure the convergence of the superconvergence conditions ( [ sup1 ] ) and ( [ sup2 ] ) .
the following formula is assumed for @xmath116 : @xmath117 which may be interpreted as the form factor for @xmath118 with @xmath119 being the threshold of the quark antiquark pair .
the parameters @xmath119 , @xmath120 and @xmath121 are taken as adjustable parameters and will be determined by the analysis of experimental data . for the time - like momentum , we perform the analytic continuation of the regularized effective coupling constant @xmath122 to @xmath123 through the equation @xmath124 + i\,{\rm im}[\alpha_r(s)].\ ] ] we express the qcd part as the power series expansion in @xmath123 @xmath125 the summation in ( [ cqcd ] ) begins in the second order in the effective coupling constant so as to realize the logarithmic decrease of the nucleon form factors
. imaginary part of ( [ cqcd ] ) is obtained to be @xmath126 \nonumber \\ & & \quad+c_{i,4}^{qcd,\,i}[4({\rm re}\,\alpha_r)^3{\rm im}\,\alpha_r -4{\rm re}\,\alpha_r({\rm im}\,\alpha_r)^3 ] \nonumber \\ & & \quad+\cdots , \label{imqcd}\end{aligned}\]]and @xmath127 we write the low energy part , intermediate resonance part and asymptotic qcd parts of form factors as @xmath128 , @xmath129 and @xmath110 , respectively , which are given by the dispersion integral with the imaginary parts ( [ imh ] ) , ( [ imbw ] ) and ( [ imqcd1 ] ) .
the form factors @xmath130 are defined by adding them up .
we impose the conditions ( [ sup1 ] ) and ( [ sup2 ] ) on @xmath131 so that the qcd conditions are satisfied .
we analyzed the experimental data of nucleon electromagnetic form factors @xmath132 , @xmath133 , @xmath134 @xmath135 and the ratio @xmath136 for the space - like momentum transfer , and @xmath137 and @xmath138 in refs .
@xcite- @xcite for the time - like momentum transfer and the above mentioned recent experimental data @xmath4 for the space - ike and @xmath139 for the time - like momentum transfer .
the parameters appearing in the formulas are determined so as to minimize @xmath140 .
as was mentioned in the introduction we analyze by taking account of the recent experimental data : ( a ) @xmath4 for @xmath141 ( gev / c)@xmath8 ( clas collaboration ) and ( b ) @xmath142 ( babar collaboration ) . in order to see how the situation changes by taking account of these new experiments in addition to the other data , we perform analysis for the following two cases in the @xmath140 analysis : + case i : both of the experimental data , ( a ) @xmath143 for the time - like momentum and ( b ) new data for @xmath4 for the space - like momentum , are added .
+ case ii : only the data ( a ) @xmath143 for the time - like momentum are added .
+ let us remark on the experiments for the time - like momentum @xcite , @xcite , @xcite , where the form factors @xmath137 and @xmath138 are determined by using the formula for the cross section @xmath144 for the processes @xmath145 or @xmath146 , which is given as @xmath147 here @xmath148 is the fine structure constant and @xmath149 is the nucleon velocity .
@xmath150 are estimated from @xmath151 under the assumption @xmath152 or @xmath153 .
@xmath144 is expressed in terms of @xmath154 and @xmath155 as follows : @xmath156 equating ( [ exp ] ) and ( [ theory ] ) , we have @xmath157 substituting our calculated result of form factors to the right hand side of ( [ timedata ] ) , we obtain the theoretical value for @xmath151 , which is compared with the experimental data for the magnetic form factor obtained under the assumption @xmath152 .
the parameters appearing in our analysis are the following : residues at resonances , coefficients appearing in the expansion by the qcd effective coupling constants , cut - offs for the intermediate region @xmath158 .
in addition to them we have parameters in the breit - wigner formula and the convergence factor @xmath159 of qcd contribution , @xmath119 , @xmath73 , @xmath160 .
we have taken the masses and the widths of resonances as adjustable parameters . as the superconvergence constraints impose very stringent conditions on the form factors , it was necessary to take the masses and widths as parameters .
we give in tables 1 , 2 and 3 the results for the parameters for the cases i and ii obtained by the @xmath140 analysis ; in table 1 the masses and widths of resonances and in table 2 residues at resonance poles and in table 3 the coefficients @xmath161 @xmath162 in the expansion in terms of the effective coupling constant @xmath163 of qcd defined by ( [ cqcd ] ) .
the number of flavor is taken as @xmath164 .
@xmath165 is cut - off at @xmath166 gev@xmath8 and the breit - wigner formulas at @xmath158 = 26.0 gev .
the qcd parameter is fixed at @xmath85 = 0.216 gev .
the other parameters are determined as follows : + case i : @xmath72 = 4@xmath167 , @xmath120 = 0.243@xmath168 gev@xmath8 , @xmath121 = 0.237@xmath168 gev@xmath8 , @xmath73 = 0.2260@xmath169 gev@xmath8 , @xmath119 = 0.202@xmath170 gev@xmath8 .
+ case ii : the same as in the case i except for @xmath73 = 0.2253@xmath169 gev@xmath8 .
the value of @xmath140 is obtained to be @xmath171 for the case i and @xmath172 for the case ii , which includes both the data of space - like and time - like regions .
the total number of data is 245 for the case i and 236 for the case ii .
number of parameters is 36 so that dof/@xmath173 = 1.88 for the case i and 1.54 for the case ii .
.masses and widths determined by the @xmath140 analysis for the cases i and ii.@xmath174 [ cols="^,^,^,<,^,^,<,<",options="header " , ] we illustrate in figs . 1 - 9 the calculated results for the form factors . the results for the case i is given by the solid curve and case ii by the dashed one .
figs . 1 - 4
the results for the space - like momentum are illustrated : fig .
1 the proton magnetic form factors @xmath132 , fig . 2 proton electric form factor @xmath133 , fig . 3 the neutron magnetic form factor @xmath175 and fig . 4 the neutron electric form factor . in fig .
5 we illustrate the ratio of proton electric and proton magnetic form factors @xmath176 . we find that @xmath10 at @xmath177 ( gev / c)@xmath8 for the case i and @xmath178 ( gev / c)@xmath8 for the case ii .
the form factor for the time - like momentum @xmath151 is given in fig .
6 for the proton and in fig . 7 for the neutron .
the result for the proton form factor agrees with the experimental data , but for the neutron the calculated one becomes larger than the experiments for large @xmath2 . in fig .
8 we compare the calculated result for the neutron magnetic form factor @xmath134 with the recent experiments .
the solid curve agrees with the experimental data very well .
the dashed one becomes a little larger than the result obtained by the clas collaboration .
however , the deviation is not very large . in fig .
9 we illustrate the result for @xmath179 for the time - like momentum .
there seems to be some discrepancy between the experimental data : the ratio obtained by bardin el al .
@xcite is smaller than that of aubert et al . .
our result coincides with the result of bardin et al . for small @xmath2 and that of aubert el al . for large @xmath2 .
the solid curve is the result for case i and the dashed one for the case ii . ] for the time - like momentum .
the solid curve is the result for case i and the dashed one for the case ii . ]
the experimental data for the neutron magnetic form factor for the space - like momentum with @xmath180 ( gev / c)@xmath8 @xcite , mentioned in sec . 1 , are reproduced very well by our calculation .
the absorptive parts of the form factors for the asymptotic region are approximated by the power series in the effective coupling constant of qcd , which begins @xmath181 as is given in ( [ qcd2 ] ) .
we have taken three terms in the expansions ; the terms of order up to @xmath182 are necessary to reproduce the experiments as in the case of deep inelastic electron scattering processes .
it is remarked here that the electromagnetic form factor of bosons , both for the space - like and time - like momentums , can be explained with recourse to the superconvergent dispersion relation with the qcd constraints @xcite . for the electric form factor of proton
there are deviation of the dispersion theoretical calculation from the experimental data for large @xmath2 , where the data were obtained by using the rosenbluth formula .
the discrepancy may imply the necessity of correction of two photon processes to the experimental data @xcite @xcite .
we used the experimental data for the helicity amplitudes obtained by hher and schopper in which the contribution from the @xmath64 meson is included .
as their data are limited to low @xmath183 ( gev / c)@xmath8 ) , we do not have sufficient data for the region @xmath184 .
we supplemented the unphysical region for @xmath185 state by introducing vector bosons with the mass , @xmath186 gev / c@xmath8 .
for the isoscalar state we also introduced a vector boson with the mass about @xmath187 gev / c@xmath8 .
in our calculation we treated all of the vector boson masses and widths as parameters .
if they are kept at experimental values , we get poor results .
the superconvergence conditions are so strong that the value of @xmath140 is very sensitive to the mass and width .
the masses are obtained to be smaller than the experimental value and the existence of vector bosons with the masses around 1.2 @xmath188 1.4 gev / c@xmath8 are necessary both for the @xmath189 and @xmath190 states . to conclude the paper we remark on the mass around 1.2 gev / c@xmath8 .
we have introduced the vector boson to supplement the lack of information on the the small @xmath2 .
however , both for @xmath191 and @xmath185 states there are indications of resonances observed by the processes @xmath192 , @xmath193 and @xmath194 @xcite .
incorporation of further resonances may improve results for the time - like momentum .
+ the authors wish to express gratitude to professor m. ishida for the valuable discussions and comments .
we also would like to thank dr .
t. komada for the information on the vector bosons with the mass around 1.2 gev / c@xmath8 .
jones et al . , phys .
lett * 84 * , 1398 ( 2000 ) .
o. gayou et al .
lett . * 88 * , 092301 ( 2002 ) .
g. maclachlan et al .
* a764 * , 261 ( 2006 ) .
crawford et al . , phys .
lett . * 98 * , 052301 ( 2007 ) .
g. ron et al .
lett . * 99 * , 202002 ( 2007 ) .
b. anderson et al . , phys .
c * 75 * , 034003 ( 2007 ) .
j. lachniet et al . , phys .
lett . * 102 * , 192001 ( 2009 ) .
g. bardin et al . , nucl . phsy . *
b411 * , 3 ( 1994 ) .
b. aubert et al .
d * 73 * , 012005 ( 2006 ) ; hep - ex/0512020 .
s.j . brodsky and g.r .
farrar , phys .
lett . * 31 * , 1153 ( 1973 ) ; phys .
d * 11 * , 1309 ( 1975 ) ; g.p .
lepage and s.j .
brodsky , phys .
d * 22 * , 2157 ( 1980 ) ; v.l .
chernyak and i.r .
* b246 * , 52 ( 1984 ) . h. f. jones and i.l .
solovstov , phys . lett . *
b349 * , 519 ( 1995 ) .
yu.l . dokshitzer and b.r .
webber , phys .
b352 * , 451 ( 1995 ) .
dokshitzer , g. marchesini and b.r .
webber , nucl .
* b469 * , 93 ( 1996 ) . s. furuichi and k. watanabe , prog .
82 * , 581 ( 1989 ) ; p. mergell , ulf - g .
meiner , d. drechsel , nucl .
a*596 * , 367 ( 1996 ) .
belushikin , h .- w .
hammer and u .- g.meissner , phys .
rev , c * 75 * , 035202 ( 2007 ) ; hep - ph/0608337 .
g. hhler and h.h .
schopper , landolt - brnstein , i/9b 2 , ( 1983 ) .
s. furuichi and k. watanabe , prog .
. phys . * 92 * , 339 ( 1994 ) .
m. nakagawa and k. watababe , nouvo cimento * 112 * a , 873 ( 1999 ) ; phys .
c * 61 * , 055207 ( 2000 ) ; m. nakagawa , h. ishikawa and k. watanabe , acta phyica polonica b * 31 * , 2539 ( 2000 ) ; arxiv:0809.3334 .
bosted et al .
, phys . rev .
c * 42 * , 38 ( 1990 ) . s. rock et al . , phys .
* 49 * , 1139 ( 1982 ) .
h. anklin et al . , phys . lett . * b428 * , 248 ( 1998 ) . m. meyerhoff et al . , phys . lett . * b327 * , 201 ( 1994 ) .
j. becker et al .
j. a * 6 * , 329 ( 1999 ) . c. herberg et al .
j a * 5 * , 131 ( 1999 ) .
i. passchier et al . , phys .
lett . * 82 * , 4988 ( 1999 ) .
m. ostrick et al .
lett * 83 * , 276 ( 1999 ) . c. herberg et al .
j a * 5 * , 131 ( 1999 ) .
g. kubon et al . , phys . lett . *
b524 * , 26 ( 2002 ) .
w. xu et al . , phys .
c * 67 * , 012201(r ) ( 2003 ) .
d. rohe et al .
lett . * 83 * , 4257 ( 1999 ) .
h. zhu et al . , phys .
* 87 * , 081801 ( 2001 ) .
m. ablikim et al . ,
b630 * , 14 ( 2005 ) .
a. antonelli et al . , nucl .
b517 * , 3 ( 1998 ) .
f. borkowski et al . , nucl . phys . *
b93 * , 461 ( 1975 ) .
bosted et al . ,
68 * , 3841 ( 1992 ) .
berger et al . , phys . lett . *
b35 * , 87 ( 1971 ) .
price et al . , phys .
d * 4 * , 45 ( 1971 ) .
w. bartel et al . , nucl .
* b58 * , 429 ( 1973 ) .
a.f . sill et al . , phys .
d * 48 * , 29 ( 1993 ) .
r. madey et al .
lett . * 91 * , 122002 ( 2003 ) .
d. borisyuk and a. kobushkin , phys .
c * 76 * , 022201(r ) ( 2007 ) .
belushkin , h .- w . hammer and ulf - g .
meissner , phys .
b658 * , 138 ( 2008 ) . on the vector boson with the mass around 1.3 gev@xmath8 ,
see t. komada , hep - ph/0612339 , ( kek - proceedings , p93 ) ; w .-
yao et al .
( particle data group ) , j. phys .
g * 33 * , 1 ( 2006 ) . | elastic electromagnetic form factors of nucleons are investigated both for the time - like and the space - like momentums by using the unsubtracted dispersion relation with qcd constraints .
it is shown that the calculated form factors reproduce the experimental data reasonably well ; they agree with recent experimental data for the neutron magnetic form factors for the space - like data obtained by the clas collaboration and are compatible with the ratio of the electric and magnetic form factors for the time - like momentum obtained by the babar collaboration.x pacs no .
13.40.gp , 12.38.-tm , 11.55.fv |
as a function of @xmath22 . the critical value @xmath23 is indicated with a dotted line.,width=264 ] the existence of the seap , which is obtained from fourier transformation of poisson s equation with the quasi - stationary electron density perturbation deduced from the linearized continuity and generalized momentum equation @xcite for non - relativistic , degenerate electrons in a dense quantum plasma , critically depends on the electron number density through the parameter @xmath24 , where @xmath25 .
the parameter @xmath26 measures the quantum electron recoil effect caused by the quantum bohm potential @xcite @xmath27 compared to the quantum statistical fermi electron pressure and the electron - exchange and electron - correlation effects arising from @xmath1-spin of the degenerate electrons . here
@xmath28 is the electron fermi speed and @xmath29^{1/2}$ ] includes the effects of electron exchange and electron correlations , where @xmath30 represents the average inter - electron distance .
the expression for @xmath31 is derived by linearizing the sum of the electron exchange and electron correlation potentials @xcite @xmath32 $ ] .
we note that @xmath26 depends only on @xmath22 , as @xmath33 ^ 2 $ ] .
@xcite found that attractive potentials between ions exist only for @xmath34 .
figure 1 displays the value of @xmath26 as a function of @xmath22 , where one observes that it is above a critical value @xmath35 only for a limited range @xmath36 , corresponding to an electron number density in the range @xmath37 ( with @xmath38 ) .
the maximum value is @xmath39 at @xmath40 , corresponding to the electron number density @xmath41 , a few times below solid densities .
the validity of the seap has been further expanded @xcite for wider density ranges by including chandrasekhar s ( @xcite ) generalized pressure law for degenerate electron fluids and salpeter s ( @xcite ) electron - exchange and electron - correlation potentials that are of astrophysical interest ( e.g. the cores of white dwarf stars ) .
the nonlinear shielding effects on the seaf is discussed by @xcite . as a function of @xmath42 for @xmath43 ( solid curve ) , @xmath44 ( dashed curve ) and @xmath45 ( dash - dotted curve ) .
the value 0.627 is the maximum possible value of @xmath26 in our model , obtained for @xmath40.,width=264 ] for @xmath46 , the profile of the electric potential as a function of distance @xmath42 around a stationary test ion charge @xmath47 is @xcite @xmath48 \exp(-k_r r ) , \label{pot1}\ ] ] which is referred to as the seap . here
@xmath47 is the ion charge , @xmath49 , @xmath50 , and @xmath51 , with @xmath52 being the modified inverse tf screening length .
the spatial profile of the seap in fig .
2 shows a distinct minimum for the case @xmath53 . the negative part of the seap , given by eq .
( 1 ) , resembles the lj potential and leads to a short - range seaf between neighboring ions . on the other hand ,
the potential distribution around a test ion for @xmath54 reads @xcite @xmath55,\ ] ] where @xmath56 , and @xmath57 . for this case , the potential is positive and monotonically decreasing ( cf .
2 ) , giving rise to only a repulsive force ( similar to the tf force ) between ions . in the high- or low - density limit , where @xmath58 , we recover the modified tf screened coulomb potential @xmath59 .
we here present a computer simulation study of the dynamics of a system of ions interacting with each other under the action of the seaf . for this purpose
, we numerically solve the equations of motion for a system of ions with equal charges and masses , given by @xmath60 where instantaneous position of each ion is determined from @xmath61 . here
@xmath62 is the radius vector between particle @xmath63 and @xmath64 , @xmath65 the position and @xmath66 the velocity of the @xmath64th ion , @xmath67 the ion mass , @xmath68 denotes the gradient of @xmath69 at position @xmath70 , and @xmath71 denotes an effective collision frequency , which tends to retard the ion motion .
the external confining potential , @xmath72 ) , of charged particles may have different amplitudes @xmath73 and @xmath74 perpendicular and parallel to the @xmath75-axis , respectively , where @xmath76 . in order to demonstrate the clustering of ions under the seaf
, we now carry out particle simulations of eq .
( 3 ) with 1000 particles , initially randomly placed in space , as shown in fig .
3 . in the first set of simulations , displayed in fig .
4 , we consider interactions of the ions in the absence of the external confining potential @xmath77 , viz .
@xmath78 , and with @xmath79 .
the positions of ions at the end of the simulations are shown in fig . 4 at time @xmath80 . here
the positions and time are in units of @xmath81 and @xmath82 , respectively , where @xmath83 is the ion plasma frequency , and @xmath84 the equilibrium ion number density , related to the electron number density @xmath85 via the quasi - neutrality condition @xmath86 , where @xmath87 is the ion charge state . for @xmath43 , which leads to a potential minimum ( cf .
2 ) , we observe the clustering of ions and the formation of large - scale ionic structures .
the clustering of ions is a relatively slow process in comparison with the ion plasma period @xmath88 .
ion pairs and smaller clusters are initially formed , and later the larger ion clusters are gradually formed by the agglomeration of smaller ion clusters . as a contrast , for @xmath44 , we see in fig .
4 that there does not exist condensation / coalescence of ions .
this is due to the fact that there is no potential minimum for this value of @xmath26 ( cf .
2 ) , and hence the inter - ion force is always repulsive for this case .
furthermore , we have carried out a set of simulations with a symmetric potential @xmath77 with @xmath89 . here , as seen in fig . 5 , almost spherical
non - coulombic ion crystals are formed for both @xmath43 and @xmath44 .
we also performed simulations including an asymmetric external potential @xmath77 with @xmath90 and @xmath91 , providing a stronger confinement in the perpendicular direction .
figure 6 displays the final state for @xmath43 and @xmath44 , where ions , in both cases , form non - coulombic ion crystals elongated along the @xmath75-direction . in general , the formation of non - coulombic ion crystals is attributed to the balance between the external and inter - ion potentials , where the system tends to a configuration of a minimum potential energy .
we note that similar configurations have been previously reported for both charged macro - particles @xcite and ions @xcite confined by external potentials in the paul trap . finally , in figs . 7 and 8
, we continued the simulations in figs . 5 and 6 and set the external confining potential to zero at time @xmath92 .
for both cases , with symmetric and asymmetric potentials , we see that ion crystals remained tightly packed for @xmath43 , where it performed damped oscillations as elastic solids before settling down to the final states in figs .
7(a ) and 8(a ) .
in contrast , for @xmath44 , the ion cloud expanded to larger and less dense ion clouds , as seen in figs 7(b ) and 8(b ) at @xmath93 .
hence the short - range potential around ions for @xmath43 ( cf .
2 ) lead to the solidification of the ions , which is not the case for the the short - range repulsive potential for @xmath44 .
in this paper , we have carried out particle simulations to demonstrate clustering of ions due to the newly found seaf arising from collective interactions between an ensemble of degenerate electrons that shield ions in dense quantum plasmas .
specifically , the seaf leads to clustering / condensation or coagulation of ions in the absence of an external confining potential for charged particles .
however , ion clustering can be put on the firm footing by calculating the dynamical ion structure factor ( disf ) based on the fluctuation - dissipation theorem and the dielectric constant of degenerate electrons and strongly correlated ions in a viscous quantum plasma .
it may well turn out that the disf will reveal long - range correlations between ions .
it is our belief that the formation of ion nano - clusters is going to play a valuable role in the area of high - density compressed plasmas with degenerate electrons @xcite for icf to succeed , and also in the emerging field of nano - material sciences ( e.g. nanodiodes , metallic nanostructures for thin films @xcite , nanowires ) , where closely - packed ions will lend support to enhanced fusion probabilities ( with anomalous fusion cross - sections ) for controlled thermonuclear icf , and may also influence the electric properties ( e.g. resistivity ) of new high density plasma materials at relatively high temperatures .
finally , we stress that the cooper pairing of ions at atomic dimensions shall provide possibility of novel superconducting plasma - based nanotechnology , since the electron transport in nanostructures would be rapid due to shortened distances between ions in the presence of the novel seaf . 99 akbari - moghanjoughi , m. 2012 shukla - eliasson attractive force : revisited .
_ j. plasma phys . _ * 78 * , in press ; doi : http://dx.doi.org/s00223778120000839 ; e - print : arxiv:1207.5154v2 [ astro-ph.ep ] .
avinash , k. 2007 mean - field theory of critical phenomenon for mutually repelling particles in complex plasmas .
lett . _ * 98 * , 095003 .
barkan , a. & merlino , r. l. 1995 confinement of dust particles in a double layer .
plasmas _ * 2 * , 32613265 .
berg , t. g. o. & gaukler , t. a. 1969 apparatus for the study of charged particles and droplets .
j. phys . _
* 37 * , 10131018 .
bohm , d. 1952 a suggested interpretation of the quantum theory in terms of `` hidden '' variables .
i. _ phys .
rev . _ * 85 * , 166179 .
bohm , d. & pines , d. 1953 a collective description of electron interactions : iii .
coulomb interactions in a degenerate electronggas .
rev . _ * 92 * , 609625 .
brey , l. , dempsey , j. , johnson , n. f. & halperin , b. i. 1990 infrared optical absorption in imperfect parabolic quantum wells .
b _ * 42 * , 12401247 .
brodin , g. , marklund , m. & manfredi , g. 2008 quantum plasma effects in the classical regime .
lett . _ * 100 * , 175001 .
chandrasekhar , s. 1931 the maximum mass of ideal white dwarfs .
. j. _ * 74 * , 8182 .
chandrasekhar , s. 1939 _ an introduction to the study of stellar structure _
( the university of chicago press , chicago ) .
chu , j. h. & i , l. 1994 direct observation of coulomb crystals and liquids in strongly coupled rf dusty plasmas . _
lett . _ * 72 * , 40094012 .
crandall , r. s. & williams , r. 1971 crystallization of electrons on the surface of liquid helium .
lett . a _ * 34 * , 404405 . crouseilles , n. , hervieux , p. a. & manfredi , g. 2008 quantum hydrodynamic model for the nonlinear electron dynamics in thin metal films .
b _ * 78 * , 155412 .
debye , p. , & hckel , e. 1923 zur theorie der electrolyte i : gefrierpunktserniedrigung und verwandte erscheinungen .
z. _ * 24 * , 185206 .
deshpande , v. v. , & bockrath , m. 2008 the one - dimensional wigner crystal in carbon nanotubes .
_ nature phys . _ * 4 * , 314318 .
drewsen , m. , brodersen , c. , hornekr , l. , hangst , j. s. , & schiffer , j. p. 1998 large ion crystals in a linear paul trap .
* 81 * , 28782881 .
fermi , e. 1927 un metodo statistico per la determinazione di alcune propriet dellatomo .
* 6 * , 602607 .
fortov , v. e. 2009 extreme state of matter on earth and in space .
usp . _ * 52 * , 615647 .
fortov v. e. , nefedov , a. p. , torchinsky , v. m. , molotkov , v. i. , petrov , o. f. , samarian , a. a. , lipaev , a. a. & khrapak , a. g. , crystalline structure of strongly coupled dusty plasmas in dc glow discharge strata , _ phys .
a _ * 229 * , 317322 .
gardner , c. l. & ringhofer , c. 1996 smooth quantum potential for the hydrodynamic model .
e _ * 53 * , 157167 .
glenzer s. h. & redmer , r. 2009 x - ray thomson scattering in high energy density plasmas .
phys . _ * 81 * , 16251663 .
glenzer , s. h. , landen , o. l. , neumayer , p. , lee , r. w. , widmann , k. , pollaine , s. w. , wallace , r. j. , gregori , g. , holl , a. , bornath , t. , thiele , r. , schwarz , v. , kraeft , w. d. , redmer , r. 2007 observations of plasmons in warm dense matter .
* 98 * , 065002 .
grimes , c. c. & adams , g. 1979 evidence for a liquid - to - crystal phase transition in a classical , two - dimensional sheet of electrons , _ phys .
lett . _ * 42 * , 795798 .
haas , f. 2011 _ quantum plasmas : an hydrodynamical approach _
( springer , new york ) .
hayashi , y. & tachibana , k. 1994 observation of coulomb - crystal formation from carbon particles grown in a methane plasma .
phys . _ * 33 * , l804l806 .
hedin , l. & lundqvist , b. i. 1971 explicit local exchange - correlation potentials .
_ j. phys .
c : solid state phys . _ * 4 * , 20642083 .
klimontovich , yu . l. & silin , v. p. 1952a
nauk sssr _ ,
* 82 * , 361 .
klimontovich , yu . l. & silin , v. p. 1952b
concerning the spectra of systems of interacting particles . _
_ * 23 * , 151160 .
klumov , b. 2010 on melting criteria for complex plasma .
_ * 53 * , 10531065 .
kremer , k. , robbins , m. o. , & grest , g. s. 1986 phase diagram of yukawa systems : model for charge - stabilized colloids .
* 57 * , 26942697 .
landau , l. d. & lifshitz , e. m. 1980 _ statistical physics _ ( butterworth - heinemann , oxford ) .
langmuir , i. 1929 the interaction of electron and positive ion space charges in cathode sheaths .
rev . _ * 33 * , 954989 .
malkin , v. m. , fisch , n. j. , & wurtele , j. s. 2007 compression of powerful x - ray pulses to attosecond durations by stimulated raman backscattering in plasmas .
e _ * 75 * , 026404 .
manfredi , g. & haas , f. 2001 self - consistent fluid model for a quantum electron gas .
b _ * 64 * , 075316 .
manfredi , g. 2005 how to model quantum plasmas .
_ fields inst .
* 46 * , 263287 .
melrose , d. b. 2008 _ quantum plasmadynamics : unmagnetized plasmas _
( springer , berlin ) .
mendona , j. t. 2011 wave kinetics of relativistic quantum plasmas .
plasmas _ * 18 * , 062101 .
mohideen , u. , rahman , h. u. , smith , m. a. , rosenberg , m. , & mendis , d. a. 1998 intergrain coupling in dusty - plasma coulomb crystals .
_ * 81 * , 349352 .
mlhave , k. & drewsen , m. 2000 formation of translationally cold mgh@xmath94 and mgd@xmath94 molecules in an ion trap .
a _ * 62 * , 011401 .
nambu , m. , vladimirov , s. v. , & shukla , p. k. 1995 attractive forces between charged particulates in plasmas .
lett . a _ * 203 * , 4042 . resendes , d. f. , mendona , j. t. , & shukla , p. k. 1998 formation of dusty plasma molecules .
a _ * 239 * , 181186 .
robertson , s. & younger , r. 1999 coulomb crystals of oil droplets . _ am .
_ * 67 * , 310315 .
salpeter , e. e. 1961 energy and pressure of a zero - temperature plasma .
. j. _ * 134 * , 669682 .
shaikh d. & shukla , p. k. 2007 fluid turbulence in quantum plasmas .
lett . _ * 99 * , 125002 .
shukla , p. k. & eliasson , b. 2006 formation and dynamics of dark solitons and vortices in quantum electron plasmas .
lett . _ * 96 * , 245001 .
shukla , p. k. & eliasson , b. 2007 nonlinear interactions between electromagnetic waves and electron plasma oscillations in quantum plasmas _ phys .
lett . _ * 99 * , 096401 .
shukla , p. k. & eliasson , b. 2009 fundamentals of dust plasma interactions _ rev . mod .
phys . _ * 81 * , 2550 .
shukla , p. k. & eliasson , b. 2010 nonlinear aspects of quantum plasma physics .
* 53 * , 5176 .
shukla , p. k. & eliasson , b. 2011 nonlinear collective interactions in quantum plasmas with degenerate electron fluids .
phys . _ * 83 * , 885906 .
shukla , p. k. & eliasson b. 2012a novel attractive force between ions in quantum plasmas .
lett . _ * 108 * , 165007 .
shukla , p. k. & eliasson b. 2012b erratum : novel attractive force between ions in quantum plasmas [ phys .
108 , 165007 ( 2012 ) ] .
lett . _ * 108 * , 219902(e ) .
shukla , p. k. & eliasson b. 2012c erratum : novel attractive force between ions in quantum plasmas [ phys .
108 , 165007 ( 2012 ) ] .
lett . _ * 109 * , 019901(e ) .
shukla , p. k. , akbari - moghanjoughi , m. & eliasson b. 2012 comment on `` on novel attractive forces between ions in quantum plasmas failure of linearized quantum hydrodynamics '' _ arxiv:1206.3456 [ physics.plasm - ph]_. shukla , p. k. & mamun , a. a. 2002 _ introduction to dusty plasma physics _ ( institute of physics , bristol ) . shukla , p. k. & rao , n. n. 1996 coulomb crystallization in colloidal plasmas with streaming ions and dust grains .
plasmas _ * 3 * , 17701772 .
son s. & fisch , n. j. 2004 aneutronic fusion in a degenerate plasma . _
a _ * 329 * , 7682 .
son s. & fisch , n. j. 2005 current - drive efficiency in a degenerate plasma . _
lett . _ * 95 * , 225002 .
son s. & fisch , n. j. 2006 controlled fusion with hot - ion mode in a degenerate plasma .
a _ * 356 * , 6571 .
staanum , p. f. , hjbjerre , k. , skyt , p. s. , hansen a. k. & drewsen , m. 2010 rotational laser cooling of vibrationally and translationally cold molecular ions .
_ nature phys . _
* 6 * , 271274 .
tan , j. n. , bollinger , j. j. , jelenkovic , b. & wineland , d. j. 1995 long - range order in laser - cooled , atomic - ion wigner crystals observed by bragg scattering .
lett . _ * 75 * , 41984201 .
thomas , h. , morfill , g. e. , demmel , v. , goree , j. , feuerbacher , b. & mhlmann , d. 1994 plasma crystal : coulomb crystallization in a dusty plasma .
lett . _ * 73 * , 652655 .
thomas , l. h. 1927 the calculation of atomic fields .
cambridge phil .
* 23 * , 542548 .
tsintsadze , n. l. & tsintsadze , l. n. 2009 novel quantum kinetic equations of the fermi particles .
_ europhys .
_ * 88 * , 35001 .
vladimirov , s. v. & nambu , m. 1995 attraction of charged particulates in plasmas with finite flows .
e _ * 52 * , r2172r2174 .
vladimirov , s. v. & tyshetskiy , yu .
o. 2011 on description of a collisionless quantum plasma .
* 54 * , 12431256 .
watanabe , h. 1956 experimental evidence for the collective nature of the characteristic energy loss of electrons in solids
studies on the dispersion relation of plasma frequency. _ j. phys .
jpn . _ * 11 * , 112119 .
wigner , e. 1934 on the interaction of electrons in metals .
* 46 * , 10021011 .
wilhelm , h. e. 1971 wave - mechanical formulation of plasma dynamics in longitudinal electric fields .
_ z. phys . _ * 241 * , 18 .
winter h. & ortjohann , h. w. 1991 simple demonstration of storing macroscopic particles in a `` paul trap '' .
_ * 59 * , 807813 .
wuerker , r. f. , shelton , h. & langmuir , r. v. 1959 electrodynamic containment of charged particles .
_ j. appl .
phys . _ * 30 * , 342349 .
yukawa h. 1935 on the interaction of elementary particles i. _ proc .
phys .- math .
* 17 * , 4857 . | by means of particle simulations of the equations of motion for ions interacting among themselves under the influence of newly discovered shukla - eliasson attractive force ( seaf ) in a dense quantum plasma , we demonstrate that the seaf can bring ions closer at atomic dimensions .
we present simulation results of the dynamics of an ensemble of ions in the presence of the seaf without and with confining external potentials and collisions between the ions and degenerate electrons .
our particle simulations reveal that under the seaf , ions attract each other , come closer and form ionic clusters in the bath of degenerate electrons that shield the ions .
furthermore , an external confining potential produces robust ion clusters that can have cigar - like and ball - like shapes , which remain stable when the confining potential is removed .
the stability of ion clusters is discussed .
our results may have applications to solid density plasmas ( density exceeding @xmath0 per cubic centimeters ) , where the electrons will be degenerate and quantum forces due to the electron recoil effect caused by the overlapping of electron wave functions and electron tunneling through the bohm potential , electron - exchange and electron - exchange and electron correlations associated with electron-@xmath1 spin effect , and the quantum statistical pressure of the degenerate electrons play a decisive role . during the early thirties
, there were several discoveries related to non - coulombic shielded potential distributions that exhibit the role of collective interactions between electrons and ions in electro - chemistry @xcite ( viz .
electrolytes and colloidal suspensions ) , solid state @xcite and gaseous @xcite plasmas , and between neutrons and protons in elementary particle physics @xcite .
the screened non - coulombic potentials , which were obtained by using linearized theory based on the assumption that the potential energy between the particles is much smaller than the particle kinetic energy , are now known as the debye - hckel ( dh ) , thomas - fermi ( tf ) , and yukawa potentials in the context of electro - chemistry and plasma physics , condensed matter physics , and nuclear physics , respectively .
the dh , tf , and yukawa potentials describe short - range ( of the order of the dh radius , the tf radius , and the yukawa radius , which are fixed by the size of a shielded cloud ) repulsive interactions between two particles that have the same polarity .
the dh theory has also been extended to dusty plasma physics where charged dust particles are shielded by non - degenerate electrons and ions .
the dh , tf , and yukawa interaction potentials , which significantly deviate from the long - range coulomb interaction potential , have important applications to the understanding of phase transitions @xcite in different areas of physical sciences . in order for charged particles
to form ordered structures under the influence of coulombic , dh , tf and yukawa repulsive forces , one must confine the like - charged particles in an external potential , so as to bring them to a minimum energy state .
examples include the wigner crystals @xcite composed of an ensemble of electrons on the surface of liquid helium , ion crystals in laser cooled paul @xcite and penning ( electromagnetic ) traps @xcite , charged dust particle crystals @xcite , which were formed when like - charged dust particles were kept together via external confining potentials despite short - range coulombic or shielded coulombic repulsive forces between charged dust particles . in fact , both electron and ion crystals , as well as crystals of colloidal suspensions and oil droplets have been observed experimentally under different physical circumstances @xcite .
moreover , an ensemble of strongly correlated micron - sized negative dust particles form dust coulomb crystals @xcite when they were confined by the sheath parabolic potential in low - temperature laboratory dusty plasma discharges @xcite .
the condensation of charged dust particles occurs since the dusty plasma @xmath2 ( the ratio between the coulomb energy between highly charged dust grains and the average dust particle kinetic energy ) becomes relatively large due to the high dust charge state and low dust temperature .
the attraction between like - charged dust particles forming dust coulomb crystals may also be attributed to attractive forces arising from overlapping of the dusty plasma debye spheres @xcite , ion focusing and wakefield effects @xcite , and dust dipole - dipole interactions @xcite .
the cooper s pairing of charged dust particles , which are glued by ions , led to the discovery of a soft- condensed matter of dust particle crystals in low - density and low - temperature classical plasma with maxwell - boltzmann distributions for electrons and ions .
it turns out that several milestones were reached in the areas of ordered crystalline structures composed of charged particles ( e.g. an ensemble of in low - temperature physical systems electrons , ions , as well as charged colloidal and dust particles ) in physical systems , which share some common physics .
however , solid density plasmas are of fundamental importance for industrial applications ( e.g. semiconductors , nano - diodes and metallic nanostructures for thin films ) , for inertial confinement fusion ( icf ) schemes that utilizes high density compressed ( hdc ) plasmas , as well as for planetary systems [ e.g. the core of jupiter @xcite ] and superdense astrophysical objects ( e.g the cores of white dwarf stars , warm dense matter ) .
in dense plasmas , one has to account for degeneracy @xcite of electrons which obey the fermi - dirac distribution function .
correspondingly , quantum mechanical effects play a vital role since in such dense plasmas the wigner - seitz radius @xmath3 is comparable to the thermal de broglie wavelength @xmath4 ( which is a measure of the extent of electron wave functions ) , where @xmath5 is planck s constant divided by @xmath6 , @xmath7 the electron mass , @xmath8 the electron thermal speed due to random electron motions , and @xmath9 the boltzmann constant .
also , in dense plasmas with degenerate electrons , @xmath10 turns out to be much smaller than the landau length @xmath11 , which can be conveniently expressed as @xmath12 , where @xmath13 is the bohr radius of a hydrogen atom .
the electron degeneracy effects at nanoscales in dense plasmas can thus be captured through the consideration of the fermi - dirac statistics for electrons with spin-@xmath1 ( fermions ) , and overlapping of electron wavefunctions due to heisenberg s uncertainty principle and pauli s exclusion principle , as well as electron - exchange and electron - correlations .
hence , there are quantum forces @xcite associated with the quantum statistical electron pressure @xcite , the quantum bohm potential @xcite through which degenerate electrons can tunnel through ( often known as the quantum electron recoil effect ) , as well as the electron - exchange and electron - correlations potentials @xcite .
it has been found that the above mentioned quantum forces acting on degenerate electrons in quantum plasmas introduce new dispersive features to electron plasma oscillations @xcite with frequencies in the x - ray regime , which can be assessed by using collective x - ray spectroscopic scattering techniques @xcite .
in fact , @xcite have reported observations of electron plasma oscillations ( epos ) in warm dense matter ( with the peak electron number density @xmath14 @xmath15 and the equilibrium electron and ion temperatures of 12 ev ( @xmath16 degrees kelvin ) .
the latter is different from the electron fermi temperature @xmath17 degrees kelvin , corresponding to an electron number density @xmath18 ) .
thus , the recent experiments of @xcite have unambiguously demonstrated the importance of the quantum statistical pressure and quantum electron recoil effects on the frequency spectra of epos @xcite , although a previous experimental investigation @xcite had already established the quantum dispersion properties of epos in metals .
very recently , @xcite discovered an oscillating shielded coulomb ( osc ) potential , ( also referred to as the shukla - eliasson attractive potential ( seap ) @xcite ) , which is valid for the potential energy much smaller than @xmath19 and @xmath20 , around a stationary test ion in an unmagnetized quantum plasma , where @xmath21 is the speed of light in vacuum . the seap arises due to collective interactions between an ensemble of degenerate electrons that shield an isolated ion at atomic dimensions .
the profile of the seap in quantum plasmas resembles the lennard - jones ( lj ) potential in atomic gases . the shukla - eliasson attractive force ( seaf ) , defined as minus the gradient of the seap , brings ions closer to form ion clusters in quantum plasmas . in this paper , we demonstrate the formation of ion clusters at atomic dimensions by performing computer simulations of the equations of motion for an ensemble of ions that are interacting with each other through the seaf . |
in 1998 , the cdf collaboration observed the pseudoscalar bottom - charm @xmath0 mesons through the decay modes @xmath1 and @xmath2 in @xmath3 collisions at the energy @xmath4 at the fermilab tevatron , the measured mass is @xmath5 @xcite . in 2007 ,
the cdf collaboration observed the pseudoscalar @xmath0 mesons with a significance exceeds @xmath6 through the decay modes @xmath7 in @xmath3 collisions at the energy @xmath8 using the collider detector at fermilab ( cdf ii ) , the measured mass is @xmath9 @xcite . in 2008 ,
the d0 collaboration reconstructed the @xmath7 decays and observed the pseudoscalar @xmath0 mesons with a significance more than @xmath10 , the measured mass is @xmath11 @xcite . now
the average value listed in the review of particle physics is @xmath12 @xcite .
other @xmath0 mesons , such as the scalar , vector , axialvector , tensor @xmath0 mesons , have not been observed yet , but they are expected to be produced at the large hadron collider ( lhc ) in the future @xcite .
the heavy quarkonium states and triply - heavy baryon states play an important role both in studying the interplays between the perturbative and nonperturbative qcd and in understanding the heavy quark dynamics due to the absence of the light quark contaminations .
the bottom - charm quarkonium states @xmath0 , which consist of the heavy quarks with different flavors , are of special interesting .
the ground states and the excited states lying below the @xmath13 , @xmath14 , @xmath15 , @xmath16 thresholds can not annihilate into gluons , and therefore are more stable than the corresponding charmonium and bottomonium states , and would have widths less than a hundred @xmath17 @xcite . the excited states can undergo radiative or hadronic transitions to the ground state pseudoscalar @xmath0 mesons , which decay weakly . there have been several theoretical works on the mass spectroscopy of the @xmath0 mesons , such as the relativized ( or relativistic ) quark model with an special phenomenological potential @xcite , the nonrelativistic quark model with an special phenomenological potential @xcite , the semi - relativistic quark model using the shifted large-@xmath18 expansion @xcite , the perturbative qcd @xcite , the nonrelativistic renormalization group @xcite , the lattice qcd @xcite , etc .
the qcd sum rules is a powerful theoretical tool in studying the heavy quarkonium states @xcite , and the existing works focus on the @xmath19-wave heavy quarkonium states @xmath20 , @xmath21 , @xmath22 , @xmath23 , and the @xmath24-wave spin - triplet heavy quarkonium states @xmath25 , @xmath26 , @xmath27 @xcite .
the pseudoscalar @xmath0 mesons have been studied by the full qcd sum rules @xcite and the potential approach combined with the qcd sum rules @xcite , while the vector @xmath0 mesons ( @xmath28 ) have been studied by the full qcd sum rules @xcite , and the axialvector @xmath0 mesons have not been studied yet . in ref.@xcite ,
colangelo , nardulli and paver took the leading - order approximation , obtained the values @xmath29 and @xmath30 , and did not present the value @xmath31 . in ref.@xcite , narison took into account the next - to - leading - order perturbative contributions by assuming that one quark had zero mass , and obtained the values @xmath32 , @xmath33 , @xmath34 , the predicted mass @xmath35 is much larger than other theoretical calculations @xcite .
those studies based on the qcd sum rules were preformed before the pseudoscalar @xmath0 mesons were observed by the cdf collaboration , the predictions should be updated .
now we can take the experimental data as guides to choose the suitable borel parameters and continuum threshold parameters .
naively , we expect that the masses of the pseudoscalar , vector and axialvector @xmath0 mesons have the hierarchy : @xmath36 , the @xmath37 , @xmath38 and @xmath39 denote the spin - parity @xmath40 .
furthermore , the calculations based on the nonrelativistic renormalization group indicate that @xmath41 @xcite . in this article
, we carry out the operator product expansion by including the next - to - leading - order perturbative contributions , study the masses and decay constants of the vector and axialvector @xmath0 mesons with the qcd sum rules , and make reasonable predictions for the masses and decay constants , furthermore , we calculate the leptonic decay widths .
the decay constants are basic input parameters in studying the exclusive processes of the @xmath0 mesons with the three - point vacuum correlation functions .
the article is arranged as follows : we derive the qcd sum rules for the masses and decay constants of the vector and axialvector @xmath0 mesons in sect.2 ; in sect.3 , we present the numerical results and discussions ; and sect.4 is reserved for our conclusions .
in the following , we write down the two - point correlation functions @xmath42 in the qcd sum rules , @xmath43 where @xmath44 , the vector and axialvector currents @xmath45 and @xmath46 interpolate the vector and axialvector @xmath0 mesons , respectively . we can insert a complete set of intermediate hadronic states with the same quantum numbers as the current operators @xmath47 into the correlation functions @xmath42 to obtain the hadronic representation @xcite . after isolating the ground state contributions come from the vector and axialvector @xmath0 mesons , we get the following result , @xmath48 where the decay constants @xmath49 are defined by @xmath50 and the @xmath51 are the polarization vectors of the vector and axialvector @xmath0 mesons .
we can use dispersion relation to express the hadronic ( or phenomenological ) representation of the correlation functions @xmath52 in the following form , @xmath53 where the @xmath54 are the continuum threshold parameters .
now , we briefly outline the operator product expansion for the correlation functions @xmath42 .
we contract the quark fields in the correlation functions @xmath55 ( here we add the indexes @xmath56 and @xmath57 to denote the vector and axialvector currents respectively ) with wick theorem firstly , @xmath58 where the @xmath59 and @xmath60 are the full @xmath61 and @xmath62 quark propagators , and can be written as @xmath63 collectively , @xmath64 and @xmath65 , the @xmath66 is the gell - mann matrix , the @xmath67 , @xmath68 are color indexes , @xmath69 , and the @xmath70 is the gluon condensate @xcite ; then complete the integrals both in the coordinate space and in the momentum space , which corresponds to calculate the feynman diagrams in figs.1 - 3 ; finally obtain the correlation functions @xmath42 ( or @xmath52 ) at the level of the quark - gluon degrees of freedom . in calculations ,
we have used the equation of motion , @xmath71 , and taken the approximation @xmath72 to obtain the contributions of the four - quark condensates .
the contributions of the four - quark condensates are depressed by inverse powers of the large euclidean momentum @xmath73 ( thereafter the borel parameter @xmath74 ) and play minor important roles , we neglect other diagrams contribute to the four - quark condensates of the order @xmath75 .
we also neglect the contributions come from the three gluon condensates , as they are also depressed by inverse powers of the large euclidean momentum @xmath73 .
the feynman diagrams for the next - to - leading - order perturbative contributions are shown in fig.4 .
we calculate the diagrams using the cutkosky s rule to obtain the spectral densities .
there are two routines in application of the cutkosky s rule ( or optical theorem ) , we resort to the routine used in ref.@xcite , not the one used in ref.@xcite .
there are ten possible cuts , the six cuts shown in fig.5 attribute to virtual gluon emissions and correspond to the self - energy corrections and vertex corrections , while the four cuts shown in fig.6 correspond to real gluon emissions , for technical details , one can consult ref.@xcite . .
, width=188 ] once analytical expressions of the spectral densities at the quark level are obtained , then we take the quark - hadron duality and perform the borel transforms with respect to the variable @xmath76 to obtain the following qcd sum rules , @xmath77 \exp\left(-\frac{s}{t^2}\right ) \
, , \end{aligned}\ ] ] where @xmath78 @xmath79{m_b^7m_c^7s}}{\lambda(s , m_b^2,m_c^2)}+\frac{2(s - m_b^2-m_c^2)}{\sqrt{\lambda(s , m_b^2,m_c^2)}}\log\left(\frac{1+\omega}{1-\omega}\right ) \log\frac{\lambda(s , m_b^2,m_c^2)}{m_b m_c s}\nonumber\\ & & \left .
-\frac{2(s - m_b^2-m_c^2)}{3\sqrt{\lambda(s , m_b^2,m_c^2)}}\log\left(\frac{1+\omega}{1-\omega}\right)-r_{12}^1(s)\right\ } \nonumber\\ & & + \frac{4}{3}\frac{\alpha_s}{\pi}\left\ { \frac{s-(m_b\mp m_c)^2}{4s\pi^2}\left[\log\left(\frac{1+\omega}{1-\omega}\right)(s - m_b^2-m_c^2)-\sqrt{\lambda(s , m_b^2,m_c^2)}\right ] \right . \nonumber\\ & & + \frac{\sqrt{\lambda(s , m_b^2,m_c^2)}}{16s\pi^2}r_{12}^2(s)\left[2+\frac{(m_b\mp m_c)^2}{s}\right]-\frac{1}{\pi^2}r_0(s)+\frac{\sqrt{\lambda(s , m_b^2,m_c^2)}^3}{s^2}\nonumber\\ & & \left.\left[\frac{1}{12\pi^2 } \left(1-\frac{s - m_b^2-m_c^2}{\sqrt{\lambda(s , m_b^2,m_c^2)}}\log\left(\frac{1+\omega}{1-\omega}\right ) \right ) -\frac{(m_b\pm m_c)\left(f_{1\pm}(s)+f_{2\pm}(s ) \right)}{32\pi^2 } \right]\right\ } \ , , \end{aligned}\ ] ] @xmath80\delta(s-\widetilde{m}_q^2 ) \nonumber\\ & & \pm \frac{m_b m_c}{8t^2}\langle\frac{\alpha_sgg}{\pi}\rangle\int_0 ^ 1dx\left [ \frac{1}{x^2}+\frac{1}{(1-x)^2}\right ] \delta(s-\widetilde{m}_q^2)\nonumber\\ & & -\frac{s}{24t^4}\langle\frac{\alpha_sgg}{\pi}\rangle\int_0 ^ 1dx \left [ \frac{(1-x)m_c^2}{x^2}+\frac{x m_b^2}{(1-x)^2}\right]\delta(s-\widetilde{m}_q^2 ) \nonumber\\ & & -\frac{1}{12}\langle\frac{\alpha_sgg}{\pi}\rangle\int_0 ^ 1dx \left [ 1+\frac{s}{2t^2}\right]\delta(s-\widetilde{m}_q^2)+\frac{4\alpha_s^2\langle\bar{q}q\rangle^2}{81t^2}\int_0 ^ 1dx \left[\frac{2}{x(1-x)}+\frac{m_b^2 m_c^2}{x^2(1-x)^2t^4 } \right.\nonumber\\ & & \left.+\frac{m_b^2 + m_c^2\pm 9m_b m_c}{3x(1-x)t^2 } -\frac{2}{3}\left(1+\frac{s}{t^2}-\frac{s^2}{t^4 } \right)+\frac{5s}{3x(1-x)t^2 } \right]\delta(s-\widetilde{m}_q^2 ) \ , , \end{aligned}\ ] ] where @xmath81 \pm 2m_bm_c\nonumber\\ & & \left [ v_{10}(s)+v_{01}(s)\right]+2m_b^2\left [ v_{10}(s)-v_{20}(s)\right]+2m_c^2\left [ v_{01}(s)-v_{02}(s)\right]\ , , \nonumber\\ f_{1\pm}(s)&=&4m_bv_{20}(s)\mp 4m_cv_{01}(s)\pm 4m_cv_{11}(s)\ , , \nonumber\\ f_{2\pm}(s)&=&\pm 4m_cv_{02}(s)-4m_bv_{10}(s)+4m_bv_{11}(s)\ , , \nonumber\\ \omega&=&\sqrt{\frac{s-(m_b+m_c)^2}{s-(m_b - m_c)^2}}\ , , \end{aligned}\ ] ] @xmath82 , @xmath83 , and the @xmath74 is the borel parameter .
the explicit expressions of the @xmath84 , @xmath85 , @xmath86 , @xmath87 , @xmath88 , @xmath89 , @xmath90 , @xmath91 , @xmath92 , @xmath93 , @xmath94 , @xmath95 and @xmath96 are given in the appendix .
we can eliminate the decay constants @xmath97 and obtain the qcd sum rules for the masses of the vector and axialvector @xmath0 mesons , @xmath98\exp\left(-\frac{s}{t^2}\right)}{\int_{(m_b+m_c)^2}^{s_0 } ds \left[\rho^0_{\pm}(s)+\rho^{1}_{\pm}(s)+\rho^{\rm con}_{\pm}(s)\right]\exp\left(-\frac{s}{t^2}\right)}\ , , \end{aligned}\ ] ] then use the resulting masses as input parameters to obtain the decay constants @xmath97 .
the mass of the pseudoscalar @xmath0 meson is @xmath12 from the particle data group @xcite , while the calculations based on the nonrelativistic renormalization group indicate that @xmath41 @xcite .
we can tentatively take the continuum threshold parameters as @xmath99 and @xmath100 , and search for the ideal values , where we have assumed that an additional @xmath24-wave results in mass - shift @xmath101 and the energy gap between the ground states and the first radial excited states is @xmath101 .
the quark condensate is taken to be the standard value @xmath102 at the energy scale @xmath103 @xcite .
the quark condensate evolves with the renormalization group equation , @xmath104^{\frac{4}{9}}$ ] . the value of the gluon condensate @xmath105 has been updated from time to time , and changes greatly @xcite
, we use the recently updated value @xmath106 @xcite . in this article , we study the vector and axialvector @xmath0 mesons with both the @xmath107 masses and pole masses .
the @xmath107 masses have been studied extensively by the qcd sum rules and lattice qcd @xcite .
the values listed in the review of particle physics are @xmath108 and @xmath109 @xcite , which correspond to the pole masses
@xmath110 and @xmath111 . the recent studies based on the qcd sum rules @xcite , the nonrelativistic large - n @xmath22 sum rules with renormalization group improvement @xcite and the lattice qcd @xcite indicate ( slightly ) different values .
we take the @xmath107 masses @xmath112 and @xmath113 from the particle data group @xcite .
furthermore , we take into account the energy - scale dependence of the @xmath107 masses from the renormalization group equation , @xmath114^{\frac{12}{25 } } \ , , \nonumber\\ \overline{m}_b(\mu^2)&=&\overline{m}_b(\overline{m}_b^2)\left[\frac{\alpha_{s}(\mu)}{\alpha_{s}(\overline{m}_b)}\right]^{\frac{12}{23 } } \ , , \nonumber\\ \alpha_s(\mu)&=&\frac{1}{b_0t}\left[1-\frac{b_1}{b_0 ^ 2}\frac{\log t}{t } + \frac{b_1 ^ 2(\log^2{t}-\log{t}-1)+b_0b_2}{b_0 ^ 4t^2}\right]\ , , \end{aligned}\ ] ] where @xmath115 , @xmath116 , @xmath117 , @xmath118 , @xmath119 , @xmath120 and @xmath121 for the flavors @xmath122 , @xmath123 and @xmath124 , respectively @xcite . , where @xmath125 is the squared mass of the pseudoscalar meson @xmath0.,width=264 ] in fig.7 , we plot the threshold @xmath126 with variations of the energy scales . from the figure , we can see that the threshold @xmath126 decreases quickly with increase of the energy scale , the energy scale should be larger than @xmath127 for the @xmath128 or @xmath129 system , we can take the typical energy scale @xmath130 , which corresponds to the threshold @xmath131 . on the other hand , if we take the pole masses @xmath132 and @xmath133 from the particle data group @xcite , the threshold @xmath134 is larger than the value @xmath125 of the squared mass of the pseudoscalar meson @xmath0 .
we have to choose much smaller values , @xmath135 and @xmath136 , which corresponds to the threshold @xmath137 .
furthermore , we choose the uncertainties as that of the @xmath107 masses from the particle data group tentatively @xcite .
the pole masses and the @xmath107 masses have the relation @xmath138 $ ] , we maybe expect that a simple replacement of the corresponding quantities in the spectral densities @xmath139 , @xmath140 and @xmath141 can lead to analogous results , such an expectation is sensible only in the case that the integral ranges @xmath142 and @xmath143 are large enough , the variations @xmath144 are small enough so as to be neglected . in the present case , the integral ranges are small , we have to fit the parameters independently . for the @xmath107 masses ,
we observe that the ideal parameters are @xmath145 [ @xmath146 and @xmath147 [ @xmath148 for the vector [ axialvector ] @xmath0 mesons , the corresponding pole contributions and the resulting masses and decay constants are presented in table 1 and figs.8 - 9 . for the pole masses , we observe that the ideal parameters are @xmath149 [ @xmath150 and @xmath147 [ @xmath148 for the vector [ axialvector ] @xmath0 mesons , the corresponding pole contributions and the resulting masses and decay constants are also presented in table 1 and figs.8 - 9 .
the threshold parameters and predicted masses satisfy the relations @xmath151 and @xmath152 , which are compatible with our naive expectation that the energy gap between the ground state and first radial excited is about @xmath101 .
the calculations based on the nonrelativistic renormalization group indicate that @xmath41 @xcite , the present prediction @xmath153 is satisfactory .
. the borel parameters , continuum threshold parameters , pole contributions , masses and decay constants of the vector and axialvector @xmath0 mesons .
the wide - hat denotes that the pole masses are used . [ cols="^,^,^,^,^,^,^,^",options="header " , ] for the heavy quarkonium states , the relative velocities @xmath154 of the quarks are small , we should account for the coulomb - like @xmath155 corrections .
after taking into account all the coulomb - like contributions , we obtain the coefficient @xmath156 to dress the leading - order spectral densities @xmath157 @xcite , @xmath158 if we take the approximation @xmath159 , then @xmath160 . in fig.10
, we plot the ratio @xmath161 of the leading - order spectral densities , where the @xmath107 masses are used . from the figure we can see that @xmath162 .
the terms in the next - to - leading order spectral density @xmath163 can not be factorized as @xmath164 lead to the behavior @xmath165 , where @xmath166 is a formal notation . of the leading - order spectral densities.,width=264 ] the next - to - leading order
spectral density @xmath167 can be approximated by @xmath168 .
we account for all the coulomb - like contributions by multiplying the leading - order spectral density @xmath169 by the coefficient @xmath156 tentatively , and obtain the central values @xmath170 with the @xmath107 masses ( pole masses ) , those predictions are also shown in tables 2 - 3 .
the mass - shifts are about @xmath171 , while the decay constant shifts are about @xmath172 .
in this article , we study the vector and axialvector @xmath0 mesons by including the next - to - leading order perturbative contributions in the operator product expansion with the qcd sum rules , and make reasonable predictions for the masses and decay constants , then calculate the leptonic decay widths .
the present predictions for the masses and decay constants can be confronted with the experimental data in the future at the lhc .
we can also take the masses and decay constants as basic input parameters and study other phenomenological quantities , such as the semi - leptonic , non - leptonic and radiative decays .
this work is supported by national natural science foundation , grant number 11075053 , and the fundamental research funds for the central universities .
the notations in the next - to - leading order spectral densities , @xmath173 c. h. chang , y. q. chen , g. p. han and h. t. jiang , phys . lett .
* b364 * ( 1995 ) 78 ; k. kolodziej , a. leike and r. ruckl , phys . lett .
* b355 * ( 1995 ) 337 ; c. h. chang , y. q. chen and r. j. oakes , phys . rev .
* d54 * ( 1996 ) 4344 ; k. cheung and t. c. yuan , phys . rev .
* d53 * ( 1996 ) 1232 ; k. cheung and t. c. yuan , phys . rev . *
d53 * ( 1996 ) 3591 ; i. p. gouz , v. v. kiselev , a. k. likhoded , v. i. romanovsky and o. p. yushchenko , phys .
* 67 * ( 2004 ) 1559 ; c. h. chang and x. g. wu , eur .
j. * c38 * ( 2004 ) 267 ; a. v. berezhnoy , a. k. likhoded and a. a. martynov , phys .
* d83 * ( 2011 ) 094012 .
e. bagan , h. g. dosch , p. gosdzinsky , s. narison and j. m. richard , z. phys .
* c64 * ( 1994 ) 57 .
m. chabab , phys . lett .
* b325 * ( 1994 ) 205 .
p. colangelo , g. nardulli and n. paver , z. phys .
* c57 * ( 1993 ) 43 .
s. narison , phys . lett .
* b210 * ( 1988 ) 238 .
k. g. chetyrkin , j. h. kuhn , a. maier , p. maierhofer , p. marquard , m. steinhauser and c. sturm , phys . rev .
* d80 * ( 2009 ) 074010 ; s. bodenstein , j. bordes , c. a. dominguez , j. penarrocha and k. schilcher , phys
. rev . * d83 * ( 2011 ) 074014 ; s. bodenstein , j. bordes , c. a. dominguez , j. penarrocha and k. schilcher , phys
* d85 * ( 2012 ) 034003 ; b. dehnadi , a. h. hoang , v. mateu and s. m. zebarjad , arxiv:1102.2264 . | in this article , we study the vector and axialvector @xmath0 mesons with the qcd sum rules , and make reasonable predictions for the masses and decay constants , then calculate the leptonic decay widths .
the present predictions for the masses and decay constants can be confronted with the experimental data in the future .
we can also take the masses and decay constants as basic input parameters and study other phenomenological quantities with the three - point vacuum correlation functions via the qcd sum rules .
+ zhi - gang wang + department of physics , north china electric power university , baoding 071003 , p. r. china pacs number : 14.40.pq , 12.38.lg key words : @xmath0 mesons , qcd sum rules |
in this paper we aim to determine the multiplicity - free permutation representations of the symmetric groups .
this question of which subgroups of the symmetric group have a multiplicity - free permutation representation arose from the combinatorial problem of determining the largest collection of set - partitions with a specific type of intersection .
we will start by describing the combinatorial problem and stating our results and then explain how it is related to the multiplicity - free representations of the symmetric group .
all the calculations performed in this paper are done using the computational discrete algebra system gap @xcite .
let @xmath0 be integers with @xmath1 .
a _ uniform @xmath2-partition of an @xmath3-set _ is a set - partition of the @xmath3-set into @xmath2 classes each of cardinality @xmath4 .
we say two such partitions @xmath5 and @xmath6 are _ qualitatively independent _ if each cell of @xmath5 and each cell of @xmath6 have at least one element in common . for practical reasons ,
we want large sets of pairwise qualitatively independent partitions @xcite . the _ qualitative independence graph _ @xmath7 is the graph whose vertices are the uniform @xmath2-partitions , where two partitions are adjacent if they are qualitatively independent . in this
setting , our problem is to find the cliques of maximum size in @xmath7 .
for @xmath8 the problem is trivial since the graph @xmath9 is a clique .
for @xmath2 a prime power , the size of the maximum clique of @xmath10 is known to be @xmath11 ( @xcite , section 5.4 ) .
very few bounds on the clique size of @xmath7 are known for other values of @xmath3 and @xmath2 .
one approach is to make use use of a family of graphs that generalises the graphs @xmath7 . if @xmath5 and @xmath6 are two uniform @xmath2-partitions , their _ meet table _ is the @xmath12 array with the @xmath13 entry @xmath14 .
two meet tables are _ isomorphic _ if we can obtain one from the other by permuting its rows and columns .
different orderings on the classes in the partitions @xmath5 and @xmath6 could produce different , but isomorphic meet tables . note that @xmath5 and @xmath6 are qualitatively independent if and only if all entries of their meet table are positive . for each possible meet table
@xmath15 , we define a graph whose vertices are the uniform partitions , where two partitions are adjacent if their meet table is isomorphic to @xmath15 . for example , if @xmath16 and all entries in @xmath15 are equal to one , then the corresponding graph is @xmath10 .
thus we have a family of graphs , one for each meet table , whose edges partition the edges of the complete graph on the uniform partitions .
if the adjacency matrices of these graphs commute , then standard eigenvalue methods can be used to obtain bounds on the maximum sizes of cliques .
mathon and rosa @xcite show that for @xmath17 and @xmath18 that the family of adjacency matrices of these graphs is commutative .
this result is used by godsil and newman to prove that the graph @xmath19 is a core @xcite .
it is conjectured in @xcite that the graph @xmath10 is a core for all values of @xmath2 .
a first step towards this is to determine if the adjacency matrices of these graphs for @xmath16 commute for all values of @xmath2 .
how can we decide if these matrices commute ?
the key fact is that the symmetric group @xmath20 acts as a group of permutations on the set of uniform partitions , and the matrices we have just defined commute with each other if and only if this permutation representation is _ multiplicity free _ , that is , each irreducible representation of @xmath20 occurs in it with multiplicity 0 or 1 .
( see @xcite , section 29 . )
we have determined all multiplicity - free permutation representations of the symmetric group .
saxl @xcite gives a list of all the possible subgroups of @xmath20 for @xmath21 which could be multiplicity - free .
our work extends saxl s list not only by including groups acting on fewer than 18 points but also by determining exactly which groups in his list are multiplicity free .
mark wildon @xcite has also determined which of the groups in saxl s list are multiplicity free , his focus is on groups acting on at least 66 points .
( saxl @xcite)[saxl ] let @xmath22 and @xmath23 .
if @xmath24 is multiplicity free , then one of the following holds : a. @xmath25 for some @xmath2 with @xmath26 ; b. @xmath27 and @xmath28 ; c. @xmath27 and @xmath29 with @xmath30 \leq 4 $ ] ; d. @xmath31 and @xmath24 fixes a point of @xmath32 $ ] and is one of the subgroups in ( ii ) or ( iii ) on the rest of @xmath32 $ ] ; e. @xmath33 where @xmath34 or 9 and @xmath35 is @xmath36 , @xmath37 or @xmath38 respectively .
bannai @xcite determined all the maximal subgroups with permutation rank less than or equal to five of the finite symmetric groups .
all of these representations are multiplicity free and , naturally , occur in our list as well .
there are several infinite families of groups whose permutation representations is multiplicity free and a long list of groups that do not belong to any of these families .
a complete list of the groups can be found in the tables in the appendix .
table [ table : transitive ] is a list of all the transitive primitive subgroups .
table [ table : imprimitive ] contains all the transitive imprimitive subgroups and the intransitive groups are given in tables [ table : intrans ] and [ table : intrans2 ] .
we list the infinite families here .
many of the groups in the list are subgroups of others in the list , we organized the list so a subgroups are below the groups that contains them .
[ thm : infinitefam ] the table below is a complete list of the infinite families of subgroups of the symmetric groups whose permutation representation is multiplicity free . [ cols="<,<,^",options="header " , ]
the eigenvalues for the association schemes were calculated using akihide hanaki s gap package `` elementary functions for association schemes on gap''@xcite . c. j. colbourn , j. h. dinitz , and d. r. stinson .
applications of combinatorial designs to communications , cryptography , and networking . in
_ surveys in combinatorics , 1999 _ , pages 37100 .
cambridge univ . press ,
cambridge , 1999 . | we determine all the multiplicity - free representations of the symmetric group .
this project is motivated by a combinatorial problem involving systems of set - partitions with a specific pattern of intersection . |
molecules are usually treated as purely coulombic systems , while the strong interaction between their nuclear constituents is assumed to play a negligible role .
however any coulombic molecular level lying above the lower threshold of the nuclear subsystem , is embedded in the continuous spectrum of the nuclear sub - hamiltonian .
the coupling between the molecular and nuclear channels , hence , turns this level into a resonance ( see , e. g. , refs .
@xcite and references cited therein ) .
of course , due to the wide coulombic barrier between the nuclei and the short - range character of the nuclear interaction , this coupling , and thus the width of the resonance , which determines the fusion probability of the nuclear constituents of the molecule , is in general extremely small .
however , as pointed out in @xcite , the situation may be rather different if the nuclear subsystem of a molecule has a sufficiently narrow near - threshold resonance .
examples of such nuclear systems may be read off from the data presented in @xcite . among them
are even customary systems like @xmath4o and @xmath5o @xcite , i.e. , the nuclear constituents of the water molecule @xmath6 or the hydroxyl ion @xmath7 with @xmath8 being the isotope @xmath9o . for @xmath10 and @xmath6 the influence of near - threshold nuclear resonances on the molecular properties has been studied in @xcite by estimating the overlap integrals between the corresponding molecular and nuclear wave functions .
the best known example of such phenomena is the muon catalyzed fusion of deuteron and triton in the @xmath11 molecule , where the near - threshold nuclear resonance @xmath12he@xmath13 plays a decisive role @xcite .
being motivated by the above special cases , we deal in the present work like in @xcite with a rather general model hamiltonian related to the ones considered by friedrichs in @xcite .
this hamiltonian consists of a nuclear part , a molecular part with eigenvalues embedded in the continuous spectrum of the nuclear part , and a weak coupling term which turns these unperturbed eigenvalues into molecular resonances . since the model is explicitly solvable
, the mechanism of formation of the resonances becomes clearly visible .
the following property pointed out in @xcite appears , in particular , as a general feature : if the nuclear channel itself has a narrow resonance with a position close to the molecular energy , then the width ( the imaginary part ) of the resulting molecular resonance is found to be inversely proportional to the nuclear width . in other words , a large increase of the decay rate of the molecular state , i.e. of the fusion probability , is observed in this case .
such a coincidence of nuclear and molecular energies is , of course , a very rare phenomenon in nature .
influence of the narrow pre - threshold resonances on the properties of infinite crystalline molecular structures was studied in @xcite . in the present work we concentrate on more realistic finite crystals .
a goal of this work is to show that the decay rate of a molecular state with the energy close to a near - threshold resonance may be considerably enhanced when arranging molecular clusters within a finite crystalline structure .
the reason is that in such a configuration the original discrete molecular energy turns into a set of energy levels .
that is , even if the position of the nuclear resonance differs from the original molecular level , it can get within this set .
this allows for a fine tuning by exciting the crystalline structure to energies as close as possible to the energy of the nuclear resonance .
we show that the lattice states , which correspond to such an initial choice of their quasimomentum distribution , decay exponentially with a rate which is again inversely proportional to the width of the nuclear resonance .
in this section we recall our main reasoning @xcite regarding an influence of a near - threshold nuclear resonance on the width of a molecular resonance in the case of a single molecule .
let us consider a two - channel hilbert space @xmath14 consisting of a nuclear hilbert space @xmath15 ( channel 1 ) and a one - dimensional molecular space @xmath16 ( channel 2 ) .
the elements of @xmath17 are represented as vectors @xmath18 where @xmath19 and @xmath20 ( @xmath21 is simply a complex number ) .
the inner product @xmath22 in @xmath17 is naturally defined via the inner products @xmath23 in @xmath15 and @xmath24 in @xmath25 . as a hamiltonian in @xmath17 we consider the @xmath26 operator matrix @xmath27 where @xmath28 is the ( self - adjoint ) `` nuclear hamiltonian '' in @xmath15 , and @xmath29 a trial molecular energy .
a vector @xmath30 provides the coupling between the channels .
it should be mentioned that the hamiltonian ( [ h2 ] ) resembles one of the well known friedrichs models @xcite . if there is no coupling between the channels , i.e. for @xmath31 , the spectrum of @xmath32 consists of the spectrum of @xmath28 and the additional eigenvalue @xmath33 .
we assume that the continuous spectrum @xmath34 of the hamiltonian @xmath28 is not empty and that the eigenvalue @xmath33 is embedded in @xmath34 .
it is also assumed that @xmath33 is not a threshold point of @xmath34 , and that this spectrum is absolutely continuous in a sufficiently wide neighborhood of @xmath33 .
a nontrivial coupling @xmath35 between the channels will , in general , shift the eigenvalue @xmath33 into an unphysical sheet of the energy plane .
the resulting perturbed energy appears as a resonance , i.e. , as a pole of the analytic ( or , more precisely , meromorphic ) continuation of the resolvent @xmath36 taken between suitable states ( see , e.g. , @xcite ) . in the present work
we assume that such a continuation through the absolutely continuous spectrum of @xmath28 in some neighborhood of @xmath33 is possible at least for the matrix element @xmath37 of the resolvent @xmath38 .
then , from the explicit representation for the resolvent @xmath39 @xcite , one can easily see that the operator - valued function @xmath40 admits meromorphic continuation to the same neighborhood , too .
the poles of @xmath39 on the physical sheet are either due to zeros of the _ transfer function _
( see @xcite ) @xmath41 or due to poles of the resolvent @xmath42 ( see @xcite ) .
the latter correspond to the discrete spectrum of the operator @xmath28 which may determine part of the point spectrum of @xmath32 .
this is true , in particular , for the multiple eigenvalues of @xmath28 . in any case
it is obvious that the perturbation of the eigenvalue @xmath33 only corresponds to solutions of the equation @xmath43 , i.e. , of @xmath44 this equation has no roots @xmath45 with @xmath46 on the physical sheet .
therefore , being eigenvalues of the self - adjoint operator @xmath32 , they have to be real .
thus , eq . ( [ h2res ] ) may have solutions only on the real axis and in the unphysical sheet(s ) of the riemann surface of the resolvent @xmath42 .
we start with a brief discussion of the case where the nuclear channel hamiltonian @xmath28 generates no resonances close to @xmath33 in a domain @xmath47 of the unphysical sheet which ajoins the physical sheet from below the cut .
this assumption implies that for a wide set of unit vectors @xmath48 the quadratic form @xmath49 can be analytically continued in @xmath47 . moreover , under certain smallness conditions for @xmath50 , eq .
( [ h2res ] ) is uniquely solvable @xcite in @xmath47 providing in the main order ( see , e.g. , @xcite ) @xmath51 the real and imaginary parts of the resonance @xmath52 , thus , are given by @xmath53 our main interest concerns the opposite case of a nuclear resonance @xmath54 , @xmath55 , with a real part @xmath56 close to @xmath33 . for the sake of simplicity we assume the corresponding pole of @xmath42 to be of first order .
let the element @xmath30 be such that the function @xmath57 admits an analytic continuation into a domain @xmath47 which contains both points @xmath33 and @xmath58 .
this domain , moreover , is assumed to belong to the unphysical sheet which adjoins the physical sheet along the upper rim of the cut . in @xmath47
the function @xmath57 , thus , can be written as @xmath59 with @xmath60 being a holomorphic function . for a fixed `` structure function '' we have @xmath61 with a constant @xmath62 determined by the residue of @xmath42 at @xmath63 . note that this residue is usually expressed in terms of resonance ( gamow ) functions ( see for example @xcite ) .
in fact , we assume that the resonance corresponds to an `` almost eigenstate '' of @xmath28 . that is , in principle a limiting procedure @xmath64 is possible so that the resonance turns into a usual eigenvalue with an eigenvector @xmath65 more precisely , we assume @xmath66 with @xmath67 .
this can be achieved , e.g. , if the hamiltonian @xmath28 itself has a matrix representation of the form ( [ h2 ] ) and the resonance @xmath58 is generated by a separated one - dimensional channel .
in such a case we would have @xmath68 ( for details see ref .
@xcite , sec.ii ) .
let @xmath69 and , for @xmath70 , @xmath71 with constants @xmath72 and @xmath73 .
furthermore , the coupling between the channels in the hamiltonian ( [ h2 ] ) is assumed to be so weak that @xmath74 itself is very small . after inserting ( [ betarepres ] ) for @xmath57 , eq .
( [ h2res ] ) turns into the `` quadratic '' equation @xmath75 which can be `` solved '' , i.e. , can be rewritten in form of two equations @xmath76 by banach s fixed point theorem , each of the equations ( [ h2roots ] ) has only one solution in the domain @xmath47 .
in case of the sign `` @xmath77 '' we denote the root of ( [ h2roots ] ) by @xmath78 , in case of the sign `` @xmath79 '' by @xmath80 . according to @xcite and @xcite , the roots @xmath78 and @xmath80 of ( [ h2roots ] )
are essentially given by @xmath81 from the second condition ( [ h2conditions ] ) follows @xmath82 .
consequently , this term provides in @xmath78 a very small perturbation of the initial nuclear resonance @xmath58 .
as compared to @xmath83 it represents also in @xmath80 a very weak perturbation of the molecular energy @xmath33 .
however , as compared to the result ( [ h2resonance ] ) , valid in case of a missing nearby nuclear resonance , it can be rather large .
in particular , if the molecular energy @xmath33 coincides with the real part @xmath56 of the nuclear resonance @xmath58 , then @xmath84 with @xmath85 _ the width of the molecular resonance @xmath80 in the presence of a nearby nuclear resonance @xmath58 , thus , turns out to be inversely proportional to the nuclear width @xmath83_. the second inequality ( [ h2conditions ] ) , chosen as a condition for @xmath50 reflects the fact that the `` usual '' molecular width @xmath86 is much smaller than the width of a usual nuclear resonance @xmath83 , @xmath87 this can practically always be assumed for concrete molecules . under condition ( [ alimit ] )
the value of @xmath88 differs from zero , @xmath89 , as @xmath64 .
therefore , in the presence of a narrow ( @xmath90 ) nuclear resonance close to @xmath33 the molecular width @xmath0 is much larger than the molecular width @xmath86 observed in absence of such a resonance .
in fact , this ratio is determined by the large quotient @xmath91 .
let us assume that the `` molecules '' described by the hamiltonian ( [ h2 ] ) are arranged in form of an finite one - dimensional linear ( chain ) crystalline structure . to describe such a crystal we introduce the lattice hilbert space @xmath92 representing an orthogonal sum of the hilbert spaces associated with the individual cells @xmath93 here the subspaces @xmath94 and @xmath95 are exactly the same ones as in sec .
[ twochannel ] and , thus , @xmath96 .
the elements of the total hilbert space @xmath97 are represented by the sequences @xmath98 with components @xmath99 where @xmath100 and @xmath101 .
the inner product in @xmath17 is defined by @xmath102 the subspaces @xmath103 and @xmath104 , with @xmath105 , represent respectively the pure nuclear and pure molecular channels . in the present section
we will first deal with the hamiltonian @xmath106 acting in @xmath17 according to @xmath107 where only the interaction between neighboring cells is taken into account and the interaction operator @xmath108 is chosen in the simplest form @xmath109 with @xmath110 being a positive number .
such a choice of the interaction corresponds to the natural assumption that the cells interact between each other via the molecular states , while the direct interaction between nuclear constituents belonging to different cells is negligible .
we assume that the closed interval is totally embedded in the continuous spectrum @xmath34 of @xmath28 and , moreover , that no thresholds of @xmath34 belong to this interval .
for the sake of simplicity we also assume that the interval belongs to the domain @xmath47 introduced in sec .
[ twochannel ] and that for any @xmath111 obviously , the hamiltonian ( [ 1dimlatham ] ) is a self - adjoint operator on the domain @xmath112 with @xmath113 .
the resolvent @xmath114 of @xmath106 possesses a natural block structure , @xmath115 , @xmath116 .
the blocks @xmath117 satisfy the equations @xmath118 where @xmath119 stands for the kronecker delta and @xmath120 for the identity operator in the hilbert space @xmath17 of cells .
hereafter we assume @xmath46 so that the value of @xmath45 automatically belongs to the resolvent set of the operator @xmath106 .
the blocks @xmath117 themselves possess a @xmath26 matrix structure , @xmath121 , @xmath122 corresponding to the decomposition .
the set of the sequences @xmath123 , @xmath124 , with the elements @xmath125 forms an orthonormal basis for the ( @xmath126-dimensional ) hilbert space @xmath127 of @xmath126-element sequences of the form @xmath128 , @xmath129 , @xmath130 .
the fourier transform @xmath131 in @xmath97 reduces eq .
( [ rjk ] ) to @xmath132 and the numbers @xmath133 represent the matrix elements of the resolvent @xmath134 in this representation . from ( [ rpp ] )
it immediately follows that @xmath135 where @xmath136 here , the scalar function @xmath137 reads @xmath138 the numbers @xmath139 given by ( [ pk ] ) represent the quasimomenta of the finite crystalline structure under consideration .
consider now the time evolution of the system described by the hamiltonian @xmath106 starting from a pure molecular state @xmath140 , @xmath141 , @xmath142 , with @xmath143 and @xmath144 .
the probability to find the system at a time @xmath145 in the molecular channel is given by @xmath146 where @xmath147 is the orthogonal projection in @xmath97 on the pure molecular subspace @xmath148 .
obviously , one can represent the time evolution operator @xmath149 in terms of the resolvent @xmath114 , @xmath150 where the integration is performed along a counterclockwise contour @xmath151 in the physical sheet encircling the spectrum of the hamiltonian @xmath106 . according to eqs .
( [ rfactor ] ) and ( [ trfunctmodif ] ) the operator @xmath152 acts in quasi - momentum representation as the multiplication operator , @xmath153 here @xmath154 stands for the values of the fourier transform ( [ fourier ] ) of the vector @xmath155 hence @xmath156 with @xmath157 repeating almost literally the analysis of section iii in @xcite one finds that the asymptotics of the term @xmath158 as @xmath159 reads as follows @xmath160 \\ \label{jptbehavior } & \quad\quad + & \exp\{-{\rm i}z_{\rm nucl}(p ) t\ } \\
\nonumber & & \times\biggl[{\displaystyle}\frac{a}{\bigl(\widetilde{\lambda}_2(p_k)-z_1- \beta^{\rm reg}(z_1)\bigr)^2 } + o\bigl(\varepsilon^4(p_k , z_1)\bigr)\biggr ] + \widetilde{\varepsilon}(p_k , t)\ , , \ ] ] where @xmath161 ^ 2}\ , . \ ] ] the function @xmath162 is always small , @xmath163 . by ( [ znucl ] ) and ( [ zmol ] ) for the positions of the resonance poles we then obtain @xmath164 further ,
let us assume that the number @xmath126 of cells in the lattice is large and the real part @xmath56 of the nuclear resonance @xmath58 belongs to the interval @xmath171 $ ] , that is @xmath172 .
then , one can always prepare an initial molecular state @xmath173 which decays via the nuclear channel with a rate close to @xmath174 ( cf .
formula ( [ gmfin ] ) ) . indeed , under the assumption ( [ reaima ] ) , this maximum is given by @xmath175 the most appropriate is the monochromatic molecular state @xmath173 with the only nonzero component @xmath176 associated with the quasimomentum @xmath177 closest to @xmath178 in particular , if the values of @xmath154 are nonzero only for quasimomenta @xmath139 restricted by @xmath179 with some small @xmath180 , then the width @xmath0 given by the relation ( [ gammarm ] ) varies in an interval lying approximately between @xmath181 and @xmath182 . in a similar way one
also treats a one - dimensional cyclic crystalline structure . in this case
the hilbert space @xmath97 is the same as in ( [ cg ] ) but the operator @xmath106 reads @xmath183 if the intercellur interaction is still given by ( [ w1dim ] ) , the only difference in the analysis will be the use of another complete orthonormal set in the space @xmath127 . instead of the sequences ( [ fk ] )
one now employs the ortonormal sequences @xmath184 with quasimomenta @xmath139 given by @xmath185 after the fourier transform ( [ fourier ] ) with @xmath186 the matrix elements of the resolvent @xmath187 again acquire the form ( [ rfactor ] ) , ( [ gpz ] ) with the transfer function @xmath188 given by ( [ trfunctmodif ] ) .
hence , one concludes with formulas like in ( [ p2phifactor])([gammarm ] ) and then observes that if the number of cells is large enough it is possible to prepare pure molecular states that decay with the rate close to @xmath189 . in the same way one can also consider the finite two- and three - dimensional crystalline structures arranged of the molecular cells described by the hamiltonian ( [ h2 ] ) . if the cell has a sharp near - threshold nuclear resonance with energy embedded into the convex hull of the arising crystalline molecular levels , one again will find an enhancement of the decay rate for particular molecular states . as in the case of the one - dimensional lattices
these molecular states should decay with the rate close to @xmath189 . | we consider finite linear or cyclic crystalline structures with molecular cells having narrow pre - threshold nuclear resonance . we prove that if the real part of such a nuclear resonance lies within the energy band ( the convex hull of the energy levels ) of the crystalline structure arising of a separated molecular level , then there exist molecular crystalline states that decay exponentially in time and the decay rate @xmath0 of these states in the main order is described by the formula @xmath1 where @xmath2 is the value of the residue of the molecular channel transfer function at the nuclear resonance point and @xmath3 is the nuclear resonance width . |
the milky way contains @xmath3 of stars .
approximately 90% of those stars lie in the disk , 9% in the bulge , and 1% in the halo @xcite .
however , these structures did not form their stars en masse .
the sites of star formation are far less massive than any of these components .
for example , star formation in the disk occurs in open clusters , which typically have just @xmath4 of stars @xcite .
dwarf galaxies , which may have built the bulge and halo @xcite , range in mass from @xmath5 ( e.g. , segue 1 ; * ? ? ?
* ) to more than @xmath6 ( e.g. , fornax ; * ? ? ?
* ; * ? ? ?
the pockets of star formation within the dwarf galaxies likely contained just as little mass as the open clusters in the galactic disk .
dynamical processes assemble the many disparate sites of star formation into the primary galactic components , such as the disk , bulge , and halo .
therefore , the key to discovering the origins of stars lies in studying stellar populations in less massive structures .
the dwarf galaxies that orbit the milky way ( mw ) present the opportunity to study the formation of stellar populations most similar to the mw stellar halo . unlike most open clusters , dwarf galaxies contain old , metal - poor stars .
unlike most globular clusters , dwarf galaxies enable the study of temporally extended star formation , which ultimately results in chemical evolution .
furthermore , galaxies of this class probably created at least some of the stellar halo of the mw by gravitational dissolution ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
if so , then the stellar population of the present mw halo contains a mixture of many different galaxies .
the individual stellar populations are difficult to disentangle , especially in the inner halo ( e.g. , * ? ? ?
surviving dwarf galaxies may offer a look at single , though evolved , counterparts to an accreted component of the halo .
the stellar populations in surviving and accreted dwarfs differ because the accretion time is correlated with stellar mass and therefore star formation history .
nonetheless , the observed abundance differences can be compared with models of surviving galaxies and hierarchically formed halos ( e.g. , * ? ? ?
* ; * ? ? ?
alternatively , the halo may have been created by a combination of monolithic collapse @xcite and hierarchical assembly .
comparisons of the abundances of dwarf galaxy stars and the chemodynamics of halo stars ( e.g. , * ? ? ?
* ; * ? ? ?
* ) show that only some of the stars in the halo are consistent with accretion of dwarf galaxies similar to the surviving dwarfs .
although it is clear that the nearby halo stars have different abundance patterns than dwarf satellite galaxies , the number of dwarf galaxy stars with published multi - element abundance measurements is less than for the halo .
the relative distance of dwarf galaxies compared to the nearest observable halo stars has prevented the sample of dwarf galaxy abundances from growing as fast as for halo stars .
studying the role of dwarf galaxies in building the halo would benefit from a larger sample of dwarf galaxy elemental abundances .
( @xcite @xcite points out that even the halo abundances have not been well - sampled because the more distant halo stars have not been observed nearly as completely as nearby halo stars . )
stellar elemental abundances reveal many characteristics of the stellar population of a galaxy .
its metallicity distribution , usually represented by the stellar iron content , is a function of the galaxy s total mass , the amount of gas infall and outflow , the chemical yield of supernovae , whether the galaxy formed out of gas pre - enriched with metals , and other variables . beyond the one - dimensional metallicity distribution , the ratios of other elements to iron reflect other details of the star formation history . in particular , the ratios of alpha elements , such as magnesium , to iron depend on the duration and intensity of star formation .
a large sample of multi - element abundances for individual stars in dwarf galaxies would be a useful tool in studying star formation in small systems .
however , obtaining elemental abundances is expensive .
high - quality , high - resolution spectroscopy ( hrs , @xmath7 ) allows the measurements of tens of elements to precisions of @xmath8 dex for stars with @xmath9 .
this magnitude limit restricts the measurement of elemental abundances to the upper 12 magnitudes of the red giant branch for most mw dwarf galaxies .
in addition , most current high - resolution spectrographs operate on one star at a time . the exposure time required for adequate signal to derive accurate abundance measurements has limited the sample of stars in intact mw dwarf satellite galaxies ( excluding sagittarius ) with published high - resolution spectroscopy to fewer than 200 . with current technology , a statistical analysis of the elemental abundance distributions in dwarf galaxies
must rely on medium - resolution spectroscopy ( mrs , @xmath10 ) .
mrs relaxes the magnitude limit to @xmath11 while permitting multiplexing of more than one hundred stars in one exposure . building a large sample of multi - element abundances in mw dwarf galaxies
requires both of these capabilities .
the disadvantages of mrs are that fewer elements are accessible than with hrs , and the measurements are less precise .
@xcite presented one of the first applications of mrs to dwarf galaxy metallicity distributions .
they based their metallicity measurements on a calibration of infrared calcium triplet equivalent width to [ fe / h ] ( e.g. , * ? ? ?
* ; * ? ? ?
however , spectral synthesis is required to measure the abundances of multiple elements .
several recent spectral synthesis - based studies have employed mrs to begin building the sample of abundances in mw dwarf galaxies .
@xcite discovered extremely metal - poor ( @xmath12 } } < -3 $ ] ) stars in the ultra - faint dwarf galaxies using keck / deimos spectroscopy .
@xcite measured fe , mg , ca , and ti abundances for 27 stars in the leo ii galaxy using keck / lris spectroscopy .
finally , ( * ? ? ?
* paper i ) measured fe , mg , si , ca , and ti abundances for 388 stars in the sculptor dwarf galaxy using keck / deimos spectroscopy . in this article , we present a catalog of elemental abundance measurements of 2961 stars in eight mw dwarf galaxies .
the technique is very similar to that developed by ( * ? ? ?
* kgs08 ) and modified as noted in @xcite .
we describe the observations in section [ sec : observations ] . in section [ sec : measurements ] , we briefly summarize the technique and present the catalog . in section [ sec : accuracy ] , we explore the accuracy of the measurements using repeat observations and comparison to hrs measurements of the same stars .
we summarize our work and discuss our articles that interpret these data in section [ sec : summary ] .
the next papers in this series apply the data presented here to deducing the past chemical evolution histories of the eight dwarf spheroidal galaxies ( dsphs ) .
@xcite @xcite includes galactic chemical evolution models of the metallicity distributions .
@xcite @xcite shows how the [ @xmath13/fe ] ratios change with [ fe / h ] , a diagnostic of the past intensity of star formation .
both papers examine trends with the total luminosity of the dsph .
this large data set , encompassing a wide luminosity range of mw satellites , will uniquely enable an exploration of the dependence of chemical evolution on dwarf galaxy mass .
lc@ c@ c@ c + ngc 288 & @xmath14 & @xmath15 & 8.8 & 14.74 + ngc 1904 ( m79 ) & @xmath16 & @xmath17 & 12.9 & 15.56 + ngc 2419 & @xmath18 & @xmath19 & 84.3 & 19.63 + ngc 5904 ( m5 ) & @xmath20 & @xmath21 & 7.5 & 14.37 + ngc 6205 ( m13 ) & @xmath22 & @xmath23 & 7.6 & 14.42 + ngc 6341 ( m92 ) & @xmath24 & @xmath25 & 8.2 & 14.58 + ngc 6838 ( m71 ) & @xmath26 & @xmath27 & 4.0 & 13.02 + ngc 7006 & @xmath28 & @xmath29 & 41.4 & 18.09 + ngc 7078 ( m15 ) & @xmath30 & @xmath31 & 10.3 & 15.06 + ngc 7089 ( m2 ) & @xmath32 & @xmath33 & 11.5 & 15.30 + pal 13 & @xmath34 & @xmath35 & 25.8 & 17.05 + ngc 7492 & @xmath36 & @xmath37 & 25.8 & 17.06 + sculptor & @xmath38 & @xmath39 & 85 & 19.67 + fornax & @xmath40 & @xmath41 & 139 & 20.72 + leo i & @xmath42 & @xmath43 & 254 & 22.02 + sextans & @xmath44 & @xmath45 & 95 & 19.90 + leo ii & @xmath46 & @xmath47 & 219 & 21.70 + canes venatici i & @xmath48 & @xmath49 & 210 & 21.62 + ursa minor & @xmath50 & @xmath51 & 69 & 19.18 + draco & @xmath52 & @xmath53 & 92 & 19.84 + the medium - resolution abundances in this article depend on spectroscopy obtained with the deep imaging multi - object spectrograph ( deimos , * ? ? ?
* ) on the keck ii telescope and photometry from a variety of sources .
this section explains how spectroscopic targets were selected from photometric catalogs and how the deimos observations were performed .
every spectroscopic target is an individual star in one of three types of stellar system : globular cluster , the milky way halo , and dwarf galaxy .
all of the target systems except the mw halo field stars are listed in table [ tab : targets ] .
the mw halo field stars are included in table [ tab : hrscompare ] .
the globular cluster and halo field stars are interesting in their own right .
the globular cluster stars in particular will be examined in further detail in a future work .
however , we examine the globular cluster and halo field stars here only to assess the accuracy of the abundance measurement technique . the observations most relevant to this article
s scientific focus are the individual dsph stellar spectra
. stars in most globular clusters ( gcs ) are excellent metallicity standards because every star in a given cluster has the nearly the same iron abundance and the same heliocentric distance .
exceptions include @xmath54 cen @xcite and m22 @xcite .
we targeted individual stars in twelve monometallic gcs , listed in table [ tab : targets ] .
we check stellar abundances within each cluster for internal consistency , and we also check the deimos - derived abundances against hrs abundance measurements .
@xcite has made photometry publicly available for all of these clusters as part of his photometric standard field database .
every star has been measured in at least two of the johnson - cousins @xmath55 , @xmath56 , @xmath57 , and @xmath58 filters .
this database does not contain every star in each field , but only those stars suitable for photometric standardization .
therefore , p. b. stetson ( private communications , 2007 , 2008 ) generously provided complete photometric data for our targets in m13 , m71 , ngc 7006 , m15 , and ngc 7492 .
we supplemented stetson s photometry with additional photometry for five clusters , mostly to increase the field of view so that it spanned a full deimos slitmask .
for m79 , we utilized the @xmath59 measurements of @xcite and the @xmath60 measurements of @xcite .
for m5 , we supplemented stetson s photometry with the @xmath61 sloan digital sky survey ( sdss ) crowded field photometry of @xcite . for consistency with the other data in the johnson - cousins system , we converted sdss @xmath61 to johnson - cousins @xmath62 following the global , metallicity - independent transformations of @xcite .
all magnitudes were corrected for extinction based on @xmath63 from @xcite s ( 1996 , updated 2003 ) catalog of gc properties . in order to arrive at his values , harris averaged measurements of @xmath63 by @xcite , @xcite , @xcite , and the authors listed in table [ tab : targets ] .
not all targets could be observed spectroscopically due to the limited field of view and slitmask design constraints .
first , we attempted to maximize the number of target stars previously observed with high - resolution spectroscopy ( hrs ) by placing the slitmasks in regions with a high density of hrs targets .
we selected the remaining targets based on the clusters color - magnitude diagrams ( cmds ) . in order of priority
, we filled each slitmask with stars from the ( 1 ) upper red giant branch ( rgb ) , ( 2 ) lower rgb , ( 3 ) red clump , and ( 4 ) blue horizontal branch . for the gcs with angular sizes smaller than the deimos field of view , we filled the slits at the edges of the slitmask far from the center of the gcs with objects having similar colors and magnitudes to stars on the rgb .
@xcite prioritized targets for the n1904 and n2419 slitmasks ( see table 2 ) differently .
we refer the reader to their article for further information .
the most metal - poor mw globular cluster known is m15 ( @xmath12}}= -2.38 $ ] ; * ? ? ? * ; * ? ? ?
* ; * ? ? ?
dwarf galaxies contain even more metal - poor stars ( e.g. , * ? ? ?
* ; * ? ? ?
in order to verify that the abundance measurements based on mrs are accurate at these very low metallicities , we obtained deimos spectra of metal - poor milky way halo field stars having hrs data .
we chose several studies of metal - poor halo field stars : the keck high - resolution echelle spectrograph ( hires ) and lick / hamilton spectrograph measurements of @xcite at @xmath64 ; the keck / hires measurements of @xcite , @xcite , and @xcite ; the kitt peak national observatory coud ' e spectrograph measurements of @xcite ; and the keck echellette spectrograph and imager ( esi ) measurements of @xcite at @xmath10 .
each of these authors selected targets from low - resolution surveys for metal - poor stars in the galactic halo ( * ? ? ?
? * ; * ? ? ?
? * and additional references from @xcite @xcite ) .
these abundance measurements are based on equivalent width analysis and/or spectral synthesis .
not all of the hundreds of stars in these studies could be observed .
the highest priority was assigned to the most metal - poor stars . strictly speaking , @xcite did not conduct an hrs study because their keck / esi spectra had a spectral resolution of @xmath65 .
we still include their sample in the comparison along with the truly high - resolution studies .
@xcite describes in detail the target selection for stars in the sculptor dsph .
for convenience , we reproduce here the map of targets on the sky and the sculptor cmd ( fig .
[ fig : scl ] ) .
we adopted the @xmath66 photometric catalog of @xcite , and we selected targets based on position in the cmd .
magnitudes were corrected star by star for extinction determined from the @xcite dust maps .
( the same dust maps were not used for gcs because gcs subtend a smaller solid angle than dsphs .
therefore , we used a single value for each gc rather than interpolation in the @xcite dust maps . )
targets were drawn from seven polygons surrounding the rgb and horizontal branch . in order from highest to lowest priority ,
the targets were selected ( 1 ) from @xmath67 to the tip of the rgb , ( 2 ) @xmath68 , ( 3 ) @xmath69 , ( 4 ) @xmath70 , and ( 5 ) @xmath71 .
the red and blue edges of the polygons extended until the stellar density reached the background level , with a typical width of @xmath72 centered on the rgb .
figure [ fig : for ] shows the fornax field with the spectroscopic targets highlighted in red . to minimize confusion ,
only targets with @xmath73 are plotted .
the five deimos slitmasks are for1b , for3b , for4b , for6 , and for7 ( see table 2 ) .
each slitmask except for4b included at least one of 18 duplicate targets included on other slitmasks .
figure [ fig : for ] also shows the cmd of the targets within the right ascension and declination ranges of the axes in the left panel of fig .
[ fig : for ] .
some stars on the extreme blue end of the rgb did not pass the spectroscopic selection criteria .
these stars may be asymptotic giant branch ( agb ) stars , or they may be extremely young or extremely metal - poor red giants .
this potential selection bias should be kept in mind when considering the derived metallicity distribution of fornax .
twenty stars , listed in table [ tab : hrscompare ] , have previously published hrs abundance measurements @xcite , and all 20 were observed .
one of these stars , m12 ( not shown in the right panel of fig .
[ fig : for ] ) , was absent from the @xmath66 photometric catalog . for calculation of photometric temperature and surface gravity
, we adopted the same extinction - corrected @xmath60 magnitudes used by @xcite .
we used @xcite s ( @xcite ) deimos spectra of individual stars in leo i and @xmath74 magnitudes of the spectroscopic targets .
we converted @xmath75 and @xmath76 to cousins @xmath56 and @xmath58 magnitudes in the same manner as for the sculptor photometric catalog .
magnitudes were corrected star by star for extinction determined from the @xcite dust maps . for other details on the photometric catalog and spectroscopic target selection ,
we refer to @xcite s published paper on the deimos slitmasks leoi_1 and leoi_2 .
slitmasks .
figure [ fig : leoi ] shows the sky chart and cmd for spectroscopic targets .
the black points in the right panel of the figure show other stars from the @xmath60 catalog of @xcite for context , but the spectroscopic targets were not drawn from this catalog . @xcite
measured spectroscopic abundances of two stars in leo i. one of these , m5 , was observed on three different deimos slitmasks .
we selected spectroscopic targets from the deep , wide - field @xmath77 catalog of @xcite .
they assumed a constant reddening of @xmath78 , and we corrected all of the magnitudes and colors accordingly .
we selected rgb members by overlaying yonsei - yale isochrones @xcite between 2 and 14 gyr , @xmath12}}= -3.76 $ ] and @xmath79 , and @xmath80}}= 0.0 $ ] and @xmath81 on the cmd .
all stars that lay between the bluest and reddest of these isochrones within photometric errors were considered for spectroscopic selection .
an additional 0.05 mag was allowed on the blue edge of the bluest isochrones to account for possible systematic error in the isochrones which might have excluded extremely metal - poor stars . when forced to choose between multiple rgb candidates
, we selected the brightest one .
figure [ fig : sex ] shows the sextans field with the spectroscopic targets highlighted in red .
only targets with @xmath82 are plotted to minimize confusion .
the sex3 slitmask contained 25 targets also included on other slitmasks .
figure [ fig : sex ] also shows the cmd of the targets within the right ascension and declination ranges of the axes in the left panel of fig .
[ fig : sex ] .
some extremely red stars were targeted in order to fill the slitmask with targets . although these stars are unlikely to be sextans members , they were included because they could potentially be metal - rich rgb stars in sextans .
additionally , sextans is near enough to permit a significant number of horizontal branch stellar spectra , although the spectroscopic abundance measurement technique used here does not yet work for horizontal branch stars .
five stars , listed in table [ tab : hrscompare ] , have previously published hrs abundance measurements @xcite , and all five were observed .
@xcite published measurements of an additional six stars at high resolution after we designed the deimos slitmasks .
these stars were not included in our slitmasks .
deimos observations of leo ii were conducted in the same program as observations of leo i. the photometry and spectroscopy were treated identically . for additional information ,
see sec .
[ sec : leoi ]
. figure [ fig : leoii ] shows the spectroscopic target selection .
we used @xcite s ( @xcite ) deimos spectra for individual stars in canes venatici i. photometry and extinction corrections were taken from the sdss data release 5 @xcite . in order to use the same isochrones for determining effective temperatures and surface gravities as we used for other galaxies , we converted sdss @xmath61 to johnson - cousins @xmath62 following the global , metallicity - independent transformations of @xcite .
we refer to @xcite s article for the details of the spectroscopic target selection , shown in figure [ fig : cvni ] .
no stars in canes venatici i have been observed yet at high spectral resolution .
we selected spectroscopic targets from the @xmath60 catalog of @xcite .
magnitudes were corrected star by star for extinction determined from the @xcite dust maps .
we selected rgb members by following the same procedure as for sextans , except that we allowed an additional 0.1 mag on the blue side of the rgb instead of 0.05 mag because the extra contamination was negligible .
figure [ fig : umi ] shows the ursa minor field with the spectroscopic targets highlighted in red .
every slitmask contained several of the 33 targets also included on other slitmasks .
figure [ fig : umi ] also shows the cmd of the targets within the right ascension and declination ranges of the axes in the left panel of fig . [
fig : umi ] .
the discussion of the sextans cmd similarly applies to the ursa minor cmd .
six stars , listed in table [ tab : hrscompare ] , have previously published hrs abundance measurements @xcite , and all six were observed in our study .
@xcite devoted an extremely wide - field photometric survey to draco .
they observed draco with three different telescopes , but we used only their @xmath83 catalog from the cfht megacam camera for its large field of view . in a manner similar to the target selection for sextans , we chose targets from the @xmath84 cmd between the bluest and reddest padova isochrones @xcite between 2 and 14 gyr , @xmath12}}= -2.23 $ ] and @xmath79 , and @xmath80}}= 0.0 $ ] and @xmath81 .
we allowed an additional 0.05 mag on the blue side of the cmd to account for stars more metal - poor than the most metal - poor padova isochrone ( @xmath12}}= -2.23 $ ] ) .
figure [ fig : dra ] shows the large field on which eight draco slitmasks were placed .
only stars with @xmath85 and @xmath86 are shown to minimize confusion .
the dra7 and dra8 slitmasks were located on the periphery of the field to search for draco members close to the tidal radius .
the dra2 slitmask included 11 targets also observed on the dra1 and dra4 slitmasks .
figure [ fig : dra ] also shows the cmd of targets that lay within the outlines of the slitmasks shown in the left panel of fig . [
fig : dra ] .
the discussion of the sextans cmd similarly applies to the draco cmd .
fourteen stars in draco have been observed with high - resolution spectroscopy @xcite .
we observed 12 of these .
they are listed in table [ tab : hrscompare ] .
we wished to use the same isochrones to determine effective temperature and surface gravities that we used for other galaxies .
@xcite showed that cfht @xmath83 magnitudes are indistinguishable from the usno standard @xmath87 magnitudes .
therefore , we transformed cfht @xmath83 magnitudes to sdss @xmath83 magnitudes following the prescription of @xcite .
then , we transformed sdss @xmath83 magnitudes to johnson - cousins @xmath88 magnitudes @xcite .
@xcite give color transformations from @xmath88 to cfht @xmath89 but not absolute zero points for that conversion .
therefore , we used the two - step transformation . because the cfht and sdss filter transmission curves are very similar , the first transformation from the cfht system to the sdss system introduced negligible error .
the galaxies ursa major ii , leo iv , coma berenices , and hercules are not included in the catalog presented here ( table [ tab : catalog ] ) .
however , some stars in these four galaxies have been observed both with hrs and mrs .
like canes venatici i , these galaxies were observed with deimos by @xcite .
@xcite observed at high resolution three stars each in ursa major ii and coma berenices ; @xcite observed at high resolution one star in leo iv ; and @xcite observed at high resolution two stars in hercules , but only one overlaps our sample .
we used these stars in the comparison between hrs and mrs abundance measurements ( sec .
[ sec : hrscompare ] ) .
we designed the deimos slitmasks with the ` dsimulator ` iraf software module .
each slitmask subtended approximately @xmath90 . in order to subtract night sky emission lines adequately
, we required a minimum slit length of @xmath91 .
the minimum space between slits was @xmath92 .
although this spacing is less than the typical seeing fwhm , light contamination between slitlets was negligible . because slits had the freedom to be placed along the dispersion axis of the slitmask , the end of a slit
was rarely located within 1 of a neighboring slit . even in those rare instances ,
the stars were located far enough from the ends of their slits that light spillage into a neighboring slit did not affect the spectral extraction or sky subtraction . when the slitmask design constraints forced the selection of one among multiple possible red giant candidates , we invoked priorities explained in secs .
[ sec : gcobs][sec : dra ] ( usually the brightest candidate ) . for most dsph slitmasks ,
the slitmasks were designed to be observed at approximately the parallactic angle at the anticipated time of observation .
this choice minimized the small light losses due to differential atmospheric refraction .
most slitmasks sky position angle ( pa ) was offset by @xmath93 from the slit pa .
the resulting tilt of the night sky emission lines relative to the ccd pixel grid increased the subpixel wavelength sampling and improved sky subtraction .
the dsph slitmasks contained many duplicate observations , which provide estimates of uncertainties in abundance measurements ( sec .
[ sec : duplicate ] ) .
the spectral coverage of each slit is not the same .
the minimum and maximum wavelengths of spectra of targets near the long , straight edge of the deimos footprint can be up to 400 lower than for targets near the irregularly shaped edge of the footprint ( upper left and lower right of the slitmask footprints in fig .
[ fig : scl ] , respectively ) .
furthermore , spectra of targets near either extreme of the long axis of the slitmask suffered from vignetting , which reduced the spectral range .
it is important to keep these differences of spectral range in mind when interpreting the differences of measurements derived from duplicate observations .
our observing strategy was nearly identical to that of @xcite and @xcite . in summary , we used the 1200 lines mm@xmath94 grating at a central wavelength of @xmath95 .
the slit widths were @xmath96 ( except for @xmath97 slit widths for m79 , ngc 2419 , and leo i and ii ) , yielding a spectral resolution of @xmath98 fwhm ( resolving power @xmath10 at 8500 ) .
the og550 filter blocked diffraction orders higher than @xmath99 .
the spectral range was about 64009000 with variation depending on the slit s location along the dispersion axis .
exposures of kr , ne , ar , and xe arc lamps provided wavelength calibration , and exposures of a quartz lamp provided flat fielding .
table 2 lists the slitmasks observed , the number of targets on each slitmask , the dates of observations , the airmass , the approximate seeing ( if known ) , and the exposure times . for the multi - slit masks ,
seeing was estimated by measuring the fwhm of a gaussian fit to the one - dimensional profiles of alignment stars . because halo stars are not clustered on the sky
, only one star could be observed in each exposure . therefore , these observations did not require custom slitmasks .
instead , the spectra were obtained through longslits .
the instrument configuration for the longslit exposures was identical to the configuration for custom multi - slit masks . to estimate the seeing
, we measured the fwhm from the guider camera of the star used to the focus the telescope .
all but two of the spectra of halo field stars presented here were obtained during nights of poor transparency in 2008 apr and may .
star vii-18 in the gc m92 was also observed in this manner .
the raw frames were reduced into one - dimensional spectra using version 1.1.4 of ` spec2d ` , the deimos data reduction pipeline developed by the deep galaxy redshift survey .
in addition to the procedure described by @xcite , kgs08 , and in @xcite , we tweaked the arc lamp wavelength solution by using the pipeline s optional ` skytweak_1d ` procedure .
the procedure evaluates small wavelength shifts between the observed and expected wavelengths of night sky emission lines in 100 regions .
a polynomial fit to the individual shifts adjusts the wavelength array of the observed spectrum . for any given pixel ,
a typical wavelength shift is 0.1 .
the fornax slitmasks for1b , for6 , and for7 ( see table 2 ) required special attention . because exposures for the same masks were taken months apart , the heliocentric velocity corrections changed enough that frames from the different months could not be stacked without affecting spectral line widths significantly . in order to overcome the velocity shift , we extracted one set of 1-d spectra for each night .
the wavelength arrays for the second epoch were shifted by the difference in the heliocentric velocity correction between the two nights .
the velocity - shifted spectra were rebinned to match the wavelength arrays of the spectra from the first epoch .
then , the 1-d spectra were coadded .
the rest of the analysis proceeded in the same way as for the slitmasks observed on single nights .
figure [ fig : snexamples ] shows small spectral regions for eight different dsph stars .
the stars were chosen to show a range of signal - to - noise ratio ( s / n ) and metallicity .
some absorption lines used in the abundance determination are labeled .
some prominent lines are not labeled because they are not included in the abundance measurements .
for example , @xmath100 and its blended neighbor , @xmath101 , are not labeled because the mg line is too strong for a local thermodynamic equilibrium abundance measurement .
we determined membership in each dsph by the radial velocities of the stars .
( table [ tab : catalog ] does not include radial velocities .
geha et al .
will publish the radial velocities in a separate work . ) in order to measure radial velocities , we cross - correlated each one - dimensional spectrum with several different stellar templates ( see * ? ? ?
we adopted the velocity of the cross - correlation peak of the best - fitting template .
then , we fit a gaussian to the velocity distribution of each dsph .
every star within @xmath102 of the peak ( 99.7% of the true members , assuming that they are normally distributed in velocity ) was considered a member .
stars outside of this range were discarded , and they are not included in the catalog in sec .
[ sec : catalog ] . finally , horizontal branch stars and stars obviously not belonging the red giant branch of the dsph were excluded .
these steps reduce contamination by foreground and background mw stars , but it is impossible to eliminate contamination entirely . the small remaining contamination may influence the interpretation of the abundance measurements , but it does not our affect our estimation of their accuracy and precision ( sec . [
sec : accuracy ] ) .
the number of unique member stars across all eight dsphs is 2961 .
kgs08 and @xcite discuss the details of the mrs technique used to create the catalog in this article . here
we present only a summary and some modifications to the method .
other than the changes described in sections [ sec : grid][sec : offsets ] , the details of the procedure remain identical to the method of kgs08 with the modifications in @xcite .
lccc + mg@xmath103 & 6319.237 & 5.110 & @xmath104 + mg@xmath103 & 6319.495 & 5.108 & @xmath105 + mg@xmath103 & 7691.553 & 5.753 & @xmath106 + mg@xmath103 & 7811.133 & 5.946 & @xmath107 + mg@xmath103 & 7811.141 & 5.946 & @xmath107 + mg@xmath103 & 8047.720 & 5.932 & @xmath108 + mg@xmath103 & 8098.719 & 5.946 & @xmath109 + mg@xmath103 & 8098.727 & 5.946 & @xmath110 + mg@xmath103 & 8213.034 & 5.753 & @xmath111 + mg@xmath103 & 8346.106 & 5.946 & @xmath112 + each observed spectrum was compared to a large grid of synthetic spectra at a variety of effective temperatures ( @xmath113 ) , surface gravities ( @xmath114 ) , metallicities ( represented by [ fe / h ] , but all elements heavier than he were modulated ) , and alpha enhancements ( [ @xmath13/fe ] , an additional modulation for mg , si , ca , and ti ) . section [ sec : grid ] gives additional details on the grid .
the spectra were synthesized with the local thermodynamic equilibrium , plane - parallel spectrum synthesis code moog @xcite , atlas9 model atmospheres @xcite , and atomic and molecular transition data from the vienna atomic line database ( vald , * ? ? ?
* ) , modified as described by kgs08 .
table [ tab : linelist ] gives a subset of the line list for the lines visible in a high s / n deimos spectrum of a red giant with @xmath12}}\sim -0.5 $ ] , such as arcturus .
a levenberg - marquardt algorithm ( the idl code ` mpfit ` by @xcite @xcite ) minimized @xmath115 to find the best - fitting synthetic spectrum in several steps . in the first step , @xmath113 , [ fe / h ] , and [ @xmath13/fe ] were determined iteratively .
surface gravity was fixed by photometry .
only regions sensitive to fe absorption were used to determine fe and only regions sensitive to mg , si , ca , or ti absorption were used to determine [ @xmath13/fe ] ( see @xcite ) . the observed spectrum s continuum was refined between each iteration by removing a high - order spline fit to the quotient of the observed spectrum and the best - fitting synthetic spectrum of the previous iteration .
the temperature was allowed to vary , but within the bounds of a photometrically determined temperature based on available photometry .
( the spectroscopic refinement of the temperature leads to more accurate abundances when compared to hrs measurements . because lines of ionized species are sparse in these far - red spectra of red giants , spectroscopic refinement of surface gravity is not possible . ) in the next steps , the individual mg , si , ca , and ti abundances were determined by fitting the spectral regions sensitive to absorption from each element .
special consideration was made for the mw halo field star sample . because the distances to the halo field stars are poorly known , surface gravities could not be fixed by photometry .
however , every star in the mw halo field star sample was chosen because previous authors had already observed it with hrs .
we adopted the high - resolution measurements of surface gravities for all of the mw halo field stars . in @xcite , we announced the discovery of one star in sculptor at @xmath12}}= -3.80 $ ] , which @xcite confirmed with the magellan / mike high - resolution spectrograph .
the lower [ fe / h ] limit of the spectral grid of kgs08 and @xcite was @xmath12}}= -4.0 $ ] .
if any star reached the lower [ fe / h ] limit , we discarded it under the presumption that the s / n was inadequate to yield an [ fe / h ] measurement of useful precision .
however , the confirmed existence of a star very close to the grid limit prompted us to expand the spectral grid to @xmath12}}= -5.0 $ ] .
although we have not recovered any ultra metal - poor stars ( @xmath116 } } < -4 $ ] ) , the expansion gives us confidence that we have not discarded any such stars because of a limitation in the grid .
expanding the grid required computing model atmospheres at @xmath117 } } < -4 $ ] with @xmath113 , @xmath114 , and [ @xmath13/fe ] on the grid described by table 3 of @xcite .
for each value of @xmath113 and @xmath114 in that table , we computed odfnew opacity distribution functions @xcite using r. l. kurucz s dfsynthe code ( described by * ? ? ? * ) at @xmath12}}= -4.5 $ ] and @xmath118 for each of 21 steps in [ @xmath13/fe ] ( 0.1 dex between @xmath80}}= -1.2 $ ] and @xmath119 ) and 2 steps in microturbulent velocity ( @xmath120 ) .
the values of @xmath120 were chosen to be the two velocities that bracket the microturbulent velocity appropriate for the star s surface gravity ( eq . 2 of @xcite ) .
the values of [ fe / h ] and [ @xmath13/fe ] in the computation of the opacity distribution functions are given relative to the solar abundances of @xcite , except that the abundance of iron is @xmath121 ( see * ? ? ?
* for an explanation ) .
we computed the model atmospheres with @xcite s ( @xcite ) atlas9 code ported into linux @xcite .
convective overshooting was turned off , and the mixing length parameter for convection was @xmath122 .
these convection parameters are the same as for @xcite s ( @xcite ) grid of atlas9 atmospheres .
we linearly interpolated the atmospheres in the [ fe / h ] and @xmath120 dimensions to populate the grid described in @xcite .
following @xcite , we modified the method of kgs08 to perform better continuum determination in @xcite .
we iteratively determined the continuum by fitting a high - order spline to the quotient of the observed spectrum and the best - fitting synthetic spectrum of the previous iteration . in @xcite , we incorporated this technique into the levenberg - marquardt algorithm that determines the best - fitting synthetic spectrum . during each levenberg - marquardt iteration , any remaining fluctuations in the continuum level were removed before the @xmath115 between the observed spectrum and a trial synthetic spectrum was evaluated .
instead , we decided that it was more appropriate to determine the best - fitting synthetic spectrum without modifying the observed spectrum in the minimization of @xmath115 .
the continuum refinement occurred between separate levenberg - marquardt determinations of the best - fitting synthetic spectrum .
the continuum refinement was considered converged when [ fe / h ] and the bulk [ @xmath13/fe ] changed by less than 0.001 dex and @xmath113 changed by less than 1 k. after the continuum refinement , we added a new step not used by kgs08 or in @xcite : masking pixels whose absolute difference from the best - fit synthetic spectrum exceeded 2.5 times their variance .
the remaining steps to measure [ fe / h ] were performed on the clipped spectrum .
the sigma clipping excludes spectral regions with artifacts , such as improperly removed cosmic rays .
it also insulates the abundance measurements from fe absorption lines with incorrect oscillator strengths .
the comparison between mrs and hrs abundances ( sec .
[ sec : hrscompare ] ) improved slightly as a result of including sigma clipping .
we experimented with the threshold value , and we found that @xmath123 included @xmath124 of the pixels while reducing the mean and standard deviations of the differences between mrs and hrs abundances . previously ( @xcite ) , we measured [ mg / fe ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] by searching the spectral grid at fixed , previously determined @xmath113 and [ fe / h ] . because each element was not a separate dimension , abundances of all of the @xmath13 elements ( o , ne , mg
, si , s , ar , ca , and ti ) varied together .
individual elements were measured by isolating the spectral regions containing those lines .
this resulted in a different value of [ @xmath13/fe ] for the model atmospheres used in the measurement of each element .
however , the parameter [ @xmath13/fe ] controls not only the strength of @xmath13 element lines , but also important atmospheric quantities , such as the free electron fraction .
the resulting changes in the continuous opacity in the red spectral region affect absorption lines of all elements
. for red giants , increasing the atmospheric value of [ @xmath13/fe ] at fixed [ fe / h ] raises the continuous opacity by increasing the number density of h@xmath125 ions .
metal lines become weaker . in the present catalog ,
we have determined the individual @xmath13 element ratios by fixing the atmospheric value of [ @xmath13/fe ] , which we call @xmath126}_{\rm atm}$ ] .
we synthesized a subgrid of spectra with an additional dimension , @xmath126}_{\rm abund}$ ] .
the new dimension is the [ @xmath13/fe ] ratio for the abundances of mg , si , ca , and ti at fixed @xmath126}_{\rm atm}$ ] . like @xmath127}_{\rm atm}$ ] , @xmath126}_{\rm abund}$ ] is spaced at 0.1 dex from @xmath128 to @xmath129 .
it was not necessary to synthesize spectra with @xmath126}_{\rm atm } = { \rm [ \alpha / fe]}_{\rm abund}$ ] because the existing grid already contained such spectra . to save computation time
, the new subgrid included only the spectral regions used in the measurement of mg , si , ca , and ti . we determined @xmath126}_{\rm atm}$ ] by fitting to the mg , si , ca , and ti lines simultaneously . then , the values of [ mg / fe ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] were determined by variation of @xmath127}_{\rm abund}$ ] with @xmath113 , @xmath114 , [ fe / h ] , and @xmath127}_{\rm atm}$ ] held fixed .
the complete list of steps to determine the abundances is largely the same as in sec .
4.7 of @xcite .
the following is the updated list . 1 .
@xmath113 and [ fe / h ] , first pass : an observed spectrum was compared to a synthetic spectrum with @xmath113 and @xmath114 determined photometrically . only spectral regions most susceptible to fe absorption were considered .
@xmath113 was loosely constrained by photometry .
photometry alone determined @xmath114 .
@xmath126}_{\rm atm}$ ] , first pass : @xmath113 , @xmath114 , and [ fe / h ] were fixed , but @xmath126}_{\rm atm}$ ] and @xmath130}_{\rm abund}$ ] varied together . only the spectral regions susceptible to absorption by mg , si , ca , or ti were considered .
continuum refinement : the continuum - divided , observed spectrum was divided by the synthetic spectrum with the parameters determined in steps 1 and 2 .
the result approximated a flat noise spectrum . to better determine the continuum
, we fit a b - spline with a breakpoint spacing of 150 pixels to the residual spectrum .
we divided the observed spectrum by the spline fit .
steps 13 were repeated until @xmath113 changed from the previous iteration by less than 1 k and [ fe / h ] changed by less than 0.001 dex .
[ fe / h ] , second pass : we repeated step 1 with the revised , sigma - clipped spectrum , but @xmath113 was held fixed at the previously determined value .
@xmath126}_{\rm atm}$ ] , second pass : we repeated step 2 with the revised spectrum and the value of [ fe / h ] from step 5 . 7 .
[ fe / h ] , third pass : we repeated step 5 with the value of @xmath127}_{\rm atm}$ ] from step 6 .
[ mg / fe ] : we determined [ mg / fe ] by varying @xmath130}_{\rm abund}$ ] at fixed @xmath126}_{\rm atm}$ ] .
only spectral regions subject to mg absorption were considered .
[ si / fe ] : we repeated step 9 for si instead of mg . 10
. [ ca / fe ] : we repeated step 9 for ca instead of mg . 11
. [ ti / fe ] : we repeated step 9 for ti instead of mg . in @xcite , we found that adding 0.15 dex to the mrs measurement of [ fe / h ] was necessary to bring @xmath12}}_{\mathrm{mrs}}$ ] into agreement with @xmath12}}_{\mathrm{hrs}}$ ] for nine stars with both mrs and hrs measurements in sculptor .
the modification to the continuum division and sigma clipping mitigated the need for the artificial offset in [ fe / h ] .
there is still an offset of @xmath131 dex , but it is small enough that we do not correct it .
an offset of 0.15 dex between @xmath132}}_{\mathrm{mrs}}$ ] and @xmath132}}_{\mathrm{hrs}}$ ] still exists , and we added 0.15 dex to @xmath132}}_{\mathrm{mrs}}$ ] for better agreement .
this offset is unique to si , and does not affect other [ @xmath13/fe ] ratios . therefore , the offset does not arise from continuum division uncertainties .
we share eight lines in common with the high - resolution studies of cohen et al .
( e.g. , * ? ? ?
* ; * ? ? ?
our oscillator strengths for these eight lines are all larger , with an average difference of 0.16 dex .
therefore , our si measurements are lower than cohen et al.s si measurements . rather than adjust the oscillator strengths and recompute the spectral grid at large computational expense , we added 0.15 dex to our [ si / fe ] measurements .
the maximum @xmath114 in the dsph catalog is 3.5 .
some of the gc masks observed in 2009 ( see table 2 ) included stars below the main sequence turn - off .
we excluded gc and mw halo stars with @xmath133 in order to restrict the comparison sample to stars of the same spectral type as the dsph catalog .
lr & 0.113 + & 0.095 + & 0.104 + & 0.118 + & 0.083 + table [ tab : catalog ] gives the measurements of 2961 unique stars ( 376 in sculptor , 675 in fornax , 827 in leo i , 141 in sextans , 258 in leo ii , 174 in canes venatici i , 212 in ursa minor , and 298 in draco ) .
the table lists the right ascension , declination , extinction - corrected @xmath88 magnitudes where available ( sometimes converted from @xmath61 following * ? ? ?
* ) , temperature , surface gravity , microturbulent velocity , [ fe / h ] , and the [ mg / fe ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] ratios where measurable .
uncertainties are also given for all of the measurable quantities except the space coordinates and microturbulent velocity .
for @xmath113 , two types of uncertainty are given .
the first type , @xmath134 , is the uncertainty from random spectral noise .
( in detail , it is the square root of the diagonal element of the covariance matrix corresponding to @xmath135 when both @xmath113 and [ fe / h ] are allowed to vary . ) the second type , @xmath136 , is the uncertainty on the photometric estimate of @xmath113 , accounting for random photometric uncertainty and systematic error from isochrone modeling ( eq . 6 of @xcite ) . because @xmath114 is determined from photometry alone , the uncertainty on @xmath114 is derived analogously to @xmath137 , but the error on the distance modulus to each stellar system is included in the error estimate of @xmath114 as well .
the uncertainties on the abundance measurements are the @xmath138 uncertainties from the spectral fit added in quadrature with the constant error floors , given in table [ tab : syserr ] . as an example , consider the first star in table [ tab : catalog ] .
the first column identifies its origin as sculptor , and the second column gives the star s identification number . in the online version ,
the next two columns give its coordinates . @xmath56 and
@xmath58 magnitudes are given , but @xmath55 and @xmath57 magnitudes are not available .
we have measured @xmath139 k with an uncertainty of 101 k based on spectral noise . from photometry alone ,
the uncertainty is 258 k. we have also measured a photometric surface gravity of @xmath140 .
the microturbulent velocity , deduced solely from the surface gravity , is 1.66 km s@xmath94 .
the remaining columns give [ fe / h ] , [ mg / fe ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] and their estimated uncertainties .
the range of abundances and size of our sample demand rigorous checks to validate the accuracy and precision of our measurements . in this section , we estimate measurement uncertainties , check those estimates against repeat measurements of the same stars , and check the accuracy of our measurements against previously published high - resolution spectroscopic abundance measurements .
we estimated abundance measurement uncertainties from individual stars in globular clusters .
because we have excluded gcs suspected to exhibit fe abundance variations , such as m22 @xcite , we assumed that the spread of [ fe / h ] in each cluster is much smaller than the measurement uncertainty , but we made no such assumption for other elements . although no monometallic cluster has been found to have a measurable spread of ca or ti abundances except ngc 2419 @xcite , many clusters exhibit star - to - star mg variations ( see the review by * ? ? ?
kgs08 and @xcite explicitly described the procedure for determining measurement uncertainties . for each elemental abundance measurement
, they added an extra error component in quadrature with the random error from spectral noise .
this error floor for [ fe / h ] was determined from the excess spread in the measured [ fe / h ] in each gc after accounting for random error due to spectral noise . for this purpose , we excluded m92 because we observed only one star in it , and we excluded ngc 2419 because we suspect that it may contain a spread in heavy element abundances @xcite .
the error floor in the four [ @xmath13/fe ] ratios was determined from comparison to high - resolution measurements .
ngc 2419 was included for this purpose .
the error floor is the value needed to account for differences between the mrs and hrs measurements after removing the mrs random error and the published value of the hrs measurement uncertainty .
table [ tab : syserr ] gives the updated values for the error floors for each element .
the differences between this table and table 5 of @xcite arise from the modifications described in sec .
[ sec : measurements ] and the inclusion of gc observations obtained since the publication of kgs08 . from changes in microturbulent velocity ( @xmath120 ) of @xmath141 ( _ dashed lines _ ) and @xmath142 ( solid lines ) vs. surface gravity .
the 20 stars analyzed are selected from four different globular clusters over a range of metallicities and from a range of surface gravities .
the figure legend gives the mean [ fe / h ] for each gc from our own medium - resolution measurements.[fig : vttest],scaledwidth=40.0% ] errors in @xmath113 , @xmath114 , and @xmath120 have the potential to change the measured abundances significantly .
the error floor for @xmath12}}_{\rm mrs}$ ] accounts for uncertainty in the atmospheric parameters , but we also quantify the response of [ fe / h ] to forced changes in @xmath113 and @xmath114 .
differences in @xmath113 , @xmath114 , and @xmath120 affect the @xmath13 element abundance ratios less than they affect [ fe / h ] . to first order , the measurement of an @xmath13 element s abundance responds to the atmospheric parameters in the same way as the measurement of the fe abundance . as a result
, the division of one element s abundance by the fe abundance somewhat masks errors in atmospheric parameters .
we recomputed abundances at 125 k and 250 k above and below the best - fitting @xmath113 for every star in the gc , mw halo , and dsph samples .
the new abundance was determined by searching the grid for the synthetic spectrum that best matched the observed spectrum with @xmath113 and @xmath114 held fixed .
we also recomputed abundances with @xmath114 held fixed at 0.3 dex and 0.6 dex above and below the best - fitting @xmath114 .
the measured abundance of [ fe / h ] is higher for a larger value of @xmath113 or a larger value of @xmath114 .
conversely , the measured abundance of [ fe / h ] is lower for a lower value of @xmath113 or a lower value of @xmath114 .
( see sec .
6.3 of kgs08 for a more thorough discussion . )
the globular cluster star sample provides some insight on the general response of [ fe / h ] to changes in atmospheric parameters .
the average change in [ fe / h ] is @xmath143 per @xmath144 change in @xmath113 .
the average change in [ fe / h ] is @xmath145 per @xmath146 change in @xmath114 .
the [ fe / h ] measurements are relatively insensitive to changes in @xmath114 because the abundance analysis uses many lines but virtually no lines , which are few and weak in red giants .
lines vary little with @xmath114 .
ideally , we would use only lines because they are not subject to overionization problems @xcite .
however , we must use lines because lines are not visible in deimos spectra of red giants .
the slopes of @xmath147}}/ \delta { t_{\rm eff}}$ ] and @xmath148}}/ \delta ( { \log g})$ ] are approximately linear within @xmath149 and @xmath150 .
table [ tab : atmparerr ] gives the response of [ fe / h ] [ mg / fe ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] to a change of 125 k and 250 k in @xmath113 and 0.3 dex and 0.6 dex in @xmath114 .
the values given are the averages of the absolute values of the increase and decrease in @xmath113 or @xmath114 .
for example , consider a star with @xmath12}}= -2.00 $ ] and @xmath151 .
if the column for @xmath147}}(125~{\rm k})$ ] lists 0.12 , then it can be assumed that the measured abundance of the star would be @xmath12}}= -1.88 $ ] if @xmath113 were truly 4625 k , and the measured abundance would be @xmath12}}= -2.12 $ ] if @xmath113 were truly 4375 k. we could not estimate the effect of errors in microturbulent velocity as easily as for @xmath113 and @xmath114 because our grid of spectra did not have a separate dimension for @xmath120 . instead , we fixed @xmath120 based on a calibration to @xmath114 ( eq . 2 of @xcite ) . to investigate the effect of errors in @xmath120 on [ fe / h ] , we reanalyzed 20 stars from four gcs at different metallicities ( m71 , m5 , m13 , and m15 ) .
we chose the stars with @xmath114 nearest to 0.5 , 1.0 , 1.5 , 2.0 , 2.5 , and 3.0 , where available .
we synthesized a new , small grid of spectra for each star .
we generated new atmospheres by linear interpolation in our atlas9 grid ( sec .
[ sec : grid ] ) .
the values of @xmath113 and @xmath114 were chosen to be the best - fit values we already determined for each star .
values of [ fe / h ] were spaced at every 0.1 dex from at least 0.4 dex below to at least 0.4 dex above the best - fit value .
values of @xmath120 were @xmath152 and @xmath153 around the value of @xmath120 from the surface gravity calibration .
we computed spectra with moog . for each of the four displacements of @xmath120
, we found the value of [ fe / h ] that minimized @xmath115 between the new grid of spectra and the observed spectrum .
as before , we used a levenberg - marquardt minimization algorithm with linear interpolation to compute @xmath115 between [ fe / h ] grid points .
figure [ fig : vttest ] shows the changes in the measurement of [ fe / h ] induced by changes in @xmath120 for each of the 20 gc stars . reducing @xmath120 increases [ fe / h ] because the absorption lines saturate at lower abundance .
a larger abundance is necessary to compensate for the weaker lines .
the abundance change is most severe for stars of low surface gravity ( low temperature ) and high metallicity because those stars have strong lines and are most susceptible to line saturation .
the absolute change in [ fe / h ] does not exceed 0.1 dex , even for changes of @xmath154 .
this result agrees with our later assessment of the effect of microturbulent velocity error ( fig .
[ fig : covarvtfeh ] ) , where we deduce a slope of @xmath155}}/\delta\xi = { -0.20}~\rm{dex}/(\rm{km~s}^{-1})$ ] . for comparison , the standard deviation between @xmath156 and @xmath157 is @xmath158 ( see sec . [
sec : hrscompare ] and table [ tab : hrsdiff ] ) .
the repeat observations of 167 dsphs stars provided insight into the effect of random error on the measurements of [ fe / h ] and @xmath0\rangle$ ] ( an average of [ mg / fe ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] ) .
repeat measurements of @xmath0\rangle$ ] were possible for both stars in 141 pairs .
figures [ fig : fehdup ] and [ fig : alphadup ] summarize the comparisons of measurements of different spectra of the same stars .
they show the distributions of the absolute difference between the measured [ fe / h ] and @xmath0\rangle$ ] for each pair of spectra divided by the expected error of the difference . in calculating the expected error of the difference
, we apply the error floor to only one of the two stars .
even though the same technique is used to measure abundances in both stars including the same photometric estimate of surface gravity-the extra error , which accounts for sources of error beyond random spectral noise , is appropriate because the wavelength range within a pair of spectra differs by up to 400 .
the different fe lines in these ranges span a different range of excitation potentials , and the levenberg - marquardt algorithm converges on different solutions . figures [ fig : fehdup ] also shows the best - fit gaussian ( @xmath159 ) and a gaussian with unit variance .
the best - fit gaussian in fig .
[ fig : alphadup ] has @xmath160 .
the areas of the gaussians are normalized to the number of stars .
our estimate of uncertainty is accurate because the variances of the best - fit gaussians are close to unity . lr mg & 7.58 + si & 7.55 + ca & 6.36 + ti & 4.99 + fe & 7.52 + lccccc + @xcite & atlas9 & moog & phot & phot & spec + @xcite & @xcite & width2 & phot & phot & spec + @xcite & marcs & moog & spec & spec & spec + @xcite & atlas9 & width9 & spec & spec & spec + @xcite & atlas9 & moog & phot & phot & spec + @xcite & marcs & moog & spec & spec & spec + @xcite & marcs & moog & spec & phot & spec + @xcite & atlas9 & unknown & phot & phot & spec + @xcite & atlas9 & moog & phot & phot & phot + @xcite & atlas9 & moog & spec & spec & spec + @xcite & atlas9 & moog & spec & spec & spec + @xcite & atlas9 & moog & phot & phot & evolutionary + @xcite & marcs & moog & phot & phot & spec + @xcite & atlas9 & moog & spec & spec & spec + @xcite & atlas9 & moog & spec & spec & spec + @xcite & atlas9 & moog & phot & phot & spec + @xcite & marcs & moog & phot & spec & spec + @xcite & atlas9 & moog & spec & spec & spec + @xcite & marcs & calrai & phot & phot & spec + @xcite & atlas9 & sptool & spec & spec & spec + @xcite & marcs & moog & spec & spec & spec + @xcite & atlas9 & moog & phot & phot & spec + ccccc @xmath161 ( k ) & @xmath162 ( @xmath163 ) & @xmath164 ( @xmath165 ) & @xmath166 ( @xmath167 ) & @xmath168 ( @xmath169 ) + @xmath170 ( cm s@xmath171 ) & @xmath172 ( @xmath173 ) & & @xmath174 ( @xmath175 ) & @xmath174 ( @xmath176 ) + @xmath177 ( km s@xmath94 ) & @xmath178 ( @xmath179 ) & @xmath180 ( @xmath181 ) & @xmath180 ( @xmath179 ) & @xmath182 ( @xmath179 ) + @xmath183}}\rangle$ ] & @xmath184 ( @xmath185 ) & @xmath186 ( @xmath187 ) & @xmath188 ( @xmath1 ) & @xmath131 ( @xmath1 ) + @xmath189}\rangle$ ] & @xmath131 ( @xmath190 ) & @xmath174 ( @xmath190 ) & @xmath191 ( @xmath173 ) & @xmath192 ( @xmath193 ) + @xmath194}\rangle$ ] & @xmath195 ( @xmath187 ) & @xmath196 ( @xmath173 ) & @xmath197 ( @xmath198 ) & @xmath188 ( @xmath199 ) + @xmath200}\rangle$ ] & @xmath201 ( @xmath2 ) & @xmath188 ( @xmath187 ) & @xmath202 ( @xmath203 ) & @xmath204 ( @xmath193 ) + @xmath205}\rangle$ ] & @xmath188 ( @xmath206 ) & @xmath201 ( @xmath207 ) & @xmath192 ( @xmath190 ) & @xmath188 ( @xmath208 ) + @xmath209}}\rangle\rangle$ ] & @xmath210 ( @xmath206 ) & @xmath201 ( @xmath187 ) & @xmath192 ( @xmath211 ) & @xmath212 ( @xmath2 ) + = .45 except for an average of the @xmath13 elements .
the average includes only those [ @xmath13/fe ] measurements with estimated uncertainties less than 0.5 dex.[fig : alphaavg],scaledwidth=40.0% ] the most reliable test of the mrs atmospheric parameter and abundance estimates is to compare with completely independent high - resolution observations and analyses of the same stars .
table [ tab : hrscompare ] lists the names , references , and coordinates of stars with previously published hrs measurements that we have also observed with deimos .
the gc and dsph stars were some of the targets on multi - slit masks , and the mw halo stars were observed through a longslit mask .
the one star in m92 was also observed through a longslit .
different authors prefer different measurements of the solar compositions .
we have placed all abundance measurements in this paper on the same scale by adjusting the published hrs values by the differences between the solar composition adopted by each hrs study and the solar composition given in table [ tab : solar ] .
the note at the bottom of the table explains our choice of solar composition .
all hrs studies share some techniques in common , but some aspects of the analyses differ .
table [ tab : hrsmethod ] summarizes the components of the hrs methods that change from study to study .
the hrs studies we cite here all measure equivalent widths ( ews ) of individual metal lines , and they compute abundances from those ews , model atmospheres , and line lists .
the line lists vary from study to study , causing typically small changes in derived abundances ( @xmath213 dex ) .
also , different studies use different codes to calculate model atmospheres , e.g. , atlas9 @xcite and marcs @xcite . the different codes used to calculate the abundances
are given in footnote b of table [ tab : hrsmethod ] .
most importantly , different hrs studies determine @xmath113 and @xmath114 in different ways .
sometimes effective temperature is determined through excitation equilibrium , wherein @xmath113 is adjusted to minimize the trend of abundance with the excitation potential of the fe transition .
surface gravity is sometimes determined through ionization balance , wherein @xmath114 is adjusted until the abundance measured from lines of agrees with the abundance measured from lines of .
@xcite explains these methods in detail .
some authors of hrs studies use photometric determinations of @xmath113 and @xmath114 only , relying on empirical calibrations or theoretical isochrones . for a discussion on the merits and disadvantages of photometrically determined atmospheric parameters , we refer the reader to @xcite . it is tempting to assume that our hrs comparison set is a flawless standard , but the heterogeneity of the sources of these abundance estimates introduces systematic error . measurement uncertainties from hrs are generally smaller than from mrs , but systematic offsets between studies arise from different choices of methods of determining atmospheric parameters , line lists , model atmospheres , and spectral synthesis codes .
table [ tab : hrsmethod ] is meant to illustrate the diversity of ways to measure elemental abundances spectroscopically .
these differences should be borne in mind when examining the hrs measurements presented below .
the sample sizes are often too small to make a meaningful statistical quantification of bias between hrs studies of the same stars or stellar systems .
careful attention should be paid to potential systematic offsets between hrs studies . although our analysis is also subject to its own random uncertainties and systematic errors , one of the principal advantages of our sample is that the abundances of all of the stars have been measured in a homogeneous fashion , eliminating most relative systematic offsets from star to star .
table [ tab : hrsdiff ] lists the differences and standard deviations between mrs and hrs measurements for all three types of stellar systems .
the quantities are very similar to comparisons between different hrs measurements of the same stars ( e.g. , * ? ? ?
* their appendix b ) .
figures [ fig : teff ] to [ fig : alphaavg ] show the comparison between hrs measurements ( @xmath214-axes ) and mrs measurements ( @xmath215-axes ) .
the points are coded by their membership in gcs , the mw halo field , or dsphs . coding the points by individual system such as the identity of the gc or dsph and also by hrs reference is also instructive , but the plots contain too many points for this coding to be feasible .
tables [ tab : hrscompare ] and [ tab : hrsmethod ] contain more complete descriptions of the comparisons between mrs and hrs measurements . the bottom panels of figs . [ fig : teff][fig : alphaavg ] show the differences between mrs and hrs measurements ( @xmath216 ) . instead of plotting @xmath216 versus @xmath214
, we have plotted @xmath216 versus @xmath217 , the average of @xmath214 and @xmath215 .
this is effectively a scaled orthogonal distance from the one - to - one line .
( a pure rotation of the upper panel , @xmath215 versus @xmath214 , would be @xmath218 versus @xmath219 , but @xmath216 is easier to interpret than @xmath218 . ) we have chosen the orthogonal distance as the abscissa because a plot of @xmath216 versus @xmath214 would show trends even if @xmath214 and @xmath215 are drawn from the same distribution with random scatter .
in fact , any random uncertainty in @xmath214 would cause a downward sloping trend in @xmath216 versus @xmath214 . in the orthogonal residual plots , random scatter along the one - to - one line
does not produce a trend as long as the magnitude of the scatter in @xmath214 is close to magnitude of the scatter in @xmath215 .
the uncertainties in the mrs measurements are generally slightly larger than the uncertainties in the hrs measurements , but not enough to produce these trends in the orthogonal residuals .
figure [ fig : teff ] compares @xmath220 and @xmath221 .
the average difference in @xmath113 between mrs and hrs depends on the source of the hrs measurement .
as one example , consider the mw halo field star sample .
@xcite calculate @xmath113 spectroscopically , whereas @xcite calculate @xmath113 photometrically .
our measurements of @xmath113 are typically @xmath222 k below those of @xcite and @xmath222 k above those of @xcite .
the differences possibly result from the different methods of measuring @xmath113 .
as another example of a trend seen with the source of the hrs measurements , consider the study of @xcite .
five of the six stars in the ursa major ii and coma berenices data set have @xmath221 at least 175 k higher than @xmath220 .
furthermore , @xmath156 is between 0.3 and @xmath223 less than @xmath157 for all of the stars in @xcite s sample .
our measurements of @xmath12}}_{\rm mrs}$ ] for the five stars with the significantly different temperature measurements exceed @xmath12}}_{\rm hrs}$ ] by 0.10.4 dex .
in contrast to our measurements , @xcite used a blue spectral range , higher resolution , and spectroscopically derived surface gravities and microturbulent velocities .
our choices are not better , simply different .
the differences partly explain our different offsets from different studies .
the standard deviation of @xmath12}}_{\rm mrs } - { { \rm [ fe / h]}}_{\rm hrs}$ ] is @xmath1 . the minimum uncertainty that we quote on @xmath12}}_{\rm mrs}$ ] ( in the limit of infinite spectral s / n ) is @xmath224 .
the small difference between our minimum estimate of uncertainty and the typical difference between mrs and hrs measurements which includes the error on @xmath12}}_{\rm hrs}$ ] in addition to the error on @xmath12}}_{\rm mrs}$]indicates that we have not underestimated our measurement uncertainties , even on an absolute scale .
furthermore , fig .
[ fig : feh ] shows no systematic trend in @xmath12}}_{\rm mrs } - { { \rm [ fe / h]}}_{\rm hrs}$ ] as a function of [ fe / h ] .
therefore , our mrs measurements of [ fe / h ] are consistent with hrs measurements at least over the range @xmath225 } } < -0.5 $ ] . the mrs @xmath13 element abundances also agree with hrs measurements .
( however , we added a constant to @xmath226}_{\mathrm{mrs}}$ ] to force better agreement with @xmath226}_{\mathrm{hrs}}$ ] .
see sec .
[ sec : continuum ] . ) the standard deviation of the differences between mrs and hrs @xmath0\rangle$ ] ( an average of the four measured [ @xmath13/fe ] ratios ) is @xmath2 , about the same as the standard deviation for [ fe / h ] .
the number of detectable mg , si , ca , and ti absorption lines in the deimos spectral range is comparable to the number of fe absorption lines for red giants .
therefore , it is reassuring that the precision of the @xmath227}}\rangle_{\rm mrs}$ ] measurements is comparable to the precision of the @xmath12}}_{\rm mrs}$ ] measurements .
we point out in particular an interesting feature of the individual alpha element ratios concerning gc stars ( red points in fig .
[ fig : alpha ] ) . internal variations in ca and ti have not been detected in the gcs presented here .
furthermore , the [ ca / fe ] and [ ti / fe ] ratios vary little from cluster to cluster .
therefore , it is expected that we see small or zero correlation between mrs and hrs measurements for these elements .
in fact , the linear pearson correlation coefficients between mrs and hrs measurements are @xmath176 and @xmath199 for [ ca / fe ] and [ ti / fe ] , respectively . on the other hand , [ mg / fe ] shows a significant spread within many gcs ( e.g. , * ? ? ?
as a result , we see a correlation between @xmath228}_{\rm mrs}$ ] and @xmath228}_{\rm hrs}$ ] .
the correlation coefficient is @xmath175 . in other words ,
the @xmath13 element ratio measurements from deimos can sort out which of the four [ x / fe ] ratios has a dispersion between gcs .
the large sample of stars observed with at least two independent measurements provides a unique opportunity to examine the influence of errors in atmospheric parameters on [ fe / h ] .
figures [ fig : covartefffeh][fig : covarvtfeh ] show how differences in @xmath113 , @xmath114 , and @xmath120 affect the measurement of [ fe / h ] .
each figure shows the least - squares linear fit .
the vertical error bars are the quadrature sum of the total errors on [ fe / h ] from mrs and hrs , including the error introduced by uncertainty in atmospheric parameters .
horizontal error bars are not shown because most hrs studies do not include errors on atmospheric parameters .
not surprisingly , @xmath113 is the atmospheric parameter that most affects the measurement of [ fe / h ] . in general
, underestimating @xmath113 leads to an underestimate of [ fe / h ] .
reassuringly , the intercept of the least - squares linear fit to @xmath229}}_{\rm mrs } - { { \rm [ fe / h]}}_{\rm hrs})$ ] versus @xmath230 is close to zero ( @xmath231 ) . in other words ,
if @xmath113 were determined perfectly , @xmath12}}_{\rm mrs}$ ] and @xmath12}}_{\rm hrs}$ ] would agree extremely well .
surface gravity does not have a strong influence on the measurement of [ fe / h ] , and microturbulent velocity has a moderately strong influence .
underestimating @xmath120 leads to an overestimate of [ fe / h ] . in order to quantify the total effect of errors on atmospheric parameters ,
we have identified the linear combination of @xmath113 , @xmath114 , and @xmath120 differences that minimizes the scatter about the least - squares linear fit to @xmath229}}_{\rm mrs } - { { \rm [ fe / h]}}_{\rm hrs})$ ] .
figure [ fig : leastsquares ] shows the result . to remove dimensionality , @xmath135 , @xmath232 , and @xmath233 have been normalized by their standard deviations .
as expected , @xmath113 has the most influence on the [ fe / h ] measurement by far .
surface gravity and @xmath120 have about one half of the influence of @xmath113 .
the rms scatter about the line is 0.13 dex , compared to 0.15 dex , which is the rms scatter in @xmath229}}_{\rm mrs } - { { \rm [ fe / h]}}_{\rm hrs})$ ] without removing the effect of errors from atmospheric parameters .
therefore , uncertainty in @xmath113 and @xmath120 do inflate the error on [ fe / h ] .
this result possibly indicates that s / n ( for these bright comparison stars ) and spectral resolution are not limiting the precision of @xmath12}}_{\rm mrs}$ ] . instead , improving the determinations of @xmath113 and @xmath120 would do much to improve the measurement of @xmath12}}_{\rm mrs}$ ] .
the problem of measuring @xmath113 and @xmath120 is not unique to mrs .
table [ tab : hrsmethod ] shows that many hrs studies , particularly for the more distant stars , employ photometry at least in part for determining @xmath113 , @xmath114 , and @xmath120 , as we do .
we have presented a catalog of fe , mg , si , ca , and ti abundance measurements for 2961 stars in eight dwarf satellite galaxies of the milky way .
medium - resolution spectroscopy from the keck / deimos spectrograph provided the throughput and multiplexing necessary to perform these measurements for the large number of faint , distant stars .
the majority of these stars are inaccessible to high - resolution spectrographs even on the largest telescopes .
the measurements relied on a spectral synthesis technique discussed previously ( kgs08 and @xcite ) with some modifications detailed in this paper .
we have estimated the uncertainty on every measurement of @xmath113 , @xmath114 , [ fe / h ] , [ mg / fe ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] .
the uncertainties on the abundances include the effect of spectral noise , spectral modeling uncertainties , and uncertainties in atmospheric parameters ( effective temperature and surface gravity ) .
the estimated uncertainties were shown to be accurate based on duplicate observations of stars in the scientific targets , dwarf galaxies . finally , we have quantified the accuracy of our medium - resolution measurements by observing with deimos a sample of stars with high - resolution spectroscopic measurements .
we deliberately targeted stars in dwarf galaxies with previous hrs measurements , and we observed stars in globular clusters and in the field of the milky way stellar halo . the mean difference in [ fe / h ] and @xmath0\rangle$ ] between the hrs and mrs measurements of these 132 stars is @xmath131 and @xmath212 , with standard deviations of @xmath1 and @xmath2 .
the next paper in this series focuses on the metallicity evolution of the individual dsphs .
we fit chemical evolution models to the metallicity distributions .
the shapes of the distributions and parameters of the fits show trends with dsph luminosity .
the following paper in the series addresses the star formation timescale and chemical evolution as revealed by the distribution of the @xmath13 elements .
the authors thank the referee , piercarlo bonifacio , for his thoughtful suggestions , which improved this manuscript .
we gratefully acknowledge sandra faber , ricardo schiavon , and michael cooper of the deep2 team for acquiring keck / deimos spectroscopy of bright milky way halo field stars during nights of poor transparency .
we also thank peter stetson for providing additional globular cluster photometry and bob kraft for helpful discussions and for providing some of the globular cluster spectroscopy for this work .
the generation of synthetic spectra made use of the university of california santa cruz pleiades supercomputer and the yale high performance computing cluster bulldog .
we thank joel primack for sharing his allocation of pleiades supercomputer time and mario juri for a helpful discussion on the mass of the components of the milky way .
support for this work was provided by nasa through hubble fellowship grant hst - hf-51256.01 awarded to enk 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 .
pg acknowledges nsf grants ast-0307966 , ast-0607852 , and ast-0507483 .
mg acknowledges support from nsf grant ast-0908752 .
cs acknowledges nsf grant ast-0909978 .
jgc acknowledges nsf grant ast-090109 .
mhs was supported at psu by nasa contract nas5 - 00136 .
this research used the facilities of the canadian astronomy data centre operated by the national research council of canada with the support of the canadian space agency .
the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community .
we are most fortunate to have the opportunity to conduct observations from this mountain .
gustafsson , b. , edvardsson , b. , eriksson , k. , jrgensen , u. g. , mizuno - wiedner , m. , & plez , b. 2003 , in asp conf .
ser . 288 , modelling of stellar atmospheres , ed .
i. hubeny , d. mihalas , & k. werner ( san francisco : asp ) , 331 llcr@ c@ lccl ngc 288 & n288 & 119 & 2008 & nov & 24 & 1.92 & @xmath234 & 300 s , 2 @xmath235 420 s + & & & 2008 & nov & 25 & 1.86 & @xmath236 & 4 @xmath235 300 s + m79 & n1904 & 22 & 2006 & feb & 2 & 1.42 & unknown & 2 @xmath235 300 s + & ng1904 & 104 & 2009 & feb & 22 & 1.40 & @xmath237 & 2 @xmath235 600 s , 2 @xmath235 1200 s + ngc 2419 & n2419 & 70 & 2006 & feb & 2 & 1.21 & unknown & 4 @xmath235 300 s + & n2419c & 94 & 2009 & oct & 13 & 1.15 & @xmath238 & 1200 s , 900 s + & & & 2009 & oct & 14 & 1.15 & @xmath239 & 200 s , 900 s + m5 & ng5904 & 181 & 2009 & feb & 22 & 1.05 & @xmath240 & 1200 s , 900 s , 600 s , 480 s + m13 & n6205 & 93 & 2007 & oct & 12 & 1.35 & unknown & 3 @xmath235 300 s + m92 & lvmslits & 1 & 2008 & may & 6 & 1.09 & unknown & 500 s , 2 @xmath235 300 s + m71 & n6838 & 104 & 2007 & nov & 13 & 1.09 & @xmath241 & 3 @xmath235 300 s + ngc 7006 & n7006 & 105 & 2007 & nov & 15 & 1.01 & @xmath242 & 2 @xmath235 300 s + m15 & n7078 & 63 & 2007 & nov & 14 & 1.01 & @xmath243 & 2 @xmath235 300 s + & n7078d & 164 & 2009 & oct & 13 & 1.01 & @xmath244 & 3 @xmath235 900 s + & n7078e & 167 & 2009 & oct & 14 & 1.01 & @xmath245 & 3 @xmath235 900 s + m2 & n7089b & 91 & 2009 & oct & 13 & 1.09 & @xmath242 & 3 @xmath235 900 s + pal 13 & pal13 & 33 & 2009 & oct & 13 & 1.48 & @xmath246 & 2 @xmath235 900 s + & & & 2009 & oct & 14 & 1.50 & @xmath247 & 900 s , 822 s + ngc 7492 & n7492 & 38 & 2007 & nov & 15 & 1.30 & @xmath242 & 2 @xmath235 210 s + he 0012 - 1441 & lvmslits & 1 & 2010 & aug & 11 & 1.75 & @xmath248 & 2 @xmath235 900 s + hd 88609 & lvmslits & 1 & 2008 & may & 6 & 1.22 & unknown & 30 s , 2 @xmath235 100 s + hd 115444 & lvmslits & 1 & 2008 & may & 6 & 1.04 & unknown & 200 s , 100 s + bs 16467 - 062 & lvmslits & 1 & 2008 & may & 6 & 1.00 & unknown & 4 @xmath235 1000 s + hd 122563 & lvmslits & 1 & 2008 & may & 6 & 1.20 & unknown & 3 @xmath235 5 s + bs 16550 - 087 & lvmslits & 1 & 2008 & may & 6 & 1.04 & unknown & 4 @xmath235 1000 s + cs 30325 - 028 & lvmslits & 1 & 2008 & may & 6 & 1.10 & unknown & 3 @xmath235 500 s + cs 30329 - 129 & longmirr & 1 & 2008 & apr & 11 & 1.52 & unknown & 2 @xmath235 120 s + bd @xmath2495 3098 & lvmslits & 1 & 2008 & may & 6 & 1.07 & unknown & 3 @xmath235 50 s + bs 16084 - 160 & longmirr & 1 & 2008 & apr & 11 & 1.29 & unknown & 2 @xmath235 60 s + bd @xmath2499 3223 & lvmslits & 1 & 2008 & may & 6 & 1.02 & unknown & 3 @xmath235 50 s + bs 16080 - 054 & longmirr & 1 & 2008 & apr & 11 & 1.36 & unknown & 3 @xmath235 120 s + cs 22878 - 101 & longmirr & 1 & 2008 & apr & 8 & 1.10 & unknown & 3 @xmath235 120 s + bs 16080 - 093 & longmirr & 1 & 2008 & apr & 11 & 1.33 & unknown & 3 @xmath235 120 s + bd @xmath24923 3130 & longmirr & 1 & 2008 & apr & 11 & 1.00 & unknown & 2 @xmath235 45 s + hd 165195 & lvmslits & 1 & 2008 & may & 6 & 1.05 & unknown & 3 @xmath235 10 s + hd 186478 & lvmslits & 1 & 2008 & may & 6 & 1.47 & unknown & 2 @xmath235 50 s , 300 s + bd @xmath25018 5550 & lvmslits & 1 & 2008 & may & 6 & 1.30 & unknown & 3 @xmath235 100 s + bd @xmath25017 6036 & lvmslits & 1 & 2008 & may & 6 & 1.47 & unknown & 3 @xmath235 200 s + cs 22880 - 086 & lvmslits & 1 & 2008 & may & 6 & 1.40 & unknown & 4 @xmath235 1000 s + he 2323 - 0256 & lvmslits & 1 & 2010 & aug & 11 & 1.18 & @xmath248 & 2 @xmath235 600 s + sculptor & scl1 & 86 & 2008 & aug & 3 & 1.79 & @xmath251 & 3 @xmath235 1200 s + & scl2 & 106 & 2008 & aug & 3 & 1.68 & @xmath251 & 2 @xmath235 900 s + & scl3 & 87 & 2008 & aug & 4 & 1.67 & @xmath252 & 462 s + & & & 2008 & aug & 31 & 1.67 & @xmath243 & 1000 s , 834 s + & scl5 & 95 & 2008 & sep & 1 & 1.73 & @xmath253 & 3 @xmath235 720 s + & scl6 & 91 & 2008 & sep & 1 & 1.88 & @xmath236 & 3 @xmath235 720 s + fornax & for1b & 166 & 2008 & sep & 1 & 2.00 & @xmath254 & 3 @xmath235 500 s + & & & 2008 & nov & 25 & 1.86 & @xmath255 & 2 @xmath235 1200 s + & for3b & 169 & 2008 & sep & 1 & 1.71 & @xmath251 & 3 @xmath235 500 s + & for4b & 164 & 2008 & nov & 26 & 1.74 & @xmath256 & 1200 s , 1020 s + & for6 & 169 & 2008 & aug & 31 & 1.73 & @xmath257 & 3 @xmath235 500 s + & & & 2008 & nov & 25 & 2.31 & @xmath258 & 2 @xmath235 1200 s + & for7 & 169 & 2008 & aug & 31 & 1.71 & @xmath259 & 2 @xmath235 500 s , 460 s + & & & 2008 & sep & 30 & 1.71 & unknown & 3 @xmath235 600 s + leo i & leoi_1 & 42 & 2003 & oct & 29 & 1.86 & unknown & 4800 s + & leoi_2 & 83 & 2004 & oct & 15 & 1.9 & unknown & 6900 s + & lin1_1 & 112 & 2006 & feb & 2 & 1.82 & unknown & 3600 s ( total ) + & lin1_2 & 100 & 2006 & feb & 2 & 1.23 & unknown & 3600 s ( total ) + & lin1_3 & 92 & 2006 & feb & 2 & 1.04 & unknown & 3600 s ( total ) + & lin1_4 & 100 & 2006 & feb & 2 & 1.01 & unknown & 2520 s ( total ) + & lin2_1 & 104 & 2006 & feb & 3 & 1.83 & unknown & 4800 s ( total ) + & lin2_2 & 98 & 2006 & feb & 3 & 1.20 & unknown & 4800 s ( total ) + & lin2_3 & 67 & 2006 & feb & 3 & 1.01 & unknown & 3600 s ( total ) + & lin2_4 & 102 & 2006 & feb & 3 & 1.07 & unknown & 3600 s ( total ) + & lin3_1 & 86 & 2006 & feb & 4 & 1.43 & unknown & 3011 s ( total ) + & lin3_2 & 67 & 2006 & feb & 4 & 1.21 & unknown & 4800 s ( total ) + & lin3_3 & 69 & 2006 & feb & 4 & 1.02 & unknown & 4500 s ( total ) + & lin3_4 & 88 & 2006 & feb & 4 & 1.04 & unknown & 1800 s ( total ) + sextans & sex1 & 108 & 2009 & feb & 22 & 1.53 & @xmath260 & 3 @xmath235 1200 s , 396 s + & sex2 & 85 & 2009 & feb & 22 & 1.21 & @xmath261 & 3 @xmath235 1200 s + & sex3 & 88 & 2009 & feb & 22 & 1.09 & @xmath262 & 4 @xmath235 1200 s + & sex4 & 109 & 2009 & feb & 23 & 1.51 & @xmath263 & 4 @xmath235 1200 s + & sex6 & 100 & 2009 & feb & 23 & 1.19 & @xmath264 & 3 @xmath235 1200 s + leo ii & l2a & 72 & 2006 & feb & 2 & 1.00 & unknown & 3300 s ( total ) + & l2b & 56 & 2006 & feb & 3 & 1.29 & unknown & 1080 s ( total ) + & l2c & 77 & 2006 & feb & 4 & 1.01 & unknown & 3960 s ( total ) + & l2d & 70 & 2006 & feb & 2 & 1.09 & unknown & 3600 s ( total ) + & l2e & 57 & 2006 & feb & 3 & 1.07 & unknown & 3600 s ( total ) + & l2f & 62 & 2006 & feb & 4 & 1.13 & unknown & 3400 s ( total ) + canes venatici i & cvn1 - 1 & 91 & 2007 & feb & 14 & 1.14 & unknown & 4140 s ( total ) + & cvn1 - 2 & 94 & 2007 & feb & 14 & 1.06 & unknown & 4140 s ( total ) + & cvn1 - 3 & 90 & 2007 & feb & 14 & 1.03 & unknown & 4860 s ( total ) + & cvn1-dp & 115 & 2007 & feb & 15 & 1.03 & unknown & 9000 s ( total ) + ursa minor & umi1 & 125 & 2009 & feb & 22 & 1.99 & @xmath265 & 3 @xmath235 1200 s + & umi2 & 134 & 2009 & feb & 22 & 1.79 & @xmath266 & 2 @xmath235 1200 s , 1400 s + & umi3 & 137 & 2009 & feb & 23 & 1.48 & @xmath267 & 3 @xmath235 1200 s + & umi6 & 137 & 2009 & feb & 23 & 1.57 & @xmath262 & 3 @xmath235 1200 s + draco & dra1 & 151 & 2009 & may & 23 & 1.27 & @xmath268 & 3 @xmath235 1200 s + & dra2 & 167 & 2009 & may & 23 & 1.28 & @xmath247 & 1200 s , 1000 s , 900 s + & dra3 & 140 & 2009 & may & 23 & 1.35 & @xmath269 & 3 @xmath235 960 s + & dra4 & 140 & 2009 & may & 23 & 1.50 & @xmath270 & 2 @xmath235 960 s , 700 s + & dra5 & 78 & 2009 & may & 24 & 1.28 & @xmath271 & 3 @xmath235 1200 s + & dra7 & 75 & 2009 & may & 24 & 1.35 & @xmath272 & 3 @xmath235 960 s , 1080 s + & dra8 & 60 & 2009 & may & 24 & 1.55 & @xmath242 & 3 @xmath235 840 s + & dra9 & 106 & 2009 & may & 24 & 1.28 & @xmath273 & 3 @xmath235 960 s + llcccccccc scl & 1002473 & @xmath274 & @xmath275 & 1.66 & @xmath276 & @xmath277 & @xmath278 & @xmath279 & @xmath280 + scl & 1002447 & @xmath281 & @xmath282 & 1.84 & @xmath283 & @xmath284 & @xmath285 & @xmath286 & @xmath287 + scl & 1002888 & @xmath288 & @xmath289 & 1.70 & @xmath290 & & @xmath291 & @xmath292 & @xmath293 + scl & 1003386 & @xmath294 & @xmath295 & 1.79 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath300 + scl & 1003505 & @xmath301 & @xmath302 & 1.93 & @xmath303 & @xmath304 & @xmath305 & @xmath305 & @xmath306 + scl & 1003443 & @xmath307 & @xmath308 & 1.84 & @xmath309 & @xmath310 & @xmath311 & @xmath312 & @xmath313 + scl & 1003537 & @xmath314 & @xmath315 & 2.01 & @xmath316 & @xmath317 & @xmath318 & @xmath319 & @xmath320 + scl & 1003694 & @xmath321 & @xmath322 & 2.07 & @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 + scl & 1003702 & @xmath328 & @xmath329 & 1.77 & @xmath330 & & @xmath331 & @xmath332 & @xmath333 + scl & 1003967 & @xmath334 & @xmath335 & 1.88 & @xmath336 & & @xmath337 & @xmath338 & @xmath339 + llccccccccccccccccccccc scl & 1002473 & 0.10 & 0.20 & 0.00 & 0.00 & 0.02 & 0.04 & 0.03 & 0.06 & 0.06 & 0.16 & 0.00 & 0.03 & 0.06 & 0.11 & 0.03 & 0.06 & 0.05 & 0.08 & 0.02 & 0.03 + scl & 1002447 & 0.13 & 0.26 & 0.00 & 0.00 & 0.09 & 0.12 & 0.02 & 0.03 & 0.07 & 0.16 & 0.00 & 0.01 & 0.07 & 0.12 & 0.02 & 0.03 & 0.02 & 0.08 & 0.04 & 0.04 + scl & 1002888 & 0.13 & 0.24 & 0.00 & 0.00 & 0.01 & 0.01 & 0.04 & 0.06 & 0.09 & 0.17 & 0.01 & 0.00 & 0.05 & 0.09 & 0.02 & 0.03 & 0.03 & 0.06 & 0.01 & 0.01 + scl & 1003386 & 0.13 & 0.26 & 0.02 & 0.04 & 0.03 & 0.04 & 0.02 & 0.05 & 0.08 & 0.19 & 0.04 & 0.03 & 0.04 & 0.09 & 0.01 & 0.04 & 0.06 & 0.11 & 0.04 & 0.04 + scl & 1003505 & 0.15 & 0.29 & 0.01 & 0.01 & 0.09 & 0.16 & 0.01 & 0.02 & 0.11 & 0.21 & 0.02 & 0.03 & 0.07 & 0.11 & 0.01 & 0.03 & 0.08 & 0.18 & 0.00 & 0.00 + scl & 1003443 & 0.13 & 0.25 & 0.01 & 0.02 & 0.11 & 0.11 & 0.01 & 0.01 & 0.08 & 0.14 & 0.01 & 0.02 & 0.03 & 0.05 & 0.01 & 0.02 & 0.03 & 0.06 & 0.01 & 0.02 + scl & 1003537 & 0.16 & 0.31 & 0.01 & & 0.08 & 0.13 & 0.02 & 0.02 & 0.12 & 0.24 & 0.01 & 0.02 & 0.06 & 0.12 & 0.01 & 0.02 & 0.08 & 0.16 & 0.00 & 0.01 + scl & 1003694 & 0.09 & 0.16 & 0.07 & & 0.02 & 0.01 & 0.05 & 0.06 & 0.12 & 0.23 & 0.02 & 0.01 & 0.09 & 0.12 & 0.06 & 0.08 & 0.20 & 0.29 & 0.02 & 0.06 + scl & 1003702 & 0.14 & 0.27 & 0.01 & 0.02 & 0.00 & 0.00 & 0.00 & 0.00 & 0.08 & 0.16 & 0.00 & 0.01 & 0.05 & 0.06 & 0.00 & 0.00 & 0.04 & 0.10 & 0.05 & 0.03 + scl & 1003967 & 0.14 & 0.27 & 0.01 & 0.02 & 0.00 & 0.00 & 0.00 & 0.00 & 0.08 & 0.15 & 0.01 & 0.02 & 0.04 & 0.08 & 0.01 & 0.01 & 0.01 & 0.08 & 0.01 & 0.02 + lllcccccccccc m79 & n1904-s80 & @xcite & 4250 & 0.75 & 2.50 & @xmath340 & @xmath341 & 4055 & 0.56 & 2.01 & @xmath342 & @xmath343 + ngc 2419 & n2419-s1305 & @xcite & 4275 & 0.70 & 2.10 & @xmath344 & @xmath345 & 4395 & 0.79 & 1.96 & @xmath346 & @xmath347 + ngc 2419 & n2419-s1305 & @xcite & 4275 & 0.70 & 2.10 & @xmath344 & @xmath345 & 4366 & 0.79 & 1.96 & @xmath348 & @xmath349 + m5 & iii-149 & @xcite & 4200 & 0.91 & 1.75 & @xmath350 & & 4218 & 1.05 & 1.89 & @xmath296 & @xmath351 + m5 & g18155_0228 & @xcite & 5270 & 3.25 & 1.44 & @xmath352 & @xmath353 & 5286 & 3.25 & 1.37 & @xmath354 & @xmath355 + m5 & ii-59 & @xcite & 4450 & 1.27 & 1.30 & @xmath350 & & 4443 & 1.32 & 1.83 & @xmath356 & @xmath357 + m5 & g18447_0453 & @xcite & 5275 & 3.15 & 1.44 & @xmath358 & @xmath359 & 5279 & 3.06 & 1.42 & @xmath360 & @xmath361 + m5 & 1 - 31 & @xcite & 4880 & 2.25 & 1.64 & @xmath362 & @xmath363 & 4903 & 2.14 & 1.63 & @xmath364 & @xmath365 + m5 & iv-59 & @xcite & 4229 & 0.79 & 2.10 & @xmath366 & & 4254 & 0.99 & 1.91 & @xmath367 & @xmath368 + m5 & g18484_0316 & @xcite & 4995 & 2.50 & 1.58 & @xmath369 & @xmath370 & 4980 & 2.41 & 1.57 & @xmath371 & @xmath351 + | we present a catalog of fe , mg , si , ca , and ti abundances for 2961 red giant stars that are likely members of eight dwarf satellite galaxies of the milky way ( mw ) : sculptor , fornax , leo i , sextans , leo ii , canes venatici i , ursa minor , and draco . for the purposes of validating our measurements , we also observed 445 red giants in mw globular clusters and 21 field red giants in the mw halo .
the measurements are based on keck / deimos medium - resolution spectroscopy combined with spectral synthesis .
we estimate uncertainties in [ fe / h ] by quantifying the dispersion of [ fe / h ] measurements in a sample of stars in monometallic globular clusters .
we estimate uncertainties in mg , si , ca , and ti abundances by comparing our medium - resolution spectroscopic measurements to high - resolution spectroscopic abundances of the same stars . for this purpose ,
our deimos sample included 132 red giants with published high - resolution spectroscopy in globular clusters , the mw halo field , and dwarf galaxies .
the standard deviations of the differences in [ fe / h ] and @xmath0\rangle$ ] ( the average of [ mg / fe ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] ) between the two samples is @xmath1 and @xmath2 , respectively . this catalog represents the largest sample of multi - element abundances in dwarf galaxies to date .
the next papers in this series draw conclusions on the chemical evolution , gas dynamics , and star formation histories from the catalog presented here .
the wide range of dwarf galaxy luminosity reveals the dependence of dwarf galaxy chemical evolution on galaxy stellar mass . |
the boolean convolution of probability measures on the real line was introduced and studied in @xcite . in particular , it turned out that all probability measures on the real line are infinitely divisible for the boolean convolution . in this note
we give a negative answer to schrmann s question , if the same is also true for the multiplicative boolean convolution of probability measures on the unit circle . in section [ sec - indep ]
we recall the definition of boolean independence and prove a crucial theorem on the distribution of two unitary operators @xmath0 and @xmath1 , if @xmath2 and @xmath3 are boolean independent , see theorem [ theo - f ] .
the boolean convolution for probability measures on the unit circle is then defined in section [ sec - conv ] .
we motivate the definition by the relation between the boolean product and the conditionally free product @xcite .
but definition [ def - bool - conv ] also agrees with the theory of lvy processes on dual groups .
i.e. , if we take the algebra of polynomials on the unit circle as a dual group in the sense of voiculescu @xcite and define lvy processes on this dual group as in @xcite , then the distributions of the increments of these processes form convolution semigroups in the sense of definition [ def - bool - conv ] . in propositions
[ prop - f ] and [ prop - mult ] we show that the transform @xmath4 with @xmath5 where @xmath6 for @xmath7 , reduces the multiplicative boolean convolution to the multiplication of holomorphic functions in the unit disk .
finally , in theorem [ theo - inf - div ] we give a lvy - khintchin formula that describes all probability measures on the unit circle that are infinitely divisible for the boolean convolution .
let @xmath8 be a hilbert space and @xmath9 a normal state on @xmath10 .
two operators @xmath11 are called _ boolean independent _ w.r.t . @xmath12 , if @xmath13 holds for all @xmath14 , @xmath15 , @xmath16 . for a unitary operator @xmath0 on @xmath8
we define functions @xmath17 and @xmath18 on the unit disk by @xmath19 and @xmath20 the following theorem will be crucial for our study of the boolean convolution of probability measures in section [ sec - conv ] .
[ theo - f ] let @xmath21 be two unitary operators on @xmath8 such that @xmath22 and @xmath23 are boolean independent w.r.t .
then we have @xmath24 for all @xmath25 .
let @xmath26 with @xmath27 .
note that the resolvent @xmath28 of @xmath29 can be written as @xmath30 we have chosen @xmath31 sufficiently large so that @xmath32 expands into the norm - convergent series @xmath33 introducing the operators @xmath34 , we get @xmath35 the operator @xmath36 can be written as a norm - convergent series @xmath37 with vanishing constant term . using equation , we therefore have @xmath38 where @xmath39 and @xmath40 .
we obtain the following formula for the cauchy transform @xmath41 of the distribution of @xmath42 in the state @xmath12 , @xmath43 since @xmath44 , we can rewrite this as @xmath45 substituting @xmath46 into this , we get after some simplification and a shift of the argument @xmath47 for @xmath27 . substituting now @xmath48 for @xmath49 and replacing @xmath50 by @xmath51 ,
this becomes finally @xmath52 for @xmath53 .
but by uniqueness of analytic continuation the identity holds for all @xmath54 .
if @xmath55 is self - adjoint , then there exists a uniquely determined compactly supported probability measure @xmath56 on @xmath57 such that @xmath58 holds for all @xmath59 .
we call this measure the distribution of @xmath60 in the state @xmath12 and denote it by @xmath61 .
let @xmath60 and @xmath62 be two self - adjoint operators on @xmath8 and assume that @xmath60 and @xmath62 are boolean independent w.r.t .
. then the joint moments @xmath63 of the pair @xmath60 and @xmath62 are uniquely determined by the distributions of @xmath60 and @xmath62 , since equation can be used to compute them from the moments of @xmath60 and @xmath62 .
in particular , the distribution of the sum @xmath64 is uniquely determined by the marginal distributions @xmath65 and @xmath66 and given by the boolean convolution @xmath67 it was actually this property that motivated the definition of the boolean convolution in @xcite .
let now @xmath0 be a unitary operator on @xmath8 .
then there exists a unique probability measure @xmath56 on the unit circle such that @xmath68 holds for all @xmath59
. we will denote this measure by @xmath69 and call it the distribution of @xmath0 in the state @xmath12 .
if @xmath21 are two boolean independent unitary operators , then the moments of their product are given by @xmath70 since only the first moments of @xmath0 and @xmath1 enter , this does not lead to an interesting convolution of probability measures on the unit circle .
the right definition can be deduced by recalling that the boolean product is a special case of the conditionally free product @xcite , which is defined for pairs of states .
looking at the algebra of continuous functions on the real line and taking evaluation at the origin for the second state , on recovers the additive boolean convolution .
the important property of the choice of the second state is that the dirac mass at the origin is an idempotent for the free convolution . if instead one takes the algebra of continuous functions on the unit circle and integration against the dirac measure at @xmath71 for the second state , one arrives at the following definition of the multiplicative boolean convolution .
[ def - bool - conv ] let @xmath72 be two unitary operators with distributions @xmath73 , @xmath74 .
assume furthermore that @xmath2 and @xmath3 are boolean independent .
then we call the probability measure @xmath75\cup\end{array}\!}\nu=\mathcal{l}_\varphi(uv)\ ] ] the _ boolean convolution _ of @xmath56 and @xmath76 .
it is uniquely determined by @xmath56 and @xmath76 .
we call a probability measure @xmath56 on the unit circle _ infinitely divisible _ w.r.t . the boolean convolution , if there exists a probability measure @xmath77 with @xmath78\cup\end{array}\!}\mu_n{\!\begin{array}{c } { \scriptstyle\times } \\[-12pt]\cup\end{array}\!}\cdots{\!\begin{array}{c } { \scriptstyle\times } \\[-12pt]\cup\end{array}\!}\mu_n } \\ & & \hspace*{10mm}n \mbox { times}\end{aligned}\ ] ] for all @xmath79 .
let @xmath56 be a probability measure on the unit circle .
then it is uniquely characterized by the function @xmath80 we shall also need the function @xmath81 [ exa - f ] let @xmath82 be a dirac measure in @xmath83 , @xmath84
. then we get @xmath85 and @xmath86 for @xmath54 .
let now @xmath87 , with @xmath88 and @xmath89 .
then we get @xmath90 and @xmath91 for @xmath54 , where @xmath92 .
denote by @xmath93 the uniform distribution on @xmath94 , i.e. the haar measure .
then we get @xmath95 and @xmath96 for @xmath54 .
consider now @xmath97 , where @xmath98 , for @xmath79 .
this is the haar measure on the cyclic subgroup @xmath99 of @xmath94 .
here we get @xmath100 and @xmath101 for @xmath54 .
the following proposition shows that the map @xmath4 establishes an isomorphism between probability measures on the unit circle and holomorphic functions in the unit disk .
[ prop - f ] the following are equivalent for a function @xmath102 on the unit disk @xmath103 .
we have @xmath104 and @xmath102 is holomorphic .
there exists a probability measure @xmath56 on the unit circle @xmath105 such that @xmath106 where @xmath107 the probability measure @xmath56 is uniquely determined by @xmath102 . if @xmath102 is not constant , then it is open and we can replace @xmath108 in part ( i ) by @xmath109 . if @xmath102 is constant and takes a value in @xmath110 , then the associated measure @xmath56 is the dirac measure in that point .
the function @xmath111 in ( ii ) has a zero at the origin and therefore @xmath112 on @xmath113 has a unique analytic extension to @xmath109 ( i)@xmath114(ii ) : let @xmath104 be holomorphic .
then @xmath115 is also holomorphic and has non - negative real part .
therefore @xmath116 has a herglotz representation @xmath117 see ( * ? ? ?
* theorem v.4.10 ) . here
@xmath118 and @xmath119 , i.e. @xmath76 is probability measure on the unit circle .
this implies @xmath120 where @xmath121 is the image of @xmath76 under the map @xmath122 .
the uniqueness of @xmath56 follows from the uniqueness of the herglotz representation .
( ii)@xmath114(i ) : the mbius transformations @xmath123 map the unit disk to the half plane with real part bigger @xmath124 for all @xmath125
. therefore @xmath126 is holomorphic and satisfies @xmath127 .
this implies that @xmath128 is holomorphic and maps the unit disk to itself . since @xmath129
, we also have @xmath130 and @xmath131 , @xmath132 , can be extended analytically to the origin .
the function @xmath102 satisfies @xmath133 letting @xmath134 go to zero , one gets @xmath135 for all @xmath7 by the maximum principle
. we can now restate theorem [ theo - f ] for the boolean convolution .
[ prop - mult ] let @xmath136 be probability measures on the unit circle
. then we have @xmath137\cup\end{array}}\!\!}\nu}(z)=f_\mu(z)f_\nu(z)\ ] ] for @xmath7 .
follows immediately from theorem [ theo - f ] and definition [ def - bool - conv ] .
the haar measure @xmath93 has @xmath138 .
it is infinitely divisible for the boolean convolution , since @xmath139\cup\end{array}}\!\!}n}=\lambda$ ] for all @xmath79 , and satisfies @xmath140\cup\end{array}\!}\mu=\lambda$ ] for all probability measures @xmath56 on the unit circle .
the following theorem describes all other infinitely divisible probability measures by a lvy - khintchin - type formula .
[ theo - inf - div ] a probability measure @xmath141 on the unit circle is infinitely divisible for the boolean convolution if and only if there exists a real number @xmath142 and a finite measure @xmath143 on @xmath94 such that @xmath144 with @xmath145 for @xmath7 .
the pair @xmath146 is called the _ characteristic pair _ of @xmath56 , it is uniquely determined by @xmath56 .
a probability measure @xmath56 on the unit circle is infinitely divisible for the boolean convolution if and only if there exist holomorphic functions @xmath147 with @xmath148 for all @xmath79 . if @xmath102 is not identically zero , then this happens if and only if @xmath149 . in this case
we can take the logarithm of @xmath102 , i.e. there exists a holomorphic function @xmath150 with @xmath151 for @xmath54 such that @xmath152 this function @xmath150 is uniquely determined up to an integer multiple of @xmath153 .
we choose @xmath150 such that @xmath154 .
since the real part of @xmath155 is non - negative , it has a herglotz representation @xmath156 with @xmath157 and @xmath143 a finite measure on the unit circle .
conversely , given a pair @xmath146 of a real number @xmath142 and @xmath143 a finite measure on @xmath94 , the functions @xmath158 with @xmath156 give by proposition [ prop - f ] probability measures @xmath56 and @xmath77 which satisfy @xmath159\cup\end{array}}\!\!}n}\ ] ] by proposition [ prop - mult ] . therefore @xmath56 is infinitely divisible for the boolean convolution .
the uniqueness follows from the uniqueness of the herglotz representation .
it follows that not all probability measures on the unit circle are infinitely divisible for the boolean convolution .
let @xmath141 .
if the first moment of @xmath56 vanishes , then @xmath160 and @xmath111 have a zero of order at least two at the origin .
therefore @xmath161 , but @xmath162 , and we can not take arbitrary holomorphic roots of @xmath163 .
this is equivalent to @xmath56 not being infinitely divisible .
it follows that every infinitely divisible probability measure @xmath56 can be embedded into a continuous convolution semigroup , just take for @xmath164 the convolution semigroup with the measures @xmath165 given by @xmath166 for @xmath167 , if @xmath146 is the characteristic pair of @xmath56 .
the embedding is not unique , because replacing @xmath168 by @xmath169 , @xmath170 , leads to another convolution semigroup @xmath171 which also satisfies @xmath172 .
probability measures with @xmath173\cup\end{array}\!}\mu=\mu$ ] are called idempotents . for the classical convolution on the unit circle
these are exactly the haar measures of compact subgroups of @xmath94 , i.e. the probability measures given in example [ exa - f ] ( d ) .
idempotents for the boolean convolution are characterized by the condition @xmath174 for @xmath7 .
since @xmath163 has to be holomorphic , this leaves only the constant functions @xmath175 and @xmath176 , which correspond to the dirac measure at @xmath71 and the haar measure on @xmath94 .
analytic functions on the unit disk have a unique factorization into a product @xmath177 , where @xmath178 is a blaschke product , @xmath179 is a singular function , i.e. of the form @xmath180 with a measure @xmath181 that is singular w.r.t . to the haar measure , and
@xmath182 is an outer function , i.e. of the form @xmath183 with @xmath184 , @xmath185 , and @xmath186 a real - valued integrable function , see , e.g. , @xcite . if @xmath187 for @xmath7 , then @xmath186 is non - negative .
the function @xmath178 has the form @xmath188 where @xmath189 are non - negative integers and @xmath190 is a sequence such that @xmath191 , @xmath192 for @xmath193 , and @xmath194 converges . by proposition [ prop - mult ] and theorem [ theo - inf - div ]
, we get a corresponding factorization @xmath195\cup\end{array}\!}\mu_s{\!\begin{array}{c } { \scriptstyle\times } \\[-12pt]\cup\end{array}\!}\mu_o\ ] ] of any probability measure @xmath56 on the unit circle into the boolean convolution of a not - infinitely divisible measure @xmath196 , an infinitely divisible singular measure @xmath197 and an infinitely divisible measure @xmath198 that is absolutely continuous w.r.t .
@xmath93 . for @xmath199
we get the dirac measure @xmath200 .
let @xmath201 and @xmath202 , @xmath203 .
then we get @xmath204 and @xmath56 is characterized by @xmath205 for @xmath7 .
since the real part of @xmath206 goes to zero almost everywhere when @xmath207 goes to the unit circle , the measure @xmath56 is singular with respect to the haar measure @xmath93 .
denote by @xmath208 the zeros of @xmath209 on the unit circle .
one can show that they have an accumulation point at @xmath210 . since @xmath116 has single poles in each @xmath211 , the measure @xmath56 has a dirac mass @xmath212 at @xmath213 .
let @xmath214 , @xmath215 , then the coefficient @xmath216 is given by @xmath217 we do not know , if @xmath56 also charges the point @xmath218 .
consider now the constant function @xmath219 with @xmath220 , @xmath84 .
this function also belongs to an infinitely divisible measure .
it is absolutely continuous with respect to the haar measure @xmath93 , with density given by the poisson kernel @xmath221 in @xmath222 .
the measure @xmath143 in the characteristic pair of @xmath56 is determined by the condition @xmath223 it is equal to @xmath224 , i.e. it is a scalar multiple of the haar measure .
i wish to thank rolf gohm for his remarks and advice .
m. schrmann .
non - commutative probability on algebraic structures . in h.
heyer , editor , _ proceedings of xi oberwolfach conference on probability measures on groups and related structures _ , pages 332356 .
world scientific , 1995 .
r. speicher and r. woroudi .
boolean convolution . in d.
voiculescu , editor , _ free probability theory .
papers from a workshop on random matrices and operator algebra free products , toronto , canada , march 1995 _ , volume 12 of _ fields inst .
_ , pages 267279 .
american mathematical society , providence , ri , 1997 . | we introduce the boolean convolution for probability measures on the unit circle . roughly speaking
, it describes the distribution of the product of two boolean independent unitary random variables .
we find an analogue of the characteristic function and determine all infinitely divisible probability measures on the unit circle for the boolean convolution .
la convolution boolenne de deux probabilits sur le cercle est dfinie comme la distribution du produit de deux oprateurs unitaires @xmath0 et @xmath1 tels que @xmath2 et @xmath3 .
un analogue de la fonction caractristique est donne et les lois infiniment divisibles pour cette convolution sont caractrises . |
in quantum cryptography unconditional security proofs are derived under the assumption that alice s and bob s apparata ( private spaces ) are completely inaccessible by an eavesdropper who , therefore , can only attack the signal systems which are transmitted through the quantum communication channel connecting the two parties . under this assumption ,
secret - key rates and security thresholds are derived in both discrete and continuous variable quantum key distribution .
one potential loophole in the security proofs is related to how a theoretical protocol is actually implemented experimentally .
any redundant information encoded in extra degrees of freedom or extra hilbert space dimensions outside the theoretical prescription can allow for so - called side - channel attacks
. by their nature , such attacks may be of classical or quantum degrees of freedom and are insidious because even quantifying their threat appears to involve understanding what have been called unknown unknowns about the vulnerability of the experimental set - up .
progress has been made on eliminating side - channel attacks in the quantum communication channels between private spaces , but this leaves open potential attacks on the private spaces through their quantum communication ports .
let us therefore take a step back and consider private spaces in more detail : what goes on in alice s and bob s private spaces involves a significant amount of classical information processing ; at the very least the key itself will be generated and stored as classical information .
now with virtually any technology we have today classical information is stored , processed and transmitted in a highly redundant fashion ( many electrons are used to charge a capacitor to represent a bit value , or many electrons must pass through the base junction of a transistor to effect a logical switching operation , tapping on a keyboard produces sound waves and electromagnetic signals in addition to the ` legitimate ' electrical signals in the wires , etc ) . in principle
any of this redundant information may leak out of the private space through a parasite channel .
an eavesdropper might therefore ignore the quantum communication channel and directly attack alice s and bob s apparata by exploiting the presence of parasite channels : this is also a side - channel attack .
the implicit assumption in quantum cryptography is that we could always improve technology in such a way that alice s and bob s private spaces are not affected by the presence of parasite channels , so that the legitimate participants do indeed have access to absolutely private spaces .
( for instance , alice and bob could simulate the classical information processing on a quantum computer .
a hacked operating system on such a machine could be tested for by randomly running subroutines that confirm that coherence is preserved and that no information is copied out to where it can be stored or transmitted by a trojan program see also ref .
however , even if you rely on a perfect isolation technology , there remains a potential chink in this armor , which is the quantum communication port used either to transmit a quantum state out of your private space or to accept a quantum state for detection into it .
if you open a communication port for quantum states to enter or leave you must explicitly deal with side channels which can be probing these links to your private space .
eve can potentially send trojan systems through alice s and bob s communication ports and detect their reflection to infer both state preparation and measurement settings . as an example , in the standard bb84 protocol , eve can irradiate alice s apparatus by using optical modes at slightly different frequencies . then , from reflection , eve can infer the polarization chosen in each round of the protocol .
thanks to this information , eve can measure each signal system in the correct basis .
another example regards the so - called plug - and - play systems , where trojan systems can be reflected together with signal systems , as discussed in ref .
@xcite .
our paper shows how to overcome the problem of the open quantum communication ports , therefore making feasible the notion of absolutely private spaces .
note that this problem is not addressed by current device - independent quantum cryptography , where such attacks on the private space ports are simply considered illegitimate as they violate the strong private space assumption .
the key point of our scheme is that detectors are no longer in line with the quantum communication port of the private space .
for this reason , it is not possible for an external party to probe the port and obtain detector settings or readouts from the processing of parasite systems . in order to explain this key feature in detail
, we analyze the problem of the quantum communication ports by comparing standard protocols with our scheme . in fig .
[ ex1 ] , we depict a general prepare - and - measure protocol , where alice s variable @xmath39 is encoded in a quantum state @xmath107 by modulation . bob s
variable @xmath40 is the output of a quantum measurement . here
, eve can attack the quantum communication ports by using two trojan systems @xmath44 and @xmath108 . by means of @xmath44 ,
eve can retrieve information about the state preparation @xmath109 . by means of @xmath108
, she can retrieve information about the measurement apparatus of bob and , therefore , about @xmath40 . in fig .
[ ex2 ] , we depict a general entanglement - based protocol , where an untrusted party ( eve ) distributes entanglement between two parties .
this is done by distributing an entangled state @xmath110 , where system @xmath50 is sent to alice and system @xmath60 is sent to bob .
alice and bob can perform entanglement distillation and measure the output distilled systems to derive two correlated classical variables , @xmath39 and @xmath40 , respectively . in this scenario ,
eve can decide not to attack the source @xmath2 but directly the two quantum communication ports of alice and bob .
eve can probe these ports by using two trojan systems @xmath44 and @xmath108 , which can retrieve information about alice s and bob s distilling and detecting apparata . as a result ,
eve can infer information about @xmath39 and @xmath40 . in fig .
[ ex3 ] , we depict our protocol where an untrusted party ( eve ) represents an entanglement swapper between alice and bob .
this is generally done by measuring two _ public _
systems , @xmath111 and @xmath112 , received from alice and bob , processing the outcome of the measurement , and classically communicating the processed data back to alice and bob . as a result
the two private systems , @xmath50 and @xmath60 , become correlated , so that alice and bob can extract two correlated classical variables , @xmath39 and @xmath40 , by applying suitable measurements . in particular ,
if alice and bob can access quantum memories , then they can extract a secret key at a rate which is at least equal to the coherent information between @xmath50 and @xmath60 .
eve can attempt a side - channel attack against the two ports by sending two trojan systems @xmath44 and @xmath108 . in this case , however , the apparata which detect the two private systems @xmath50 and @xmath60 are inaccessible to eve . by exploiting reflections from the ports , eve can only retrieve information regarding the reduced states @xmath113 and @xmath114 of the two public systems @xmath111 and @xmath112 .
however , these reduced states contain no useful information about the private system @xmath50 or @xmath60 or alice s or bob s detector settings or outputs .
to understand better how the full isolation of the private systems might be achieved , we may consider the procedure depicted in fig .
[ ports ] .
it is explained for alice s private space , but steps are identical for bob . in the first step ( a ) ,
alice s port is closed and she prepares an entangled state @xmath115 where system @xmath50 is directed towards a quantum memory ( qm ) , while system @xmath111 is directed towards a delay line ( dl ) . in step ( b ) , once system @xmath50 is stored in the memory and while system @xmath111 is trapped in the delay line , a shutter is used to fully separate the delay line from the rest of alice s apparatus .
note that a virtual channel between @xmath50 and @xmath111 has been created . in step ( c ) ,
alice s quantum communication port is opened and system @xmath111 is transmitted to eve . during this stage , trojan systems may enter the port but no detector is in line with the port . in step ( d ) , the port is closed with the private system @xmath50 kept in the memory . the previous steps ( a)-(d )
are repeated many times , so that alice collects many private systems in her quantum memory .
we therefore reach step ( e ) of the figure .
finally , once alice has received all the classical communications , she applies a collective quantum measurement on her quantum memory to retrieve the classical variable @xmath39 .
this measurement can include or be anticipated by an entanglement distillation .
in part of the derivation we adopt the enlarged hilbert space ( ehs ) representation , where stochastic classical variables are embedded in quantum systems . consider a stochastic variable @xmath55 which is encoded into an ensemble of states of some quantum system @xmath50 , i.e. , @xmath116 this ensemble may be equivalently represented by the classical - quantum ( cq ) state @xmath117 where the stochastic variable @xmath39 is embedded into the dummy quantum system @xmath118 , by using an orthonormal basis @xmath119 in the hilbert space @xmath120 of @xmath118 .
we denote by @xmath121 the state of a system @xmath50 which is conditioned by the value @xmath51 of a stochastic variable @xmath39 .
the notation @xmath122 refers to the conditional state @xmath121 where @xmath51 is not specified .
clearly , we have @xmath123 given a quantum system @xmath50 in a state @xmath124 , its von neumann entropy @xmath125 is also denoted by @xmath126 .
given a quantum system @xmath118 , embedding the stochastic variable @xmath39 , its quantum entropy @xmath127 is just the shannon entropy @xmath128 . given two quantum systems , @xmath50 and @xmath60 , we denote by @xmath129 their quantum mutual information
this is defined by @xmath130 where @xmath131 is the conditional quantum entropy .
note that @xmath132 can be negative and it is related to the coherent information by the relation @xmath133 for @xmath134 , the quantum mutual information @xmath135 , which is computed over the cq - state of eq .
( [ cq ] ) , corresponds to the holevo information @xmath136 , computed over the ensemble of eq .
( [ ensemble ] ) . for @xmath134 and @xmath137 , embedding two stochastic variables @xmath39 and @xmath40 , @xmath138 is just the classical mutual information @xmath139 . for three quantum systems
@xmath60 , and @xmath140 , we can consider the conditional quantum mutual information @xmath141 which is @xmath142 as a consequence of the strong subadditivity of the von neumann entropy . for a classically correlated system @xmath143
, we have a probabilistic average over mutual informations , i.e. , @xmath144 list of other useful elements : * given a tripartite quantum system @xmath145 , we can use the `` chain rule '' @xmath146 * invariance of the holevo information under addition of classical channels , i.e. , for a classical channel @xmath147 we have @xmath148 * given a markov chain @xmath149 , the classical mutual information decreases under conditioning @xcite , i.e. , @xmath150 notice that , for three general stochastic variables , we have @xmath151 , so that the so - called `` interaction information '' @xmath152 can be positive , negative or zero . * data processing inequality . for a markov chain @xmath153
, we have @xmath154
let us purify the mixed state @xmath98 into the pure state @xmath155 by introducing an ancillary system @xmath156 which is assumed to be in eve s hands ( so that eve s global system consists of @xmath157 ) .
this scenario is depicted in fig .
[ piclp2 ] .
thus , for the total state @xmath98 , we have @xmath159 .\ ] ] for the conditional state @xmath99 , generated by the measurement , we can write @xmath160 \nonumber\\ & = \frac{1}{p(x|l^{\prime})}\mathrm{tr}_{a\tilde{e } } \left [ \hat{a}% ( x)\phi_{abe\tilde{e}}(l^{\prime } ) \hat{a}(x)^{\dagger}\right ] \nonumber\\ & = \mathrm{tr}_{\tilde{e}}\left [ \phi_{be\tilde{e}}(x , l^{\prime})\right ] , \label{traceeq}%\end{aligned}\ ] ] where @xmath161 , \ ] ] represents the conditional state @xmath158 which is generated by the measurement in the purified scenario . clearly if we discard @xmath39 , we get the reduced state @xmath162 .\ ] ] because of eq .
( [ traceeq ] ) , the conditional state @xmath163 can be used to compute @xmath97 via @xmath164 where @xmath165 and @xmath166 ( the computation is exactly the same up to a trace over @xmath156 ) . in the ehs representation ,
the conditional state @xmath158 becomes @xmath167 thus , we can also set @xmath168 where @xmath169 . from the chain rule
we have @xmath170 where @xmath171 [ gamma]is the information contribution due to the purification @xcite .
in other words , the ( conditional ) holevo information can only increase with the purification , i.e. , @xmath172 as a consequence , we have @xmath173 , where @xmath174 in terms of conditional entropies , we have @xmath175 . \label{rsec}%\end{aligned}\ ] ] here @xmath176 is computed over @xmath177 discarding @xmath39 and @xmath60 , i.e. , over the reduced state @xmath178 .\ ] ] now since @xmath179 is pure , we have @xmath180 , where @xmath181 can be computed over @xmath182 $ ] . clearly , also @xmath183 can be computed over @xmath35 . as a consequence we can recognize in eq .
( [ rsec ] ) the conditional coherent information @xmath184 associated with alice and bob s conditional state @xmath35 .
thus , we can set @xmath185.\ ] ] here , we can assume that alice s measurement is a rank one povm .
as a result , @xmath166 is also a pure state , and we can set @xmath186 , so that @xmath187 . finally , we can write @xmath188 where we have used @xmath189 from its definition . b. qi , _
et al . _ , quantum inform
. comput .
* 7 * , 73 ( 2007 ) ; c .- h .
f. fung , _ et al .
_ , phys . rev . a * 75
* , 032314 ( 2007 ) ; y. zhao , _ et al .
_ , phys .
a * 78 * , 042333 ( 2008 ) ; l. lydersen , _ et al . _ , nature photonics * 4 * , 686 ( 2010 ) ; l. lydersen , _ et al .
_ , nature photonics * 4 * , 801 ( 2010 ) ; i. gerhardt , _ et al .
_ , nature comm .
* 2 * , 349 ( 2011 ) ; l.lydersen , _ et al . _ , new j. phys . * 13 * , 113042 ( 2011 ) .
d. mayers and a. yao , quantum inform .
* 4 * , 273 ( 2004 ) ; j. barrett , l. hardy and a. kent , phys .
lett . , 010503 ( 2005 ) ; a. acin , _ et al .
_ , phys .
rev.lett . * 97 * , 120405 ( 2006 ) ; a. acin , _ et al .
_ , phys.rev .
* 98 * , 230501 ( 2007 ) ; n. gisin , s. pironio and n. sangouard , phys .
lett . * 105 * , 070501 ( 2010 ) .
equivalently , we can adopt the ehs representation ( see supplementary material for details ) , where the ensemble @xmath190 and the stochastic variables @xmath39 and @xmath25 are described by a unique classical - quantum state @xmath191 .
the holevo quantity of eq .
( [ holb ] ) corresponds to the conditional quantum mutual entropy @xmath192 computed over this state .
note that the ehs representation has been mainly introduced to give the correct interpretation to the definition of @xmath195 , where a quantum system @xmath13 conditions a classical variable @xmath39 thanks to the embedding in a quantum system @xmath118 . | quantum key distribution ( qkd ) offers the promise of absolutely secure communications .
however , proofs of absolute security often assume perfect implementation from theory to experiment .
thus , existing systems may be prone to insidious side - channel attacks that rely on flaws in experimental implementation . here
we replace all real channels with virtual channels in a qkd protocol , making the relevant detectors and settings inside private spaces inaccessible while simultaneously acting as a hilbert space filter to eliminate side - channel attacks . by using a quantum memory
we find that we are able to bound the secret - key rate below by the entanglement - distillation rate computed over the distributed states . in 1982
richard feynman conjectured the use of quantum systems as a technological platform for solving difficult calculations in physics .
eventually this insight lead to the field of quantum information processing .
as part of the field s growth , it has partly diverged into the two main application domains : computation and communications , though much fundamental and technical overlap still exists .
interestingly , the key application that has started to mature and is now commercially available is quantum cryptography , or more precisely quantum key distribution ( qkd ) which has quickly moved from the purely theoretical @xcite to a practical technology @xcite .
how can we explain the impressive industrial uptake of quantum cryptography and its ultimate aim to take over classical systems ?
the answer lies in the claim of `` absolute security '' @xcite .
unfortunately , while the idea is very compelling , subtle details in implementation may introduce flaws that could , potentially , be open to attack .
specifically , attacks from so called `` side channels '' represent one of the most elusive threats in practical quantum cryptography , because a system could be vulnerable to side - channel attacks even if it is unbreakable in theory @xcite .
in fact , the recent approach of `` device - independent qkd '' @xcite makes important advances in handling imperfect implementations , and can even be made by untrusted parties , but does not directly address all possible side - channel attacks , where , for example , detectors may directly receive external probing aimed at seeding or gleaning their readout . in principle side - channel attacks
affect both classical and quantum cryptography , but could be especially devastating for quantum cryptography , precisely because of the proclaimed absolute security `` guarantee '' .
the threat from such attacks has been demonstrated in both lab and installed field settings @xcite .
thus , while practical qkd systems have been fighting a trade - off between distance and key generation rate , they are still facing the fundamental problem of guaranteed security , choosing to rely on theoretical promises of absolute security without having any way of authenticating them in practice .
0.0truein * private spaces : general model * let us consider the scenario of fig .
[ gen ] .
two authenticated parties , alice and bob , control two private spaces , @xmath0 and @xmath1 , respectively . conventionally , these spaces are assumed completely inaccessible from the outside , i.e. , no illegitimate system may enter @xmath0 or @xmath1 .
for this reason every kind of side - channel attack upon the private spaces is assumed excluded . in practice , however , any port can allow a side - channel to enter possibly probing any detector , state - generation or detector settings . to prevent or overcome such attacks
, the qkd system must effectively isolate its private spaces : the private space must not be directly involved in either state preparation ( for sending ) or detection ( of incoming states ) . to overcome such probing side - channel attacks , we propose performing state - generation by collapse of a bipartite entangled state , so that any probe from outside is perfectly isolated from the state - generation `` machinery '' ( see supplementary material for an extended discussion ) .
thus , in a manner akin to teleportation , we replace all real channels with virtual channels .
this allows us to physically ( and `` topologically '' ) separate all detectors and settings within the private space from external probing , while also acting as a hilbert space filter @xcite against any side channel . within its own private space
, each party ( alice or bob ) has a bipartite state @xmath2 which entangles two systems : @xmath3 for alice , and @xmath4 for bob .
systems @xmath5 are kept within the private spaces , while systems @xmath6 are sent to an untrusted third party ( utp ) , whose task is to perform a quantum measurement and communicate the corresponding result .
this untrusted locc then allows the creation of correlations between the private systems @xmath5 that alice and bob can exploit to generate a secret - key . in its simplest form
an ideal side - channel free qkd scheme reduces to an entanglement swapping setup @xcite , with the dual teleportation channel acting as an ideal hilbert space filter .
what is unique about our protocol is the ability to completely protect private space settings and detectors from probing side - channel attacks . in the worst case scenario ,
the utp must be identified with eve herself , whose aim is to eavesdrop the key , or even prevent alice and bob from generating the key ( i.e. , a denial of service ) . in the most general case
, eve applies a quantum instrument @xmath7 to the incoming systems @xmath8 .
this is a quantum operation with both classical and quantum outputs . for each classical outcome @xmath9
, there is a corresponding completely positive ( cp ) map @xmath10 applied to the systems @xmath6 @xcite .
this means that the global input state @xmath11 is transformed into the conditional output state @xmath12 where @xmath13 represents an output quantum system in the hands of eve , while @xmath14 is the identity channel acting on the private systems @xmath5 .
cleary each outcome @xmath9 will be found with some probability @xmath15 , depending both on @xmath10 and the input state . as a consequence
the classical output of @xmath16 can be simply represented by the stochastic variable @xmath17 .
the quantum output of @xmath16 is represented by the system @xmath13 which is correlated with the private systems @xmath5 via the conditional state @xmath18 specified by eq .
( [ condstate ] ) .
@xmath13 is the system that eve will use for eavesdropping .
for instance , most generally eve can store all the output systems @xmath13 ( generated in many independent rounds of the protocol ) into a big quantum memory .
then , she can detect the whole memory using an optimal quantum measurement ( corresponding to a collective attack ) . according to the agreed protocol ,
the utp must send a classical communication ( cc ) to both alice and bob in order to `` activate '' the correlations . here
, eve has another weapon in her hands , i.e. , tampering with the classical outcomes . in order to decrease the correlations between the honest parties , eve
may process the output stochastic variable @xmath19 via a classical channel @xmath20 and then communicate the fake variable @xmath21 to alice and bob , where @xmath22 this process projects the private systems @xmath5 onto the conditional state @xmath23 where @xmath24 notice that , if @xmath25 is completely unrelated to @xmath19 , then eve realizes a denial of service , being the communication of the fake variable equivalent to tracing over systems @xmath8 . in other words , for @xmath26 , we have @xmath27 , where @xmath28 and @xmath29 . 0.1truein
* secret - key rate : general analysis * after @xmath30 rounds of the protocol , alice and bob will share @xmath30 copies @xmath31 .
note that , in general , alice and bob do not know anything about the physical process within the utp , i.e. , they do not know the couple @xmath32 .
for this reason , what they actually get are @xmath30 copies of an unknown state @xmath33 plus classical information @xmath25 . however , by measuring a suitable number @xmath34 of these copies , they are able to deduce the explicit form of the conditional state @xmath35 for the remaining @xmath36 copies ( here @xmath37 @xmath34 and @xmath38 are large numbers )
. then , by applying local measurements , alice on her private systems and bob on his , they are able to extract two correlated classical variables , @xmath39 and @xmath40 . finally , from these variables , they can derive a shared secret key via the classical techniques of error correction ( ec ) and privacy amplification ( pa ) .
these procedures can be implemented using one - way classical communications between these two parties .
let us bound the secret - key rate of the protocol . for simplicity
we omit here the conditioning on @xmath25 , so that eq .
( [ abstate ] ) simply becomes @xmath41 .
it is understood that the final result must be averaged over @xmath25 . independently from its generation , the ( generally ) mixed state @xmath42 can be purified in a pure state @xmath43 by introducing a suitable system @xmath44 to be assigned to eve ( this is generally larger than the @xmath13 system considered before ) .
after this purification , the scenario is the one depicted in fig .
[ abe ] . here ,
for every bipartition of the systems , @xmath45 @xmath46 , or @xmath47 , the corresponding reduced states have the same von neumann entropy .
in particular , we have @xmath48 .
now suppose that alice performs a povm @xmath49 on her system @xmath50 with classical outcome @xmath51 .
this measurement projects @xmath52 onto the conditional state@xmath53 , \label{fibe}%\ ] ] where @xmath54.\ ] ] thus alice encodes the stochastic variable @xmath55 in the nonlocal ensemble @xmath56 . given the conditional state @xmath57 of eq .
( [ fibe ] ) , bob and eve can only access their local states , respectively given by @xmath58 , \qquad \rho_{e}(x)=\mathrm{tr}_{b}\left [ \phi_{be}(x)\right ] .\ ] ] thus , on his side , bob has the ensemble @xmath59 , whose measurement estimates alice s variable @xmath39 . assuming that bob has a quantum memory
, he can collect all the private systems @xmath60 associated to the @xmath38 rounds of the protocol .
then , asymptotically for @xmath61 , bob can reach the holevo bound @xcite @xmath62.\ ] ] at the same time , eve s information is bounded by @xmath63.\ ] ] assuming one - way ccs from alice to bob ( for implementing ec and pa ) , we can write the secret - key rate as a difference of holevo informations @xcite , i.e.,@xmath64 if we now assume that alice s povm is rank one , then the conditional state @xmath57 is pure and , therefore , @xmath65 and @xmath66 have the same entropy , i.e. , @xmath67=s[\rho_{e}(x)]$ ] . as a consequence
, we can write@xmath68 where @xmath69 is the coherent information between alice and bob .
thus the secret - key rate is lower - bounded by the entanglement - distillation rate . 0.1truein * secret - key rate : detailed analysis * here we make a more detailed analysis which is more closely connected to the scenario of fig .
[ gen ] .
in fact , the rate @xmath70 of eq .
( [ rate1 ] ) comes from the general configuration of fig .
[ abe ] , which is independent from the actual process generating the final state of alice and bob .
if we explicitly consider the peculiarities of the scheme of fig .
[ gen ] , then we could achieve a larger rate @xmath71 . this new rate can be achieved if alice and bob have some knowledge of the classical unreliability of the utp , i.e. , of the amount of information which is `` absorbed '' by the classical channel @xmath72 .
thus , if eve tries to tamper with the overall security by employing fake ccs , then alice and bob can potentially extract a secret - key with rate larger than the entanglement - distillation rate . in this section ,
we take the different conditionings ( by @xmath19 and @xmath25 ) explicitly into account . after the cc of @xmath21 , alice and
bob possess the conditional state @xmath73 of eq .
( [ abstate ] ) .
let us assume that alice performs a povm @xmath49 on her system @xmath50 with classical outcome @xmath51 .
this generates the doubly - conditional state @xmath74,\ ] ] where @xmath75.\ ] ] averaging over the ccs , the output of alice s measurement is the unconditional variable @xmath55 , where @xmath76 .\ ] ] this is the secret variable to be estimated by bob . in his private system @xmath60
, bob has the ensemble @xmath77 where @xmath78 .
clearly , this ensemble depends on both @xmath39 and @xmath25 . exploiting his knowledge of @xmath79
, bob applies a conditional measurement @xmath80 to his system @xmath60 which estimates the value @xmath51 encoded by alice .
asymptotically ( i.e. , for @xmath81 ) , using a quantum memory and averaging over the ccs ( i.e. , over @xmath25 ) , bob can reach the conditional holevo information @xcite @xmath82 for eve we have to consider the different conditioning given by @xmath19 .
thus , the conditional state that eve shares with alice is @xmath83 which becomes @xmath84 after alice s projection .
explicitly this state is given by @xmath85 , \ ] ] where @xmath86 .\ ] ] thus , eve has the ensemble @xmath87 where @xmath88 .
asymptotically , eve can eavesdrop @xmath89 bits per copy @xcite . as a result ,
we can write the secret - key rate @xmath90 this quantity can be rewritten as @xmath91 , where @xmath92 and @xmath93 , quantifies the information which is `` absorbed '' by the classical channel @xmath72 .
we call @xmath94 the `` classical cheating '' by eve
. clearly , we have @xmath95 for @xmath96 .
@xmath97 is the `` apparent rate '' , which refers to the apparent scenario where alice , bob and eve are all subject to the same conditioning @xmath25 . in other words , @xmath97 is computed assuming the total state @xmath98 , which is then projected onto @xmath99 by alice s measurement ( see fig . [ piclp1 ] ) . projected onto @xmath99.,scaledwidth=42.0% ] we can now easily prove that the secret - key rate is larger than the entanglement - distillation rate .
we have the following result ( see supplementary material for the proof ) .
* theorem*. _ suppose that eve measures the incoming systems but cheats on the results using a classical channel _
@xmath100 _ then , alice and bob s secret - key rate satisfies _
@xmath101 _ where _ @xmath102 _ is the coherent information conditioned to eve s fake variable _ @xmath25 _ _ , and _ _
@xmath94 _ is the classical cheating_. our analysis leaves an intriguing open question
. it would be wonderful to provide an explicit example where simultaneously @xmath103 and @xmath104 , so that @xmath105 .
this would imply secret - key distillation without entanglement distillation .
more generally , we can not exclude the possibility that @xmath106 by using povms which are not rank one .
0.1 truein * conclusion * we have shown that virtual channels may replace real channels in the qkd setting so as to remove any possibility of side - channel attacks . in its simplest setting , our qkd protocol corresponds to an entanglement swapping experiment , where the dual teleportation channels act as ideal hilbert space filters to wipe out side - channel attacks .
the authenticated users private spaces are designed so that any incoming quantum signal is topologically excluded from access to detectors , detector settings or state - generation settings , thus side - channel probing attacks of the private spaces are eliminated .
finally , an external untrusted party performs a suitable locc ( such as a bell - state measurement ) to create correlations necessary for shared key generation .
0.1 truein * acknowledgments * the research leading to these results has received funding from epsrc under grant no .
ep / j00796x/1 ( hipercom ) .
0.1 truein * note added * we noticed the manuscript of lo _ et al . _
@xcite submitted to the arxiv just four days before our submission . at a 2009 dagstuhl seminar
, slb told the first author of that manuscript the detailed ingredients needed to turn entanglement swapping into a completely side - channel free qkd scheme . despite this , we stress that there are several important differences between our work and that of lo _ et al . _
@xcite which itself is based on previous results by biham _
et al . _ and inamori @xcite .
\1 ) _ generality _ : our work considers a qkd setup involving general quantum systems ( e.g. , they could be qubit - systems or bosonic modes ) . in the case of lo _ et al . _
@xcite , the setup is strictly related to the use of decoy states .
\2 ) _ untrusted party _
: our work does not impose any restriction for the action of the middle eavesdropper , who implements a general quantum instrument with classical information communicated to alice and bob . despite this general scenario
, we can easily prove that the secret - key rate is lowerbounded by the coherent information . in lo _
et al . _
@xcite , a very strong conjecture is made explicitly for the middle eavesdropper , who is always assumed to perform a bell measurement .
thus , in practice , the third party can not be regarded as an eavesdropper ( eve ) but more correctly as a trusted party ( charlie ) .
the real action of eve is therefore restricted to the noise present in the quantum channels between the honest users and charlie .
\3 ) _ trojan - horses _ : our work excludes the possibility that the eavesdropper is able to get information by sending trojan - horses into alice s and bob s apparata .
this is the result of a hilbert space filtering which derives from a suitable use of the entanglement in alice s and bob s private spaces ( this aspect is further discussed in our supplementary material ) .
by contrast , in the paper lo _
et al . _
@xcite , state preparation can be easily eavesdropped upon , being directly inline with the output ports of alice s and bob s private spaces . finally , in lo _
et al . _
@xcite there is a disconnect between the security proof ( which relies on alice and bob creating entanglement ) and the implementation .
we note however that the authors cite ref .
@xcite as a possible way to avoid quantum information processing but do not seriously discuss this as part of their proof in any detail . supplementary material |
understanding the formation and evolution of galaxies is difficult because it involves many different processes that are coupled in a complex way . in order to simulate a scenario properly
, one has to consider the dynamics of stars , gas and dark matter as well as star formation , the interaction with the interstellar medium ( ism ) , and chemical enrichment @xcite . since this problem includes many unknown parameters , it is more effective to decouple various processes and to closely inspect observational constraints which are relevant to the different parameters .
chemical evolution models constrain star formation histories , supernova rates , and abundances in the ism , in the stars , and in the intracluster medium ( icm ) .
thus , trying to reproduce element abundances in chemical simulations already puts significant constraints on galaxy formation without considering complicated dynamical aspects .
hence , in the present stage it is more effective to decouple the dynamical and the chemical approach .
different timescales for the duration of the star forming phase cause different abundance ratios in the stars and in the ism . in short phases of star formation short - living ,
massive stars govern the enrichment of the ism . thus ,
in those formation scenarios the abundance ratios reflect the type ii supernova ( sn ii ) production ( hashimoto , iwamoto , & nomoto 1993 ) . in order to obtain the various element abundances in the solar neighbourhood , a modest and continuous star formation rate ( sfr )
is necessary @xcite . in the standard models ( e.g. matteucci & greggio 1986 ; timmes , woosley , & weaver 1995 ; tsujimoto et al .
1995 ) , this kind of star formation history is obtained with infall of primordial gas on a timescale of several gyr and an sfr which depends on the gas density via the schmidt - law @xcite . in this scenario ,
the disk of the galaxy is assumed to form out of slowly accreting gas .
since star formation occurs over a long timescale of @xmath1 yr , the chemical evolution is noticeably influenced by sn ia . in bulges and elliptical galaxies ,
[ @xmath2/fe]-ratios seem to be enhanced with respect to solar abundances ( peletier 1989 ; worthey , faber , & gonzalez 1992 ; davies , sadler , & peletier 1993 ; mcwilliam & rich 1994 ) .
since the @xmath2-elements are mainly produced in massive stars experiencing sn ii explosions @xcite , and iron is substantially contributed by type ia supernovae ( sn ia ) , the chemical history of the light dominating component of the stellar population in bulges and ellipticals must be dominated by massive stars . in chemical evolution models ,
this can be realized by either @xcite 1 . a flat imf or 2 . a short phase of star formation or 3 . a low fraction of close binary systems experiencing sn ia
chemical evolution models have to constrain and quantify these different possibilities .
although in the pure chemical approach the number of input parameters is already reduced , there are still plenty of uncertainties in the calculations .
typical input parameters are the shape and slope of the initial mass function ( imf ) , the sfr , the infall rate and the fraction of close binary systems producing iron via type ia supernovae . however , chemical evolution is also very sensitive to the adopted stellar yields , especially of sn ii ( see also gibson 1997 ) .
thus , besides the parameters above , chemical evolution models should always take into account different stellar nucleosynthesis prescriptions , which are strongly affected by uncertainties of stellar evolution models ( thielemann , nomoto , & hashimoto 1996 ) . in this paper , we compare two recently published nucleosynthesis calculations for sn ii by : 1 .
woosley & weaver , hereafter ww95 2 .
thielemann et al . and nomoto et al . , hereafter tnh96 .
we focus on the question , if the considered sets of models are able to explain an important observed feature of galaxy formation : the [ mg / fe ] overabundance .
there is a broad consensus that metal - poor halo stars in our galaxy have magnesium enhanced abundance ratios ( gratton & sneden 1988 ; magain 1989 ; edvardsson et al .
1993 ; axer , fuhrmann , & gehren 1994 ; axer , fuhrmann , & gehren 1995 ; fuhrmann , axer , & gehren 1995 ) .
the exact value of the enhancement is still debatable , but it seems to converge to @xmath3 dex @xcite .
this observation can be easily understood taking into account that metal - poor stars form in the early stages of the galaxy formation , when the enrichment due to sn ii is dominating chemical evolution .
however , we will show that there are still unresolved problems caused by uncertainties in stellar nucleosynthesis .
as already noted , in elliptical galaxies there are strong indications from spectra in the visual light that there is a magnesium overabundance of at least 0.2 dex in nuclei of these galaxies ( worthey et al .
1992 ; weiss , peletier , & matteucci 1995 ) . however , while the halo in the solar neighbourhood has low metallicities ( @xmath4[fe / h]@xmath5 ) , the stars which dominate the visual light in the nuclei of ellipticals have solar or super - solar @xmath6 @xcite .
therefore , the [ mg / fe ] overabundance is realized at both low and high @xmath6 in two considerably different systems . while a detailed inspection of element abundances in elliptical galaxies will be the subject of a forthcoming paper , in this work we concentrate on the chemical evolution of the solar neighbourhood using ww95 and tnh96 sn ii yields . in order to calibrate our code
, we use the same approach as in the most common chemical evolution models ( matteucci & greggio 1986 ; timmes , woosley , & weaver 1995 ) for the solar neighbourhood , performing the calculation for both sets of nucleosynthesis prescriptions .
taking finite stellar lifetimes into account , the classical numerical models relax the instantaneous _ recycling _ approximation , but usually assume the stellar ejecta to mix instantaneously with the ism @xcite .
only few attempts have been made in the literature to relax the instantaneous _ mixing _ approximation in numerical models of chemical evolution ( see discussion in timmes et al .
we present a modification of the basic equations @xcite splitting the gaseous mass into two different phases , one including the stellar ejecta and the second being cool and well mixed .
the mixing process is characterized by a gas flow from the first , inactive to the active , star forming gas phase .
we additionally present the results of simulations considering this modification . in section [ sec : general ]
we summarize the most important aspects of chemical evolution , while in section [ sec : yields ] we discuss stellar yields from sn ii explosions , comparing ww95 and tnh96 .
we analyse their influence on chemical evolution by calculating the ssp - yields of various elements in section [ sec : ssp ] . in section [ sec : solar ] we present our model for the chemical evolution in the solar neighbourhood . in the conclusion
we summarize the main results .
non - primordial elements develop in a cycle of birth and death of stars .
these form out of the ism , process elements , and eject them during the late stages of their evolution in the form of stellar winds , planetary nebulae ( pn ) , or supernovae ( sn ia and sn ii ) , depending on their main sequence mass @xmath7 . the formation of stars and the re - ejection of gas can be described by the following phenomenological equations @xcite : @xmath8 the total baryonic mass @xmath9 is governed by infall or outflow @xmath10 of material , being either primordial or enriched gas .
the total stellar mass @xmath11 is _ increasing _ according to the sfr @xmath12 and _ decreasing _ due to re - ejection @xmath13 of gas .
the total gaseous mass @xmath14 behaves exactly contrary to @xmath11 with the additional component of in - falling or out - flowing gas @xmath10 . the ejection rate @xmath13 is obtained by integrating the ejected mass fraction @xmath15 , folded with the sfr and the normalized imf @xmath16 , from the turnoff mass @xmath17 to the maximum stellar mass @xmath18 . here
, @xmath19 denotes the mass fraction of the remnant . @xmath20
the quantity @xmath21 is the stellar lifetime of a star with mass @xmath7 . in the instantaneous _ recycling _ approximation ,
@xmath21 is assumed to be negligible in comparison to the time @xmath22 .
this approximation is relaxed in numerical simulations as well as in our evolution code .
this must not be confused with the instantaneous _ mixing _ approximation ( i m a ) which is assumed in most chemical evolution models @xcite . in section [ sec :
solar ] , we also relax this assumption and take a delay in the mixing of the stellar ejecta into account .
parallel to equation [ eq : gas ] , the mass production of the element @xmath23 in the ism ( @xmath24 ) is expressed in the equation below : @xmath25 here , @xmath26 is the abundance of element @xmath23 in the ism , @xmath27 is the abundance of element @xmath23 in the in - falling or out - flowing gas
. the element ejection rate @xmath28 is obtained by integrating the ejected mass fraction @xmath29 ( including both initial abundance and newly produced material ) of the element @xmath23 , again folded with the sfr and imf over the appropriate mass range , equivalent to equation [ eq : ejection ] : @xmath30 equations [ eq : gas ] and [ eq : gasmetal ] can be combined to describe the progression of the abundance @xmath26 of element @xmath23 in the ism .
@xmath31 the key value in this equation is the stellar yield @xmath29 hidden in the element ejection rate @xmath32.we neglect stellar winds during the evolution , and assume that the stars enrich the ism at the time when they die .
depending on their initial mass , they either experience a sn ii explosion ( @xmath33 ) or become a white dwarf blowing off their envelopes ( @xmath34 ) .
elements heavier than oxygen are mainly processed in supernovae .
a substantial fraction of iron is contributed by sn ia . adopting the description of the supernova rates from greggio & renzini , the element ejection rate integrated over the total mass range
can be described with the following equation as in matteucci & greggio ( 1986 ) and timmes et al .
( 1995 ) : @xmath35 in this equation , the enrichment due to stars in the mass range @xmath36 is splitted into a contribution by type ii supernovae ( @xmath37 ) plus planetary nebulae ( @xmath38 ) , due to single stars , and type ia supernovae ( @xmath39 ) , assumed to be the end product of close binary evolution . in the formulation of greggio & renzini ( 1983 ) @xmath40 is the ratio between the mass of the secondary and the total mass @xmath7 of the system .
the maximum fraction of the secondary is 0.5 by definition , while the minimum mass @xmath41 is dependent on the turnoff as defined in the following equation @xcite : @xmath42\ ] ] one has to integrate the distribution of the secondary component @xmath43 , folded with the yield ( independent of the mass of the system ) and the sfr over the appropriate mass range .
the clock of the sn ia explosion is given by the lifetime of the secondary @xmath44 , whose mass can be as low as @xmath45 @xcite .
thus , the enrichment due to sn ia is substantially delayed with respect to sn ii .
the degree of influence by sn ia highly depends on the fraction @xmath46 of close binaries , which is a free parameter in chemical evolution .
greggio & renzini calibrate @xmath46 on the ratio between the current type ii to type ia supernova rates in the galaxy .
it should be noticed that equation [ eq : all_enrich ] is not completely consistent , since @xmath7 refers to the mass of single and binary stars .
nevertheless , as long as the parameter @xmath46 is small ( as in our case ) type ia events can be regarded as a small perturbation .
hence , equation [ eq : all_enrich ] is an acceptable approximation , and it allows us to describe the delayed release of iron from sn ia . in these terms , the rates of type ii and ia supernovae can be described by the following equations : @xmath47 @xmath48 in spite of the several unknown parameters ( imf slope @xmath49 , sfr as a function of time , stellar yields ) , the following arguments show how they can be constrained by different observational information , step by step .
this is useful for the interpretation of the results from chemical evolution calculations . assuming the instantaneous _ recycling _ approximation
, one can take @xmath12 out of the integral in eqn .
[ eq : ejection ] . for large @xmath22 ,
the residual integral expressing the returned fraction @xmath50 depends strongly on the imf slope @xmath49 but only marginally on turnoff mass ( and then time ) .
equation [ eq : stars ] can then be written as : @xmath51 the integrated solution demonstrates that the final total mass of stars depends on the _ time - average _ star formation rate @xmath52 and the imf slope . @xmath53 where the subscript zero refers to the initial conditions .
similar considerations show that the final mass of gas in the ism depends on a mean sfr @xmath52 , a mean infall rate @xmath54 and the imf slope .
now we need an approximation to pin down @xmath49 . for this purpose
, we consider the element abundances at the time @xmath22 . the above approximation applied to equation
[ eq : gasmetal ] leads to @xmath56 with @xmath57 as the returned mass fraction of element @xmath23 mostly dependent on the imf slope and the stellar yield .
if we consider an element whose stellar yield is well known , we get a constraint on @xmath49 ( with known abundance of the in - falling gas ) , for a given minimum mass of the stellar population @xmath58 . to summarize , the stellar and gaseous mass at the current epoch and the abundance of a specific element whose yield is relatively certain constrain @xmath59 and @xmath49 .
we can then draw conclusions on the nucleosynthesis of other elements , whose stellar yields are relatively uncertain . in other words , from chemical evolution of galaxies
one can get a constraint on the stellar evolution models .
this is a convincing example how tight these two disciplines are coupled .
since the approximation @xmath60 is especially valid for elements produced mainly by type ii supernovae , we will use this strategy in section [ sec : solar ] to fix the imf slope with the element oxygen and constrain the necessary magnesium yield in sn ii .
the imf is as usual assumed to be a declining function of mass , according to a power law : @xmath61 .
since the imf is usually normalized _ by mass _ , the actual amount of mass created in one generation of stars is controlled by the sfr @xmath12 .
@xmath62 in these terms , the slope @xmath63 corresponds to the salpeter - value @xcite .
to avoid uncertain extrapolations of the stellar yields to the high mass end , we have adopted @xmath64 which is the maximum mass for which ww95-models are computed .
tnh96 do give the yields for a @xmath65-star .
for the comparison between the two sets of models we keep @xmath18 fixed at @xmath66 .
the effect of adopting @xmath67 is explored in section [ sec : cutoff ] and [ par : cutoff ] .
the lower cutoff of one generation of stars is assumed to be @xmath68 .
the higher the minimum mass , the larger is the fraction of massive stars , thus more metals are produced .
abundance ratios , however , are not affected by the choice of @xmath58 .
alternative formulations of the imf with different slopes at different mass ranges exist in the literature ( e.g. scalo 1986 ; kroupa , tout , & gilmore 1995 ; gould , bahcall , & flynn 1997 ) .
however , in order to keep the number of free parameters low , we have decided to fix @xmath68 and use one specific slope @xmath49 for the whole mass range .
the value of @xmath49 , instead , is treated as a free parameter .
in our calculations , we use the results in renzini & voli ( 1981 ) for the enrichment due to intermediate mass single stars ( @xmath69 ) .
in particular we select the models with @xmath70 , @xmath71 .
type ia sne are assumed to occur in close binary systems @xcite . in this model ,
the explosion is caused by a carbon - deflagration of the material accreting degenerate white dwarf ( hansen & wheeler 1969 ; nomoto 1980 ; weaver & woosley 1980 ; nomoto 1981 ) .
we adopt the results of the nucleosynthesis from the classical w7-model by nomoto , thielemann , & yokoi ( 1984 ) .
low mass stars , in the range @xmath72 to @xmath73 , do not contribute to the enrichment of o , mg and fe @xcite .
type ia supernovae produce significantly more iron than oxygen or magnesium as can be seen in table [ tab : snia - yields ] .
one can see that @xmath74fe is clearly dominating the ejecta .
it follows that sn ii must be the main contributor to the @xmath2-elements enrichment .
.the most abundant elements ejected in a type ia supernova . in the calculations , nomoto et al
( 1984 ) assume accreting white dwarfs in close binary systems to be the progenitors of sn ia events .
the data refer to the w7 model , the values are given in @xmath75 .
the numbers show that the ejecta of sn ia are clearly dominated by @xmath76 . [ cols= " <
, > , < , > , < , > " , ] [ tab_helium ] the difference in hydrogen ( and then helium ) yields comes from two causes : a different @xmath77-@xmath7 relation at he ignition and the fact that tnh96 neglect the h - shell burning occurring after he ignition . in this respect
we notice that in ww95 models , the @xmath77 of a @xmath78 star is @xmath79 , @xmath80 larger than that adopted by tnh96 for the same initial mass , on the basis of the @xmath77-@xmath7 relation by sugimoto & nomoto .
the he - yield of this star is @xmath81 larger than in tnh96 models , reflecting the he - production due to the h - burning shell .
a fair comparison between the predictions of the two sets of models should be done at constant @xmath77 . however , since we lack the @xmath77-values for ww95 models for masses other than @xmath78 , we proceed comparing the elements production for the same initial mass . the yield of oxygen and total ejected mass of all elements heavier than helium ( @xmath6 ) are plotted in figure [ oxy_z ] .
the figure shows that @xmath6 is clearly dominated by oxygen .
both depend only weakly on initial metallicity except for the @xmath82 case .
the results of ww95 and tnh96 are similar , except that tnh96 produce more oxygen in the higher mass range .
this can be understood in terms of the higher @xmath83-rate in the tnh96 models .
it is worth noting that the large difference in the yields of high mass stars may also result from the fact that ww95 consider fall back of material , whereas tnh96 do not .
the o - yield of tnh96 increases rapidly with mass , the ww95-yields , instead , seem to saturate .
this discrepancy already indicates that there is a huge uncertainty concerning the stellar yields of high mass stars ( @xmath84 ) . at the lower mass range ( @xmath85 ) ,
ww95 yields tend to be slightly larger for the same metallicity ( @xmath86 ) , possibly because of the larger @xmath77 .
however , the similarity of the results of the two sets of models suggests that the uncertainty in the oxygen yield from sn ii is small . both , ww95 and tnh96 produce a mg - yield with a rapid rise in a certain mass range at @xmath87 for tnh96 and at @xmath88 for ww95 . as a consequence , in the mass range @xmath89 ,
the mg yield of tnh96 is larger by about a factor of @xmath90 .
we will briefly investigate the origin of this discrepancy , which , as we will show , is very significant in the context of chemical evolution .
@xmath91 is mainly produced during hydrostatic carbon - burning .
thus , table [ tab : diffmag ] gives the yields of @xmath92 and of the main carbon - burning products @xmath93 @xcite .
@xmath91 is produced in the following reaction @xcite : @xmath94 hence , the model producing more carbon should also produce more magnesium .
table [ tab : diffmag ] shows , that the carbon yields are systematically higher in ww95 for all stellar masses .
this is reasonable when taking into account the larger helium cores and the lower @xmath83-rate of ww95 .
however , for the yields of @xmath95 , @xmath96 and @xmath91 this is not the case for all masses . in general , for low mass stars ( @xmath97 ) the yields of the carbon - burning products @xmath95 and @xmath96 are higher in ww95 models as well . with the exception of the @xmath66-star , the higher masses exactly inverse this pattern .
the yield of @xmath91 behaves similar , but the effect is much stronger with the largest discrepancy for the @xmath98-star .
ww95 argue that the larger extent of the convective shells in the tnh96-models ( schwarzschild criterion ) is responsible for the above behaviour . since the observations of magnesium overabundance can be better explained with high mg - yields in sn ii ( see following sections ) , this could be interpreted as an argument in favour of the schwarzschild criterion in convection theory .
rrllllll & & @xmath99&@xmath100&@xmath101&@xmath98&@xmath78&@xmath66 + + @xmath102c & ww95 & @xmath103&@xmath104&@xmath105&@xmath106&@xmath107&@xmath108 + & tnh96 & @xmath109&@xmath110&@xmath111&@xmath112&@xmath113&@xmath114 + @xmath115o & ww95 & @xmath116&@xmath117 & 1.13 & 1.94 & 3.25 & 6.03 + & tnh96 & @xmath118&@xmath119&@xmath120 & 1.61 & 3.15 & 9.34 + @xmath121ne & ww95 & @xmath122&@xmath123&@xmath124&@xmath125&@xmath126 & 1.24 + & tnh96 & @xmath127&@xmath128&@xmath129&@xmath130&@xmath131&@xmath132 + @xmath133na & ww95 & @xmath134&@xmath135&@xmath136&@xmath137&@xmath138&@xmath139 + & tnh96 & @xmath140&@xmath141&@xmath142&@xmath143&@xmath144&@xmath145 + @xmath146 mg & ww95 & @xmath147&@xmath148&@xmath149&@xmath150&@xmath151&@xmath152 + & tnh96 & @xmath153&@xmath154&@xmath155&@xmath156&@xmath157&@xmath158 + [ tab : diffmag ] similar to the o - yields , the mg - yields of ww95 seem to saturate or even decline for increasing mass above @xmath66 , due to re - implosion .
according to the tnh96 calculations , instead , a huge amount of magnesium is ejected by high mass stars .
figure [ iron_mag ] shows , that the iron yield declines for masses between @xmath99 and @xmath98 in both sets of models .
the iron yields in the lower mass range are very similar , both models match the observational constraints at @xmath159 ( sn1993j ; e.g. baron , hauschildt , & young 1995 ; nomoto , iwamoto , & suzuki 1995 ) and @xmath98 ( sn1987a ; e.g arnett et al .
. table [ tab : ni_obs ] shows that both groups can reproduce the observed @xmath160 of the supernova events , which is dominating the iron yield .
thus , in the lower mass range , tnh96 and ww95 basically agree in the fe - yield .
llll @xmath161 ( @xmath75 ) & observation & ww95 ( @xmath86 ) & tnh96 + + @xmath162 & @xmath163 & @xmath164 & @xmath165 + @xmath166 & @xmath167 & 0.088 & 0.074 + [ tab : ni_obs ] however , ww95 produce significantly more iron than tnh96 in stars of @xmath168 , especially in model b and c. thus , mainly this mass range will be responsible for discrepancies in the fe - yields of the total mass range of sn ii ( ssp - yields , see section [ sec : ssp ] ) . for the discussion of chemical evolution , it is more meaningful to consider stellar yields weighted by the imf .
normalized on the sn ii yield of the whole mass range , these values give the relative contribution of a @xmath72 interval to the total sn ii yield . to show the role of various mass intervals to the enrichment of a certain element , we plot the imf weighted yields of the elements oxygen , magnesium , iron , and metallicity for different imf slopes and both nucleosynthesis models ww95 and tnh96 . in more detail , we plot the following quantity : @xmath169 the figures are given in the appendix [ ap : cont ] . summarizing the plots , we obtain the following results : 1 .
there is no specific mass range dominating the o- and @xmath6-enrichment significantly for all considered imf slopes . 2 .
the imf weighted mg - yield is slightly peaked at @xmath170 ( ww95 ) and @xmath98 ( tnh96 ) , but again these masses do not dominate the ssp - yield significantly .
3 . the fe - enrichment due to sn ii , instead
, is clearly governed by stars of @xmath85 . altogether ,
the plots demonstrate the increasing weight of the higher mass range with decreasing imf slope @xmath49 . in figure [ cont_mgfe ] , we plot the abundance ratio [ mg / fe ] produced in ww95 ( @xmath86 ) and tnh96 as a function of stellar mass .
the figure shows that the ratio of magnesium to iron is basically increasing with mass . in the intermediate mass range , the overabundance in tnh96 models exceeds the results of ww95 significantly . the maximum overabundance in tnh96
is reached in the most massive star ( @xmath171 ) , whereas mg / fe of the ww95-models peaks at @xmath172 . according to ww95 , a magnesium overabundance
is only produced in stars with @xmath173 ( except for the small peak at @xmath174 ) .
thus , in the first @xmath175 myr , when the turnoff is above @xmath98 ( see upper x - axis in figure [ cont_mgfe ] ) , these stars will enrich the ism with highly magnesium overabundant ejecta .
but already 30 myr after the beginning of star formation , the turnoff of @xmath176 is reached and the whole sn ii generation of stars is contributing to the enrichment .
thus , the key value for the discussion of chemical evolution is the ssp - yield .
we calculated ssp - yields of the elements oxygen , magnesium , iron in the mass range of type ii sne for different imf slopes and both sets of sn ii nucleosynthesis .
the tables in the appendix [ ap : tables ] give the abundances of the considered elements in the ejecta of sn ii explosions of one ssp ( @xmath64 ) .
the basic conclusion for the discussion of the yields are : 1 .
the highest [ mg / fe]-ratio in ww95 is produced in model b assuming an initial metallicity of @xmath177 .
this ratio is lowest for models c because of the high iron yield
the second highest value for [ mg / fe ] is produced in the models with @xmath178 .
the results for @xmath179 and @xmath180 are in between .
hence , [ mg / fe ] is neither increasing nor decreasing with initial metallicity .
the high and low ww95 metallicities _ do not _ bracket the expected ssp - yields as claimed by gibson , loewenstein , & mushotzky .
3 . tnh96 produce systematically higher [ mg / fe]-ratios than ww95 . for salpeter - imf
, the magnesium abundance in the sn ii ejecta is 0.13 dex higher with tnh96-models than with ww95-models .
the iron abundance , instead , is 0.08 dex lower . in total
, this leads to a [ mg / fe]-ratio , which is 0.21 dex higher for tnh96 nucleosynthesis .
we define the _ time dependent _ ssp - yield for element @xmath23 at time @xmath22 as @xmath181 this equation describes the abundance of element @xmath23 in the ejecta of one generation of stars of one single metallicity at the time @xmath22 .
with progression of time , the turnoff mass decreases and @xmath182 converges to the standard ssp - yield , integrated over the whole mass range .
we consider enrichment due to pn , sn ii , and sn ia ( see also eqn .
[ eq : all_enrich ] ) .
since tnh96 models are computed only for solar metallicities , we consider ww95-yields of solar initial metallicity .
the stellar lifetimes are taken from schaller et al . .
for the following computations we have extrapolated tnh96 yields to @xmath183 , and neglected the contribution from sn ii coming from stars with mass in the range @xmath184 ( see ww95 ) .
the fraction of close binary systems is @xmath185 .
as discussed in section [ sec : solar ] this value is calibrated in the chemical evolution model of the solar neighbourhood .
since the value of this parameter is very small , the yields of sn ii are only marginally affected by the choice of @xmath46 .
the exact number becomes important when star formation timescales of @xmath175 gyr and enrichment due to sn ia are considered .
for the remnant masses , we adopt renzini & voli ( 1981 ) up to @xmath73 , and either ww95 or tnh96 from @xmath183 to @xmath66 . in the range
@xmath184 the mass of the remnant is taken to be @xmath186 .
figure [ sspt_mg ] shows the abundance of magnesium in the ejecta of one dying generation of stars as a function of turnoff mass .
the abundances are normalized to solar values and plotted on a logarithmic scale .
the magnesium abundance in the ejecta is significantly super - solar .
the upper x - axis shows the progression of time which is not linear with the turnoff mass .
the turnoff of @xmath187 is reached after 0.341 gyr , but it takes more than 7 gyr until stars of @xmath72 contribute to the enrichment as well .
the different line styles belong to various imf slopes .
the solid line indicates the salpeter - imf .
one can see that the magnesium abundance in the total ejecta is decreasing with turnoff mass for @xmath188 .
this is due to the fact that most magnesium is processed in stars more massive than @xmath98 ( figure [ iron_mag ] ) .
the sn ii - ssp - yield is reached at @xmath189 . for @xmath190 , sn ia and pn begin to contribute .
but since both events do not eject a significant amount of magnesium ( see table [ tab : snia - yields ] ) , the abundance is still decreasing with decreasing turnoff mass and increasing time .
the most striking aspect of this diagram is that the magnesium abundance due to tnh96 nucleosynthesis is @xmath191 dex higher ( salpeter imf , model b ) .
this is caused by the significant difference in the yields of @xmath89 stars as shown in figure [ iron_mag ] .
hence , the discrepancy between ww95 and tnh96 is maximum at @xmath192 . since iron is mainly synthesized in stars of lower masses , the time dependent ssp - yield is roughly constant ( ww95 ) or even increasing ( tnh96 ) with increasing turnoff mass ( fig .
[ sspt_fe ] ) until the sn ii value is reached . since stars between @xmath184 are assumed not to contribute to the enrichment of heavy elements , there is a peak at @xmath183 . at late times
, the contribution due to type ia sne comes into play , and the iron abundance in the ejecta is rising again .
it is important to recognize that the iron abundance in the ejecta of sn ii is higher for a steeper imf ( see dotted curve ) .
the results from tnh96-models are more strongly dependent on the slope of the imf than ww95 , because the contribution of high mass stars to the iron production is smaller in tnh96 ( see also figure [ cont_mgfe ] ) . for the same reason , the difference between ww95 and tnh96 increases for a flatter imf . 1 . in the first 10 myr ,
the produced magnesium overabundance is fairly high , the difference between ww95 and tnh96 is extremely large .
ww95 yields reach [ mg / fe]@xmath193 at a turnoff @xmath194 , after 10 myr . even considering a flat imf with @xmath195 , the minimum overabundance in ellipticals of 0.2 dex @xcite
is reached at @xmath196 myr when the contribution of type ii sne is not yet complete .
tnh96 provide the same value of [ mg / fe ] after 7.3 gyr when sn ia explosion are already tearing the ratio down .
obviously , this strongly affects the timescales of star formation of a system showing a [ mg / fe ] overabundance .
one generation of sn ii exploding stars can not produce the magnesium overabundance in metal - poor stars in the solar neighbourhood , when considering salpeter imf and ww95-sn ii yields . assuming a value of [ mg / fe]@xmath197 which is already a lower limit all low - metallicity stars should have been born in the first 9 million years ( see also section [ sec : solar ] ) .
3 . with progression of time , the overabundance is decreasing , because more and more low mass stars ( @xmath198 ) with higher iron and lower magnesium yields are contributing to the enrichment . once the turnoff mass is @xmath73 , the final sn ii - ssp value is reached .
4 . the magnesium overabundance in the sn ii output increases according to flattening of the imf .
this is because of giving more weight to magnesium producing high mass stars .
the dependence of [ mg / fe ] on the imf slope is increasing with time .
this is understandable , because for a larger considered mass range the role of the imf slope becomes more important .
the tnh96-models provide a [ mg / fe]-ratio of 0.26 dex , ww95-models lead to [ mg / fe]=0.05 dex , both for salpeter - imf .
ww95 specify an uncertainty of a factor of 2 in the iron yield . since this causes a shift by 0.3 dex , one could argue that simply taking half of the iron yield would solve the overabundance problem .
we want to show that this is not the case .
since the iron yields of stars below @xmath98 are very similar in the two sets of models , and since both sets reproduce the observations of sn1987a and sn1993j very well ( see table [ tab : ni_obs ] ) , it is reasonable only the halve the iron yields of stars above @xmath98 in ww95 .
re - calculating the ssp - yields with the modified fe - yield of ww95 , it turns out that the total [ fe ] is only shifted by 0.08 dex to lower values .
figure [ sspt_mgfe2 ] shows the [ mg / fe]-ratio as a function of the turnoff mass in this experiment .
the plot shows , that it remains difficult to reproduce the observed magnesium overabundances with ww95 nucleosynthesis .
it is important to mention that also tnh96 magnesium yields may not suffice to explain observed [ mg / fe ] overabundances in elliptical galaxies .
there are several indications that [ mg / fe ] in nuclei of ellipticals does even exceed 0.4 dex @xcite .
as demonstrated in figure [ sspt_mgfe ] , this value can not be theoretically produced by one sn ii - exploding generation of stars .
hence , claiming [ mg / fe]@xmath199 dex , the star forming phase in giant ellipticals must be of the order @xmath200 yr , even for tnh96-yields and a flat imf ( @xmath201 ) .
a detailed exploration of star formation timescales , imf - slopes and stellar yields in elliptical galaxies will be the subject of a forthcoming paper .
finally , one should not forget , that the number of input parameters in the calculations is very small .
the above conclusions do not depend on galaxy formation scenarios , on star formation histories , on infall models , or on binary fractions .
the only considered parameters are the imf slope and stellar yields .
figure [ sspt_oxfe ] shows the time dependent ssp yield as a function of turnoff mass for the abundance ratio [ o / fe ] . since both , oxygen and magnesium ,
are produced mainly in type ii supernovae , one would expect similar values for the overabundance . for tnh96 ,
this is exactly the case .
the contribution of low mass stars ( @xmath202 ) to the enrichment of oxygen manifests itself in an elongation of the sn ii plateau down to @xmath203 .
although sn ia produces 6 times more oxygen than magnesium , the sn ia tears the curve down by approximatly the same amount because of the dominant role of iron in the ejecta .
the ww95-[o / fe]-ratio , instead , is 0.12 dex higher than [ mg / fe ] .
the discrepancy between ww95 and tnh96 [ o / fe]-ratios now originates mainly from the discrepancy in the iron yields , while oxygen yields differ by 0.02 dex .
this leads again to the conclusion , that ww95 may underestimate the magnesium yield . in this paragraph
, we will investigate the influence of a variation of the upper mass cutoff on the calculated ssp - yields .
for this purpose , we include the results of tnh96 for the @xmath65-star . in order to compare the different nucleosynthesis prescriptions
, we have to extrapolate the ww95-yields to higher masses , hence the result has to be interpreted with cautions . in the ww95-models most heavy elements re - implode for the massive stars , so that the contribution of these stars to the enrichment is negligible . indeed , the plots in figures [ oxy_z ] and [ iron_mag ] show this trend for the elements oxygen and magnesium , respectively ( also model b ) .
tnh96 do not consider fall back , thus their o - mg - yields increase with mass up to @xmath171 .
table [ tab : delta ] gives the variation of the _ abundances _ of various elements in the sn ii ejecta of one ssp , if @xmath67 with respect to the @xmath64 case .
these reflect both the metal production and the total ejected mass in the range @xmath204 .
lrrrrrrrrrrrr & & & & + + & @xmath205[o ] & @xmath205[mg ] & @xmath205[fe ] & @xmath205[o ] & @xmath205[mg ] & @xmath205[fe ] & @xmath205[o ] & @xmath205[mg ] & @xmath205[fe ] & @xmath205[o ] & @xmath205[mg ] & @xmath205[fe ] + tnh96 & @xmath206 & @xmath207 & @xmath208 & @xmath209 & @xmath210 & @xmath208 & @xmath211 & @xmath212 & @xmath208 & @xmath213 & @xmath214 & @xmath215 + ww95 & @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath217 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & @xmath191 & @xmath224 & @xmath225 + 1 .
the iron abundance in the ejecta decreases for all models .
this effect is strongest for flatter imf and ww95-models .
2 . the oxygen abundance , instead , increases for all models .
again the effect is strongest for a flatter imf , but more important in tnh96 .
the behaviour of the magnesium abundance is more complex .
for tnh96 the increase of magnesium becomes more significant with a flatter imf , whereas for ww95 the pattern is inverse . in total ,
the [ mg / fe]-ratio in the ssp - ejecta increases significantly only for tnh96-models . in the ww95-models ,
the effect of fall back prevents a significant change .
this is confirmed in figure [ fig : sspt_mgfe3 ] in which we show the abundance ratios as a function of turnoff mass and time with @xmath67 considered .
the diagram shows that assuming a larger value for @xmath18 the problem of the magnesium overabundance is relaxed , _ if the high mg - yield calculated by tnh96 in high mass stars is correct .
_ this seems to us still controversial .
it is of great importance in the future to improve the knowledge on the stellar yields of these stars , too .
furthermore , it is worth noting that the abundance ratio of magnesium to oxygen decreases with the inclusion of stars more massive than @xmath66 .
this is important because it becomes even more difficult to reproduce the solar mg / o ratio .
we give a detailed exploration on this aspect in paragraph [ par : cutoff ] of the next section .
the tables [ tab : a1.70 ] to [ tab : c0.70 ] show the abundances of oxygen , magnesium and iron in the ejecta of one generation of sn ii - exploding stars ( ssp - yields ) .
one table refers to a certain imf - slope and explosion model of ww95 ( a , b , c ) .
the values are normalized on ( meteoritic ) solar abundances from anders & grevesse and expressed on a logarithmic scale .
rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath369 & @xmath370 & @xmath371 & @xmath372 & @xmath373 & @xmath374 + [ mg ] & @xmath369 & @xmath375 & @xmath376 & @xmath377 & @xmath378 & @xmath379 + [ fe ] & @xmath380 & @xmath376 & @xmath381 & @xmath382 & @xmath374 & @xmath371 + [ o / fe ] & @xmath383 & @xmath384 & @xmath385 & @xmath386 & @xmath218 & @xmath387 + [ mg / fe ] & @xmath383 & @xmath388 & @xmath389 & @xmath390 & @xmath391 & @xmath208 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath369 & @xmath392 & @xmath393 & @xmath394 & @xmath395 & @xmath396 + [ mg ] & @xmath369 & @xmath397 & @xmath398 & @xmath380 & @xmath373 & @xmath382 + [ fe ] & @xmath380 & @xmath398 & @xmath399 & @xmath374 & @xmath395 & @xmath400 + [ o / fe ] & @xmath383 & @xmath401 & @xmath402 & @xmath389 & @xmath386 & @xmath403 + [ mg / fe ] & @xmath383 & @xmath404 & @xmath405 & @xmath406 & @xmath223 & @xmath386 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath369 & @xmath407 & @xmath393 & @xmath408 & @xmath409 & @xmath369 + [ mg ] & @xmath369 & @xmath410 & @xmath411 & @xmath407 & @xmath373 & @xmath372 + [ fe ] & @xmath380 & @xmath408 & @xmath400 & @xmath396 & @xmath412 & @xmath409 + [ o / fe ] & @xmath383 & @xmath391 & @xmath413 & @xmath223 & @xmath414 & @xmath415 + [ mg / fe ] & @xmath383 & @xmath390 & @xmath208 & @xmath416 & @xmath417 & @xmath418 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath420 & @xmath415 & @xmath407 & @xmath408 & @xmath395 & @xmath421 + [ mg ] & @xmath422 & @xmath403 & @xmath399 & @xmath410 & @xmath411 & @xmath407 + [ fe ] & @xmath380 & @xmath423 & @xmath376 & @xmath400 & @xmath393 & @xmath424 + [ o / fe ] & @xmath425 & @xmath426 & @xmath427 & @xmath428 & @xmath370 & @xmath429 + [ mg / fe ] & @xmath425 & @xmath430 & @xmath415 & @xmath418 & @xmath220 & @xmath370 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath420 & @xmath431 & @xmath432 & @xmath433 & @xmath434 & @xmath435 + [ mg ] & @xmath422 & @xmath431 & @xmath374 & @xmath373 & @xmath409 & @xmath408 + [ fe ] & @xmath380 & @xmath400 & @xmath436 & @xmath394 & @xmath395 & @xmath373 + [ o / fe ] & @xmath425 & @xmath437 & @xmath438 & @xmath428 & @xmath428 & @xmath402 + [ mg / fe ] & @xmath425 & @xmath437 & @xmath439 & @xmath440 & @xmath389 & @xmath415 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath420 & @xmath441 & @xmath432 & @xmath433 & @xmath434 & @xmath422 + [ mg ] & @xmath422 & @xmath400 & @xmath374 & @xmath441 & @xmath409 & @xmath395 + [ fe ] & @xmath380 & @xmath432 & @xmath441 & @xmath434 & @xmath442 & @xmath433 + [ o / fe ] & @xmath425 & @xmath418 & @xmath443 & @xmath208 & @xmath218 & @xmath428 + [ mg / fe ] & @xmath425 & @xmath391 & @xmath444 & @xmath391 & @xmath225 & @xmath220 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath446 & @xmath428 & @xmath382 & @xmath396 & @xmath433 & @xmath434 + [ mg ] & @xmath447 & @xmath403 & @xmath424 & @xmath411 & @xmath393 & @xmath382 + [ fe ] & @xmath380 & @xmath448 & @xmath449 & @xmath411 & @xmath441 & @xmath424 + [ o / fe ] & @xmath450 & @xmath451 & @xmath452 & @xmath427 & @xmath385 & @xmath453 + [ mg / fe ] & @xmath454 & @xmath455 & @xmath456 & @xmath386 & @xmath389 & @xmath456 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath446 & @xmath457 & @xmath458 & @xmath435 & @xmath420 & @xmath459 + [ mg ] & @xmath447 & @xmath460 & @xmath432 & @xmath409 & @xmath434 & @xmath433 + [ fe ] & @xmath380 & @xmath382 & @xmath371 & @xmath394 & @xmath408 & @xmath393 + [ o / fe ] & @xmath450 & @xmath461 & @xmath462 & @xmath403 & @xmath403 & @xmath429 + [ mg / fe ] & @xmath454 & @xmath416 & @xmath463 & @xmath464 & @xmath443 & @xmath456 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath446 & @xmath421 & @xmath442 & @xmath422 & @xmath420 & @xmath447 + [ mg ] & @xmath447 & @xmath394 & @xmath396 & @xmath421 & @xmath412 & @xmath369 + [ fe ] & @xmath380 & @xmath434 & @xmath395 & @xmath458 & @xmath422 & @xmath442 + [ o / fe ] & @xmath450 & @xmath220 & @xmath385 & @xmath464 & @xmath389 & @xmath443 + [ mg / fe ] & @xmath454 & @xmath225 & @xmath370 & @xmath414 & @xmath440 & @xmath218 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath466 & @xmath428 & @xmath441 & @xmath412 & @xmath458 & @xmath442 + [ mg ] & @xmath467 & @xmath375 & @xmath380 & @xmath373 & @xmath395 & @xmath441 + [ fe ] & @xmath378 & @xmath468 & @xmath376 & @xmath469 & @xmath372 & @xmath371 + [ o / fe ] & @xmath470 & @xmath471 & @xmath463 & @xmath429 & @xmath439 & @xmath472 + [ mg / fe ] & @xmath473 & @xmath474 & @xmath403 & @xmath413 & @xmath428 & @xmath403 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath466 & @xmath379 & @xmath475 & @xmath459 & @xmath476 & @xmath477 + [ mg ] & @xmath467 & @xmath478 & @xmath458 & @xmath434 & @xmath422 & @xmath442 + [ fe ] & @xmath378 & @xmath372 & @xmath424 & @xmath394 & @xmath394 & @xmath374 + [ o / fe ] & @xmath470 & @xmath391 & @xmath479 & @xmath452 & @xmath383 & @xmath480 + [ mg / fe ] & @xmath473 & @xmath417 & @xmath425 & @xmath456 & @xmath427 & @xmath375 + rrrrrrr & tnh96 & ww95 & ww95 & ww95 & ww95 & ww95 + & & @xmath364 & @xmath365 & @xmath366 ) & @xmath367 ) & @xmath368 + + [ o ] & @xmath466 & @xmath412 & @xmath481 & @xmath459 & @xmath447 & @xmath482 + [ mg ] & @xmath467 & @xmath432 & @xmath458 & @xmath434 & @xmath420 & @xmath435 + [ fe ] & @xmath378 & @xmath442 & @xmath396 & @xmath435 & @xmath475 & @xmath420 + [ o / fe ] & @xmath470 & @xmath215 & @xmath405 & @xmath415 & @xmath370 & @xmath385 + [ mg / fe ] & @xmath473 & @xmath414 & @xmath413 & @xmath440 & @xmath208 & @xmath208 +
the chemical evolution of our galaxy is treated in the literature several times ( e.g. matteucci & greggio 1986 ; timmes et al .
1995 ; pagel & tautvaisiene 1995 ; tsujimoto et al .
1995 , yoshii et al .
the model predictions fit the data quite well , the main observational features can be reproduced . in the classical numerical models ,
the chemical evolution of the ism in the solar neighbourhood is described in a one - zone model of homogeneous and instantaneously mixing gas .
the latter assumption is called the instantaneous _ mixing _ approximation . the instantaneous _ recycling _ approximation , which neglects the stellar lifetimes is relaxed in these models as well as in our calculations . in principle
, the accretion of gaseous matter over a timescale of @xmath226 gyr enables us to avoid the formation of extremely metal - poor stars with [ fe / h]@xmath227 , known as the g - dwarf problem @xcite .
the formation of the disk in the solar vicinity due to accretion @xmath228 is described in the following equation @xcite : @xmath229\ ] ] @xmath230}\ , \label{eq : infall}\ ] ] with @xmath231 and @xmath232 as the surface densities ( @xmath233 ) of the total mass ( stars+gas ) today and at the beginning of the disk formation , respectively .
the accretion timescale for the formation of the disk is controlled by the parameter @xmath234 .
the sfr is assumed to depend on the gas density of the ism @xcite with @xmath235 ( @xmath236 ) as the efficiency of star formation ( free parameter ) .
@xmath237^k \label{eq : sfr}\ ] ] in the literature , the adopted value for the exponent @xmath238 varies between @xmath239 and @xmath240 ( e.g. matteucci & franois 1989 ) . in the following section
, we will show the influence of this parameter on the observational features .
the schmidt - law together with the infall of gas over a relatively long timescale guarantee a roughly continuous star formation during the evolution of the solar neighbourhood .
furthermore , the enrichment of the ism due to pn , sn ii and sn ia is considered , using supernova rates as described in section [ sec : general ] .
the parameter @xmath46 in equations [ eq : snii ] and [ eq : snia ] is a free parameter .
it is calibrated on the current supernova rates in our galaxy .
as shown in the previous sections , especially the yields of type ii sne are affected by many uncertainties .
hence , we treat the sn ii - yields as a parameter in the sense that we consider the different sn ii nucleosynthesis prescriptions presented in section [ sec : yields ] ( ww95 and tnh96 ) .
tnh96-yields consist of the given yield from the evolution of the helium core plus the initial abundance of the element in the envelope ( see also section [ sec : yields ] ) . in the simulations of the chemical evolution ,
the initial element abundances of the envelopes correspond to the element abundances in the ism when the star forms . in these terms , the tnh96-yields become metallicity dependent , although the evolution of the helium core is only calculated for solar element abundances . the basic equations of chemical evolution are explained in section [ sec : general ] .
since sn ia explode delayed with respect to sn ii @xcite , the element abundances in metal - poor stars are determined mainly by sn ii .
hence , the adopted standard model for the chemical evolution in the solar neighbourhood can easily explain the enhancement of @xmath2-elements in metal - poor stars , assuming the [ mg / fe]-ratios given by sn ii nucleosynthesis are high enough .
there are several observational features in the solar neighbourhood basically constraining different parameters . in the subsections below
, we will discuss in detail the influence of the parameters on the abundance distribution function ( adf ) , the age - metallicity relation ( amr ) , the current supernova rates , and the element abundances in the sun .
the parameters have to be adjusted to provide the best possible _ simultaneous _ fit to the existing observational data . in table
[ tab : parameters ] we summarize how the various parameters can be constrained by the different observational features .
the right column of the table gives the final adopted values .
lll parameter & observational constraint & adopted value + + stellar yields & element abundances of the sun & tnh96 + imf slope @xmath49 & solar abundance ratios & 1.36 + close binary fraction @xmath46 & relative frequency of type ii and ia sne & 0.035 + & amr & + star formation efficiency @xmath235 & current fraction of gaseous mass & @xmath241 + schmidt exponent @xmath238 & adf & 2 + accretion timescale @xmath234 & current infall rate & 4 gyr + & adf & + the calculations are performed using the stellar yields of tnh96 .
additional computations for ww95-yields under the same conditions are made in order to work out the influence of stellar nucleosynthesis .
the galactic age is assumed to be @xmath242 gyr @xcite , the age of the sun is 4.5 gyr .
the value of the surface density in the solar neighbourhood is assumed to be @xmath243 @xcite .
stellar lifetimes are taken from schaller et al . .
the differential adf gives the number of stars that are born per unit metallicity as a function of metallicity .
pagel & patchett derived this relation for the solar vicinity ( @xmath244 pc ) with a sample of 132 g - dwarfs .
the most important feature of the adf is the paucity of extremely metal - poor stars . assuming a closed box for the chemical evolution , the so called _ simple model _ predicts too many metal - poor stars @xcite .
this deviation is known as the g - dwarf problem .
the considerations of pre - enrichment @xcite or infall of material @xcite help to avoid the formation of low - metallicity stars .
the latter possibility is used in the adopted model for the solar neighbourhood , assuming the disk to form due to accretion of primordial gas ( see equation [ eq : infall ] ) .
the shape of the resulting theoretical adf depends basically on dynamical parameters like the accretion timescale @xmath234 and the schmidt - law exponent @xmath238 @xcite .
figure [ fig : adf ] shows the results for different choices of the parameter @xmath238 . to garantuee that in all computations the same total number of stars is formed , the star formation efficiency @xmath235 ( see equation [ eq : sfr ] ) is reduced for smaller @xmath238 .
the diagram demonstrates the following : 1 . the inclusion of infall solves the g - dwarf problem in the sense that the extremely high amount of metal - poor stars as predicted by the cosed box model ( long - dashed line ) is significantly decreased .
the general shape of the adf , the peak at intermediate metallicities , is reproduced by the model .
the smaller the exponent @xmath238 is , the more stars of high and low metallicity are formed .
since the adf predicted by the model is already too flat , @xmath240 may be the best choice
. a better fit to the adf - data requires an improvement of the adopted model .
since there are both , too many metal - poor and metal - rich stars , a different description of the infall - term may be necessary .
in addition , the consideration of pre - enrichment of the infalling gas further reduces the number of low metallicity stars .
since the aim of this work is to inspect the influence of different stellar yields on the chemical evolution in the solar vicinity in the framework of the standard infall - model model , we simply use the adf to constrain the model parameters without improving the model to obtain better fits .
figure [ fig : adf_tau ] shows the adf for different accretion timescales . for @xmath245 gyr and @xmath246 gyr the number of metal - poor and metal - rich stars is overestimated , respectively .
thus , we use @xmath247 gyr in our simulations .
the age metallicity relation ( amr ) shows the ratio [ fe / h ] indicating the metallicity as a function of the ages of the stars @xcite . since different element abundances of the ism at different times
are locked in stars of different ages , this corresponds to the evolution of [ fe / h ] in the ism as a function of time . in the first 2 gyr of the evolution when the sfr is at its maximum
, [ fe / h ] rises very steeply to a value of @xmath248 dex .
the increase flattens out significantly and emerges to solar metallicity at @xmath249 gyr .
figure [ fig : amr ] shows , that this behaviour is well reproduced by the simulations .
star formation ( dotted line ) is occurring over the whole range of 15 gyr with a peak of @xmath250 at @xmath251 gyr .
fitting an exponential law like @xmath252 to the range @xmath253 gyr where the sfr is decreasing leads to @xmath254 gyr . the iron abundance at large @xmath22 is significantly determined by the contribution of type ia sne .
the simulations excluding enrichment by sn ia clearly underestimate the production of iron .
, since the abundance _ ratio _ of iron to oxygen is underestimated without sn ia . ] furthermore , the fraction of iron contributed by the different types of sne depends on the adopted stellar yields .
ww95 models ( dashed lines ) produce @xmath255 dex more iron than tnh96 ( see section [ sec : ssp ] ) .
using tnh96 yields , 60 per cent of the produced iron comes from type ia sne , according to ww95 models this amount decreases to 50 per cent .
because of the higher iron yield , ww95 models fit the amr - relation worse , but are still within the error bars .
the total amount of iron in the ism highly depends on the fraction @xmath46 of close binary systems .
the amr could be slightly better fitted for a reduced iron production , thus for a lower parameter @xmath46 .
however , this parameter is additionally constrained by the relative frequency of the different types of supernovae
. unfortunately , the current rates of both types of sne ( ia and ii ) in spiral galaxies and in the solar neighbourhood are still uncertain @xcite .
since there is no consensus , the range allowed by observations is fairly large .
the theoretical relative frequency of sn ii and sn ia is mainly determined by the parameter @xmath46 . in their review paper , van den bergh and tammann claim @xmath256 for sab sb galaxies and @xmath257 for sc scd galaxies . since our galaxy
is assumed to have a hubble type between sb and sc @xcite , a relative frequency of @xmath258 seems to be a reasonable estimate .
this value is in agreement with our calculations .
while the relative frequency of sn ii and sn ia basically constrains the parameter @xmath46 , the absolute number of type ii sne occuring today depends on the parameters @xmath235 , @xmath238 , and @xmath234 , whose values are already chosen to fit the adf . assuming that sne ii occur in stars above @xmath73 , tammann estimates a surface density of @xmath259 in the solar neighbourhood from historical data .
however , because of the small size of the sample , this value is quite uncertain .
indeed , van den bergh claims that the historical data may overestimate the absolute number of sn ii significantly .
thus , the calculated value of @xmath260 is still acceptable . in section [ sec :
general ] , we showed how to constrain the sfr ( parameters @xmath235 and @xmath238 ) , the imf slope ( @xmath49 ) and the stellar yield from observational data of the accretion rate , the current gas fraction and the element abundances .
the timescale for disk formation @xmath234 is constrained by the accretion rate which is observed today ( see caption of table [ tab : constraints ] ) estimated from observations of high velocity h@xmath263i clouds ( see timmes et al .
1995 and references therein ) allows @xmath264 gyr .
the more specific value of 4 gyr is constrained by the adf . ] .
having fixed @xmath228 and @xmath58 , the current fraction of gaseous mass constrains the mean sfr ( @xmath265 ) , depending on the imf slope .
we showed that , considering an element whose yield is relatively certain , the calculated solar abundance of this element depends on @xmath52 and @xmath49 .
thus , @xmath234 , @xmath235 , @xmath238 , and @xmath49 are fixed .
figure [ oxy_z ] shows that ww95 and tnh96 differ only slightly in the calculated oxygen yield .
furthermore , oxygen is mainly produced in massive stars of small stellar lifetimes @xmath21 , thus the neglection of @xmath21 in the arguments in section [ sec : general ] is valid .
hence , we assume this yield to be the most certain and use oxygen to pin down the imf slope .
having done this , we can analyse if the stellar yields of various elements are in agreement with observations .
table [ tab : constraints ] shows the comparison between the calculated quantities and their observational constraints .
llll & tnh96 & ww95 & observation + + @xmath266 & 0.13 & 0.13 & @xmath267 + @xmath268 & 0.46 & 0.46 & @xmath269 + + solar @xmath6 & @xmath270 & @xmath271 & @xmath272 + solar @xmath273 & @xmath274 & @xmath275 & @xmath276 + solar @xmath277 & @xmath278 & @xmath279 & @xmath280 + solar @xmath91 & @xmath281 & @xmath282 & @xmath283 + solar @xmath76 & @xmath284 & @xmath285 & @xmath286 + the element abundances of @xmath287 and @xmath6 are best reproduced ( due to the above strategy ) : ww95 and tnh96 differ only marginally , the deviations from observational data are between 1 and 4 per cent . however , in the case of magnesium , the situation is different : the calculated @xmath91-abundance deviates from observational data by 7 per cent ( tnh96 ) and 29 per cent ( ww95 ) .
reproducing the solar oxygen abundance , the calculated magnesium abundance is too low , especially with ww95 nucleosynthesis .
hence , the predicted ratio between the two element abundances is not in agreement with observations . since sn ii is the main contributor to the mg - enrichment
, we can conclude that the magnesium yield of type ii sne is clearly underestimated by ww95 .
we now turn to consider the element abundances observed in stars of various metallicities in the solar neighbourhood , being the last important observational constraint on theoretical models .
gratton & sneden and magain determined [ mg / fe ] in metal - poor halo stars , edvardsson et al . determined [ mg / fe ] in disk stars with [ fe / h]@xmath288 .
these data together with the theoretical predictions from the model using the parameters of table [ tab : parameters ] are plotted in figure [ fig : mgfe_fe ] .
the scatter of the data for [ fe / h]@xmath289 is extremely large .
a reasonable average in the range @xmath4[fe / h]@xmath5 seems to be @xmath290[mg / fe]@xmath291 dex . while tnh96-[mg
/ fe]-yields are high enough to fit this value , ww95 fail to reproduce such large values .
the same preliminary conclusion was already made in section [ sec : ssp ] .
timmes et al
. also realized , that the produced [ mg / fe]-ratio in ww95 is too low to explain the data , and suggested a reduction of the sn ii iron yield by a factor of 2 .
on the other hand , as discussed in section [ sec : ssp ] , the iron yields of stellar masses smaller than @xmath98 are in good agreement with the observational data from sn1987a and sn1991j .
thus , it is reasonable to halve the iron yield of stellar masses _ greater _ than @xmath98 .
we showed that this results in a shift of 0.08 dex to higher [ mg / fe]-values , which is not enough to explain a [ mg / fe ] overabundance of @xmath3 dex . since
a flatter imf would result in an overestimation of the solar metallicity and oxygen abundance , the observed trend can only be produced with an increased mg - yield .
since it is the mg / o ratio which is underestimated , a variation of @xmath58 is not suitable to solve the problem , either .
timmes et al .
claim that a small contribution of type ia sne or intermediate- and low - mass stars to the magnesium enrichment could solve the problem without increasing the sn ii - yields . but both alternatives seem to be unlikely for the following reasons : 1 .
low - mass stars ( @xmath292 ) form co - wd and therefore do not burn carbon to magnesium @xcite .
intermediate - mass stars ( @xmath184 ) may even produce a lower [ mg / fe]-ratio than high - mass stars , because the ratio decreases with decreasing mass ( see figure [ cont_mgfe ] ) .
if this trend can be approximately extrapolated to lower masses , those stars do not increase the value of [ mg / fe ] in the ism .
3 . sn ia may be a candidate for a higher magnesium production .
on the other hand , the [ mg / fe]-ratio is underestimated in a regime at low metallicities , where type ii sn products dominate and type ia sne do not play any role .
the observational data points in figure [ fig : mgfe_fe ] show two different slopes at different metallicity ranges .
the progression of [ mg / fe ] is very flat in the low [ fe / h]-region and only slightly decreasing with increasing metallicity .
this belongs to the regime in the first 70 myr ( see scale at the upper x - axis ) , where type ii supernovae are dominating the enrichment of the ism . for [ fe / h]@xmath293 , type ia supernovae enter the game and drive down the [ mg / fe]-ratio because of their iron dominated ejecta .
the decrease of [ mg / fe ] with increasing [ fe / h ] becomes notedly steeper .
although the theoretical curves reflect this behaviour roughly , using tnh96 yields the slope at low metallicities is still too steep , especially for @xmath4[fe / h]@xmath294 .
this decrease becomes flatter for the smaller schmidt exponent @xmath239 as indicated by the dashed - dotted line .
however , while the choice of @xmath239 may improve the agreement with the data in the [ mg / fe]-[fe / h]-diagram , the adf is worse reproduced ( see figure [ fig : adf ] ) .
the offset to the data of @xmath295 dex seems to have its origin in a too low magnesium yield . since
the model parameters are chosen to reproduce the solar oxygen abundance , it is interesting to consider the [ o / fe ] ratio as a function of [ fe / h ] , too .
again we compare the theoretical results with observations by edvardsson et al .
( 1993 ) at high metallicities and gratton & ortolani ( 1986 ) in the low metallicity regime ( figure [ fig : oxfe_fe ] ) .
these data points are very few and show a large scatter .
thus we concentrate on the discussion of the edvardsson et al . data .
these are well fitted by the model using tnh96 nucleosynthesis , hence the oxygen abundance in the solar neighbourhood is reproduced as well .
the calculation with a lower schmidt - exponent @xmath238 ( tnh96 yields , dashed - dotted line ) clearly fails to match the observed [ o / fe ] .
although ww95-models suffice to produce a solar [ o / fe ] ratio , they give a bad fit to the data for @xmath296}<0 $ ] . as already mentioned , we additionally performed calculations with @xmath67 and tnh96 nucleosynthesis .
the fitting parameters had to be re - adjusted , the new values and the results for the calculated solar abundances are given in table [ tab : const_cutoff ] .
again , the parameters were chosen to match the observational constraints simultaneously .
llll & @xmath66 & @xmath65 & observation + + @xmath266 & 0.13 & 0.13 & @xmath267 + @xmath268 & 0.46 & 0.46 & @xmath269 + + solar @xmath6 & @xmath270 & @xmath297 & @xmath272 + solar @xmath273 & @xmath274 & @xmath298 & @xmath276 + solar @xmath277 & @xmath278 & @xmath299 & @xmath280 + solar @xmath91 & @xmath281 & @xmath300 & @xmath283 + solar @xmath76 & @xmath284 & @xmath301 & @xmath286 + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the solar mg abundance is even worse reproduced , since the ratio of magnesium to oxygen decreases with the inclusion of @xmath65-stars ( see also table [ tab : delta ] ) .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ a further remarkable effect is the stronger influence of type ia sne on the enrichment of iron ( parameter @xmath46 ) , because iron is the only element ( of the considered ones ) that is not additionally ejected by extremely high mass stars .
the ratio @xmath302 is still within the range allowed by observation , it may be even a better fit to the historical data ( see discussion above ) .
the effect on the [ mg / fe]-[fe / h]-diagram is shown by the long - dashed curve in figure [ fig : mgfe_fe ] .
the ratio [ mg / fe ] is higher by @xmath295 dex in the low metallicity regime , whereas it is still too small at higher metallicities ( @xmath303}\leq 0.3 $ ] ) .
the long - dashed line in figure [ fig : oxfe_fe ] shows that the oxygen abundance in the solar neighbourhood can still be reproduced .
the other observational constraints are also matched .
the differential adf changes as if there was a kind of pre - enrichment : it increases more rapidly at the lower @xmath6 .
however already at @xmath304 it presents too many objects , with respect to the observations .
the upper x - axis in figure [ fig : mgfe_fe ] shows that the steep decrease of the tnh96-curve at low metallicities comes from the short timescales in this regime . on the other hand , this is a consequence from the i m a , assuming that the stellar ejecta mix immediately with the ism .
although there is no doubt that this assumption is not realistic @xcite , most chemical evolution calculations hold this approximation . the validity of the i m a depends on the timescale of the mixing process .
malinie et al .
( 1993 ) claim that due to chemical inhomogeneities in the disk , re - mixing and star formation may be delayed by @xmath305 yr .
we will include the consideration of delayed mixing in our calculations and inspect the influence of different mixing timescales on the observational constraints discussed above .
we distinguish between two different phases of the gas component : the _ active _ and the _ inactive _ phase .
the _ inactive _ gas consists of the enriched stellar ejecta .
since this component is hot and not homogeneously distributed , stars can not form out of this phase .
the _ active _ phase , instead , is assumed to be cool and well mixed , hence , star formation is possible only in the _ active _ gas phase . in order to keep the circle of star formation and chemical enrichment alive ,
the _ inactive _ phase converts to the _ active _ star forming phase on a certain timescale , which includes both the _ cooling _ and the _ mixing _ process . the timescale is treated as a free parameter in the simulations . to include this scenario in the calculations , we modify the equations [ eq : gas ] and [ eq : abundance ] presented in section [ sec : general ] : @xmath306 to keep the equations as simple as possible , we assume the mass flow between the two gas phases to be proportional to the total amount of _ inactive _ gas divided by the mixing timescale . we now have to distinguish between the abundance in the _ active _ and the abundance in the _ inactive _ gas phase . @xmath307 @xmath308 @xmath309 the sfr described by the schmidt - law is then dependent on the density of the _ active _ gas : @xmath310^k \label{eq : sfr_mix}\ ] ] the infalling material is assumed to mix instantaneously with the _ active _ gas .
we now show the influence of the different mixing timescales on the observational constraints discussed in the previous subsections .
the values of the parameters in table [ tab : parameters ] are not changed .
since the delay due to the mixing processes elongates the timescales ( see upper x - axis for the case @xmath311 gyr ) , the curve becomes flatter .
the effect is maximum at early epochs and becomes negligible at solar ages .
while the fit to the data in the [ mg / fe]-[fe / h ] diagram has become better , the constraint on the amr relation is still fulfilled .
even the results for mixing timescales of the order @xmath312 gyr are still in agreement with observations .
table [ tab : const_delay ] gives the element abundances in both gas phases for different mixing timescales .
since the sun forms out of _ active _ gas at @xmath313 gyr , these abundances have to match the solar values given by observation ( table [ tab : constraints ] ) .
the abundances in the _ inactive _ gas are systematically higher .
the numbers show that the abundances of the elements h , o , and @xmath6 are well reproduced for the same set of parameters given in table [ tab : parameters ] , especially for @xmath314 gyr llll @xmath315 & 0.01 gyr & 0.1 gyr & 1 gyr + + @xmath316 & @xmath317 & @xmath317 & @xmath318 + @xmath319 & @xmath320 & @xmath321 & @xmath322 + + @xmath323 & @xmath272 & @xmath324 & @xmath325 + @xmath326 & @xmath327 & @xmath328 & @xmath329 + h@xmath330 & @xmath331 & @xmath132 & @xmath332 + h@xmath333 & @xmath334 & @xmath334 & @xmath335 + o@xmath330 & @xmath336 & @xmath337 & @xmath338 + o@xmath333 & @xmath339 & @xmath340 & @xmath341 + mg@xmath330 & @xmath342 & @xmath343 & @xmath344 + mg@xmath333 & @xmath345 & @xmath346 & @xmath285 + fe@xmath330 & @xmath347 & @xmath348 & @xmath349 + fe@xmath333 & @xmath350 & @xmath350 & @xmath351 + since for a larger delay in the mixing , at the end less gas is formed into stars , the fractions of both gas phases increase .
however , more metal - poor stars are formed in the beginning . as a consequence , the element abundances at @xmath352 ( i.e. when
the sun is born ) become higher with increasing @xmath315 . at @xmath353 , instead , a larger delay causes lower abundances in the ism . for the case of the iron abundance
, this pattern is demonstrated in figure [ fig : amr_mix ] .
the formation of more low - metallicity stars , though , has consequences for the derived adf . the discussion in section [ sec : adf ] showed that the adopted infall - model can not fit the adf in the whole metallicity range . at both ends of low and high metallicity , too many stars are formed .
certainly , the inclusion of delayed mixing worsens the situation . figure [ fig : adf_mix ] shows the results for the different mixing timescales .
there is no doubt , that the consideration of delayed mixing processes in the disk gives a more realistic approach to the chemical evolution in the solar neighbourhood @xcite . since the formation of more metal - poor stars
can hardly be avoided with a delayed mixing , the additional consideration of pre - enrichment of the disk due to early halo evolution ( burkert , truran , & hensler 1992 ) is necessary to solve the g - dwarf problem .
however , to treat this scenario properly , more sophisticated evolution models , calculating halo and disk evolution
seperately , have to be considered .
using two different sets of models for sn ii yields ( ww95 and tnh96 ) , we analysed the influence of stellar nucleosynthesis on the chemical evolution of galaxies , in particular the element abundances in the solar neighbourhood .
it turns out that there is a good agreement in the sn ii yields of oxygen and total metallicity between ww95 and tnh96 over the whole mass range of sn ii .
however , from the point of view of galactic chemical evolution , there are significant differences in the magnesium yields in the mass range @xmath89 . for a @xmath98-star ,
the mg - yield calculated by tnh96 is @xmath354 times higher than the result of ww95 .
we showed that , since the imf is giving more weight to smaller masses , the results of chemical evolution models are very sensitive to this discrepancy .
the iron yield , instead , is mainly uncertain in the upper mass range . ww95 and tnh96 agree very well in the lower mass range ( @xmath355 ) which is well constrained by the observed lightcurves of sn ii events ( sn1987a , sn1991j ) .
however , in high mass stars with @xmath168 , ww95 models give significantly higher fe - yields than tnh96 . in total , this leads to lower [ mg / fe]-ratios produced by ww95-nucleosynthesis .
a significantly super - solar value is only reached in high mass stars ( figure [ cont_mgfe ] ) which are dominating the enrichment in the first few @xmath356 yr of chemical evolution .
only 0.04 gyr after the birth of the first stars , the complete generation of sn ii - exploding stars in the mass range @xmath357 is enriching the ism .
we calculated the sn ii ssp - yields of o , mg , and fe for different imf slopes and both nucleosynthesis prescriptions .
the result is that tnh96-nucleosynthesis leads to @xmath358}=0.26 $ ] dex for salpeter - imf , while the ratio with ww95 ( @xmath86 , model b ) is 0.05 dex .
we showed that this discrepancy is due to a lower mg - ssp - yield of 0.13 dex and a higher fe - ssp - yield of 0.08 dex in ww95 .
even for a flat imf with @xmath201 , the ssp - value of [ mg / fe ] is 0.12 dex with ww95-yields . without any impact from complex evolution models , from these numbers one can already conclude , that the [ mg / fe ] overabundance in both , ellipticals and the solar neighbourhood , can not be explained with the stellar yields of the ww95-models . applying
the standard infall - model @xcite on the chemical evolution of the solar neighbourhood confirms the conclusions drawn from the discussion of the ssp yields .
both , the [ mg / fe ] overabundance in metal - poor stars and the magnesium abundance of the sun can be better reproduced with the mg - yields of tnh96 . in addition to this , we discussed the relaxation of the instantaneous _ mixing _ approximation for the chemical evolution in the solar neighbourhood .
for this purpose , we modified the basic equations of chemical evolution seperating the gaseous component into two different gas phases . while the _
inactive _ phase is enriched by the stellar ejecta , stars can only form out of the _ active _ , well mixed phase . a mass flow from the _ inactive _ to the _ active _ gas phase on a variable timescale represents the mixing process . for different mixing timescales of the order @xmath359 yr , we investigated the influence of a delayed mixing on the reproduction of the observational features .
it turns out that a delay in the mixing supports the approximately constant value of [ mg / fe ] in the [ mg / fe]-[fe / h]-diagram in the low - metallicity range , while the agreement with the age - metallicity relation ( amr ) and solar element abundances is not violated .
however , since a delayed mixing causes the formation of more low - metallicity stars , the abundance distribution function ( adf ) is less well reproduced . on the other hand ,
the instantaneous mixing of the stellar ejecta is certainly an unrealistic assumption , and the inclusion of a delay is a necessary step to improve chemical evolution models .
hence , the solution of the g - dwarf problem in the solar neighbourhood may require a combination of infall and pre - enrichment .
since tnh96 include the @xmath65-star in their computations , we additionally investigated the influence of a variation of the upper mass cutoff on the theoretical ssp - yields and on the chemical evolution in the solar neighbourhood .
applying tnh96 nucleosynthesis , the [ mg / fe ] ratio in the ejecta of one ssp is significantly increased .
this result is highly uncertain , however , because tnh96 do not consider fall back , which may play an important role for the nucleosynthetic contribution from high mass stars . indeed , extrapolating the results of ww95 to @xmath65 leaves the ssp yields basically unchanged .
the problem of the underestimation of the solar magnesium abundance remains the same also for tnh96-yields , since the mg / o ratio in stars more massive than @xmath66 is even smaller .
however , it is important to investigate quantitatively the metal contribution of stars more massive than @xmath66 , since they could play an important role for chemical evolution . in general
, we demonstrated the sensitivity of galactic chemical evolution on nucleosynthesis prescriptions of type ii supernovae .
different stellar yields can significantly alter conclusions on the parameters of chemical evolution models like imf slope or star formation timescales .
as long as the stellar nucleosynthesis of important elements like magnesium and iron is affected by so many uncertainties , the results from simulations of chemical evolution have to be interpreted considering the whole range of up to date nucleosynthesis calculations .
we would like to thank b. pagel , the referee of the paper , for carefully reading the first version and giving important remarks on the subject .
he especially motivated us to explore the influence of the upper mass cutoff on the results of the calculations .
we also thank f .- k .
thielemann for usefull and interesting discussions .
the sfb 375 of the dfg and the alexander von humboldt stiftung ( l.g . ) are acknowledged for support .
the figures show the relative contribution of a @xmath72 mass interval to the total sn ii yields of the elements oxygen , magnesium , iron , and metallicity for different imf slopes and nucleosynthesis prescriptions ( ww95 , tnh96 ) .
the three models of different explosion energies ( a , b , c ) . ] and four different initial metallicities ( @xmath360 ) in ww95 are considered .
each figure shows the results for one certain element and one specified ww95-model ( a , b , c ) .
the four panels of each figure show the results for the four different imf slopes .
the four different initial metallicities of ww95 and tnh96-yields are plotted together in each panel .
the meaning of the linestyles and symbols are specified in figure [ eject_hydro ] .
the quantity @xmath361 is determined according to equation [ eq : cont ] .
it is normalized such that the integration over the total mass range of sn ii ( @xmath362 ) is equal 1 .
the x - axis give the initial stellar mass on the main sequence . to obtain the fractional contribution of a mass interval
, one has to multiply the width of the interval ( @xmath75 ) with the mean value of @xmath361 in this mass range . | we analyse two recent computations of type ii supernova nucleosynthesis by woosley & weaver ( 1995 , hereafter ww95 ) and thielemann , nomoto , & hashimoto ( 1996 , hereafter tnh96 ) , focusing on the ability to reproduce the observed [ mg / fe]-ratios in various galaxy types .
we show that the yields of oxygen and total metallicity are in good agreement .
however , tnh96-models produce more magnesium in the intermediate and less iron in the upper mass range of type ii supernovae than ww95-models . to investigate the significance of these discrepancies for chemical evolution ,
we calculate _ simple stellar population_-yields for both sets of models and different imf slopes .
we conclude that the mg - yields of ww95 do not suffice to explain the [ mg / fe ] overabundance neither in giant elliptical galaxies and bulges nor in metal - poor stars in the solar neighbourhood and the galactic halo . calculating the chemical evolution in the solar neighbourhood according to the standard infall - model ( e.g. matteucci & greggio 1986 ; timmes , woosley , & weaver 1995 ; yoshii , tsujimoto , & nomoto 1995 )
, we find that using ww95 and tnh96 nucleosynthesis , the solar magnesium abundance is underestimated by 29 and 7 per cent , respectively .
we include the relaxation of the instantaneous _ mixing _ approximation in chemical evolution models by splitting the gas component into two different phases . in additional simulations of the chemical evolution in the solar neighbourhood , we discuss various timescales for the mixing of the stellar ejecta with the interstellar medium .
we find that a delay of the order of @xmath0 years leads to a better fit of the observational data in the [ mg / fe]-[fe / h ] diagram without destroying the agreement with solar element abundances and the age - metallicity relation .
stars : stellar yields stars : type ia and type ii supernovae stars : simple stellar population galaxies : element abundances galaxies : solar neighbourhood chemical evolution : instantaneous mixing approximation |
the evidence that supermassive black - holes are ubiquitous in nearby inactive galaxies is now extremely strong ( eg . van der marel 1999 , kormendy & richstone 1995 ) .
furthermore , several studies have concluded that the masses of these dormant black - holes are directly proportional to the mass of the galaxy bulge ( spheroidal ) component ( eg . kormendy & richstone 1995 , magorrian et al .
although a significant scatter is present in the correlation ( @xmath5 dex at a given bulge mass ) , the data appear to suggest that @xmath6 .
while studies of nearby inactive galaxies have seen a reasonably coherent picture emerge , the situation with regard to active galaxies is less certain .
although the black - hole mass relation observed in nearby inactive galaxies can be comfortably reproduced by combined galaxy+quasar formation models ( eg . wilman , fabian & nulsen 2000 , kauffmann & haehnelt 2000 ) , observational studies to investigate the form of the @xmath1 relation in active galaxies have led to seemingly contradictory results . in a study of the @xmath1 relation in quasars , laor ( 1998 ) used the fwhm of the broad @xmath0 emission line , together with a calibration between the broad - line region ( blr ) radius and quasar uv luminosity , to obtain virial black - hole mass estimates for 14 pg quasars . using estimates of the host - galaxy bulge luminosities from the hst study of bahcall et al .
( 1997 ) , laor found a @xmath7 correlation which agreed reasonably well with the results of magorrian et al .
in contrast , wandel ( 1999 ) found that virial black - hole mass estimates for a sample of 17 seyfert1 galaxies , based on the @xmath0 fwhm and reverberation mapping measurements of the blr radius , were substantially smaller than would be expected from the @xmath1 relation seen in nearby inactive galaxies . using estimates of the seyfert galaxy bulge luminosities from the empirical morphology - based formula of simien & de vaucouleurs ( 1986 ) ,
wandel found that @xmath8 , a factor of between ten and twenty lower than found for both nearby inactive galaxies ( magorrian et al .
1998 ) and for luminous quasars ( laor 1998 ) .
the implication of the wandel ( 1999 ) result is that there exists a difference in the formation or triggering mechanism at work in seyfert galaxies and quasars .
this possibility is explored by wang , biermann & wandel ( 2000 ) who present a model in which @xmath9 , where @xmath10 is the velocity dispersion of the accreting gas . in this model
the high values of @xmath11 found in quasars and nearby bulges are the asymptotic , high - velocity case , possibly produced by a violent merger event . in contrast , the much lower ratios found in seyfert galaxies are the result of the accretion of lower velocity gas triggered by tidal disruption . in this paper the @xmath1 relation is examined over a wide range in agn luminosity by combining new results for a sample of 30 luminous ( m@xmath12 ) quasars with a re - analysis of the bulge masses attributed to the wandel ( 1999 ) seyfert galaxy sample .
first , in section 2 we briefly review the complications and potential uncertainties in deriving black - hole mass estimates from emission - line widths , and explore how best to model the geometry of the broad - line region for the purpose of black - hole estimation in agn .
next , in section 3 we describe the sample of objects selected for study , present the new spectroscopic observations , and explain how reliable bulge luminosities have been extracted from existing hst images . in section
[ main ] three important questions are addressed .
firstly , do powerful agn actually display a @xmath1 correlation consistent with that found in nearby inactive galaxies ?
secondly , do the apparent differences between seyfert galaxies and quasars actually stem from fundamental differences between the two agn classes , or simply from systematic errors in estimating seyfert bulge luminosities ? and finally , do differences in black - hole mass and gas accretion rates play a crucial role in determining the radio properties of agn ? in section [ fp ] stellar velocity estimates for a sub - sample of the quasars are used to compare our virial @xmath0 black - hole mass estimates with those of the @xmath2 relations discovered by gebhardt et al .
( 2000a ) and merritt & ferrarese ( 2000 ) .
finally , our main conclusions are summarized in section 6 . unless otherwise specified all cosmological calculations performed in this paper assume @xmath13 kms@xmath14 mpc@xmath14 , q@xmath15 and @xmath16 .
the basic theory of obtaining virial black - hole mass estimates from quasar emission lines has been discussed extensively in the literature ( eg . wandel , peterson & malkan 1999 ) and consequently only a brief outline is provided here .
the underlying assumption is that the broad emission lines are produced by material which is gravitationally bound and orbiting with keplerian velocities . if this assumption is applicable then an estimate of the central mass is given by @xmath17 , where @xmath18 is the radius of the blr and v is the velocity of the line - emitting material .
there are two standard methods of estimating the radius of the broad - line region .
the first of these , and presumably the more accurate , is to use the time delay between continuum and line variations . for the sample of seyfert galaxies studied by wandel , peterson & malkan ( 1999 ) and wandel ( 1999 ) ,
long - term monitoring has provided such reverberation mapping estimates of the radius of the broad - line region . in the absence of reverberation mapping data for the quasar sample
it is necessary to use a more indirect estimate of the broad - line radius .
the method used here is the correlation between @xmath18 and monochromatic luminosity at @xmath19 found by kaspi et al .
( 2000 ) , from their reverberation mapping results for 17 pg quasars . when converted to the cosmology adopted in this paper , the kaspi et al .
correlation becomes : @xmath20 the other necessary component for the virial mass estimate is a measure of velocity of the line emitting material . to obtain this , wandel ( 1999 )
adopted the assumption of random orbits , in which case v @xmath21 fwhm .
however , the assumption of random orbits seems unrealistic for quasars . in particular , for radio - loud quasars where the radio core : lobe ratio provides an independent measure of orientation , there exists strong evidence that the velocity - field of the broad - line region is better represented by a combination of a random isotropic component , with characteristic velocity @xmath22 , and a component only in the plane of the disk , with characteristic velocity @xmath23 ( wills & browne 1986 ) . in this case
, the observed fwhm will be given by @xmath24 where @xmath25 is the angle between the disc normal , and the line of sight to the observer .
in the case of radio - quiet objects we have no guide as to what precise value of @xmath25 should be adopted for a particular object .
however , attempts to unify radio - loud quasars and galaxies ( e.g. barthel 1989 ) , and imaging polarimetry studies of radio - quiet agn ( e.g. antonucci & miller 1985 ) support a picture in which an active nucleus is likely to appear to the observer as a quasar provided @xmath26 . in order to test the viability of this disc - model we have attempted to reproduce the form of the cumulative fwhm distribution of the combined quasar+seyfert sample . as can be seen from fig 1b . , the cumulative distribution of the data is convex in shape .
this form of distribution can not in fact be produced by a single disc model in which the angle between the disc normal and observer s line of sight is assumed to be randomly distributed between @xmath27 and some adopted upper limit @xmath28 .
the reason for this is simply that , given a random distribution of theta ( and @xmath23 significantly larger than @xmath22 ) the cumulative distribution of fwhm grows proportional to solid angle @xmath29 .
however , we have found that the form of the observed distribution can in fact be almost perfectly reproduced by a model which includes only one additional free parameter .
specifically we have relaxed the assumption of a purely random distribution of theta for @xmath30 by exploring whether the data can be described by a combination of two randomly - oriented disc - model populations with the same characteristic orbital velocity but different angular constraints .
the best - fitting model parameters were found to be @xmath31 km@xmath32 , @xmath33 , @xmath34 , with the two models contributing @xmath35 and @xmath36 of the combined model distribution respectively . in both cases
the model agn are assumed to lie randomly at angles between @xmath37 to the line of sight .
although two - component disc - models with non - zero random velocity components and independent orbital velocity components were considered , it was found that the introduction of these extra free parameters did not produce significantly improved fits .
the best - fitting model distribution is shown as the thin line in fig 1b . and can be seen to be an excellent representation of the data , with an application of the kolmogorov - smirnov test showing the two distributions to be indistinguishable ( @xmath38 ) . by substituting the fitted model parameters into eqn 2 .
it is now possible to determine the average relationship between observed fwhm and actual orbital velocity @xmath23 by calculating @xmath39 as weighted by solid angle .
we therefore find that on average @xmath40 fwhm , and it is this relation which is adopted throughout the remainder of this paper . as can be seen from the best - fitting parameters presented above , the dominant population of objects in this two - component disc model of the blr is constrained to lie within @xmath41 to the line of sight , as expected in the unified model ( eg .
barthel 1989 ) .
in contrast , the second population of objects is required to lie within a smaller range of angles to the line of sight than is normally assumed in the unified model ( ie.@xmath42 ) .
however , one can argue that this is entirely reasonable given that the present sample is largely dominated by powerful , optically - selected broad - line agn .
however , it is worth noting the implications of the assumptions in this simple model , and the effect of relaxing them . due to the fact that the virial black - hole mass estimate is proportional to @xmath43
, the adoption of the standard model in which blr velocities are purely random ( @xmath44fwhm ) leads to black hole mass estimates three times _ smaller _ than derived in this paper .
further evidence in support of the appropriateness of the blr model adopted here is discussed in section 4.1 and section 5 .
the sample of 30 luminous quasars ( m@xmath12 ) studied in this paper consists of two optically - matched sub - samples of 17 radio - quiet quasars and 13 radio - loud quasars .
the unique feature of this sample is that all of its members have accurate bulge luminosities available from detailed two - dimensional modelling of hst images .
nineteen of the objects are taken from the sample studied by mclure et al .
( 1999 ) and dunlop et al .
( 2000 ) , with a further eleven taken from the sample of bahcall et al .
this is therefore the best dataset currently available for this type of investigation .
the sample of seyfert galaxies consists of fifteen of the seventeen objects studied by wandel ( 1999 ) .
the two objects which are not included , akn 120 & mrk 110 , have been excluded on the basis that no suitable hst archive data exists for these objects , and that a reliable disc / bulge decomposition could not be identified in the literature .
the new observations presented in this paper consist of nuclear spectra of 13 of the quasar sample which were obtained in december 1999 on the 2.5 m isaac newton telescope on la palma . all observations used the intermediate dispersion spectrograph ( ids ) with the 235 mm camera and the eev10 @xmath45 pixel ccd as the detector .
integration times on individual objects were @xmath46 seconds , with the rv300 grating providing spectra centred on @xmath47 with a resolution of @xmath48 . at the average redshift of the quasar sample ( @xmath49 )
this ensured that the spectra were always approximately centred on the h@xmath50 line , with the wavelength coverage of @xmath51 allowing an accurate continuum determination .
the reduction of the spectra was performed in iraf in the standard way , including flux calibration , and correction for atmospheric extinction and galactic reddening .
the final wavelength calibrated spectra were interpolated onto a linear dispersion of 1.5 / pix and are shown in fig [ spectra ] .
the h@xmath50 fwhm measurements for 16/17 of the remaining quasars in the sample are taken from the boroson & green ( 1992 ) and corbin ( 1997 ) studies .
given that the h@xmath50 fwhm enters into the virial mass estimate quadratically , it is obviously important that there is no large systematic offset between the two sets of spectra . fortunately , five of the quasars for which new spectra have been obtained also feature in the boroson & green study . for these five objects the fwhm values determined from the new spectra agree well with those of boroson & green , @xmath52 , providing confidence that large variations in measured fwhm values do not introduce large systematic errors into the virial black - hole estimates . the @xmath0 fwhm measurement for one final object , 1635 + 119 ,
is taken from the spectrum of wills & browne ( 1986 ) .
a substantial effort has been invested in providing accurate estimates of the bulge luminosities for all the objects included in this study .
all 30 objects in the quasar sample have bulge luminosities derived from two - dimensional modelling of hst data from the deep host - galaxy imaging studies of dunlop et al .
( 2000 ) and bahcall et al .
the addition of the dunlop et al .
( 2000 ) host - galaxy sample means that the number of quasars with reliable bulge luminosities is a factor of two greater than was available to laor ( 1998 ) . of equal importance to this study
are revised estimates of the bulge luminosities within the seyfert galaxy sample .
the bulge luminosities used by wandel ( 1999 ) were taken from whittle ( 1992 ) and are based upon an application of the simien & de vaucouleurs ( 1986 ) formula relating bulge / disc ratio to galaxy morphology . as well as the large amount of scatter associated with this empirical formula , it is not clear that it is directly applicable to seyfert galaxies . indeed ,
if a typical early - type colour of @xmath53 is assumed , then the bulge magnitudes adopted by wandel correspond to an average of @xmath54 , half a magnitude brighter than the recent determination of @xmath55 by lin et al .
( 1996 ) . in light of this
it seems at least plausible that overestimation of the seyfert galaxy bulge luminosities may contribute to the lower @xmath11 ratios found by wandel ( 1999 ) . in order to properly quantify the extent to which such overestimates of the seyfert bulge luminosities
can explain the wandel ( 1999 ) result , it is important to obtain bulge luminosity estimates of comparable quality to those of the quasar sample .
suitable quality hst archive data is available for 9/15 of the seyfert galaxies , predominantly from the large - scale seyfert galaxy imaging study of malkan , gorjian & tam ( 1998 ) .
the images of these objects were decomposed into disc and bulge components in an identical fashion to that utilised in the quasar host study of dunlop et al .
( 2000 ) , and details of the modelling technique used can be found in mclure , dunlop & kukula ( 2000 ) .
the remaining six seyfert galaxies have bulge luminosity estimates taken from the disc / bulge decompositions of baggett , baggett & anderson ( 1998 ) and kotilainen , ward & williger ( 1993 ) .
although determined from ground - based data , these published bulge luminosities are still expected to be more accurate than a morphology based estimate .
the best estimates of the bulge luminosities for the quasar and seyfert samples are listed in tables [ big ] & [ small ] respectively . [ cols="^,^,^,^,^,^,^,^",options="header " , ]
the black - hole mass vs. bulge luminosity relation for the combined quasar and seyfert galaxy sample is plotted in fig [ mainfig ] .
the two quantities can be seen to be well correlated , with the rank - order coefficient of @xmath56 having a significance of @xmath57 . in order to obtain a reliable fit to the relation ,
a @xmath58 minimization was used which properly accounts for the uncertainties in both the estimated black - hole masses and bulge luminosities ( press et al .
the best - fitting relation ( @xmath59 for 45 d.o.f . ) is found to be : @xmath60 and is shown as the solid line in the left - hand panel of fig [ mainfig ] . in terms of black - hole mass
the scatter around the best - fitting relation of @xmath61 is large ( see fig [ scatter ] ) , although still comparable to that found by both magorrian et al .
( 1998 ) and laor ( 1998 ) .
in contrast to the findings of wandel ( 1999 ) there is no evidence from fig [ mainfig ] that the seyfert galaxy @xmath62 relation is different to that of quasars .
the removal of this discrepancy is due to the improved estimates of the seyfert galaxy bulge luminosities use here . as mentioned in section [ bulge ] ,
it was suspected _ a priori _ that the bulge luminosities adopted by wandel ( 1999 ) were probably overestimated .
this suspicion is supported by the mean bulge luminosity derived for the seyfert galaxies here , m@xmath63 , which is 1.3 magnitudes fainter than estimated by wandel ( 1999 ) .
it appears therefore that the evidence from this data - set supports a straightforward unification of seyfert galaxies and quasars in which black - hole mass scales with bulge luminosity . in this scenario
fig [ mainfig ] suggests that a black - hole mass of @xmath64 needed to power a luminous quasar requires a host galaxy with a bulge luminosity of m@xmath65 .
this restriction naturally predicts that the hosts of luminous quasars will be predominantly massive early - type galaxies , in good agreement with the findings of recent studies ( eg .
mclure et al . 1999 ,
dunlop et al .
2000 , boyce et al .
1998 ) . by combining bulge luminosity with a suitable mass - to - light ratio it is possible to test whether or not the observed @xmath7 relation is consistent with a simple linear scaling between black - hole and bulge mass .
the mass - to - light ratio adopted here is the @xmath66 relation determined by j@xmath67rgensen , franx & kj@xmath68rgaard ( 1996 ) , hereafter jfk96 , from their gunn-@xmath69 fundamental plane study . under the assumption that @xmath70 , the @xmath7 relation is expected to have the form : @xmath71 where the constant @xmath72 results from assuming an average bulge colour of @xmath73 ( fukugita et al .
1995 ) and an absolute gunn-@xmath69 magnitude for the sun of @xmath74 ( j@xmath67rgensen 1994 ) . a relation of this form , with a value of @xmath75 as determined by magorrian et al .
( 1998 ) , is shown as the solid line in the right - hand panel of fig [ mainfig ] .
it can be seen that , for a given bulge luminosity adopting the magorrian et al .
normalization predicts black - hole masses approximately 2.5 times larger than determined via the @xmath0 fwhm .
it is also noteworthy however that the slope of @xmath76 expected from a linear @xmath1 relation is not inconsistent with the best - fitting value of @xmath77 determined above .
consequently , a least - squares fit of the data was undertaken with an enforced slope of @xmath76 , and is shown as the dashed line in the right - hand panel of fig [ mainfig ] .
this can be seen to be a reasonable representation of the data , suggesting that the scaling between black - hole and bulge mass is consistent with being linear .
the least - squares fit with an enforced slope of @xmath76 has a intercept of @xmath78 which , when compared with equ [ predict ] , suggests that the constant of proportionality between black - hole and bulge mass is @xmath79 .
it is worth emphasising that this value for the constant of proportionality results , at least in part , from our use of a simple , but realistic disc - like model for the blr , incorporating the known orientation biases of optically luminous agn .
if instead one simply uses the random orbit relation @xmath80fwhm , then as discussed in section 2 , the inferred black - hole masses would be @xmath81 times smaller , yielding the relation @xmath82 .
this is the origin of the findings of ho ( 1999 ) who studied the @xmath7 relation in a sample 16 seyfert galaxies ( 12 of which are common to the sample studied here ) using published @xmath83-band photometry and virial black - hole mass estimates from reverberation mapping and @xmath0 fwhm measurements .
ho found that the reverberation mapping black - hole estimates were lower than expected from the @xmath7 relation , suggesting a value of @xmath84 .
further evidence that simple random - orbit calculations for the blr lead to systematic underestimation of black - hole mass is discussed in section 5 .
fig [ lion ] shows the correlation between bulge luminosity and continuum ionizing luminosity for the full sample , where we have adopted the relation @xmath85 which was found to be the average of a large scatter by wandel , peterson & malkan ( 1999 ) .
these two quantities are well correlated ( @xmath86 ) , in apparent agreement with the results of the combined seyfert galaxy and quasar study of mcleod & rieke ( 1995 ) .
mcleod & rieke found a relation between @xmath87-band host luminosity and @xmath83-band nuclear luminosity in the form of a lower limit , such that @xmath88(host)@xmath89(nucleus ) .
however , it is also noticeable from fig [ lion ] that the correlation between @xmath90 and @xmath91 is entirely dependent upon the low luminosity seyfert galaxies , with the quasar sample on its own displaying no significant correlation , @xmath92 .
the conclusion that the correlation between @xmath90 and @xmath91 breaks down at higher nuclear luminosities is supported by the results of a study of high - luminosity quasars by percival et al .
( 2000 ) . although including quasars up to a factor of ten more luminous than those studied here , percival et al .
find very similar host galaxy luminosities .
the suggestion is therefore that high luminosity quasars operate at higher accretion rates , rather than harbouring significantly more massive black - holes .
the correlation between @xmath90 and @xmath91 amongst the seyfert galaxies raises the possibility that the observed correlation between @xmath93 and @xmath91 could in fact be simply due to independent correlations with @xmath90 . in order to investigate this possibility
the partial spearman rank correlation test ( macklin 1982 ) has been used to quantify the correlation between @xmath93 and @xmath91 , independent of @xmath90 .
the correlation between @xmath93 and @xmath91 at constant @xmath90 has a coefficient of @xmath94 which has a significance of @xmath95 .
therefore , although the @xmath93 and @xmath91 correlation is still clearly present , its significance is substantially lowered .
unfortunately , due to the heterogeneous nature of the sample of seyfert galaxies which currently have reverberation mapping data available , it is difficult to quantify properly the significance of the @xmath7 relation . in order to overcome this problem ,
a substantially larger sample of agn is required , selected from a narrow slice in both nuclear luminosity and redshift .
the black - hole mass estimates for the 17 radio - quiet and 13 radio - loud quasars studied here provide an opportunity to determine the influence ( if any ) of black - hole mass on quasar radio luminosity .
the ionizing continuum luminosity distributions of the two quasar sub - samples are indistinguishable , with a kolmogorov - smirnov ( ks ) test returning a probability of @xmath96 that both sub - samples are drawn from the same underlying distribution .
consequently , any difference in the distribution of black - hole masses between the two sub - samples is presumably linked to the difference in radio properties . with the results of section [ linear ] showing that agn black - hole and bulge masses are consistent with being directly proportional , it is expected that the @xmath97 magnitude difference in median bulge luminosity between the quasar sub - samples should translate directly into a corresponding difference in average black - hole mass .
this is indeed the case , with the median black - hole mass of the radio - loud sub - sample a factor of three larger than that of the radio - quiet sub - sample .
a natural division between the quasar sub - samples appears to occur at @xmath98 .
only 2/13 of the radio - loud quasars have @xmath99 , while only 4/17 of the radio - quiet have @xmath100 .
this difference in black - hole mass distributions is shown to be significant at the @xmath101 ( @xmath102 ) level by a ks test .
the implication from the quasar sample is therefore that ( albeit with substantial overlap ) for a given nuclear luminosity , the probability of a source being radio - loud increases with black - hole mass , or alternately decreases with @xmath103 .
fig [ ledd ] shows a plot of estimated ionizing luminosity vs. black - hole mass , together with lines indicating the location of objects radiating at 100% , 10% and 1% of the eddington limit .
two aspects of this figure are noteworthy .
firstly , it can be seen that the vast majority of the objects appear to be radiating in the range @xmath104 of the eddington limit , with a mean value of @xmath105 .
this is despite the fact that the sample spans a range in ionizing luminosity of four orders of magnitude , and is in contrast to the results of wandel , peterson & malkan ( 1999 ) and kaspi et al .
( 2000 ) who both find that the @xmath106 ratio increases weakly with continuum luminosity .
secondly , there appears to be no evidence that the seyfert galaxies are operating in a different accretion regime to the quasars , contrary to the finding of lu & yu ( 1999 ) .
however , it should also perhaps be noted that differences between the uv spectra of seyfert galaxies and quasars could well mean that a global correction of the form @xmath107 is inappropriate .
in this section the possibility of calibrating the virial @xmath0 fwhm black - hole estimates , by taking advantage of the remarkably small scatter in the recently discovered @xmath2 correlation , is explored . considering that @xmath0 fwhm measurements can be obtained with relative ease for powerful agn , in comparison with determining the host - galaxy stellar velocity dispersion , this is a question of some interest .
although stellar velocity measurements are not available for the quasar sample , a comparison is still possible for the 19 objects which are taken from the host - galaxy study of mclure et al .
( 1999 ) and dunlop et al .
one of the main successes of this study was the first demonstration that the massive elliptical host galaxies of these quasars follow an identical kormendy relation ( @xmath109 ) to normal inactive ellipticals .
therefore , because the kormendy relation is simply a projection of the fundamental plane that elliptical galaxies occupy in the three dimensional space defined by @xmath110 , it is possible to use the @xmath111 & @xmath112 data for these objects to estimate their stellar velocity dispersions .
the fundamental plane parameters adopted here are those determined for the coma cluster by jfk96 : @xmath113 although there is evidence that the form of the fundamental plane changes with redshift , the form of this evolution ( @xmath114 ; j@xmath67rgensen et al .
1999 ) should not introduce significant errors at the average redshift of the quasar sample ( @xmath115 ) . due to the disagreement over what the correct form of the @xmath2 relation is , both the @xmath116 ( gebhardt et al .
2000a ) and @xmath117 ( merritt & ferrarese 2000 ) versions of the relation have been used to convert the velocity dispersion estimates into black - hole masses . in both cases
the @xmath2 relations predict black - hole masses which are in remarkably good agreement with our virial @xmath0 estimates , with median @xmath118 ratios of 1.04 & 1.31 for the gebhardt et al . and merritt & ferrarese relations respectively
. the clear implication of this comparison is that the use of a disc - like blr , combined with limits of disc orientation relative to the observer , has led to reliable virial @xmath0 fwhm estimates of black - hole mass .
in contrast , the standard assumption of purely random blr velocities lead to virial @xmath0 fwhm black - hole mass estimates which appear to be systematically low by a factor of @xmath119 .
a similar systematic problem with the assumption of random orbits was alluded to by gebhardt et al .
( 2000b ) when they compared ( random orbit ) virial black - hole mass estimates , based on reverberation mapping and @xmath0 fwhm measurements , with velocity dispersion black - hole masses for a sample of seven seyfert galaxies ( five of which are common to the sample studied here ) .
although gebhardt et al .
( 2000b ) concluded that the discrepancy between the two black - hole mass estimates was not significant , they did comment that an increase in the reverberation mapping estimates by a factor of @xmath120 would improve the agreement .
finally , in further support of the relation @xmath121 derived from our @xmath0 fwhm analysis , we note that gebhardt et al .
( 2000b ) also comment that new three integral models show that the black - hole masses originally determined by magorrian et al .
( 1998 ) were overestimated by a factor of @xmath122 , again leading to the conclusion that @xmath123 .
the black - hole masses for a sample of 30 optically - matched radio - loud and radio - quiet quasars have been estimated using @xmath0 fwhm and continuum luminosity measurements from new and previously published nuclear spectra .
reliable quasar bulge luminosities have been combined with new and published disc / bulge decompositions for a sample of seyfert galaxies with reverberation mapping black - hole estimates , in order to study the form of the @xmath7 relation over a large baseline in agn luminosity .
the results of the @xmath0 black - holes mass estimates have been compared with the @xmath2 correlation in an attempt to check the normalization of the @xmath1 relation .
the main conclusions of this study can be summarized as follows : 1 .
host - galaxy bulge luminosity and black - hole mass are found to be well correlated , albeit with scatter of 0.6 dex in @xmath93 . assuming a simple but realistic disc - like geometry for the blr , coupled with limits on viewing angle for optically - luminous agn , the form of the @xmath124 correlation is shown to be consistent with @xmath3 .
contrary to the results of wandel ( 1999 ) no evidence is found that seyfert galaxies follow a different @xmath7 relation to quasars .
the results presented here suggest that unification of seyfert galaxies and radio - quiet quasars via a simple scaling in black - hole mass is tenable .
a comparison of the virial black - hole estimates based on @xmath0 fwhm measurements with those predicted by the @xmath2 relation suggests that our @xmath0 black - hole estimates are , on average , accurate to within a factor @xmath125 .
these two , completely independent approaches to black - hole mass estimation both support a black - hole mass vs. bulge mass relation of the form @xmath3 .
4 . the distribution of black - hole masses in the radio - quiet and radio - loud quasar sub - samples are found to be significantly different ( @xmath102 ) , with the median radio - loud black - hole mass a factor of three larger than the equivalent radio - quiet value .
it appears that , for some reason , a radio - loud quasar requires a black - hole mass @xmath126 .
finally , perhaps the most important conclusion resulting from this study is that easily obtainable agn nuclear spectra can be be used to provide black - hole mass estimates which are accurate to within a factor of @xmath127 .
it is therefore possible that large - scale studies of agn black - hole demography can be undertaken without the need for more demanding and time - intensive stellar velocity dispersion measurements .
the int is operated on the island of la palma by the isaac newton group in the spanish observatorio del roque de los muchachos of the instituto de astrofisica de canarias .
based on observations with the nasa / esa hubble space telescope , obtained at the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . under nasa contract
nas5 - 26555 .
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 .
rjm acknowledges a pparc pdf .
antonucci r. , miller j.s . , 1985 ,
apj , 297 , 621 + baggett w.e . , baggett s.m . ,
anderson k.s.j . , 1998 ,
aj , 116 , 1626 + bahcall j.n . , kirhakos s. , saxe d.h . , schneider d.p . , 1997 ,
apj , 479 , 642 + boroson t.a . , green r.f . , 1992 ,
apjs , 80,109 + boyce p.j . , et al . , 1998 , mnras , 298 , 121 + corbin m.r . , 1997 , apjs , 113 , 245 + dunlop j.s . , et al . , 2001 , mnras , submitted , + fukugita m. , shimasaku k. , ichikawa t. , 1995 , pasp , 107 , 945 + gebhardt k. , et al
, 2000a , apj , 539 , l13 + gebhardt k. , et al . , 2000b , apj , 543 , l5 + ho , l.c . , 1999 , in chakrabarti s.k . , ed , proc observation evidence for the black holes in the universe , kluwer , dordrecht , p.157 + j@xmath67rgensen i. , 1994 , pasp , 106 , 967 + j@xmath67rgensen i. , franx m. , kj@xmath68rgaard p. , 1996 ,
mnras , 280 , 167 + j@xmath67rgensen i. , franx m. , hjorth j. , van dokkum p.g . , 1999 ,
mnras , 308 , 833 + kaspi s. , smith p.s . , netzer h. , maoz d. , jannuzi b.t . , giveon u. , 2000 , apj , 533 , 631 + kauffmann g. , haehnelt m. , 2000 , mnras , 311 , 576 + kormendy j. , richstone d. , 1995 , ara&a , 33 , 581 + kotilainen j.k . , ward m.j
. , williger , 1993 , mnras , 263 , 655 + laor a. , 1998 , apj , 505 , l83 + lin h. , kirshner p.p . ,
schectman s.a . ,
landy s.d .
, oemler a. , tucker lu y. , yu q. , apj , 1999 , 526 , l5 + d.l . ,
schechter p.l .
, 1996 , apj , 464 , 60 + macklin j.t . , 1982 , mnras , 199 , 1119 + mcleod k.k . , rieke g.h . , 1995 , apj , 441 , 96 + mclure r.j . , dunlop j.s . , kukula m.j.,baum s.a . , odea c.p . , hughes d.h . , 1999 , mnras , 308 , 377 + mclure r.j . , dunlop j.s . , kukula m.j . , 2000 , mnras , 318 , 693 + magorrian j. , et al . , 1998 , aj , 115 , 2285 + malkan m. , gorjian v. , tam r. , 1998 , apjs , 117 , 25 + merritt d. , ferrarese l. , 2000 , mnras , 2001 , 320 , l30 + percival w.j . , miller l. , mclure r.j . , dunlop j.s . , 2000 , mnras , in press , astro - ph/0002199 + press w.h . , teukolsky s.a . , vetterling w.t . , flannery b.p . , 1992 , numerical recipes , cambridge university press + schade d.j
boyle b.j . ,
letawsky m. , 2000 , mnras , 315 , 498 + silk j. , rees m.j . , 1998 ,
a&a , 331 , l1 + simien f. , de vaucouleurs g. , 1986 , apj , 302 , 564 + van der marel , r.p . , 1999 , in barnes j.e .
, sanders d.b . , eds , proc iau symp . 186 , galaxy interactions at low and high redshift , kluwer , dordrecht , p. 333
+ vestergaard m. , wilkes b.j . , barthel p.d . ,
2000 , apj , l103 + wandel a. , 1999 , apj , 519 , l39 + wandel a. , peterson b.m .
, malkan m.a . , 1999 ,
apj , 526,579 + wang y.p .
, biermann p.l .
, wandel a. , 2000 , a&a , in press , astro - ph/0008105 + wills b.j . ,
browne i.w.a . , 1986 ,
apj , 302 , 56 + wilman r.j .
, fabian a.c . , nulsen p.e.j . , 2000 ,
mnras , in press , astro - ph/0008019 + whittle m. , 1992 , apjs , 79 , 49 + | the central black - hole masses of a sample of 30 luminous quasars are estimated using @xmath0 fwhm measurements from a combination of new and previously - published nuclear spectra .
the quasar black - hole mass estimates are combined with reverberation - mapping measurements for a sample of seyfert galaxies ( wandel , peterson & malkan 1999 ) in order to study agn black - hole masses over a wide range in nuclear luminosity .
the link between bulge luminosity and black - hole mass is investigated using two - dimensional disc / bulge decompositions of the host - galaxy images , the vast majority of which are high resolution hst observations .
it is found that black - hole mass and bulge luminosity are well correlated and follow a relation consistent with that expected if black - hole and bulge mass are directly proportional . contrary to the results of wandel ( 1999 )
no evidence is found that seyfert galaxies follow a different @xmath1 relation to quasars .
however , the black - hole mass distributions of the radio - loud and radio - quiet quasar sub - samples are found to be significantly different , with the median black - hole mass of the radio - loud quasars a factor of three larger than their radio - quiet counterparts .
finally , utilizing the elliptical galaxy fundamental plane to provide stellar velocity dispersion estimates , a comparison is performed between the virial @xmath0 black - hole mass estimates and those of the @xmath2 correlations of gebhardt et al .
( 2000a ) and merritt & ferrarese ( 2000 ) . with the disc geometry of the broad - line region adopted in this paper , the virial @xmath0 black - hole masses indicate that the correct normalization of the black - hole vs. bulge mass relation is @xmath3 , while the standard assumption of purely random broad - line velocities leads to @xmath4 .
the normalization of @xmath3 provided by the disc model is in remarkably good agreement with that inferred for our quasar sample using the ( completely independent ) @xmath2 correlations .
galaxies : active galaxies : nuclei galaxies : seyfert quasars : general |
in this paper we introduce new obstructions to links being concordant , which are obtained from twisted torsion invariants . recall that an _ @xmath0component ( oriented ) link _ is an embedded ordered collection of @xmath0 disjoint ( oriented ) circles in @xmath1 .
given an oriented link @xmath2 we denote by @xmath3 the same link with the orientation of each component reversed .
we say that two @xmath0component oriented links @xmath4 and @xmath5 are _ concordant _ if there exists a collection of @xmath0 disjoint , locally flat , oriented annuli @xmath6 in @xmath7 $ ] such that @xmath8 .
we say that an @xmath0component link is _ slice _ if it is concordant to the trivial @xmath0component link .
equivalently a link is slice if it bounds @xmath0 disjoint locally flat disks in @xmath9 .
let @xmath10 be an oriented @xmath0component link . throughout this paper
we write @xmath11 , where @xmath12 is a tubular neighborhood of @xmath2 , and @xmath13 .
let @xmath14 be a non trivial homomorphism to a free abelian group and let @xmath15 be a representation with @xmath16 a subring of @xmath17 .
we can then consider the twisted homology module @xmath18^k)$ ] and we define @xmath19 } h_1(e_l;r[h]^k),\ ] ] where @xmath20 } a$ ] denotes the dimension of @xmath21 } q(h)$ ] over the quotient field @xmath22 of the integral domain @xmath23 $ ] .
if @xmath24 , then it turns out that @xmath25 is @xmath26acyclic , i.e. , @xmath27 , and we can define the torsion @xmath28 which is well defined up to multiplication by an element of the form @xmath29 where @xmath30 and @xmath31 .
this invariant generalizes the invariants introduced by reidemeister , milnor and turaev and later by lin and wada . in general
though @xmath32 will be non
zero , for example this is the case for any boundary link with at least two components ( cf .
theorem [ thm : tauboundary ] ) and the trivial representation .
suppose that @xmath33 is closed under complex conjugation .
in that case the ring @xmath23 $ ] has a natural involution given by complex conjugation and @xmath34 for @xmath35 , this involution furthermore extends naturally to an involution on @xmath22 .
now suppose that @xmath15 is a unitary representation and suppose that @xmath36 is non
trivial on each meridian of @xmath2 . under these assumptions
we define a torsion invariant @xmath37 even if @xmath38 . here
@xmath39 denotes the subgroup of norms of the multiplicative group @xmath40 , i.e. , @xmath41 .
the torsion @xmath42 viewed as an element in @xmath43 is again well defined up to multiplication by an element of the form @xmath29 where @xmath30 and @xmath35 .
the invariant @xmath44 is the twisted version of an invariant first introduced by turaev ( * ? ? ?
* section 5.1 ) .
if @xmath45 is a knot in @xmath46 , @xmath47 is an isomorphism , and @xmath48 is the trivial representation , then @xmath49 and @xmath50 ( cf .
@xcite ) . in general ,
if @xmath24 , then @xmath51 can be expressed in terms of twisted alexander polynomials ( cf .
@xcite and @xcite ) .
if @xmath52 is the trivial representation and if @xmath53 , then the torsion is related to the alexander polynomial defined using @xmath54}(h_1(e_l;\z[h]))$ ] ( cf .
* theorem 5.1.1 ) ) .
we expect that a similar result also holds in the twisted case .
we can now state our sliceness obstruction .
first , if @xmath2 is the @xmath0component unlink in @xmath46 with meridians @xmath55 , then given @xmath52 and @xmath36 as above we will show in corollary [ cor : unlink ] that @xmath56 and @xmath57 with @xmath30 and @xmath31 .
the following theorem says that the torsion invariant of a slice link is given by the above expression if the representation factors through a @xmath58-group .
this is a special case of our main result , theorem [ mainthm ] , which is a more general result for link concordance .
[ mainthmintro ] let @xmath2 be an @xmath0component oriented slice link .
let @xmath33 be a subring closed under complex conjugation and let @xmath59 be a representation which factors through a finite group of prime power order .
let @xmath14 be an epimorphism onto a free abelian group which is non
trivial on each meridian of @xmath2
. then @xmath60 and @xmath61 for some @xmath30 and @xmath35 , where @xmath55 are meridians of @xmath2 .
let @xmath62 be a knot .
fox and milnor @xcite have shown that if @xmath45 is slice , then the untwisted alexander polynomial of @xmath45 factors as @xmath63 for some polynomial @xmath64 and some @xmath65 .
it is well known that for a knot all representations which factor through a @xmath58group are necessarily abelian , in particular one can easily show that the sliceness obstruction of theorem [ mainthmintro ] reduces to the fox
milnor obstruction . on the other hand for knots kirk and livingston @xcite used twisted torsion corresponding to ` casson gordon'type representations to give sliceness obstructions which go beyond fox
we also refer to @xcite , @xcite , @xcite , @xcite and also @xcite for more on twisted alexander polynomials of knots and their relation to knot concordance .
if @xmath52 is the trivial representation , then theorem [ mainthmintro ] was proved in the one
variable case by murasugi @xcite ( cf .
also @xcite ) and in the multi
variable case it was proved independently by kawauchi ( * ? ? ?
* theorem b ) and nakagawa @xcite .
these results were reproved by turaev ( * ? ? ?
* theorem 5.4.2 ) using torsion .
theorem [ mainthmintro ] is related in spirit to ( * ? ? ?
* theorem 2.2 ) where a sliceness obstruction for links is given using the atiyah
singer eta invariant and unitary representations factoring through @xmath58groups .
the metabelian case was considered earlier by the first author and ko @xcite .
we expect the obstruction of theorem [ mainthmintro ] to be closely related to the discriminant part of hirzebruch - type invariants developed by the first author @xcite .
we prove a generalization of milnor s classical duality theorem for reidemeister torsion @xcite and use it as a key ingredient in the proof of the above sliceness obstruction .
for an even dimensional compact oriented manifold @xmath66 endowed with a homomorphism @xmath67 \to q$ ] into a field @xmath68 with involution such that @xmath66 is _
@xmath68acyclic _ , milnor proved that the reidemeister torsion of @xmath69 over @xmath68 is of the form @xmath70 , where @xmath71 is the reidemeister torsion of @xmath66 .
in particular , up to norms , the torsion of @xmath72 is trivial .
a generalization of this to the twisted acyclic case is given in @xcite .
we generalize these prior results to the _ non - acyclic _ case : _ the twisted torsion of an odd dimensional manifold @xmath72 is equal to the determinant of the twisted intersection pairing of a bounding even - dimensional manifold @xmath66 ( up to norms)_. for a precise statement , see theorem [ thm : torsion - duality ] .
we hope that this is of independent interest and useful for further applications .
the paper is organized as follows . in section [ sec : twisted invariants ] we introduce twisted torsion of 3-manifolds and links with non - acyclic twisted homology and we show how these invariants relate to intersection forms of bounding 4-manifolds . in section [ section : cs ]
we study the torsion invariant of homology cobordant 3-manifolds and in section [ section : lc ] we apply these results to link concordance where we prove our main theorem ( which implies theorem [ mainthmintro ] ) . in section [ section :
prep ] we discuss and to a certain degree classify complex representations which factor through @xmath58groups . in section
[ section : boundarylinks ] we show how to compute twisted torsion of boundary links using a boundary link seifert matrix and we discuss the behavior of our invariants for links which are boundary slice .
finally in section [ section : satellitebing ] we prove a formula for the twisted invariants of satellite links and we use this formula to reprove the fact ( first proved by the first author @xcite ) that the bing double of the figure 8 knot is not slice .
the following are understood unless it says specifically otherwise : ( i ) groups are finitely generated , ( ii ) manifolds are compact , orientable and connected , ( iii ) rings are understood to be associative , commutative with unit element , and ( iv ) homology and cohomology groups are taken with integral coefficients .
we would like to thank baskar balasubramanyam , david cimasoni , taehee kim , vladimir turaev and liam watson for helpful comments and conversations .
we also thank an anonymous referee for useful suggestions .
the first author was supported by the national research foundation of korea ( nrf ) grants funded by korean government(mest ) ( grant no .
20090094069 and r012007000116870 ) .
in this section we define twisted torsion invariants of odd dimensional manifolds for finite dimensional unitary representations for which the corresponding twisted homology groups are not necessarily acyclic .
we furthermore investigate twisted torsion of odd dimensional manifolds cobounding a manifold , along the way we generalize a theorem of milnor to the nonacyclic case . in the last section
we then specialize to the case of 3-manifolds and link exteriors .
let @xmath73 be a cw - pair with @xmath74 , and @xmath75 a finite dimensional representation over a ring @xmath16 .
denote by @xmath76 the universal covering of @xmath77 and write @xmath78 . recall that @xmath79 acts on the left of @xmath80 as the deck transformation group .
we consider the cellular chain complex @xmath81 as a right @xmath82-module via @xmath83 for a chain @xmath84 .
the @xmath85coefficient cellular chain complex of @xmath73 is defined to be @xmath86 where @xmath85 is viewed as a @xmath87bimodule via the representation @xmath88 .
we define _ the @xmath89th twisted homology group _ to be the @xmath16-module @xmath90 if the representation is understood clearly , then we just write @xmath91 instead of @xmath92 .
now assume that @xmath16 is a ring with a ( possibly trivial ) involution @xmath93 . given an @xmath16module @xmath94 we denote by @xmath95 the opposite @xmath16module , i.e. @xmath96 as abelian groups , and the multiplication by @xmath97 on @xmath95 is given by the multiplication by @xmath98 on @xmath94 . finally recall that a representation @xmath99 is called _ unitary _ if @xmath100 for all @xmath101 .
we now have the following twisted poincar duality theorem ( cf .
also @xcite , ( * ? ? ?
* lemma 4.12 ) , and @xcite ) . here , given a vector space @xmath102 over a field @xmath68 ,
we denote the dual vector space by @xmath103 .
[ thm : duality ] let @xmath66 be an @xmath104manifold with @xmath105 , @xmath72 and @xmath106 submanifolds such that @xmath107 .
let @xmath108 be a unitary representation over a field @xmath68 with involution
. then there exists a natural isomorphism @xmath109 of @xmath68vector spaces .
the theorem can be proved along the same lines as the standard proof of poincar duality .
for the reader s convenience we give an outline of the proof .
choose a cell structure of @xmath110 to define @xmath111 and let @xmath112 be the chain complex of the dual cell structure .
then the equivariant intersection pairing ( e.g. , see @xcite ) gives an isomorphism of chain complexes @xmath113 applying the universal coefficient theorem over @xmath68 , the conclusion follows .
we begin by recalling the algebraic setup for torsion invariants .
suppose @xmath114 is a based chain complex over a field @xmath68 and @xmath115 is a basis for @xmath116 , i.e. , @xmath117 is a basis of the @xmath68vector spaces @xmath118 .
the _ torsion _
@xmath119 is defined as in @xcite .
( see also milnor s classic introduction to torsion @xcite , but note that we follow turaev s convention , which gives the reciprocal of the torsion in @xcite . )
if @xmath116 is identically zero , then we will just write @xmath120 for the torsion . in the following theorem , we collect well - known algebraic properties of torsion which will be useful later . in the theorem , given two bases @xmath121 and @xmath122 of a vector space
, we will denote the basis change matrix from @xmath121 to @xmath122 by @xmath123 , as in milnor @xcite and turaev @xcite .
[ thm : alg - prop - torsion][thm : mi66 ] suppose @xmath124 is a based chain complex over @xmath68 and @xmath125 is a basis of @xmath116 . 1 .
suppose @xmath126 is another basis of @xmath116 .
then @xmath127 2 .
let @xmath128 be the dual based chain complex given by @xmath129 and @xmath130 be the basis of @xmath131 dual to @xmath132
. then @xmath133 .
3 . suppose @xmath134 is acyclic .
choose a basis of the @xmath104th cycle submodule @xmath135 , and view @xmath136 as acyclic based chain complexes ( indexed so that @xmath137 , @xmath138 )
. then @xmath139 .
4 . suppose @xmath140 is a short exact sequence of based chain complexes and @xmath130 , @xmath132 , and @xmath141 are bases of @xmath142 , @xmath116 and @xmath143 , respectively .
we view the associated homology long exact sequence as an acyclic complex , say @xmath144 , based by @xmath132 , @xmath130 , @xmath141 .
then @xmath145 let @xmath73 be a finite cw - pair with @xmath74 and let @xmath146 be a representation over a field @xmath68 .
let @xmath147 be a basis of @xmath148 .
the universal cover @xmath149 has a natural cell structure , and the chain complex @xmath150 can be based over @xmath82 by choosing a lift of each cell of @xmath73 and orienting it .
this , together with the standard basis of @xmath151 , gives rise to a basing of @xmath152 over @xmath68 .
we can then define the _ twisted torsion _
@xmath153 to be the torsion of @xmath152 with respect to @xmath132 .
we will drop @xmath132 from the notation if @xmath154 .
standard arguments ( e.g. , see @xcite ) show that @xmath155 is well - defined up to multiplication by an element in @xmath156 , and is invariant under simple homotopy preserving @xmath132 . by chapman s theorem @xcite
the invariant @xmath155 only depends on the homeomorphism type of @xmath73 .
in particular when @xmath157 is a manifold pair , we can define @xmath158 by picking any finite cw
structure for @xmath157 .
let @xmath72 be an odd dimensional manifold with @xmath159 and let @xmath160 be a representation over a field @xmath68 .
in many interesting cases @xmath161 will be nontrivial .
for example the untwisted multivariable alexander module of a boundary link with at least two components is always non - torsion . in order to define an invariant for the nonacyclic case without referring to a basis of homology , we now assume the following two conditions hold : 1 .
@xmath68 is endowed with a ( possibly trivial ) involution , and @xmath88 is unitary .
@xmath162 is @xmath151-acyclic , i.e. , @xmath163 .
since @xmath88 is unitary , we have the poincar duality isomorphism @xmath164 of @xmath68vector spaces , by theorem [ thm : duality ] and by condition ( 2 ) .
since @xmath72 is odd dimensional , one can pick a basis @xmath115 for @xmath161 with the following property : for each @xmath89 , @xmath165 is the dual basis of @xmath166 via the above poincar duality isomorphism .
we call such a basis @xmath167 a _ self - dual basis _ for @xmath161 .
[ lem : changebasis ] suppose @xmath132 and @xmath130 are self - dual bases for @xmath161 .
then for some @xmath168 , up to the indeterminacy of the torsion , @xmath169 let @xmath170 be the determinant of the base change matrix from @xmath117 to @xmath171
. then @xmath172 .
the desired conclusion follows immediately from theorem [ thm : alg - prop - torsion ] .
we define the _ norm subgroup _ of @xmath173 to be @xmath174 , and we say @xmath175 is a _ norm _ when @xmath176 .
we define @xmath177 where @xmath132 is a self - dual basis of @xmath161 . by lemma [ lem : changebasis ]
the invariant @xmath178 is well - defined up to multiplication by an element in @xmath179 .
this invariant was first introduced by turaev ( * ? ? ?
* section 5.1 ) in the untwisted case . in the following ,
given @xmath180 , we will sometimes write @xmath181 , to indicate that there exists a representative of @xmath182 which equals @xmath183 . in this subsection
we prove a _ non - acyclic _ generalization of a well - known theorem of milnor ( * ? ? ?
* theorem 2 ) and of its twisted analogue due to kirk and livingston ( * ? ? ?
* theorem 5.1 and corollary 5.3 ) .
let @xmath66 be a @xmath184dimensional manifold with ( possibly disconnected ) boundary @xmath72 .
let @xmath185 be a unitary representation .
now consider the map @xmath186 where the first map is the isomorphism given by theorem [ thm : duality ] and the second map is induced by @xmath187 .
this map gives rise to a pairing @xmath188 which is well - known to be @xmath189-hermitian .
this pairing is called the _ equivariant intersection form of @xmath190_. this form is in general singular , in fact for any @xmath191 and @xmath192 we have @xmath193 . in particular
@xmath194 gives rise to a pairing on @xmath195 which turns out to be non - singular .
we now pick a basis @xmath196 for @xmath197 and we compute @xmath198 .
note that if we change the basis , then the determinant of the form changes by a norm .
put differently , we obtain a well - defined invariant @xmath199 note that @xmath200 since @xmath194 is @xmath189-hermitian .
[ thm : torsion - duality ] suppose @xmath66 is a @xmath184dimensional manifold with ( possibly disconnected ) boundary @xmath72 .
let @xmath185 be a unitary representation .
then @xmath201 up to the indeterminacy @xmath202 . in the following
we will only apply the theorem to the case that @xmath203 , but we hope that the general case is of independent interest . to prove theorem [ thm : torsion - duality ] we need the following duality of torsion , which is essentially due to milnor @xcite .
( see also kirk and livingston @xcite for the twisted case . )
[ lem : torsion - duality ] suppose @xmath66 is an @xmath104manifold and @xmath72 , @xmath106 are submanifolds of @xmath204 such that @xmath105 and @xmath205 .
suppose @xmath206 is a unitary representation , @xmath115 is a basis for @xmath207 , and @xmath130 is the dual basis for @xmath208 .
then @xmath209 the lemma follows immediately from the duality @xmath210 ( see the proof of theorem [ thm : duality ] ) and theorem [ thm : alg - prop - torsion ] . choose a basis @xmath132 for @xmath211 , and choose a self - dual basis @xmath130 for @xmath161 .
let @xmath141 be the basis of @xmath212 which is dual to @xmath132 .
from ( the proof of ) theorem [ thm : duality ] , theorem [ thm : alg - prop - torsion ] , and lemma [ lem : torsion - duality ] , it follows that @xmath213 where @xmath214 is the torsion of the acyclic chain complex ( @xmath215 homology long exact sequence ) @xmath216 which is based by @xmath132 , @xmath130 , @xmath141 .
we break the long exact sequence at @xmath217 as in theorem [ thm : alg - prop - torsion ] : let @xmath218 be the image of @xmath219 , choose a basis @xmath121 for @xmath218 , and let @xmath220 then by theorem [ thm : alg - prop - torsion ] ( 3 ) we have @xmath221 . since @xmath130 is self - dual , @xmath222 is canonically isomorphic , as a based chain complex , with the dual chain complex @xmath223 of @xmath128 except @xmath224
. therefore we have @xmath225 where @xmath226 is the dual basis of @xmath121 for @xmath227 .
it follows that @xmath228 note that by definition we have @xmath229 . since the intersection form is @xmath189-hermitian , @xmath230 and so @xmath231 in @xmath232 .
this completes the proof . if @xmath66 is @xmath151acyclic , then @xmath233 automatically and @xmath234 in the above proof .
so @xmath235 up to @xmath236 , where @xmath237 .
this shows our non - acyclic result specializes to milnor s theorem ( * ? ? ?
* theorem 2 ) and its twisted analogue due to kirk and livingston ( * ? ? ?
* theorem 5.1 and corollary 5.3 ) . in this section
we now specialize to the case of 3-manifolds .
let @xmath72 be a 3-manifold with empty or toroidal boundary ( e.g. a link exterior ) .
we write @xmath159 .
let @xmath238 be an epimorphism onto a nontrivial free abelian group @xmath144 .
let @xmath16 be a domain with involution .
we equip @xmath23 $ ] with the involution given by @xmath239 for @xmath97 and @xmath35 .
this extends to an involution on @xmath22 , the quotient field of @xmath23 $ ] .
now let @xmath240 be a unitary representation . using @xmath52 and @xmath36
, we define a left @xmath241$]-module structure on @xmath23^k : = r^k\otimes_r r[h]$ ] as follows : @xmath242 where @xmath101 , @xmath243 and @xmath244 $ ] .
this extends to a @xmath241$]-module structure on @xmath26 .
it can be seen that @xmath245 is unitary since @xmath246 is unitary .
we will several times make use of the following simple fact : @xmath247 we now say that @xmath238 is _ admissible _ if @xmath36 restricted to any boundary component of @xmath72 is nontrivial .
note that @xmath36 is always admissible if @xmath72 is closed .
if @xmath36 is admissible then condition ( 2 ) in section [ section : twitor ] is satisfied ( e.g. , see ( * ? ? ?
* proposition 3.5 ) and ( * ? ? ?
* section 3.3 ) ) and by the discussion of the previous section we thus obtain an invariant @xmath248 well - defined up to multiplication by an element of the form @xmath29 with @xmath249 and @xmath35 . note that if @xmath250 is nonzero , then in the above setting we have @xmath251 for @xmath252 ( see e.g. @xcite ) , so that it suffices to choose @xmath253 in the definition of torsion .
we now specialize even further , namely to the case of link exteriors .
let @xmath10 be an oriented @xmath0component link .
we denote the exterior by @xmath25 and the ( oriented ) meridians by @xmath55 . using the basis @xmath55
we can now naturally identify @xmath254 with @xmath255 .
we say that a homomorphism @xmath256 is _ admissible _ if @xmath36 is an epimorphism onto a nontrivial free abelian group @xmath144 such that the epimorphism is nontrivial on each subsummand of @xmath257 . in this case
the induced map @xmath258 is admissible in the above sense , so that we obtain a well - defined torsion invariant @xmath259
in this section we investigate the behaviour of the twisted torsion under homology cobordism .
recall that two @xmath260manifolds @xmath72 and @xmath106 ( possibly with nonempty toroidal boundary ) are said to be _ homology cobordant _ if there exists a @xmath261manifold @xmath66 containing @xmath72 and @xmath106 as submanifolds such that @xmath262 , @xmath263 , and the inclusions @xmath264 and @xmath265 induce isomorphisms on @xmath266 . from now on
, @xmath33 is always assumed to be a subring closed under complex conjugation .
we denote by @xmath267 the set of representations @xmath268 factoring through a @xmath58-group .
we will continuously make use of the well - known fact that complex representations factoring through finite groups are necessarily unitary . for @xmath269 and @xmath270 , we define @xmath271 our main result is the following : [ thm : homology - cob - inv ] suppose two @xmath260manifolds @xmath72 and @xmath106 are homology cobordant and @xmath144 is a free abelian group
. then there are bijections @xmath272 such that for any @xmath269 , @xmath36 is admissible if and only if @xmath273 is admissible , and in this case , for any @xmath274 , we have @xmath275 and @xmath276 for some @xmath277 and @xmath35 . in the proof of theorem [ thm : homology - cob - inv ] we will see that for a given homology cobordism @xmath66 between @xmath72 and @xmath106 , the maps @xmath278 and @xmath279 are both induced by @xmath280 , in particular @xmath278 and @xmath279 are ` related ' bijections .
we first construct the bijections in theorem [ thm : homology - cob - inv ] using stallings theorem @xcite .
suppose @xmath66 is a homology cobordism between @xmath72 and @xmath106 .
obviously the induced isomorphisms on @xmath281 gives rise to bijections @xmath282 their composition will be @xmath278 in the above statement .
since @xmath283 it follows that @xmath269 is admissible if and only if so is @xmath273 .
the bijection @xmath279 is constructed as follows . for a group @xmath284
, we denote the @xmath71th lower central subgroup by @xmath285 , which is defined inductively via @xmath286 , @xmath287 $ ] .
for any @xmath288 , @xmath246 factors through @xmath289 for some @xmath71 , since any @xmath58-group is nilpotent . by stallings theorem @xcite , we have @xmath290 therefore @xmath246 induces a representation @xmath291 which factors through @xmath292 .
it is easily seen that this induces a bijection @xmath293 .
similarly for @xmath106 and @xmath66 , and composing them , we obtain a bijection @xmath294 . in our proof of the conclusion on torsions in theorem [ thm : homology - cob - inv ] we will compute ( the quotient of ) the torsions of @xmath72 and @xmath106 in terms of the @xmath26-coefficient intersection pairing of a cobordism @xmath66 , appealing to theorem [ thm : torsion - duality ] .
the key point is that one can prove that the intersection pairing is trivial when @xmath66 is a homology cobordism .
this is done following a standard strategy , namely by controlling the size of the underlying @xmath26-coefficient homology of @xmath110 , appealing to a chain contraction argument originally due to vogel . in lemma [ lem : cohn - local - prop ] below
we denote by @xmath295 \to \z$ ] the augmentation map given by @xmath296 .
[ lem : cohn - local - prop ] [ thm : levine ] suppose @xmath79 is a group and @xmath16 is a domain with characteristic zero , @xmath297 , and @xmath298 is an epimorphism onto a nontrivial free abelian group @xmath144 .
if @xmath94 is a square matrix over @xmath241 $ ] such that @xmath299 is invertible over @xmath300 , then @xmath301 is invertible over @xmath22 . another way of phrasing lemma [ lem : cohn - local - prop ] is to say that the @xmath241$]-module @xmath26 is in fact a module over the _
cohn localization _ of @xmath241 $ ] .
our proof of lemma [ lem : cohn - local - prop ] depends heavily on a result originally due to strebel @xcite and levine @xcite .
since @xmath16 has characteristic zero and since the quotient ring of @xmath16 gives the same @xmath22 , we may assume that @xmath16 contains @xmath302 .
suppose @xmath246 factors through @xmath303 where @xmath218 is a @xmath58-group .
we have the following commutative diagram : @xmath304}(\z[\pi]^n ) } \arrow{e , t}{\alpha\otimes\psi } \arrow{s , l}{f } \arrow{se , t}{\alpha } \node{\end_r(r[h]^{nk } ) } \arrow{e , t}{\det } \arrow{s , r}{\epsilon_r } \node{r[h ] } \arrow{s , r}{\epsilon_r } \\
\node{\end_{\q[p]}(\q[p]^n ) } \arrow{e } \node{\end_r(r^{nk } ) } \arrow{e , t}{\det } \node{r } \end{diagram}\ ] ] where the induced maps are denoted by the same symbol , as an abuse of notation .
the augmentation @xmath23
\to r$ ] is denoted by @xmath305 .
suppose @xmath306}(\z[\pi]^n)$ ] is an @xmath307 matrix over @xmath241 $ ] such that @xmath299 is invertible .
due to strebel ( * ? ? ?
* lemma 1.10 ) and levine ( * ? ? ?
* lemma 4.3 ) this implies that @xmath308 is invertible over @xmath309 $ ] , i.e. , a unit in @xmath310}(\q[p]^n)$ ] . since the maps in the above diagram are multiplicative , it follows that @xmath311 is invertible .
therefore @xmath312 and @xmath313 .
it follows that @xmath94 is injective over @xmath23 $ ] and hence invertible over the field @xmath22 .
a standard well - known chain contraction argument originally due to vogel ( see also levine ( * ? ? ?
* proposition 4.2 ) and cochran - orr - teichner @xcite ) shows the following statement as a consequence of lemma [ lem : cohn - local - prop ] : [ lem : homology - cohn - local - coeff ] suppose @xmath314 is a finite cw - pair , @xmath16 is of characteristic zero , @xmath315 , and @xmath316 is an epimorphism onto a nontrivial free abelian group . if @xmath317 , then @xmath318 .
suppose @xmath66 is a homology cobordism between @xmath72 and @xmath106 .
recall that we have constructed bijections @xmath319 fix @xmath320 and @xmath321 , and as an abuse of notation , we denote by @xmath246 and @xmath36 the corresponding representations and homomorphisms of @xmath322 and @xmath323 as well . for convenience write @xmath324 .
since @xmath325 , we have @xmath326 by lemma [ lem : homology - cohn - local - coeff ] , and so @xmath327 . similarly for @xmath106 and @xmath66 , and from this the conclusion on the ranks follows immediately .
now pick a self dual basis @xmath328 for @xmath161 and pick a self dual basis @xmath329 for @xmath330 . recall that @xmath331 since @xmath36 is admissible ( see section [ section : twi3mfd ] ) .
in particular @xmath328 gives rise to a basis for @xmath332 which we also denote by @xmath328 .
note that @xmath328 and @xmath329 give rise to a self dual basis for @xmath333 which we denote by @xmath334 .
since @xmath326 , we have @xmath203 .
therefore , by theorem [ thm : torsion - duality ] , we have @xmath335 . applying theorem [ thm : alg - prop - torsion ] to the excision short exact sequence @xmath336 and then applying lemma [ lem : torsion - duality ] we obtain @xmath337 where @xmath214 is the torsion of the homology long exact sequence @xmath338 which is based by @xmath339 and @xmath328 .
it follows easily that @xmath340 .
therefore , multiplying @xmath341 by @xmath342 , the conclusion follows .
it is well known that if two links are concordant , then their exteriors are homology cobordant . more precisely ,
if @xmath0-component links @xmath2 and @xmath343 in @xmath46 are concordant , then the exterior @xmath66 of a concordance has boundary @xmath344 ) \cup -e_{l'}$ ] where we glue the @xmath89-th meridian ( longitude ) of @xmath2 to the @xmath89-th meridian ( longitude ) of @xmath343 ( recall that links are assumed to be _ ordered _ collections of circles ) .
rounding corners and extending collars outward , we can assume that @xmath345 and @xmath346 .
using alexander duality it is straightforward to verify that @xmath66 is indeed a homology cobordism . therefore , we can apply the invariance of torsion ( theorem [ thm : homology - cob - inv ] ) immediately .
furthermore , in case of links , we can make further observations on the correspondence between representations .
the main aim of this section is to provide a precise description of this correspondence .
first , recall that given an oriented link @xmath10 with @xmath0 components , we can naturally identify @xmath347 with @xmath255 in such a way that the @xmath89th positive meridian @xmath348 of @xmath2 represents the @xmath89th standard basis of @xmath255 .
therefore @xmath349 is identified with @xmath350 , where @xmath144 denotes a nontrivial free abelian group as in the previous sections .
suppose two links @xmath2 and @xmath343 are concordant , in particular the concordance exterior @xmath66 is a homology cobordism between @xmath25 and @xmath351 .
recall that in theorem [ thm : homology - cob - inv ] , we defined a bijection @xmath352 using @xmath353 . since the meridians of @xmath2 and @xmath343 are freely homotopic in @xmath66 , it is easily seen that the bijection @xmath278 induces the identity on @xmath350 , that is , a homomorphisms @xmath258 and its image under @xmath278 are identified with the same homomorphism @xmath354 . recall that a homomorphism @xmath355 is _ admissible _ if @xmath36 is an epimorphism which is nontrivial on each @xmath348 . in this case
the corresponding map @xmath258 , which will also be denoted by @xmath88 , is admissible in the sense of section [ sec : twisted invariants ] .
thus the twisted torsion @xmath356 can be defined for any unitary representation @xmath357 .
the correspondence between representations of @xmath358 for concordant links is better described in terms of representations of the lower central quotients @xmath359 .
it causes no loss of generality , since any representation of @xmath79 factoring through a @xmath58-group factors through @xmath359 for some @xmath71 since @xmath58groups are nilpotent . to proceed
we will need some technicalities discussed in the following subsection . in this subsection we state some known results and some folklore results on link concordance and lower central series .
let @xmath10 be an ordered , oriented @xmath0component link .
we denote the meridians by @xmath55 and by abuse of notation we will denote the corresponding elements in @xmath360 by @xmath55 as well .
we now denote by @xmath361 the free group on @xmath0 generators @xmath362 .
following ideas of levine ( see @xcite ) we define an _
@xmath363-structure _ for @xmath2 to be a homomorphism @xmath364 such that for any @xmath365 the element @xmath366 is a conjugate of @xmath367 .
( here recall that given a group @xmath79 we denote by @xmath368 the @xmath71th lower central subgroup . )
furthermore , we refer to a link @xmath2 equipped with an @xmath363structure as an _
@xmath363link_. a _ special automorphism of @xmath363 _ is an automorphism @xmath369 of @xmath363 such that @xmath370 is a conjugate of @xmath367 for each @xmath89 .
[ lemmapsi ] let @xmath2 be an ordered oriented @xmath0-component link .
we write @xmath371 .
any @xmath363-structure @xmath372 induces an isomorphism @xmath373 .
if @xmath374 and @xmath375 are @xmath363structures for @xmath2 , then there exists a special automorphism @xmath376 such that @xmath377 .
we denote by @xmath378 the map which sends @xmath367 to @xmath348 .
we also pick a map @xmath379 such that the map @xmath380 agrees with @xmath381 . finally note that @xmath382 descends to a map @xmath383 which we denote by @xmath384
we now have the following commutative diagram , where the vertical maps are the obvious projection maps : @xmath385{e , t}{\psi } \arrow{s}\arrow{se , t}{\gamma } \node{}\node{f}\arrow{s } \\
\node{f / f_q}\arrow{e , t}{\ol{\gamma } } \node{\pi/\pi_q}\arrow{e , t}{\varphi}\node{f / f_q.}\end{diagram}\ ] ] it follows from stallings theorem @xcite applied to @xmath36 and from the commutativity of the diagram that the map @xmath386 is an isomorphism . on the other hand it follows from milnor s theorem ( *
* theorem 4 ) that @xmath384 is surjective .
we now conclude that @xmath374 is an isomorphism .
this concludes the proof of ( 1 ) .
the second statement is an immediate consequence of ( 1 ) .
note that it follows from lemma [ lemmapsi ] ( 1 ) and @xcite that a link admits an @xmath363-structure if and only if milnor s @xmath387-invariants of the form @xmath388 defined in @xcite vanish for @xmath2 .
now we are ready to derive the following theorem : [ thm : mainthm][mainthm ] suppose two @xmath0component ordered links @xmath2 and @xmath343 are concordant , @xmath256 is an admissible homomorphism , and @xmath16 is a subring of @xmath17 closed under complex conjugation .
if @xmath374 and @xmath375 are arbitrary @xmath363structures for @xmath2 and @xmath343 , respectively , then there exists a special automorphism @xmath376 of @xmath363 such that for any @xmath389 we have @xmath390 and @xmath391 for some @xmath392 and @xmath35 .
let @xmath66 be the exterior of a concordance , @xmath393 , and @xmath394 . since @xmath395 we can apply stallings theorem @xcite to conclude that the inclusion maps induce isomorphisms @xmath396 observe that the composition sends meridians of @xmath2 to ( conjugates of ) meridians of @xmath343 .
it follows that the composition @xmath397 gives an @xmath363structure on @xmath343 which we denote by @xmath398 . for @xmath399
, @xmath400 corresponds to @xmath401 under the bijection @xmath279 in theorem [ thm : homology - cob - inv ] , by the definition of @xmath363concordance and the definition of @xmath279 ( see subsection [ subsec : correspondence - of - representations ] ) .
now by theorem [ thm : homology - cob - inv ] , we have @xmath402 by lemma [ lemmapsi ] there exists a special automorphism @xmath376 such that @xmath403 . the theorem
now immediately follow from these observations .
the following corollary is equivalent to theorem [ mainthmintro ] given in the introduction .
[ cor : maincor ] suppose @xmath2 is an @xmath0component slice link , @xmath355 is an admissible homomorphism , and @xmath16 is a subring of @xmath17 closed under complex conjugation . then for any @xmath404 , we have @xmath405 and @xmath406 for some @xmath30 and @xmath35 .
. let @xmath407 . by definition @xmath52 factors through a @xmath58-group .
since @xmath58-groups are nilpotent it follows that there exists a @xmath71 such that @xmath52 factors through @xmath359 . since @xmath2 is slice , by the proof of theorem [ thm : mainthm ] there exists an @xmath363structure @xmath408 on @xmath2 .
in particular there exists @xmath409 such that @xmath410 . since @xmath2 is concordant to the trivial link , we may by theorem [ mainthm ] assume without loss of generality that @xmath2 is already the trivial link .
( precisely speaking , this requires the change of @xmath411 to @xmath412 for some special automorphism @xmath376 of @xmath363 , but this can be ignored since @xmath413 is left unchanged when @xmath367 is replaced with its conjugate . )
the corollary now follows from the explicit calculation for the unlink , which will be done in theorem [ cor : unlink ] .
in this section we collect a few basic facts of the theory of representations factoring through @xmath58-groups
. this summary will be useful in the later discussion of examples .
given @xmath414 we denote by @xmath415 the subgroup of permutation matrices , i.e. matrices with exactly one non trivial entry in each row and in each column , and all the non trivial entries are equal to one .
let @xmath33 be a subring which is closed under complex conjugation .
we then write @xmath416 for the subgroup of diagonal matrices .
[ proppequalspd ] [ lem : ppowerrep ] let @xmath79 be a group with generators @xmath417 .
let @xmath418 be a representation .
we write @xmath419 .
let @xmath58 be a prime .
then @xmath52 factors through a @xmath58group if and only if there exists a matrix @xmath420 such that for @xmath421 we can write @xmath422 with @xmath423 and @xmath424 such that 1 . @xmath425 generate a subgroup of @xmath58power order , and 2 .
the diagonal entries of @xmath426 are @xmath58power roots of unity .
we will say that a matrix @xmath427 is @xmath58diagonal if @xmath427 is diagonal and if all the diagonal entries are @xmath58power roots of unity . we will several times make use of the following two basic observations : 1 . if @xmath428 and @xmath429 with @xmath430 , then @xmath431 and @xmath432 .
2 . let @xmath428 and @xmath429 , then @xmath433 and @xmath434 is a diagonal matrix , furthermore , if @xmath435 are @xmath58diagonal , then @xmath434 is also @xmath58diagonal .
now assume that we have a representation @xmath436 with the following property : there exists a matrix @xmath437 such that for @xmath421 we can write @xmath438 with @xmath439 which satisfy conditions ( 1 ) and ( 2 ) . we claim that @xmath440 is a @xmath58group . without loss of generality
we can assume that @xmath68 is the identity .
let @xmath441 be any element in the group @xmath440 , we can write @xmath442 for some @xmath443 . we have to show that @xmath441 has @xmath58power order .
let @xmath444 . by assumption @xmath445 for some @xmath446 .
using the above observations we can write @xmath447 for some @xmath58diagonal matrix @xmath427 .
it follows that @xmath448 ( and hence @xmath441 ) has @xmath58power order .
this shows that @xmath440 is a @xmath58group .
now let @xmath449 be a representation which factors through a @xmath58group @xmath218 .
we write @xmath419 . by (
* corollary 4.9 ) any representation @xmath450 of the @xmath58-group @xmath218 is induced from a representation of degree 1 .
this means that there exists a finite index subgroup @xmath451 and a one dimensional representation @xmath452 such that @xmath52 is given by the natural @xmath218left action on @xmath453\otimes_{\c [ q ] } \c$ ] .
note that the one dimensional representation is necessarily of @xmath58power order .
now pick representatives @xmath454 for @xmath455 .
these representatives defines a basis for @xmath453\otimes_{\c [ q ] } \c$ ] , and writing @xmath52 with respect to this basis we see that @xmath52 is of the required type .
boundary links form an important subclass of links which has been intensely studied over the years . in this section
we will show how to calculate twisted invariants for boundary links and we will study the twisted invariants of boundary links that are ` boundary slice ' . a _
boundary link _ is an @xmath0component oriented link in @xmath46 which has @xmath0 disjoint seifert surfaces , i.e. there exist @xmath0 disjoint oriented surfaces @xmath456 such that @xmath457 .
we refer to such a disjoint collection of seifert surfaces as a _ boundary link seifert surface_. not every link is a boundary link .
for example it is clear that a boundary link has trivial linking numbers .
furthermore , milnor s @xcite @xmath458invariants are trivial for a boundary link , and in fact trivial for any link concordant to a boundary link .
cochran and orr @xcite showed that there also exist links with vanishing @xmath458invariants which are not concordant to a boundary link ( cf . also @xcite and @xcite ) . throughout this section
let @xmath361 be the free group on the generators @xmath362 .
@xmath361link _ is a pair @xmath459 where @xmath2 is an oriented @xmath0component link in @xmath1 and @xmath460 is an epimorphism sending an @xmath461 meridian to @xmath367 .
note that @xmath462 gives rise to a map from @xmath25 to the wedge of @xmath0 circles , and by a standard transversality argument such a map then gives rise to a boundary link seifert surface .
conversely , the existence of a boundary link seifert surface @xmath102 for @xmath2 produces such an epimorphism @xmath463 by the thom - pontryagin construction .
given an oriented boundary link @xmath2 with disjoint oriented surfaces @xmath456 we can define the ` boundary link seifert matrix ' as follows : let @xmath464 be the genus of @xmath465 . for @xmath365 pick a basis @xmath466 for @xmath467 .
for @xmath468 we define @xmath469 to be the @xmath470matrix given by @xmath471 where @xmath472 denotes the positive push off of @xmath473 from @xmath474 into @xmath475 .
we now view @xmath476 as an @xmath477matrix of matrices and we refer to it as a _ boundary link seifert matrix_. we refer to @xcite and @xcite for details and for more information on boundary link seifert matrices .
we now turn to the computation of twisted invariants of boundary links .
let @xmath478 be an @xmath361-link and @xmath479 a corresponding boundary link seifert matrix where @xmath480 is an @xmath481matrix .
let @xmath482 and @xmath483 we view @xmath484 as an @xmath485matrix over @xmath486 $ ] .
let @xmath355 be an admissible epimorphism to a nontrivial free abelian group @xmath144 and let @xmath487 be a representation .
note that all entries of the matrix @xmath488 are sums of monomials in one variable , we can therefore unambiguously define the matrix @xmath489 to be the @xmath490matrix over @xmath23 $ ] given by replacing each summand @xmath491 of an entry of @xmath484 by the matrix @xmath492,k)$ ] .
( here and throughout this section we will identify the additive group @xmath255 with the multiplicative group generated by @xmath493 . ) [ thm : compboundary][thm : tauboundary ] let @xmath478 be an @xmath361link .
let @xmath94 be a boundary link seifert matrix corresponding to seifert surfaces given by @xmath374 .
let @xmath494 be an admissible epimorphism to a nontrivial free abelian group @xmath144 .
let @xmath33 be a subring closed under complex conjugation , and let @xmath495 be a unitary representation which factors through the epimorphism @xmath374 . then the following hold : 1 .
we have @xmath496 and equality holds if and only if @xmath497 , 2 . if @xmath498 , then @xmath499 3 . if @xmath52 factors through a @xmath58group , then @xmath500 .
let @xmath501 be a properly embedded boundary link seifert surface for @xmath2 corresponding to @xmath374 together with bases @xmath466 for @xmath467 such that @xmath502 is the corresponding seifert matrix , i.e. @xmath471 . let @xmath503 .
note that the induced maps @xmath504 and @xmath505 are trivial , in particular @xmath52 and @xmath36 restricted to @xmath506 and @xmath507 are trivial .
we therefore get the following exact mayer vietoris sequence : @xmath508^k \to h_1(c)\otimes_\z r[h]^k \to h_1(e_l;r[h]^k )
\\ \to \bigoplus\limits_{i=1}^m h_0(v_i)\otimes_\z r[h]^k \to h_0(c)\otimes_\z r[h]^k \to h_0(e_l;r[h]^k ) \to 0 \end{gathered}\ ] ] ( we refer to ( * ? ? ?
* section 3 ) for details . )
now let @xmath509 be the basis for @xmath510 dual to the basis @xmath511 , i.e. @xmath512 . using these bases it is well known that the map @xmath513^k{\mathchoice{{\ext@arrow 0359\longrightarrowfill@ { } { } } } { { \ext@arrow 0359\rightarrowfill@{}{}}}{{\ext@arrow 0359\rightarrowfill@{}{}}}{{\ext@arrow 0359\rightarrowfill@ { } { } } } } h_1(c)\otimes_\z r[h]^k\ ] ] is represented by @xmath488 .
we write @xmath514 .
note that @xmath488 is an @xmath307-matrix .
we now have @xmath515 in particular @xmath516 , and equality holds if and only if the rank of the @xmath307-matrix @xmath517 is equal to @xmath104 .
this proves ( 1 ) .
we now turn to the calculation of @xmath44 .
suppose that @xmath498 .
for @xmath365 we pick a point @xmath518 .
note that @xmath519 defines a basis for the free @xmath300module @xmath520 and @xmath521 since @xmath522 are the only units in the ring @xmath300 .
we now denote by @xmath523 the standard basis of @xmath151 . for @xmath365
we then let @xmath524 note that for @xmath365 the set @xmath525 defines a basis for the free @xmath22module @xmath526 and @xmath527 we now also pick a point @xmath528 and for @xmath529 we denote by @xmath530 the result of gluing @xmath531 and @xmath532 together along @xmath533 $ ] .
note that @xmath534 are a basis for @xmath535 .
we now let @xmath536 note that the set @xmath537 defines a basis for the free @xmath22module @xmath538 and a similar argument as above shows that @xmath539 now note that @xmath56 implies by ( 1 ) that the map @xmath540 is injective .
this implies that the map @xmath541 is an isomorphism , and we denote by @xmath542 be the image of @xmath543 under this isomorphism . by a slight abuse of notation we also write @xmath544 . for @xmath365
we now write @xmath545 . furthermore for @xmath365
we let @xmath348 be a meridian of @xmath546 which is based at @xmath121 and which intersects @xmath465 geometrically once in the positive direction and which is disjoint from @xmath547 .
we now denote by @xmath548 the @xmath361cover corresponding to @xmath374 .
we pick a point @xmath549 over @xmath121 and for @xmath365 we denote by @xmath550 the component of @xmath551 such that @xmath552 . for @xmath553 and @xmath554
we define @xmath555 then the following hold : 1 .
@xmath556 is a basis for @xmath557 , 2 . for any @xmath558 and @xmath559 we have @xmath560 3 . for any @xmath89 and @xmath561 we have @xmath562 in order to prove the claim , first note that @xmath563 it therefore follows that @xmath564 the calculation of @xmath565 follows easily from the definitions
this shows in particular that the @xmath566 are linearly independent and hence form a basis for @xmath557 finally it is straightforward to verify that @xmath562 this concludes the proof of the claim . for @xmath529 and @xmath567
we now let @xmath568 we write @xmath569 .
it follows immediately from the above claim that @xmath570 is a basis , and that @xmath570 is dual to @xmath543 .
we now consider the following short exact sequence of @xmath22complexes : @xmath571 together with the above bases .
it follows from theorem [ thm : mi66 ] ( 4 ) together with the above calculations and from the definition of torsion that @xmath572 where @xmath124 is the following complex with the canonical bases : @xmath573 it now follows ( cf .
e.g. ( * ? ? ?
* theorem 2.2 ) ) that @xmath574 this concludes the proof of the second statement of the theorem . finally note that @xmath575 ( see @xcite ) , in particular if @xmath52 factors through a @xmath58group , then it follows immediately from theorem [ thm : levine ] that @xmath576
this concludes the proof of part ( 3 ) of the theorem .
the following corollary follows immediately from taking @xmath94 to be the trivial matrix .
[ cor : unlink ] let @xmath10 be the @xmath0component unlink in @xmath46 with meridians @xmath55 .
let @xmath577 be an admissible homomorphism .
let @xmath33 be a subring closed under complex conjugation , and let @xmath15 be a unitary representation
. then @xmath60 and @xmath578 with @xmath30 and @xmath31 .
let @xmath10 be an @xmath0component boundary link .
we say @xmath2 is _ boundary slice _ if there exist disjointly embedded 3manifolds @xmath579 such that for @xmath365 , the boundary @xmath580 is the union of a seifert surface and a slice disk for @xmath546 .
it is known that a boundary slice link is boundary slice with respect to _ any _ boundary link seifert surface .
we note that boundary slice links are slice ; it is a long - standing open question whether the converse holds for boundary links .
obstructions to being boundary slice has been studied by several authors , including cappell and shaneson @xcite , duval @xcite , ko @xcite , mio @xcite , sheiham @xcite .
the following torsion obstruction to being boundary slice is a consequence of these works .
[ thm : boundaryslice ] let @xmath478 be an @xmath0-component @xmath361-link which is boundary slice .
let @xmath581 be a unitary representation where @xmath33 is a subring closed under complex conjugation .
let @xmath355 be an admissible homomorphism .
if @xmath405 , then @xmath582 for some @xmath30 and @xmath35 .
let @xmath583 be a seifert surface for @xmath2 corresponding to @xmath374 and let @xmath502 be the corresponding boundary link seifert matrix .
we denote by @xmath464 the genus of @xmath465 . by the first remark of this subsection
, there are disjointly embedded 3manifolds @xmath584 in @xmath585 such that @xmath586a slice disk for @xmath587 .
by ko ( * ? ? ?
* lemma 3.3 ) , it follows that @xmath502 is metabolic .
this means that for @xmath365 there exists an invertible @xmath588 matrix @xmath589 such that each @xmath590 is of the form @xmath591 where @xmath592 is a @xmath593-matrix .
( in fact ko @xcite has shown that in higher odd dimensions @xmath94 being metabolic is in fact a sufficient condition for being boundary slice . )
it follows easily that @xmath488 is equivalent to a matrix of the form @xmath594 where @xmath35 and @xmath77 and @xmath441 are matrices over @xmath23 $ ] of size @xmath595 .
it now follows that @xmath596 for some @xmath597 .
the theorem now follows immediately from theorem [ thm : tauboundary ] and the assumption that @xmath598 .
let @xmath2 be a boundary link . according to theorem [ mainthm ] the twisted torsion corresponding to any representation factoring through a @xmath58group gives a sliceness obstruction .
on the other hand , according to theorems [ thm : compboundary ] and [ thm : boundaryslice ] the twisted torsions corresponding to many more representations give an obstruction to being boundary slice .
this is very similar to the situation in @xcite where the vanishing of the eta invariant for ( boundary ) slice links is investigated , as well as @xcite ( cf .
* theorem 4.8 ) and ( * ? ? ?
* theorem 4.3 ) ) .
levine @xcite used work of vogel @xcite to resolve the apparent discrepancy in @xcite between the vanishing of the eta invariants for boundary links which are slice and those which are boundary slice .
let @xmath10 be an @xmath0component oriented link and let @xmath599 be a knot .
let @xmath600 be a simple closed curve , unknotted in @xmath46 .
then @xmath601 is a solid torus .
let @xmath602 be a diffeomorphism which sends a meridian of @xmath94 to a longitude of @xmath45 , and a longitude of @xmath94 to a meridian of @xmath45 .
the space @xmath603 is diffeomorphic to @xmath46 . the image of @xmath2 is denoted by @xmath604 .
we say @xmath605 is the _ satellite link with companion @xmath45 , orbit @xmath2 and axis @xmath94_. put differently , @xmath605 is the result of replacing a tubular neighborhood of @xmath45 by a link in a solid torus , namely by @xmath606 .
note that @xmath605 inherits an orientation from @xmath2 .
we now return to the discussion of satellite links in general .
note that the abelianization map @xmath609 gives rise to a map of degree one @xmath610 to @xmath611 which is a diffeomorphism on the boundary .
in particular we get an induced map @xmath612 which we denote by @xmath183 .
note that @xmath183 is a diffeomorphism on the boundary and that @xmath183 induces an isomorphism of homology groups .
[ lemmahofsat ] let @xmath614 and @xmath183 as above .
let @xmath68 be a subfield of @xmath17 and let @xmath615 be a unitary representation .
we denote the representation @xmath616 by @xmath52 as well .
let @xmath577 be an admissible homomorphism such that @xmath617 .
denote by @xmath618 the eigenvalues of @xmath619 and let @xmath620 be a fixed representative of the alexander polynomial of @xmath45 . then the following hold : [ corhofsat ] let @xmath2 be an oriented @xmath0component slice link and let @xmath626 and @xmath604 as above
. let @xmath68 be a subfield of @xmath17 closed under complex conjugation which has the unique factorization property and let @xmath495 be a representation which factors through a @xmath58group .
let @xmath577 be an admissible homomorphism .
denote by @xmath618 the eigenvalues of @xmath619 and let @xmath620 be a fixed representative of the alexander polynomial of @xmath45 .
assume that @xmath617 . if @xmath604 is slice , then @xmath622 for @xmath623 and @xmath627 for some @xmath30 and @xmath628 .
we write @xmath629 .
consider the following commutative diagram of short exact sequences of cellular chain complexes ( where we write @xmath630 ) : @xmath631 we assumed that @xmath617 .
it follows that the map @xmath36 restricted to @xmath632 and restricted to @xmath610 is trivial .
in particular @xmath633 and @xmath634 .
it is a consequence of ( * ? ? ? * section 5 ) that @xmath635 is always surjective and it is an isomorphism if and only if @xmath622 for @xmath623 .
the first statement of the lemma now follows immediately from the commutative diagram of long exact homology sequences corresponding to the above diagram .
we already saw above that @xmath635 is an isomorphism .
note that @xmath638 is homotopy equivalent to a 2complex and that @xmath632 is homotopy equivalent to 1complex .
in particular @xmath639 and it follows from the long exact sequence that @xmath640 as well . note that @xmath641 is surjective , it follows that @xmath642 is surjective as well .
on the other hand we have that @xmath643 , hence @xmath644 i.e. @xmath645 is in fact an isomorphism . this concludes the proof of the claim .
let @xmath646 be @xmath68bases for @xmath647 .
we endow @xmath648 with the corresponding basis @xmath649 .
it follows from the long exact sequence corresponding to @xmath650 that @xmath651 .
it follows from theorem [ thm : mi66 ] ( 4 ) that @xmath652 ( here and in the following lines all equalities of torsion are up to multiplication by an element of the form @xmath29 with @xmath653 and @xmath35 . )
note that the @xmath646 are also @xmath68bases for @xmath654 .
in particular the calculations in the previous lines also work for the torsions corresponding to twisting by @xmath655 .
let @xmath656 be a seifert matrix for the knot @xmath45 .
it follows easily from ( * ? ? ?
* section 5 ) and ( * ? ? ?
* theorem 2.2 ) that @xmath657 now pick a @xmath22basis @xmath328 for @xmath658 and denote the dual basis for @xmath659 by @xmath329 . finally pick bases @xmath660 for @xmath661 .
note that the bases @xmath662 give rise to bases @xmath663 for @xmath664 .
it now follows from theorem [ thm : mi66 ] ( 4 ) that @xmath665 ( here and in the next line the equalities of torsion are up to multiplication by an element of the form @xmath29 with @xmath653 and @xmath35 . ) combining these results we now obtain that @xmath666 this implies the desired equality .
we consider the bing double @xmath608 of a knot @xmath45 , in order to provide an example for a detailed computation .
note that if @xmath45 is a slice knot , then @xmath608 is in fact a slice link ( cf .
* section 1 ) ) .
the converse is a well known folklore conjecture . given any knot @xmath45 all the ` classical ' sliceness obstructions ( e.g.multivariable alexander polynomial , levine
tristram signatures ) of its bing double are trivial ( cf .
* section 1 ) for a beautiful survey ) .
it is therefore an interesting question to determine which methods and invariants detect the non - sliceness of bing doubles . for interesting results ,
we refer to cimasoni @xcite , harvey @xcite , cha @xcite , and cha - livingston - ruberman @xcite .
in particular in @xcite it was proved that if @xmath45 is not algebraically slice , @xmath608 is not slice .
also it is known that there are algebraically slice knots with non - slice bing doubles @xcite .
now let @xmath667 be the figure 8 knot .
it is well known that @xmath45 is not slice . for example
this follows from @xcite since the alexander polynomial @xmath668 is not a norm , namely not of the form @xmath669 . in what follows
we show that an appropriate twisted torsion invariant of the bing double @xmath670 is not a norm , and consequently @xmath670 is not slice .
( the fact that @xmath670 is not slice had first been shown in @xcite using the discriminant invariant . )
let @xmath671 be the meridians of the two component unlink @xmath2 and denote by @xmath94 the unknotted curve in the definition of bing doubles .
it is well - known that @xmath94 represents the commutator @xmath672 $ ] .
we write @xmath673 .
note that @xmath674 $ ] is a ufd ( cf .
* theorem 11.1 ) ) . consider the representation @xmath675,8)$ ] given by @xmath676 and @xmath677 it follows easily from proposition [ lem : ppowerrep ] that @xmath52 factors through a 2group .
we denote the representation @xmath678,8)$ ] by @xmath52 as well , and we let @xmath36 be the isomorphism @xmath679 .
( here @xmath667 . )
we compute the eigenvalues of @xmath680)$ ] to be @xmath681 we then calculate @xmath682 to equal @xmath683 .
if @xmath608 was slice , then by corollary [ corhofsat ] we would have @xmath684 for some @xmath30 and for some @xmath71 in the quotient field of @xmath674 $ ] .
note that @xmath685 is a positive real number and note that @xmath686 .
it therefore follows that if @xmath608 is slice , then @xmath687 .
since @xmath674 $ ] is a ufd we furthermore deduce that @xmath71 actually lies in @xmath674 $ ] .
now note that @xmath688 . since @xmath674 $ ] is a ufd we only have to show that @xmath689 is not a norm .
a direct calculation shows that any norm in @xmath674 $ ] is of the form @xmath690 for some @xmath691 .
one can now easily deduce that @xmath689 is not a norm in @xmath674 $ ] .
this concludes the proof of the claim .
note that the process of bing doubling can be iterated .
we refer to the work of the first author @xcite , the first author and kim @xcite , cochran , harvey and leidy @xcite and van cott @xcite for interesting results on iterated bing doubling .
j. levine , _ links invariants via the eta invariant _ , comm . math .
69 ( 1994 ) , 82119 .
j. levine , _ concordance of boundary links _
, journal of knot theory and its ramifications 16 ( 2007 ) , no .
9 , 11111120 .
d. sheiham , _ invariants of boundary link cobordism . the blanchfield - duval form _ , noncommutative localization in algebra and topology , lms lecture notes series 330 ( 2006 ) , cambridge university press , 143219 . | the twisted torsion of a 3-manifold is well - known to be zero whenever the corresponding twisted alexander module is non - torsion . under mild extra assumptions we introduce a new twisted torsion invariant which is always non - zero .
we show how this torsion invariant relates to the twisted intersection form of a bounding 4-manifold , generalizing a theorem of milnor to the non - acyclic case . using this result
, we give new obstructions to 3-manifolds being homology cobordant and to links being concordant .
these obstructions
are sufficiently strong to detect that the bing double of the figure eight knot is not slice . subjclassname@1991=subjclassname@2000= |
we estimate @xmath3 from a variety of simulated galaxy distributions , created by applying simple local and deterministic biasing algorithms to the mass distributions obtained from cosmological n - body simulations .
the * cosmological models * that we use are : ( 1 ) @xmath6 , cdm model with a tilted power spectrum designed to simultaneously satisfy both cobe and cluster normalization constraints .
( 2 ) @xmath7 , cdm , cluster normalized model . ( 3 ) @xmath8 , cdm , cluster normalized model .
cluster normalization requires that @xmath9 ( white , efstathiou & frenk 1993 ) .
the * biasing algorithms * that we use are : ( 1 ) semi - analytic : an empirical bias prescription derived by narayanan et al .
( 1999 , in prep . ) which characterizes the relation between the galaxy and mass density fields in the semi - analytic galaxy formation models of benson et al .
( 2 ) sqrt - exp .
( square - root exponential ) : a biasing prescription in which @xmath10 .
( 3 ) power - law : a biasing prescription in which @xmath11 . here ,
@xmath12 and @xmath13 are the galaxy and mass over - densities , respectively .
all the bias models are designed to yield galaxy distributions with an rms fluctuation , in top - hat spheres of radius @xmath14mpc , of @xmath15 .
the values of @xmath16 for the galaxy distributions are 0.6 , 0.58 , and 0.56 for @xmath6 , 0.4 , and 0.2 , respectively .
all the results have been averaged over four independent realizations of the cosmological models .
the simulation volumes are periodic cubes , with a box size of @xmath17mpc . the first method that we use to estimate @xmath3 involves the anisotropy of the redshift - space power spectrum .
kaiser ( 1987 ) and cole , fisher & weinberg ( 1994 ) showed that in linear theory , and under the plane - parallel approximation , the multipole moments of the redshift - space power spectrum may be used to estimate @xmath3 . here
we employ two different @xmath3 estimators : @xmath18 , the ratio of the angle - averaged redshift - space power spectrum ( monopole ) to the real - space power spectrum , and @xmath19 , the ratio of the quadrupole and monopole moments of the redshift - space power spectrum .
these ratios are , in principle , measurable , and they are related to @xmath3 as follows : @xmath20 , @xmath21 .
these ratios yield an estimate of @xmath3 at each wavenumber @xmath22 , and the scale dependence of this estimate is caused mainly by non - linearity in the peculiar velocity field .
it is possible to obtain a global estimate of @xmath3 by modeling this non - linearity .
we use two such models in our analysis : ( 1 ) exponential velocity distribution model ( cole , fisher & weinberg 1995 ) , where we assume that the galaxies have uncorrelated small scale peculiar velocities which are drawn from an exponential distribution .
we use this model in our analysis of @xmath18 ( the results of which are shown in fig .
( 2 ) empirical model ( hatton & cole 1999 ) , where the scale dependence of @xmath19 is found empirically by examining a large number of n - body simulations spanning a broad range of parameter space .
the second method that we use to estimate @xmath3 involves a direct comparison between the mass and galaxy density fields . the potent method ( bertschinger & dekel 1989 ) reconstructs the full three dimensional peculiar velocity field @xmath23 from the observed radial peculiar velocity field , allowing a measurement of @xmath3 directly from the linear theory relation @xmath24 .
we estimate @xmath3 using an idealized potent method , in which we have perfectly reconstructed the three dimensional peculiar velocity field .
we obtain the value of @xmath3 by fitting a line of the form @xmath25 to the @xmath26 vs. @xmath27 relation .
we perform this fit in the region : @xmath28 . by repeating this procedure for different smoothing of the velocity and density fields
, we can measure @xmath3 as a function of scale ( see fig .
figure 1 compares @xmath3 estimates from the two methods described previously to the function @xmath29 . here , @xmath5 is the bias function defined as @xmath30 , where @xmath31 and @xmath32 are the rms fluctuations of the galaxy and mass density fields , smoothed with a gaussian filter of radius @xmath33 , of the galaxy and mass distributions , respectively . in our full analysis , we also compare our @xmath3 estimates to the function @xmath34 . here
@xmath35 is the bias function defined as @xmath36 , where @xmath37 and @xmath38 are the power spectra of the galaxy and mass distributions , respectively . in the case of a simple , linear bias model , @xmath39 , all these definitions of @xmath2 are equivalent , i.e. , @xmath40 . in most cases ,
we find that the bias factor @xmath4 derived from redshift - space anisotropy or potent is similar to the large - scale value of @xmath35 or @xmath5 defined by rms fluctuation amplitudes . discrepancies at the 10 - 20% level may reflect non - linear dynamics as much as non - linearity of bias , since they also occur for unbiased models . in greater detail : ( 1 ) measurements of @xmath3 derived by fitting the cole et al .
( 1995 ) exponential velocity distribution model to @xmath18 underestimate the large - scale values of @xmath41 by about 10 - 20% for all cosmological models and biasing prescriptions .
( 2 ) measurements of @xmath3 derived by fitting the hatton & cole ( 1999 ) empirical model to @xmath19 accurately reproduce the large - scale values of @xmath41 for all cosmological models and biasing prescriptions .
( 3 ) measurements of @xmath42 derived from a potent - like comparison of @xmath26 , the negative divergence of the peculiar velocity field , to @xmath27 , the galaxy density field , underestimate the large - scale values of @xmath43 by about 10 - 20% for all cosmological models and biasing prescriptions , with the exception of the sqrt - exp . biasing prescription .
the sqrt - exp .
model gives erratic results because the non - linearity of the bias model makes the estimated @xmath3 sensitive to the range of @xmath27 used in the linear fit .
( 4 ) estimates of @xmath3 from the anisotropy of the redshift - space power spectrum approximately agree with those from the potent method on large scales , except in the case of the sqrt - exp . model .
benson , a. j. , cole , s. , frenk , c. s. , baugh , c. m. & lacey , c. g. 1999 , , submitted bertschinger , e. & dekel , a. 1989 , , 336l , 5 cole , s. , fisher , k. b. & weinberg , d. h. 1994 , , 267 , 785 cole , s. , fisher , k. b. & weinberg , d. h. 1995 , , 275 , 515 hatton , s. j. & cole , s. 1999 , , submitted kaiser , n. 1987 , 227 , 1 narayanan et al .
, in preparation white , s. d. m. , efstathiou , g. & frenk , c. s. 1993 , 262 , 1023 | measurements of the cosmological density parameter ( @xmath0 ) using techniques that exploit the gravity - induced motions of galaxies yield , in linear perturbation theory , the degenerate parameter combination @xmath1 , where the linear bias parameter @xmath2 is the ratio of the fluctuation amplitudes of the galaxy and mass distributions .
however , the relation between the mass and the galaxy density fields depends on the complex physics of galaxy formation , and it can in general be non - linear , non - local , and stochastic .
there is a growing consensus that the one - parameter model for bias is oversimplified .
this leaves us with the following question : what is the quantity that is actually being measured by the different techniques ? in order to address this question , we present estimates of @xmath3 from galaxy distributions constructed by applying a variety of biased galaxy formation models to cosmological n - body simulations .
we compare the values of @xmath3 estimated using two different techniques : the anisotropy of the redshift - space power spectrum and an idealized version of the potent method . in most cases , we find that the bias factor @xmath4 derived from redshift - space anisotropy or potent is similar to the large - scale value of @xmath5 defined by the ratio of rms fluctuation amplitudes of the galaxy and mass distributions .
however , non - linearity of bias and residual effects of non - linear gravitational evolution both influence @xmath3 estimates at the 10 - 20% level . |
the study of identified hadron spectra at large transverse momentum ( @xmath0 ) in @xmath1+@xmath1 collisions can be used to test the predictions from perturbative quantum chromodynamics ( pqcd ) @xcite . in the framework of models based on qcd
, the inclusive production of single hadrons is described by the convolution of parton distribution functions ( pdfs ) , parton interaction cross - sections and fragmentation functions ( ffs ) .
the pdf provide the probability of finding a parton ( a quark or a gluon ) in a hadron as a function of the fraction of the hadron s momentum carried by the parton .
the ffs @xcite give the probability for a hard scattered parton to fragment into a hadron of a given momentum fraction .
these are not yet calculable from the first principles and hence are generally obtained from experimental data ( e.g. , @xmath10+@xmath11 collisions ) .
the factorization theorem for cross - sections assumes that ffs are independent of the process in which they have been determined and hence represent a universal property of hadronization .
it is therefore possible to make quantitative predictions for other types of collision systems ( e.g. , @xmath1+@xmath1 ) .
comparisons between experimental data and theory can help to constrain the quark and gluon ffs that are critical to predictions of hadron spectra in @xmath1+@xmath1 , @xmath1+a , and a+a collisions .
the simultaneous study of identified hadron @xmath0 spectra in @xmath1+@xmath1 and @xmath2+au collisions may also provide important information on the pdfs @xcite of the nucleus . the identified particle spectra in @xmath1+@xmath1 and @xmath2+au collisions also provide reference spectra for particle production at high @xmath0 in au+au collisions .
moreover , studies of identified particle production and their ratios as a function of @xmath0 in high - energy heavy - ion collisions have revealed many unique features in different @xmath0 regions @xcite and between baryons and mesons @xcite .
a good description of both identified pion and proton spectra in @xmath1+@xmath1 and @xmath2+au collisions at intermediate and high @xmath0 by nlo pqcd will provide a solid ground for models based on jet quenching @xcite and quark recombination @xcite .
these emphasize the need for a systematic study of @xmath0 spectra from @xmath1+@xmath1 and @xmath2+au collisions at the same energy as the nucleus - nucleus collisions . in this letter
, we present the @xmath0 spectra for identified pions , protons and anti - protons in @xmath1+@xmath1 and @xmath2+au collisions at @xmath3 = 200 gev as measured by the star experiment at rhic .
the results are compared to nlo pqcd calculations and a phenomenological model .
we also study the @xmath8-scaling in @xmath1+@xmath1 collisions and the nuclear modification factors in @xmath2+au collisions .
the star experiment consists of several detectors to measure hadronic and electromagnetic observables spanning a large region of the available phase space at rhic .
the detectors used in the present analysis are the time projection chamber ( tpc ) , the time - of - flight ( tof ) detector , a set of trigger detectors used for obtaining the minimum bias data , and the forward time projection chamber for the collision centrality determination in @xmath2+au collisions .
the details of the design and other characteristics of the detectors can be found in ref .
@xcite .
a total of 8.2 million minimum bias @xmath1+@xmath1 collision events and 11.7 million @xmath2+au collision events have been analyzed for the present study .
the data set was collected during the years 2001 and 2003 .
the details of minimum bias trigger conditions for @xmath1+@xmath1 and @xmath2+au collisions can be found in the refs .
the minimum - bias trigger captured @xmath12 of the @xmath13 barn @xmath2+au inelastic cross - section .
the trigger efficiency was determined from a cross study of two sets of trigger detectors : two zero - degree calorimeters ( zdcs ) and two beam - beam counters ( bbcs ) .
the absolute cross - section is derived from a monte carlo glauber calculation .
these results are consistent with other recent measurements @xcite .
the trigger for the minimum bias @xmath1+@xmath1 collisions required a coincidence measurement of the two bbcs covering 3.3 @xmath6 @xmath14 @xmath6 5.0 @xcite .
this trigger was sensitive to color exchange hadronic and doubly - diffractive events ; here , these are labelled `` non - singly diffractive ( nsd ) events '' . using pythia(v6.205 )
@xcite and herwig @xcite , it was determined that the trigger measured 87% of the @xmath15 mb nsd cross - section , which was measured via a vernier scan @xcite .
distribution normalized by pion @xmath16 at 4.5 @xmath6 @xmath0 @xmath6 5.0 gev/@xmath7 and @xmath14 @xmath6 0.5 , and shifted by @xmath175 for positively and negatively charged particles , respectively .
the distributions are for minimum bias @xmath2+au collisions .
the pion , proton and anti - proton peak positions are indicated by arrows . ] the data from tof are used to obtain the identified hadron spectra for @xmath18 gev/@xmath7 . the procedure for particle identification in tof
has been described in ref .
@xcite . for @xmath19
gev/@xmath7 , we use data from the tpc . particle identification at high @xmath0 in the tpc comes from the relativistic rise of the ionization energy loss ( @xmath16 ) .
details of the method are described in ref .
@xcite . at @xmath20 gev/@xmath7 ,
the pion @xmath16 is about 1020% higher than that of kaons and protons due to the relativistic rise , resulting in a few standard deviations ( 1 - 3@xmath21 ) separation between them .
since pions are the dominant component of the hadrons in @xmath1+@xmath1 and @xmath2+au collisions at rhic , the prominent pion peak in the @xmath16 distribution is fit with a gaussian to extract the pion yield @xcite .
the proton yield is obtained by integrating the entries ( @xmath22 ) in the low part of the @xmath16 distribution about @xmath23 away from the pion @xmath16 peak .
the integration limits were varied to check the stability of the results .
[ fig0 ] shows a typical @xmath16 distribution normalized by the pion @xmath16 at 4.5 @xmath6 @xmath0 @xmath6 5.0 gev/@xmath7 and @xmath14 @xmath6 0.5 .
the gaussian distribution used to extract the pion yield and the pion , proton and anti - proton peak positions are also shown in the figure .
the kaon contamination is estimated via either of the equations given below .
the uncorrected proton yield is @xmath24 or @xmath25 where @xmath26 and @xmath27 are the proton and kaon efficiencies from the integration described above , derived from the @xmath16 calibration , resolution and the bichsel function @xcite . in the first case
the kaon contamination is estimated through the yields of the inclusive hadrons ( @xmath28 ) and pions , in case two from known yields from @xmath29 measurements @xcite .
the typical values of @xmath26 for a de / dx cut slightly away from the proton peak position is 0.4 and the @xmath27 values decrease from 0.2 to 0.08 with @xmath0 in the range 2.5 @xmath6 @xmath0 @xmath6 10 gev/@xmath7 . at high @xmath0 , the yields of other stable particles ( i.e. , electrons and deuterons ) are at least two orders of magnitude lower than those of pions , and are negligible in our studies .
the two results are consistent where star @xmath29 measurements are available .
the @xmath0-dependence of the reconstruction efficiency , background and the systematic uncertainties for pions , protons and anti - protons for low @xmath0 in @xmath1+@xmath1 and @xmath2+au collisions are described in ref .
@xcite . at high @xmath0 ( @xmath30 2.5 gev/@xmath7 ) ,
the efficiency is almost independent of @xmath0 in both @xmath1+@xmath1 and @xmath2+au collisions .
the tracking efficiencies are @xmath31 88% and 92% in @xmath1+@xmath1 and @xmath2+au collisions , respectively .
the difference in tracking efficiency arises because of worse vertex determination in @xmath1+@xmath1 collisions than @xmath2+au collisions .
the background contamination to pion spectra for @xmath0 @xmath30 2.5 gev/@xmath7 , primarily from @xmath29 weak decay is estimated from pythia / hijing simulations with full geant detector descriptions to be @xmath31 4% .
the charged pion spectra are corrected for efficiency and background effects .
the inclusive proton and anti - proton spectra are presented with efficiency corrections and without hyperon feed - down corrections .
the integrated @xmath32/@xmath1-ratio is estimated to be @xmath6 25% @xcite .
additional corrections are applied for primary vertex reconstruction inefficiency as discussed in refs .
the momentum resolution is given as @xmath33/@xmath0 = 0.01 + 0.005@xmath0/(gev/@xmath7 ) and has @xmath6 4% effect on the yields at the highest @xmath0 value .
the spectra are not corrected for momentum resolution effects , but they are included in the systematic errors . the total systematic uncertainties associated with pion yields are estimated to be @xmath34 15% .
this systematic uncertainty is dominated by the uncertainty in modeling the detector response in the monte carlo simulations .
protons from hyperon ( @xmath32 and @xmath35 ) decays away from the primary vertex can be reconstructed as primordial protons at a slightly higher @xmath0 than their true value , but with worse momentum resolution .
this results in an uncertainty of the inclusive proton yield of @xmath36 at @xmath0 = 3 gev/@xmath7 and @xmath37 at @xmath0 = 10 gev/@xmath7 . for proton and anti - proton yields at high @xmath0
an additional systematic error arises from the uncertainties in the determination of the efficiencies , @xmath26 and @xmath27 , under a specific @xmath16 selection for integration .
this is due to the uncertainties in the mean @xmath16 positions for protons and kaons . the total systematic uncertainty in obtaining the proton and anti - proton yields for @xmath0 @xmath30 2.5 gev/@xmath7 increases with @xmath0 from 12% to 23% ( at @xmath0 = 10 gev / c ) in both @xmath1+@xmath1 and @xmath2+au collisions .
the errors shown in the figures are statistical , and the systematic errors are plotted as shaded bands . in addition , there are overall normalization uncertainties from trigger and luminosity in @xmath1+@xmath1 and @xmath2+au collisions of 14% and 10% , respectively @xcite .
these errors are not shown .
@xmath5@xmath4 @xmath6 0.5 ) transverse momentum spectra for charged pions , proton and anti - proton in @xmath1+@xmath1 and @xmath2+au collisions for various event centrality classes .
minimum bias distributions are fit to levy functions which are shown as dashed curves . ]
figure [ fig1 ] shows the invariant yields of charged pions , protons and anti - protons for the @xmath0 range of 0.3 @xmath6 @xmath0 @xmath6 10 gev/@xmath7 in minimum bias @xmath1+@xmath1 collisions and for various centrality classes in @xmath2+au collisions .
the yields span over eight orders of magnitude .
the minimum bias distributions are fit with a levy distribution @xcite of the form @xmath38 , where @xmath39 = @xmath40 and @xmath41 is the mass of the hadron .
the levy distribution essentially takes a power - law form at higher @xmath0 and has an exponential form at low @xmath0 . for the @xmath1 and @xmath42 spectra ,
fit with a power - law function gives a worse @xmath43 compared to the fit with the levy function . for @xmath2+au collisions the @xmath43 for the power - law fit to @xmath1(@xmath42 )
spectra is 68.55/20(86.77/20 ) and the corresponding value for the fit with the levy function is 21.19/20(26.4/20 ) .
the nuclear modification factor ( @xmath44 ) can be used to study the effects of cold nuclear matter on particle production .
it is defined as a ratio of the invariant yields of the produced particles in @xmath2+au collisions to those in @xmath1+@xmath1 collisions scaled by the underlying number of nucleon - nucleon binary collisions .
@xmath45 where @xmath46 is the average number of binary nucleon - nucleon ( nn ) collisions per event , and @xmath47 is the nuclear overlap function @xmath48 @xcite .
the value of @xmath49 is taken to be 42 mb . , for charged pions @xmath50 and @xmath1+@xmath42 at @xmath4@xmath5@xmath4 @xmath6 0.5 in minimum bias and 0 - 20% central @xmath2+au collisions . for comparison results on inclusive charged hadrons ( star ) from ref .
@xcite at @xmath4@xmath51@xmath4 @xmath6 0.5 are shown by dashed curves .
the first two shaded bands around @xmath52 correspond to the error due to uncertainties in estimating the number of binary collisions in minimum bias and 0 - 20% central @xmath2+au collisions respectively .
the last shaded band is the normalization uncertainty from trigger and luminosity in @xmath1+@xmath1 and @xmath2+au collisions . ]
the left panel of fig .
[ fig2 ] shows @xmath44 values for charged pions ( ( @xmath53+@xmath54)/2 ) in minimum bias and 0 - 20% central collisions at @xmath4y@xmath4 @xmath6 0.5 .
the @xmath44 for 0 - 20% central collisions are higher than @xmath44 for minimum bias collisions .
the result @xmath44 @xmath30 1 indicates a slight enhancement of high @xmath0 charged pion yields in @xmath2+au collisions compared to binary collision scaled charged pion yields in @xmath1+@xmath1 collisions within the measured ( @xmath5 , @xmath0 ) range .
the right panel of fig .
[ fig2 ] shows the @xmath55 of baryons ( @xmath1+@xmath42 ) for the minimum bias collisions at @xmath4y@xmath4 @xmath6 0.5 .
the @xmath44 for @xmath56 is again greater than unity for @xmath0 @xmath30 1.0 gev/@xmath7 and is larger than @xmath44 for charged pions .
the @xmath44 of pions for 2@xmath6 @xmath0 @xmath6 5 gev/@xmath7 is 1.24 @xmath17 0.13 and that for @xmath1+@xmath42 is 1.49 @xmath17 0.17 in minimum bias collisions .
identified hadron @xmath44 are sensitive to nuclear modification of the pdf from processes such as nuclear shadowing and parton saturation as well as to transverse momentum broadening , energy loss in cold nuclear matter and hadronization through recombination , thereby further constraining the models @xcite .
, @xmath57 , @xmath58 , @xmath59 at midrapidity ( @xmath4@xmath5@xmath4 @xmath6 0.5 ) as a function of @xmath0 in @xmath1+@xmath1 minimum bias collisions . for comparison the results from lower energies at isr @xcite and fnal @xcite
are also shown for @xmath58 and @xmath59 ratios .
the dotted curves are the results from pythia .
the shaded bands below the @xmath60 and @xmath57 ratios are the point to point correlated errors in the yields associated with the ratio . ] the particle ratios at midrapidity as a function of @xmath0 for @xmath1+@xmath1 and @xmath2+au minimum bias collisions are shown in figs . [ fig3a ] and [ fig3b ] respectively .
correlated errors are shown as the shaded bands below the data points .
the @xmath60-ratio has a value @xmath31 1 and is independent of @xmath0 in both @xmath1+@xmath1 and @xmath2+au collisions .
the @xmath57-ratio for @xmath1+@xmath1 collisions is also independent of @xmath0 within the range studied and has a value of 0.81 @xmath17 0.1 at 2.5 @xmath6 @xmath0 @xmath6 6.5 gev/@xmath7 .
however , in @xmath2+au collisions we observe a clear decrease of @xmath57 for @xmath616 gev/@xmath7 . in quark fragmentation
, the leading hadron is more likely to be a particle rather than an anti - particle , and there is no such preference from a gluon jet .
a decrease in the antiparticle / particle ratio with @xmath0 would then indicate a significant quark jet contribution to the baryon production .
it is , however , not clear whether the same effect exists in @xmath1+@xmath1 collisions or whether the decrease of @xmath57 is due to additional nuclear effects in @xmath2+au collisions .
calculations from pythia(v6.319 ) predict somewhat more prominent @xmath0-dependence @xcite . for d+au minimum bias collisions . for comparison the @xmath58-ratio from lower energies at fnal @xcite are shown . ] at rhic , the @xmath58 and @xmath59 ratios increase with @xmath0 up to 2 gev/@xmath7 and then start to decrease for higher @xmath0 in both @xmath1+@xmath1 and @xmath2+au collisions .
the @xmath59-ratio rapidly approaches a value of 0.2 , which is between the values in @xmath10+@xmath11 collisions for quark and gluon jets @xcite .
the @xmath58 and @xmath59 ratios from pythia are constant at high @xmath0 in contrast to a decreasing trend oberserved in the data . the @xmath58-ratios in @xmath1+@xmath1 collisions compare well with results from lower energy isr and fnal fixed target experiments @xcite .
meanwhile , @xmath59-ratios at high @xmath0 have a strong energy dependence with larger values at higher beam energies . in @xmath2+au collisions the @xmath58-ratio at high @xmath0 is lower for @xmath1+a collisions at fnal energy than at rhic .
+ @xmath54)/2 and ( @xmath1+@xmath42)/2 at high @xmath0 for minimum bias @xmath1+@xmath1 and @xmath2+au collisions compared to results from nlo pqcd calculations using kkp @xcite ( pdf : cteq6.0 ) and akk @xcite ( pdf : cteq6 m ) sets of fragmentation functions and results from the epos model @xcite .
the pdfs for @xmath2 and au - nucleus are taken from refs .
@xcite and @xcite respectively .
all results from nlo pqcd calculations are with factorization scale is @xmath62 = @xmath0 . ,
title="fig : " ] + @xmath54)/2 and ( @xmath1+@xmath42)/2 at high @xmath0 for minimum bias @xmath1+@xmath1 and @xmath2+au collisions compared to results from nlo pqcd calculations using kkp @xcite ( pdf : cteq6.0 ) and akk @xcite ( pdf : cteq6 m ) sets of fragmentation functions and results from the epos model @xcite .
the pdfs for @xmath2 and au - nucleus are taken from refs .
@xcite and @xcite respectively .
all results from nlo pqcd calculations are with factorization scale is @xmath62 = @xmath0 . , title="fig : " ] in fig .
[ fig4 ] we compare ( @xmath53 + @xmath54)/2 and ( @xmath1+@xmath42)/2 yields in minimum bias @xmath1+@xmath1 and @xmath2+au collisions at midrapidity for high @xmath0 to those from nlo pqcd calculations and the phenomenological parton model ( epos ) @xcite .
the results from epos agree fairly well with our data for charged pions and proton+anti - proton in @xmath1+@xmath1 and @xmath2+au collisions .
the nlo pqcd results are based on calculations performed with two sets of ffs , the _ kniehl - kramer - potter ( kkp ) _
@xcite and the _ albino - kniehl - kramer ( akk ) _ set of functions @xcite .
the factorization scale for all the nlo pqcd calculations shown is for @xmath62 = @xmath0 .
the charged pion data for @xmath0 @xmath30 2 gev/@xmath7 in @xmath1+@xmath1 collisions are reasonably well described by the nlo pqcd calculations using the kkp and akk set of ffs .
a similar observation for @xmath63s using kkp ffs was made by the phenix collaboration @xcite . for @xmath2+au collisions nlo pqcd calculations with kkp ffs
are consistent with the data for @xmath0 @xmath30 4 gev/@xmath7 while those with akk ffs underpredict the measured charged pion yields .
the proton+anti - proton yield at high @xmath0 in @xmath1+@xmath1 and @xmath2+au collisions is much higher than the results from nlo pqcd calculations using the kkp set of ffs and lower compared to calculations using akk ffs .
the relatively better agreement of nlo pqcd calculations with akk ffs compared to those with kkp ffs for proton+anti - proton yields shows the importance of the flavor - specific measurements in @xmath10+@xmath11 collisions in determining the ffs for baryons .
one may further improve the nlo pqcd calculations by an all - order resummation of large logarithmic corrections to the partonic cross - sections @xcite .
the invariant cross - sections of inclusive pion production in high energy @xmath1+@xmath1 collisions have been found to follow the scaling laws @xcite : @xmath64 where @xmath8 = 2@xmath65 and @xmath66 and @xmath67 are some functions of @xmath8 .
similar scaling has been observed in @xmath10+@xmath11 collisions , but without the @xmath68 or @xmath69 factor @xcite . the value of the power @xmath70 ranges from 4 to 8 @xcite . in the general scaling form @xmath31 1/@xmath71 ,
@xmath70 depends on the quantum exchanged in the hard scattering . in parton models , it is related to the number of point - like constituents taking an active role in the interaction .
the value reaches 8 in the case of a quark - meson scattering by exchanging a quark . with the inclusion of qcd
, the scaling law follows as @xmath31 1/@xmath68 , where @xmath70 becomes a function of @xmath8 and @xmath72 .
the value of @xmath70 depends on the evolution of the structure function and ffs .
-scaling of pions , protons and anti - protons .
the data from other experiments are from the following references , fnal : refs @xcite , isr : ref .
@xcite , phenix : ref .
@xcite , and ua2 @xcite .
the inset shows the @xmath73-scaling of the invariant yields for charged pions and protons+anti - protons in @xmath1+@xmath1 and @xmath2+au collisions . ]
@xmath70=4 is expected in more basic scattering processes ( as in qed ) @xcite .
figure [ fig5 ] shows the @xmath8-scaling of pions , protons and anti - protons .
the value of @xmath70 obtained for the scaling with @xmath68 of the invariant cross - section is 6.5 @xmath17 0.8 .
the star data covers the range 0.003 @xmath6 @xmath8 @xmath6 0.1 .
the data points deviate from the scaling behavior for @xmath74 gev/@xmath7 for pions and protons , which could be interpreted as a transition region from soft to hard processes in the particle production .
the deviations start at a higher @xmath0 for the anti - protons . the available data on pion and proton invariant cross - sections at various center - of - mass energies @xcite for @xmath0 @xmath30 2 gev/@xmath7 are compiled and fitted using the function @xmath75 .
the value of @xmath70 ranges from 6.0 to 7.3 for @xmath3 between 19 gev and 540 gev , while that for @xmath76 ranges between 13 and 22 .
the average value of @xmath70 for pions is 6.8 @xmath17 0.5 and that for protons and anti - protons is 6.5 @xmath17 1.0 .
the variations in @xmath70 and @xmath76 values may lead to differences in details of scaling behaviour at different energies when the cross - section is multiplied by @xmath77 @xcite .
this feature is not observed in the scaling shown in fig .
[ fig5 ] due to the data spanning several orders of magnitude .
the inset of fig .
[ fig5 ] shows the @xmath73 scaling at @xmath74 gev/@xmath7 , consistent with possible transition between soft and hard processes at around @xmath78 gev/@xmath7 .
the @xmath73-scaling also indicates that flow effects in @xmath1+@xmath1 and @xmath2+au collisions are negligible @xcite .
the presented data suggests that the transition region from soft to hard physics occurs around @xmath0@xmath792 gev/@xmath7 in @xmath1+@xmath1 collisions .
we have presented transverse momentum spectra for identified charged pions , protons and anti - protons from @xmath1+@xmath1 and @xmath2+au collisions at @xmath3 = 200 gev .
the transverse momentum spectra are measured around midrapidity ( @xmath4@xmath5@xmath4 @xmath6 0.5 ) over the range of 0.3 @xmath6 @xmath0 @xmath6 10 gev/@xmath7 with particle identification from the ionization energy loss and its relativistic rise in the time projection chamber , as well as the time - of - flight in star .
the following conclusions can be drawn from the present study : ( a ) the nuclear modification factor around midrapidity is enhanced in @xmath2+au collisions to about 1.5 for pions and to about 2 for protons and antiprotons at intermediate @xmath0 ( 2 @xmath6 @xmath0 @xmath6 5 gev/@xmath7 ) .
( b ) identified particle ratios were measured up to @xmath0 of 7 gev/@xmath7 in @xmath1+@xmath1 and 10 gev/@xmath7 in @xmath2+au reactions .
their dependence on species , @xmath0 and collisions energy was shown to be sensitive to the relative contributions from quark and gluon fragmentation as well as to their fragmentation functions .
( c ) the nlo pqcd calculations describe the high @xmath0 data for charged pions reasonably well in @xmath1+@xmath1 collisions and @xmath2+au collisions . in general
, baryon production has historically been difficult to describe by pqcd and hadronization @xcite .
use of the recently published akk ffs results in a much improved description of the measured @xmath1 and @xmath42 spectra .
( d ) the proton and pion spectra in @xmath1+@xmath1 collisions follow @xmath8-scaling with a beam - energy dependent factor @xmath31 @xmath80 above @xmath0@xmath31 2 gev/@xmath7 .
the pion and proton spectra follow transverse mass scaling for @xmath73 @xmath6 2 gev/@xmath81 in both @xmath1+@xmath1 and @xmath2+au collisions , suggesting the transition region from soft to hard process domination occurs at @xmath0@xmath31 2 gev/@xmath7 in these collision systems .
the measurements presented in this paper provide better constraints on jet quenching and quark recombination models which are presently the best candidates for explaining particle production in the intermediate @xmath0 region .
we would like to thank simon albino , stefan kretzer and werner vogelsang for providing us the nlo pqcd results , klaus werner for the epos results and j. raufeisen for useful discussions .
we thank the rhic operations group and rcf at bnl , and the nersc center at lbnl for their support .
this work was supported in part by the henp divisions of the office of science of the u.s .
doe ; the u.s .
nsf ; the bmbf of germany ; in2p3 , ra , rpl , and emn of france ; epsrc of the united kingdom ; fapesp of brazil ; the russian ministry of science and technology ; the ministry of education and the nnsfc of china ; irp and ga of the czech republic , fom of the netherlands , dae , dst , and csir of the government of india ; swiss nsf ; the polish state committee for scientific research ; staa of slovakia , and the korea sci . & eng . foundation .
j. c. collins and d. e. soper , ann .
nucl . part .
37 ( 1987 ) 383 ; j. c. collins , d. e. soper and g. sterman in perturbative quantum chromodynamics ( world scientific , 1989 ) edited by a. h. muller and adv .
high energy phys . 5 ( 1988 ) 1 ; g. sterman et al . ,
67 ( 1995 ) 157 .
d. e. soper , nucl .
b ( proc . suppl . ) 53 ( 1997 ) 69 ; a. d. martin , w. j. stirling and r.g .
roberts , phys .
b 354 ( 1995 ) 155 ; m. gluck , e. reya and a. vogt , z. phys .
c 67 ( 1995 ) 433 ; h. l. lai et al .
d 51 ( 1995 ) 4763 .
brahms collaboration , i. arsene et al .
phys . a 757 ( 2005 ) 1 ; phobos collaboration , b.b .
back et al . , nucl .
phys . a 757 ( 2005 ) 28
; star collaboration , j. adams et al .
phys . a 757 ( 2005 ) 102 ; phenix collaboration , k. adcox et al . ,
phys . a 757 ( 2005 ) 184 .
star collaboration , c. adler et al .
( 2003 ) 082302 ; phenix collaboration , s.s .
adler et al . , phys .
91 ( 2003 ) 072301 .
star collaboration , j. adams et al .
92 ( 2004 ) 052302 .
wang and m. gyulassy , phys .
68 ( 1992 ) 1480 .
r. vogt , phys .
c 70 ( 2004 ) 064902 ; j. qiu and i. vitev , arxiv : hep - ph/0405068 ; d. kharzeev , y. kovchegov and k. tuchin , phys .
b 599 ( 2004 ) 23 ; r. c. hwa and c.b .
yang , phys .
93 ( 2004 ) 082302 .
e706 collaboration , l. apanasevich et al .
d 68 ( 2003 ) 052001 ; p.
b. straub et al . , phys .
d 45 ( 1992 ) 3030 .
k. werner , f. liu and t. pierog , arxiv : hep - ph/0506232 .
the epos model includes two important nuclear effects through elastic and inelastic parton ladder splitting .
the elastic splitting is related to screening and saturation while the inelastic splitting is related to the hadronization process .
the proton+anti - proton spectra from epos model is not corrected for the hyperon feed - down corrections .
b. a. kniehl , g. kramer and b. potter , nucl .
b 597 ( 2001 ) 337 .
the kkp ffs are obtained from @xmath10+@xmath11 collision data where the observed hadron is identified as a @xmath82 or @xmath83 or @xmath1(@xmath42 ) and the emitting parton is identified as either a gluon , light ( @xmath84,@xmath2 , and @xmath85 ) quark , @xmath7 quark or @xmath86 quark .
s. albino , b. a. kniehl and g. kramer , nucl .
b 725 ( 2005 ) 181 .
the akk ffs use the recent data from the opal collaboration @xcite on light - flavor separated measurements of light charged hadron production in @xmath10+@xmath11 collisions , thereby allowing the extraction of the flavor - dependent ffs of light quarks .
b. a. kniehl and l. zwirner , nucl .
b 637 ( 2002 ) 311 . for nlo pqcd calculations in @xmath2+au collisions ,
the parton distributions for the deuteron are obtained via the average of the distributions in protons and neutrons ( ( @xmath1+@xmath70)/2 ) .
l. frankfurt , v. guzey and m. strikman , phys . rev .
d 71 ( 2005 ) 054001 ; d. de florian and r. sassot , phys .
d 69 ( 2004 ) 074028 .
the parton distributions for gold nuclei are obtained by convoluting the free nucleon parton densities with simple weight functions that are meant to parametrize nuclear effects .
these are further constrained by comparing to deep inelastic scattering data off nuclei .
phenix collaboration , s.s .
adler et al , phys .
91 ( 2003 ) 241803 .
a. g. clark et al .
b 74 ( 1978 ) 267 ; ccor collaboration , a. l. s. angelis et al . ,
b 79 ( 1978 ) 505 ; phenix collaboration , s.s . adler et al . , phys .
c 69 ( 2004 ) 034910 .
tpc collaboration , h. aihara et al .
61 ( 1988 ) 1263 ; aleph collaboration , d. buskulic et al . , z. phys .
c 66 ( 1995 ) 355 ; argus collaboration , h. albrecht et al .
, z. phys .
c 44 ( 1989 ) 547 .
s. m. berman , j. d. bjorken and j. b. kogut , phys .
d 4 ( 1971 ) 3388 ; r. blankenbecler , s. j. brodsky and j.f .
gunion , phys .
b 42 ( 1972 ) 461 ; j.f .
owens , e. reya , m. gluck , phys .
d 18 ( 1978 ) 1501 . | we present the transverse momentum ( @xmath0 ) spectra for identified charged pions , protons and anti - protons from @xmath1+@xmath1 and @xmath2+au collisions at @xmath3 = 200 gev .
the spectra are measured around midrapidity ( @xmath4@xmath5@xmath4 @xmath6 0.5 ) over the range of 0.3 @xmath6 @xmath0 @xmath6 10 gev/@xmath7 with particle identification from the ionization energy loss and its relativistic rise in the time projection chamber and time - of - flight in star .
the charged pion and proton+anti - proton spectra at high @xmath0 in @xmath1+@xmath1 and @xmath2+au collisions are in good agreement with a phenomenological model ( epos ) and with next - to - leading order perturbative quantum chromodynamic ( nlo pqcd ) calculations with a specific fragmentation scheme and factorization scale .
we found that all proton , anti - proton and charged pion spectra in @xmath1+@xmath1 collisions follow @xmath8-scaling for the momentum range where particle production is dominated by hard processes ( @xmath0 @xmath9 2 gev/@xmath7 ) .
the nuclear modification factor around midrapidity is found to be greater than unity for charged pions and to be even larger for protons at 2 @xmath6 @xmath0 @xmath6 5 gev/@xmath7 . , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
, and ( star collaboration ) particle production , perturbative quantum chromodynamics , fragmentation function , cronin effect and @xmath8-scaling . |
though @xmath0 violation has only been observed in the neutral kaon system , both the standard model ( sm ) and its many extensions predict large @xmath0-violating effects in @xmath2 decays .
one of the main programs at the @xmath2 factories is to test the sm cabibbo - kobayashi - maskawa ( ckm ) paradigm of @xmath0 violation through , for example , measurement of the angles of unitarity triangle . in particular ,
the angle @xmath4 can be cleanly related within the sm to the @xmath0 asymmetry in the gold - plated " decay of @xmath3 .
the current sm fit already places quite a nontrivial constraint on this angle : @xmath5 @xcite .
therefore direct experimental measurement of the angles of the unitarity triangle with values in gross disagreement with the sm expectation will be a clear signal for new sources of @xmath0 violation .
it is thus important to explore the pattern of @xmath0 violation in @xmath2 decays in different models of @xmath0 violation . two higgs doublet models ( 2hdm ) without natural flavor conservation @xcite ( nfc ) , also called model iii @xcite , contain in general both tree - level flavor - changing neutral higgs ( fcnh ) couplings and new @xmath0-violating phases beside the ckm phase .
the phenomenology of this class of models thus differs significantly from that of the much studied 2hdm s with nfc ( _ i.e. , _ models i and ii @xcite ) where neither tree level fcnh couplings nor new @xmath0 violating phases are present . it is worth noting that the abandonment of nfc is not against any fundamental principle and in fact it opens up many possibilities for flavor physics . depending on the ansatz for fcnh couplings , type iii 2hdm need not run into problems with the low energy data such as @xmath1 mixing . in this talk , we present a @xmath0 violation study of a specific type iii 2hdm introduced in @xcite where the top quark is treated differently than the other quarks .
as will be shown , the charged higgs sector of the model could lead to a distinctive pattern of @xmath0 violation in @xmath2 decays even after imposing all the experimental constraints .
the top - quark 2hdm ( t2hdm ) @xcite distinguishes itself from 2hdm s with nfc by the special status it gives to the top quark , as evidenced in the yukawa couplings , @xmath6 where @xmath7 @xmath8 , and where @xmath9 , @xmath10 and @xmath11 are @xmath12 matrices in generation space ; @xmath13 ; @xmath14 ; and @xmath15 and @xmath16 are the usual left - handed quark and lepton doublets .
note that among the right - handed quarks , only @xmath17 couples to @xmath18 , and that the mass hierarchy between the top quark and the other quarks is understood as a result of the hierarchy between the vev s of @xmath18 and @xmath19 .
this form of the yukawa interactions can be seen as a consequence of some discrete symmetry . denoting the ratio of the two vev s by @xmath20 ,
the model then requires a large @xmath21 in accordance with the large mass ratio of the top and bottom quarks , @xmath22 . in our analysis
, we will always take @xmath23 . the absence of nfc in the model results in tree level fcnh couplings among the up - type quarks , whose contribution to @xmath24 mixing is dependent on the mixing among the right - handed up - type quarks@xcite .
for the charged higgs sector , we obtain the following yukawa lagrangian @xcite , @xmath25 + \overline{u}_r m_u v d_l\left[g^+ -\tan\beta h^+\right ] \right .
\nonumber \\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\ ; \left .
+ \overline{u}_r \sigma^\dagger v d_l\left[\tan\beta + \cot\beta\right]h^+ + { \rm h.c.}\right\ } , \label{eq : chiggs}\end{aligned}\ ] ] where @xmath26 and @xmath27 represent the would - be goldstone bosons and the physical charged higgs bosons , respectively . here
@xmath28 and @xmath29 are the diagonal up- and down - type quark mass matrices , @xmath30 is the usual ckm matrix , and @xmath31 with @xmath32 being the unitary rotation that brings the right - handed up - type quarks from gauge eigenstates to mass eigenstates . unlike 2hdm s with nfc ,
the t2hdm depends on the @xmath32 rotation . for our analysis , we take the simple parameterization for @xmath32 @xcite , @xmath33 where @xmath34 and @xmath35 is a complex number of order unity . the @xmath36 matrix
then depends only on the unknown parameter @xmath37 , @xmath38 for numerical analysis , we will assume @xmath39 .
several features of the model can be noted .
first , it contains a new cp violating phase @xmath40 through its dependence on @xmath32 . by comparison
, 2hdm s with nfc ( models i and ii ) involve only the ckm phase . secondly , charm quark yukawa couplings of the type @xmath41 ( @xmath42 ) are proportional to @xmath21 and thus are enhanced in this model .
in contrast , the corresponding couplings in models i and ii of the 2hdm are suppressed by @xmath43 .
it is the presence of the new phase that leads to a novel cp - violating pattern in @xmath2 decays , as will be shown later . on the other hand , the enhanced charm yukawa s
draw our immediate attention to the @xmath1 system , @xmath44 and @xmath45 .
recall that there exists no tree level fcnh coupling among the down - type quarks in the present model .
the most important contribution to the @xmath1 system from the charged higgs sector is associated with the @xmath46 box diagram which has a huge enhancement factor of @xmath47 . therefore ,
both the @xmath1 mass difference @xmath48 and the cp violation parameter @xmath49 are expected to place stringent constraints on the parameter space of the model . * @xmath50 * in the sm the short - distance contribution to @xmath48 is dominated by the charm quark and the long distance contribution can not be estimated reliably though is expected to be quite significant @xcite . including both the sm and the charged higgs effects , we obtain in the t2hdm the total short distance contribution to the mass difference as , @xmath51\ ] ] where the first term is the sm charm contribution and the second term is due to the dominant @xmath46 box diagram . here
@xmath52 is the usual bag factor , @xmath53 , and @xmath54 and @xmath55 are the qcd corrections to the two box diagrams .
the sm top quark contribution is a few percent of the charm quark effect and is not included here .
similarly , in comparison to the @xmath46 box diagram , the contributions from @xmath56 and other box diagrams are small in the large @xmath57 limit and are ignored . to numerically deduce the allowed parameter space subject to the @xmath48 constraint , we use the method described in @xcite for error analysis . assuming the magnitude of the long distance contribution to @xmath48 to be no larger than @xmath58 and simply taking @xmath59
, we find the @xmath60 c.l .
limit , @xmath61 which is valid for @xmath62 . as a result of the @xmath47 dependence in @xmath48 ,
this constraint puts a severe lower bound on the charged higgs mass when @xmath57 is large . by comparison ,
the charged higgs effect on @xmath48 in 2hdm s with nfc is suppressed by @xmath63 and is thus negligible .
note also that the short distance effect is independent of the mixing parameter @xmath37 of the t2hdm . * @xmath64 * unlike @xmath48 , the @xmath0 violation parameter @xmath49 is short distance dominated . in the sm , the dominant component of @xmath49 comes from the @xmath65 and @xmath66 box diagrams .
the complete expression and its numerical evaluation , including the next - to - leading - order ( nlo ) qcd corrections , can be found in @xcite .
the charged higgs contribution to @xmath49 is still dominated by the charm quark , with the imaginary part of the @xmath46 box diagram now dependent on both the magnitude and the phase of @xmath37 .
more explicitly , this contribution from charged higgs exchange is given by , @xmath67 where @xmath68 , @xmath69 , @xmath70 , and @xmath71 are the ckm parameters in the wolfenstein parameterization @xcite , @xmath72 is the ckm phase , and @xmath73 . due to its dependence on the new phase @xmath40 ,
@xmath49 no longer restricts the ckm angle @xmath74 .
in fact , current data allows a wide range of values for @xmath74 .
we can obtain bounds on the parameter @xmath75 for any given value of @xmath74 by allowing @xmath76 to vary within its @xmath77 uncertainties derived from @xmath78 . for a real ckm matrix ( @xmath79 ) , we obtain the @xmath60 c.l bound @xmath80 ; the bound becomes @xmath81 for @xmath82 . if @xmath74 takes its sm central value of @xmath83 @xcite , the bound becomes nearly symmetric about zero as expected : @xmath84 . unlike @xmath48 , the cp violation parameter @xmath49 imposes a constraint on the @xmath85 plane that is dependent on @xmath86 and @xmath40 . as for @xmath48 , the charged higgs contribution to @xmath49 in 2hdm s with nfc is suppressed by @xmath63 and is thus negligible when @xmath21 is large . the agreement between the experimental measurement and the sm prediction for the @xmath44 decay rate places stringent constraints on possible flavor physics beyond the sm . in model ii for example , where the most important correction comes from the top quark and charged higgs loop ,
a @xmath21-independent lower bound of about @xmath87 can be imposed on the charged higgs mass @xcite .
this picture changes dramatically in the t2hdm where the charm yukawa s are greatly enhanced relative to those in model ii .
consequently , both top and charm loops involving charged higgs exchange become important .
furthermore , both amplitudes become complex due to their dependence on the new phase @xmath40 . as a result ,
the charged higgs amplitude can interfere either constructively or destructively with the sm amplitude depending on the phase @xmath40 .
the bound on the charged higgs mass will be dependent on both @xmath57 and @xmath37 , and a relatively light higgs can still be allowed . for the numerical estimate , we will simply neglect the effect of the scalar operator @xmath88 induced from higgs exchange , and concentrate on the effect due to the charged higgs correction to the matching condition for the leading - order ( lo ) wilson coefficient @xmath89 at the scale @xmath90 : @xmath91 \label{eq : delc7 } \\ \!\ ! & & \!\!\ !
\times\left\{b(y_u ) + a(y_u ) \left[-1 + \left(\sigma^\dagger v\right)_{ub } \left(\cot^2\beta + 1\right)/ m_u v_{ub } \right]/6 \right\ } .
\nonumber\end{aligned}\ ] ] in this expression @xmath92 for @xmath93 , @xmath94 , and @xmath68 and @xmath2 are the standard expressions @xcite . the effective wilson coefficient at the scale @xmath95 is modified according to @xmath96 the @xmath44 constraint on the @xmath97 plane , together with the constraints from @xmath48 and @xmath49 , are presented in fig . 1 for three representative choices of the ckm phase @xmath74 .
the experimental data on @xmath98 and on other decays do not lead to new exclusion regions after imposing the @xmath44 and @xmath1 constraints .
.5 cm 6 truein
in the sm , the time - dependent @xmath0 asymmetry @xmath99/ \left[\gamma(b(t))+\gamma(\overline{b}(t))\right]$ ] for @xmath100 is free of hadronic uncertainties and is given by @xmath101 @xcite , where @xmath102 is the mass difference between the neutral @xmath2 mesons . existing experimental data
indirectly constrains @xmath5 @xcite within the sm .
note that this @xmath0 asymmetry in the sm arises from the @xmath103 mixing @xcite and not from the decay amplitudes , as can be seen easily from the wolfenstein parameterization of the ckm matrix . in the t2hdm ,
there is one additional diagram due to charged higgs exchange that mediates the decay @xmath104 .
integrating out the @xmath105 and @xmath106 gives us the effective lagrangian , @xmath107 + { \rm h.c .
} , \label{eqn : leff}\ ] ] where @xmath108 with @xmath109 taken to be real and positive .
terms which are subdominant in the large @xmath57 limit are not included in the above equation .
note that the charged - higgs - mediated decay amplitude is complex and its interference with the sm amplitude could significantly modify the @xmath0 asymmetry in @xmath3 @xcite . on the other hand ,
the charged higgs has a vanishingly small effect on the @xmath103 mixing amplitude . to proceed , we assume factorization and use fierz rearrangement to evaluate the hadronic matrix elements of the two four - fermi operators .
the total amplitudes are then obtained , ( bk_s ) _ sm , ( |bk_s ) _ sm where the sm amplitudes satisfy @xmath110 .
therefore the ratio @xmath111 gets a phase , /a = ( - 2 i ) , = / ( 1- ) instead of measuring @xmath112 , the @xmath0 asymmetry nows measures @xmath113 although the magnitude of @xmath114 can be at most of order @xmath115 after taking into account the experimental constraints , there could be large deviations of the @xmath0 asymmetry , @xmath116 , from the sm prediction .
this can be understood as follows .
the charged higgs contribution to @xmath49 could be comparable in size to the sm effect , and this basically sets free the ckm angles @xmath117 and @xmath74 .
when @xmath117 and @xmath74 take negative values , the @xmath0 asymmetry in @xmath118 can be of opposite sign relative to the sm expectation .
the cp asymmetry in the present model is illustrated in fig . 2 for three representative choices of @xmath74 .
as can be seen from the figure , @xmath119 can take almost any value in the t2hdm .
2.5 truein .2 cm
the analysis for @xmath118 can be straightforwardly extended to other @xmath2 decay modes @xcite . of particular interest are those decays where the sm prediction for the cp asymmetry is zero or very small .
these include @xmath120 , @xmath121 , and some others . in the t2hdm , it is not hard to see that the @xmath0 asymmetries for these two decays are @xmath122 , and thus could be of order tens of percent .
indeed , a simple @xmath0-violating pattern can be obtained for the various @xmath2 decays that is quite distinctive from the sm prediction .
the complex , enhanced charm yukawa for @xmath123 provides an interesting scenario for generating a sizable neutron electric dipole moment ( edm ) via a one loop diagram with virtual charm and @xmath106 .
two vertices of the diagram involve charm quarks of opposite chirality in order for the diagram to have an imaginary piece . using the non - relativistic quark model relation between the neutron edm and quark edm , one has , d_n d_d = a || ^4 ( _ ckm - ) ( - ) . taking @xmath29 as the current quark mass and imposing the experimental constraints on the model , we find that the neutron edm can be as large as @xmath124 , not too far below the present experimental limit .
it is worth pointing out that this estimate is valid even if the new phase @xmath40 vanishes , as the @xmath125 mixing always gives the @xmath123 coupling a complex piece that is proportional to @xmath126 .
the top - quark 2hdm provides one example for new @xmath0-violating phases and non - standard charged higgs yukawa couplings which should be generally present in multi - higgs doublet models when natural flavor conservation is not enforced .
although one need not take this particular model too seriously , the interesting phenomenological implications of flavor violation in general deserve further investigation .
cheng and m. sher , phys .
d * 35 * , 3484 ( 1987 ) ; + l. hall and s. weinberg , phys .
d * 48 * , r979 ( 1993 ) ; + y. l. wu and l. wolfenstein , phys .
lett . * 73 * , 1762 ( 1994 ) , and y. l. wu , hep - ph/9404241 .
see , e.g. , m. ciuchini , _
et al_. , nucl . phys . *
b527 * , 21 ( 1998 ) .
b. grinstein , r. springer , and m. wise , nucl .
b339 * , 269 ( 1990 ) .
see , e.g. , the article by h. quinn in review of particle physics , c. caso _ et al_. , eur . phys .
j. c * 3 * , 1 ( 1998 ) .
for a recent analysis of new physics effects on @xmath0 asymmetries through @xmath0 violation in @xmath103 mixing , see for example , m. gronau and d. london , phys .
d * 55 * , 2845 ( 1997 ) , and references therein . | in the absence of natural flavor conservation , multi - higgs doublet models generally contain new sources of @xmath0 violation and anomalous charged higgs yukawa couplings .
we present a charged - higgs @xmath0 violation study of one such two - higgs doublet model ( 2hdm ) which treats the top quark differently from the other quarks .
the phenomenological implications for the @xmath1 system and for @xmath2 decays differ significantly from those of the standard model ( sm ) and of the 2hdm s with natural flavor conservation .
in particular , the sm phase in this model could take a wide range of values , and the @xmath0 asymmetry in the gold - plated " decay mode @xmath3 could be quite different from and even of opposite sign relative to the sm prediction .
a new mechanism is also noted for generating a large neutron electric dipole moment which is close to the present experimental limit . |
the theory of random matrices , introduced by wigner in the 1950s , has proved to be a very useful tool in many fields of physics , from localisation theory to quantum transport ( see e.g. @xcite and references therein ) . in quantum chaos , a well accepted conjecture states that wigner - dyson random matrix ensembles describe statistical properties of spectra of quantum systems whose classical counterpart is chaotic @xcite , while statistics of integrable systems is best described by poisson statistics of independent random variables @xcite .
the corresponding wave functions are extended in the chaotic case and localised in the integrable case .
the choice of the random matrix ensemble suited to describe the statistical behaviour of a system depends on the symmetries of that system . in the usual setting @xcite ,
standard random matrix ensembles consist of matrices @xmath1 with independent gaussian random elements whose measure is invariant over conjugation @xmath2 where @xmath3 is an arbitrary matrix belonging to one of the three following groups of matrices : unitary , orthogonal , or symplectic .
the unitary group defines the gaussian unitary ensemble ( gue ) , which is supposed to describe statistical properties of energy levels of chaotic systems without time - reversal invariance .
the orthogonal group corresponds to gaussian orthogonal ensemble ( goe ) , used for time - reversal invariant chaotic systems .
the symplectic group gives rise to gaussian symplectic ensemble ( gse ) , applicable to time - reversal chaotic systems with half - integer spin without rotational symmetry .
though many extensions and generalisations of random matrices have been proposed in order to best describe various models @xcite , the existence of a large invariance group as in ( [ invariance ] ) remains their characteristic feature . without such invariance
it is very difficult to connect analytically simple properties of matrix elements with complex properties of matrix eigenvalues . for all random matrix ensembles with invariance group
it is possible to integrate over unnecessary variables in order to get explicitly the joint distribution of eigenvalues under the form @xmath4 with @xmath5 a system - dependent potential and @xmath6 a parameter . for the gaussian ensembles the potential is quadratic and the parameter @xmath6 is equal to 1 for goe , 2 for gue , and 4 for gse .
all correlation functions for invariant ensembles can be calculated analytically @xcite .
however the resulting formulas are cumbersome . for the nearest - neighbour distribution @xmath7 , instead of
the exact expression one often uses a simple surmise proposed by wigner .
this surmise has correct functional dependence at small and large argument and takes the form @xmath8 with constants @xmath9 and @xmath10 determined from the normalisation conditions @xmath11 the wigner - type surmise for the probability @xmath12 that between two eigenvalues separated by @xmath13 there exist exactly @xmath14 other levels ( with @xmath15 ) is @xcite @xmath16 and @xmath17 , @xmath18 are fixed by the normalisations @xmath19 while for chaotic systems it is possible to argue that eigenstates may statistically be invariant under rotations , this is not the case for more general models . in order to describe statistical properties of such systems
one has to consider non - invariant ensembles of random matrices .
one of the most investigated examples is the three - dimensional anderson model @xcite , with on - site disorder and nearest - neighbour coupling .
depending on the strength of the disorder , it can display metallic behaviour well described by standard random matrix ensembles , or insulator behaviour with poisson - like spectrum .
however , at the metal - insulator transition , spectral statistics are of an intermediate type and are not described by invariant ensembles @xcite .
similar behaviours have been observed in pseudo - integrable billiards @xcite , quantum maps corresponding to diffractive classical maps @xcite , or quantum hall transitions @xcite .
models have been proposed to describe such intermediate statistics @xcite , and random matrix ensembles which possess similar features have been constructed , e.g. power - law random banded matrix ensembles @xcite .
the main purpose of this paper is to construct random matrix ensembles which are not invariant over rotations of eigenstates , but whose joint distributions of eigenvalues can nevertheless be calculated analytically .
a short version of the paper has been published in @xcite .
all of these ensembles have intermediate statistics , and for certain of them spectral correlation functions , e.g. the nearest - neighbour distribution , are obtained explicitly .
eigenfunctions of these ensembles are neither localised ( as for integrable systems ) nor extended ( as for chaotic models ) but have fractal properties @xcite .
random matrices of the proposed critical ensembles are constructed from the lax matrices of classical integrable models .
these models are systems of @xmath0 classical particles labelled by an index @xmath20 , @xmath21 in a one - dimensional space .
each particle @xmath20 is characterised by its position in space @xmath22 and its momentum @xmath23 .
the dynamics of the particles is entirely described by the hamiltonian @xmath24 , where @xmath25 and @xmath26 .
the characteristic property of these models is the existence of a pair of @xmath27 matrices @xmath28 and @xmath1 , called the lax pair of the system @xcite , such that the equations of motion ( the hamilton equations , derived from the system hamiltonian ) are equivalent to @xmath29 the lax matrix @xmath28 is a matrix depending on momenta @xmath30 and coordinates @xmath31 . we propose to consider these lax matrices as random matrices with a certain natural measure of random variables @xmath32 and @xmath33 @xmath34 the explicit form of this measure depends on the system and will be discussed below .
we do not impose any dynamics on variables @xmath30 and @xmath31 .
the only information we use from the integrability of the underlying classical system is the existence and explicit form of action - angle variables @xmath35 and @xmath36 .
in particular , it is well known that the transformation from momenta and coordinates to action - angle variables is canonical , so that @xmath37 direct proof that the transformation is canonical is difficult in general , and implicit methods have been used to establish it for specific systems @xcite-@xcite . in the models we consider here , action variables turn out to be the eigenvalues @xmath38 of the lax matrix , or a simple function of them . the canonical change of variables from momenta and coordinates to action - angle variables in ( [ measure_pq ] ) leads to a formal relation @xmath39 where @xmath40 .
the exact joint distribution of eigenvalues is then obtained by integration over angle variables , which can easily be performed in all cases considered , and yields @xmath41 this scheme is general and can be adapted to many different models .
in this paper we consider in detail four typical models of @xmath0-particle classical integrable systems .
the three first , labelled cm@xmath42 , cm@xmath43 , and cm@xmath44 , correspond to the rational , hyperbolic , and trigonometric calogero - moser models @xcite .
the fourth model , labelled rs , is a trigonometric variant of the ruijsenaars - schneider model @xcite .
the calogero - moser models are defined by the hamiltonian @xmath45 where @xmath46 is a potential depending on the distance between particles and @xmath47 is a constant @xcite . for the models considered here it has the form @xmath48 , where @xmath49 the hamiltonian of our fourth model , the trigonometric ruijsenaars - schneider model , is @xcite @xmath50}{\sin^2\big [ \mu ( q_j - q_k)/2 \big ] } \right ) ^{1/2 } \hspace{1 cm } \mathrm{model\ rs}\ ; . \label{hamiltonian_rs}\ ] ] the plan of the paper is the following .
sections [ model_gm_r ] , [ model_cm_h ] , and [ model_cm_t ] are devoted to the construction of critical ensembles related respectively with the rational , hyperbolic , and trigonometric calogero - moser models . in each of these sections
we briefly present the construction of the action - angle variables and choose a natural measure of random momenta and coordinates which allows an easy change of variables as in ( [ measure_lphi ] ) .
we then give explicit formulas for the joint distribution of eigenvalues for the resulting critical ensembles of lax matrices . in section [ model_rs ]
this scheme is applied to the ruijsenaars - schneider model . for this model
the joint distribution of eigenvalues takes a form which makes it suitable for the application of the transfer operator formalism .
this approach is detailed in section [ transfer_op ] , and in section [ correlation_functions ] it is applied to the analytic calculation of nearest - neighbour distributions for the rs model .
the spectral compressibility for this model is obtained in section [ compressibility ] .
+ for clarity we state below the principal results for the four models considered in this paper .
the cm@xmath42 ensemble is defined as the ensemble of @xmath27 hermitian matrices of the form @xmath51 with @xmath47 a real constant .
positions @xmath31 and momenta @xmath30 are random variables distributed according to the density @xmath52\ , \ ] ] with @xmath53 and @xmath54 arbitrary positive constants .
the joint distribution of eigenvalues for this ensemble is then given by @xmath55\ .\ ] ] a characteristic property of this ensemble is the exponentially strong level repulsion : the nearest - neighbour spacing distribution @xmath7 is characterised by @xmath56 we propose the following wigner - type surmise for the next - to - nearest - neighbour spacing distributions @xmath12 , depending on four parameters : @xmath57 it contains two fitting constants depending on @xmath58 .
the other two are fixed by the normalisation .
the cm@xmath43 ensemble is defined as the ensemble of @xmath27 hermitian matrices of the form @xmath59 } \label{lax_cm_h}\ ] ] with @xmath47 and @xmath60 real constants , and @xmath31 and @xmath30 distributed according to the density @xmath61 } \big ) -b\sum_j \cosh \mu q_j\right ] \ .\ ] ] the exact joint distribution for this model is @xmath62 where @xmath63 is the modified bessel function of the second kind .
the nearest - neighbour spacing distribution has an exponential asymptotic similar to ( [ p_cm_r ] ) but with @xmath64 leading term instead of @xmath65 , namely @xmath66 the wigner - type surmise for cm@xmath43 is @xmath67 matrices from this ensemble correspond to a situation where @xmath60 in eq .
( [ lax_cm_h ] ) is allowed to take pure imaginary values .
they are of the form @xmath68 } \label{lax_cm_t}\ ] ] with @xmath47 and @xmath60 real constants , and @xmath31 and @xmath30 distributed according to the density @xmath69 } \big ) \right ] \ ] ] with the restrictions that all @xmath33 are between @xmath70 and @xmath71 .
the exact joint distribution of eigenvalues for this ensemble is @xmath72 where the function @xmath73 is equal to 1 if @xmath74 and @xmath75 for all @xmath76 .
the nearest - neighbour spacing distribution is given by a shifted poisson distribution of the form @xmath77 with @xmath10 some fitting constant .
the rs ensemble is defined as the ensemble of @xmath27 matrices of the form @xcite @xmath78 } { \sin \big [ \mu ( q_k - q_r+g\sigma)/2\big ] } \tilde{v}_r^{1/2}\mathrm{e}^{\mathrm{i}\sigma p_r/2 } \label{lax_ps}\ ] ] with @xmath79}{\sin \big [ \mu ( q_k - q_j)/2\big ] } \ , \qquad \tilde{w}_k=\prod_{j\neq k}\frac{\sin \big
[ \mu ( q_k - q_j+g\sigma)/2\big ] } { \sin \big [ \mu ( q_k - q_j)/2\big ] } \ .\ ] ] let @xmath80 be the set of @xmath31 such that for all @xmath81 the sign of both @xmath82 and @xmath83 is the same as the sign of @xmath84 .
the matrix @xmath28 is unitary if and only if @xmath85 . the variables @xmath31 and @xmath30 are chosen to be distributed according to the uniform density in the region where @xmath28 is unitary .
that is , we choose momentum variables @xmath32 independent and uniformly distributed between @xmath70 and @xmath86 and coordinate variables @xmath31 uniformly distributed over @xmath80 . in this case
eigenvalues of the lax matrices ( [ lax_ps ] ) are also uniformly distributed over @xmath80 .
choosing @xmath87 , @xmath88 and @xmath89 , we compute correlation functions of eigenvalues of matrix ( [ lax_ps ] ) for fixed @xmath9 and @xmath90 .
the results strongly depend on the integer part of @xmath9 . for @xmath91
the nearest - neighbour spacing distribution is similar to ( [ shifted_poisson ] ) with constant @xmath10 now equal to @xmath9 . for @xmath92
the nearest - neighbour distribution takes the form @xmath93 constants @xmath53 and @xmath94 are determined from the normalisation conditions .
other correlation functions are also obtained in section [ correlation_functions ] .
the results presented above are quite robust with respect to alterations in the distribution of @xmath31 and @xmath30 . in all models
considered we chose ( as explained above ) a distribution of coordinates such that the @xmath33 are confined to a finite interval while having a strong repulsion between each other .
as may be expected physically ( though we do not have a rigorous proof for this ) , numerical evidence shows that , if we keep these two characteristic features , spectral properties for @xmath95 depend only weakly on the precise choice for the distribution of @xmath31 and @xmath30 . from these considerations
it is thus natural to use , rather than the exact complicated distribution of @xmath31 , the picket - fence configuration when all coordinates are just fixed and equally spaced .
as all definitions of our ensembles involve only differences @xmath96 multiplied by a parameter ( @xmath60 or @xmath47 , depending on the model ) , we can without loss of generality choose to take @xmath97 , @xmath98 .
for numerical implementation we chose @xmath97 , and @xmath32 as independent gaussian variables with zero mean and with variance equal @xmath99 ( cm ensembles ) or independent variables uniformly distributed between @xmath70 and @xmath100 ( rs ensemble ) .
for concreteness , we fixed @xmath101 for cm@xmath43 and cm@xmath44 and @xmath102 for rs .
for such a choice , the @xmath27 lax matrices take the form @xmath103 } , & \mathrm{model\ cm_h}\\ l_{kr}=p_k\delta_{kr}+\mathrm{i}g\dfrac{2\pi(1-\delta_{kr})}{n\sin\big [ 2\pi(k - r)/n\big ] } , & \mathrm{model\ cm_t}\\ l_{kr}=\mathrm{e}^{\mathrm{i}p_k}\dfrac{1-\mathrm{e}^{2\mathrm{i}\pi g } } { n(1-\mathrm{e}^{2\mathrm{i}\pi(k - r+g)/n } ) } , & \mathrm{model\ rs}\;\ ; \end{array } \ .
\label{simplified}\ ] ] for cm@xmath44 matrices with even @xmath0 , to avoid the singularity we changed @xmath104 in the above formula . as the figures in the next sections
show , despite this particular choice for the distribution of @xmath31 and @xmath30 , the agreement between the computed spectral statistics and analytical formulas is remarkable .
the first model we consider is the rational calogero - moser model cm@xmath42 @xcite , characterised by the lax matrix @xmath105 it depends on a real constant @xmath47 and on a set of @xmath106 random variables @xmath107 and @xmath108 whose distribution will be specified later on .
we are interested in eigenvalues @xmath38 and eigenfunctions @xmath109 of this matrix ( here and below we will use the greek letters to label eigenvalues and corresponding eigenfunctions ) @xmath110 to construct angle - action variables let us define the new quantities @xmath111 from eq .
one gets @xmath112 multiplying both sides by @xmath113 and summing over @xmath81 and @xmath114 one gets @xmath115 where @xmath116 for @xmath117 , eq . implies that @xmath118 , and one can choose the overall phase of the eigenvector @xmath109 in such a way that @xmath119 .
let @xmath120 be new variables defined by @xmath121 then from eq .
( [ qmn ] ) we have @xmath122 the matrix @xmath123 can be seen as the dual matrix of @xmath28 , with @xmath120 playing the role of momenta and @xmath38 the role of positions . in @xcite
it was proved that there is a canonical transformation from position and momentum variables @xmath124 to action and angle variables @xmath125 .
showing that the transformation is canonical is a rather technical mathematical result .
however one can easily check that the new variables @xmath38 and @xmath120 verify hamilton - jacobi equations ( see appendix [ hj ] ) .
we now consider an ensemble of hermitian matrices of the form with random variables @xmath107 and @xmath108 drawn according to the measure @xmath126\prod_k\mathrm{d}p_k\ , \mathrm{d}q_k\ , \label{measureinr}\ ] ] where @xmath53 and @xmath54 are given constants and @xmath127 a normalisation factor . the first term in eq
. is the analog of the usual gaussian weight of rmt ; the second term is a quadratic confinement potential .
since the action - angle transformation is canonical one has @xmath128 from eq .
, using orthogonality of eigenvectors one gets @xmath129 using these relations one can rewrite the distribution ( [ measureinr ] ) in action - angle variables @xmath38 and @xmath120 as @xmath130\prod_{\alpha}\mathrm{d}\lambda_{\alpha}\ , \mathrm{d}\phi_{\alpha}\,.\ ] ] integration over the @xmath120 gives a constant .
we thus obtain the joint distribution of eigenvalues for the ensemble of random matrices @xmath28 with the measure ( [ measureinr ] ) as @xmath131\ .
\label{distributioninr}\ ] ] note that , similarly as in the standard rmt case , this joint eigenvalue distribution can be interpreted via the coulomb gas model as the partition function of an ensemble of particles on a line , here with inverse square repulsion .
after rescaling @xmath132 , equilibria positions of the particles at positions @xmath38 are given by @xmath133 such a relation characterises the zeros of hermite polynomials of degree @xmath0 ( see also eq . ( 10.3 ) of @xcite ) .
it is known from rmt @xcite that the distribution of eigenvalues of gaussian random ensembles has a similar property , which implies that the asymptotic density of eigenvalues is given by wigner s semi - circle law .
an immediate consequence of the distribution is the unusual very strong level repulsion at small distances . for all standard random matrix ensembles the nearest - neighbour distribution @xmath7 behaves as @xmath134 at small @xmath13 .
by contrast , in our case it follows from ( [ distributioninr ] ) that @xmath135 as the potential between eigenvalues decreases as the inverse square of the distance between them , the probability of having a gap of size @xmath13 for large @xmath13 is exponentially small .
we could not calculate exactly correlation functions for the distribution ( [ distributioninr ] ) .
however , combining the two asymptotics above , we build a wigner - type surmise for the nearest - neighbour spacing distribution of the form @xmath136 where @xmath10 is a fitting constant , and constants @xmath9 and @xmath137 are determined from the normalisation conditions ( [ normalization ] ) .
for the @xmath58th nearest - neighbour spacing distributions @xmath12 , with @xmath138 , we conjecture , by analogy with the wigner surmise for standard random matrices , the form @xmath139 with two fitting constants @xmath10 and @xmath140 . to assess this conjecture we compare the analytical expressions with numerical results , with the choice of parameters detailed in section [ numericalimplementation ] .
results displayed in fig .
[ psinr ] show that the agreement is remarkable . for the random matrices of model
cm@xmath42 with @xmath141 ( top left ) , @xmath142 ( top right ) , @xmath143 ( bottom left ) , and @xmath99 ( bottom right ) , averaged over the central quarter of the unfolded spectrum for @xmath144 realisations of matrices of size @xmath145 .
solid lines are numerical results , dashed lines indicate the fits ( [ surmise ] ) for and for @xmath12 with ( in each panel from left to right ) @xmath146 ( red ) , @xmath147 ( green ) and @xmath148 ( blue ) .
[ psinr ] ]
the lax matrix for the hyperbolic calogero - moser model cm@xmath43 reads @xcite @xmath149 let us define two matrices @xmath123 and @xmath150 by @xmath151 and two vectors @xmath152 from one can get the two equivalent equations @xmath153 multiplying both sides by @xmath154 and summing over all @xmath81 and @xmath114 one gets @xmath155 which implies that matrices @xmath123 and @xmath150 take the form @xmath156 by their definition , matrices @xmath123 and @xmath150 are inverse of each other , so that @xmath157 for all @xmath158 . for @xmath117
this condition implies that @xmath159 ( in particular , it follows that all @xmath160 and @xmath161 are nonzero ) . for @xmath162
one obtains @xmath163 using the identity @xmath164\dfrac{1}{\lambda_{\alpha}-\lambda_{\gamma}}\ , \ ] ] valid for @xmath165 , one concludes that @xmath166 where @xmath137 is a certain constant independent on @xmath76 . according to this equation the quantities @xmath167 obey a system of linear equations of the form @xmath168 with @xmath169 and @xmath170 .
this equation coincides with eq . in appendix
[ app_a ] . from it follows that @xmath171 where @xmath172 while implies that @xmath173 it readily follows from the definition of @xmath160 and @xmath161 that @xmath174 , thus the value of @xmath137 is fixed by @xmath175 .
equation is then fulfilled as a direct consequence of .
finally we have @xmath176 let @xmath120 be new variables defined from diagonal elements of matrix @xmath123 by @xmath177 then from and it follows that @xmath178 in the definition of @xmath160 it is convenient to choose the overall phase of the eigenvector @xmath109 in such a way that @xmath160 be real . as @xmath179
one has @xmath180 using eq . , the matrix @xmath123 can now be expressed in terms of the new variables @xmath38 and @xmath120 , as @xmath181 as in the case of model cm@xmath42 , the matrix @xmath123 can be seen as the dual matrix of @xmath28 .
indeed , @xmath123 coincides with the lax matrix of the rational ruijsenaars - schneider model with coordinates @xmath38 and momenta @xmath120 @xcite . in @xcite it has been proved that the transformation from position and momentum variables @xmath182 to action - angle variables @xmath125 is canonical .
again one can check that the new variables @xmath38 and @xmath120 verify hamilton - jacobi equations .
we now consider an ensemble of hermitian matrices of the form with random variables @xmath107 and @xmath108 drawn according to the measure @xmath183\prod_k\mathrm{d}p_k\ , \mathrm{d}q_k\ .
\label{measureiinr}\ ] ] as in the case of model cm@xmath42 , eq . contains a standard rmt gaussian weight and a confinement potential which can be rewritten as @xmath184 . using , , and the fact that the transformation is canonical , we get the distribution in terms of the new variables @xmath38 and @xmath120 as @xmath185\mathrm{d}\phi_{\alpha}= \frac{1}{\mu}k_0\left(b|v_{\alpha}|\right)\ ] ] where @xmath186 is the modified bessel function of the second kind .
this yields the joint distribution of eigenvalues for model cm@xmath43 as @xmath187 this expression is exact but difficult to handle . in order to find a wigner - type surmise for the nearest - neighbour distributions we consider the limiting behaviour @xmath188 when two nearby eigenvalues @xmath189 and @xmath190 get close to each other . setting @xmath191
we see that the factor @xmath192 in the case of model cm@xmath42 is replaced by a factor @xmath193 we therefore expect the nearest - neighbour spacing distribution to behave as @xmath194 in fig .
[ psiinr ] we show the results of numerical computations of the nearest - neighbour spacing distributions for matrices of the form with the choice of parameters and variables detailed in section [ numericalimplementation ] .
the surmise perfectly reproduces numerical results . for the random matrices of model
cm@xmath43 with @xmath101 and @xmath141 ( top left ) , @xmath142 ( top right ) , @xmath143 ( bottom left ) , and @xmath99 ( bottom right ) , averaged over the central quarter of the spectrum for @xmath195 realisations of matrices of size @xmath196 .
solid lines are numerical results , dashed lines indicate the fit for @xmath12 with ( in each panel from left to right ) @xmath146 ( red ) , @xmath147 ( green ) and @xmath148 ( blue ) .
[ psiinr ] ] in order to assess better the validity of the exponentially strong level repulsion for models cm@xmath42 and cm@xmath43 , we compare in fig .
[ psinr - iinr ] the beginning of the distributions @xmath7 for these models .
clearly the @xmath65 repulsion for cm@xmath42 and the @xmath64 repulsion for cm@xmath43 fit numerical curves very well .
however , the precision of our numerical results does not permit to confirm or reject the presence of the logarithmic term @xmath197 in @xmath7 for cm@xmath43 model . for the random matrices cm@xmath42
( @xmath198 , black circles ) and @xmath199 , red squares ) and cm@xmath43 ( @xmath200 , green triangles up , and @xmath199 , blue triangles down ) , with @xmath101 .
symbols are numerical results , solid lines indicate the fit @xmath201 ( model cm@xmath42 , eq . ) and @xmath202 ( model cm@xmath43 , eq . ) .
logarithm is natural .
[ psinr - iinr ] ]
for the trigonometric calogero - moser model cm@xmath44 the lax matrix is @xcite @xmath203 the only difference with the previous model , ( [ iinr ] ) , is the @xmath204 function which replaces the @xmath205 .
the matrix ( [ iiinr ] ) can be obtained from ( [ iinr ] ) by the substitution @xmath206 .
however the fact that positions of the particles are now defined on a circle ( because of the sin function ) makes the resulting spectral statistics entirely different from the previous models cm@xmath42 and cm@xmath43 . to construct action - angle variable we introduce , as in the previous section , two matrices @xmath207 and two vectors @xmath208 following the same steps as above with @xmath60 replaced by @xmath209 , one gets @xmath210 again , using the fact that @xmath123 is the inverse of @xmath150 we obtain that @xmath211 with certain constants @xmath212 and @xmath213 independent on @xmath76 . repeating the same arguments as in the previous section and using results of appendix [ app_a ]
one concludes that @xmath214 and @xmath215 with @xmath216 the new variables @xmath120 are defined as above from the diagonal elements of matrix @xmath123 as follows @xmath217 in the definition of @xmath160 one has a freedom to choose the overall phase of the eigenvector @xmath109 . since from one
must have @xmath218 , one can choose phases , for example , as follows @xmath219 then the matrix @xmath123 can be expressed in terms of new variables @xmath38 and @xmath120 as @xmath220 the inverse matrix @xmath150 plays a symmetric role , as it is obtained from @xmath123 by exchanging @xmath60 to @xmath221 . again , there is a canonical transformation from position and momentum variables @xmath182 to action and angle variables @xmath125 @xcite .
an important consequence of is that for all @xmath76 we should have @xmath222 these inequalities impose non - trivial restrictions on eigenvalues @xmath38 , as we will see now .
we label eigenvalues so that @xmath223 , and we consider the function @xmath224 it has @xmath0 poles at @xmath225 , and according to eqs . and it has @xmath0 zeros at @xmath226 .
if all numerators @xmath227 are positive then the derivative of @xmath228 is positive , and it is easy to check from the graph of the function that between two consecutive poles there is one and only one zero .
suppose @xmath229 .
then for @xmath230 the function @xmath231 has a strictly positive limit .
the lowest zero @xmath232 must thus lie in the interval @xmath233\lambda_1,\lambda_2[$ ] .
more generally one must have @xmath234\lambda_{\alpha},\lambda_{\alpha+1}[$ ] for @xmath235 , while the largest zero @xmath236 lies in the interval @xmath233\lambda_n,\infty[$ ] .
thus eigenvalues fulfill the inequalities @xmath237 conversely , if these inequalities are fulfilled then trivially all @xmath227 are positive .
therefore , are the necessary and sufficient conditions for the positivity of all @xmath227 . in particular , eigenvalues of the lax matrix obey inequalities for any choice of the @xmath31 .
these results adapt straightforwardly to the case where @xmath238 is negative .
since in the @xmath108 only appear as an argument in the @xmath204 function , there is no need to choose a confining potential for the particle distribution as in models cm@xmath42 and cm@xmath43 .
we consider the probability distribution of @xmath30 and @xmath31 in the form @xmath239 } \big ) \right ] \ ] ] with the restrictions that all @xmath33 are between @xmath70 and @xmath71 .
since the change of variables from @xmath30 and @xmath31 to @xmath38 and @xmath120 is canonical and the restrictions do not depend on phase variables the joint distribution of eigenvalues is @xmath240 where the function @xmath73 is equal to 1 if is fulfilled for all @xmath76 , and 0 otherwise .
it turns out that model cm@xmath44 is very similar to a fourth model , the ruijsenaars - schneider model , that we will consider in the next section .
therefore we postpone analytical calculations of the nearest - neighbour spacing distributions to section [ transfer_op ] .
the nearest - neighbour spacing distributions @xmath12 are shifted poisson distributions of the form @xmath241 with some numerical constant @xmath10 . in fig .
[ psiiinr ] we show the results of numerical computations for matrices of the form with the choice of parameters and variables detailed in section [ numericalimplementation ] . for model
cm@xmath44 with @xmath12 the shifted poisson distribution .[psiiinr ] ]
the calogero - moser models considered in the previous sections are such that there exists a matrix @xmath123 which is , in a certain sense , dual to the lax matrix @xmath28 .
namely , the canonical transformation from variables @xmath242 to action - angle variables @xmath243 is such that the action variables @xmath38 are eigenvalues of @xmath28 and angle variables @xmath120 are related to @xmath244 in a simple way .
in fact , the matrix @xmath123 is also a lax matrix , corresponding to a possibly different hamiltonian , and @xmath28 plays the role of a matrix dual to @xmath123 for the inverse of the canonical transformation @xcite .
the rational calogero - moser system is self - dual since matrices @xmath28 and @xmath123 , given by and , are equal up to labelling of the variables . the model we consider in this section is related to the above models in that its lax matrix @xmath28 and the dual matrix @xmath123 are a kind of generalisation of those of model cm@xmath44 .
the treatment of this model closely follows the previous section .
the lax matrix for ruijsenaars - schneider model is the @xmath27 unitary matrix given by @xcite @xmath78 } { \sin \big [ \mu ( q_k - q_r+g\sigma)/2\big ] } \tilde{v}_r^{1/2}\mathrm{e}^{\mathrm{i}\sigma p_r/2}\ . \label{iv}\ ] ] with @xmath245}{\sin \big [ \mu ( q_k - q_j)/2\big ] } \ , \qquad \tilde{w}_k=\prod_{j\neq k}\frac{\sin \big [ \mu ( q_k - q_j+g\sigma)/2\big ] } { \sin \big [ \mu ( q_k - q_j)/2\big ] } \ .\ ] ] this lax matrix is related with the hamiltonian by @xmath246 it is convenient to introduce the vectors @xmath247 so that matrix @xmath28 can be rewritten as @xmath248}{\sin \big [ \mu ( q_k - q_r+g\sigma)/2\big ] } \mathrm{e}^{-\mathrm{i}\mu q_r/2}\tilde{e}_r\ . \label{lefiv}\ ] ] as in the previous section , the condition that the matrix is unitary imposes certain restrictions on the coordinates @xmath31 , which we will discuss later . assuming that the lax matrix is unitary
, we choose to denote its eigenvalues by @xmath249 .
the dual matrices for the ruijsenaars - schneider model are defined by @xcite @xmath250 and the vectors @xmath160 and @xmath161 by @xmath251 from eq .
one has @xmath252=\tilde{f}_k^*\sin \big [ \mu g\sigma/2\big ] \tilde{e}_r\ .\ ] ] multiplying this expression by @xmath253 and summing both sides over @xmath81 and @xmath114 leads to @xmath254 e_{\beta}\ , \ ] ] which yields the analogue of , @xmath255}{\sin \big[\sigma ( \lambda_{\beta}-\lambda_{\alpha}+g\mu)/2 \big ] } \mathrm{e}^{-\mathrm{i}\sigma\lambda_{\beta}/2}e_{\beta}\ .
\label{q_iv_new}\ ] ] let us rewrite @xmath256 where we have set @xmath257 . from its definition it is clear that @xmath123 has to be an unitary matrix , i.e. @xmath258 .
selecting terms with @xmath117 and @xmath165 yields the two equations @xmath259 where we have set @xmath260 . using the identity @xmath261
it follows from that there exists a constant @xmath137 such that @xmath262 equations allow to obtain @xmath263 while from eq .
one obtains , using eq .
@xmath264 with new vectors @xmath227 and @xmath265 defined by @xmath266}{\sin \big [ \sigma ( \lambda_{\alpha}-\lambda_{\beta})/2\big ] } \ , \qquad w_{\alpha}=\prod_{\beta\neq \alpha}\frac{\sin\big [ \sigma ( \lambda_{\alpha}-\lambda_{\beta}-g\mu)/2\big ] } { \sin\big [ \sigma ( \lambda_{\alpha}-\lambda_{\beta})/2\big ] } \ . \label{vwiv}\ ] ] using we get @xmath267 there is some overall freedom in the definition of vectors @xmath268 and @xmath269 in eq .
, as one could multiply @xmath270 by some constant factor and divide @xmath271 by the same factor .
this in turn entails the same freedom for vectors @xmath160 and @xmath161 in .
if one chooses for instance @xmath272 then from unitarity of the transformation in one has @xmath273 , which fixes the value @xmath274 and thus @xmath275 is positive .
a convenient choice is to take @xmath276 since @xmath277 and @xmath278 are non - negative one concludes that @xmath227 and @xmath279 have the same sign as @xmath280 for all @xmath76 .
the choice implies that @xmath281 the new action variables are the @xmath38 , and angle variables @xmath120 are defined by @xmath282 equation implies that @xmath283 . in view of and
one can choose the phases of the eigenvectors @xmath109 such that @xmath160 and @xmath161 defined by can be expressed as @xmath284 in terms of the new variables , @xmath123 thus reads @xmath285}{\sin \big [ \sigma ( \lambda_{\beta}-\lambda_{\alpha}+g\mu)/2\big ] } v_{\beta}^{1/2}\mathrm{e}^{\mathrm{i}\mu \phi_{\beta}/2}\ .
\label{qiv}\ ] ] comparing and one sees that for model rs the dual matrix matrix @xmath123 is obtained from @xmath28 by changing @xmath286 , @xmath287 , @xmath288 , and @xmath289 .
it means that this model is self - dual ( the matrices @xmath28 and @xmath123 are the same up to a change of notation ) .
one can show that the transformation from @xmath182 to action and angle variables @xmath125 is canonical @xcite . as mentioned
, a consequence of the unitarity of the matrix @xmath123 is that @xmath227 and @xmath279 have the same sign as @xmath280 for all @xmath76 .
this implies that certain inequalities have to be verified by the @xmath38 , which we now derive in a way similar as in the previous section .
let us define the function @xmath290-t\cot ( n\tau/2\,)\ , \ ] ] which is periodic with period @xmath100 and can be considered as a function on the unit circle .
it has @xmath0 poles at @xmath291 .
taking the imaginary part of eq .
, with @xmath137 given by and @xmath160 given by , one obtains that @xmath231 has @xmath0 zeros at @xmath292 .
when all @xmath293 are positive , the same arguments as in the previous section imply that between two consecutive poles of @xmath231 there must be exactly one zero .
it means that between two nearby eigenvalues @xmath294 there is one and only one number of the form @xmath295 .
a similar reasoning starting from matrix @xmath296 leads to the conclusion that between two consecutive @xmath297 there must also be one and only one number of the form @xmath298 .
these two conditions ( shift by @xmath299 or @xmath300 ) are equivalent , thus we can restrict ourselves to a shift by @xmath299 , i.e. the condition that the sets @xmath301 and @xmath302 intertwine on the unit circle .
it is a necessary and sufficient condition for the matrix to be an unitary matrix . a similar conclusion is readily obtained when all @xmath293 are negative .
the conditions implied by this type of intertwining have been discussed in @xcite .
there is an fundamental difference between these eigenvalue conditions in the rs model and in the cm@xmath44 model discussed in the previous section . in the cm@xmath44 case
the lax matrix is hermitian , and the @xmath303 can take values on the whole real axis . in the rs case , as the lax matrix is unitary , the @xmath304 lie on the unit circle .
therefore poles and zeros can not be ordered in a simple way as in the previous case , and the analysis of the previous section does not apply . for completeness
we shortly repeat the arguments of the papers @xcite .
let us put all eigenvalues @xmath294 of the lax matrix on the unit circle and divide the circle into sectors with angle @xmath305 .
denote the ( positive ) angular distance from the boundaries of the @xmath81th sector in counter clockwise direction by @xmath306 and in clockwise direction by @xmath307 , as in fig .
[ sectors ] . .
only sectors numbers @xmath81 , @xmath308 , and @xmath309 are indicated .
black circles indicate the positions of the @xmath304 , where @xmath310 are the eigenvalues of the rs lax matrix . ] after a shift by @xmath305 , the intertwining relations imply that only one of the two points corresponding to @xmath306 and @xmath307 will fall in - between points corresponding to @xmath311 and @xmath312 .
the first case corresponds to @xmath313 and @xmath314 . in the second case
the inequalities are reversed and @xmath315 and @xmath316 . in both cases
the inequality @xmath317 is fulfilled .
let us consider consecutive sectors of angle @xmath305 as in fig .
[ sectors ] .
denote the number of eigenvalues in each sector by @xmath318 . after a shift by @xmath305 ,
eigenphases from the @xmath81th sector will move into the @xmath319th sector .
the @xmath318 shifted points divide this sector into @xmath320 intervals . as was proved above
, eigenphases in the @xmath319th sector have to intertwine with these shifted eigenphases .
therefore all @xmath320 intervals except the first and the last will be occupied .
the first will be occupied if @xmath321 and the last interval will be occupied provided @xmath322 .
these statements can be rewritten in the form of the recurrence relation @xmath323 where @xmath324 is the heaviside step function , @xmath325 when @xmath326 and @xmath327 for @xmath328 . from it
follows that this relation can be rewritten in the form @xmath329 we now specialise to the case where @xmath305 depends on the size @xmath0 of the lax matrix .
we set @xmath330 with fixed @xmath9 .
the case of integer @xmath9 is trivial and we thus assume that @xmath9 is not an integer .
the total number of sectors of angle @xmath305 in the unit circle is @xmath331\ ] ] where @xmath332 $ ] denotes the integer part of @xmath333 .
suppose the beginning of the first sector lies at position @xmath334 .
we choose to consider that this eigenvalue does not belong to the first sector , i.e. @xmath335 .
. for @xmath336 implies that @xmath337 . thus necessarily
@xmath338 and for @xmath339 one easily gets from eq . that @xmath340 the total number of eigenvalues lying into all @xmath341 sectors obeys the inequalities @xmath342 the right - hand side inequality comes from the fact that when @xmath9 is not an integer the union of all @xmath341 intervals does not overlap the whole circle : in particular , it does not contain the first eigenvalue @xmath334 from which we start our sectors .
the left - hand side inequality is a consequence of the fact that if eigenvalues were shifted by @xmath300 , the first sector in the opposite direction would have exactly the same number of eigenvalues as the second sector , i.e. @xmath343 eigenvalues : as all @xmath341 sectors does not cover the whole circle and the uncovering region is smaller than the sector of angle @xmath305 , it follows that the number of eigenvalues in all @xmath341 sectors is larger than @xmath344
. from one easily obtains a second inequality @xmath345 from these two inequalities it follows that @xmath346 by definition of @xmath341 we have @xmath347 substituting in the right - hand side of the minimum of @xmath341 and in the left - hand side the maximum value of @xmath341 one gets @xmath348 which entails that for @xmath0 large enough @xmath349 . since @xmath350 has to be an integer , @xmath351 $ ] .
this means that for sufficiently large @xmath0 , within an interval of length @xmath305 from any eigenvalue there are always exactly @xmath352 $ ] eigenvalues .
the inverse statement is also true .
if within an interval of length @xmath305 from any eigenvalue there exist exactly @xmath352 $ ] other eigenvalues then all @xmath227 and @xmath265 have the sign of @xmath353 . to see this
let us consider e.g. @xmath227 given by eq . .
it is a product of terms @xmath354 where @xmath355 .
as the distance between two eigenvalues may be restricted , @xmath356 , @xmath357 obeys inequality @xmath358 .
therefore @xmath359 , and @xmath360 is negative when @xmath361 and positive when @xmath362 .
if within an interval of length @xmath305 from @xmath363 there are exactly @xmath352 $ ] eigenvalues , then in the product formula for @xmath227 there are exactly @xmath352 $ ] negative terms , so that its total sign is @xmath364}$ ] . for @xmath0 large enough the sign of @xmath353 with @xmath365
is the sign of @xmath366 , which is precisely @xmath364}$ ] .
the above arguments prove that for sufficiently large @xmath0 ( whose value depends only on @xmath9 ) the necessary and sufficient condition for the unitarity of matrix @xmath123 is that at distance @xmath367 from any eigenvalue there exist @xmath352 $ ] other eigenvalues . as matrices @xmath28 and @xmath123 are dual , the unitarity condition for @xmath28 can be readily deduced : it is that at distance @xmath367 from one coordinate @xmath368 there are exactly @xmath352 $ ] other coordinates .
note that in @xcite only the case @xmath91 had been considered .
these restrictions determine the allowed region in coordinate space .
we choose as the natural measure of momenta and coordinates the uniform distribution for momenta ( between @xmath70 and @xmath86 ) and coordinates uniformly distributed in the allowed region as explained above . after the change of variables from coordinates and momentum to action - angle variables it follows that the resulting distribution of eigenvalues will be also uniform but in the allowed region of eigenvalues with the only restriction that any interval @xmath233\sigma
\lambda_\alpha,\sigma \lambda_\alpha+2\pi a / n[$ ] contains exactly @xmath352 $ ] eigenvalues .
in next section we will show that it is possible to calculate asymptotic expressions for the joint distribution of eigenvalues by using a transfer operator technique . for numerical investigations
, we consider an ensemble of unitary matrices of the form , with @xmath107 chosen as independent random variables uniformly distributed between @xmath70 and @xmath100 and the picket fence distribution of coordinates @xmath97 , @xmath369 ( see section [ numericalimplementation ] ) .
choosing constants such that @xmath102 , @xmath88 , and @xmath89 , so that @xmath370 , direct calculations yield @xmath371 with a slightly different choice of phases in , matrix @xmath28 simplifies to @xmath372 where we denote @xmath373 .
this is a particular specialisation of the lax matrix for the ruijsenaars - schneider model . in the form but with @xmath374 with fixed @xmath10
it first appeared in @xcite as a result of the quantisation of a classical parabolic map on the torus proposed in @xcite .
when @xmath10 is a rational number the map considered in @xcite corresponds to a pseudo - integrable map of exchange of two intervals . for this particular case where @xmath9 depends on @xmath0 ,
the spectral statistics of the unitary matrix has been obtained analytically in @xcite and @xcite without knowledge of the relation with the lax matrix of the ruijsenaars - schneider model . in the present case , where @xmath9 is a fixed parameter independent on @xmath0 ,
results are quite different .
analytical calculations of spectral correlation functions for this model are performed in the next sections . to illustrate the accuracy of the analytical results that we derive in the next sections , we show in fig .
[ psiv ] results of numerical computations for matrices of the form for different values of the parameter @xmath9 .
agreement is remarkable for all parameter values . for the random matrices of model rs with @xmath102 and @xmath375
( top left ) , @xmath376 ( top right ) , @xmath377 ( bottom left ) , and @xmath378 ( bottom right ) , averaged over @xmath379 realisations of matrices of size @xmath380 .
solid lines are numerical results , dashed lines indicate the analytical expressions for @xmath375 , ( [ nearest])([p3s ] ) for @xmath381 and @xmath382 , and for @xmath383 . in each panel
( from left to right ) @xmath146 ( red ) , @xmath147 ( green ) and @xmath148 ( blue ) .
we now calculate asymptotic expressions for the joint distribution of eigenvalue spacings for the lax matrix ensemble corresponding to the rs model .
as discussed in the previous section , eigenvalues are such that within an interval of length @xmath9 from any eigenvalue there are exactly @xmath352 $ ] other eigenvalues .
we introduce the rescaled nearest - neighbour spacings @xmath384 the answer strongly depends on the integer part of @xmath9 .
therefore we consider different cases separately .
the simplest case corresponds to @xmath91 . in this case
the only restriction is that the distance between the nearest eigenvalues is larger than @xmath9 , namely @xmath385 .
it is convenient to rewrite this restriction as follows .
the joint probability density of having @xmath386 eigenvalues inside an interval of length @xmath28 is given by @xmath387 where @xmath388 and @xmath389 is the normalisation constant @xmath390 we are interested in the joint probability distribution of @xmath58 consecutive spacings @xmath391 when @xmath392 with mean level spacing @xmath393 remaining constant . in the following we set @xmath394 .
we shall proceed as it was done in @xcite .
the multiple integrals in are easily calculated by introducing the function @xmath395 whose laplace transform reads @xmath396 here @xmath397 is the laplace transform of @xmath398 .
the inverse laplace transform of @xmath399 is then @xmath400 the large-@xmath0 behaviour of is obtained by saddle - point approximation .
the value @xmath137 corresponds to the solution of the saddle - point equation @xmath401 ( recall that @xmath394 ) .
the solution reads @xmath402 similarly , the inverse laplace transform of @xmath403 can be calculated by saddle - point approximation . rather than calculating explicitly all prefactors coming from the integration
, it is easier to observe that the large-@xmath0 behaviour of the normalisation factor @xmath404 is obtained similarly .
one finally gets @xmath405 we now consider the case @xmath92 . according to previous section for sufficiently large matrix size , any eigenvalue @xmath306 is such that there exists exactly @xmath99 eigenvalue in the interval @xmath233x_k , x_k+2\pi a / n[$ ] . in other words ,
the constraints on @xmath306 are that @xmath406x_k , x_k+2\pi a / n[$ ] and @xmath407 . in terms of the differences between consecutive eigenvalues ,
the above restriction are equivalent to @xmath408 introduce the function @xmath409 by @xmath410 and @xmath398 as in .
then the joint probability density of @xmath386 eigenvalues inside an interval of length @xmath28 is given by the following expression @xmath411 where @xmath389 is the normalisation constant @xmath412 the large-@xmath0 behaviour of the joint probability distribution of @xmath58 consecutive spacings can be then obtained as above , following @xcite .
we review the main steps of the procedure .
introducing the function @xmath413 its laplace transform reads @xmath414 this quantity can be seen as a product of transfer operators @xmath415 where the transfer operator , @xmath416 is defined as @xmath417 this is a real symmetric operator .
its real eigenvalues , @xmath418 , and eigenfunctions , @xmath419 , verify @xmath420 as this operator is real symmetric its eigenfunctions can be chosen to be orthonormal @xmath421 the transfer operator can be expanded over the basis of eigenfunctions as @xmath422 in the large-@xmath0 limit the dominant contribution to is given by the largest eigenvalue @xmath423 . then using the orthogonality property of the @xmath424 we get @xmath425 the large-@xmath0 behaviour of @xmath426 is obtained by performing the inverse laplace transform of @xmath427 . as in the leading term is obtained by saddle - point approximation ; here the saddle - point is such that @xmath428 ( again we take @xmath394 ) . calculating in a similar way the large-@xmath0 behaviour of the normalisation factor @xmath404 ,
one finally gets @xmath429 the general case @xmath430 can be treated in a similar way . in this case
each interval @xmath233x_k , x_k+2\pi a / n[$ ] contains exactly @xmath431 eigenvalues @xmath432 . in terms of the rescaled differences between consecutive eigenvalues
this condition is equivalent to the following two inequalities @xmath433 in analogy with eq .
, the joint probability of eigenvalues spacings thus reads @xmath434 with @xmath435 and @xmath47 defined by and .
the large-@xmath0 behaviour is calculated as above by introducing a transfer operator , which in this case depends on two sets of variables @xmath436 ) and @xmath437 shifted by one unit i.e. @xmath438 , @xmath439 , @xmath440 , @xmath441 ( see e.g. @xcite ) .
the explicit form of the transfer operator is the following @xmath442 the eigenvalue equation @xmath443 reduces to a one - dimensional equation because of the @xmath444-functions appearing in the definition of the transfer operator .
this equation can be written in the form @xmath445 here it is implicitly assumed that all variables @xmath446 and @xmath447 as above the largest eigenvalue , @xmath423 as well as the corresponding eigenfunction @xmath448 , calculated at point @xmath449 obeying the same saddle - point condition eq .
, determine all correlation functions in the limit of large @xmath0 .
the joint probability of @xmath58 consecutive spacings takes a different form for @xmath450 and @xmath451 . for @xmath450 @xmath452
while for @xmath451 @xmath453
in the previous section we have derived expressions for the joint distribution of eigenvalue spacings @xmath454 . from these expressions
the @xmath58th nearest - neighbour spacing distribution can be calculated as @xmath455 for @xmath91 these integrals are easily calculable ( e.g. by laplace transform ) and from the joint distribution eq .
we obtain @xmath456 comparison with numerical simulations is displayed at fig .
[ psiv ] . when @xmath92 the joint distribution is given by eq .
what remains is to calculate the largest eigenvalue of the transfer operator @xmath457 , as well as its associated eigenfunction . as @xmath457 is a positive operator ,
the analog of the perron - frobenius theorem states that the eigenvector corresponding to the largest eigenvalue is positive .
orthogonality of the eigenfunctions implies that the converse is also true .
thus if one finds a positive eigenfunction then the corresponding eigenvalue is the largest one .
the eigenvalue equation is equivalent to @xmath458 let us look for solutions of eq .
positive on @xmath459 $ ] under the form @xmath460 , with @xmath94 some unknown complex parameter . since eq .
should hold for all @xmath461 $ ] we get the necessary condition @xmath462 when @xmath463 , eq
. admits two real solutions @xmath464 .
thus @xmath465 with @xmath466 solution of eq . is a positive solution of eq . .
if @xmath467 , eq .
admits two pure imaginary solutions @xmath468 , thus @xmath469 with @xmath466 and @xmath470 solution of eq . is a positive solution of eq . .
finally if @xmath471 , @xmath472 is the unique solution to eq . .
in that case @xmath473 is a solution of eq . which is positive on @xmath459 $ ] .
thus for all @xmath333 we have a positive solution to eq . .
properly normalised , this solution gives the eigenvector @xmath474 .
the corresponding eigenvalue is given by @xmath475 with @xmath94 an implicit function of @xmath333 .
the saddle - point @xmath137 is a solution of eq . . for @xmath476 given by eq
the condition becomes @xmath477 and the saddle - point @xmath137 is obtained from @xmath94 through eq . .
equivalently , this condition can be expressed as @xmath478 .
dashed vertical line indicates the abscissa equal @xmath479 . ] in fig .
[ az ] we plot @xmath9 as a function of @xmath480 . for @xmath481
. has a unique real solution @xmath466 , and @xmath482 is a positive eigenfunction of the transfer operator . for @xmath483 eq .
has a unique pure imaginary solution @xmath470 with @xmath466 .
furthermore in that latter case @xmath4840,\pi[$ ] , so that @xmath485 is an eigenfunction of the transfer operator which is positive on @xmath459 $ ] . at @xmath486
the unique solution is @xmath472 and @xmath487 is a positive eigenfunction of the transfer operator .
the @xmath58th nearest - neighbour spacing distribution can now be calculated from eq . . in the case
@xmath488 it directly gives us the nearest - neighbour spacing distribution @xmath489 , where @xmath53 is the normalisation constant .
it is nonzero only for @xmath490 $ ] , where it takes the following form @xmath491 constants @xmath53 and @xmath94 can be determined either by solving eq . and normalizing the eigenfunction @xmath492 , or equivalently by imposing the normalisation conditions .
the next - to - nearest distribution , @xmath147 is non - zero only when @xmath493 and within this interval it is given by @xmath494 in particular for @xmath486 all integrals can be calculated analytically and @xmath147 has the form @xmath495 in a similar manner one can obtain the higher nearest - neighbour functions .
for example , @xmath148 is given by the formula @xmath496 it is non - zero only when @xmath497 .
in particular for @xmath486 we obtain @xmath498 in fig .
[ psiv ] these formulas are compared with numerical simulations and show a remarkable agreement . for @xmath499 the joint distribution is given by .
the largest eigenvalue and corresponding eigenfunction of the transfer operator are solution of the eigenvalue equation ( [ eigen_equation ] ) . for @xmath500
it takes the form @xmath501 let us look for solutions of the form similar to bethe ansatz @xmath502 where we have set @xmath503 . as eq .
has to be fulfilled for all @xmath504 , this function is a solution of ( [ t_3 ] ) if and only if the following conditions are valid @xmath505 from the first equality in eq
. one can express @xmath60 as a function of @xmath76 and @xmath6 .
after inspection we found that the solutions of the above equations have the following form @xmath506 with real parameters @xmath507 , @xmath508 , and @xmath509 . under this substitution
the eigenfunction ( [ eigenfunction ] ) is transformed to @xmath510 where from @xmath507 , @xmath508 , and @xmath509 must fulfill the following equalities @xmath511 and depend on time @xmath333 through the relation @xmath512 this implies that @xmath513 and , consequently , @xmath507 is related with @xmath508 as follows @xmath514 the eigenfunction corresponding to the largest eigenvalue of the transfer operator is thus given by with @xmath507 , @xmath508 , and @xmath509 real parameters depending on @xmath333 , which must verify and and be such that @xmath515 is a positive function over @xmath459 ^ 2 $ ] .
the saddle - point condition is again given by eq .
( [ point_c ] ) . using
we get a second relation between @xmath507 and @xmath508 , namely @xmath516 equations ( [ x1x2 ] ) and ( [ eqn_a ] ) determine parameters @xmath507 and @xmath508 at a given @xmath9 . to get a positive eigenfunction
@xmath517 it is necessary to get the solutions in the intervals @xmath518 the knowledge of these parameters allows us to calculate the eigenfunction , from which the nearest - neighbour distributions can be deduced through eq . .
the first distributions read @xmath519 and @xmath520 with @xmath53 the normalisation constant @xmath521 these analytical expressions perfectly agree with numerical simulations , as shown in fig .
the expressions for the joint distribution of eigenvalue spacings @xmath454 obtained in section [ transfer_op ] allow to derive formulas for the level compressibility @xmath522 , which characterises the asymptotic behaviour of the number variance .
the number variance @xmath523 is the average variance of the number of energy levels in an interval of length @xmath28 .
it is defined from the two - point correlation function @xmath524 as @xmath525 for systems with intermediate spectral statistics , @xmath526 for large @xmath28 . in order to obtain the large-@xmath0 behaviour of the level compressibility we calculate the laplace transform of the two - point correlation function .
it has a series expansion of the form @xmath527 which allows us to obtain @xmath522 ( see @xcite for more detail ) .
the @xmath58th nearest - neighbour spacing distributions are given by eq . .
summation over @xmath58 gives the two - point correlation function , and its laplace transform is readily obtained , yielding @xmath528 small-@xmath333 expansion of @xmath529 gives @xmath530 the functions @xmath12 are given by eqs . and .
their laplace transform reads @xmath531 where as in the previous sections @xmath532 is the largest eigenvalue of the transfer operator and @xmath533 its associated eigenfunction , both taken at the saddle - point @xmath137 . using the definition of the transfer operator , we see that @xmath534 can be rewritten @xmath535 replacing the transfer operator by its expansion and summing over @xmath58 we get @xmath536 one can check , using normalisation of the eigenfunctions and the saddle - point condition , that the leading - order term is given by @xmath537 .
the next - order term can be simplified using the normalisation of @xmath538 .
it yields @xmath539 from which one gets @xmath540 here @xmath423 is given by ( with @xmath94 depending on @xmath333 through ) , and @xmath137 is given by condition . after calculation
, @xmath522 can be expressed as a function of @xmath94 at the saddle - point .
we get @xmath541 with @xmath94 the real positive solution @xmath542 of for @xmath481 or the pure imaginary solution @xmath470 of for @xmath483 . for @xmath486 , the limit @xmath543 in
gives @xmath544 .
as in the previous case @xmath522 is given by with @xmath423 given by , with @xmath507 , @xmath508 , @xmath509 and @xmath333 related through . from , parameter @xmath507 can be expressed as @xmath545 differentiating both and with respect to time we obtain @xmath546 and @xmath547 as a function of @xmath507 and @xmath508 , and then similarly @xmath548 and @xmath549 .
using , the saddle - point condition can be rewritten @xmath550 and from @xmath522 can then be expressed as @xmath551 using the expression obtained @xmath548 we finally obtain @xmath522 as a function of @xmath507 and @xmath508 , with @xmath507 , @xmath508 obtained as solution of . inverting we get @xmath552 after some manipulation @xmath522 simplifies to @xmath553\ .\nonumber\end{aligned}\ ] ] figure [ chi_ruij ] is a plot of the level compressibility @xmath522 . the theoretical prediction obtained from , and agrees with numerical data . .
black diamonds are the numerical values extracted from a cubic fit of @xmath523 over the range @xmath554 $ ] , with @xmath523 the variance of the number of levels in an interval of length @xmath28 averaged over 50 windows of length @xmath28 , calculated from the unfolded spectrum with mean level spacing @xmath394 , for 10000 realisations of the random matrix with matrix size @xmath196 .
red curve is the theoretical prediction , and .
[ chi_ruij ] ] for integer @xmath9 the spectrum is rigid and thus the level compressibility is expected to take the value 0 . here
we consider the first - order expansion of @xmath522 in the vicinity of integer @xmath9 .
we will show that at lowest order the expansion of @xmath522 around @xmath555 is given by @xmath556 .
we first consider the expansion around @xmath557 .
let @xmath558 .
for @xmath559 we have @xmath560 , thus expansion is trivial . for @xmath561 @xmath522
is given by with @xmath9 and @xmath562 related by .
at @xmath557 the solution of is @xmath563 .
an asymptotic expansion of and yields @xmath564 where @xmath565 is some rational fraction in @xmath562 .
thus at first order @xmath566 expanding @xmath522 for large @xmath562 to lowest order gives @xmath567 with @xmath568 is some rational fraction in @xmath562 .
using one gets @xmath569 .
suppose now that @xmath570 .
for @xmath571 @xmath522 is given by , with @xmath572 solution of .
equivalently , @xmath522 is given by @xmath573 with @xmath562 the real positive solution of @xmath574 at point @xmath575 the solution is @xmath576 . expanding both sides of at lowest order in @xmath577 with @xmath578
we get @xmath579 . inserting this expansion for @xmath562 in
we obtain @xmath580 . for @xmath581 @xmath522
is given by , with @xmath507 and @xmath508 specified by . at @xmath575 , we have @xmath582 which vanishes for @xmath583 . for @xmath584
we have the expansion @xmath585 equation is equivalent to @xmath586 and the small-@xmath577 expansion of both members of this equation reads @xmath587 ( we use to obtain the expansion of @xmath507 ) .
this implies that the two first terms in the expansion must vanish , thus @xmath588 and @xmath589 .
putting these values into gives @xmath580 .
the same result can be obtained from and for @xmath590 . at @xmath591 again @xmath508 takes the value @xmath479 , and an expansion @xmath592 gives @xmath593 , whence @xmath594 .
in this paper we construct new random matrix ensembles with unusual properties .
random matrices from these ensembles are lax matrices of @xmath0-body integrable classical systems with a certain measure of momenta and coordinates .
though such matrices are not invariant over rotation of the basis ( as usual random matrix ensembles ) the joint distribution of their eigenvalues can be calculated analytically .
four different models are considered in detail .
three of them correspond to rational , hyperbolic , and trigonometric calogero - moser models .
the fourth is related to the trigonometric ruijsenaars - schneider model . for the trigonometric calogero - moser model and the ruijsenaars - schneider model spectral correlation functions
are calculated explicitly . for rational and hyperbolic calogero - moser models
wigner - type surmises are proposed .
our formulas are in a good agreement with results of direct numerical calculations .
in this appendix we check that the action - angle variables @xmath38 and @xmath120 for model cm@xmath42 verify hamilton - jacobi equations by calculating their time derivative .
we use the fact that for the lax pair @xmath595 the matrix @xmath1 can be seen as a time derivative operator for the eigenfunctions of the matrix @xmath28 .
namely , if @xmath596 is a normalised eigenvector of @xmath28 , then @xmath597 ( here @xmath598 denotes complex conjugation ) . for model cm@xmath42 , the lax matrix @xmath1 is given by @xmath599 one can easily check that for @xmath600 one has @xmath601 deriving with respect to time , using the definition of @xmath123 , yields @xmath602 \ .\ ] ] from hamilton - jacobi equations @xmath603 . using , we get that the time derivative of @xmath120 is given by @xmath604u_r(\alpha ) & = & \sum_{k , r}u_k^{*}(\alpha ) l_{kr } u_r(\alpha)=\lambda_{\alpha}\ .\end{aligned}\ ] ] the time derivative of @xmath38 is easily obtained from , and , yielding @xmath605 .
this shows that the @xmath38 and @xmath120 verify hamilton - jacobi equations .
the purpose of the appendix is to give , for completeness , the proofs of certain often used formulas .
let coefficients @xmath606 obey the following system of linear equations for all @xmath607 with known @xmath608 and @xmath609 @xmath610 then @xmath606 for all @xmath611 are expressed through @xmath608 and @xmath609 by using e.g. the cauchy determinants @xmath612 the following identities are also useful . for all @xmath613 one has @xmath614 @xmath615 and @xmath616 a simple way to check ( [ bm ] ) is to consider the function @xmath617 asymptotically this function decreases as @xmath618 when @xmath619 , so that the integral over a large contour encircling all poles equals 1 . rewriting this integral as the sum over all finite poles gives @xmath620 which proves ( [ bm ] ) .
the equality ( [ sum_bm2 ] ) can be obtained by the integration of the function @xmath621 over a contour which includes all poles . as this
function decreases as @xmath622 when @xmath619 the integral equals zero .
taking the sum over poles at @xmath623 with all @xmath607 and at @xmath624 one verifies ( [ sum_bm2 ] ) .
let us now consider the function @xmath626 it decreases as @xmath618 when @xmath619 and has poles at @xmath627 and @xmath628 with @xmath611 . integrating it over a contour encircling all poles one obtains ( [ sum_norm ] ) .
99 t. guhr , a. mller - groeling , h. a. weidenmller , _ random matrix theories in quantum physics : common concepts _ , phys . rep . * 299 * , 189 ( 1998 ) .
o. bohigas , m .- j .
giannoni and c. schmit , _ characterization of chaotic quantum spectra and universality of level fluctuation laws _ , phys .
* 52 * , 1 ( 1984 ) .
m. v. berry and m. tabor , _ level clustering in the regular spectrum _ , proc .
a * 356 * , 375 ( 1977 ) .
m. l. mehta , _ random matrix theory _ , springer , new york ( 1990 ) .
a. altland and m. r. zirnbauer , _ novel symmetry classes in mesoscopic + normal - superconducting hybrid structures _ , phys .
b * 55 * , 1142 ( 1997 ) . a. y. abul - magd and m. h. simbel , _ wigner surmise for high - order level spacing distributions of chaotic systems _ , phys .
e * 60 * , 5371 ( 1999 ) .
p. w. anderson , _ absence of diffusion in certain random lattices _ , phys
109 * , 1492 ( 1958 ) .
b. i. shklovskii , b. shapiro , b. r. sears , p. lambrianides , and h. b. shore , _ statistics of spectra of disordered systems near the metal - insulator transition _ , phys .
b * 47 * , 11487 ( 1993 ) .
e. b. bogomolny , o. giraud and c. schmit , _ nearest - neighbor distribution for singular billiards _ ,
e * 65 * , 056214 ( 2002 ) .
o. giraud , j. marklof and s. okeefe , _ intermediate statistics in quantum maps _
, j. phys .
a * 37 * , l303 ( 2004 ) .
b. huckestein , _ scaling theory of the integer quantum hall effect _ , rev .
* 67 * , 357 ( 1995 ) .
e. bogomolny , u. gerland , and c. schmit , _ models of intermediate spectral statistics _ ,
phys . rev .
e * 59 * , r1315 ( 1999 ) a. d. mirlin , y. v. fyodorov , f .-
dittes , j. quezada , and t. h. seligman , _ transition from localized to extended eigenstates in the ensemble of power - law random banded matrices _ , phys .
e * 54 * , 3221 ( 1996 ) . v. e. kravtsov and k. a. muttalib , _ new class of random matrix ensembles with multifractal eigenvectors _ , phys .
lett . * 79 * , 1913 ( 1997 ) .
e. bogomolny , o. giraud , and c. schmit , _ random matrix ensembles associated with lax matrices _ , phys .
lett . * 103 * , 054103 ( 2009 ) .
e. bogomolny and o. giraud , _
perturbation approach to fractal dimensions for certain critical random matrix ensembles _ , in preparation ( 2011 ) .
lax , _ integrals of nonlinear equations of evolution and solitary waves _ , comm .
pure applied math . *
21 * , 467 ( 1968 ) .
ruijsenaars , _ action - angle maps and scattering theory for some finite - dimensional integrable systems i. the pure soliton case _ , commun .
115 * , 127 ( 1988 ) .
s.n.m . ruijsenaars , _ action - angle maps and scattering theory for some finite - dimensional integrable systems ii .
solitons , antisolitons , and their bound states _ , publ .
rims , kyoto univ . *
30 * , 865 ( 1994 ) .
s. ruijsenaars , _ action - angle maps and scattering theory for some finite - dimensional integrable systems , iii .
sutherland type systems and their duals _ , publ .
rims , kyoto univ . *
31 * , 247 ( 1995 ) .
f. calogero , _ solution of the one - dimensional n - body problem with quadratic and/or inversely quadratic pair potentials _ , j. math .
* 12 * , 419 ( 1971 ) ; erratum : j. math . phys . * 37 * , 3646 ( 1996 ) .
j. moser , _ three integrable hamiltonian systems connected with isospectral deformations _
, advances in math .
, * 16 * , 197 ( 1975 ) .
ruijsenaars and h. schneider , _ a new class of integrable systems and its relation to solitons_. ann . phys .
( ny ) * 170 * , 370 ( 1986 ) .
olshanetsky and a.m. perelomov , _ classical integrable finite - dimensional systems related to lie algebras _ , phys
. rep . * 71 * , 313 ( 1981 ) .
e. dhoker and d.h .
phong , _ lax pairs and spectral curves for calogero - moser and spin calogero - moser systems _ , arxiv : hep - th/9903002 , ( 1999 ) , regul .
chaotic dyn . * 3 * , 27 ( 1998 ) .
krichever , _ elliptic solutions of the kadomtsev - petviashvili equation and integrable systems of particles _ , func .
* 14 * , 282 ( 1980 ) .
e. bogomolny and c. schmit , _ spectral statistics of a quantum interval - exchange map _ , phys .
93 * , 254102 ( 2004 ) .
e. bogomolny , r. dubertrand , and c. schmit , _ spectral statistics of a quantum interval - exchange map : the general case _ , nonlinearity , * 22 * , 2101 ( 2009 ) .
e. bogomolny , u. gerland , and c. schmit , _ short - range plasma model for intermediate spectral statistics _ ,
j. b * 19 * , 121 ( 2001 ) .
ruijsenaars , _ complete integrability of relativistic calogero - moser systems and elliptic function identities _ ,
phys . * 110 * , 191 ( 1987 ) .
j. gibbons and t. hermsen , _ a generalization of the calogero - moser system _ , physica * 11d * , 337 ( 1984 ) . | we propose new classes of random matrix ensembles whose statistical properties are intermediate between statistics of wigner - dyson random matrices and poisson statistics .
the construction is based on integrable @xmath0-body classical systems with a random distribution of momenta and coordinates of the particles .
the lax matrices of these systems yield random matrix ensembles whose joint distribution of eigenvalues can be calculated analytically thanks to integrability of the underlying system .
formulas for spacing distributions and level compressibility are obtained for various instances of such ensembles . |
the purpose of whi jop204 was to characterize the behavior of the quiet sun ; aiming to build a better picture of the relationship between chromospheric dynamic spicules , their extension into the transition region and ( possibly ) corona .
in essence we hope to fully assess the role of magneto - convective driving mass and energy from the lower atmosphere into the quiet corona or solar wind .
jop 204 observations took place between april 10th through april 16th , 2008 involving a host of observatories on the ground ( nso ) and in space ( _ soho _ , _ trace _ , _ hinode _ , _ stereo _ ) .
we will discuss the preliminary analysis of data from _ hinode_/sot , _ trace _ and _
soho_/sumer @xcite on april 12th 2008 on the south - west limb that ran from 14:00 - 17:00ut . the correspondence between above - the - limb structure in _
ca ii h and trace 1600 passband image sequences is striking - noting that the 1600 passband is dominated by 1548 and 1550 emission lines of c iv there .
immediately , we see that while the emission observed in ca ii is considerably finer in structure the gross behavior and dynamics present are repeated in the mid - transition region indicating that material is at least heated to those temperatures as a spicule . by far
the most easily identifiable event to transpire was the passage of a large macro - spicule across the sumer slit at about 16:08ut , see figure f1 .
space - time analysis of the macro - spicule in the sot and trace image sequences shows that the ejected material follows a parabolic trajectory ( characteristic of type - i spicules ) with super - sonic launch and receding velocities of @xmath040km / s seen some 7 minutes later .
this event is clearly on a much larger spatial scale than the typical type - i and ii spicules which dominate the limb chromosphere and transition region .
we suggest that the structure seen is a macro - spicule driven by the large - scale reconnection of the photospheric magnetic field on the supergranular scale ( e.g. , * ? ? ?
* ) with the material driven upward by a shock , given the parabolic spicule evolution observed studying the dynamic behavior of this event is made complex by the non - trivial evolution of the line profiles with the increasing formation temperatures of the lines as shown in figs .
f1 and f2 . the right side of fig .
f1 shows spectral snapshots while fig .
f2 shows the space - time evolution of the fitted ( single gaussian fit ) spectra in the same lines .
we see the upward phase of the event in _ all _ the emission lines as the line centers are blue - shift by @xmath040km / s - consistent with the measurements from sot - indeed the cooler emission ( n iv and o vi ) show parabolic behavior in the space - time plots .
further , we also see that the profiles are asymmetric and get significantly broadened ( @xmath060km / s , a non - thermal line width of @xmath040km / s at these wavelengths - with sumer detector b ) over the entire duration of the event , indicative of unresolved velocities or active heating of the structure . perhaps the most startling thing , and
best seen in the bottom row of fig .
f2 , is that the ne viii emission of the macro - spicule is extremely weak ( only a few of percent of the background ne viii emission above the limb ) but as the material starts to fall , and we start to see red - shift , the excess hot emission is _ gone _ ; both the doppler and line width signature of the macro - spicule disappears abruptly while the emission from the cooler lines show downward motion .
the apparent disappearance of the hot emission provides a lot of questions for subsequent work and motivates the discussion below .
we have been able to successfully co - align data from _ hinode_/sot , _ trace _ and _
soho_/sumer for the whi qs period on april 12 2008 .
the data provide a tantalizing look at the dynamic evolution of ejected transition material and its interaction with the corona which surrounds it @xcite , the macro - spicule observed clearly reaches the formation temperature of ne viii ( 600,000k ; * ? ? ?
further , we speculate that the hottest material seen , based on the lack of returning ne viii emission , has actually become a constituent of the corona .
it is probably not a coincidence that this emission line is formed above the temperature at which thermal conduction is critical in the energy balance of the plasma @xcite and that ne viii is the coolest transition region species for which we see outflow in coronal holes ( e.g. , * ? ? ?
we suggest that the physical process required for such a `` hot mass - injection '' to the coronal magnetic system is intrinsically not one that can be studied in an single - fluid ideal mhd framework .
subsequent work will show that this is the proverbial tip of the iceberg in terms of our understanding of the transition region and its role in the heating and mass - loading of the outer solar atmosphere .
we would like to thank alfred de wijn ( sot ) , werner curdt ( sumer ) and dawn myers ( _ trace _ ) for conducting the observations discussed .
the work presented here is supported by a grant nnx08al22 g to swm and bdp .
bdp is supported by the nasa _ hinode _ grant to lmsal ( nnm07aa01c).the national center for atmospheric research is sponsored by the national science foundation . | we explore joint observations of the south - east limb made by _ hinode _ , _ trace _ and _ soho_/sumer on april 12 , 2008 as part of the whole heliosphere interval ( whi ) quiet sun characterization targeted observing program . during the sequence a large , 10 mm long
, macro - spicule was sent upward and crossed the line - of - sight of the sumer slit , an event that affords us an opportunity to study the coupling of cooler chromospheric material to transition region emission formed as hot as 600,000k .
this short article provides preliminary results of the data analysis . the solar optical telescope ( sot ; * ? ? ?
* ) of _ hinode _
@xcite has revolutionized our view of the dynamic chromosphere revealing the existence of at least two types of spicule @xcite . ``
type - i '' spicules are long - lived ( 3 - 5 minutes ) and exhibit longitudinal motions of the order of 20km / s that are driven by shocks resulting from the leakage of p - modes in regions around strong magnetic flux concentrations @xcite .
`` type - ii '' spicules , on the other hand , are a relative unknown that show much shorter lifetimes ( 50 - 100s ) , higher velocities ( @xmath0100 km / s ) , are considerably taller ( 5 - 8 mm ) and rarely exhibit any recession ( downfall ) of material that is ejected upward .
it has been demonstrated that both types of spicules undergo alfvnic motion @xcite .
however , in terms of chromosphere / transition region connectivity , there are several questions that are inadequately explained at present , but may be critical in understanding the interface of the cool and hot solar atmospheres : + 1 .
what is the relationship between chromospheric spicules and the emitting transition region structures observed above the limb ?
+ 2 .
how are these spicules heated and to what temperature ?
+ 3 . is the transition region really composed of two physically different `` components '' and do these spicule types play a role ?
+ 4 . what mechanism drives type - ii spicules ?
+ 5 .
do the unresolved transverse and longitudinal motions of the spicules produce the observed non - thermal line widths ?
+ these are points we wish to address with the whi joint observations and the detailed work is in its early stages .
the present paper discusses some of the interesting preliminary analysis pertaining to points 1 and 2 . |
the electronic structure and spectroscopic properties of quantum dots , including excitons,@xcite charged excitons,@xcite multiexcitons,@xcite and excitonic fine - structure,@xcite all depend on the size and shape of the dots .
this dependence reflects both quantum confinement effects , as well as shape - induced band - folding and inter - band coupling.@xcite _ simulations _ of electronic structure and spectroscopic properties of quantum dots@xcite must naturally assume the size and shape of the dot .
on the other hand , _ measurements _ of spectroscopic properties of a dot@xcite are rarely accompanied by accurate measurements of the size and shape of the gaas - covered dot , except in rare cases where detailed cross - sectional scanning tunneling microscopy experiments are performed , such as in refs . .
this situation creates a significant difficulty , if not a crisis , in interpreting spectroscopic data on quantum dots , and in critically testing various theoretical approaches .
thus , in reality one is often forced to address the inverse problem,@xcite namely , fit the spectroscopic data to a theoretical model by using the size and shape as adjustable parameters .
the difficulty with this approach is threefold : first , in this approach all theories , no matter how nave , ultimately work by virtue of forcing a fit to experiment , even if the assumptions entering the theory may seem unjustified in their own right ( e.g. assuming single - band effective mass ; neglect of strain ; neglect of spin - orbit coupling . ) second , there are usually too many free - parameters , which involve not only non - trivial shapes , but also unknown composition profiles ( e. g. , in@xmath6ga@xmath7as dots ) .
third , since the relationship between shape and spectroscopic properties is model dependent , unrealistic shapes are often deduced .
for example , in simple effective - mass,@xcite k@xmath2p,@xcite or parabolic models the `` p '' and `` d '' levels of electrons and holes are degenerate if one assumes spherical , lens - shaped , cubic , or cylindrical dots .
in contrast , in more advanced
_ atomistic _
models like empirical pseudopotential@xcite or tight binding@xcite those levels are split even for the above mentioned ideal shapes , resulting in clear spectroscopic signatures . to fit measured spectroscopic signatures of actual dots by simple theoretical models , one needs to assume at the outset irregular shapes .
for instance , dekel _
et al._@xcite needed to assume a parallepipedal box to explain their multiexciton data on non - alloyed inas / gaas dots ; and ferreira assumed shape distortions to explain fine structure.@xcite such assumptions are not needed in atomistic approaches to modeling . in this work we have used a high - level atomistic approach to predict the spectroscopic characteristics of in@xmath8ga@xmath1as / gaas self - assembled dots as a function of the most crucial geometric parameter , namely the heigh .
we calculate strain profiles , `` p- '' and `` d-''level splittings as well as electron and hole wavefunctions .
while ultimately it will be necessary for experimentalists to report the size , shape , and composition profile to which their spectroscopic data correspond , the type of study reported here may be used to bridge , in the interim , spectroscopy with theory without clouding the issue by severe theoretical approximations .
similar studies were carried out by the group of bimberg in refs .
, by shumway _
_ in ref . , by williamson _
_ in ref . , by kim _
_ in ref . and by pryor in ref .
among several geometries , stranski - krastanow quantum dots growth in lens shape.@xcite hence , we focus on lens - shaped , self - assembled in@xmath8ga@xmath1as / gaas quantum dots ( qds ) . in addition , we include a 1-monolayer - thick in@xmath8ga@xmath1as wetting layer ( wl ) .
figure 1 shows a sketch of the geometry of the nanostructure .
( qd+wl+gaas matrix ) .
we focus on dots with @xmath9 and pure inas .
the qds have circular base with diameter @xmath10 and height @xmath11 in the interval @xmath12-@xmath13 .
sketch of the geometry of self - assembled in@xmath0ga@xmath1as / gaas quantum dots .
a section of the gaas matrix ( grey ) has been ommited for clarity .
the dot ( white ) is lens shaped with base diameter @xmath14 and height @xmath11 .
the dot lies on top of a 1-monolayer - thick in@xmath8ga@xmath1as wetting layer .
two auxiliary features are shown : ( i ) a plane that bisects the dot ( dashed lines ) and ( ii ) the line @xmath15 , which pierces the dot through the center along the [ 001 ] direction.,width=283 ] the position - dependent strain profiles present in the nanostructure are usually addressed within ( harmonic ) _ continuum _ elasticity ( ce ) theory.@xcite notwithstanding , strain profiles can also be calculated within atomistic elasticity ( ae ) in the form of a valence force field.@xcite in ae the strain energy is expressed in terms of atomic positions and interatomic force constants . while the force constants are fit to the elastic constants @xmath16 , @xmath17 and @xmath18 , much like in ce , expressing the strain in terms of atomic positions contains much more information than expressing strain in terms of ce coordinates .
for example , a square - based pyramid has a @xmath19 point group symmetry if described by ce .
this means that the @xmath5 $ ] and @xmath20 $ ] facets are taken to be equivalent . in this case , all `` p '' energy levels are degenerate .
however , in ae the point group symmetry of a square - based pyramid made of inas is @xmath21 . in this case ,
the [ 110 ] and @xmath20 $ ] facets are symmetry - inequivalent and `` p '' levels are split
. in general , three interrelated sources are responsible for the splitting of the `` p '' states.@xcite a quantum dot with a base that globally possesses inversion symmetry in the ( 001 ) plane like a lens , a truncated pyramid or a truncated cone has in reality the lower @xmath21 symmetry ( for pure , non - alloyed inas / gaas dots ) , which originates from the underlying zinc - blende atomic lattice .
thus , the _ first source _ is the lack of inversion symmetry in the @xmath21 point group , which manifests itself around the dot - matrix interfaces where the [ 110 ] and [ 1@xmath220 ] directions are inequivalent .
the _ second source _ is the propagation of the strain field towards the center of the dot as a consequence of the atomic relaxation .
the _ third source _ is the piezoelectric effect ( which magnitude is presently under debate@xcite ) that arises from the strain field of @xmath21 symmetry .
it should be noted that approximations used in previous calculations of the piezoelectric effect @xcite have been shown to be crude and further investigations have been called for.@xcite our choice of ae is based on a generalization of keating model to 3 terms bond - bending , bond stretching and their cross terms.@xcite we fit the elastic constants @xmath16 , @xmath17 and @xmath18 of zinc - blende inas and gaas , as well as the correct dependence of the young s modulus with pressure for both materials . in the in@xmath8ga@xmath1as alloy system ,
the bond - stretching and the bond - stretching / bond - bending cross - term parameters for the mixed cation ga - as - in bond angle are taken as the algebraic average of the in - as - in and ga - as - ga values . while the ideal bond angle is @xmath23 for the pure zinc - blende crystal , to satisfy vegard s law for the alloy volume , the value of @xmath24 was used for the mixed bond angle .
ae is superior to ce in that it does not assume harmonicity in fact , anharmonic effects can be explicitly included@xcite into the valence force field.@xcite here , we calculate the position - dependent strain tensor @xmath25 within the atomistic elasticity approach . to calculate the strain tensor cubic components @xmath26 ( @xmath27 , @xmath28 , @xmath29 )
, we proceed in two steps : ( 1 ) we relax the atomic positions within the supercell in order to minimize the elastic energy , which is given by a generalized valence force field.@xcite ( 2 ) we relate the relaxed ( equilibrium ) atomic positions with the unrelaxed atomic positions via strain tensor.@xcite at each equilibrium position @xmath30 of atom @xmath31 we identify the tetrahedron formed by its 4 nearest neighbors .
this tetrahedron is distorted in comparison to the unrelaxed tetrahedron .
thus , the three edges of these tetrahedra that are determined by the vectors connecting the four neighbors can be related by the strain tensor @xmath32 as follows . @xmath33 where @xmath34 , @xmath35 , and @xmath36 are the 3 vectors that connect , respectively , neighbors 1 and 2 , 2 and 3 , and 3 and 4 in the equilibrium , distorted tetrahedron that encloses atom @xmath31 .
@xmath37 , @xmath38 , and @xmath39 are the corresponding vectors ( edges ) in the unrelaxed tetrahedron . from eq .
( [ strain.eq ] ) , the cubic strain tensor components are calculated by a matrix inversion .
figure [ fig_2 ] shows the biaxial strain @xmath40 and the hydrostatic distortion ( isotropic strain ) @xmath41 as a function of dot height .
the results presented correspond to the strain values on a planar section in the [ 100 ] direction that bisects the dot ( see fig .
[ fig_1 ] ) .
the spikes in the dot and wetting layer region are a consequence of the random arrangement of in and ga atoms .
thus , the spikes structure significantly changes from dot to dot .
two additional features are also prominent .
( i ) the biaxial strain at the dot - matrix interface , located near the top of the dot , increases with height and slowly decays off .
this decay is faster as the dot becomes taller .
the latter can be understood by considering the dot as a spherical cap inscribed in a sphere of radius @xmath42/2h$ ] ( the _ taller _ the dot , the _ smaller _ @xmath43 ) and by assuming that the decay of @xmath44 outside the dot is qualitatively well described by @xmath45 , which is the radial dependence of the biaxial strain outside a continuum , elastic sphere of radius @xmath43 embedded in a lattice - mismatched medium.@xcite here , @xmath46 is the radial distance from the sphere surface . in our simulations , @xmath47 shows a similar trend as @xmath44 .
( ii ) due to the underlying zincblende atomic structure , the biaxial strain drops to zero along the @xmath48 $ ] and @xmath49 $ ] directions , as well as along their equivalent crystallographic directions .
( see white arrows in fig .
[ fig_2 ] . )
this feature is robust upon changes in height , being present for all the quantum dots we considered . to calculate the energies and wavefunctions of electron and hole states in the quatum dot
, we use the empirical pseudotential method of wang and zunger.@xcite this approach combines a pseudopotential description of the single - particle hamiltonian with the linear combination of bulk bands ( lcbb ) method to solve for the energies and wavefunctions.@xcite in this method , the hamiltonian reads @xmath51 where @xmath52 is an empirical parameter that accounts for non - locality effects ; @xmath53 is a non - local empirical operator that describes the spin - orbit interaction;@xcite @xmath54 is a screened pseudopotential ( for atom of type @xmath55 ) that depends on strain ; and @xmath56 is the relaxed vector position of atom @xmath31 of type @xmath55 .
the dependence of the atomic pseudopotential on strain transfers to the electronic hamiltonian the information on atomic displacements .
the strain - dependent pseudopotential reads @xmath57,\ ] ] where @xmath58 is a fitting parameter .
it should be noted that @xmath59 is fit to _ bulk _ properties of gaas and inas , including bulk band structures , experimental deformation potentials and effective masses , and lda - determined band offsets . in order to improve the transferability of pseudopotential @xmath59 , a simple dependence on the chemical environment of atom @xmath55 is introduced .
for instance , for @xmath60 in an environment of @xmath61 ga atoms and @xmath62 in atoms we use @xmath63 the pseudopotentials used in this work have been successfully tested for quantum wells.@xcite the wavefunction of state @xmath64 is @xmath65 , which satisfies @xmath66 , is expanded in bulk bloch states @xmath67 of material @xmath68 .
[ it should be noted that the bulk materials ( @xmath68 ) can be strained.@xcite ] namely , @xmath69,\ ] ] where , @xmath70 and @xmath71 indicate the band - index and wave - vector of the bloch state , respectively ; and @xmath72 is the number of primary cells contained in a supercell that encloses the quantum dot .
thus , by diagonalizing the hamiltonian [ eq .
( [ hamiltonian.lcbb ] ) ] in the bloch states basis we find the coefficients @xmath73 .
the calculated wavefunctions @xmath74 are 2-fold , kramers degenerate , so we have omitted the spin index @xmath75 .
this representation for the wave function is very different from the familiar k@xmath2p method in that in the latter approach the basis set is constructed only from states near the brillouin zone center ( @xmath76 ) , while here we use a full - zone description .
further , in k@xmath2p one is restricted to just one or two bloch bands at @xmath76 ( @xmath77 representating two bloch bands ) while here we consider @xmath70 bloch bands . in appendix
[ levels.app ] , we present an assessment of the convergence of energy levels @xmath78 as a function of the expansion parameters in eq .
( [ psi.lcbb ] ) . while we describe the effects of strain on the electronic structure atomisticallyas shown in eq .
[ hamiltonian.lcbb ] , here , for _ illustrative purposes only _ , we calculate the conduction ( electron ) and valence ( hole ) strain - modified band offsets ( confining potentials ) present in the quantum dot by coupling strain to k@xmath2p - like equations . at each 8-atom unit cell in the supercell we diagonalize the conduction and valence ( including spin - orbit coupling ) band strain hamiltonians that wei and zunger put forward in ref . .
figure [ fig_3 ] shows the calculated strain - modified band offsets for electrons and holes along the line @xmath15 indicated in fig .
[ fig_1 ] .
electron band offset appears in the upper panel , whereas thick and thin lines in the lower panel show the first and second holes offsets , respectively .
we note the following features .
( i ) the band offsets ( fig .
[ fig_3 ] ) inherit the jagged nature of the strain fields ( fig .
[ fig_2 ] ) . inside the quantum dot ( region `` d '' in fig .
[ fig_3 ] ) , alloy fluctuations lead to a small mixing in the heavy - hole ( hh ) and light - hole ( lh ) character of the band offsets .
( ii ) regardless of the dot height , the higher energy hole band offset has hh character inside the dot and lh outside .
conversely , the lower energy hole band offset has lh character inside and hh outside .
( iii ) the increase of biaxial strain at the dot - matrix interface that occurs as the height increases ( fig .
[ fig_2 ] ) is reflected in the increase ( decrease ) of the higher ( lower ) energy hole band offset .
( see arrows in fig .
[ fig_3 ] ) .
in particular , for tall dots , the decrease of the lower energy hole offset at the interface leads to the formation of a _ pocket _ in the band offset .
the single - band effective mass model predicts that in the continuum limit , in which lens - shaped dots have full axial symmetry , the confined energy levels will form shells of degenerate states with _
definite _ angular momentum @xmath79 .
@xcitethus , the _ orbital _
character of the confined levels will be _ pure _ `` s '' , `` p '' , `` d '' , etc .
however , when the correct symmetry of the dot is taken in to account , as it is the case when using our atomistic pseudopotential method , the confined energy levels can have mixed orbital character . to quantify the degree of mixing ,
we analyse the single - particle wavefunctions @xmath74 by projecting their _ envelopes _ on eigenfunctions of the axial ( @xmath80 , see fig .
[ fig_1 ] ) angular momentum @xmath81 and determine the orbital ( angular ) character .
the latter given by @xmath82 in eq .
( [ angular.char ] ) , we have used cylindrical coordinates , i. e. @xmath83 , and written the @xmath76-derived envelope function as @xmath84 the bloch character of the hole wavefunctions @xmath85 in the quantum dot depends on strain as well as band coupling . to study the extent to which the bloch character is hh , lh , and so , we proceed similarly and project the envelope of @xmath85 on the total - angular - momentum basis @xmath86 ( @xmath87 ) .
we calculate the exciton energy levels @xmath88 by using the configuration interaction method as proposed in ref . .
briefly , in this method , the exciton states @xmath89 where @xmath90 denotes a basis of single - substitution slater determinants ( configurations ) , in which an electron is promoted from @xmath91 to @xmath92 .
the ci method would deliver the exact exciton ground and excited states in the case of a complete basis .
however , we _ truncate _ this basis and consider all the possible configurations build out of @xmath93 electron and @xmath94 hole confined states .
the coefficients @xmath95 arise from the diagonalization of the exciton hamiltonian .
the direct ( @xmath96 ) and exchange ( @xmath97 ) electron - hole coulomb integrals that enter the calculation are derived from the coulomb scattering matrix elements @xmath98^{*}\big[\psi^{(e)}_j({\bf r}^{\prime})\big]^{*}\psi^{(e)}_{k}({\bf r}^{\prime})\psi^{(h)}_{l}({\bf r})}{\epsilon({\bf r},{\bf r}^{\prime})|{\bf r}-{\bf r}^{\prime}|}\ ] ] the microscopic , phenomenological dielectric constant @xmath99 that screens the interaction is calculated within the thomas - fermi model proposed by resta.@xcite we do not use simple perturbation theory where the exciton is described via a coulomb - corrected single - particle band gap @xmath100 ; here , @xmath101 and @xmath102 are the energies of electron level @xmath64 and hole level @xmath103 , respectively .
we now present the effects of height and composition on the energies of confined levels and their splittings , wavefunctions of selected confined electron and hole levels , and the lowest transition energy of the exciton .
the confined electron and hole energy levels are respectively labeled as @xmath104 and @xmath105 , where @xmath106 is an orbital quantum number .
the corresponding wavefunctions are @xmath107 and @xmath108 . each of the confined states are 2-fold , kramers degenerate .
we label the @xmath109 electron and hole states as lumo and homo , respectively .
the first 20 electron and hole energy levels of an in@xmath110ga@xmath111as / gaas dot appear in figure [ fig_4 ] .
the electron and hole energies are measured , respectively , from the bottom of the conduction band ( cbm ) @xmath112 ( calculated bulk electron affinity ) and from the top of the valence band ( vbm ) @xmath113 ( calculated bulk ionization potential ) of bulk gaas . for all heights ,
the confined electron states form groups ( _ shells _ ) of _ non - degenerate _
levels.@xcite in turn , this shell structure can only be identified for the first few hole levels ( near homo ) in shallow dots ( up to @xmath114 ) . for taller dots ,
the holes show no shell structure .
the number of confined electron states is significantly smaller than that of holes .
while for the tallest dot there are 10 confined electron states , more than 150 hole states are confined in all the considered dots .
when comparing the results for @xmath115 with the first 20 electron and hole energy levels ( not shown ) in a lens - shaped , non - alloyed inas / gaas dot , we find that ( i ) the electron levels form shells that have a bigger average separation ( @xmath116 vs @xmath117 in in@xmath118ga@xmath111as / gaas ) , ( ii ) the hole energy level structure near the homo is significantly different ( see splittings in fig . [ fig_5 ] ) and ( iii ) the single - particle gap @xmath119 is smaller ( see fig .
[ fig_9 ] ) .
first 20 electron ( top panel ) and hole ( bottom ) energy levels @xmath120 and @xmath121 for in@xmath122ga@xmath111as / gaas dots with different heights .
the electron and hole energies are measured from gaas cbm [ @xmath112 ] and gaas vbm [ @xmath124 , respectively . as height increases , the number of confined levels ( @xmath125 and @xmath126 ) increases and the single - particle gap ( @xmath127 ) decreases .
the confined electronic energy levels group in _ non - degenerate shells _ for all dots .
`` s '' , `` p '' , and `` d '' indicate the predominant orbital character of selected states . for holes ,
up to @xmath114 the second and third levels are `` p''-like , while for larger heights these levels are `` s''-``p '' hybridized.,width=321 ] _ energy splittings._the `` s - p '' energy splitting @xmath128 and `` p - p '' splitting @xmath129 for electrons and holes are shown in figure [ fig_7 ] as a function of the dot height .
( the dots single - particle gap is also indicated . )
two features emerge .
first , `` s - p '' splitting for electrons is bigger than for holes . for electrons ,
the magnitude of the splitting is about @xmath117 and depends weakly on height . on the other hand , for holes ,
this splitting changes from @xmath130 to nearly @xmath131 when height changes from @xmath12 to @xmath13 .
second , the `` p - p '' splitting shows the opposite behavior .
namely , for holes the magnitude of the splitting remains nearly constant at approximately @xmath132 , and for electrons it changes non - monotonically .
it should be noted that the electronic `` p - p '' splitting is sensitive to alloy fluctuations and it can change by almost a factor of two by changing the alloy realization in the dot.@xcite for comparison , `` s - p '' and `` p - p '' splittings in a lens - shaped , non - alloyed inas / gaas dots with @xmath10 also appear in fig . [ fig_5 ] .
the pure dot has a `` s - p '' splitting for holes that is nearly the same as in the @xmath115 in@xmath118ga@xmath111as / gaas dot , while this splitting for electrons is nearly 20% bigger and depends weakly on height . in contrast ,
the `` p - p '' splitting for holes in the pure ( non - alloyed inas / gaas ) dot is about twice as big as in the @xmath115 alloy dot , and for electrons the `` p - p '' splitting depends weakly on height and is similar in magnitude to the splittings in alloy dots .
it should be noted that in pure ( non - alloyed inas / gaas ) dots the _ hole _ energy levels undergo a localization crossover ( see discussion in sec .
[ local.strain ] ) for tall dots , which render meaningless the notion of `` s - p '' and `` p - p '' splittings .
( such a localization crossover is absent for electrons . ) for this reason , we have compared the splittings for the holes in a shallow ( @xmath115 )
non - alloyed inas / gaas dot only , while comparing the splitting for electrons in a range of heights .
height dependence of the `` s - p '' ( @xmath133 ) and `` p - p '' ( @xmath129 ) energy splittings for electron and hole levels in in@xmath118ga@xmath111as / gaas .
the single - particle gap is also shown .
the `` s - p '' splitting for electrons is bigger than for holes ; in turn , the `` p - p '' splitting shows the opposite trend .
the height dependence of the `` p - p '' splitting for electrons is not monotonic , due to alloy fluctuations , while for holes this splitting remains nearly unchanged . for comparison , we show the `` s - p '' and `` p - p '' splittings for electron ( solid squares ) and hole ( open squares ) in non - alloyed inas / gaas dots.,width=321 ] figure [ fig_5 ] compares the wavefunctions of lumo ( lowest electron ) and homo ( highest hole ) states as a function of height in in@xmath118ga@xmath136as / gaas dots with base @xmath10 . to make the comparison ,
we plot isosurfaces that enclose @xmath137 of the total charge density , and show contour plots taken at @xmath138 above the dot s base .
in addition , the `` _ s_''-orbital character of the lumo and homo is indicated as well as the homo s heavy- and light - hole character .
further , for each height , the energy @xmath139 of lumo is shown relative to @xmath140 and the energy of homo @xmath141 relative to @xmath142 .
prominent results are the following .
_ lumo : _ the lateral spatial extent of the wavefunction depends weakly on height . in contrast , as height decreases , the wavefunction extends into the barrier along the vertical direction . the `` _
s_''-orbital character of lumo remains at nearly @xmath143 .
_ homo : _ wavefunctions are more sensitive to height , showing a spatial extension that gets reduced significantly both in the lateral and vertical direction when height changes from @xmath13 to @xmath12 .
this reduction leads to a strong localization at the center of the dot for @xmath144 .
the `` s''-orbital character of the homo remains at about 90% up to @xmath114 , for taller dots the character of the homo becomes `` s '' and `` p '' mixed .
this mixing reflects the reduction of the hole charge density near the center of the dot .
the heavy- ( hh ) and light - hole ( lh ) character also change with height in a similar manner as the `` s '' and `` p '' character .
namely , for the three smaller dots ( @xmath12 , @xmath145 , @xmath146 ) the homo is mostly of the hh type , but as the height increases the lh character increases .
we expect this behavior since the lh band - offset increases within the dot as height increases .
( see fig .
[ fig_3 ] . ) height dependence of lowest electron level ( lumo ) and highest hole level ( homo ) wavefunctions .
isosurfaces enclose 75% of the total charge density .
countours are taken at a plane @xmath147 nm above the base .
the energy relative to @xmath140 and @xmath142 is shown .
the `` s- '' and `` p-''orbital character ( see eq .
[ angular.char ] ) , and the heavy - hole ( hh ) and light - hole ( hh ) character are indicated .
for the tallest dot , the wavefunctions are entirely confined within the dot and the effective size of the dot spatial extent of the wavefunctions is nearly the same for lumo and homo . for shorter dots the effective size
gets significantly reduced for the homo while it remains nearly unchanged for the lumo.,width=321 ] _ excited hole states._figure [ fig_7 ] compares wavefunctions for excited hole states ; namely , second ( homo@xmath148 ) and third ( homo@xmath149 ) hole states . as in fig .
[ fig_5 ] , we show isosurfaces and contour plots , present the orbital and hh / lh character of these states , and the energies @xmath150 and @xmath151 for homo@xmath1521 and homo@xmath1522 , respectively .
the character of homo@xmath148 and homo@xmath149 are nearly the same at all heights .
in addition , these states have a dominant `` p''-orbital character for _ all _ heights , regardless the absence of `` p''-shell structure for tall dots ( @xmath153 and @xmath13 ) .
( see fig .
[ fig_4 ] ) as in the case of the homo state , the homo@xmath148 and homo@xmath149 states have increasing light - hole character with increasing height .
however , the percentage of light - hole character is almost twice that of homo .
for instance , at @xmath154 , the lh character of the homo@xmath149 is 17% while the homo lh character is 11% .
wavefunctions of the first 2 excited hole levels homo-1 and homo-2 for different dot s height . as in fig .
[ fig_6 ] , the isosurface encloses @xmath137 of the total charge density and the countours are taken at 1 nm above the base .
labels indicate the `` p '' orbital character , the heavy- and light - hole characters and the energy of the states relative to @xmath142.,width=321 ] for dots containing both ga and in ( in@xmath134ga@xmath135as / gaas ) with @xmath10 and heights in the range 20 - 75 , we have shown that the wavefunctions of homo ( @xmath141 ) and homo@xmath1521 ( @xmath150 ) , as well as other low - lying excited hole states , are localized inside the dot ( see figs .
[ fig_6 ] and [ fig_7 ] ) and that these levels have a sizeable `` s - p '' splitting @xmath155 ( see fig . [ fig_5 ] ) .
in contrast , in lens - shaped pure , non - alloyed inas / gaas quantum dots , localization of the low - lying hole states at the dot interface develops as the height of the dot increases .
in addition , homo and homo@xmath1521 become nearly degenerate [ e. g. , @xmath156 and @xmath157 at @xmath158 ] , as well as homo@xmath1522 and homo@xmath1523 .
( it should also be noted that for the flat dot the energies of homo and homo@xmath1521 are bigger than in the alloy dot of the same size . )
remarkably , homo and homo-1 are _ polarized _ along @xmath4 $ ] ( fig .
[ fig_8 ] ) while homo@xmath1522 and homo@xmath1523 ( not shown ) are polarized along @xmath5 $ ] .
figure [ fig_8 ] shows the development of this interfacial localization and the energy of homo and homo@xmath1521 relative to @xmath142 . as in figs .
[ fig_6 ] and [ fig_7 ] , isosurfaces enclose 75% of the homo and homo@xmath1521 total charge density .
homo and homo-1 wavefunctions in pure , non - alloyed inas / gaas lens - shaped quantum dot as a function of height .
the dots have the same base diamater @xmath159 .
the aspect ratio @xmath160 is shown , as well as the energy relative to @xmath142 .
as before , the isosurfaces enclose 75% of the charge density and the countours are taken at @xmath161 above the base . as height increases , interface localization takes plave , and the homo and homo@xmath1521 become nearly degenerate.,width=321 ] it should be noted that the pure , non - alloyed inas / gaas dots have the _ same _ base size and _ same _ height range as the alloy dots we have previously discussed .
hence , hole localization does not have its origin on geometrical aspects of the quantum dots .
instead , the localization of the hole wavefunctions is driven by the biaxial strain present in the nanostructure ( qd+gaas matrix ) . to explain this result , we plot in fig . [ fig_8prime](a ) the valence band offsets of heavy - hole ( hh ) and light - hole ( lh ) character ( inside the dot ) along the line @xmath15 ( see figs .
[ fig_1 ] and [ fig_3 ] ) and the energy of homo and homo@xmath1521 ( thick dashes ) .
in addition , we plot in fig . [ fig_8prime](b ) a 3-dimensional rendering of the higher energy hole band offset values at a plane normal to @xmath15 , cut slightly above the base of the dot . _ first _ , we see that as the dot heigh increases the pocket structure [ indicated with arrows in fig . [ fig_8prime](a ) ] that appears in the higher energy band offset dramatically widens the confining potential ( given by the band offset ) along @xmath162 and that the band offsets values decrease .
further , the band offset become assymetric with the values at the top of the dot _ smaller _ that at the _ base_. as a consequence , it becomes clear why the energy of the homo and homo-1 become bigger and we also expect the wavefunctions of these states to be localized near the base of the dot .
it is important to mention that in the calculation of the energy of the homo and homo@xmath1521 we do not utilize the band offset we present in fig .
[ fig_8prime](a ) . _
second _ , the band offsets shown in fig .
[ fig_8prime ] present a `` crown '' structure at the dot - matrix interface.@xcite this crown structure becomes more significant as height increases , as a consequence of the values of band offset along the @xmath5 $ ] and @xmath4 $ ] direction presenting a weak dependence on height .
_ finally _ , the localization of the low - lying hole wave functions that develops as the height of pure , non - alloyed inas / gaas lens - shaped quantum dots is a results of the peculiar characteristics of the higher energy valence - band offsets , which in turn are determined by the biaxial strain profile in the nanostructure .
finally , it should be noticed that electron states do not experience interface localization ; instead , they continue to form shells and the levels in these shells have a predominant , not - mixed orbital character ( `` s '' , `` p '' , @xmath3 ) .
the reason for the absence of electron localization is that electron states are sensitive to the ( hydrostatic ) isotropic strain rather than to the biaxial strain .
figure [ fig_9 ] shows the lowest bright transition ( gap ) of the exciton as a function of dot height , calculated at the single - particle ( sp ) level ( @xmath119 ) and by using the many - body , configuration interaction ( ci ) method.@xcite in the latter , we use a basis consisting of 12 electron and 20 hole confined levels @xmath163 for heights @xmath164-@xmath13 and @xmath165 for @xmath144 .
as expected , the transition energy decreases as height increases , due to confinement .
however , the scaling with height differs significantly from predictions of single - band , effective mass ( @xmath166).@xcite namely , by fitting our results for the height dependence of the gap to the function @xmath167 we find @xmath168 and @xmath169 .
the value of the gap for large heights correspond , respectively , to @xmath170 and @xmath171 in the single - particle and ci approach .
we expect the sp and ci scaling exponents to be different , as in the single - particle calculation the scaling is dictated by the scaling of the lumo and homo , whereas in the ci calculation the electron - hole matrix elements are also included and the magnitude of these matrix elements decreases with height .
the discrepancy between the scaling ( @xmath172 ) in single - band effective mass and both our sp and ci calculations can be attributed to non - trivial effects that are naturally accounted for within our atomistic approach such as non - parabolicity and multiband effects , and the position - dependent strain present in the nanostructure .
the values of @xmath173 at large heights is also expected to be different , and the difference is attributed to correlation effects that are present in the many - body , configuration approach .
thus , we expect a small difference ( @xmath174 ) between @xmath175 and @xmath176 . for comparison , we also present results for the lowest exciton transition [ squares , open ( sp ) and solid ( ci ) ] in a lens - shaped , non - alloyed inas / gaas quantum dot .
the values of the transition energies are smaller than in the alloy dots .
exciton gap as a function of height .
single - particle ( open circles ) and configuration interaction ( ci , solid circled ) results are shown .
the ci basis is @xmath163 for @xmath164-@xmath13 and @xmath165 for @xmath144 .
dashed lines represent fits to the function @xmath167 .
fitting parameters are indicated . for comparison , we present results [ squares ; open ( sp ) and solid ( ci ) ] for a pure , non - alloyed inas / gaas dot.,width=283 ]
by using a high - level atomistic approach , we have predicted spectroscopic characteristics of self - assembled in@xmath8ga@xmath1as / gaas quantum dots as a function of height and composition .
several prominent features emerged .
\(i ) the biaxial strain at the dot - gaas matrix interface increases with height , whereas , regardless of height , the strain is negligibly small along the @xmath48 $ ] , @xmath49 $ ] and crystallographically equivalent directions .
\(ii ) regardless of height and composition , the confined electron energy levels group in shells of nearly degenerate states .
the average energy splitting among these shells depends weakly on height ; however , this splitting is larger in pure , non - alloyed inas / gaas .
in contrast , the confined hole energy levels form shells only in flat dots and near the highest hole level ( homo ) .
\(iii ) in alloy dots , the _ electrons_
`` s - p '' splitting depends weakly on height , while the `` p - p '' splitting depends non - monotonically due to alloy fluctuations . in pure , non - alloyed inas / gaas dots , _ both _ these splittings depend weakly on height .
further , the `` s - p '' splitting is larger while the `` p - p '' has nearly the same magnitude . for _ holes
_ levels in alloy dots , the `` s - p '' splitting decreases with increasing height ( the splitting in tall dots being about 4 times smaller than in flat dots ) , whereas the `` p - p '' splitting remains nearly unchaged .
shallow pure dots have a `` s - p '' splitting of nearly the same magnitude , whereas the `` p - p '' splitting is about three times larger .
\(iv ) as height increases , the `` s '' and `` p '' character of the wavefunction of the homo becomes mixed , and so does the heavy- and light - hole character .
\(v ) in alloy dots , regardless of height , the wavefunction of low - lying ( near the homo ) hole states are localized inside the dot . remarkably , in pure , non - alloyed inas / gaas dots , as the dot height increases , these states become localized at the dot - matrix interface and nearly degenerate .
further , the localized states are polarized along @xmath4 $ ] ( homo and homo-1 ) and @xmath5 $ ] ( homo-2 and homo-3 ) .
this localization effect is driven by the peculiarities of the biaxial strain present in the nanostructure .
\(vi ) the lowest exciton transition energy ( gap ) decreases with height , but the scaling ( roughly @xmath177 ) differs significantly from the prediction of single - band effective mass ( @xmath178 ) . the study we presented here may be used to bridge spectroscopy results with theory without the need for severe theoretical approximations . this work has been supported by doe - sc - bes - dms under grant no .
de - ac36 - 99go10337 .
the lcbb method contains several convergence parameters : supercell size , number of bulk materials ( @xmath68 ) , number of bands ( @xmath70 ) , and number of @xmath71 points .
the choice of bulk materials and number of bands can be physically motivated and justified , whereas the choice of supercell size and number of * k * points is not as clear . for this reason
, we conducted a converge assesment on supercell size and number of * k * points .
as for the bulk materials , we choose ( i ) _ unstrained _ gaas and ( ii ) inas subject to strain values of @xmath179 , and @xmath180 .
these values are typical of strain inside a non - alloyed inas / gaas lens - shaped quantum dot . for the bands @xmath70 ,
when simulating the _ electron _ and _ hole _ energy levels we use respectively the lowest _ conduction _ band and the three highest valence bands of both bulk materials .
_ supercell size._we consider cubic supercells of size @xmath181 , where @xmath182 ( @xmath183 ) is measured in units of the lattice parameter of bulk gaas @xmath184 .
based on strain calculations ( see above ) , we selected a supercell size such that the strain profile within the quantum dot remained unchanged upon changing size .
notwithstanding , it should be noted that although the strain may be converged within the dot the strain values at the boundary of the supercell may not be zero due to the slow relaxation of the strain fields .
this residual strain introduces small changes ( @xmath185 ) in the energy levels .
for instance , figure [ fig_1 ] shows the effect of reducing the residual strain by changing the supercell size from @xmath186 to @xmath187 for a @xmath144 dot .
_ number of * k * points._the * k*-point meshes enclose all the @xmath71 vectors around the @xmath76 point that lie within an ellipsoid with main axis equal to @xmath188 , @xmath189 , and @xmath190 , respectively .
p , q , and k give the number of * k * points taken along each cartesian direction . in our convergence assesment @xmath191 , @xmath192 , and @xmath193 ranges from 8 - 18 . figure [ fig_a1 ] shows the electron energy levels for several dots as a function of * k*-point mesh .
we distinguish several features .
( a ) when compared with the energy of the levels , differences in energies are quicker to converge .
( b ) high - energy levels require larger * k*-point meshes to converge within a given threshold .
( c ) the taller the quantum dots the smaller the number of * k * points needed to converge .
( d ) the use of @xmath194 and @xmath195 * k*-point meshes results in similar energies for the lowest electron level .
the discrepancy between the predictions made with these two meshes increases for high - energy levels . to present details of the convergence with respect to * k*-point mesh , figure [ fig_a2 ]
shows a convergence assessment of homo and lumo s energy in flat ( @xmath144 ) and tall ( @xmath154 ) dots .
the energy of homo converges visibly quicker than that of lumo .
this behavior holds for all hole energy levels , and arises from the higher size ( three @xmath196 bands ) of the basis used in the simulation .
convergence of lumo ( @xmath139 ) and homo ( @xmath141 ) energies relative to their value simulated with the highest number of * k * points on a basis of the form @xmath194 . to simulate the homo energy we include the 3 @xmath196 bands .
the size of the supercell is indicated.,width=283 ] to summarize , the results we present in this work for each quantum dot derive from simulations with * k*-point meshes that provide with energy levels converged within @xmath197 .
table [ table_1 ] shows the supercell sizes and * k*-point meshes we have used to simulate the electronic structure of each of the dots we have studied .
it should be noted that _ energy differences _ are converged to much less than this lower bound .
.[table_1]summary of supercell sizes @xmath181 ( lengths in units of @xmath184 ) and * k*-point meshes @xmath198 used in the simulations .
dot s base @xmath10 .
[ fig_1 ] for dot s geometry . ) [ cols="^,^,^",options="header " , ] s. fafard , r. leon , d. leonard , j. l. merz and p. m. petroff , phys .
b * 52 * , 5752 ( 1995 ) ; s. fafard , d. leonard , j. l. merz , and p. m. petroff , appl .
lett . * 65 * , 1388 ( 1994 ) ; j .- y .
marzin , j .-
grard , a. izrel , d. barrier , g. bastard , phys .
* 73 * , 716 ( 1994 ) ; m. grundmann , j. christen , n. n. ledenstov , j. bhrer , d. bimberg , s. s. ruvimov , p. werner , u. richter , u. gsele , j. heydenreich , v. m. ustinov , a. yu .
egorov , a. e. zhukov , p. s. kopev , and zh .
alferov , phys .
* 74 * , 4043 ( 1994 ) .
f. guffarth , r. heitz , a. schliwa , o. stier , m. geller , c. m. a. kapteyn , r. sellin , and d. bimberg , phys .
b * 67 * , 235304 ( 2003 ) ; j. j. finley , a. d. ashmore , a. lematre , d. j. mowbray , m. s. skolnick , i. e. itskevich , p. a. maksym , m. hopkinson , and t. f. krauss , phys .
b * 63 * , 073307 ( 2001 ) ; r. j. warburton , c. s. drr , k. karrai , j. p. kotthaus , g. medeiros - ribeiro , and p. m. petroff , phys . rev . lett . *
79 * , 5282 ( 1997 ) .
m. bayer , o. stern , p. hawrylak , s. fafard , and a. forchel , nature * 405 * , 923 ( 2000 ) ; e. dekel , d. gershoni , e. ehrenfreund , j. m. garcia , and p. petroff , phys .
b * 61 * , 11009 ( 2000 ) ; f. findeis , a. zrenner , g. bhm , and g. abstreiter , solid state comm .
* 114 * , 227 ( 2000 ) ; l. landin , m. s. miller , m .- e . pistol , c. e. pryor , and l. samuelson , science * 280 * , 262 ( 1998 ) ; m. bayer , t. gutbrod , a. forchel , v. d. kulakovskii , a. gorbunov , m. michel , r. steffen , and k. h. wang , phys . rev .
b * 58 * , 4740 ( 1998 ) .
b. urbaszek , r. j. warburton , k. karrai , b. d. gerardot , p. m. petroff , and j. m. garcia , phys .
lett . * 90 * , 247403 ( 2003 ) ; m. bayer , g. ortner , o. stern , a. kuther , a. a. gorbunov , a. forchel , p. hawrylak , s. fafard , k. hinzer , t. l. reinecke , s. n. walck , j. p. reithmaier , f. klopf , and f. schfer , phys . rev .
b * 65 * , 195315 ( 2002 ) ; j. j. finley , d. j. mowbray , m. s. skolnick , a. s. ashmore , c. baker , a. f. g. monte , and m. hopkinson , phys . rev .
b * 66 * , 153316 ( 2002 ) . c. pryor , phys . rev .
b * 57 * , 7190 ( 1998 ) ; t. b. bahder , phys . rev .
b * 45 * , 1629 ( 1992 ) ; t. b. bahder , phys . rev .
b * 41 * , 11992 ( 1990 ) [ see also t. b. bahder , phys . rev .
b * 46 * , 9913 ( 1992 ) ] and references therein .
s. lee , o. l. lazarenkova , p. von allmen , f. oyafuso , and g. klimeck , phys . rev .
b * 70 * , 125307 ( 2004 ) ; r. santoprete , b. koiler , r. b. capaz , p. kratzer , q. k. k. liu , and m. scheffler , phys . rev .
b * 68 * , 235311 ( 2003 ) ; g. w. bryant and w. jasklski , phys . rev .
b * 67 * , 205320 ( 2003 ) ; s. lee , l. jnsson , j. w. wilkins , g. w. bryant , g. klimeck , phys . rev .
b * 63 * , 195318 ( 2001 ) ; kevin leung and k. b. whaley , phys .
b * 56 * , 7455 ( 1997 ) .
d. leonard , m. krishnamurthy , c. m. reaves , s. p. denbaars , and p. m. petroff , appl .
. lett . * 63 * , 3203 ( 1993 ) ; d. leonard , k. pond , and p. m. petroff , phys .
b * 50 * , 11687 ( 1994 ) ; s. raymond , s. fafard , p. j. poole , a. wojs , p. hawrylak , and s. charbonneau , d. leonard , r. leon , p. m. petroff , and j. l. merz , phys .
b * 54 * , 11548 ( 1996 ) ; z. r. wasilewski , s. fafard , and j. p. mccaffrey , j. cryst .
growth * 201 - 202 * , 1131 ( 1999 ) ; a. g. cullis , d. j. norris , t. walther , m. a. migliorato , and m. hopkinson , phys .
b * 66 * , 081305 ( 2002 ) ; s. anders , c. s. kim , b. klein , m. w. weller , and r. p. mirin , phys .
b * 66 * , 125309 ( 2002 ) . these `` crown '' structure has been predicted by he and coworkers in ref .
@xcite and has been argued that it is this structure that is responsible for similar localization of hole wavefunctions found in large , non - alloyed inas / gaas _ spherical _ quantum dots.@xcite for single - band effective mass calculations of the single - particle electronic levels in lens - shaped semiconductor quantum dots with aspect ratio @xmath199 and @xmath200 see j. even and s. loualiche , j. phys .
a : math . gen . * 36 * , 11677 ( 2003 ) and a. h. rodriguez , c. trallero - giner , s. e. ulloa , and j. marn - antua , phys .
b * 63 * , 125319 ( 2001 ) , respectively . | while electronic and spectroscopic properties of self - assembled in@xmath0ga@xmath1as / gaas dots depend on their shape , height and alloy compositions , these characteristics are often not known accurately from experiment .
this creates a difficulty in comparing measured electronic and spectroscopic properties with calculated ones . since simplified theoretical models ( effective mass , k@xmath2p , parabolic models )
do not fully convey the effects of shape , size and composition on the electronic and spectroscopic properties , we offer to bridge the gap by providing accurately calculated results as a function of the dot height and composition .
prominent features of our results are the following .
( i ) regardless of height and composition , the confined electron energy levels form shells of nearly degenerate states with a predominant orbital `` s '' , `` p '' , @xmath3 character . in contrast , the confined hole energy levels form shells only in flat dots and near the highest hole level ( homo ) .
( ii ) in alloy dots , the _ electrons_
`` s - p '' splitting depends weakly on height , while the `` p - p '' splitting depends non - monotonically due to alloy fluctuations . in pure , non - alloyed inas / gaas dots , _ both _ these splittings depend weakly on height .
further , the `` s - p '' splitting is larger while the `` p - p '' has nearly the same magnitude . for _ holes
_ levels in alloy dots , the `` s - p '' splitting decreases with increasing height ( the splitting in tall dots being about 4 times smaller than in flat dots ) , whereas the `` p - p '' splitting remains nearly unchaged .
shallow , pure non - alloyed dots have a `` s - p '' splitting of nearly the same magnitude , whereas the `` p - p '' splitting is about three times larger .
( iii ) as height increases , the `` s '' and `` p '' character of the wavefunction of the homo becomes mixed , and so does the heavy- and light - hole character .
( iv ) in alloy dots , regardless of height , the wavefunction of low - lying hole states are localized inside the dot .
remarkably , in non - alloyed inas / gaas dots these states become localized at the interface as height increases .
the localized states are nearly degenerate and polarized along @xmath4 $ ] and @xmath5 $ ] .
this localization is driven by the peculiarities of the biaxial strain present in the nanostructure . |
it is becoming increasingly apparent that biological evolution of organisms life history traits can occur on a similar time scale to their population dynamics @xcite , and that such rapid evolution can shape the ecological dynamics of interacting species @xcite .
in particular , predator - prey and host - parasite interactions can be dramatically affected by rapid genetic variation within the populations @xcite , and we need to take this fact into account in food - web and epidemiological models .
for instance , it was demonstrated in @xcite that rapid genetic variation can enlarge the period and modify the phase relations in predator - prey cycles .
recently , using a parsimonious eco - evolutionary predator - prey model , morozov _ et al . _
have shown that the interplay between a fast genetic variation within the prey and the selectivity of consumption of the prey by its predator can suppress large amplitude oscillations in the system @xcite .
interestingly , the reported stabilization could occur even in the case when the carrying capacity of the prey is infinitely large .
in contrast , the same system with non - changing life history traits of species would be globally unstable , see e.g. @xcite .
thus , a rapid evolution could potentially provide an extra mechanism of persistence for trophic interaction in a highly eutrophic environment ( known also as the top - down control ) @xcite .
this mechanism would answer the long standing open question in theoretical ecology about possible solutions of the paradox of enrichment : what factors can stabilize predator - prey interactions in the case the supply of resource for prey is large ?
@xcite .
the previous results on stabilization in the eco - evolutionary predator - prey model with prey structuring in a highly eutrophic ecosystems were obtained by directly simulating the underlying model equations for particular parameterisations of the model functions @xcite .
obviously , this can not be considered as a rigorous proof of stability .
furthermore , considering other functional parameterizations of the inheritance kernel as well as the selectivity of predation can potentially affect the results obtained only using simulations .
thus , the central question is whether or not the main conclusions in @xcite on the importance of within population structuring on persistence and stability will be generic .
the main objective of this paper is to explore in detail the mathematical properties of the eco - evolutionary predator - prey model @xcite .
our analytical investigation is centred around two main issues : the existence of non - trivial stationary state(s ) of the model ( assuring the coexistence of the prey and predator ) , and the stability of the stationary state(s ) for a general family of functions ( kernels ) describing genetic inheritance within the prey .
we also establish global existence and positivity of solutions of the model .
mathematically we consider the following system of integro - differential equations ( which is a generalization of that in @xcite ) : @xmath0 where @xmath1 describes the distribution of prey individuals across the cohorts according to a certain life history trait @xmath2 , which can vary between @xmath3 and @xmath4 .
hence the integration of @xmath1 over @xmath5 $ ] gives the total biomass of prey in the population , while @xmath6 is the total biomass of the predator at time @xmath7 . here
we consider the structuring life history trait @xmath2 to be an abstract parameter : in practice this can be the growth rate , body length ( at the adult stage ) , individual mobility @xcite or it can be the animals personality @xcite .
the life trait is genetically inherited , i.e. it does not change throughout the lifetime . for the sake of simplicity we assume that the predator population is genetically uniform ( i.e. there is no structuring of the predator population ) .
the integral term in equation ( [ eq : zoo ] ) stands for the growth rate of prey due to the reproduction of all cohorts .
the kernel @xmath8 describes genetic inheritance and mutation .
each cohort can produce offspring with the life trait within @xmath5 $ ] .
the contribution to the growth rate of cohort @xmath2 due to reproduction of the group of cohorts with life traits in the interval @xmath9 $ ] will be given by the product of the population size of this group @xmath10 and the kernel @xmath8,with @xmath11 being sufficiently small .
thus the kernel function includes both the reproduction rate of group of cohorts centred at @xmath12 and the probability of producing offspring with trait @xmath2 .
the total contribution of all cohorts to the production of the cohort @xmath2 is given by integration of @xmath13 over the the entire interval @xmath5 $ ] .
the mortality of prey in the model is due to predation only ( i.e. we neglect natural mortality ) , which is achieved via a standard holling type ii parametrization , see e.g. @xcite , where @xmath14 characterises the saturation of predation at high densities of the prey .
the vulnerability of the prey to predation is cohort - dependent and is described by the function @xmath15 , i.e. there is prey selectivity by the predator according to the life trait @xmath2 .
@xmath16 is the food conversion coefficient describing the efficiency of transformation of the consumed prey biomass .
the predator is subject to a natural mortality @xmath17 which is assumed to be constant .
model ( [ eq : zoo])-([eq : fish ] ) can be considered as a standard rosenzweig - macarthur model ( see @xcite ) in which genetic structuring and rapid evolution has been incorporated .
our model is similar to that of @xcite , where evolution of cohorts of predator and prey in a chemostat was considered . however , unlike the model in @xcite , here we allow mutations of the clones .
in particular , the kernel @xmath8 can be constant which can model the scenario of perfect genetic mixing within the population .
it is of importance to mention that we intentionally consider the carrying capacity of prey population to be infinitely large , i.e. we model predator - prey interactions in a highly eutrophic environment .
it is well known that the same model without population structuring is globally unstable and the species persistence is impossible @xcite , thus the main question explored using model ( [ eq : zoo])-([eq : fish ] ) is whether or not population structuring and rapid evolution ( as well as selectivity of the predator ) can eventually stabilize this otherwise unstable system . note that by rapid evolution we understand here variation in the cohort distribution @xmath1 which occurs on the same time scale as the ecological dynamics ( i.e. variation of the total biomass of prey and predator ) .
_ et al . _
@xcite , we explore the model for arbitrary parametrisations of the vulnerability to predation @xmath18 and the inheritance kernel @xmath8 .
we only make the following natural assumptions regarding the model parameters and @xmath19 and @xmath20 @xmath21 global coexistence and positivity of solutions of model ( [ eq : zoo])-([eq : fish ] ) , based on the above assumptions for the parameters and integral kernels can be readily established for example using methods developed in @xcite .
it is worthwhile to point out that unlike in the case of standard " physiologically structured population models , see for example the classical size - structured models in @xcite ( and the references therein ) , in our model individuals are structured with respect to a genetically inherited property , which implies that there is no transport term in equation .
this allows us to work in the framework of bounded operators . on the other hand ,
recruitment is represented by an integral operator , which is of infinite rank , in general .
eigenvalue problems involving integral operators are often difficult to analyse .
see for example the recent papers @xcite , where structured population models with distributed recruitment processes ( modelled by an integral operator ) were analysed .
we also note that even though the operators arising from equations - are bounded , positivity of solutions can not be immediately established for example from the variation of constants formula .
this is due to the negative feedback between predator and prey populations . in the next two sections we show the existence of the nontrivial stationary state ( coexistence state ) of the model for a particular class of kernels , which is actually a separable kernel of a finite rank .
then we derive the characteristic equations to obtain the stability condition of the nontrivial stationary state . using the obtained generic condition we consider the stability property in a few particular biologically relevant cases , for instance , in the case where the kernel is constant .
finally , we address the important question of the necessity of predation selectivity for the stabilization of model ( [ eq : zoo])-([eq : fish ] ) , i.e. whether or not the stability of the coexistence state is possible in the case @xmath22 .
we summarize our results in the discussion section .
note that model ( [ eq : zoo])-([eq : fish ] ) always admits the trivial stationary state @xmath23 . throughout the paper we assume that the kernel @xmath20 is strictly positive , but for the sake of completeness we note
that if the kernel @xmath20 is not strictly positive then our model may admit stationary states of the form @xmath24 , where @xmath25 satisfies the integral equation @xmath26 for example if @xmath20 vanishes for every @xmath12 on some interval @xmath27 $ ] then any function @xmath25 which is supported on the interval @xmath27 $ ] will be , together with @xmath28 , a non - trivial stationary state . in the next two sections we discuss the existence of strictly positive ( coexistence or `
non - trivial ' ) stationary states of model ( [ eq : zoo])-([eq : fish ] ) .
we start with the relatively simple case of a separable kernel @xmath20 . in this case
the integral operator describing the recruitment process is of rank one ( at most ) .
then , we will discuss the case of a finite rank approximation of the kernel @xmath20 , and finally , for a general kernel @xmath20 , we will reformulate the steady state problem as an eigenvalue problem for a bounded operator .
assume that @xmath29 , for some functions @xmath30 , i.e. that the kernel is separable .
we look for a strictly positive stationary solution @xmath31 .
the stationary state equations read : @xmath32 we introduce the following notation : @xmath33 assuming that @xmath34 we obtain from equations ( [ sseq1 ] ) the following set of scalar equations for the variables @xmath35 .
@xmath36 one can easily see that a unique positive solution @xmath31 of ( [ sseq1])-([sseq2 ] ) exists if and only if the 4-dimensional scalar system ( [ scalar1])-([scalar4 ] ) above has a strictly positive solution .
we introduce the notation @xmath37 hence , we obtain from ( [ scalar2])-([scalar4 ] ) a unique positive solution @xmath38 , as long as @xmath39 .
substituting this into equation ( [ scalar3 ] ) yields a unique positive solution @xmath40 . from ( [ scalar1])-([scalar4 ] )
we obtain @xmath41 , and finally from equation ( [ scalar4 ] ) we obtain @xmath42 . hence the model admits a unique positive stationary state if @xmath43 the condition above is satisfied , for instance , in the case when the saturation @xmath14 is not very large and vulnerability to predation @xmath18 is large enough .
the approach above may be extended to the more general case of the kernel : @xmath44 i.e. for a finite rank approximation of the kernel @xmath20 with non - negative functions @xmath45 . with @xmath46 and @xmath47 as defined in ( [ z1z2 ] ) , and introducing the new variables : @xmath48 the stationary state equations of model ( [ eq : zoo])-([eq : fish ] ) read : @xmath49 equations ( [ sseq3 ] ) lead to the following @xmath50 dimensional scalar system : @xmath51 condition ( [ sscond1 ] ) for a separable kernel suggests that for small enough values of the parameter @xmath14 one should be able to establish existence ( but not necessarily uniqueness ) of a positive stationary state . in particular , for @xmath52 the first and third equation of ( [ sseq1sep ] ) together yield @xmath53 with this , the @xmath54-dimensional nonlinear scalar system [ sseq2sep ] can be cast in the form : @xmath55 where @xmath56 once a non - negative solution of system ( [ sseq3sep ] ) is found it can be substituted into equation ( [ sseqsepf ] ) to determine @xmath57 , which using the second equation of ( [ sseq1sep ] ) determines a unique positive @xmath46 . to establish existence of a positive solution of system ( [ sseq3sep ] ) we utilise an idea which was employed for infinite dimensional problems recently , for example , in @xcite .
the key idea is to recast the non - linear problem ( [ sseq3sep ] ) as an eigenvalue problem for a parameterised family of matrices .
that is , we rewrite system ( [ sseq3sep ] ) as @xmath58 where @xmath59 note that for any @xmath60 the matrix @xmath61 is non - negative .
it follows from perron - frobenius theory that the spectral radius @xmath62 is an eigenvalue with a corresponding non - negative eigenvector .
also note that the function @xmath63 is continuous for @xmath64 .
it follows for example from gershgorin s circle theorem that as @xmath65 we have @xmath66 . hence if there exists a value @xmath67 such that @xmath68 , then there exists a @xmath69 such that @xmath70 and therefore @xmath71 is an eigenvalue with a corresponding non - negative eigenvector @xmath72 .
we then normalize this eigenvector such that it satisfies @xmath73 note that the existence of a @xmath74 such that @xmath68 holds depends on the @xmath75 values , i.e. on the approximation of the kernel @xmath8 .
in particular if the kernel @xmath20 is such that for some @xmath2 value it is concentrated on the diagonal point @xmath76 then again gershgorin s circle theorem implies the existence of a @xmath74 such that @xmath68 . also note that if the kernel @xmath20 is such ( typically strictly positive ) that it can be approximated with @xmath77 such that @xmath78 for @xmath79 ,
then the matrix @xmath61 above is positive , and the perron - frobenius theorem guarantees the existence of a strictly positive eigenvector @xmath80 .
also if the matrix @xmath61 is positive , then it is clear that for any @xmath81 we have @xmath82 , hence it follows from the intermediate value theorem that there exists a @xmath69 such that @xmath70 .
we summarize our findings in the following lemma .
assume that @xmath83 , such that @xmath84 for @xmath85
. then model ( [ eq : zoo])-([eq : fish ] ) admits a strictly positive stationary state for @xmath52 .
next we rewrite system - as follows : @xmath86 where @xmath87 equations ( [ sseq4sep])-([sseq6sep ] ) can be recast in the more economic form : @xmath88 where @xmath89 is well - defined ( via the left hand - sides of equations ( [ sseq4sep])-([sseq6sep ] ) ) on @xmath90 and continuously differentiable .
we would like to apply the implicit function theorem for @xmath89 to show that if the equation @xmath91 has a strictly positive solution then it also has a strictly positive solution for some small positive values of @xmath14 .
to this end we compute the jacobian of @xmath89 at @xmath92 : @xmath93 hence if the determinant of the jacobian matrix evaluated at the a strictly positive stationary state is not zero then by the implicit function theorem a strictly positive stationary state also exists for small enough values of @xmath14 .
note that the value of the determinant of @xmath94 depends on the particular finite rank approximation of the kernel @xmath20 .
we briefly discuss here how the steady state problem can be formulated in case of a general ( i.e. , non - separable ) kernel @xmath20 .
this case is challenging from the mathematical point of view since the integral operator describing the recruitment process is of infinite rank , in general .
for a positive stationary state @xmath31 we define @xmath95,\quad z^*=\int_{x_1}^{x_2}z^*(x)\,dx.\ ] ] with this notation the ( positive ) steady state problem can be formulated at least for sufficiently small values of @xmath14 ( e.g. for @xmath52 ) as follows @xmath96 , \label{eqop1 } \\ f^ * & = \frac{k}{\mu}\int_{x_1}^{x_2}\kappa^*(x)\,dx,\label{eqop2 } \\ z^ * & = \frac{\int_{x_1}^{x_2}\frac{\kappa^*(x)}{a(x)}\,dx}{f^*-\beta\int_{x_1}^{x_2}\frac{\kappa^*(x)}{a(x)}\,dx}. \label{eqop3}\end{aligned}\ ] ] problem - can be considered as an eigenvalue problem for a bounded operator @xmath97 , where @xmath98 is defined via the right hand side of equations - , and @xmath99 is an appropriately choosen banach space , for example @xmath100 .
more precisely , if @xmath98 has eigenvalue @xmath71 with a corresponding strictly positive eigenvector @xmath101 then the system admits a positive steady state with @xmath102.\ ] ] note that for @xmath52 the operator @xmath98 is positive , and the eigenvalue problem for @xmath98 can be analysed using similar arguments as in section 2.2 above for infinite dimensional problems see e.g. @xcite .
in particular , for the special case of @xmath52 and @xmath103 we have the following result . [ ssgeneral ] in the case of @xmath52 and @xmath103 model
- admits a unique positive stationary state .
the proof of theorem [ ssgeneral ] is included in appendix a. it is worth pointing out that , unlike in the previous sections , we are able to prove uniqueness of the coexistence steady state . to prove the existence of a positive steady state in the most general case , i.e. without the assumptions @xmath52 and @xmath103
, we could apply results from the forthcoming paper @xcite , but due to the technical difficulties involved this is outside the scope of the present paper .
in this section we shall address the local stability of the coexistence stationary state of the model .
to this end we linearise the model around the steady state , and analyse the arising eigenvalue problems .
our model is semi - linear , and so we may invoke theorem 5.1.1 and theorem 5.1.3 from @xcite to justify that the stability results obtained in the section are valid .
we linearise equations ( [ eq : zoo])-([eq : fish ] ) in the vicinity of the stationary state @xmath104 and consider perturbations @xmath105 and @xmath106 of this state .
we obtain @xmath107 where @xmath108 and @xmath109 .
we note that the linear problem above is governed by an analytic semigroup , hence the spectrum may contain only eigenvalues of finite multiplicity . to determine the possible eigenvalues @xmath110 we look for solutions of the linearised equations in the standard form : @xmath111 , and @xmath112 .
this ansatz leads to the following eigenvalue problem ( with the assumption of @xmath113 ) : @xmath114 as we can see the eigenvalue problem ( [ eq : stability_1])-([eq : stability_2 ] ) is rather complicated , in general , since it contains integral equations .
nevertheless in the following subsections we discuss some interesting special cases when the eigenvalue problem becomes tractable , and we can deduce analytical stability or instability results .
the eigenvalue problem for the case of a constant kernel @xmath115 reads : @xmath116 we integrate equation ( [ evalue1_1 ] ) , then multiply it by @xmath16 and add it to equation ( [ evalue1_2 ] ) to obtain : @xmath117 since any non - trivial eigenvector is determined up to a constant multiplier , we may assume that the eigenvector @xmath118 is such that @xmath119 .
note that since we only want show the existence of a positive eigenvalue with a corresponding non - trivial eigenvector ( hence we are not characterising the whole point spectrum ) we do not need to worry about possible eigenvalues with eigenvectors for which @xmath120 . with @xmath119 from ( [ evalue3_1 ] ) we obtain ( for @xmath121 ) : @xmath122 using ( [ gvalue ] ) we obtain from ( [ evalue1_1 ] ) @xmath123 @xmath124 the equation for the eigenvalues can be derived by integrating ( [ main ] ) over @xmath5 $ ] which should give unity ( according to our choice of @xmath125 ) : @xmath126 we can simplify the expression above by taking into account the explicit formulae for the stationary densities @xmath127 in the case of a constant kernel @xmath20 . without losing generality
we can assume that @xmath128 .
in this case , @xmath129 has the meaning of the growth rate of the whole prey population ( for a constant kernel the growth rate does not depend on the particular cohort ) . using ( [ scalar1])-([scalar4 ] )
, we have the stationary densities of species : @xmath130 we substitute the stationary values ( [ biomassz])-([densf ] ) into , and after some simplification we obtain : @xmath131 where @xmath132 .
we summarize our finding in the following theorem .
[ k - const ] the positive stationary state @xmath127 of ( [ eq : zoo])-([eq : fish ] ) with a constant kernel @xmath133 is asymptotically stable if all of the solutions of ( [ main2 ] ) have negative real parts , and it is unstable if there exists at least one solution of with positive real part . by using a particular parametrisation of @xmath18
the characteristic equation becomes analytically tractable and one can deduce sufficient conditions for the stability of the stationary state .
a particularly important case , however , is when the saturation in the predation rate @xmath14 is small ( @xmath134 ) . in this case
we can set @xmath52 , and the characteristic equation ( [ main2 ] ) takes the following simple form : @xmath135 utilising this equation , we can analytically explore stability properties of the stationary state for an arbitrary function @xmath18 in the case when the trait interval is small , i.e. when @xmath136 .
we can use a first order approximation of @xmath18 given by @xmath137 .
we substitute this approximation of @xmath18 into ( [ main3 ] ) , and after integration we obtain : @xmath138 \frac{h}{a_1 r i_1}&=&1.\end{aligned}\ ] ] we compute the taylor expansion of the left hand side of ( [ linear_a ] ) up to the second order of @xmath139 and multiply it by @xmath140 to obtain a fourth order polynomial function in terms of @xmath110 .
we can do it since we look for only @xmath110 with @xmath141 @xmath142 where the polynomial coefficients @xmath143 are given by : @xmath144 to determine stability of the polynomial ( [ linear_a2 ] ) we can utilise the routh - hurwitz stability criterion , see e.g. in @xcite : @xmath145 verification of the above conditions show that we have always @xmath146 and @xmath147 thus the system is locally stable for small @xmath139 .
interestingly , the actual sign of @xmath148 , i.e. the gradient of the dependence of vulnerability to predation on the life history trait @xmath2 , does not affect the stability result : the system becomes stabilised for any @xmath149 .
note also that keeping only the linear part in terms of @xmath139 in the expansion of ( [ linear_a ] ) would be misleading since some roots @xmath110 of the polynomial will be purely complex resulting in a degenerated case : small variations of this equation ( e.g. considering @xmath150 ) will perturb those solutions resulting in the appearance of the non - zero real parts . in the case when the interval @xmath5 $ ] is not small and we can not always assume that @xmath151 ( but we can still neglect the saturation in the predation rate , i.e. set @xmath52 ) we need to solve the full equation ( [ main3 ] ) , which in general is a difficult task since @xmath110 has both real and imaginary parts . however , we know that for a sufficiently small @xmath139 the stationary state is always stable and due to the continuous dependence of the model dynamics on the parameters ( by invoking theorem 3.4.1 and theorem 3.4.4 from @xcite ) , the stability loss ( if any ) of the stationary state for a larger value of @xmath139 should take place via a hopf bifurcation with @xmath152 . to find the bifurcation point we substitute @xmath153 into ( [ main3 ] ) and separate the real and imaginary part of this equation . after some rearrangement we obtain the following system of equations : @xmath154 ^ 2}dx & = & \frac{h \mu}{r+\mu},\\ \label{re_1 } \int_{x_1}^{x_2}\frac{a(x)\frac{r}{h}i_1}{\omega^2+\left[a(x)\frac{r}{h}i_1\right]^2}dx & = & \frac{h}{r+\mu},\end{aligned}\ ] ] where the integral @xmath155 is defined earlier .
one can further explore the possibility of solving equations ( [ im_1])-([re_1 ] ) for a particular function @xmath18 . in our study
we have considered the generic linear and parabolic functions given by @xmath137 and @xmath156 .
note that for those functions it is easy to calculate the above integrals analytically . in each case
we investigated the possibility of solving the system ( [ im_1])-([re_1 ] ) , i.e. to find @xmath157 satisfying both equations .
such @xmath157 will correspond to a hopf bifurcation point .
our results show that for the given functions there is no solution for any combination of the other model parameters .
we do not show here the results for the sake of brevity .
thus , we can conclude that for linear and parabolic parameterisations of @xmath18 the stationary state @xmath127 is always locally asymptotically stable for any trait interval @xmath5 $ ] , whenever it exists .
it remains to address whether or not this stability result can be extended to the case of an arbitrary function @xmath18 and @xmath52 . on the other hand , in the case of a sufficiently large saturation in predation ( i.e. when we can not neglect @xmath14 ) the stationary state @xmath127 is always unstable , at least for small values of @xmath139 .
this can be shown directly by expanding equation ( [ main2 ] ) into taylor series and keeping only the linear part with respect to @xmath139 .
@xmath158 the eigenvalues of this equation have positive real parts since the coefficients of the polynomial have different signs ( @xmath159 ) , thus the stationary state is unstable .
since for small @xmath14 the stationary state is stable ( at least for small @xmath139 , see above ) , but it becomes unstable for larger @xmath14 , there should be a critical value of @xmath160 , where the stability switch occurs .
the function @xmath161 defines a hopf bifurcation curve , which can be obtained numerically from ( [ main2 ] ) by substituting @xmath153 .
the equation for the bifurcation curve is defined by the following system , which we obtained by equating to zero both the imaginary and the real parts of equation ( [ main2 ] ) : @xmath162 ^ 2}dx \hspace{45 mm } \nonumber \\ = & \frac{\mu h^3 k^2 ( r+\mu ) } { h^2k^2 \omega^2(r+\mu)^2+r\mu \beta i_1 \left[\mu r \beta i_1 ( \mu^2+\omega^2)+2hk\omega^2 ( r+\mu ) \right ] } , \label{im_2}\\ & \int_{x_1}^{x_2}\frac{a(x)\frac{r}{h}i_1 dx}{\omega^2+\left[a(x)\frac{r}{h}i_1\right]^2}dx \hspace{45 mm } \nonumber \\ = &
\frac{\left [ hk\omega^2(r+\mu)+\mu r \beta i_1 ( \mu^2 + \omega^2 ) \right]}{h^2k^2 \omega^2(r+\mu)^2+r\mu \beta i_1 \left[\mu r \beta i_1 ( \mu^2+\omega^2)+2hk\omega^2 ( r+\mu ) \right ] } , \label{re_2}\end{aligned}\ ] ] with @xmath157 being a positive real value . by choosing a particular function @xmath18 and solving ( [ im_2])-([re_2 ] )
one can construct the hopf bifurcation curve .
note that formally one need to check the standard requirement of , see e.g. @xcite , i.e. that @xmath163 holds , with @xmath164 being a bifurcation parameter ( we can verify this condition numerically ) .
we emphasize that in ( [ im_2])-([re_2 ] ) we do not use the assumption that @xmath151 . as an illustrative example
, we constructed a family of hopf bifurcation curves numerically for a linear vulnerability function @xmath137 in the @xmath165 plane , which is shown in figure 1 .
note that for the given @xmath18 , the integrals in ( [ im_2])-([re_2 ] ) can be easily calculated analytically and the system becomes a system of transcendental equations .
for each curve in the figure , the region corresponding to the stable stationary state is located on the left - hand side of the curve .
the different curves correspond to different lengths of the interval @xmath166 $ ] .
one can see that stability loss due to an increase in @xmath14 can be compensated for either by increasing the range of @xmath2 , or by increasing the absolute value of the gradient of @xmath18 .
this is in agreement with earlier results in @xcite , which were obtained via direct simulation of the model equations . )
-([eq : fish ] ) with a constant kernel @xmath133 and the linear vulnerability @xmath137 constructed using equations ( [ im_2])-([re_2 ] ) .
different curves correspond to different values of @xmath4 .
the other model parameters are : @xmath167 . for a given @xmath4 the stability region
is located on the left - hand side of each curve .
the dashed horizontal lines show the lower limits of @xmath148 beyond which @xmath18 becomes negative.,width=453 ] note that equations ( [ im_2])-([re_2 ] ) only account for the fact that the stationary state loses it stability , however , without providing information about which type of hopf
bifurcation actually occurs : a subcritical or a supercritical one .
the knowledge about the type of hopf bifurcation can be of importance since after a supercritical bifurcation the trajectories remain bounded close to the equilibrium whereas they can become unbounded ( no population persistence , see @xcite ) in case of subcritical hopf bifurcation .
numerical simulations carried out with particular parametrisations of @xmath18 show that we can have both types of bifurcations ( depending on the model parameters ) .
however , revealing a particular type of hopf bifurcation would require a rather lengthy expressions and should be a matter of a separate study .
the main technique of stability analysis is similar to the one implemented in the case of a constant kernel .
equations ( [ eq : stability_1])-([eq : stability_2 ] ) now take the following form : @xmath169 we integrate equation ( [ evalue1 ] ) , multiply it by @xmath16 and add it to ( [ evalue2 ] ) to obtain : @xmath170 where @xmath171 .
we can again assume that @xmath119 ( as before ) and we do not need to worry about possible eigenvalues with eigenvectors for which @xmath120 .
we obtain for @xmath121 : @xmath172 using ( [ gvalue_1 ] ) we obtain from ( [ evalue1 ] ) @xmath123 @xmath173 we multiply equation ( [ main_4 ] ) by @xmath174 and integrate it over @xmath5 $ ] , this should give us the value of @xmath175 : @xmath176 we can simplify equation ( [ main_4 ] ) by using the explicit expressions for the stationary densities @xmath127 given by equations ( [ scalar1])-([scalar4 ] ) .
the stationary values of @xmath177 and @xmath25 will be the same as in ( [ biomassz ] ) and ( [ densityz ] ) .
for @xmath57 we have : @xmath178 we substitute these stationary values into ( [ main_5 ] ) , and after some simplification we obtain : @xmath179 where @xmath180 .
then we obtain the solution @xmath175 as follows : @xmath181 where @xmath182 .
finally , we substitute @xmath175 into the equation for @xmath183 and integrate over @xmath166 $ ] , this should give us 1 since we set @xmath119 .
after some simplification , we have : @xmath184 we summarize our findings in the following theorem .
[ k(x ) ] the positive stationary state @xmath127 of ( [ eq : zoo])-([eq : fish ] ) with a kernel @xmath185 is stable if the eigenvalues of characteristic equation ( [ main_7 ] ) have negative real parts , and it is unstable if there is at least one eigenvalue with positive real parts .
note that it is possible ( using the same method as in the previous section ) to prove the stability of the system in the case when @xmath14 and the length of the trait interval @xmath139 are small .
again , this stability does not depend on the sign of the gradient of @xmath18 . on the contrary , in the case
when @xmath14 is not small , but @xmath139 is small , we can prove that the stationary state is unstable .
thus again , there should exist a hopf bifurcation curve which separates the regions of stability and instability . for any particular function @xmath18 , construction of such a curve
should be based on equation ( [ main_7 ] ) , where we replace @xmath110 with @xmath186 .
finally , we can consider the situation where the vulnerability to predation is constant : @xmath187 . in this case
the characteristic equation reduces to : @xmath188 which can be re - written as an equivalent equation ( we are interested to find only @xmath110 with @xmath189 ) @xmath190 it is easy to see that for @xmath191 we shall always have eigenvalues with positive real part , thus the equilibrium will be unstable for any @xmath185 .
we can extend the previous techniques of stability analysis to deal with a more general case , where the kernel is given by ( [ finite_rank ] ) .
we assume that we have already calculated the values of @xmath192 and @xmath57 ( see section 2.2 for details ) .
this can be done , for instance , by using numerical methods .
equations ( [ eq : stability_1])-([eq : stability_2 ] ) now read : @xmath193 we introduce the notation @xmath194 , multiply equation ( [ evalue_f1 ] ) by @xmath16 , integrate this equation and add it to equation ( [ evalue_f2 ] ) .
we obtain : @xmath195 where @xmath196 .
we assume again that @xmath119 , and express @xmath197 as follows : @xmath198 using equation ( [ gvalue_f ] ) we obtain from ( [ evalue_f1 ] ) with @xmath199 given by ( [ sseq1sep ] ) @xmath200
\displaystyle\sum_{i=1}^n z_{i+2}k_i(x)}{\lambda+\frac{f^*a(x)}{1+\beta z^*}}.\end{aligned}\ ] ] to find the coefficients @xmath201 we multiply equation ( [ main_f ] ) by @xmath202 and integrate it over @xmath166 $ ] for @xmath203 .
we obtain the following system of linear equations : @xmath204\sum_{i=1}^n z_{i+2 } i_{ij},\end{aligned}\ ] ] where @xmath205 after re - arrangement , system ( [ main_ff ] ) reads : @xmath206,\label{main_fff}\end{aligned}\ ] ] with @xmath207 and @xmath208 , for @xmath209 .
we summarize our findings in the following theorem .
let us assume that system ( [ sseq1sep])-([sseq2sep ] ) has a unique solution @xmath210 and the determinant of the linear system ( [ main_fff ] ) is non - zero .
then the characteristic equation of the linearised system ( [ eq : stability_general1])- ( [ eq : stability_general2 ] ) for a separable kernel ( [ finite_rank ] ) is given by @xmath211 \sum_{i=1}^n z_{i+2}k_i(x)\right)\right .
\nonumber \\ & \left.\hspace{45mm}\left(\lambda+\frac{f^*a(x)}{1+\beta z^*}\right)^{-1}\right]dx=1,\label{main_fm}\end{aligned}\ ] ] where @xmath212 are the solutions of equations ( [ main_fff ] ) . in particular , the hopf bifurcation curve can be found by substituting @xmath213 into equation ( [ main_fm ] ) and solving the system of two equations for the imaginary and the real parts in a similar way as in ( [ im_2])-([re_2 ] ) .
although theorem 3.3 allows a complete characterisation of the eigenvalues , its practical implementation might require some advanced computational methods which can be as much complicated as solving the original differential equations .
however , in a particular case , when the kernel is separable , i.e. @xmath214 , the characteristic equation becomes much simpler .
we can use the explicit expressions for the stationary densities of species given in section 2.1 . after substituting the densities into ( [ main_fff ] )
we obtain the scalar equation for @xmath175 , which we substitute into ( [ main_fm ] ) to obtain @xmath215 where @xmath216 interestingly , in the particular case @xmath187 , we can easily prove based on ( [ main_71 ] ) that the stationary state is always unstable .
we also managed to prove instability for the case @xmath187 when @xmath217 , ( the result is not shown here due to a rather lengthy proof ) ; however , for the general case with an arbitrary number of terms in the kernel @xmath20 , proving ( in)stability remains an open problem .
note that in the next section , we partially address this question by providing a sufficient condition for instability for the case of @xmath187 . in the previous section we partially addressed the question of stability of the stationary state in the case @xmath187 for a simple separable kernel and found that the state is always unstable . in this subsection
, we shall consider a more general situation for an arbitrary positive kernel @xmath218 .
the case of @xmath187 is of particular practical interest since it is important to know whether or not stability would be possible in the case of structuring in the prey population in terms of growth rate only and without the selectivity in predation according to the life trait @xmath2 . in case of @xmath187
the eigenvalue problem - reads : @xmath219 we also note that the stationary biomass of prey can be found from @xmath220 utilising the above identity we rewrite the eigenvalue problem ( [ eigva1])-([eigva2 ] ) as follows : @xmath221 any @xmath110 for which there is a non - trivial vector @xmath222 , which together with @xmath110 solves equations - is an eigenvalue . in general
it is very difficult ( if not impossible ) to completely characterize the spectrum , mainly because of the inhomogeneous integral equation .
however , to show instability we only need to prove the existence of a solution @xmath223 with a corresponding non - trivial eigenvector @xmath222 . in
what follows we establish a condition which guarantees the existence of a positive eigenvalue .
to this end we utilise some well - known results from the spectral theory of positive and compact operators . for basic definitions and results
not given here we refer the reader to @xcite . [ theorem ] if the following inequality holds : @xmath224}\left\{\int_{x_1}^{x_2}k(x , y)\,dy\right\},\end{aligned}\ ] ] then the stationary state @xmath127 is unstable . assuming that @xmath225 from equation we obtain : @xmath226 note that @xmath227 holds .
with this , from equation we have : @xmath228.\end{aligned}\ ] ] hence for @xmath229 we introduce a parametrised family of bounded linear operators : @xmath230 with domain d@xmath231 . we show that for some @xmath229 this ( bounded ) linear operator has eigenvalue @xmath71 , namely its spectral radius , with a corresponding strictly positive eigenvector @xmath232 . to this end
first we notice that in the region of the parameter plane determined by @xmath233 , the operators @xmath234 are positive and irreducible , since @xmath20 is strictly positive ( see e.g. @xcite ) .
moreover , since we assumed that the kernel @xmath20 is bounded above , i.e. @xmath235 for all @xmath236 $ ] , it is shown that the operators @xmath234 are compact . we also note that along the half - line @xmath237 the spectral radius @xmath238 is constant .
next we are going to establish that at any point of the half - line @xmath237 the spectral radius of the operator @xmath234 is greater or equal than one . to this end recall
( see e.g. @xcite ) that for any bounded linear operator @xmath239 we have the following characterisation of the spectral radius : @xmath240 we have : @xmath241}\left\{\int_{x_1}^{x_2}k(x , y)\,dy\right\}\,d x \nonumber \\ & \ge 1 .
\nonumber\end{aligned}\ ] ] similarly we obtain : @xmath242 in particular , for any @xmath243 as above we obtain @xmath244 hence @xmath245 to establish the existence of a point in the region @xmath233 ( in the positive quadrant of the parameter plane @xmath246 ) , where the spectral radius equals @xmath71 it is convenient to re - parametrise the family of operators using polar coordinates .
we write : @xmath247 with this new parametrisation we have : @xmath248 note that the condition @xmath233 corresponds to the condition @xmath249 . also note that for every @xmath250 we have : @xmath251 since we assumed that @xmath20 is bounded above by some @xmath252 it follows that the spectral radius of the integral operator , which maps @xmath232 to @xmath253 is bounded above by @xmath252 , hence for every @xmath250 @xmath254 we note that we have shown above that condition implies that @xmath255 for any @xmath250 .
it follows from the intermediate value theorem and the continuity of the spectral radius function ( see e.g. ( * ? ? ?
* ch.iv.3.5 ) ) that there is a point @xmath256 in the region @xmath257 , where the spectral radius of @xmath258 equals one .
note that the spectral radius itself is an eigenvalue .
it follows from ( * ? ? ?
* theorem 6.6 ) that there is a strictly positive unique normalised eigenvector @xmath259 corresponding to the spectral radius @xmath260 : @xmath261 hence we proved the existence of a solution : @xmath262 of the eigenvalue problem - , which implies that the stationary state @xmath127 is unstable .
thus , the above theorem provides us with the sufficient condition of instability in the case of @xmath187 , in general .
food - web models including rapid evolution are now gaining more and more popularity in the ecological literature in the recognition of the fact the rapid evolution can dramatically affect population dynamics @xcite . in this paper , we have analytically investigated the recent eco - evolutionary predator - prey model proposed by morozov _
_ in @xcite , which is given by equations ( [ eq : zoo])-([eq : fish ] ) .
the previous numerical investigation shows a counter - intuitive result that rapid evolution of an organism s life trait combined with predation selectivity could stabilize this otherwise globally unstable system ( when the carrying capacity of prey is considered to be infinitely large and the functional response of the predator is assumed to be a destabilizing holling type ii response ) . since these important results were obtained only using numerical simulation and only for some specific parametrization of the inheritance kernel @xmath8 as well as the selectivity function @xmath18 , the natural question was about the generality of conclusions made in @xcite . here
we address the above question by investigating the model analytically and for generic model functions .
the main mathematical outcomes of our study are the following : \(i ) we found conditions for the existence of the positive stationary states of model ( [ eq : zoo])-([eq : fish ] ) for a class of kernels @xmath8 given by ( [ finite_rank ] ) and the arbitrary parametrization of the vulnerability to predation @xmath18 .
\(ii ) we analytically derived stability conditions for the non - trivial stationary state in the case when the kernel is separable .
in particular , when the kernel is constant and there is a perfect genetic mixing in the offspring in the prey population .
the obtained characteristic equations allow us to construct the hopf bifurcation curve without the need for a direct simulation of equations ( [ eq : zoo])-([eq : fish ] ) .
although , to construct the hopf bifurcation curve one still need to solve transcendental equations ( e.g. ( [ im_2])-([re_2 ] ) ) , this task becomes substantially faster compared to direct simulations of the integro - differential equations .
\(iii ) for an arbitrary function @xmath18 , describing selectivity of predation , we analytically proved the stability of the coexistence stationary state in the case when the saturation of the predation @xmath14 as well as the length @xmath139 of the interval of variation of the life trait @xmath166 $ ] are small . moreover , for simple kernels ( e.g. @xmath263 or @xmath264 ) , by choosing concrete parametrisations for the vulnerability to predation @xmath18 , we can easily explore the stability properties in the case when the interval @xmath166 $ ] is not small .
in particular , our investigation shows that for @xmath137 and @xmath156 , the stationary state turns out to be always stable for a constant kernel @xmath20 .
\(iv ) in contrast , for a constant @xmath187 , we obtained that stabilisation is never possible in the case of @xmath264 or even in some more complicated cases of the separable kernel .
for an arbitrary kernel @xmath218 we derived sufficient conditions of instability with @xmath187 ( see theorem [ theorem ] ) .
\(v ) the framework of stability analysis and stationary states search suggested in this paper is rather generic and can be implemented to other eco - evolutionary models involving integro - differential equations . from the biological perspective
, the analytical results obtained have clear interpretation .
stabilisation of the otherwise globally unstable predator - prey system is possible as a result of rapid evolution , and the necessary ingredients of stability are : ( a ) sufficiently large genetic variation within the prey population ( for @xmath191 the length of the interval @xmath166 $ ] should be large enough to guarantee stability , see figure 1 ) ; ( b ) there should be some selectivity of consumption of prey by its predator with respect to the life trait , i.e. @xmath15 , the sign of selectivity has no pronounced influence on stability ( see figure 1 , as well as characteristic equation ( [ linear_a2 ] ) ) ; ( c ) the saturation in predation @xmath14 should not be too large .
these requirements for stability are in a perfect agreement with the numerical results of @xcite .
we should mention here that our study of model ( [ eq : zoo])-([eq : fish ] ) is by no means complete : several key issues are still need to be addressed . in particular , it would be important to prove ( or disprove ) the stability of the stationary state in the case @xmath18 is an arbitrary positive function and when the saturation @xmath14 is small ( see section 3.2 for details ) .
derivation of an exact condition for the type of hopf bifurcation ( i.e. subcritical or supercritical ) would be of a great practical use .
verifying the requirement that @xmath18 should not be a constant to assure stability is an important challenge for future investigation .
it would also be necessary to consider another class of demographic kernels @xmath8 , which are gaussian - shaped and close to zero within large portion of the interval of @xmath166 $ ] except the area around their maximum .
biologically such kernels model the situation where the reproduction of the cohort with the life trait @xmath2 would be centred around @xmath2 with only small probability for offspring to deviate from this value .
this represents a biologically relevant case of asexual reproduction .
we are planning to address the above issues in future work .
finally , it would be rather natural to incorporate diffusion into model - .
this would allow modelling of small random variations in the genetic trait in a deterministic fashion , for instance , this can account for random variation in the trait within the life time of an organism @xcite .
structured population models with diffusion in the state space are attracting interest , see e.g. the recent papers @xcite . since diffusion has a smoothing effect , in general , it would be an interesting question to investigate whether the hopf bifurcation shown here can be sustained in the analogue model with diffusion .
here we prove theorem [ ssgeneral ] stated in section 2.3 . in the special case of @xmath52 and @xmath103 the steady state problem reads @xmath265 hence we define a parametrised family of bounded linear integral operators as follows @xmath266 with domain @xmath267 since the kernel @xmath20 is strictly positive and bounded , the integral operator @xmath268 is irreducible and compact for every @xmath269 , see e.g. ( * ? ? ?
* chapter v ) , and the frchet - kolmogorov theorem e.g. in ( * ? ? ?
* chapter x ) .
theorem 6.6 in ( * ? ? ? * chapter v ) implies that the spectral radius @xmath270 of @xmath268 is an isolated and dominant simple eigenvalue with a corresponding strictly positive eigenfunction .
it also follows from the proof of theorem 2.2 in @xcite that the spectral radius is the only eigenvalue with a positive eigenvector .
also note that the spectral radius function @xmath271 is strictly monotone decreasing , and it is continuous .
continuity of the spectral radius function is a consequence of theorem 3.16 in ( * ? ? ?
* chapter iv ) .
it follows that there exists a unique value @xmath57 for which the spectral radius @xmath272 with a corresponding strictly positive eigenvector @xmath273 .
that is we have @xmath274 we normalise this eigenvector @xmath273 , so that the stationary prey population density , which also satisfies equation , is given by @xmath275.\end{aligned}\ ] ] * acknowledgements * j. z. farkas was partly funded by a university of stirling research and enterprise support grant .
we thank professors a. gorban and s. petrovskii ( university of leicester ) for helpful discussions and comments .
we also thank the anonymous referee for helpful suggestions and comments .
e. bouin , v. calvez , n. meunier , s. mirrahimi , b. perthame , g. raoul , and r. vouituriez .
_ invasion fronts with variable motility : phenotype selection , spatial sorting and wave acceleration_. c. r. math .
paris , 350 ( 2012 ) , 761766 .
d. dube , k. , kim , a. p. alker , c.d .
_ size structure and geographic variation in chemical resistance of sea fan corals gorgonia ventalina to a fungal pathogen_. mar .
ser , . 231
( 2002 ) , 139150 .
w. gentleman , a. leising , b. frost , s. storm , j. murray .
_ functional responses for zooplankton feeding on multiple resources : a review of assumptions and biological dynamics_. deep - sea res .
ii top stud .
, oceanog . , 50 ( 2003 ) , 28472875 .
a. yu . morozov .
_ incorporating complex foraging of zooplankton in models : role of micro and mesoscale processes in macroscale patterns_. in dispersal , individual movement and spatial ecology : a mathematical perspective ( eds m lewis , p maini & s petrovskii ) , pp .
new york , ny : springer , 2011 .
morozov , e.g. arashkevich , a. nikishina , k solovyev .
_ nutrient - rich plankton communities stabilized via predator - prey interactions : revisiting the role of vertical heterogeneity_. math .
biol . , 28 ( 2011 ) , 185215 d. n. reznick , c. k. ghalambor , k. crooks .
_ experimental studies of evolution in guppies : a model for understanding the evolutionary consequences of predator removal in natural communities .
17 ( 2008 ) , 97107 .
yu . v. tyutyunov , o. v. kovalev , l. i. titova .
_ spatial demogenetic model for studying phenomena observed upon introduction of the ragweed leaf beetle in the south of russia_. math .
( 2013), . | in this paper we explore the eco - evolutionary dynamics of a predator - prey model , where the prey population is structured according to a certain life history trait .
the trait distribution within the prey population is the result of interplay between genetic inheritance and mutation , as well as selectivity in the consumption of prey by the predator .
the evolutionary processes are considered to take place on the same time scale as ecological dynamics , i.e. we consider the evolution to be rapid .
previously published results show that population structuring and rapid evolution in such predator - prey system can stabilize an otherwise globally unstable dynamics even with an unlimited carrying capacity of prey . however , those findings were only based on direct numerical simulation of equations and obtained for particular parameterizations of model functions , which obviously calls into question the correctness and generality of the previous results .
the main objective of the current study is to treat the model analytically and consider various parameterizations of predator selectivity and inheritance kernel .
we investigate the existence of a coexistence stationary state in the model and carry out stability analysis of this state .
we derive expressions for the hopf bifurcation curve which can be used for constructing bifurcation diagrams in the parameter space without the need for a direct numerical simulation of the underlying integro - differential equations .
we analytically show the possibility of stabilization of a globally unstable predator - prey system with prey structuring .
we prove that the coexistence stationary state is stable when the saturation in the predation term is low .
finally , for a class of kernels describing genetic inheritance and mutation we show that stability of the predator - prey interaction will require a selectivity of predation according to the life trait . |
fast radio bursts ( frbs ) are jansky - level millisecond duration transient events at @xmath4 ghz frequencies of unknown physical origin discovered in pulsar surveys @xcite .
they have an all - sky rate of @xmath5 to @xmath6 above fluence @xmath7 @xcite .
very recently , the location of one of these bursts frb 121102 , a repeater , has been determined to an accuracy of @xmath8 by interferometry with the jansky very large array @xcite .
this frb is found to be associated with a dwarf star - forming host galaxy at redshift z = 0.19273 @xcite .
the frb is also associated with a persistent radio source @xcite , and their separation is further pinned down to @xmath9 ( @xmath10 in physical distance ) by the european vlbi network @xcite .
it is not surprising that the distance of frb 121102 is @xmath11 gpc considering that many observational evidences had already suggested that frbs are at cosmological distances ( see * ? ? ?
* for a brief review ) .
first of all , their dispersion measures ( the column density of free electrons along the line of sight @xmath12 ) are much larger than the contribution from the interstellar medium ( ism ) in the milky way by roughly an order of magnitude .
second , if the dm is mostly contributed by an ionized nebula with electron temperature @xmath13 , then in order to avoid free - free absorption of ghz waves the size of the nebula must be larger than @xmath14 @xcite .
moreover , the strength of h@xmath15 and uv continuum flux limits combined with the free - free absorption argument suggest that at least the `` lorimer burst '' frb 010724 is at a distance @xmath16 @xcite .
confirmation of the cosmological origin means that frbs are very energetic events .
if the frb sources are isotropic ( the effect of anisotropy will be included later on ) , then the energy release is @xmath17 where @xmath18 is the fluence , @xmath19 is the luminosity distance and @xmath20 is the width of the frb spectrum . if frb radiation is incoherent , for burst duration @xmath21 , the area of the emitting region in the transverse direction for a source with sub - relativistic speed is @xmath22 , so the brightness temperature is @xmath23 where @xmath24 is the peak specific flux , @xmath25 is the angular diameter distance ( @xmath26 cm ) , @xmath27 is the lab - frame frequency , @xmath28 is speed of light , and @xmath29 is the boltzmann constant ; we use the convenient notation @xmath30 throughout the paper . note that if the source is moving toward the earth at lorentz factor @xmath31 , then @xmath32 ( the comoving frame temperature ) is smaller by a factor of @xmath33 than estimated in equation ( [ tb ] ) . for any reasonable lorentz factor ,
the brightness temperature exceeds the compton catastrophe limit of @xmath34 k. this means that frb radiation must be coherent and the emitting particles relativistic , as pointed out by @xcite .
regarding the nature of frbs , the major unknowns are their progenitors and radiation mechanism .
many progenitor models have been proposed , including : collapsing neutron stars , neutron star or white dwarf mergers @xcite , magnetar bursts @xcite , supergiant pulses from young pulsars @xcite , ( galactic ) flaring stars @xcite , relativistic jets running into clouds @xcite , asteroids colliding with neutron stars @xcite , close neutron star - white dwarf binaries @xcite , lightning in the neutron star magnetosphere @xcite , and plasma stream sweeping across the neutron star magnetosphere @xcite .
these works are only based on considerations of energetics and timescales , and the authors simply _ assumed _ that ( a fraction of ) the free energy available in the system is radiated away at ghz frequencies by coherent charge `` patches '' via e.g. curvature or synchrotron processes .
however , the properties of the emitting particles / plasma and conditions for coherent emission have not been discussed in detail and will be the focus of this paper .
@xcite used coherent curvature radiation to explain both frbs and the mjy shot pulses from the crab pulsar , but the formation and stability of the relativistic near neutral coherent patches ; particles of positive and negative charges are moving in the same direction .
the basic reason for requiring an extremely small fractional charge is that the energy of frb electromagnetic waves comes from particles initial kinetic energy in the model of @xcite .
the cooling time is much shorter than 1 ns ( see 3 ) unless an amount of net charge @xmath35 is associated with kinetic energy @xmath36 .
another reason for quasi - neutrality is the electrostatic repulsion . ] in their model were not discussed .
@xcite derived the general conditions for synchrotron maser emission ; ghz synchrotron emission requires weak ( and ordered ) magnetic field ( @xmath37 g for electrons and @xmath38 g for protons ) which can be found at distances @xmath39 from a neutron star for a dipole field configuration @xmath40 ; where @xmath31 is the lorentz factor of particles perpendicular to the b - field .
@xcite pointed out that the energy requirement for frbs may be hard to satisfy for the synchrotron maser model if frbs are at cosmological distances . in an earlier paper , @xcite proposed that synchrotron maser emission may be produced when a magneto - hydrodynamic wave generated in a magnetar giant flare interacts with the surrounding pulsar wind nebula at a distance of @xmath41 ( here @xmath42 ) .
the common problem of @xcite and @xcite models is that they did not discuss how the particle distribution ( population inversion ) for efficient maser emission can be achieved .
most frbs models are based on neutron stars ( ns ) , which could naturally explain their short durations , large energy requirement , ordered magnetic field ( hereafter b - field ) needed for coherent emission , and repetitions with intervals between @xmath43 to @xmath44 seconds ( the only repeater so far is frb 121102 , but others may also be repeating , see * ? ? ?
* ) . in this paper
, we use general physical considerations to show that the b - field in the frb source plasma should be @xmath45 and hence frbs are most likely to be produced in the magnetosphere of a ns or stellar - mass black hole .
we also note that the persistent radio source near frb 121102 is consistent with being a supernova remnant ( snr ) energized by a young ns / magnetar @xcite , although the possibility of an active galactic nuclei ( agn ) can not be ruled out @xcite .
much of the work presented in this paper was done in the summer of 2016 , but the writing up of the paper took a long time due to some pressing matter .
we have made every effort to cite papers related to this work that have been published in the meantime in this fast developing field . in 2 , we provide estimates for the size of the source region , and the strength of the electric field associated with the frb radiation at the source . in 3 , we show that coherent curvature radiation in neutron star magnetosphere can explain the frb properties , and we also provide a number of arguments to show that the b - field strength should be @xmath46 g. some predictions of this model are discussed in 4 , and a summary of the main results are provided in 5 .
we provide a broad - brush picture of the frb source in this section based on general physics considerations . considering that the frb radiation is coherent
, the size of the source along the line of sight can not be larger than the wavelength ( @xmath47 ) of the radiation we observe .
if the source is moving toward us with a lorentz factor @xmath31 , the frequency in the source frame is smaller than in the lab frame by a factor @xmath31 , and the transverse size corresponding to the wavelength is @xmath48
. since particle velocities are @xmath49 , separation of @xmath47 along the line of sight can be maintained over a lab - frame time @xmath50 . the transverse source size is often taken to be @xmath51 .
however , this need not be the case . the transverse size can be larger than @xmath48 by factor @xmath52 and still coherence can be maintained ; @xmath53 .
this is because the time delay between the arrival of photons at the observer from the opposite ends of the source in the transverse direction @xmath54 is smaller than @xmath55 provided that @xmath56 $ ] ; where @xmath57 is the angular diameter distance to the source , and @xmath58 is the distance between the source and the _ trigger point _ located behind the source from which the signal originates and propagates outward and triggers different points in the source to start radiating . for frbs at distances @xmath59 ,
the requirement on @xmath60 is @xmath61 moreover , the shape of the coherent patch in the transverse direction is unknown . in the following ,
we take the area of the coherent patch in the transverse direction to be @xmath62 where @xmath63 contains two multiplication factors @xmath64 and @xmath65 corresponding to the directions of the two principal axes in the transverse plane .
thus , the maximum coherent volume in the lab frame is @xmath66 and in the comoving frame @xmath67 this coherent patch contributes to the frb radiation for a time duration of order @xmath68 in the observer s frame , during which the source has traveled a distance @xmath69 ( in the lab frame ) toward the observer .
the patch turns off after this time and another patch lights up so that the intrinsic frb duration is much longer than @xmath70 .
there could be more than one coherent patches ( @xmath71 ) adding up incoherently , and a continuous plasma flow could produce coherent radiation for a duration longer than @xmath70 ; two sources that are separated by a distance larger than the size of a coherent zone are treated as two independent patches , and even within a coherent zone we can have multiple patches if the radiation they produce is not in phase . if the same source produces coherent radiation continuously for a time duration longer than @xmath70 ( in observer s frame ) , then we tag it as a different source after time interval @xmath70 and yet another source after @xmath72 and so on .
the total isotropic equivalent of energy release for frbs is @xmath73 erg and the isotropic equivalent luminosity is @xmath74 erg s@xmath75 ( for a cosmological distance of @xmath76 cm ) . in the far field ( fraunhofer diffraction )
limit , the solid angle within which electromagnetic ( em ) waves add up coherently is , the beaming solid angle @xmath77 . equation ( [ eq:15 ] ) is only correct in the limit @xmath78 and @xmath79 . for smaller @xmath60
s , the energy density associated with electromagnetic waves is larger than our current , conservative , estimate , because more energy is produced in a smaller coherent volume . ]
@xmath80 if there are @xmath71 distinct coherent patches , each of volume @xmath81 , contributing to the total observed frb flux _ at any given time _ , the beaming corrected total energy is @xmath82 the average lab - frame energy density in em waves in the radiating plasma is @xmath83 which is of order 10@xmath84 erg @xmath85 . therefore , the electric and magnetic field strengths associated with this em wave energy density are @xmath86 where the b - field strength is in gauss .
the electric field is very strong in the sense that @xmath87 .
thus , unless this field is almost exactly perpendicular to the local magnetic field it will accelerate electrons to highly relativistic speeds and drain the energy out of the frb radiation .
since all the photons from a patch must be traveling in nearly the same direction and in phase in order that their fields add coherently , we therefore infer that the electric and magnetic fields associated with the observed frb radiation calculated above is in a direction roughly perpendicular to the patch s velocity vector .
guided by general considerations of the last section that the b - field in the region where the observed radio photons from frbs are produced is large , we restrict ourselves to the magnetosphere of a neutron star or a stellar - mass black hole .
the curvature radiation radiation produced when charged particles stream along curved b - field lines is an efficient way in these conditions to produce large coherent radiation . in this section
, we describe a detailed curvature radiation model that can reproduce observed frb properties without any fine tuning of parameters .
let us consider an electron moving with lorentz factor @xmath31 along a b - field line of local curvature radius @xmath88 .
due to its acceleration along the b - field curvature , the electron produces em waves at frequency @xmath89 and the power radiated is @xmath90 the second expression for @xmath91 is obtained by expressing @xmath31 in terms of @xmath92 using equation ( [ nu_r ] ) , i.e. @xmath93 the electron s radiation is beamed in the forward direction within a cone of angle @xmath94 , and the pulse time in the observer frame is smaller than the lab - frame time by a factor @xmath95 .
thus , the isotropic luminosity due to one electron , in the observer frame , is given by @xmath96 since we are considering radiation from a coherent patch where all electrons are moving with almost exactly the same velocity ( @xmath31 ) and streaming along nearly parallel b - field lines , we can take the patch to be moving with a lorentz factor @xmath31 , and electrons in the patch s comoving frame to be essentially at rest .
let us consider that the electron density in the comoving frame is @xmath97 , which we express in terms of the goldreich - julian ( gj ) density @xmath98 @xcite : @xmath99 where @xmath100 @xmath101 , @xmath102 and @xmath103 are the radius , rotation period and surface b - field strength of a neutron star , and @xmath104 is the distance from the coherent patch to the center of the star .
the spin down time for the neutron star due to dipole radiation is @xmath105 so if the object underlying frbs is a neutron star with b - field strength @xmath106 g , then it can produce radio bursts for only a relatively short period of time of order a few hundred years .
we note that the surface b - field @xmath103 of the frb progenitor star may not be of dipole configuration ; higher order multipole components give a smaller spin down power and hence a longer spin down time . as our fiducial parameter ,
we take @xmath107 and use @xmath108 hereafter , so the total number of electrons in the patch that are radiating coherently is @xmath109 the isotropic equivalent luminosity , in the observer frame , due to @xmath110 coherent patches ( each containing @xmath111 electrons ) can be calculated using equations ( [ dl_iso ] ) & ( [ ntot ] ) @xmath112 or @xmath113 in arriving at the above equation we made use of equation ( [ gam ] ) .
the observed frb isotropic luminosity is @xmath114 erg s@xmath75 , which can be easily accounted for if @xmath115 , @xmath116 or @xmath117 ( or smaller values if some combination of these variables are adjusted ) .
the beaming corrected luminosity in the lab frame is @xmath118 @xmath119 the factor in the numerator on the rhs of equation ( [ l_lab1 ] ) is the solid angle of the radiation beam from the patch divided by @xmath120 , and the factor in the denominator is the ratio of the observer frame and lab frame times .
the total kinetic energy of electrons in all radiating patches in the lab frame is @xmath121 combining equations ( [ l_lab2 ] ) & ( [ e_e_tot ] ) , we find the cooling time of electrons in the lab frame @xmath122 which is much shorter ( by a factor @xmath123 ) than the lab frame time duration ( @xmath124 ) over which the patch is radiating .
this requires that there is an electric field parallel to the b - field that can accelerate electrons and sustain their lorentz factor at @xmath31 for the time duration of @xmath125 .
the required electric field in the lab frame is given by @xmath126 we can express the electric field in terms of the frb luminosity using equation ( [ liso2 ] ) @xmath127 this field , parallel to the b - field , is of the same order as the perpendicular electric field associated with the frb radio emission ( equation [ e_em ] ) .
the radiation can be sustained for a duration longer than @xmath70 in the observer s frame provided that the electric field is maintained by some process for longer than @xmath124 ( lab frame ) such as an ongoing reconnection of b - field . in
[ current - sheet ] , we provide a possible magnetic reconnection geometry in which such a parallel electric field can be maintained .
the electric current associated with the motion of electrons in the volume @xmath128 is @xmath129 .
the number density of particles with positive charge is nearly the same as the electron number density in order that the plasma is charge neutral . and
since these positively charged particles are accelerated by the electric field in opposite direction to the electrons , the total current in the region is @xmath130 .
thus , the b - field generated by the current inside the patch is @xmath131 or @xmath132 this induced b - field is in the direction perpendicular to the primary b - field ( @xmath133 ) , and the induced field lines are closed curves that lie in planes normal to @xmath133 .
thus the superposition of the two fields @xmath134 & @xmath133 will point in different directions at different locations in the source .
particles are stuck in the lowest landau state , as shown later , and therefore their motion is along b - field lines . in order for these particles to produce coherent radiation
their velocity vectors should be nearly parallel . to be more precise
, the angle between the velocity vectors for different particles should not be larger than @xmath94 .
this constraint provides a lower limit on the strength of the original b - field @xmath103 in the region : @xmath135 it is unphysical that a large number of patches would turn on _ simultaneously _ on a timescale of @xmath136 ns , so we expect that @xmath137 .
moreover , @xmath52 is at most of order unity , as been constrained in
[ sec : general ] ( eq .
[ eq:16 ] ) .
this means that frbs are most likely produced near the surface of a neutron star or the event horizon of a stellar - mass black hole in a region where the b - field strength is @xmath138 g. in this sub - section we discuss effects on ghz waves as they propagate through the magnetosphere of a neutron star .
the main goal is to ascertain whether radio waves can propagate through this media without suffering excessive absorption , and to provide a constraint on the particle density in the region where frb radiation is produced .
we show below that wave absorption is small as long as the particle number density ( @xmath139 ) is smaller than 10@xmath140 @xmath85 , which is about @xmath141 times the gj density , i.e. @xmath142 ( note that @xmath143 and @xmath144 ) .
transverse electromagnetic waves produced by the curvature radiation are linearly polarized with the electric field vector ( @xmath145 ) perpendicular to both the local b - field ( @xmath133 ) and the wave - vector @xmath146 as long as @xmath147 , where @xmath148 is the plasma frequency , and @xmath149 is electron cyclotron frequency .
the condition for the propagation of transverse em waves , with @xmath150 perpendicular to @xmath133 , in a highly magnetized plasma , is : @xmath151 ( e.g. * ? ? ?
therefore , although the plasma frequency in the frb source region is much larger than frb radio frequency , the ghz waves can still propagate through the medium at a speed that is extremely close to @xmath28 .
it can be shown that as radio waves propagate through the magnetosphere with changing b - field direction , @xmath145 adiabatically changes its direction so that it is perpendicular to both the local @xmath133 & @xmath146 , as long as @xmath152 ( see * ? ? ?
* ) . in other words ,
the polarization vector of photons is along @xmath153 while moving through the region of the magnetosphere where @xmath152 .
the column density of electrons in the source region of frb model we have described above is @xmath154 @xmath155 ( or 300 pc @xmath85 ) .
however , the contribution of the magnetosphere to the observed dm value of frbs is negligible .
this is due to the fact that the _ effective plasma frequency _ in the presence of a strong b - field is reduced by a factor of @xmath156 , and thus the contribution to the dm from the frb source region is @xmath157 pc @xmath85 .
the frb waves propagating through the medium where they are produced are subject to the free - free absorption process , which is calculated here .
the standard result for free - free absorption ( e.g. * ? ? ?
* ) can not be used for the frb system because particle motion is restricted by the strong magnetic field and is essentially confined to one dimension .
a self - consistent derivation of free - free opacity for the frb source is presented below .
consider coulomb interaction between an electron ( @xmath158 ) and a proton ( or positron )
hereafter @xmath159 moving with lorentz factors @xmath160 and @xmath161 , respectively .
these particles are streaming along strong magnetic field lines of strength @xmath103 , and they remain in the lowest landau state during the entire interaction .
the angle between their momentum vectors is @xmath162 in the lab frame .
we will use the method of virtual quanta to calculate bremsstrahlung emission and absorption , which is carried out in the rest frame of the electron ( prime frame ) , and then transformed back to the lab frame .
the lorentz factor of @xmath159 and the angle between the momentum vectors in the prime frame are given by @xmath163 where @xmath164 .
the x - axis is taken to be along the electron - momentum , and the x - y plane is the plane of @xmath158 & p@xmath165 momenta .
the magnetic field in the prime frame at the location of the electron is @xmath166 the main idea behind the virtual quanta technique is that the electromagnetic field due to @xmath159 in the prime frame is sharply peaked in a narrow region close to the particle .
this field is fourier decomposed and each frequency component is treated as a virtual photon that gets scattered by the electron . the electric field components due to @xmath159 in the prime frame
is calculated using the linard
wiechert potential .
we prefer to calculate the field in the @xmath159 rest frame and then lorentz transform it to the prime frame followed by a rotation along the z - axis by an angle @xmath167 so as to obtain field components along the @xmath168 and @xmath169 axes .
the @xmath168 component of the electric field at the location of the electron is given by @xmath170,\ ] ] where @xmath171 and @xmath172 is the minimum distance between @xmath158 and @xmath159 in the prime frame .
for strong @xmath103 field associated with the frb model , the y - component of the electric field and the magnetic field associated with @xmath159 are too weak and of too small a frequency to excite the electron to a higher landau level and therefore these fields are ignored for the bremsstrahlung calculation .
the electron motion along the x-axis is described by @xmath173 the electric field in the radiation zone due to the acceleration of the electron is @xmath174 \over c r ' ( 1 - \hat n'\cdot \vec\beta'_e)^3 } \approx { q \over c r'}\left [ \hat n ' \cos\psi ' - \hat x'\right ] |\dot{\vec{\beta}}'_e|,\ ] ] where @xmath175 the poynting flux due to the electron acceleration is given by @xmath176 the energy release per unit frequency and per unit solid angle in a @xmath158@xmath159 scattering is @xmath177 where @xmath178 this integral can be carried out using the modified bessel functions @xmath179 and we obtain @xmath180 , \end{split } \label{acc - prime}\ ] ] where @xmath181 making use of equations ( [ s - prime ] ) & ( [ acc - prime ] ) , we obtain the energy release per unit frequency and per solid angle in one @xmath158@xmath159 encounter @xmath182 or @xmath183 .
\end{split}\ ] ] for a plasma with @xmath158 and @xmath159 densities of @xmath139 & @xmath184 the bremsstrahlung emissivity is given by @xmath185 the emission is beamed along the electron momentum vector within an angle of @xmath186 , and the spectrum peaks at @xmath187 which is a factor @xmath188 larger than the ghz frequency of frbs we are interested in . therefore , we will consider the low frequency limit where @xmath189 + \xi^\prime),\quad\quad k_1(a ) \approx 1/a,\ ] ] @xmath190 is the euler - mascheroni constant .
it can be shown that for frbs , the term containing @xmath191 in the above expression for @xmath192 is much smaller than the @xmath193 term . with these approximations ,
we finally arrive at the following simplified expression for bremsstrahlung emissivity @xmath194 ^ 4 } \ln\left[{b_{max}\over b_{min}}\right],\ ] ] where @xmath195 is the angle between electron momentum vector and the photon , and @xmath162 is the angle between @xmath158 & @xmath159 momentum vectors ; @xmath196 .
the emissivity decreases very rapidly with angle for @xmath197 .
finally , we can obtain the expression for free - free absorption coefficient , including correction for the stimulated emission , in rayleigh - jeans limit using the kirchhoff s law : @xmath198 or @xmath199 ^ 4 } , \label{kff1}\ ] ] where the electron temperature @xmath200 along the magnetic field . inside the source region where the photons are moving almost parallel to electrons that produced them
@xmath201 the optical depth corresponding to the source length @xmath202 along the line of sight is @xmath203 where we took @xmath204 as per equation ( [ gam ] ) .
the optical depth due to particles moving in the direction opposite to the electrons is much smaller than that given by equation ( [ tau - ff ] ) for a electron - positron plasma . ] . for a electron - proton plasma ,
the absorption of ghz photons by protons is even smaller and can be ignored .
equation ( [ tau - ff ] ) suggests that the particle density within the frb source region should not be much larger than @xmath205 @xmath85 to avoid free - free absorption of photons .
however , the allowed density is sufficient for producing the observed typical frb luminosity of @xmath206 erg s@xmath75 for a coherent source of transverse size @xmath207 , i.e. @xmath208 ( see eq . [ liso2 ] ) .
once particles leave the acceleration zone the waves might suffer significant absorption if the outside temperature is much lower than @xmath209 .
this is , however , unlikely .
the cooling time for electrons increases dramatically outside the _ coherence zone _ , and becomes of order @xmath210 s ) @xmath211 ( where @xmath91 is taken from equation [ pe ] ) . in this case
, the electron temperature stays relativistic and the free - free absorption at @xmath4 ghz is small as long as the _ coherence zone _ is contained inside a somewhat larger acceleration zone .
moreover , as photons travel further out in the magnetosphere , the angle between their momentum vector and the local magnetic field direction ( @xmath195 ) increases , and @xmath212 decreases rapidly for @xmath213 ( as @xmath214 see eq .
[ kff1 ] ) which keeps the free - free absorption small . the free - free absorption increases with decreasing frequency as @xmath215 , and therefore waves with @xmath216 mhz are likely to be absorbed in the source region .
this might explain the lack of detection of frbs below ghz frequencies in a few cases .
the bottom line is that the ghz photons can escape the magnetosphere of a magnetar without suffering much free - free absorption provided that the plasma density in the source region @xmath217 @xmath85 .
this upper limit to the density is larger than the gj density by a factor @xmath218 .
moreover , the density required for our model to produce the observed luminosity of a typical frb of @xmath74 erg s@xmath75 is @xmath219 @xmath85 , and the transverse size of the source is @xmath220 .
the contribution to the dm of a frb from the magnetar source region is negligible due to the modification of the em wave dispersion relation in a highly magnetized plasma ( @xmath221 ) .
what we have established thus far , using general arguments , is that the b - field in the region where frb ghz radiation is produced is @xmath138 g , and the electric field component parallel to the b - field is @xmath222 esu .
this electric field needs to be sustained for at least 1 @xmath223s ( @xmath124 ) . in this sub - section
we consider one possible way that such an electric field could arise : when inclined b - field lines in the magnetosphere of a neutron star are forced to come together and reconfigure . in this process of magnetic reconnection ,
a fraction of the b - field energy is dissipated and charged particles are accelerated by @xmath224 to relativistic speeds and produce coherent curvature radiation .
the calculations presented in this sub - section are less robust than other parts of the paper because the physics of reconnection and particle acceleration are poorly understood especially when the magnetization @xmath225 is large ( @xmath226 in the frb source region ) . fortunately
, the uncertainties associated with the calculation of reconnection and @xmath224 do not affect other parts of this paper apart from some of the predictions of our model described in [ predict ] .
let us consider the b - field directions at the opposite sides of the current sheet to be inclined at an angle @xmath227 , and b - field strength to be @xmath103 ; we consider @xmath228 rad .
the b - fields on the two sides of the sheet can be decomposed as parallel and anti - parallel components ( see fig .
[ fig - current - sh ] ) . only the anti - parallel components of the b - field , which have strengths of @xmath229 ,
are dissipated inside the current sheet , whereas the parallel components with strength @xmath230 ( or the guide field ) remain intact .
let us take the plasma speed flowing into the current sheet to be @xmath231 .
the dimension of the current sheet along the line of sight ( or along the guide field ) is @xmath232 and in the transverse direction is @xmath233 .
the thickness of the current sheet is @xmath234 .
thus , the transverse area of the coherent volume is @xmath235 .
the dissipation rate of magnetic energy is @xmath236 from the cooling time argument , we require @xmath237 and hence @xmath238 where we have used @xmath239 and @xmath240 . the radiation efficiency is @xmath241 the inflow velocity is also given by @xmath242 , i.e. @xmath243 the angle @xmath244 can be roughly estimated from the condition that @xmath245 ; where @xmath246 . since @xmath247 , we find that @xmath248 .
the system becomes charge starved for larger angles , and that would trigger rapid dissipation of magnetic fields . for @xmath249 , we have @xmath250 and hence @xmath251 .
there is considerable flexibility ( and uncertainty ) in values the parameters @xmath252 , @xmath103 , @xmath60 and @xmath88 can take , and hence a wide range of efficiencies from @xmath253 to @xmath254 are allowed in the magnetic reconnection model .
the intrinsic durations of most frbs only have upper limits @xmath255 , because they were not resolved after deconvolution of the scattering broadening and dm smearing resulting from the finite channel width of radio telescopes . however , the repeating events from frb 121102 showed various observed durations ( 2.8 - 8.7 ms ) with no clear evidence of scattering broadening @xcite . also ,
frb 121002 showed resolved double - peaked profile with separation of about @xmath256 between the peaks @xcite .
if we assume that these relatively long durations are intrinsic , the process that drives the magnetic reconnection may be relatively `` slow '' , i.e. operating on timescales much longer than the light - crossing time of the ns .
the large scale steady state structure of the force - free ns magnetosphere ( e.g. dipole or multipole with twists ) can be solved ( e.g. * ? ? ?
however , the b - field configuration very close to the surface of a ns ( most likely a magnetar for frbs ) is still poorly understood , so the mechanism for forced reconnection of magnetic fields is uncertain .
one possible scenario is that magnetic flux emerges from below the ns surface due to buoyancy and then the emergent b - field reconnects with pre - existing b - field in the magnetosphere at some inclination .
another possibility is the slow movement of the ns crust where the field lines are anchored .
for example , in the first scenario , flux emergence from under the ns surface occurs on a timescale @xmath257 where @xmath258 is the depth from which the flux emerges , @xmath259 is mass density of the surface layer , @xmath260 is the alfvn speed . in [ sec - particle ] , we calculate the radiation field from a single particle moving along a fixed b - field line , and then we add up the contribution from particles with different lorentz factors and moving along magnetic field lines pointing in different directions . taking @xmath261 as an example , we show that the curvature radiation from different particles within a patch adds up coherently provided that ( i ) dispersion of the lorentz factor of particles in the region is within a factor a few of the mean lorentz factor ( @xmath31 ) ; ( ii ) the orientations of magnetic field lines at different points in the region are within an angle @xmath94 of each other ; ( iii ) the spectrum of the emergent coherent radiation can have strong fluctuations on @xmath4 mhz scale ; ( iv ) the time delay between radiation from opposite ends of the patch should be no larger than @xmath70 . in 3.3.2 & 3.3.3 , we discuss a number of physical effects that could perturb particle trajectories , generate or inhibit the generation of coherent bunches .
consider a single particle with charge @xmath35 on a circular orbit of radius @xmath88 ( the curvature radius of the b - field line ) inclined at an angle @xmath15 with respect to the line of sight or @xmath262 axis , as shown in fig .
( [ fig : geometry ] ) .
the angular frequency is @xmath263 , where @xmath264 is the particle s speed divided by @xmath28 and we assume @xmath265 to be constant with time in order to focus on the curvature radiation . the azimuthal angle at any retarded time @xmath266 is given by @xmath267 ( we have taken @xmath268 for @xmath269 ) .
the electric field at the observer s location @xmath270 due to the particle s acceleration is @xcite @xmath271 , \\ & = \frac{q\beta\omega}{cd } \left[\frac{-\sin\theta \sin\alpha\ \hat{x } + ( \cos\theta - \beta\cos\alpha)\ \hat{y}}{(1 - \beta\cos\theta\cos \alpha)^3}\right ] \end{split}\ ] ] where we have used @xmath272 , @xmath273 , @xmath274 being the vector from the particle s position to the observer , and we have neglected high order terms @xmath275 .
we show in fig .
( [ fig : e_single ] ) the x and y components of the electric field in the far field due to a single particle .
note that @xmath276 is anti - symmetric with @xmath162 and @xmath15 while @xmath277 is symmetric .
if the particle flow is continuous on a timescale of @xmath278 , @xmath276 from particles with positive and negative @xmath162 will cancel out , but @xmath279 will survive . in the following , we only consider the @xmath280 component of @xmath281 , the magnitude of which is @xmath282 \\ & \simeq
\frac{q\gamma^4}{\rho d } \mbox { , if both $ \alpha$ and $ \theta \ll \gamma^{-1}$ } , \end{split}\ ] ] where time @xmath283 in the observer frame is given by @xmath284 the emission is strong only when @xmath285 and the duration of the emission from a single particle is @xmath286 . in fig .
( [ fig : coh_vol ] ) , we plot the @xmath279 from a coherent volume of electrons with flat distributions of lorentz factors in the interval @xmath287 , orbital angles between @xmath288 , and azimuthal angle in the interval @xmath289 .
we consider three different @xmath290 of @xmath291 , @xmath292 and @xmath293 radian ; and three different @xmath294 of @xmath295 , @xmath296 , and @xmath297 ; where @xmath298 is the mean lorentz factor .
the electric field is normalized to @xmath299 . to get the actual electric field
, one needs to multiply with the total number of electrons in the coherent patch .
the general conclusion is that coherence is destroyed and the resulting electric field @xmath300 is much weaker when the separation of particles in a patch significantly exceeds @xmath47
( i.e. @xmath301 ) or when the direction of particle motion has a spread much larger than @xmath302 ( i.e. @xmath303 ) .
particle clumps with longitudinal size @xmath304 may be produced in various plasma instabilities , such as the two - stream instability ( see 3.3.2 ) .
we note that coherence at frequencies @xmath305 is unaffected if particles have a broad lorentz factor distribution .
the peak frequency of curvature radiation depends strongly on the particle lorentz factor @xmath306 , so particles with lorentz factors larger ( or smaller ) than 30 by a factor of more than 2 will radiate most of their energy at frequencies much above ( or below ) @xmath305 . during the reconnection , particles with a wide range of lorentz factors may be produced , but that would not affect the observed luminosity of @xmath11 jy at @xmath305 provided that the number of particles with lorentz factor @xmath307 ( e.g. between 20 and 40 ) is of order @xmath308 .
we also note that frb spectrum depends on the lorentz factor distribution @xmath309 .
the radiation from particles with too large lorentz factors will be beamed away from the observer line of sight , so we expect that there should be a sharp break in the frb spectrum somewhere above @xmath310 ( but not too far above 10 ghz , since the chance that an observer is within the beaming cone of particles with larger lorentz factors decreases as @xmath311 ) .
the observed frb duration is affected by the dm smearing and broadening due to scatterings in the ism / igm .
if the intrinsic duration of an frb is @xmath312 ( which is determined by the duration of magnetic reconnection ) , the number of coherent patches needed is @xmath313 ; it should be pointed out that we receive radiation from only one of these patches at a time , and the million patches are part of a semi - continuous outflow in a current sheet which we have broken up into a large number of causally connected segments for the ease of calculation . as a simple demonstration , we show in fig .
( [ fig : ran_pat ] ) the emission from 2000 identical patches that turn on and off on a timescale of @xmath314 ( in observer frame ) ; @xmath314 is taken to have the poisson distribution with mean value @xmath315 .
the spectrum has strong fluctuations in frequency intervals as small as @xmath4 mhz ( fig .
[ fig : ran_pat ] ) . to model the observed frb spectrum with rich fine - scale structure ,
one may need to add up contributions from different patches in more realistic ways .
for example , particle flow could be episodic on @xmath223s ms timescales and different patches could have different polarizations due to the rotation of the ns or non - stationary b - field configuration .
scintillation and plasma lensing probably also play an important role in shaping the observed spectrum ( e.g. * ? ? ?
the em emission directly coming from the source region is linearly polarized , with electric field perpendicular to the primary b - field and the line of sight . due to the rotation of the ns , and the magnetic field orientation in the current sheet changing on @xmath223s timescale ,
the polarization angle is likely to be time dependent .
the polarization state of photons can also change substantially as they propagate through plasma along the line of sight in the ns magnetosphere , magnetar wind nebula , and possibly supernova remnant .
one would need to take these effects into consideration in order to model and interpret the polarization properties of frb radiation .
since electrons and protons ( or positrons ) are accelerated by the nearly static electric field parallel to @xmath133 in opposite directions , the two stream instability could have an effect on the coherent radiation generation .
we show that this instability could lead to the formation of particle bunches with longitudinal size on the order of @xmath304 on timescales much shorter than @xmath316 .
we consider an electron - proton plasma , which is cold so that the state of the particles transverse to the b - field is in the lowest landau level ; the analysis here is applicable to electron - positron plasma as well .
our choice of frame is such that protons are at rest in the absence of a wave , which simplifies the algebra , but does not restrict linear stability analysis .
the notation in this subsection is different from the rest of the paper .
we denote particle density in the fluid comoving frame without a prime ( in previous sections we had used prime ) , and all remaining quantities are measured in the rest frame of the proton and are also denoted without a prime .
variables in the lab frame will be denoted by the subscript `` lab '' .
the electron and proton unperturbed densities in their comoving frames are @xmath139 & @xmath184 , and the perturbations are @xmath317 & @xmath318 respectively . the 4-velocities for electrons and protons ( as measured in the comoving frame of unperturbed proton flow ) are taken to be @xmath319 and
@xmath320 ; @xmath321 .
the perturbed equations for particle flux conservation follows from @xmath322 , and are @xmath323 @xmath324 it follows from @xmath325 and @xmath326 ( since perturbations are being carried out in the rest frame of protons ) that @xmath327 , but @xmath328 .
moreover , perturbation of the equation @xmath329 , yields @xmath330 we use this relation to eliminate @xmath331 in equation ( [ ne - pert1 ] ) @xmath332 where @xmath333 , and @xmath334 .
the momentum equation for the cold plasma is : @xmath335 or @xmath336 ; where @xmath337 is the energy momentum tensor for cold plasma , @xmath338 is the electro - magnetic second rank tensor , @xmath339 is the radiation reaction force , and @xmath35 and @xmath340 are particle charge and mass .
the z - component of the momentum equation is @xmath341 the electric field @xmath342 is in the proton / positron comoving frame , but since particles are streaming in the @xmath343direction , the @xmath344-component of electric field in the lab and electron comoving frames are also @xmath342 .
the electric force on a charged particle is balanced by the radiation reaction force , @xmath345 , for the unperturbed system .
the perturbation to @xmath346 is a non - local quantity ( because it depends on all the particles in the patch that are radiating coherently ) , and is not easy to include in the stability analysis of the system , causes particles to be accelerated to a higher lorentz factor , which then causes a higher radiative power output curvature radiation power scales as @xmath347 ( equation [ pe ] ) and the resulting higher radiation reaction force tries to restore the equilibrium state of the system . ] .
we ignore @xmath348 in the perturbation analysis presented below , and due to this we might perhaps be overestimating the importance of the two - stream instability .
the perturbation of the z - component of the momentum equation for protons and electrons are : @xmath349 and @xmath350 finally , it follows from the maxwell equations that @xmath351 .
\label{dez}\ ] ] note that the density perturbation @xmath317 is defined in the electrons comoving frame , so we have the @xmath31 factor on the right - hand side because the equation is in the proton / positron rest frame .
equations ( [ dnp ] ) , ( [ dne ] ) , ( [ du_pz ] ) , ( [ du_ez ] ) and ( [ dez ] ) are five equations in five variables that we solve to determine the dispersion relation and two - stream instability growth time . taking all perturbed variables to be proportional to @xmath352 , these five equations can be combined and reduced to the following two equations @xmath353 and @xmath354 { \delta n_e\over n_e } + ik\left [ 1 - { \omega_p^2\over \omega^2}\right ] { \delta e_z\over 4\pi q \gamma } = 0 \label{dne - dez2}\ ] ] we finally obtain the dispersion relation , which is : @xmath355 - { \omega_e^2 \over \gamma^2 } = 0 , \label{disperse}\ ] ] where @xmath356 the electron plasma frequency is @xmath357 it should be emphasized that @xmath358 and @xmath359 are wavenumber and complex frequency in proton / positron rest frame ; @xmath360 and @xmath31 are the unperturbed velocity and lorentz factor of the electron fluid in the rest frame of proton / positron , and @xmath139 and @xmath184 are particle number densities in the rest frame of electron and proton respectively . -7.3
cm it can be shown from equation ( [ disperse ] ) that two - stream instability is present when @xmath361^{3/2}/(v\gamma k/\omega_e)<1 $ ] , i.e. for sufficiently small wavenumber the system is unstable to bunching up of particles .
the growth rate of instability , im@xmath362 ) , is obtained by finding the two complex roots of equation ( [ disperse ] ) , which are complex conjugates of each other .
the growth rate for @xmath158@xmath159 and @xmath158@xmath363 plasma for several different values of @xmath31 are shown in fig .
( [ figure-1 ] ) . since @xmath31 in equation ( [ disperse ] ) is the electron lorentz factor as measured in the rest frame of protons , it is equal to @xmath364 , where @xmath365 and @xmath366 are lorentz factors of protons and electrons measured in the lab frame .
electron clumps of longitudinal size @xmath367 cm are required for the coherent curvature radiation model considered in this work .
the electron density needed to generate a typical frb luminosity of @xmath368 erg s@xmath75 is @xmath369 @xmath85 ( see eq .
[ liso2 ] and the discussion following it but note that @xmath139 in this section is same as @xmath97 in that equation )
. therefore , @xmath370 rad s@xmath75 ( eq . [ plasma_w1 ] ) .
hence , for a @xmath158@xmath159 plasma of @xmath371 ( needed for the frb curvature radiation , as per eq .
[ gam ] ) , the clump size ( @xmath372 , for the fastest growing mode ) turns out to be @xmath373 cm , and the growth time for the instability is about 30 ns in the lab frame ( see fig .
[ figure-1 ] ) ; the time required for the formation of a clump is longer ( due to causality limit ) , and is of order @xmath374s . for a @xmath158@xmath363 plasma ,
clumps of size @xmath307 cm also form on a timescale of @xmath375 @xmath223s in the lab frame . since the emission from one coherent patch lasts for 1 ns in the observer frame , which corresponds to @xmath376 ns in the lab frame , particle clumps form on a time scale smaller than that required for the coherent curvature radiation to operate .
we note that the electric field associated with the clumping of electrons is much weaker than @xmath224 estimated in equation ( [ e_parallel1 ] ) and the electric field associated with the frb em wave energy @xmath377 in equation ( [ e_em ] ) . even in the extreme case where electrons are fully separated with protons / positrons in different clumps with longitudinal separation @xmath47 , the electric field due to clumping
is roughly given by @xmath378 , where @xmath111 is the total number of particles in one clump and @xmath379 is the transverse area of a coherent patch .
for parameters required to produce frb luminosities , @xmath380 , @xmath381 and @xmath382 , we obtain @xmath383 . in reality , the clumps are only partially charged and the electric field due to clumping is much smaller than this value . therefore , the electric field associated with clumping will not be able to disperse the clumps ; the back reaction of radiation on the clump can not disperse them either , because this reaction force is balanced by the reconnection electric field @xmath224 ( eq . [ e_parallel1 ] ) .
moreover , the electric field of charged clumps is not sufficiently strong to convert plasma waves to the em waves at frb luminosities .
we showed in [ sec - particle ] that particle velocity distribution along the b - field need not be narrow for generation of coherent radiation .
however , even a small dispersion in particle velocity perpendicular to the b - field can ruin the coherence . also , the primary b - field lines along which particles in the source region are moving should be nearly parallel to each other ( within an angle of @xmath94 ) .
we show here that despite the presence of strong perpendicular electric and magnetic fields in the source region ( @xmath384 g ; equation [ b_ind1 ] ) , and collision between particles with opposite charges moving relative to each other at close to the speed of light , the distribution of particle velocity perpendicular to the b - field @xmath133 is unaffected .
the reason for this comes down to the fact that the energy of excited landau levels are so large that particles stay in the ground state in spite of the large , time dependent , fields . and
therefore , the coherence is preserved .
the energy of relativistic electrons in a magnetized medium is ( see appendix a ) @xmath385 where @xmath386 is the component of electron s momentum along the b - field line , @xmath387 is a positive integer , and @xmath388 is the electron cyclotron frequency . in the limit @xmath389 , the expression for electron energy can be approximated as @xmath390 thus , the energy of the electron corresponding to the lowest landau level ( for @xmath391 ) is @xmath392 the plasma temperature in the direction transverse to the b - field should be cold ( @xmath393 kev ) even though electrons are streaming along magnetic field lines with a lorentz factor @xmath43
otherwise higher landau levels would be populated and the emergent radiation can not be coherent .
the motion of electrons perpendicular to the b - field is damped due to the cyclotron radiation on a very short time scale of @xmath394 s ) @xmath395 ( in the lab frame ) .
so the electron velocity dispersion perpendicular to the b - field ( which is proportional to the transverse temperature ) should be very small unless there is a mechanism or instability that operates on a much shorter timescale and excites electrons to higher landau levels . from equation ( [ e - landau ] )
we see that the effective electron momentum transverse to the b - field , in the first excited level ( @xmath396 , is @xmath397 .
this is much larger than the transverse momentum kick given to the electrons , @xmath398 ( classically ) , when electrons and protons undergo coulomb collision in the frb source .
particle collisions , therefore , can not dislodge electrons from their ground landau state . the transverse electric field associated with frb radiation
is so strong ( @xmath399 esu ; equation [ e_em ] ) that it shifts the energy of the landau levels by an amount that is much larger than @xmath400 .
however , it can be shown that this shift is exactly the same for all landau levels , and since @xmath401 is larger than the em wave frequency ( @xmath11 ghz ) by a factor @xmath402 , electrons adapt to this periodic shift in the energy levels adiabatically and remain in the ground state , i.e. the strong em field associated with the radiation does not kill the coherence .
the b - field produced by charged particles in the source region streaming along @xmath133 is also strong ( @xmath403 g ; equation [ b_ind1 ] ) , and its direction changes on a length scale smaller than @xmath47 .
however , electrons adapt to this changing b - field adiabatically since the de broglie wavelength of electrons in transverse direction , @xmath404 ( see appendix a ) , is much smaller than the curvature radius of the b - field .
waves generated near the surface of a relatively young ns travel through the supernova remnant left behind in the explosion that produced the ns .
we show in this sub - section that the contribution to dm from the supernova remnant is nearly constant for a period of a few years as required by the data for frb 121102 from which 17 bursts have been detected @xcite ; this point has already been made recently by @xcite .
we also show that the optical depth of the supernova remnant for free - free absorption is small .
let us consider that 10@xmath405 of material is ejected in a supernova that produced the ns , and the initial speed of the remnant is 10@xmath406 cm s@xmath75 .
the particle density of the medium in the vicinity of the explosion is taken to be 10 @xmath85 . in this case
the deceleration radius for the remnant is @xmath407 cm , and the deceleration time is about 200 years .
the mean particle density of the ejecta @xmath408 yrs after the explosion is @xmath409 @xmath85 , and the electron column density is @xmath410 @xmath155 , if the ejecta is significantly ionized as suggested by @xcite ; where @xmath411 yrs ) .
thus , the contribution of the ejecta to the frb dm is @xmath412 pc @xmath85 . and
the change to dm in three years is @xmath413 pc @xmath85 .
seventeen radio bursts were detected from frb 121102 in a period of 3 years , and the dm values for these events were the same ( about 557 pc @xmath85 ) during this period within the error of measurements .
thus , the age of the ns for this frb can not be smaller than 50 years .
a similar lower limit on the age is obtained from the consideration that the dm from the host galaxy ( at @xmath414 ) does not exceed @xmath218 pc @xmath85 @xcite .
the free - free absorption optical depth for the remnant at 1 ghz is @xmath415 ; where @xmath416 is the temperature of the remnant .
@xcite find @xmath417 for frb 121102 , and that too suggests the ns age to be @xmath41850 yrs .
it can be inferred from the large b - field strength of ( @xmath138 g ) that the ns is most likely a magnetar .
the empirical age of galactic magnetars is @xmath5 to @xmath419 ( e.g. * ? ? ?
* ) , which may be taken as the upper limit of the age of the frb progenitor .
we provide in this sub - section a number of predictions of the model we have proposed for frbs . in essence , the model is that radio waves are produced coherently in the magnetosphere of a ns with strong b - fields that undergo forced reconnection due to perhaps the movement of the neutron star crust ( where the fields are anchored ) or emergence of sub - surface flux tubes above the ns surface .
the distortion of the b - field lines in the crust or magnetosphere builds up over a period of time , until it reaches a critical state and becomes unstable and field lines are reconfigured , dissipating a fraction of the b - field energy in a current sheet in the process .
this is a well known process and is responsible for the energy release in solar flares ( e.g. * ? ? ?
we begin by writing down the coherent curvature luminosity and the strength of the electric field parallel to the b - field in a nearly model independent form , and then specialize to magnetic reconnection to express length scales in terms of parameters appropriate for a current sheet .
@xmath420 which is essentially the same as equation ( [ liso1 ] ) , except that we have taken @xmath421 ( as discussed in [ curvature - rad ] ) , the transverse area of the source is @xmath422 , and the expression for curvature radiation frequency ( equation [ nu_r ] ) has been used to eliminate the wavelength @xmath47 .
the electric field strength ( @xmath224 ) is calculated by balancing the acceleration due to electric field with the radiation back reaction force ( as was done in 3 leading to equation [ e_parallel ] ) and we re - express that result as : @xmath423^{1/2 } , \label{e_parallel3}\ ] ] where as before @xmath88 is the radius of curvature of the b - field .
thus far we have not made any reference to reconnection and current sheets , and so the above results are broadly applicable .
next , we specialize to current sheets , and make a few predictions . since we have a rather limited first - principle understanding of the properties of current sheets and particle acceleration , the following discussion is mostly qualitative .
the thickness of the current sheet is likely to be related to the plasma frequency ( @xmath424 ) , and accordingly we take @xmath425 ; where @xmath426 is a dimensionless parameter which could be of order @xmath14110@xmath427 as suggested by numerical simulations ( e.g. * ? ? ?
* ) . substituting for @xmath428 in equation ( [ e_parallel3 ] ) we find @xmath429 which is independent of the plasma density ( @xmath97 ) .
using equation ( [ gam_cs ] ) for @xmath31 , we calculate the frequency at which particles radiate most of their energy @xmath430^{3/2 } \propto \eta_{p}^{-3 } e_\parallel^{3/2 } \rho^{1/2 } , \label{nu_peak}\ ] ] and @xmath431 the electric field @xmath224 , @xmath103 , @xmath244 ( the angle between field lines on the opposite sides of the current sheet ) , and @xmath426 likely vary substantially from one reconnection event to another . therefore , several general predictions can be made for frbs according to the model described in this work . 1 .
electrons are likely accelerated to very different lorentz factors in different bursts or different regions in one burst .
hence , in addition to the frbs observed at a few ghz frequency , the model predicts that there are other frbs that radiate at frequencies much higher than ghz ( since @xmath432 ) ; some frbs may radiate most of the energy at 10 ghz whereas others might be at @xmath106 hz .
since @xmath433 , it is unlikely to find frbs at frequencies much larger than @xmath106 hz ; this is because @xmath434 hz requires @xmath435 esu which is of order the schwinger limit on electric field strength of @xmath436 and fields of larger strength can not be sustained for long as they lead to spontaneous creation of electron - positron pairs and breakdown of the vacuum .
high energy photons can be produced due to a process such as the resonant inverse - compton scattering if the conditions in the source region are just right , but that requires fine tuning of parameters . 2 .
frb event rate should decrease with increasing frequency roughly as , and the combined angular size of all the beams in case there are multiple beams associated with a source .
the frb rate @xmath437 should be taken as a rough estimate . ]
@xmath438 , or perhaps more steeply , because of possible enhancement to the beaming associated with the transverse source size .
the intrinsic durations of frbs at frequencies much larger than ghz may be of the same order as ghz - frbs provided that there are a large number of independent beams ( each of angular size @xmath94 ) that contribute to the total duration . on the other hand , if the radiation is produced by a single beam sweeping across the observer line of sight which is the most conservative possibility for these transients then the observed duration at higher frequencies would be smaller than a few ms since the higher frequency radiation is produced by electrons with larger lorentz factors .
the luminosity function of frbs in a fixed frequency band should be broad since @xmath439 .
this might be the reason for the large flux variations ( by at least an order of magnitude ) observed for the different outbursts of the repeating frb 121102 @xcite .
the intrinsic coherent radiation luminosity decreases with decreasing frequency , and the observed pulse width increases rapidly at lower frequencies due to wave scattering ( roughly as @xmath440 , but see * ? ? ?
* ) by turbulence in the ism / igm and dm smearing ( @xmath441 ) .
these effects cause the observed flux to decrease sharply at lower frequencies . to make matters worse , lower frequency waves may have difficulty in getting out of the supernova ejecta as a result of free - free absorption .
so there is likely to be a low frequency cut - off to the frb observations which might be of order a few hundred mhz .
the very high brightness temperature of frbs suggest that the radiation process is coherent , which means that the comoving source size can not be larger than the comoving - frame wavelength @xmath442 of the radiation we observe .
thus , the electric field associated with radiation at the source is of order 10@xmath1 esu ( and the b - field is 10@xmath1 g ) .
furthermore we show that the curvature radiation can account for the observed frb properties without any need for fine tuning .
the current associated with charged particles motion along field lines produces transverse b - field ( with roughly cylindrical shaped field lines ) that is sufficiently strong to perturb particle trajectories and destroy coherence unless the primary field is @xmath46 g ( see eq .
[ b0-lim ] ) .
the strong b - field also ensures that electrons stay in the ground state of landau levels in spite of several strong perturbations that are present in the source region that otherwise would excite particles to higher landau level and destroy coherence .
a particle density of @xmath443 @xmath85 ( in lab frame ) is sufficient to account for the observed isotropic equivalent frb luminosity 10@xmath444 erg s@xmath75 ; the total number of charged particles in one coherent patch with lorentz factors @xmath445 is of order 10@xmath446 ; a coherent patch produces radiation for @xmath447 ns , and therefore the total number of particles responsible for a typical frb transient is @xmath448 , and the mass of this matter is about a kg for an electron - positron plasma .
however , because the particles are radiating in phase , their radiative cooling time is extremely small of order 10@xmath449 s or six orders of magnitude smaller than the wave period unless there is a powerful acceleration mechanism that balances the radiative losses and maintains the particle speed .
an electric field of strength @xmath450 esu , parallel to the primary b - field , is the required mechanism to sustain the particle motion for at least the lab - frame travel time ( @xmath451 ) over which ghz em waves are produced .
such an electric field could be produced during forced magnetic reconnection near the surface of a magnetar .
this could also explain why we see the bursts repeat episodically .
the total energy release in these bursts , corrected for beaming , is estimated to be of order @xmath2 ergs , whereas the total energy in the b - field is at least @xmath452 ergs .
the intrinsic durations of some frbs may be much longer than the light - crossing time of a ns .
if this is true , the process that drives the magnetic reconnection should be relatively `` slow '' .
one possible scenario is that b - field flux emerges from below the ns surface due to buoyancy ( e.g. * ? ? ?
* ; * ? ? ?
* ) and then reconnects with pre - existing b - field in the magnetosphere .
another possibility is the slow movement of the ns crust where the fields are anchored .
it is currently unclear what this process might be .
according to the model we have described , the lorentz factor of electrons ( and hence the peak frequency of the spectrum ) , and the isotropic luminosity are dependent mainly on the thickness of the current sheet @xmath453 ( @xmath424 being the plasma frequency ) and the strength of the electric field parallel to the b - field , @xmath224 .
these two parameters could have large variations among different bursts or at different locations of one burst .
thus , according to the model we have described , the luminosity function for frbs should be very broad .
also , the mechanism we have described should produce short duration bursts at frequencies much larger than 1 ghz ( up to @xmath454 hz ) .
the rate of bursts , however , is predicted to decrease with increasing peak frequency because higher frequency photons require larger lorentz factor of particles or smaller curvature radius of b - field .
the former is a much stronger effect , and the burst rate is expected to decrease with frequency at least as @xmath455 .
there are several issues that we have not discussed in detail , which future works should address .
in particular , studying the propagation of polarized waves through the ns magnetosphere and surrounding medium in the host galaxy may help our understanding of the odd spectrum of frbs and their polarization properties .
the distortion of b - field lines , and reconnection for an extremely large magnetization parameter plasma ( @xmath456 ) are also topics that require separate investigations .
we thank m. bailes , d. lorimer , e. petroff , m. lyutikov , a. spitkovsky , a. piro , j. cordes and a. stebbins for discussions of frb observations and possible emission mechanisms , b. zhang for reading the manuscript and for his excellent comments , and f. guo and s. paban for useful discussions regarding magnetic reconnection and quantum effects in strong magnetic field respectively .
we are highly indebted to the referee , jonathan katz , for his careful reading of the manuscript and for his numerous excellent comments and suggestions that substantially improved the paper and clarified a number of conceptual issues .
wl was funded by the named continuing fellowship at the university of texas at austin .
akgn , t. , miralles , j. a. , pons , j. a. , & cerd - durn , p. 2016
, , 462 , 1894 arons , j. , & barnard , j. j. 1986 , , 302 , 120 burke - spolaor , s. , & bannister , k. w. 2014 , , 792 , 19 champion , d .- j . , petroff , e. , kramer , m. , et al . 2016 , chatterjee , s. , law , c. j. , wharton , r. s. , et al .
2017 , arxiv:1701.01098 cheng , a. f. , & ruderman , m. a. 1979 , , 229 , 348 cordes , j. m. , & wasserman , i. 2016 , , 457 , 232 connor , l. , sievers , j. , & pen , u .- l . 2016 , , 458 , l19 dai , z. g. , wang , j. s. , wu , x. f. , & huang , y. f. 2016 , arxiv:1603.08207 falcke , h. , & rezzolla , l. 2014 , , 562 , a137 geng , j. j. , & huang , y. f. 2015 , , 809 , 24 ghisellini , g. 2017 , , 465 , l30 goldreich , p. , & julian , w. h. 1969 , , 157 , 869 gu , w .- m . , dong , y .- z . , liu , t. , ma , r. , & wang , j. 2016 , , 823 , l28 hankins , t. h. , & eilek , j. a. 2007 , , 670 , 693 ho , w.c.g . , dong , l. , potekhin , a.y . , and chabrier , g. , 2003 , apj 599 , 1293 kashiyama , k. , & murase , k. 2017 , arxiv:1701.04815 kashiyama , k. , ioka , k. , & mszros , p. 2013
, , 776 , l39 katz , j. i. 2014 , , 89 , 103009 katz , j. i. 2016 , modern physics letters a , 31 , 1630013 katz , j. i. 2016 , , 826 , 226 katz , j. i. 2016 , , 818 , 19 katz , j. i. 2016 , arxiv:1611.01243 katz , j. i. 2017 , arxiv:1702.02161 keane , e. f. , & kramer , m. 2008 , , 391 , 2009 keane , e. f. , & petroff , e. 2015 , , 447 , 2852 kulkarni , s. r. , ofek , e. o. , neill , j. d. , zheng , z. , & juric , m. 2014 , , 797 , 70 loeb , a. , shvartzvald , y. , & maoz , d. 2014 , , 439 , l46 lorimer , d. r. , bailes , m. , mclaughlin , m. a. , narkevic , d. j. , & crawford , f. 2007 , science , 318 , 777 lu , w. , & kumar , p. 2016
, , 461 , l122 luan , j. , & goldreich , p. 2014
, , 785 , l26 lyutikov , m. , burzawa , l. , & popov , s. b. 2016 , arxiv:1603.02891 lyubarsky , y. 2014 , , 442 , l9 marcote , b. , paragi , z. , hessels , j. w. t. , et al .
2017 , , 834 , l8 masui , k. , lin , h .- h . ,
sievers , j. , et al .
2015 , , 528 , 523 mereghetti , s. 2008 , , 15 , 225 metzger , b. d. , berger , e. , & margalit , b. 2017 , arxiv:1701.02370 mickaliger , m. b. , mclaughlin , m. a. , lorimer , d. r. , et al .
2012 , , 760 , 64 murase , k. , kashiyama , k. , & mszros , p. 2016
, , 461 , 1498 muslimov , a. , & page , d. 1995 , , 440 , l77 pen , u .- l . , & connor , l. 2015 , , 807 , 179 petroff , e. , van straten , w. , johnston , s. , et al .
2014 , , 789 , l26 petroff , e. , johnston , s. , keane , e. f. , et al .
2015 , , 454 , 457 petroff , e. , bailes , m. , barr , e. d. , et al .
2015 , , 447 , 246 petroff , e. , barr , e. d. , jameson , a. , et al .
2016 , , 33 , e045 piro , a. l. 2012 , , 755 , 80 piro , a. l. 2016 , , 824 , l32 popov , s. b. , & postnov , k. a. 2010 , evolution of cosmic objects through their physical activity , 129 potekhin , a. y. , & chabrier , g. 2003 , apj , 585 , 955 rane , a. , lorimer , d. r. , bates , s. d. , et al .
2016 , , 455 , 2207 ravi , v. , shannon , r. m. , bailes , m. , et al .
2016 , science , 354 , 1249 romero , g. e. , del valle , m. v. , and vieyro , f. l. , 2016 , phy rev d , 93 , 3001 rybicki , g. b. , & lightman , a. p. 1979
, new york , wiley - interscience , 1979 .
393 p. scholz , p. , spitler , l. g. , hessels , j. w. t. , et al .
2016 , , 833 , 177 sironi , l. , giannios , d. , & petropoulou , m. 2016 , , 462 , 48 shibata , k. , & magara , t. 2011 , living reviews in solar physics , 8 , 6 spitler , l. g. , scholz , p. , hessels , j. w. t. , et al .
2016 , , 531 , 202 tendulkar , s. p. , kaspi , v. m. , & patel , c. 2016 , , 827 , 59 tendulkar , s. p. , bassa , c. g. , cordes , j. m. , et al .
2017 , , 834 , l7 totani , t. 2013 , , 65 , thornton , d. , stappers , b. , bailes , m. , et al . 2013 , science , 341 , 53 vigan , d. , & pons , j. a. 2012 , , 425 , 2487 vigan , d. , rea , n. , pons , j. a. , et al .
2013 , , 434 , 123 xu , s. , & zhang , b. 2016 , , 832 , 199 zhang , b. 2014 , , 780 , l21 zhang , b. 2017 , arxiv:1701.04094
we calculate the energy states of a charged particle in a strong b - field .
particle speed along the field line is highly relativistic , and the field is very strong so that the particle spin is aligned with the field .
we , therefore , ignore particle spin , and consider klein - gordan equation with magnetic potential ( the electric field is taken to be zero ) : @xmath457 or @xmath458 let us consider a uniform b - field , @xmath459 , and vector potential @xmath460 corresponding to it .
the wave function has a non - trivial dependence on the @xmath461-coordinate , and we express it as @xmath462 and substitute that in the klein - gordan equation to obtain @xmath463 = 0,\ ] ] where @xmath464 the wave function @xmath469 , for the landau state @xmath387 , is given by : @xmath470^{1/4 } e^{-{m\omega_c y_1 ^ 2\over 2 \hbar } } h_n(y_1 \sqrt{m\omega_c/\hbar}),\ ] ] where @xmath471 is the hermite polynomial of n - th order @xmath472 the spatial extent of the wave - function in the ground state is @xmath473 this is a result that we use in 3 . | we use the observed properties of fast radio bursts ( frbs ) and a number of general physical considerations to provide a broad - brush model for the physical properties of frb sources and the radiation mechanism .
we show that the magnetic field in the source region should be at least 10@xmath0 gauss .
this strong field is required to ensure that the electrons have sufficiently high ground state landau energy so that particle collisions , instabilities , and strong electro - magnetic fields associated with the frb radiation do not perturb electrons motion in the direction transverse to the magnetic field and destroy their coherent motion ; coherence is required by the high observed brightness temperature of frb radiation . the electric field in the source region required to sustain particle motion for a wave period
is estimated to be of order 10@xmath1 esu .
these requirements suggest that frbs are produced near the surface of magnetars perhaps via forced reconnection of magnetic fields to produce episodic , repeated , outbursts .
the beaming - corrected energy release in these bursts is estimated to be about @xmath2 ergs , whereas the total energy in the magnetic field is at least @xmath3 ergs .
we provide a number of predictions for this model which can be tested by future observations .
one of which is that short duration frb - like bursts should exist at much higher frequencies , possibly up to optical .
radiation mechanisms : coherent non - thermal - methods : analytical - radio : fast bursts , theory |
with the recent start of the up and coming experimental program of the large hadron collider ( lhc ) at cern , the search for new physics effects at the tev scale enters into an exciting era .
in particular , due to the expected copious production of top quark events , clues of physics beyond the fermi scale could be evidenced .
the flavor changing neutral current ( fcnc ) transitions of the top quark are good processes to look for signals of new physics , since the new dynamical effects are likely more evident in those top quark processes that are forbidden or strongly suppressed in the standard model ( sm ) @xcite . in the sm , the fcnc top transitions @xmath5 , @xmath6 , @xmath7 , and @xmath8 arise at the one loop level , being considerably gim suppressed , as they have branching ratios ranging from @xmath9 to @xmath10 @xcite .
the decays involving in the final state the quark @xmath11 instead of @xmath12 are much more suppressed due to kobayashi maskawa effects . on the other hand ,
the three body @xmath13 @xcite and @xmath14 @xcite transitions have also been studied within the context of the sm .
it was found @xcite the surprising result that the branching rations for these decays are respectively one order of magnitude larger and of the same order of magnitude when compared with the two
body @xmath5 decay . beyond the sm ,
fcnc top quark decays have been the subject of considerable interest in the literature , as they can have branching ratios much larger than the sm ones .
these type of top transitions have been studied in some extended theories such as the two higgs doublet model ( thdm ) @xcite , supersymmetric ( susy ) models with nonuniversal soft breaking @xcite , susy models with broken r
parity @xcite , models with extra @xmath15 gauge bosons @xcite , and even more exotic scenarios @xcite .
similar results were obtained within the context of effective theories @xcite . in this work ,
we are interested in using the mass difference @xmath16 in the @xmath0 mixing recently observed by the babar @xcite and belle @xcite collaborations to impose a bound on the branching ratios for the fcnc top quark @xmath1 and @xmath2 decays .
it turns out that due to electromagnetic gauge invariance , the @xmath1 transition is somewhat restricted , as it can occurs only through two electromagnetic gauge structures of dipolar type , which are characterized by lorentz tensor structures of the form @xmath17 ( transition magnetic dipole moment ) , with @xmath18 the photon momentum , or @xmath19 ( transition electric dipole moment ) .
the @xmath2 decay occurs through the analogous chromomagnetic and chromoelectric dipole moments .
these interactions are naturally suppressed , as they can arise in renormalizable theories only through quantum fluctuations of one loop or higher orders .
an outline of this paper is as follows . in section [ effective - lagrangian
] we briefly study the couplings arising from the effective lagrangian .
section [ contributions ] is devoted to obtain the leading contribution of the @xmath20 and @xmath21 couplings to the mass difference @xmath16 of the @xmath0 mixing .
finally , in section [ final - remarks ] we comment our results .
the most general structure for the @xmath22 and @xmath23 couplings arise from the following dimension six @xmath24invariant effective lagrangian : @xmath25 where @xmath26 and @xmath27 stand for left and right handed doublet and singlet of @xmath28 , respectively and a sum over quark flavor is implied . here ,
@xmath29 , @xmath30 , and @xmath31 are the gauge tensors associated with the @xmath32 , @xmath28 , and @xmath33 groups , respectively . on the other hand , @xmath34 , with @xmath35 the sm higgs doublet .
in addition , the @xmath36 are the coefficients of general @xmath37 matrices defined in the flavor space , whereas @xmath38 represents the new physics scale
. this parametrization allow us to treat in a model independent way fcnc transitions in the up quark sector and @xmath39 . ]
mediated by massless gauge bosons .
after spontaneous symmetry breaking , @xmath40 can be diagonalized as usual via the @xmath41 and @xmath42 unitary matrices , which relate gauge states to mass eigenstates . in particular , the @xmath43 and @xmath44 vertices are given by the following lagrangian : @xmath45 where @xmath46 are vectors in the flavor space .
notice the presence of the global factor @xmath47 , which arises from the higgs mechanism and is common to both electromagnetic and strong couplings . in the above expression ,
@xmath48 and @xmath49 are flavor matrices given by @xmath50 where @xmath51 gev is the fermi scale and @xmath52 ( @xmath53 ) stands for cosine ( sine ) of the weak angle . to generate vector
mediated fcnc effects at the tree level in the effective theory , it is assumed that the @xmath54 matrices diagonalize the yukawa mass matrix @xmath55 but not the matrices @xmath56 and @xmath57 . from the above expressions , it is immediate to derive the vertex functions for the @xmath20 and @xmath21 ( @xmath58 ) couplings , which can be written as @xmath59 where @xmath60 and @xmath61 .
in addition , @xmath62 and @xmath63 .
as we will see below , it is not possible to isolate the contribution of the @xmath64 and @xmath65 ( @xmath66 ) couplings from the @xmath0 mixing .
this contribution is given through an amplitude that depends symmetrically on both vertices .
due to this , it is not possible to obtain bounds for the @xmath6 and @xmath5 decays without making additional assumptions .
we now are in position of deriving the contribution of the @xmath20 and @xmath21 vertices to the mass difference @xmath16 of the @xmath0 mixing .
these contributions occur through short distance effects characterized by the box diagrams shown in fig.[fig1 ] . neglecting the external momenta ,
the @xmath20 and @xmath21 couplings generate an amplitude given by : @xmath68 the integral involved in ( [ qdiv ] ) is of the form @xmath69 with @xmath70 and @xmath71 , and it diverges quadratically at @xmath72 . as discussed in reference @xcite , the use of cutoffs to regularize quadratic or higher divergent integrals would violate the decoupling theorem , as this scheme introduce a strong dependence on the new physics scale . in our case , the use of a cutoff leads to an amplitude proportional to @xmath73 , which would overestimate our prediction .
only a logarithmic dependence on the new physics scale is physically acceptable @xcite ( for some applications of dimensional regularization in effective lagrangians see , for instance , refs .
thus , we do not use a cutoff to regularize our amplitude , instead we will use dimensional regularization , which allow one to convert the quadratic divergence into a logarithmic one through the following limit @xcite : @xmath74 . \label{eq2}\end{aligned}\ ] ] the arising integral with logarithmic divergence , @xmath75 , can be treated in the usual manner , hence a renormalization scheme can directly be applied to the amplitude in ( [ qdiv ] ) .
once the integral is solved , one obtains the result for the amplitude : @xmath76 where the ultraviolet divergence is contained in @xmath77 , with @xmath78 and @xmath79 the euler s constant . here , @xmath80 is the dimensional regularization scale .
following refs.@xcite , the divergent term in this amplitude can be absorbed by renormalizing the appropriated coefficients of the complete effective lagrangian since it already contains all the invariants allowed by the sm symmetry .
the invariants needed to absorb the divergences are of the form @xmath81 , with @xmath82 and @xmath83 , etc . using the @xmath84 renormalization scheme with @xmath85 , the renormalized amplitude can be written as follows : @xmath86 where we have defined the dimensionless function @xmath87 .
it is easy to see that this amplitude can be obtained directly from a four quark effective interaction @xmath88 where the dimension
six operator @xmath89 is well known in the literature @xcite . in the above expression , a factor of @xmath90
was introduced in order to compensate the @xmath91 wick contractions . on the other hand , the mass difference @xmath92 is given by @xmath93 which in our case takes the form @xmath94 where the expressions for @xmath95 given in ref.@xcite were used . here
, @xmath96 is the @xmath92 decay constant .
we will use the cleo collaboration determination @xmath97 @xcite .
the factor @xmath98 is unknown , but lattice calculation @xcite leads to @xmath99 , although in vacuum saturation and in the heavy quark limit it approximates to the unity . in our numerical analysis
, we will use @xmath100 . using the value @xmath101 gev ,
one obtains @xmath102 from the experimental side , the heavy flavor averaging group @xcite interpretation of the current data leads to a mass difference of @xmath0 given by @xmath103 gev @xcite . to bound the @xmath104 coefficients
, we demand that the above contribution does not exceed the experimental uncertainty , which leads to @xmath105 where it was assumed that @xmath106 .
this assumption is reasonable @xcite , as a new physics scale in the tevs region is expected . considering the electromagnetic and strong contributions one at the time , one obtains @xmath107 , so that a conservative point of view allow us to obtain the bounds @xmath108 and @xmath109 without making additional assumptions . on the other hand ,
the branching ratios for the @xmath1 and @xmath2 decays are given by @xmath110 from these expressions and from eqs.([b1],[b2 ] ) , we obtain the following bounds for these decays @xmath111 where the approximation @xmath112 gev for the total top decay width was used .
it is worth comparing our results with those predictions obtained in some specific models .
most of the known results are on the @xmath6 and @xmath5 decays , which are strongly suppressed in the sm , with branching ratios of order of @xmath10 and @xmath113 @xcite , respectively .
the branching ratios for the decays involving the quark @xmath11 instead of @xmath12 are even more suppressed by a kobayashi maskawa factor of @xmath114 . beyond the sm ,
these branching ratios are considerably enhanced .
for instance , in the thdm - iii , the @xmath6 and @xmath5 decays can have a branching ratio in the range @xmath115 and @xmath116 @xcite , respectively . on the other hand , the respective predictions in some susy models @xcite
are of order of @xmath117 and @xmath118 .
an effective lagrangian analysis of higgs mediated fcnc leads to branching ratios for these decays of order of @xmath119 and @xmath120 @xcite .
all these predictions are consistent with our bound , as it is expected that the branching ratios for the @xmath1 and @xmath2 decays are lower by two or higher orders of magnitude than those involving the quark @xmath12 .
to conclude , we would like to comment on the possible detection of the @xmath1 and @xmath2 decays at the lhc . on the light of our bounds for their branching ratios
, one can conclude that they are within reach of this collider , given the important production of @xmath121 events of several millions per year . in a purely statistical basis those channels with branching ratios larger than about @xmath120 do have the chance of being detected .
however , the observability of a particular channel decay depends on several factors , background and systematics may reduce this value by several orders of magnitude depending of the particular signature . for instance
, the @xmath2 decay would require a large branching in order to be detected as it is swamped by hadronic backgrounds .
however , the important difference of about three orders of magnitude between the observability parameter and our bound makes quite possible the observation of this channel .
as far as the @xmath1 mode is concerned , it could be detected even with a relatively small branching ratio because it would be produced in a cleaner environment . in conclusion
, our analysis suggests that the recent observation of @xmath0 mixing does not exclude the possibility of observing the rare top quark decays @xmath1 and @xmath2 at the lhc .
since the @xmath122 and @xmath109 couplings are not directly constrained by this experimental result , their observation at the lhc would be more probable .
simultaneous contribution of the @xmath122 and @xmath123 couplings to the @xmath0 mixing .
the contribution of the @xmath109 and @xmath23 couplings also is given through this type of diagrams.,width=336 ]
we acknowledge financial support from conacyt and sni ( m ' exico ) . | the recently observed mass difference of the @xmath0 mixing is used to predict the branching ratios of the rare top quark decays @xmath1 and @xmath2 in a model independent way using the effective lagrangian approach .
it is found that @xmath3 and @xmath4 , which still may be within reach of the lhc collider .
= = # 1tempboxa= tempdima=0.5tempdima 0.5tempboxa -tempdima tempboxa # 1tempboxa= tempdima=0.5tempdima 0.5tempboxa -tempdima tempboxa # 1 |
in classical differential geometry ( `` a language the muse did not sing at my cradle '' - see below ) , _ parallel transport _ is defined in the context of a _ connection _ on a smooth bundle @xmath4 .
the latter can mean a covariant derivative operator , a differential 1-form or a set of horizontal subspaces in the tangent bundle @xmath5 the corresponding _ parallel transport _
@xmath6 is constructed by lifting a path in @xmath0 to a _ unique ! _ path in @xmath7 with specified starting point .
holonomy _ is given by the evaluation of @xmath8 on @xmath9 , the space of based loops in @xmath0 .
the _ holonomy group _ is the image as a subgroup of the structure group of the bundle . that it is a group follows from the uniqueness of the lifting .
it is well defined up to conjugation depending on the choice of base point .
if @xmath4 is only a fibration of topological spaces , the situation is different : we still can lift paths but not uniquely .
however , the lifts are related by homotopies , so that the right action of @xmath9 on the fibre @xmath10 is well behaved up to homotopy , i.e. @xmath11 is homotopy commutative , where @xmath12 is multiplication in @xmath9 and @xmath13 is the action . in fact , there is a whole sequence of coherent higher homotopies , which we review below .
( because @xmath14 denotes travelling along @xmath15 first and then along @xmath16 , the action will be written as right action : @xmath17 ) perhaps the oldest treatment in algebraic topology ( i learned it as a grad student from hilton s introduction to homtopy theory @xcite - the earliest textbook on the topic ) is to consider the long exact sequence , where @xmath10 is the fibre over a chosen base point in @xmath0 , @xmath18 ending with @xmath19 of course , exactness is very weak at the end since the last three are in general only sets , but exactness at @xmath20 is in terms of the action of @xmath21 on @xmath22 this passage to homotopy classes obscures the ` action ' of @xmath9 on @xmath10 .
initially , this was referred to as a _
homotopy action _ , meaning only that @xmath23 was homotopic to @xmath24 for @xmath25 and @xmath26 in those days , at least at princeton , there was no differential geometry until milnor gave an undergrad course my final year there .
notice that our book ` characteristic classes ' @xcite considers differential forms only in appendix c , added much later .
i think this lack of de rham theory in my graduate education was the result of serre s thesis which triumphed over characteristic 0 , cf .
colloque de topologie , bruxelles ( 1950 ) .
only recently have i learned of even greater significance of the metric geometry for the boundary map @xmath27 @xcite .
similarly , i learned only much lateer of the notion of _ thin homotopy _ which quotients @xmath9 to a group without losing so much information .
just recently , johannes huebschman led me to a paper of kobayashi ( from 1954 ! ) where he is already using what is now called thin homotopy in terms of parallel transport and holonomy for smooth bundles with connection . back in 1966 , in the mexican math bulletin @xcite , a journal not readily available@xcite (
although summarized in @xcite ) , i showed that in the topological setting of fibrations ( satisfying the homotopy lifting property ) , there was a notion of ` parallel ' transport not dependent on having a connection .
this meant not only the above homotopy action , but in fact an sh ( or @xmath28)-action , which is to say the adjoint @xmath29 was an @xmath28-map .
( not long after , nowlan @xcite studied more general @xmath28-actions of fibres on total spaces . ) note that the homotopy lifting property does _ not _ imply unique lifting , so even the composition of paths is _ not _ necessarily respected when lifting . for my purposes , it was sufficient to consider transport along based loops in the base , though the arguments allow for transport along any path in the base
. inspired by block - smith @xcite and igusa @xcite , abad and schaetz @xcite showed that ` flat ' smooth parallel transport ( e.g. as in @xcite ) can be derived from the @xmath28 version of de rham s theorem due to gugenheim @xcite ; this allows them to extend parallel transport to an @xmath28-functor .- connections data , as in @xcite . ]
they work with what is known as the @xmath3-groupoid @xmath30 of a space @xmath0 and its _ representations up to homotopy _ on a fibre bundle @xmath31 .
the @xmath3-groupoid @xmath30 can be represented as a simplicial set : the singular complex @xmath32 this led me to revisit my earlier work on ` parallel ' transport and extend work to the fibrations setting , i.e. without any smoothness or connection operator or differential form or any infinitesimal input .
what i am after here is an analog of parallel transport along paths extended to parallel transport over ( maps of ) simplices , not just 1-simplices .
just as there is only linguistic difference between an action of a group on an object and a representation of that group , the same is true in the homotopy setting , but sometimes a homotopy action means only the existence of a single homotopy and some times the usual coherent collection of higher homotopies ; when in doubt , it is best to prepend @xmath28 or _ strong homotopy_.
we first recall what are rightly known as moore paths @xcite on a topological space @xmath33 .
let @xmath34 be the nonnegative real line . for a space @xmath35 let @xmath36 be the subspace of _ moore paths _
@xmath37 of pairs
@xmath38 such that @xmath39 is constant on @xmath40 .
there are two maps * @xmath41 * @xmath42 * @xmath43 recall composition @xmath44 of moore paths in @xmath45 is given by sending pairs @xmath46 such that @xmath47 to @xmath48 which is constant on @xmath49 , @xmath50=\f|[0,r]$ ] and @xmath51 for @xmath52 . an identity function @xmath53 is given by @xmath54 where @xmath55 is the constant map on @xmath56 with value @xmath57 .
composition is continuous and gives , as is well known , a category/@xmath3-groupoid structure on @xmath45 .
if we had used the ` ancient ' poincar paths @xmath58 , we would have had to work with an @xmath28-structure on @xmath59 indeed , it was working with that standard parameterization which led to @xmath28-structures @xcite .
for a category @xmath60 , we denote by @xmath61 the set of n - tuples of composable morphisms . in partcular , we will be concerned with @xmath62 we will write @xmath63 for @xmath64 and @xmath65 for @xmath66 , back in 1988 @xcite , i referred to _ strong homotopy representations _ , but today i will use the _ representation up to homotopy _ terminology , having in mind the generalization that comes next .
[ thetas ] a _ representation up to homotopy _ of @xmath67 on a fibration @xmath31 is an @xmath28-morphism ( or shm - morphism @xcite ) from @xmath67 to @xmath68 ; that is , a collection of maps @xmath69 ( where @xmath70 consists of @xmath71-tuples @xmath72 where the @xmath73 are composable paths , constant on @xmath74 , and @xmath75 ) such that @xmath76 @xmath77 is a fibre homotopy equivalence and satisfies the usual / standard relations : * @xmath78 * @xmath79 _ that the parameterization is by cubes , as for sugawara s strongly homotopy multiplicative maps rather than more general polytopes , reflects the fact that @xmath45 and @xmath80 are strictly associative .
strictly speaking , referring to @xmath81 as an @xmath28-map raises issues about a topology on @xmath82 the adjoint formulas above avoid this difficulty .
1ex even more difficult , at least for exposition and detailed proofs , would be the use of ` poincar ' loops requiring @xmath83 to use @xmath84 instead of @xmath85 _ since our construction uses in a crucial way the homotopy lifting property , we first construct maps @xmath86 such that the desired @xmath83 are then recovered at @xmath87 _ the special case _
@xmath88 _ corresponds to a path lifting function . in a little known paper , dyer and eilenberg
call such a @xmath89 an action of the path space @xmath90 on @xmath7 and point out that its existence is equivalent to @xmath91 having the homotopy lifting property . _ the idea is that if @xmath92 has been defined satisfying these relations for all @xmath93 , the @xmath94 will fit together to define @xmath95 on all faces of the cube except for the face where @xmath87 in analogy with the horns of simplicial theory , we will talk about filling an _ open box _
, meaning the boundary of the cube minus the open face , called a _ lid _ , where @xmath96 ( compare horn - filling in the simplicial setting ) .
use the homotopy lifting property to ` fill in the box ' in @xmath7 after filling in the trivial image box in @xmath97 that box in @xmath0 is a trivial box since it is just the composite path @xmath98 it might help to consider the cases @xmath99 consider @xmath100 lift @xmath101 to a path @xmath102 starting at @xmath103 define @xmath104 by @xmath105 and @xmath106 now lift @xmath107 to a path @xmath108 starting at @xmath109 and lift @xmath110 to a path @xmath111 starting at @xmath112 these lifts fit together to define a map to @xmath7 , which will be the restriction of the desired map on the open 2-dimensional box of the desired map @xmath113 this open box has an image in @xmath0 which can trivially be filled in .
regarding the filling as a homotopy , the map to @xmath7 on the open 2-dimensional box can be filled in by lifting that homotopy .
[ parallel ] ( cf .
theorem a in @xcite ) for any fibration @xmath4 , there is an @xmath28-action @xmath114 of moore(b ) on @xmath7 such that @xmath115 is a fibre homotopy equivalence .
this action is unique up to homotopy in the @xmath28-sense .
in theorem b in @xcite , i proved further : given an @xmath28-action @xmath114 of the moore loops @xmath9 on a space @xmath10 , there is a fibre space @xmath116 such that , up to homotopy , the @xmath28-action @xmath114 can be recovered by the above procedure .
if the @xmath28-action @xmath114 was originally obtained by the above procedure from a fibre space @xmath4 , then @xmath117 is fibre homotopy equivalent to @xmath91 .
this construction gave rise to the slightly more general ( re)construction below .
it can also be generalized to give an @xmath3-version of the borel construction / homotopy quotient : @xmath118 for an @xmath28-action @xcite .
inspired by block - smith @xcite and igusa @xcite , abad and schaetz @xcite look not at just composable paths , but rather look at the singular complex @xmath2 . for a singular @xmath119-simplex @xmath120 ,
there are several @xmath119-tuples of composable paths from vertex 0 to vertex @xmath119 by restriction to edges , in fact , @xmath121 such . given @xmath122 ,
we denote by @xmath123 the fibre over vertex @xmath124 .
following e.g. abad - schaetz @xcite ( based on abad s thesis and his earlier work with crainic ) , we make the following definition of a _ representation up to homotopy _ , where we take a singular @xmath119-simplex @xmath122 to be ( the image of ) @xmath125 with the @xmath91-th face @xmath126 being @xmath127 however , we keep much of the notation above rather than switch to theirs . a _ representation up to homotopy _ of @xmath2 on a fibration @xmath31 is a collection of maps @xmath128 which assign to any @xmath119-simplex @xmath120 a map @xmath129 satisfying , for any @xmath130 , the relations : @xmath131 is the identity on @xmath132 for any @xmath133 , @xmath134 is a homotopy equivalence . for any @xmath135 and @xmath130 , @xmath136 @xmath137 @xmath138 _ in definition [ parallel ] , we worked with moore paths so that the @xmath28- map was between strictly associative spaces . here
instead the compatible 1-simplices compose just as e.g. a pair of 1-simplices and are related to a single 1-simplex only by an intervening 2-simplex .
associativity is trivial ; the subtlety is in handling the 2-simplices and higher ones for multiple compositions .
the idea of constructing a representation up to homotopy is very much like that of theorem 1 , the major difference being that instead of comparing two different liftings of the composed paths which are necessarily homotopic , we are comparing a lifting e.g. of a path from 0 to 1 to 2 with a lifting of a path from 0 to 2 if there is a singular 2-simplex @xmath139 .
however , note that @xmath140 plays the role of @xmath141 of moore paths in the above formulas . _
for any fibration @xmath4 , there is a representation up to homotopy of @xmath2 on @xmath142 _ the fact that the representation up to homotopy is by fibre homotopy equivalence ( as for the action of moore(b ) ) is justified by the following : since a simplex @xmath122 is contractible , the pullback @xmath143 is fibre homotopy trivial over @xmath122 .
choose the requisite lifts in @xmath143 using a trivialization corresponding to the homotopy we want to lift and then map back into @xmath7 .
_ _ in contrast to the smooth bundle case where a connection provides unique path lifting , the fibration case is considerably more subtle since path and homotopy lifting is far from unique . _
the proof is in essence the same as that for theorem [ parallel ] .
the desired @xmath83 will appear as the missing lid on an open box ( defined inductively ) which is filled in by homotopy liftings @xmath95 of a _ coherent _ set of maps @xmath144 where @xmath145 is the set @xmath146 are given in terms of iterated convex linear functions . for @xmath147 ,
define @xmath148 for @xmath149 , write @xmath150 for @xmath151 , define @xmath152 and then @xmath153 see figure [ cube2simplex ] .
these have probably been written elsewhere ; if you find them , let me know . ] by coherent , i mean respecting the facial structure of the cubes and simplices .
closely related are coherent maps @xmath154 where @xmath155 denotes the set of paths , i.e. @xmath156 such maps were first produced by adams @xcite in the topological context by induction using the contractability of @xmath145 .
later specific formulas were introduced by chen @xcite and , most recently , equivalently but more transparently , by igusa @xcite . more precisely : @xmath157 is the trivial path , constant at @xmath158 and @xmath159 is the ` identity ' .
for any @xmath135 , @xmath160 and @xmath161 @xmath162 one way to describe the relation between the @xmath163 and the @xmath164 in words is : travel from vertex 0 partway toward vertex 1 then straight partway toward vertex 2 then straight partway toward vertex 3 etc .
( see figure[mine ] . ) and the @xmath164 ] note that these are slighty different from the version of @xmath164 given by igusa ; see figure [ igusa ] taken from @xcite .
hopefully the pattern is clear . correspondingly , the liftings @xmath166 form a collection of maps which assign to any @xmath119-simplex @xmath120 a map @xmath167 satisfying , for any @xmath130 , the relations : @xmath168 is the identity on @xmath169 for any @xmath170 , @xmath171 is a homotopy equivalence . for any @xmath135 , @xmath172 @xmath173 @xmath174 1ex the desired @xmath83
is again recovered at @xmath87 2ex the maps @xmath163 can be interpreted as homotopies @xmath175 and so subject to the homotopy lifting property .
for example , @xmath176 is a path which can be lifted as in theorem 1 to give @xmath177 . then @xmath178 such that @xmath158 maps to the ` identity ' path @xmath179 while @xmath180 maps to the concatenated path @xmath181 ( henceforth , we will assume paths have been normalized to length 1 where appropriate . )
now lift the homotopy @xmath182 to a homotopy @xmath183 between @xmath184 and @xmath185 in particular , @xmath186 gives the desired homotopy @xmath187 the situation becomes slightly more complicated as we increase the dimension .
the case @xmath188 is illustrative . the faces @xmath189 and @xmath190
lift just as @xmath139 had via @xmath191 , but that lift must then be ` whiskered ' by a rectangle over @xmath192 which glues onto @xmath193 . in a less complicated way @xmath194 is lifted so that vertex @xmath180 agrees with the end of the ` whisker ' which is the lift of @xmath195 .
thus the total lift of @xmath196 ends with the desired @xmath197 the needed whiskering ( of various dimensions ) is prescribed by the @xmath198 relations of definition [ thetas ] to be satisfied . ]
further details are left to the industrious reader .
in @xcite , i showed how to construct a fibration from the data of an strong homotopy action of @xmath9 on a ` fibre ' @xmath10 .
if the action came from a given fibration @xmath199 , the constructed fibration was fibre homotopy equivalent to the given one . for _ representations up to homotopy _ , a similar result applies using analogous techniques , with some additional subtlety .
first we try to construct a fibration naively . over each 1-simplex @xmath122 of @xmath2 , we take @xmath200 and attempt to glue these pieces appropriately . for the one
simplices @xmath195 and @xmath201 , we have @xmath202 which tells us how to glue @xmath203 to @xmath204 at vertex 1 , but , since @xmath205 is not the composite of @xmath202 and @xmath206 , we can not simply plug in @xmath207 over @xmath139 .
however , we can plug in @xmath208 since @xmath209 will supply the glue over vertex 2 . to describe the total space of the fibration ( or at least a quasi - fibration ) , we use the special maps @xmath210 now return to the description of the fibration @xmath211 above . in greater precision , @xmath212 the attaching maps over the vertices @xmath158 and @xmath180 are obvious as are the projections to the edges of @xmath213 on @xmath214 , the attaching maps are obvious except for the face @xmath215 where it is given by @xmath216 so as to be compatible with the projection @xmath217 the result is at least a quasi - fibration @xmath218 and can be replaced up to fibre homotopy equivalence by a true fibration ( cf .
@xcite .
notice that although the definition of representation up to homotopy was in terms of fibrations , in fact it really needs only the collection of fibres @xmath219 for the 0-simplices of @xmath2 .
the equivalence in the appropriate sense between _ representations up to homotopy of @xmath2 _ and _ fibrations over @xmath0 _ follows as for theorem b in @xcite .
physics is often written in terms of smooth structures , differential forms and ` geometrically ' in terms of connections . from a topological point of view
, parallel transport is the more basic notion .
in particular , string theory and string field theory has inspired string topology , inititiated by chas and sullivan , and a variety of @xmath3-algebras .
i look forward to the corresponding representation - up - to - strong homotopy theory feeding back into physics .
james d. stasheff .
-spaces and classifying spaces : foundations and recent developments . in _ algebraic topology ( proc .
pure math .
xxii , univ .
wisconsin , madison , wis . , 1970 )
_ , pages 247272 .
soc . , providence , r.i .
, 1971 .
james d. stasheff .
parallel transport and classification of fibrations . in _
algebraic and geometrical methods in topology ( conf .
topological methods in algebraic topology , state univ .
new york , binghamton , n.y . , 1973 )
_ , pages 117 .
lecture notes in math .
, vol . 428 .
springer , berlin , 1974 . | parallel transport in a fibre bundle with respect to smooth paths in the base space@xmath0 have recently been extended to representations of the smooth singular simplicial set @xmath1 .
inspired by these extensions , i revisit the development of a notion of ` parallel ' transport in the topological setting of fibrations with the homotopy lifting property and then extend it to representations of @xmath2 on such fibrations .
closely related is the notion of ( strong or @xmath3 ) homotopy action , which has variants under a variety of names . |
qcd sum rule analysis has suggested that a mass - shift occurs in proportion to the baryon density of medium . in nuclear matter , a few to 10 % reduction of vector meson masses is expected @xcite and confirmation of this fact is of current experimental interest .
although the importance of a quantitative investigation on this subject is obvious , it seems difficult to tackle it by present lattice technology.@xcite a new attempt to measure the derivative of mass @xmath1 in terms of the chemical potential , @xmath2 at @xmath3 has been made in this work . at @xmath4 , since the density depends cubically on @xmath5 , @xmath2 vanishes . on the other hand , at finite temperature , a linear dependence on the chemical potential
is expected and we will have a chance to see the density effect .
we use 2 flavors of the wilson fermion for the fermionic part of the action .
the action is given by @xmath6 where @xmath7 and @xmath8 stand for the gauge and the fermionic parts respectively , and @xmath8 has the form , @xmath9 for @xmath7 we use the standard wilson action .
the wilson matrix @xmath10 with chemical potential is written as@xcite @xmath11 \\ \nonumber & & -\kappa [ ( 1-\gamma_4)u_4(m)e^{\mu}\delta_{m , n+\hat{4 } } \\ \nonumber & & + ( 1+\gamma_4)u_i^\dagger(n-\hat{4})e^{-\mu}\delta_{m , n-\hat{4}}]\end{aligned}\ ] ] the hadron correlation function @xmath12 is given by @xmath13 where @xmath14 may be written as @xmath15 . to derive the expression of the mass susceptibility on the chemical potential , we take a derivative of @xmath12 with respect to the chemical potential @xmath5 . > from eq.([eq1 ] ) one obtains @xmath16 on the other hand from eq.([eq2 ] ) one obtains @xmath17 .
% \label{eq4}\end{aligned}\ ] ] eqs.([eq3 ] ) and ( [ eq4 ] ) should be equal to each other .
eq.([eq3 ] ) is evaluated by monte carlo simulations .
eq.([eq4 ] ) contains 4 unknown fitting parameters for a given @xmath18 .
@xmath14 and @xmath19 are determined from a fit to the hadronic correlation function of eq.([eq2 ] ) .
then these results are used for eq.([eq4 ] ) as fixed parameters and , thus the chemical potential dependent terms @xmath20 and @xmath21 are extracted from a two parameter fit to eq.([eq4 ] ) .
in the present study we aim at obtaining the mass susceptibility at @xmath23 . at @xmath23 , @xmath24 .
thus the calculation of eq.([eq3 ] ) reduces to calculate the remaining term @xmath25 .
now we specify the hadronic operator . for the present study
we take the @xmath0 meson operator , which is given by , @xmath26 we take @xmath27 here .
the derivative of @xmath8 with respect to @xmath5 is @xmath28\psi^i(m ) \\
\nonumber & & \equiv - \kappa \sum_{i = u , d } \sum_{n , m } \bar{\psi}^i ( n)[b_1 - b_2 ] \psi^i(m ) , \end{aligned}\ ] ] where @xmath29 @xmath30 let us consider @xmath31 where @xmath32 .
this term contains a connected and disconnected diagrams .
we neglect the disconnected part , which is unimportant for the present study .
eq.([eq5 ] ) is written as @xmath33 where the chemical potential @xmath5 is set to zero .
similar calculations are applied for the remaining terms ( for @xmath34 and for @xmath35 ) .
simulations are done on a lattice size of @xmath36 at @xmath37 . to find a finite temperature transition point ,
the polyakov loop and its susceptibility are calculated , as shown in fig.1 the finite temperature transition happens around @xmath38 . the chemical potential effect on hadron masses at @xmath23 is expected to start in the vicinity of the transition .
we take @xmath39 and @xmath38 for the present study .
we use 30 configurations generated by the hybrid monte carlo algorithm for each @xmath40 .
the configurations are separated by 40 - 50 trajectories .
figs.2 and 3 show the derivative of the hadronic correlation function @xmath41 with respect to @xmath5 as a function of the largest lattice direction : @xmath42 .
the hadronic correlation function is measured in the x - direction .
although the signals within the current statistics are still too noisy , we fit the data using eqs.([eq3 ] ) and ( [ eq4 ] ) . the mass susceptibility , @xmath43 ,
is consistent with zero , while @xmath44 could be finite and positive .
the effect of chemical potential on the mass susceptibility is suppressed by @xmath45 where @xmath46 is the baryon mass .
our results may show that @xmath46 at our simulation parameter is too heavy to obtain a reasonable signal within limited statistics .
our hope is that signal may be enhanced if a small quark mass is used .
one of the authors ( t.t . ) would like to thank electric technology research foundation of chugoku for financial support .
calculations reported here were done on ap1000 at fujitsu parallel computing research facilities and sx-4 at rcnp , osaka univ . 9 i.m.barbour , s.e.morrison , e.g.kepfish , j.g.kogut and m-p.lombardo , nucl .
b ( proc . suppl . ) 60a ( 1998 ) 220 t.hatsuda and s-h.lee , phys . rev .
c46 ( 1992 ) 34 .
p.hasenfratz and f.karsch , phys .
lett . 125b ( 1983 ) 308 .
a.nakamura , phys .
lett . 149b ( 1984 ) 391 , acta physica polonica b16 ( 1985 ) 635 | we study the effects of the chemical potential on the @xmath0 meson mass at finite temperature .
our preliminary results show that some effects are seen in the vicinity of the phase transition point .
although the signal is still too noisy to obtain conclusive physical results within limited statistics , the mass susceptibility is consistent with zero . |
due to the discovery of non - zero neutrino masses by super - kamiokande through neutrino oscillation experiments in 1998 , the standard model of particle physics was forced to be modified so as to theoretically include the neutrino masses .
so far only the mass - squared differences of neutrino species have been measured by neutrino oscillation experiments , which are reported to be @xmath9 @xcite and @xmath10 @xcite . however , absolute values and their hierarchical structure ( normal or inverted ) have not been obtained yet although information on them is indispensable to build new particle physics models . in particle physics , some new ideas and new future experiments have been proposed to measure the absolute values and/or determine the hierarchy of neutrino masses , e.g. , through tritium beta decay in katrin experiment @xcite , neutrinoless double - beta decay @xcite , atmospheric neutrinos in the proposed iron calorimeter at ino @xcite and the upgrade of the icecube detector ( pingu ) @xcite , and long - baseline oscillation experiments , e.g. , no@xmath11a @xcite , j - parc to korea ( t2kk ) @xcite or oki island ( t2ko ) @xcite , and cern to super - kamiokande with high energy ( 5 gev ) neutrino beam @xcite . on the other hand , such nonzero
neutrino masses affect cosmology significantly through suppression of growth of density fluctuation because relativistic neutrinos have large large thermal velocity and erase their own density fluctuations up to horizon scales due to their free streaming behavior . by measuring power spectra of density fluctuations
, we can constrain the total neutrino mass @xmath12 @xcite and the effective number of neutrino species @xmath13 @xcite through observations of cosmic microwave background ( cmb ) anisotropies and large - scale structure ( lss ) .
we can also obtain additional information about @xmath13 from the cmb bispectrum @xcite .
furthermore , we can constraint on the neutrino isocurvature perturbations from cmb observations @xcite .
the robust upper bound on @xmath12 obtained so far is @xmath14 ev ( 95 @xmath15 c.l . ) by the cmb observation by planck ( see ref .
@xcite ) . for forecasts for future cmb observations , see also refs .
@xcite . for current constraints and forecasts for combinations of cosmological and laboratory experiments ( oscillation and neutrinoless double - beta decay ) , see also ref.@xcite .
moreover , by observing the power spectrum of cosmological 21 cm line radiation fluctuation , we will be able to obtain useful information on the neutrino masses @xcite , and its asymmetry ( lepton number asymmetry ) @xcite .
besides , we can obtain further information about the neutrino masses from observations of the hi distribution in the post - reionization era @xcite .
that is because the 21 cm line radiation is emitted ( 1 ) long after the recombination ( at a redshift @xmath16 ) and ( 2 ) before an onset of the lss formation .
the former condition ( 1 ) gives us information on smaller neutrino mass ( @xmath17 ev ) , and the latter condition ( 2 ) means we can treat only a linear regime of the matter perturbation , which can be analytically calculated unlike the lss case .
in actual analyses , it is essential that we combine data of the 21 cm line with those of cmb because the constrained cosmological parameter space is complementary to each other .
for example , the former is quite sensitive to the dark energy density , but the latter is relatively insensitive to it . on the other hand ,
the former has only a mild sensitivity to the normalization of matter perturbation , but the latter has an obvious sensitivity to it by definition . in pioneering work by @xcite , the authors tried to make a forecast for constraint on the neutrino mass hierarchy by combining planck with future 21 cm line observations in case of relatively degenerate neutrino masses @xmath18 ev . additionally , in our previous work @xcite , we investigate the detectability of the mass hierarchy by combination of a futuristic 21 cm line observation ( omniscope @xcite ) and cmb polarization observation ( polarbear or cmbpol ) when the total neutrino mass is relatively small value @xmath19 ev . the cmb b - mode polarization produced by cmb lensing gives us more detailed information on the matter power spectrum @xcite at later epochs . in that work
, we found that the combinations have enough sensitivity to distinguishing the neutrino mass hierarchy when total neutrino mass is @xmath20 ev or less . in this paper ,
as for 21 cm line observation , we particularly focus on square kilometer array ( ska ) @xcite , which is planned in the 2020s and more realistic observation than omniscope in near future , and forecast allowed parameter regions for the neutrino masses , the effective number of neutrino species and the mass hierarchy . in our analysis , we consider combinations of ska with ground - based cmb polarization observations , such as polarbear-2 or simons array , which can measure the gravitational lensing of cmb with a high degree of accuracy . besides , we take into account including the information of baryon acoustic oscillation ( bao ) observations , such as dark energy spectroscopic instrument ( desi ) @xcite .
this paper is organized as follows . in section [ sec:21 cm ] , we briefly review 21 cm line observation , analytical methods used in this paper , and how neutrino properties are constrained by the 21 cm line observation . in section [ sec
: cmb ] and [ sec : baofisher ] , we explain analytical methods of cmb and bao observation , respectively .
we show our results in section [ sec : results ] , and section [ sec : conclusion ] is devoted to our conclusion .
in this section , we briefly review basic methods to use the 21 cm line observation as a cosmological probe . for further details ,
we refer to refs .
@xcite .
the 21 cm line of neutral hydrogen atom is emitted by transition between the hyperfine levels of the 1s ground state , and the hyperfine structure is induced by an interaction of magnetic moments between proton and electron .
it can be observed as the differential brightness temperature relative to the cmb temperature @xmath21 : @xmath22 where @xmath23 is the comoving coordinate of the 21 cm radiation , @xmath24 represents the redshift , @xmath25 is the spontaneous decay rate of 21 cm line transition , @xmath26 ghz is the 21 cm line frequency , @xmath27 is the number density of hydrogen and @xmath28 is the fraction of neutral hydrogen .
@xmath29 is the spin temperature , which is defined by @xmath30 , where @xmath31 and @xmath32 are the number densities of singlet and triplet states of neutral hydrogen atom , respectively . here
@xmath33 is the temperature corresponding to 21 cm line , and @xmath34 is its wavelength .
@xmath35 is the gradient of peculiar velocity along the line of sight . from now on
, we assume that @xmath36 because we focus on the epoch of reionization ( eor ) during which this condition is well satisfied . in general ,
the brightness temperature is sensitive to details of inter - galactic medium ( igm ) .
however , with a few reasonable assumptions , we can eliminate this dependence from the 21 cm line brightness temperature @xcite . at the epoch of reionization
long after star formation begins , x - ray background produced by early stellar remnants heats the igm .
therefore , gas kinetic temperature @xmath37 becomes much higher than the cmb temperature @xmath38 .
furthermore , the star formation produces a large amount of ly@xmath39 photons sufficient to couple @xmath40 to @xmath37 through the wouthuysen - field effect @xcite . in this scenario , @xmath41
are justified at @xmath42 , and @xmath43 does not depend on @xmath40 .
next , we are going to consider fluctuations of @xmath44 . by expanding the hydrogen number density @xmath45 and the ionization fraction @xmath46 ( @xmath47 ) as @xmath48 and @xmath49
, we can rewrite eq . to be @xmath50 where we assume that @xmath51 and neglect the term including spin temperature .
@xmath52 is the spatially averaged differential brightness temperature at redshift @xmath24 and given by @xmath53 where @xmath54 is the primordial @xmath8he mass fraction . by denoting the fluctuation of @xmath55 as @xmath56 , the 21 cm line power spectrum @xmath57 in the @xmath58-space
is defined by @xmath59 by treating the peculiar velocity @xmath60 as a perturbation and using that its fourier transform is given by @xmath61 ( @xmath62 is the cosine of the angle between the wave vector and the line of sight ) , the power spectrum can be written as @xmath63 where @xmath64 and @xmath65 here , @xmath66 and @xmath67 , where @xmath68 and @xmath69 are the power spectra defined in the same manner as eq .
for the fluctuation of hydrogen number density @xmath70 and that of ionization fraction @xmath71 . therefore , @xmath72 traces the fluctuation of matter , which includes information on cosmological parameters such as neutrino mass . @xmath73 and @xmath69 can be neglected as long as we consider the era when the igm is completely neutral . however , after the reionization starts , these two spectra significantly contribute to the 21 cm line power spectrum .
although a rigorous evaluation of these power spectra may need some numerical simulations , we adopt the treatment given in ref .
@xcite , where it is assumed that @xmath74 and @xmath75 have specific forms which match simulations of radiative transfer in refs . @xcite .
the explicit forms of the power spectra are parametrized to be @xmath76^{-\gamma_{xx } / 2 } \mathcal{p}_{\delta\delta } ( k ) , \\ \label{eq : pxdelta } \mathcal{p}_{x \delta } ( k ) & = & b_{x\delta}^2 ~e^ { - \alpha_{x\delta } ( k r_{x\delta } ) - ( k r_{x\delta})^2 } \mathcal{p}_{\delta\delta } ( k),\end{aligned}\ ] ] where @xmath77 , @xmath78 , @xmath79 , @xmath80 and @xmath81 are parameters which characterize the amplitudes and the shapes of the spectra . @xmath82 and @xmath83 represent the effective size of ionized bubbles . in our analysis
, we adopt the values listed in table [ tab : pxx_xdelta ] as the fiducial values of these parameters .
.fiducial values of the parameters in @xmath84 and @xmath85 ( see eqs . and ) @xcite .
[ cols="^,^,^,^,^,^,^,^,^ " , ] now , we show the specifications of the cmb observations . in order to obtain the future constraints , we consider planck @xcite , polarbear-2 and simons array , whose experimental specifications are summarized in table [ tab : cmb_obs_mass ] .
the latter two experiments are ground - based precise cmb polarization observations . for the analysis about planck and polarbear-2 ( or simons array ) , we combine both the experiments , and assume that a part of the whole sky @xmath86 is observed by both the experiments , and the remaining region @xmath87 is observed only by planck .
therefore , we evaluate a total fisher matrix of cmb @xmath88 by summing the two fisher matrices , @xmath89 where @xmath90 is the fisher matrix of the region observed by both planck and polarbear-2 ( pb-2 ) [ or simons array ( sa ) ] , and @xmath91 is that by planck only .
in addition , we calculate noise power spectra @xmath92 of the cmb polarization ( @xmath93 or @xmath94 ) in @xmath95 with the following operation .
( 1 ) : : @xmath96@xmath97 ( 2 ) : : @xmath98 @xmath99^{-1}\end{aligned}\ ] ] since we assume that the cmb temperature fluctuation observed by polarbear-2 or simons array is not used for constraints on the cosmological parameters , the temperature noise power spectrum @xmath100 is equal to * @xmath101*. this reason is that the cmb temperature fluctuation observed by planck reaches almost cosmic variance limit .
( planck observation greatly helps to determine other cosmological parameters such as the normalization of power spectra , energy densities of the dark matter , baryons and dark energy ) .
therefore , the constraints are not strongly improved if we include the cmb temperature fluctuation observed by polarbear-2 or simons array .
in this section , we briefly explain analysis methods about the baryon acoustic oscillation ( bao ) . in the early universe , baryons and photons are strongly coupled and their fluctuations ( fourier components ) of the mixed fluid oscillate by the pressure of radiation . at the time of the decouple between them ,
a characteristic peak feature remains at the sound horizon .
the scale can be used for a standard ruler of distance .
therefore , we can get the information of the distance and the hubble expansion rate by measurements of the bao scale for matter fluctuations . in this paper ,
we especially consider galaxy surveys for the bao observation . here
we introduce the fisher matrix of bao experiments .
the observables of bao are the comoving angular diameter distance @xmath102 and the hubble parameter @xmath103 ( and more specifically , @xmath104 and @xmath105 are the observables ) .
the fisher matrix is given by @xmath106 where @xmath107 and @xmath108 are the variances of @xmath109 and @xmath110 in the bao observation respectively , @xmath111 is the error of the systematics , and we assume that the observed redshift range is divided into bins , whose width and central redshift values are represented by @xmath112 and @xmath113 , respectively . here , @xmath114 is the index of the redshift bins .
their variances @xmath109 and @xmath110 are determined by the fitting formulae of bao presented by @xcite , and they are given by @xmath115 here , @xmath116 is the comoving survey volume and expressed as @xmath117 where @xmath118 is the survey solid angle .
@xmath119 is the non - linear evolution factor , which represents the erasure of baryon oscillation features by the non - linear evolution of density fluctuations . in our analysis
, we use the following function for @xmath119 , @xmath120 where @xmath121 is the redshift at which the improvement in the baryon oscillation accuracy saturates ( for fixed survey volume and number density ) .
additionally , in the analysis of the bao observation , we use the following parameters , @xmath122 where @xmath123 is the dimensionless hubble parameter . according to @xcite , we assume the following systematic error , @xmath124 the set of cosmological parameters related to the bao observation are only @xmath125 or @xmath126 when we assume that the universe is flat and the dark energy is the cosmological constant .
we estimate the sensitivity of the bao observation only in the analysis of the neutrino mass , the number of neutrino species and the mass hierarchy . in the analysis , we focus on the dark energy spectroscopic instrument ( desi ) @xcite , which is a future large volume galaxy survey .
the survey redshift range is @xmath127 ( we do not include the ly-@xmath39 forest at @xmath128 for simplicity ) and the sold angle is @xmath129 $ ] . in our analysis , we divide the redshift range into 18 bins , in other words @xmath130 @xcite . additionally , in the same manner as @xcite , when we combine bao with the other observations , we add a 1% @xmath131 prior to the fisher matrix .
the prior of the hubble parameter is achievable in the next decade .
the fisher matrix of the hubble prior is given by @xmath132 where @xmath133 is the fiducial value of @xmath131 . if we choose the hubble parameter as a dependent parameter , it is necessary to translate the fisher matrix into that of the chosen parameter space .
under the transformation of a parameter space @xmath134 , the translated fisher matrix is give by @xcite @xmath135 by using this formula , under the translation of @xmath126 @xmath136 @xmath125 , the fisher matrix in the new parameter space is written as @xmath137
in this section , we present our results for projected constraints by the 21 cm line , cmb and bao observations on cosmological parameters , paying particular attention to parameters related to neutrino ( the total neutrino mass , the effective number of neutrino species and the neutrino mass hierarchy ) . when we calculate the fisher matrices , we choose the following basic set of cosmological parameters : the energy density of matter @xmath138 , baryons @xmath139 and dark energy @xmath140 , the scalar spectral index @xmath141 , the scalar fluctuation amplitude @xmath142 ( the pivot scale is taken to be @xmath143 @xmath144 ) , the reionization optical depth @xmath145 , the primordial value of the @xmath8he mass fraction @xmath146 and the total neutrino mass @xmath147 .
fiducial values of these parameters ( except for @xmath148 ) are adopted to be @xmath149 @xmath150 , which are the best fit values of the planck result @xcite . here
, we numerically evaluate how we can determine the effective number of neutrino species ( in section [ subsec : const_hie ] ) , and the neutrino mass hierarchy ( in section [ subsec : const_hie ] ) , by combining the 21 cm line observations ( ska phase 1 or phase 2 ) with the cmb experiments ( planck + polarbear-2 ( pb-2 ) or simons array ( sa ) ) and the bao observation ( desi ) . in the former analysis
, we fix the neutrino mass hierarchy to be the normal one , and set the fiducial value of the total neutrino mass @xmath148 and the effective number of neutrino species @xmath13 to be @xmath151 or @xmath152 and @xmath153 .
next , in the latter analysis , we fix @xmath13 to be @xmath154 , and set the fiducial values of the @xmath148 and the mass hierarchy parameter @xmath155 to be @xmath156 ( normal hierarchy ) or @xmath157 ( inverted hierarchy ) . to obtain fisher matrices
we use camb @xcite for calculations of cmb anisotropies @xmath158 and matter power spectra @xmath159 . in order to combine the cmb experiments the 21 cm line and the bao observations ,
we calculate the combined fisher matrix to be @xmath160 in this paper , we do not use information for a possible correlation between fluctuations of the 21 cm and the cmb . in figs.[fig
: nnu01_fsky0016]-[fig : nnu_noskafg_m8 ] , we plot contours of 95% confidence levels ( c.l . )
forecasts of each combination of cmb , 21 cm line and bao observations in @xmath148-@xmath13 plane .
constraints on cosmological parameters are summarized in tables [ tab : fsk0016_sigma_m_nu010 ] - [ tab : fsk065_fg_sigma_m_nu006_skanofg ] .
the upper panels of figs.[fig : nnu01_fsky0016]-[fig : nnu006_fsky065 ] show forecasts without residual foregrounds of cmb and 21 cm line ( i.e. they are completely removed ) , and the lower ones show those with given level residual foregrounds , respectively . for simons array
, we assume that its 220ghz band is used for the foreground removal when we include the contribution of residual foreground , and used for cmb observation when we do not include the residual foreground .
the fiducial value of the total neutrino mass is set to be @xmath151 ev ( figs.[fig : nnu01_fsky0016 ] , [ fig : nnu01_fsky02 ] and [ fig : nnu01_fsky065 ] ) or @xmath161 ev ( figs.[fig : nnu006_fsky0016 ] , [ fig : nnu006_fsky02 ] and [ fig : nnu006_fsky065 ] ) .
sky coverages of polarbear-2 and simons array are @xmath162 ( figs.[fig : nnu01_fsky0016 ] and [ fig : nnu006_fsky0016 ] ) , @xmath163 ( figs.[fig : nnu01_fsky02 ] and [ fig : nnu006_fsky02 ] ) , or @xmath164 ( figs.[fig : nnu01_fsky065 ] and [ fig : nnu006_fsky065 ] ) . in the left two panels of figs.[fig : nnu01_fsky0016]-[fig : nnu006_fsky065 ] , each contour represents constraints by cmb only or cmb + bao ( desi ) + hubble prior . in the right two panels of them ,
each contour represents constraints by planck only or cmb + bao ( desi ) + hubble prior + 21 cm line ( ska ) . in these six figures ,
we assume that the number of the observed fields by ska is four ( @xmath165 ) . from these figures , adding the bao experiments to the cmb ones , we see that there is a strong improvement on the sensitivities to @xmath148 and @xmath13 because several parameter degeneracies are broken by those combinations . besides , we find that the larger sky coverage improves the constraints more . however , it is difficult to detect the nonzero neutrino mass at 2@xmath166 level even by using the combination of simons array and desi . on the other hand ,
adding the 21 cm experiments ( ska phase 1 ) to the cmb observation , we see that there is a substantial improvement .
when we neglect the contribution from the residual foreground , the combination of ska phase 1 and simons array ( @xmath167 or @xmath168 ) has enough sensitivity to detect nonzero neutrino mass in the case of @xmath169 ev to be fiducial .
of course , cmb + ska phase 2 can obviously do the same job . for the effective number of neutrino species , its errors by the combination of cmb and bao with ska
are several times as small as those of only cmb + bao . in the case with a given level of residual foregrounds , planck + simons array + bao + ska phase 1 can not detect nonzero neutrino mass , and only a combination of ska phase 2 with simons array ( @xmath163 or @xmath170 ) can do it . in the case of @xmath171 ev to be fiducial ( this value corresponds to the lowest value for the normal hierarchy ) , it is difficult to detect the nonzero neutrino mass even by the combination with ska phase 2 . for the effective number of neutrino species , its errors in this case
are a few times as large as those in the case without the residual foreground .
hence , a stronger foreground removal of the 21 cm line observation is necessary if we want to obtain a stringent constraint on the effective number of neutrino species .
we can find the necessity of the foreground removal of the 21 cm line from figs.[fig : nnu_noskafg ] and [ fig : nnu_noskafg_m8 ] .
these figures show forecasts without the residual foreground of 21 cm line but with that of cmb ( i.e. the foreground of the 21 cm line is completely removed ) .
we assume that the number of observed fields of ska is @xmath165 in fig.[fig : nnu_noskafg ] and @xmath172 in fig.[fig : nnu_noskafg_m8 ] . in these cases , constraints of @xmath148 and @xmath13
are strongly improved in comparison with the lower panels of figs.[fig : nnu01_fsky0016]-[fig : nnu006_fsky065 ] , and the errors of the effective number of neutrino species are approximately @xmath173 or @xmath174 at 2@xmath166 by simons array + ska phase 1 or phase 2 , respectively . besides , from fig.[fig : nnu_noskafg ] , by planck + simons array ( @xmath163 or @xmath170 ) + bao + ska phase 1 and planck + simons array + bao + ska phase 2
, we can detect the nonzero neutrino mass of @xmath151 ev in this case .
additionally , from fig.[fig : nnu_noskafg_m8 ] , we find that more observed fields of 21 cm line observation ( @xmath175 ) are necessary if we want to detect the nonzero neutrino mass of @xmath161 ev by ska phase 2 . in fig.[fig
: nnu01_fsky02_simons_fg ] , we show two different cases about observations of simons array . at first , we assume that the 220 ghz band of simons array is used for the foreground removal ( inner solid line ) , and this assumption is used for the analysis of figs.[fig : nnu01_fsky0016]-[fig : nnu_noskafg_m8 ] .
secondary , the band is used for only the observation of cmb , and not used for the foreground removal , which is done by only planck ( outer thick dashed line ) .
this figure shows that the latter constraint is almost the same level as that of polarbear-2 .
the reason is that the strength of the residual foreground depends on only the planck sensitivity .
therefore , it is better to use the 220 ghz band of simon array for the foreground removal .
constraints on cosmological parameters in this situation are summarized in tables [ tab : fsk02_fg_sigma_m_nu010_simons ] and [ tab : fsk02_fg_sigma_m_nu006_simons ] . [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 + + pb-2 ( 0.016 ) & @xmath186 & @xmath187 & @xmath188 & @xmath189 & @xmath190 + + pb-2 + bao + @xmath131 & @xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath195 + + pb-2 + bao + @xmath131 + ska1 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 + + pb-2 + bao + @xmath131 + ska2 & @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath205 + + sa ( 0.016 ) & @xmath206 & @xmath207 & @xmath208 & @xmath209 & @xmath210 + + sa + bao + @xmath131 & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 + + sa + bao + @xmath131 + ska1 & @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 + + sa + bao + @xmath131 + ska2 & @xmath221 & @xmath222 & @xmath223 & @xmath224 & @xmath225 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath199 & @xmath230 & @xmath231 & @xmath232 + + pb-2 ( 0.016 ) & @xmath233 & @xmath234 & @xmath235 & @xmath236 + + pb-2 + bao + @xmath131 & @xmath237 & @xmath238 & @xmath239 & @xmath240 + + pb-2 + bao + @xmath131 + ska1 & @xmath241 & @xmath242 & @xmath243 & @xmath244 + + pb-2 + bao + @xmath131 + ska2 & @xmath245 & @xmath246 & @xmath247 & @xmath248 + + sa ( 0.016 ) & @xmath249 & @xmath250 & @xmath251 & @xmath252 + + sa + bao + @xmath131 & @xmath237 & @xmath253 & @xmath254 & @xmath231 + + sa + bao + @xmath131 + ska1 & @xmath255 & @xmath256 & @xmath257 & @xmath258 + + sa + bao + @xmath131 + ska2 & @xmath259 & @xmath260 & @xmath261 & @xmath262 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath263 & @xmath264 & @xmath265 & @xmath184 & @xmath185 + + pb-2 ( 0.016 ) & @xmath241 & @xmath187 & @xmath188 & @xmath189 & @xmath266 + + pb-2 + bao + @xmath131 & @xmath267 & @xmath192 & @xmath268 & @xmath194 & @xmath195 + + pb-2 + bao + @xmath131 + ska1 & @xmath269 & @xmath197 & @xmath270 & @xmath199 & @xmath200 + + pb-2 + bao + @xmath131 + ska2 & @xmath271 & @xmath222 & @xmath223 & @xmath224 & @xmath205 + + sa ( 0.016 ) & @xmath270 & @xmath207 & @xmath208 & @xmath272 & @xmath273 + + sa + bao + @xmath131 & @xmath274 & @xmath212 & @xmath275 & @xmath214 & @xmath215 + + sa + bao + @xmath131 + ska1 & @xmath276 & @xmath217 & @xmath198 & @xmath219 & @xmath220 + + sa + bao + @xmath131 + ska2 & @xmath277 & @xmath278 & @xmath279 & @xmath280 & @xmath225 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath281 & @xmath230 & @xmath240 & @xmath232 + + pb-2 ( 0.016 ) & @xmath282 & @xmath283 & @xmath284 & @xmath285 + + pb-2 + bao + @xmath131 & @xmath286 & @xmath238 & @xmath287 & @xmath240 + + pb-2 + bao + @xmath131 + ska1 & @xmath288 & @xmath268 & @xmath289 & @xmath290 + + pb-2 + bao + @xmath131 + ska2 & @xmath291 & @xmath292 & @xmath247 & @xmath293 + + sa ( 0.016 ) & @xmath219 & @xmath250 & @xmath251 & @xmath294 + + sa + bao + @xmath131 & @xmath295 & @xmath253 & @xmath296 & @xmath297 + + sa + bao + @xmath131 + ska1 & @xmath288 & @xmath298 & @xmath299 & @xmath300 + + sa + bao + @xmath131 + ska2 & @xmath301 & @xmath302 & @xmath261 & @xmath303 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath194 & @xmath304 & @xmath305 & @xmath306 & @xmath307 + + pb-2 ( 0.016 ) & @xmath308 & @xmath309 & @xmath310 & @xmath311 & @xmath312 + + pb-2 + bao + @xmath131 & @xmath313 & @xmath314 & @xmath315 & @xmath316 & @xmath195 + + pb-2 + bao + @xmath131 + ska1 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 + + pb-2 + bao + @xmath131 + ska2 & @xmath322 & @xmath323 & @xmath288 & @xmath324 & @xmath325 + + sa ( 0.016 ) & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath330 + + sa + bao + @xmath131 & @xmath331 & @xmath212 & @xmath332 & @xmath333 & @xmath334 + + sa + bao + @xmath131 + ska1 & @xmath335 & @xmath336 & @xmath337 & @xmath308 & @xmath338 + + sa + bao + @xmath131 + ska2 & @xmath339 & @xmath197 & @xmath340 & @xmath341 & @xmath342 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath343 & @xmath344 & @xmath345 & @xmath346 + + pb-2 ( 0.016 ) & @xmath219 & @xmath347 & @xmath348 & @xmath349 + + pb-2 + bao + @xmath131 & @xmath350 & @xmath351 & @xmath352 & @xmath353 + + pb-2 + bao + @xmath131 + ska1 & @xmath255 & @xmath354 & @xmath355 & @xmath356 + + pb-2 + bao + @xmath131 + ska2 & @xmath357 & @xmath358 & @xmath359 & @xmath360 + + sa ( 0.016 ) & @xmath361 & @xmath250 & @xmath251 & @xmath362 + + sa + bao + @xmath131 & @xmath363 & @xmath364 & @xmath296 & @xmath297 + + sa + bao + @xmath131 + ska1 & @xmath288 & @xmath365 & @xmath366 & @xmath367 + + sa + bao + @xmath131 + ska2 & @xmath368 & @xmath369 & @xmath370 & @xmath371 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath372 & @xmath373 & @xmath374 & @xmath306 & @xmath307 + + pb-2 ( 0.016 ) & @xmath375 & @xmath309 & @xmath376 & @xmath377 & @xmath312 + + pb-2 + bao + @xmath131 & @xmath378 & @xmath314 & @xmath379 & @xmath316 & @xmath195 + + pb-2 + bao + @xmath131 + ska1 & @xmath380 & @xmath318 & @xmath381 & @xmath382 & @xmath321 + + pb-2 + bao + @xmath131 + ska2 & @xmath383 & @xmath323 & @xmath384 & @xmath385 & @xmath386 + + sa ( 0.016 ) & @xmath368 & @xmath327 & @xmath208 & @xmath329 & @xmath387 + + sa + bao + @xmath131 & @xmath388 & @xmath212 & @xmath389 & @xmath333 & @xmath334 + + sa + bao + @xmath131 + ska1 & @xmath335 & @xmath336 & @xmath337 & @xmath390 & @xmath391 + + sa + bao + @xmath131 + ska2 & @xmath392 & @xmath217 & @xmath393 & @xmath394 & @xmath342 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath395 & @xmath396 & @xmath397 & @xmath346 + + pb-2 ( 0.016 ) & @xmath398 & @xmath399 & @xmath400 & @xmath401 + + pb-2 + bao + @xmath131 & @xmath363 & @xmath351 & @xmath402 & @xmath403 + + pb-2 + bao + @xmath131 + ska1 & @xmath324 & @xmath354 & @xmath404 & @xmath356 + + pb-2 + bao + @xmath131 + ska2 & @xmath405 & @xmath406 & @xmath407 & @xmath408 + + sa ( 0.016 ) & @xmath255 & @xmath351 & @xmath251 & @xmath409 + + sa + bao + @xmath131 & @xmath186 & @xmath364 & @xmath410 & @xmath297 + + sa + bao + @xmath131 + ska1 & @xmath384 & @xmath365 & @xmath411 & @xmath412 + + sa + bao + @xmath131 + ska2 & @xmath198 & @xmath413 & @xmath414 & @xmath415 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 + + pb-2 ( 0.2 ) & @xmath416 & @xmath417 & @xmath351 & @xmath418 & @xmath419 + + pb-2 + bao + @xmath131 & @xmath420 & @xmath421 & @xmath377 & @xmath422 & @xmath423 + + pb-2 + bao + @xmath131 + ska1 & @xmath424 & @xmath425 & @xmath245 & @xmath426 & @xmath427 + + pb-2 + bao + @xmath131 + ska2 & @xmath428 & @xmath429 & @xmath430 & @xmath431 & @xmath432 + + sa ( 0.2 ) & @xmath433 & @xmath434 & @xmath435 & @xmath436 & @xmath437 + + sa + bao + @xmath131 & @xmath280 & @xmath438 & @xmath439 & @xmath440 & @xmath441 + + sa + bao + @xmath131 + ska1 & @xmath442 & @xmath443 & @xmath301 & @xmath444 & @xmath445 + + sa + bao + @xmath131 + ska2 & @xmath446 & @xmath447 & @xmath448 & @xmath449 & @xmath450 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath199 & @xmath230 & @xmath231 & @xmath232 + + pb-2 ( 0.2 ) & @xmath324 & @xmath451 & @xmath452 & @xmath453 + + pb-2 + bao + @xmath131 & @xmath324 & @xmath275 & @xmath454 & @xmath455 + + pb-2 + bao + @xmath131 + ska1 & @xmath456 & @xmath457 & @xmath458 & @xmath459 + + pb-2 + bao + @xmath131 + ska2 & @xmath460 & @xmath461 & @xmath462 & @xmath463 + + sa ( 0.2 ) & @xmath384 & @xmath464 & @xmath465 & @xmath466 + + sa + bao + @xmath131 & @xmath384 & @xmath467 & @xmath468 & @xmath469 + + sa + bao + @xmath131 + ska1 & @xmath340 & @xmath470 & @xmath471 & @xmath472 + + sa + bao + @xmath131 + ska2 & @xmath394 & @xmath473 & @xmath474 & @xmath475 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath263 & @xmath264 & @xmath265 & @xmath184 & @xmath185 + + pb-2 ( 0.2 ) & @xmath476 & @xmath417 & @xmath351 & @xmath418 & @xmath419 + + pb-2 + bao + @xmath131 & @xmath477 & @xmath421 & @xmath478 & @xmath422 & @xmath423 + + pb-2 + bao + @xmath131 + ska1 & @xmath479 & @xmath425 & @xmath480 & @xmath481 & @xmath427 + + pb-2 + bao + @xmath131 + ska2 & @xmath482 & @xmath429 & @xmath448 & @xmath483 & @xmath432 + + sa ( 0.2 ) & @xmath484 & @xmath434 & @xmath435 & @xmath436 & @xmath215 + + sa + bao + @xmath131 & @xmath485 & @xmath438 & @xmath246 & @xmath486 & @xmath441 + + sa + bao + @xmath131 + ska1 & @xmath487 & @xmath443 & @xmath488 & @xmath444 & @xmath489 + + sa + bao + @xmath131 + ska2 & @xmath428 & @xmath447 & @xmath490 & @xmath380 & @xmath450 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath281 & @xmath230 & @xmath240 & @xmath232 + + pb-2 ( 0.2 ) & @xmath385 & @xmath491 & @xmath492 & @xmath493 + + pb-2 + bao + @xmath131 & @xmath385 & @xmath275 & @xmath494 & @xmath495 + + pb-2 + bao + @xmath131 + ska1 & @xmath496 & @xmath497 & @xmath498 & @xmath459 + + pb-2 + bao + @xmath131 + ska2 & @xmath499 & @xmath500 & @xmath501 & @xmath463 + + sa ( 0.2 ) & @xmath502 & @xmath503 & @xmath504 & @xmath505 + + sa + bao + @xmath131 & @xmath502 & @xmath467 & @xmath506 & @xmath469 + + sa + bao + @xmath131 + ska1 & @xmath393 & @xmath507 & @xmath508 & @xmath509 + + sa + bao + @xmath131 + ska2 & @xmath510 & @xmath511 & @xmath512 & @xmath513 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath194 & @xmath304 & @xmath305 & @xmath306 & @xmath307 + + pb-2 ( 0.2 ) & @xmath510 & @xmath514 & @xmath230 & @xmath515 & @xmath516 + + pb-2 + bao + @xmath131 & @xmath517 & @xmath518 & @xmath519 & @xmath363 & @xmath520 + + pb-2 + bao + @xmath131 + ska1 & @xmath521 & @xmath522 & @xmath523 & @xmath456 & @xmath524 + + pb-2 + bao + @xmath131 + ska2 & @xmath525 & @xmath526 & @xmath405 & @xmath527 & @xmath528 + + sa ( 0.2 ) & @xmath529 & @xmath530 & @xmath531 & @xmath532 & @xmath334 + + sa + bao + @xmath131 & @xmath224 & @xmath438 & @xmath497 & @xmath533 & @xmath534 + + sa + bao + @xmath131 + ska1 & @xmath535 & @xmath438 & @xmath511 & @xmath536 & @xmath537 + + sa + bao + @xmath131 + ska2 & @xmath538 & @xmath539 & @xmath488 & @xmath540 & @xmath541 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath343 & @xmath344 & @xmath345 & @xmath346 + + pb-2 ( 0.2 ) & @xmath324 & @xmath542 & @xmath543 & @xmath215 + + pb-2 + bao + @xmath131 & @xmath186 & @xmath544 & @xmath545 & @xmath546 + + pb-2 + bao + @xmath131 + ska1 & @xmath384 & @xmath464 & @xmath547 & @xmath548 + + pb-2 + bao + @xmath131 + ska2 & @xmath405 & @xmath549 & @xmath550 & @xmath551 + + sa ( 0.2 ) & @xmath393 & @xmath552 & @xmath553 & @xmath505 + + sa + bao + @xmath131 & @xmath393 & @xmath554 & @xmath555 & @xmath469 + + sa + bao + @xmath131 + ska1 & @xmath393 & @xmath556 & @xmath557 & @xmath558 + + sa + bao + @xmath131 + ska2 & @xmath206 & @xmath559 & @xmath560 & @xmath561 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath372 & @xmath373 & @xmath374 & @xmath306 & @xmath307 + + pb-2 ( 0.2 ) & @xmath394 & @xmath514 & @xmath562 & @xmath563 & @xmath516 + + pb-2 + bao + @xmath131 & @xmath564 & @xmath518 & @xmath565 & @xmath363 & @xmath520 + + pb-2 + bao + @xmath131 + ska1 & @xmath566 & @xmath522 & @xmath567 & @xmath456 & @xmath524 + + pb-2 + bao + @xmath131 + ska2 & @xmath568 & @xmath526 & @xmath198 & @xmath569 & @xmath570 + + sa ( 0.2 ) & @xmath571 & @xmath530 & @xmath531 & @xmath532 & @xmath572 + + sa + bao + @xmath131 & @xmath573 & @xmath438 & @xmath574 & @xmath575 & @xmath534 + + sa + bao + @xmath131 + ska1 & @xmath576 & @xmath438 & @xmath473 & @xmath536 & @xmath537 + + sa + bao + @xmath131 + ska2 & @xmath577 & @xmath539 & @xmath578 & @xmath564 & @xmath541 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath395 & @xmath396 & @xmath397 & @xmath346 + + pb-2 ( 0.2 ) & @xmath385 & @xmath542 & @xmath579 & @xmath580 + + pb-2 + bao + @xmath131 & @xmath385 & @xmath544 & @xmath581 & @xmath400 + + pb-2 + bao + @xmath131 + ska1 & @xmath502 & @xmath582 & @xmath583 & @xmath548 + + pb-2 + bao + @xmath131 + ska2 & @xmath198 & @xmath467 & @xmath584 & @xmath585 + + sa ( 0.2 ) & @xmath405 & @xmath552 & @xmath586 & @xmath587 + + sa + bao + @xmath131 & @xmath405 & @xmath588 & @xmath547 & @xmath589 + + sa + bao + @xmath131 + ska1 & @xmath270 & @xmath590 & @xmath591 & @xmath592 + + sa + bao + @xmath131 + ska2 & @xmath593 & @xmath439 & @xmath560 & @xmath594 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 + + pb-2 ( 0.65 ) & @xmath595 & @xmath596 & @xmath597 & @xmath598 & @xmath599 + + pb-2 + bao + @xmath131 & @xmath600 & @xmath601 & @xmath602 & @xmath603 & @xmath604 + + pb-2 + bao + @xmath131 + ska1 & @xmath605 & @xmath606 & @xmath499 & @xmath607 & @xmath453 + + pb-2 + bao + @xmath131 + ska2 & @xmath608 & @xmath609 & @xmath490 & @xmath610 & @xmath611 + + sa ( 0.65 ) & @xmath612 & @xmath613 & @xmath552 & @xmath614 & @xmath338 + + sa + bao + @xmath131 & @xmath615 & @xmath616 & @xmath617 & @xmath618 & @xmath619 + + sa + bao + @xmath131 + ska1 & @xmath620 & @xmath621 & @xmath622 & @xmath623 & @xmath205 + + sa + bao + @xmath131 + ska2 & @xmath624 & @xmath625 & @xmath626 & @xmath627 & @xmath587 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath199 & @xmath230 & @xmath231 & @xmath232 + + pb-2 ( 0.65 ) & @xmath393 & @xmath628 & @xmath629 & @xmath630 + + pb-2 + bao + @xmath131 & @xmath405 & @xmath631 & @xmath632 & @xmath235 + + pb-2 + bao + @xmath131 + ska1 & @xmath633 & @xmath634 & @xmath635 & @xmath636 + + pb-2 + bao + @xmath131 + ska2 & @xmath637 & @xmath638 & @xmath639 & @xmath640 + + sa ( 0.65 ) & @xmath270 & @xmath641 & @xmath583 & @xmath642 + + sa + bao + @xmath131 & @xmath218 & @xmath643 & @xmath644 & @xmath645 + + sa + bao + @xmath131 + ska1 & @xmath206 & @xmath646 & @xmath513 & @xmath647 + + sa + bao + @xmath131 + ska2 & @xmath648 & @xmath649 & @xmath650 & @xmath651 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath263 & @xmath264 & @xmath265 & @xmath184 & @xmath185 + + pb-2 ( 0.65 ) & @xmath652 & @xmath596 & @xmath653 & @xmath598 & @xmath599 + + pb-2 + bao + @xmath131 & @xmath490 & @xmath601 & @xmath381 & @xmath603 & @xmath604 + + pb-2 + bao + @xmath131 + ska1 & @xmath654 & @xmath606 & @xmath394 & @xmath607 & @xmath655 + + pb-2 + bao + @xmath131 + ska2 & @xmath656 & @xmath657 & @xmath658 & @xmath521 & @xmath611 + + sa ( 0.65 ) & @xmath490 & @xmath613 & @xmath552 & @xmath614 & @xmath604 + + sa + bao + @xmath131 & @xmath335 & @xmath616 & @xmath659 & @xmath618 & @xmath619 + + sa + bao + @xmath131 + ska1 & @xmath660 & @xmath661 & @xmath662 & @xmath623 & @xmath225 + + sa + bao + @xmath131 + ska2 & @xmath663 & @xmath664 & @xmath600 & @xmath665 & @xmath587 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath281 & @xmath230 & @xmath240 & @xmath232 + + pb-2 ( 0.65 ) & @xmath270 & @xmath666 & @xmath355 & @xmath630 + + pb-2 + bao + @xmath131 & @xmath218 & @xmath574 & @xmath667 & @xmath235 + + pb-2 + bao + @xmath131 + ska1 & @xmath206 & @xmath634 & @xmath668 & @xmath636 + + pb-2 + bao + @xmath131 + ska2 & @xmath669 & @xmath670 & @xmath671 & @xmath640 + + sa ( 0.65 ) & @xmath633 & @xmath457 & @xmath672 & @xmath673 + + sa + bao + @xmath131 & @xmath674 & @xmath675 & @xmath676 & @xmath677 + + sa + bao + @xmath131 + ska1 & @xmath593 & @xmath678 & @xmath679 & @xmath647 + + sa + bao + @xmath131 + ska2 & @xmath680 & @xmath681 & @xmath374 & @xmath651 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath194 & @xmath304 & @xmath305 & @xmath306 & @xmath307 + + pb-2 ( 0.65 ) & @xmath682 & @xmath683 & @xmath684 & @xmath395 & @xmath195 + + pb-2 + bao + @xmath131 & @xmath685 & @xmath686 & @xmath687 & @xmath688 & @xmath599 + + pb-2 + bao + @xmath131 + ska1 & @xmath689 & @xmath686 & @xmath690 & @xmath691 & @xmath692 + + pb-2 + bao + @xmath131 + ska2 & @xmath693 & @xmath694 & @xmath259 & @xmath695 & @xmath427 + + sa ( 0.65 ) & @xmath490 & @xmath694 & @xmath696 & @xmath697 & @xmath391 + + sa + bao + @xmath131 & @xmath698 & @xmath699 & @xmath700 & @xmath701 & @xmath692 + + sa + bao + @xmath131 + ska1 & @xmath702 & @xmath657 & @xmath703 & @xmath704 & @xmath342 + + sa + bao + @xmath131 + ska2 & @xmath705 & @xmath706 & @xmath422 & @xmath707 & @xmath205 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath343 & @xmath344 & @xmath345 & @xmath346 + + pb-2 ( 0.65 ) & @xmath384 & @xmath696 & @xmath708 & @xmath709 + + pb-2 + bao + @xmath131 & @xmath385 & @xmath710 & @xmath711 & @xmath712 + + pb-2 + bao + @xmath131 + ska1 & @xmath385 & @xmath713 & @xmath714 & @xmath715 + + pb-2 + bao + @xmath131 + ska2 & @xmath270 & @xmath272 & @xmath716 & @xmath717 + + sa ( 0.65 ) & @xmath674 & @xmath718 & @xmath719 & @xmath720 + + sa + bao + @xmath131 & @xmath721 & @xmath722 & @xmath723 & @xmath724 + + sa + bao + @xmath131 + ska1 & @xmath725 & @xmath726 & @xmath244 & @xmath727 + + sa + bao + @xmath131 + ska2 & @xmath578 & @xmath728 & @xmath729 & @xmath711 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath372 & @xmath373 & @xmath374 & @xmath306 & @xmath307 + + pb-2 ( 0.65 ) & @xmath730 & @xmath683 & @xmath731 & @xmath732 & @xmath195 + + pb-2 + bao + @xmath131 & @xmath595 & @xmath686 & @xmath733 & @xmath688 & @xmath599 + + pb-2 + bao + @xmath131 + ska1 & @xmath734 & @xmath686 & @xmath735 & @xmath691 & @xmath692 + + pb-2 + bao + @xmath131 + ska2 & @xmath693 & @xmath694 & @xmath488 & @xmath736 & @xmath445 + + sa ( 0.65 ) & @xmath223 & @xmath694 & @xmath696 & @xmath697 & @xmath391 + + sa + bao + @xmath131 & @xmath737 & @xmath738 & @xmath739 & @xmath701 & @xmath692 + + sa + bao + @xmath131 + ska1 & @xmath740 & @xmath741 & @xmath742 & @xmath704 & @xmath342 + + sa + bao + @xmath131 + ska2 & @xmath705 & @xmath743 & @xmath680 & @xmath744 & @xmath225 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath395 & @xmath396 & @xmath397 & @xmath346 + + pb-2 ( 0.65 ) & @xmath456 & @xmath696 & @xmath745 & @xmath746 + + pb-2 + bao + @xmath131 & @xmath456 & @xmath710 & @xmath747 & @xmath712 + + pb-2 + bao + @xmath131 + ska1 & @xmath340 & @xmath748 & @xmath749 & @xmath750 + + pb-2 + bao + @xmath131 + ska2 & @xmath633 & @xmath751 & @xmath752 & @xmath753 + + sa ( 0.65 ) & @xmath725 & @xmath754 & @xmath755 & @xmath756 + + sa + bao + @xmath131 & @xmath593 & @xmath602 & @xmath757 & @xmath548 + + sa + bao + @xmath131 + ska1 & @xmath480 & @xmath758 & @xmath759 & @xmath727 + + sa + bao + @xmath131 + ska2 & @xmath760 & @xmath761 & @xmath762 & @xmath763 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath313 & @xmath314 & @xmath315 & @xmath316 & @xmath195 + + ska1 ( @xmath764 ) & @xmath765 & @xmath766 & @xmath496 & @xmath767 & @xmath768 + + ska1 ( @xmath769 ) & @xmath770 & @xmath197 & @xmath680 & @xmath286 & @xmath220 + + ska2 ( @xmath764 ) & @xmath771 & @xmath772 & @xmath773 & @xmath774 & @xmath541 + + ska2 ( @xmath769 ) & @xmath775 & @xmath776 & @xmath777 & @xmath778 & @xmath205 + planck + sa(0.016 ) + bao + @xmath131 & @xmath331 & @xmath212 & @xmath332 & @xmath333 & @xmath334 + + ska1 ( @xmath764 ) & @xmath216 & @xmath217 & @xmath218 & @xmath249 & @xmath220 + + ska1 ( @xmath769 ) & @xmath779 & @xmath780 & @xmath781 & @xmath488 & @xmath342 + + ska2 ( @xmath764 ) & @xmath221 & @xmath222 & @xmath203 & @xmath224 & @xmath225 + + ska2 ( @xmath769 ) & @xmath782 & @xmath783 & @xmath784 & @xmath785 & @xmath786 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath350 & @xmath351 & @xmath352 & @xmath353 + + ska1 ( @xmath764 ) & @xmath241 & @xmath315 & @xmath749 & @xmath787 + + ska1 ( @xmath769 ) & @xmath186 & @xmath788 & @xmath789 & @xmath790 + + ska2 ( @xmath764 ) & @xmath721 & @xmath791 & @xmath792 & @xmath793 + + ska2 ( @xmath769 ) & @xmath341 & @xmath794 & @xmath795 & @xmath247 + planck + sa(0.016 ) + bao + @xmath131 & @xmath363 & @xmath364 & @xmath296 & @xmath297 + + ska1 ( @xmath764 ) & @xmath363 & @xmath796 & @xmath797 & @xmath798 + + ska1 ( @xmath769 ) & @xmath385 & @xmath799 & @xmath800 & @xmath560 + + ska2 ( @xmath764 ) & @xmath291 & @xmath801 & @xmath640 & @xmath262 + + ska2 ( @xmath769 ) & @xmath622 & @xmath802 & @xmath803 & @xmath804 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath378 & @xmath314 & @xmath379 & @xmath316 & @xmath195 + + ska1 ( @xmath764 ) & @xmath805 & @xmath766 & @xmath340 & @xmath806 & @xmath768 + + ska1 ( @xmath769 ) & @xmath807 & @xmath197 & @xmath422 & @xmath295 & @xmath220 + + ska2 ( @xmath764 ) & @xmath808 & @xmath683 & @xmath809 & @xmath810 & @xmath493 + + ska2 ( @xmath769 ) & @xmath811 & @xmath776 & @xmath784 & @xmath812 & @xmath205 + planck + sa(0.016 ) + bao + @xmath131 & @xmath388 & @xmath212 & @xmath389 & @xmath333 & @xmath334 + + ska1 ( @xmath764 ) & @xmath276 & @xmath217 & @xmath198 & @xmath282 & @xmath220 + + ska1 ( @xmath769 ) & @xmath813 & @xmath780 & @xmath814 & @xmath578 & @xmath342 + + ska2 ( @xmath764 ) & @xmath277 & @xmath278 & @xmath279 & @xmath280 & @xmath815 + + ska2 ( @xmath769 ) & @xmath816 & @xmath783 & @xmath817 & @xmath818 & @xmath786 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath363 & @xmath351 & @xmath402 & @xmath403 + + ska1 ( @xmath764 ) & @xmath288 & @xmath315 & @xmath819 & @xmath787 + + ska1 ( @xmath769 ) & @xmath502 & @xmath820 & @xmath821 & @xmath822 + + ska2 ( @xmath764 ) & @xmath480 & @xmath823 & @xmath797 & @xmath824 + + ska2 ( @xmath769 ) & @xmath394 & @xmath825 & @xmath826 & @xmath827 + planck + sa(0.016 ) + bao + @xmath131 & @xmath186 & @xmath364 & @xmath410 & @xmath297 + + ska1 ( @xmath764 ) & @xmath186 & @xmath828 & @xmath459 & @xmath258 + + ska1 ( @xmath769 ) & @xmath456 & @xmath829 & @xmath822 & @xmath830 + + ska2 ( @xmath764 ) & @xmath578 & @xmath831 & @xmath640 & @xmath262 + + ska2 ( @xmath769 ) & @xmath648 & @xmath832 & @xmath833 & @xmath261 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath517 & @xmath518 & @xmath519 & @xmath363 & @xmath520 + + ska1 ( @xmath764 ) & @xmath834 & @xmath835 & @xmath633 & @xmath836 & @xmath837 + + ska1 ( @xmath769 ) & @xmath838 & @xmath839 & @xmath814 & @xmath575 & @xmath528 + + ska2 ( @xmath764 ) & @xmath840 & @xmath841 & @xmath203 & @xmath842 & @xmath205 + + ska2 ( @xmath769 ) & @xmath843 & @xmath443 & @xmath844 & @xmath689 & @xmath432 + planck + sa(0.2 ) + bao + @xmath131 & @xmath224 & @xmath438 & @xmath497 & @xmath533 & @xmath534 + + ska1 ( @xmath764 ) & @xmath442 & @xmath539 & @xmath301 & @xmath845 & @xmath489 + + ska1 ( @xmath769 ) & @xmath846 & @xmath847 & @xmath575 & @xmath848 & @xmath493 + + ska2 ( @xmath764 ) & @xmath446 & @xmath447 & @xmath448 & @xmath849 & @xmath450 + + ska2 ( @xmath769 ) & @xmath850 & @xmath851 & @xmath852 & @xmath853 & @xmath854 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath186 & @xmath544 & @xmath545 & @xmath546 + + ska1 ( @xmath764 ) & @xmath288 & @xmath855 & @xmath856 & @xmath857 + + ska1 ( @xmath769 ) & @xmath385 & @xmath858 & @xmath243 & @xmath859 + + ska2 ( @xmath764 ) & @xmath245 & @xmath860 & @xmath861 & @xmath862 + + ska2 ( @xmath769 ) & @xmath510 & @xmath863 & @xmath864 & @xmath865 + planck + sa(0.2 ) + bao + @xmath131 & @xmath393 & @xmath554 & @xmath555 & @xmath469 + + ska1 ( @xmath764 ) & @xmath340 & @xmath866 & @xmath867 & @xmath868 + + ska1 ( @xmath769 ) & @xmath368 & @xmath869 & @xmath870 & @xmath871 + + ska2 ( @xmath764 ) & @xmath499 & @xmath728 & @xmath872 & @xmath475 + + ska2 ( @xmath769 ) & @xmath873 & @xmath874 & @xmath875 & @xmath803 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath564 & @xmath518 & @xmath565 & @xmath363 & @xmath520 + + ska1 ( @xmath764 ) & @xmath196 & @xmath835 & @xmath218 & @xmath876 & @xmath837 + + ska1 ( @xmath769 ) & @xmath877 & @xmath839 & @xmath878 & @xmath440 & @xmath570 + + ska2 ( @xmath764 ) & @xmath879 & @xmath841 & @xmath223 & @xmath809 & @xmath225 + + ska2 ( @xmath769 ) & @xmath880 & @xmath443 & @xmath817 & @xmath535 & @xmath881 + planck + sa(0.2 ) + bao + @xmath131 & @xmath573 & @xmath438 & @xmath574 & @xmath575 & @xmath534 + + ska1 ( @xmath764 ) & @xmath487 & @xmath539 & @xmath488 & @xmath845 & @xmath489 + + ska1 ( @xmath769 ) & @xmath882 & @xmath847 & @xmath575 & @xmath540 & @xmath883 + + ska2 ( @xmath764 ) & @xmath428 & @xmath447 & @xmath490 & @xmath884 & @xmath450 + + ska2 ( @xmath769 ) & @xmath885 & @xmath851 & @xmath886 & @xmath887 & @xmath854 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath385 & @xmath544 & @xmath581 & @xmath400 + + ska1 ( @xmath764 ) & @xmath384 & @xmath888 & @xmath889 & @xmath857 + + ska1 ( @xmath769 ) & @xmath340 & @xmath791 & @xmath890 & @xmath891 + + ska2 ( @xmath764 ) & @xmath488 & @xmath892 & @xmath861 & @xmath862 + + ska2 ( @xmath769 ) & @xmath637 & @xmath617 & @xmath893 & @xmath894 + planck + sa(0.2 ) + bao + @xmath131 & @xmath405 & @xmath588 & @xmath547 & @xmath589 + + ska1 ( @xmath764 ) & @xmath393 & @xmath895 & @xmath896 & @xmath472 + + ska1 ( @xmath769 ) & @xmath198 & @xmath897 & @xmath898 & @xmath899 + + ska2 ( @xmath764 ) & @xmath510 & @xmath900 & @xmath901 & @xmath475 + + ska2 ( @xmath769 ) & @xmath902 & @xmath903 & @xmath310 & @xmath803 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath685 & @xmath686 & @xmath687 & @xmath688 & @xmath599 + + ska1 ( @xmath764 ) & @xmath904 & @xmath694 & @xmath259 & @xmath416 & @xmath570 + + ska1 ( @xmath769 ) & @xmath882 & @xmath905 & @xmath906 & @xmath907 & @xmath445 + + ska2 ( @xmath764 ) & @xmath446 & @xmath908 & @xmath279 & @xmath380 & @xmath909 + + ska2 ( @xmath769 ) & @xmath850 & @xmath910 & @xmath784 & @xmath853 & @xmath911 + planck + sa(0.65 ) + bao + @xmath131 & @xmath698 & @xmath699 & @xmath700 & @xmath701 & @xmath692 + + ska1 ( @xmath764 ) & @xmath912 & @xmath913 & @xmath622 & @xmath477 & @xmath205 + + ska1 ( @xmath769 ) & @xmath914 & @xmath915 & @xmath916 & @xmath685 & @xmath815 + + ska2 ( @xmath764 ) & @xmath917 & @xmath918 & @xmath626 & @xmath812 & @xmath919 + + ska2 ( @xmath769 ) & @xmath920 & @xmath921 & @xmath886 & @xmath922 & @xmath403 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath385 & @xmath710 & @xmath711 & @xmath712 + + ska1 ( @xmath764 ) & @xmath324 & @xmath675 & @xmath509 & @xmath797 + + ska1 ( @xmath769 ) & @xmath502 & @xmath923 & @xmath924 & @xmath827 + + ska2 ( @xmath764 ) & @xmath291 & @xmath925 & @xmath926 & @xmath475 + + ska2 ( @xmath769 ) & @xmath622 & @xmath927 & @xmath803 & @xmath833 + planck + sa(0.65 ) + bao + @xmath131 & @xmath721 & @xmath722 & @xmath723 & @xmath724 + + ska1 ( @xmath764 ) & @xmath674 & @xmath928 & @xmath827 & @xmath924 + + ska1 ( @xmath769 ) & @xmath725 & @xmath316 & @xmath929 & @xmath930 + + ska2 ( @xmath764 ) & @xmath669 & @xmath931 & @xmath932 & @xmath864 + + ska2 ( @xmath769 ) & @xmath906 & @xmath933 & @xmath934 & @xmath374 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath595 & @xmath686 & @xmath733 & @xmath688 & @xmath599 + + ska1 ( @xmath764 ) & @xmath935 & @xmath694 & @xmath245 & @xmath416 & @xmath570 + + ska1 ( @xmath769 ) & @xmath936 & @xmath905 & @xmath937 & @xmath938 & @xmath489 + + ska2 ( @xmath764 ) & @xmath939 & @xmath940 & @xmath448 & @xmath317 & @xmath746 + + ska2 ( @xmath769 ) & @xmath941 & @xmath942 & @xmath852 & @xmath887 & @xmath611 + planck + sa(0.65 ) + bao + @xmath131 & @xmath737 & @xmath738 & @xmath739 & @xmath701 & @xmath692 + + ska1 ( @xmath764 ) & @xmath943 & @xmath944 & @xmath637 & @xmath945 & @xmath205 + + ska1 ( @xmath769 ) & @xmath946 & @xmath915 & @xmath916 & @xmath707 & @xmath909 + + ska2 ( @xmath764 ) & @xmath947 & @xmath918 & @xmath600 & @xmath627 & @xmath587 + + ska2 ( @xmath769 ) & @xmath948 & @xmath921 & @xmath949 & @xmath950 & @xmath403 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath456 & @xmath710 & @xmath747 & @xmath712 + + ska1 ( @xmath764 ) & @xmath385 & @xmath675 & @xmath792 & @xmath797 + + ska1 ( @xmath769 ) & @xmath496 & @xmath951 & @xmath952 & @xmath827 + + ska2 ( @xmath764 ) & @xmath578 & @xmath953 & @xmath954 & @xmath475 + + ska2 ( @xmath769 ) & @xmath955 & @xmath956 & @xmath803 & @xmath833 + planck + sa(0.65 ) + bao + @xmath131 & @xmath593 & @xmath602 & @xmath757 & @xmath548 + + ska1 ( @xmath764 ) & @xmath725 & @xmath957 & @xmath958 & @xmath924 + + ska1 ( @xmath769 ) & @xmath245 & @xmath703 & @xmath926 & @xmath930 + + ska2 ( @xmath764 ) & @xmath422 & @xmath181 & @xmath959 & @xmath864 + + ska2 ( @xmath769 ) & @xmath440 & @xmath960 & @xmath961 & @xmath374 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath194 & @xmath304 & @xmath305 & @xmath306 & @xmath307 + + pb-2 ( 0.2 ) & @xmath510 & @xmath514 & @xmath230 & @xmath515 & @xmath516 + + pb-2 + bao + @xmath131 & @xmath517 & @xmath518 & @xmath519 & @xmath363 & @xmath520 + + sa ( 1 ) ( 0.2 ) & @xmath648 & @xmath962 & @xmath963 & @xmath964 & @xmath965 + + sa ( 2 ) ( 0.2 ) & @xmath529 & @xmath530 & @xmath531 & @xmath532 & @xmath334 + + sa ( 1 ) + bao + @xmath131 & @xmath695 & @xmath966 & @xmath967 & @xmath456 & @xmath968 + + sa ( 2 ) + bao + @xmath131 & @xmath224 & @xmath438 & @xmath497 & @xmath533 & @xmath534 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath343 & @xmath344 & @xmath345 & @xmath346 + + pb-2 ( 0.2 ) & @xmath324 & @xmath542 & @xmath543 & @xmath215 + + pb-2 + bao + @xmath131 & @xmath186 & @xmath544 & @xmath545 & @xmath546 + + sa ( 1 ) ( 0.2 ) & @xmath186 & @xmath364 & @xmath367 & @xmath520 + + sa ( 2 ) ( 0.2 ) & @xmath393 & @xmath552 & @xmath553 & @xmath505 + + sa ( 1 ) + bao + @xmath131 & @xmath186 & @xmath451 & @xmath585 & @xmath969 + + sa ( 2 ) + bao + @xmath131 & @xmath393 & @xmath554 & @xmath555 & @xmath469 + [ 0.85 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck & @xmath372 & @xmath373 & @xmath374 & @xmath306 & @xmath307 + + pb-2 ( 0.2 ) & @xmath394 & @xmath514 & @xmath562 & @xmath563 & @xmath516 + + pb-2 + bao + @xmath131 & @xmath564 & @xmath518 & @xmath565 & @xmath363 & @xmath520 + + sa ( 1 ) ( 0.2 ) & @xmath637 & @xmath962 & @xmath970 & @xmath971 & @xmath965 + + sa ( 2 ) ( 0.2 ) & @xmath571 & @xmath530 & @xmath531 & @xmath532 & @xmath572 + + sa ( 1 ) + bao + @xmath131 & @xmath517 & @xmath966 & @xmath519 & @xmath456 & @xmath968 + + sa ( 2 ) + bao + @xmath131 & @xmath573 & @xmath438 & @xmath574 & @xmath575 & @xmath534 + + & @xmath226 & @xmath227 & @xmath228 & @xmath229 + planck & @xmath395 & @xmath396 & @xmath397 & @xmath346 + + pb-2 ( 0.2 ) & @xmath385 & @xmath542 & @xmath579 & @xmath580 + + pb-2 + bao + @xmath131 & @xmath385 & @xmath544 & @xmath581 & @xmath400 + + sa ( 1 ) ( 0.2 ) & @xmath385 & @xmath364 & @xmath251 & @xmath334 + + sa ( 2 ) ( 0.2 ) & @xmath405 & @xmath552 & @xmath586 & @xmath587 + + sa ( 1 ) + bao + @xmath131 & @xmath502 & @xmath451 & @xmath545 & @xmath969 + + sa ( 2 ) + bao + @xmath131 & @xmath405 & @xmath588 & @xmath547 & @xmath589 + next , we discuss whether we will be able to determine the neutrino mass hierarchies by using the future 21 cm line and the cmb observations .
constraints on cosmological parameters are summarized in tables [ tab : fsk0016_hien ] - [ tab : fsk065_fg_hiei_skanofg ] . in figs .
[ fig : hie_ellipse_fsky0016]-[fig : hie_ellipse_m8 ] , we plot @xmath972 errors of the parameter @xmath973 constrained by both the 21 cm line and the cmb observations in case of the inverted hierarchy to be fiducial ( left panels ) , and the normal hierarchy to be fiducial ( right panels ) . in these
figure , upper two panels represent results without residual foregrounds , i.e. they are completely removed . on the other hand ,
middle two panels show the results with given level residual foregrounds . here
, we assume that the amplitude of the residual foreground of the 21 cm line is 10 times as small as that of the previous section , and the 220 ghz band of simons array is used for the foreground removal .
lower two panels show the results with only the residual foreground of cmb , i.e. that of the 21 cm line is completely removed .
we also assume that the 220 ghz band of simons array is used for the foreground removal in these cases .
the number of observed field of ska is @xmath165 in figs.[fig : hie_ellipse_fsky0016]-[fig : hie_ellipse_fsky065 ] and @xmath172 in fig.[fig : hie_ellipse_m8 ] .
it is notable that the difference between @xmath155 of these two hierarchies becomes larger as the total mass @xmath148 becomes smaller .
therefore , @xmath155 is quite useful to distinguish a true mass hierarchy from the other
. allowed parameters on @xmath155 by neutrino oscillation experiments are plotted as two bands for the inverted and the normal hierarchies , respectively .
the thin solid lines inside the bands are the central values of the oscillation experiments . from these figures ,
we find that it is difficult to determine the neutrino mass hierarchy by the combination of cmb and bao with ska phase 1 . however , if we use ska phase 2 , by the combinations of them , we will be able to determine the neutrino mass hierarchy to be inverted or normal for @xmath974 ev or @xmath975 at 95 % c.l .
, respectively . in particular , the detectability becomes larger as the sky coverage of simons array and the number of observed field of ska become larger .
this improvement on the observed field of ska is more effective than that of the large integrated time per one field .
those results should be reasonable because a precise discrimination of the mass hierarchy itself may have no meaning if the masses are highly degenerate , i.e. , if @xmath976 ev @xmath977 .
once a clear signature @xmath978 were determined by observations or experiments , it should be obvious that the hierarchy must be normal without any ambiguities .
on the other hand , if the hierarchy were inverted , we could not determine it by using information of only @xmath148 .
however , it is remarkable that our method is quite useful because we can discriminate the hierarchy from the other even if the fiducial values were @xmath979 ev for both the normal and inverted cases .
this is clearly shown in these figures . in case that a fiducial value of @xmath148 is taken to be the lowest values in neutrino oscillation experiments ,
these figures indicate that by planck + simons array + bao + ska phase 2 , we can discriminate the inverted ( normal ) mass hierarchy from the normal ( inverted ) one . by using omniscope
, we can discriminate any mass hierarchies up to @xmath980 ev @xcite .
[ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath981 & @xmath982 & @xmath237 & @xmath263 & @xmath619 + + ska1 ( @xmath764 ) & @xmath983 & @xmath202 & @xmath488 & @xmath219 & @xmath200 + + ska1 ( @xmath769 ) & @xmath984 & @xmath985 & @xmath878 & @xmath760 & @xmath837 + + ska2 ( @xmath764 ) & @xmath986 & @xmath987 & @xmath279 & @xmath988 & @xmath205 + + ska2 ( @xmath769 ) & @xmath989 & @xmath990 & @xmath817 & @xmath991 & @xmath881 + planck + sa(0.016 ) + bao + @xmath131 & @xmath992 & @xmath993 & @xmath286 & @xmath994 & @xmath692 + + ska1 ( @xmath764 ) & @xmath995 & @xmath278 & @xmath301 & @xmath324 & @xmath220 + + ska1 ( @xmath769 ) & @xmath996 & @xmath841 & @xmath814 & @xmath662 & @xmath342 + + ska2 ( @xmath764 ) & @xmath997 & @xmath443 & @xmath279 & @xmath773 & @xmath225 + + ska2 ( @xmath769 ) & @xmath998 & @xmath999 & @xmath817 & @xmath844 & @xmath786 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath286 & @xmath1001 & @xmath1002 & @xmath1003 + + ska1 ( @xmath764 ) & @xmath324 & @xmath1004 & @xmath1005 & @xmath1006 + + ska1 ( @xmath769 ) & @xmath340 & @xmath319 & @xmath1007 & @xmath1008 + + ska2 ( @xmath764 ) & @xmath301 & @xmath1009 & @xmath827 & @xmath1010 + + ska2 ( @xmath769 ) & @xmath669 & @xmath1011 & @xmath1012 & @xmath1013 + planck + sa(0.016 ) + bao + @xmath131 & @xmath295 & @xmath1014 & @xmath1015 & @xmath1016 + + ska1 ( @xmath764 ) & @xmath324 & @xmath1017 & @xmath1018 & @xmath1019 + + ska1 ( @xmath769 ) & @xmath340 & @xmath1020 & @xmath1021 & @xmath1022 + + ska2 ( @xmath764 ) & @xmath341 & @xmath1023 & @xmath926 & @xmath1024 + + ska2 ( @xmath769 ) & @xmath680 & @xmath219 & @xmath1025 & @xmath1013 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath1026 & @xmath982 & @xmath1027 & @xmath418 & @xmath619 + + ska1 ( @xmath764 ) & @xmath1028 & @xmath202 & @xmath460 & @xmath219 & @xmath200 + + ska1 ( @xmath769 ) & @xmath1029 & @xmath421 & @xmath1030 & @xmath460 & @xmath386 + + ska2 ( @xmath764 ) & @xmath1031 & @xmath987 & @xmath279 & @xmath1032 & @xmath655 + + ska2 ( @xmath769 ) & @xmath1033 & @xmath990 & @xmath784 & @xmath689 & @xmath432 + planck + sa(0.016 ) + bao + @xmath131 & @xmath1034 & @xmath993 & @xmath1035 & @xmath214 & @xmath692 + + ska1 ( @xmath764 ) & @xmath1036 & @xmath278 & @xmath760 & @xmath324 & @xmath220 + + ska1 ( @xmath769 ) & @xmath1037 & @xmath596 & @xmath1038 & @xmath662 & @xmath837 + + ska2 ( @xmath764 ) & @xmath775 & @xmath443 & @xmath279 & @xmath988 & @xmath205 + + ska2 ( @xmath769 ) & @xmath1039 & @xmath1040 & @xmath784 & @xmath535 & @xmath450 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath286 & @xmath1001 & @xmath454 & @xmath1041 + + ska1 ( @xmath764 ) & @xmath324 & @xmath1042 & @xmath1043 & @xmath1044 + + ska1 ( @xmath769 ) & @xmath456 & @xmath381 & @xmath1045 & @xmath1046 + + ska2 ( @xmath764 ) & @xmath488 & @xmath1047 & @xmath640 & @xmath1048 + + ska2 ( @xmath769 ) & @xmath622 & @xmath1049 & @xmath1050 & @xmath1051 + planck + sa(0.016 ) + bao + @xmath131 & @xmath295 & @xmath1014 & @xmath1052 & @xmath1053 + + ska1 ( @xmath764 ) & @xmath324 & @xmath1017 & @xmath1054 & @xmath1044 + + ska1 ( @xmath769 ) & @xmath340 & @xmath1055 & @xmath800 & @xmath1046 + + ska2 ( @xmath764 ) & @xmath301 & @xmath1056 & @xmath1057 & @xmath1048 + + ska2 ( @xmath769 ) & @xmath955 & @xmath233 & @xmath474 & @xmath1051 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath1058 & @xmath1059 & @xmath233 & @xmath703 & @xmath619 + + ska1 ( @xmath764 ) & @xmath1060 & @xmath212 & @xmath405 & @xmath931 & @xmath537 + + ska1 ( @xmath769 ) & @xmath1061 & @xmath197 & @xmath460 & @xmath1062 & @xmath220 + + ska2 ( @xmath764 ) & @xmath1063 & @xmath522 & @xmath1064 & @xmath1065 & @xmath541 + + ska2 ( @xmath769 ) & @xmath880 & @xmath985 & @xmath1066 & @xmath430 & @xmath205 + planck + sa(0.016 ) + bao + @xmath131 & @xmath1067 & @xmath993 & @xmath398 & @xmath214 & @xmath537 + + ska1 ( @xmath764 ) & @xmath1068 & @xmath1069 & @xmath218 & @xmath1070 & @xmath325 + + ska1 ( @xmath769 ) & @xmath1071 & @xmath202 & @xmath836 & @xmath384 & @xmath837 + + ska2 ( @xmath764 ) & @xmath1072 & @xmath839 & @xmath1073 & @xmath704 & @xmath655 + + ska2 ( @xmath769 ) & @xmath920 & @xmath539 & @xmath698 & @xmath849 & @xmath432 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath288 & @xmath696 & @xmath1074 & @xmath1075 + + ska1 ( @xmath764 ) & @xmath385 & @xmath1076 & @xmath1077 & @xmath1078 + + ska1 ( @xmath769 ) & @xmath340 & @xmath1079 & @xmath1080 & @xmath1081 + + ska2 ( @xmath764 ) & @xmath480 & @xmath1082 & @xmath289 & @xmath1083 + + ska2 ( @xmath769 ) & @xmath394 & @xmath1084 & @xmath1085 & @xmath1086 + planck + sa(0.016 ) + bao + @xmath131 & @xmath186 & @xmath1087 & @xmath1088 & @xmath1089 + + ska1 ( @xmath764 ) & @xmath502 & @xmath1090 & @xmath889 & @xmath1091 + + ska1 ( @xmath769 ) & @xmath357 & @xmath1092 & @xmath798 & @xmath1093 + + ska2 ( @xmath764 ) & @xmath488 & @xmath1094 & @xmath867 & @xmath1095 + + ska2 ( @xmath769 ) & @xmath637 & @xmath1096 & @xmath894 & @xmath1097 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath1098 & @xmath1059 & @xmath199 & @xmath316 & @xmath619 + + ska1 ( @xmath764 ) & @xmath1099 & @xmath212 & @xmath198 & @xmath1100 & @xmath768 + + ska1 ( @xmath769 ) & @xmath1101 & @xmath197 & @xmath836 & @xmath436 & @xmath220 + + ska2 ( @xmath764 ) & @xmath1102 & @xmath518 & @xmath652 & @xmath603 & @xmath489 + + ska2 ( @xmath769 ) & @xmath885 & @xmath1103 & @xmath737 & @xmath1104 & @xmath655 + planck + sa(0.016 ) + bao + @xmath131 & @xmath1105 & @xmath993 & @xmath1035 & @xmath333 & @xmath537 + + ska1 ( @xmath764 ) & @xmath1106 & @xmath1107 & @xmath674 & @xmath1070 & @xmath325 + + ska1 ( @xmath769 ) & @xmath1108 & @xmath222 & @xmath598 & @xmath186 & @xmath837 + + ska2 ( @xmath764 ) & @xmath1109 & @xmath1110 & @xmath1073 & @xmath1111 & @xmath453 + + ska2 ( @xmath769 ) & @xmath1112 & @xmath539 & @xmath335 & @xmath431 & @xmath709 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath288 & @xmath696 & @xmath1113 & @xmath1114 + + ska1 ( @xmath764 ) & @xmath385 & @xmath1115 & @xmath1116 & @xmath1117 + + ska1 ( @xmath769 ) & @xmath340 & @xmath1118 & @xmath757 & @xmath1119 + + ska2 ( @xmath764 ) & @xmath593 & @xmath1120 & @xmath1121 & @xmath1122 + + ska2 ( @xmath769 ) & @xmath760 & @xmath1123 & @xmath1124 & @xmath1125 + planck + sa(0.016 ) + bao + @xmath131 & @xmath186 & @xmath1087 & @xmath1126 & @xmath1053 + + ska1 ( @xmath764 ) & @xmath502 & @xmath1127 & @xmath258 & @xmath1128 + + ska1 ( @xmath769 ) & @xmath496 & @xmath1129 & @xmath1045 & @xmath1130 + + ska2 ( @xmath764 ) & @xmath259 & @xmath1131 & @xmath1132 & @xmath969 + + ska2 ( @xmath769 ) & @xmath598 & @xmath1133 & @xmath1134 & @xmath1125 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath813 & @xmath841 & @xmath384 & @xmath422 & @xmath427 + + ska1 ( @xmath764 ) & @xmath1135 & @xmath429 & @xmath836 & @xmath1136 & @xmath489 + + ska1 ( @xmath769 ) & @xmath221 & @xmath1137 & @xmath937 & @xmath1138 & @xmath493 + + ska2 ( @xmath764 ) & @xmath1139 & @xmath1140 & @xmath448 & @xmath317 & @xmath881 + + ska2 ( @xmath769 ) & @xmath1141 & @xmath1142 & @xmath852 & @xmath1143 & @xmath1144 + planck + sa(0.2 ) + bao + @xmath131 & @xmath846 & @xmath438 & @xmath340 & @xmath440 & @xmath489 + + ska1 ( @xmath764 ) & @xmath1145 & @xmath447 & @xmath598 & @xmath476 & @xmath493 + + ska1 ( @xmath769 ) & @xmath428 & @xmath1146 & @xmath533 & @xmath907 & @xmath453 + + ska2 ( @xmath764 ) & @xmath1147 & @xmath1148 & @xmath490 & @xmath1149 & @xmath786 + + ska2 ( @xmath769 ) & @xmath1150 & @xmath1151 & @xmath886 & @xmath740 & @xmath1152 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath385 & @xmath319 & @xmath1153 & @xmath1154 + + ska1 ( @xmath764 ) & @xmath368 & @xmath1155 & @xmath1156 & @xmath1157 + + ska1 ( @xmath769 ) & @xmath633 & @xmath1158 & @xmath924 & @xmath1159 + + ska2 ( @xmath764 ) & @xmath510 & @xmath1160 & @xmath1161 & @xmath1162 + + ska2 ( @xmath769 ) & @xmath781 & @xmath245 & @xmath932 & @xmath1163 + planck + sa(0.2 ) + bao + @xmath131 & @xmath456 & @xmath1164 & @xmath1165 & @xmath1166 + + ska1 ( @xmath764 ) & @xmath270 & @xmath1167 & @xmath1168 & @xmath1169 + + ska1 ( @xmath769 ) & @xmath674 & @xmath1170 & @xmath859 & @xmath1171 + + ska2 ( @xmath764 ) & @xmath637 & @xmath249 & @xmath1172 & @xmath1173 + + ska2 ( @xmath769 ) & @xmath906 & @xmath1174 & @xmath310 & @xmath1175 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath1176 & @xmath841 & @xmath241 & @xmath955 & @xmath427 + + ska1 ( @xmath764 ) & @xmath1177 & @xmath429 & @xmath598 & @xmath1178 & @xmath489 + + ska1 ( @xmath769 ) & @xmath1179 & @xmath1137 & @xmath575 & @xmath1180 & @xmath493 + + ska2 ( @xmath764 ) & @xmath1181 & @xmath1140 & @xmath448 & @xmath449 & @xmath432 + + ska2 ( @xmath769 ) & @xmath1182 & @xmath1183 & @xmath817 & @xmath1184 & @xmath587 + planck + sa(0.2 ) + bao + @xmath131 & @xmath1185 & @xmath438 & @xmath324 & @xmath575 & @xmath489 + + ska1 ( @xmath764 ) & @xmath1186 & @xmath905 & @xmath662 & @xmath476 & @xmath493 + + ska1 ( @xmath769 ) & @xmath1187 & @xmath1188 & @xmath313 & @xmath517 & @xmath453 + + ska2 ( @xmath764 ) & @xmath1189 & @xmath1190 & @xmath490 & @xmath1191 & @xmath450 + + ska2 ( @xmath769 ) & @xmath1192 & @xmath1193 & @xmath852 & @xmath1194 & @xmath854 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath385 & @xmath319 & @xmath1195 & @xmath1196 + + ska1 ( @xmath764 ) & @xmath393 & @xmath1197 & @xmath800 & @xmath1198 + + ska1 ( @xmath769 ) & @xmath218 & @xmath1199 & @xmath859 & @xmath1200 + + ska2 ( @xmath764 ) & @xmath836 & @xmath1201 & @xmath833 & @xmath495 + + ska2 ( @xmath769 ) & @xmath878 & @xmath593 & @xmath1202 & @xmath1203 + planck + sa(0.2 ) + bao + @xmath131 & @xmath456 & @xmath461 & @xmath644 & @xmath1204 + + ska1 ( @xmath764 ) & @xmath405 & @xmath1205 & @xmath1206 & @xmath1207 + + ska1 ( @xmath769 ) & @xmath326 & @xmath1208 & @xmath247 & @xmath346 + + ska2 ( @xmath764 ) & @xmath598 & @xmath1209 & @xmath1210 & @xmath495 + + ska2 ( @xmath769 ) & @xmath1211 & @xmath1212 & @xmath1213 & @xmath1214 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath1215 & @xmath966 & @xmath255 & @xmath363 & @xmath837 + + ska1 ( @xmath764 ) & @xmath1216 & @xmath1217 & @xmath206 & @xmath480 & @xmath528 + + ska1 ( @xmath769 ) & @xmath1218 & @xmath783 & @xmath510 & @xmath669 & @xmath570 + + ska2 ( @xmath764 ) & @xmath1219 & @xmath1220 & @xmath774 & @xmath1221 & @xmath205 + + ska2 ( @xmath769 ) & @xmath1222 & @xmath694 & @xmath335 & @xmath1223 & @xmath709 + planck + sa(0.2 ) + bao + @xmath131 & @xmath1224 & @xmath438 & @xmath456 & @xmath533 & @xmath883 + + ska1 ( @xmath764 ) & @xmath1036 & @xmath686 & @xmath291 & @xmath1136 & @xmath453 + + ska1 ( @xmath769 ) & @xmath1225 & @xmath613 & @xmath648 & @xmath476 & @xmath655 + + ska2 ( @xmath764 ) & @xmath1226 & @xmath1227 & @xmath1228 & @xmath1032 & @xmath881 + + ska2 ( @xmath769 ) & @xmath1229 & @xmath942 & @xmath615 & @xmath1230 & @xmath587 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath385 & @xmath1231 & @xmath1232 & @xmath1233 + + ska1 ( @xmath764 ) & @xmath340 & @xmath1234 & @xmath1235 & @xmath1236 + + ska1 ( @xmath769 ) & @xmath393 & @xmath1237 & @xmath889 & @xmath1238 + + ska2 ( @xmath764 ) & @xmath291 & @xmath1239 & @xmath1168 & @xmath1240 + + ska2 ( @xmath769 ) & @xmath598 & @xmath1241 & @xmath827 & @xmath1242 + planck + sa(0.2 ) + bao + @xmath131 & @xmath270 & @xmath1243 & @xmath1244 & @xmath1245 + + ska1 ( @xmath764 ) & @xmath633 & @xmath728 & @xmath924 & @xmath1246 + + ska1 ( @xmath769 ) & @xmath206 & @xmath1247 & @xmath458 & @xmath1248 + + ska2 ( @xmath764 ) & @xmath836 & @xmath1249 & @xmath1250 & @xmath1251 + + ska2 ( @xmath769 ) & @xmath1174 & @xmath219 & @xmath1252 & @xmath1253 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath1254 & @xmath966 & @xmath219 & @xmath255 & @xmath837 + + ska1 ( @xmath764 ) & @xmath1255 & @xmath1217 & @xmath725 & @xmath593 & @xmath528 + + ska1 ( @xmath769 ) & @xmath1256 & @xmath783 & @xmath637 & @xmath662 & @xmath570 + + ska2 ( @xmath764 ) & @xmath1257 & @xmath1220 & @xmath1073 & @xmath1258 & @xmath655 + + ska2 ( @xmath769 ) & @xmath1259 & @xmath694 & @xmath1066 & @xmath449 & @xmath746 + planck + sa(0.2 ) + bao + @xmath131 & @xmath1260 & @xmath438 & @xmath288 & @xmath378 & @xmath493 + + ska1 ( @xmath764 ) & @xmath1261 & @xmath686 & @xmath301 & @xmath1136 & @xmath883 + + ska1 ( @xmath769 ) & @xmath1262 & @xmath613 & @xmath680 & @xmath191 & @xmath453 + + ska2 ( @xmath764 ) & @xmath1263 & @xmath1227 & @xmath1228 & @xmath1264 & @xmath432 + + ska2 ( @xmath769 ) & @xmath1265 & @xmath942 & @xmath1266 & @xmath1267 & @xmath919 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath385 & @xmath1268 & @xmath1269 & @xmath1270 + + ska1 ( @xmath764 ) & @xmath340 & @xmath209 & @xmath1271 & @xmath1272 + + ska1 ( @xmath769 ) & @xmath393 & @xmath1273 & @xmath300 & @xmath1274 + + ska2 ( @xmath764 ) & @xmath259 & @xmath1158 & @xmath859 & @xmath969 + + ska2 ( @xmath769 ) & @xmath876 & @xmath1249 & @xmath1275 & @xmath1125 + planck + sa(0.2 ) + bao + @xmath131 & @xmath405 & @xmath1243 & @xmath1276 & @xmath1277 + + ska1 ( @xmath764 ) & @xmath633 & @xmath1278 & @xmath824 & @xmath1279 + + ska1 ( @xmath769 ) & @xmath674 & @xmath1280 & @xmath762 & @xmath1281 + + ska2 ( @xmath764 ) & @xmath394 & @xmath1282 & @xmath1283 & @xmath348 + + ska2 ( @xmath769 ) & @xmath680 & @xmath233 & @xmath1284 & @xmath1125 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath1285 & @xmath1140 & @xmath206 & @xmath603 & @xmath205 + + ska1 ( @xmath764 ) & @xmath1286 & @xmath1287 & @xmath422 & @xmath1111 & @xmath815 + + ska1 ( @xmath769 ) & @xmath1288 & @xmath1289 & @xmath527 & @xmath477 & @xmath909 + + ska2 ( @xmath764 ) & @xmath1259 & @xmath1290 & @xmath658 & @xmath778 & @xmath919 + + ska2 ( @xmath769 ) & @xmath1291 & @xmath1292 & @xmath886 & @xmath922 & @xmath1293 + planck + sa(0.65 ) + bao + @xmath131 & @xmath1294 & @xmath616 & @xmath291 & @xmath618 & @xmath746 + + ska1 ( @xmath764 ) & @xmath201 & @xmath1295 & @xmath873 & @xmath420 & @xmath709 + + ska1 ( @xmath769 ) & @xmath1296 & @xmath921 & @xmath916 & @xmath204 & @xmath881 + + ska2 ( @xmath764 ) & @xmath1297 & @xmath1298 & @xmath600 & @xmath818 & @xmath505 + + ska2 ( @xmath769 ) & @xmath1299 & @xmath1300 & @xmath949 & @xmath1301 & @xmath1302 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath633 & @xmath646 & @xmath1303 & @xmath1304 + + ska1 ( @xmath764 ) & @xmath593 & @xmath1305 & @xmath1085 & @xmath1306 + + ska1 ( @xmath769 ) & @xmath291 & @xmath767 & @xmath303 & @xmath1307 + + ska2 ( @xmath764 ) & @xmath422 & @xmath326 & @xmath1012 & @xmath1308 + + ska2 ( @xmath769 ) & @xmath575 & @xmath313 & @xmath1309 & @xmath1310 + planck + sa(0.65 ) + bao + @xmath131 & @xmath725 & @xmath1311 & @xmath891 & @xmath1312 + + ska1 ( @xmath764 ) & @xmath291 & @xmath1313 & @xmath795 & @xmath1314 + + ska1 ( @xmath769 ) & @xmath460 & @xmath219 & @xmath958 & @xmath1128 + + ska2 ( @xmath764 ) & @xmath1315 & @xmath1212 & @xmath1316 & @xmath1317 + + ska2 ( @xmath769 ) & @xmath1318 & @xmath476 & @xmath961 & @xmath1319 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath1320 & @xmath601 & @xmath218 & @xmath603 & @xmath205 + + ska1 ( @xmath764 ) & @xmath1321 & @xmath1287 & @xmath1212 & @xmath1322 & @xmath815 + + ska1 ( @xmath769 ) & @xmath1296 & @xmath1289 & @xmath1323 & @xmath1324 & @xmath909 + + ska2 ( @xmath764 ) & @xmath1325 & @xmath1290 & @xmath658 & @xmath1267 & @xmath1326 + + ska2 ( @xmath769 ) & @xmath1299 & @xmath1292 & @xmath852 & @xmath1327 & @xmath466 + planck + sa(0.65 ) + bao + @xmath131 & @xmath1328 & @xmath1329 & @xmath593 & @xmath618 & @xmath909 + + ska1 ( @xmath764 ) & @xmath221 & @xmath1295 & @xmath781 & @xmath420 & @xmath709 + + ska1 ( @xmath769 ) & @xmath1330 & @xmath921 & @xmath614 & @xmath573 & @xmath432 + + ska2 ( @xmath764 ) & @xmath1331 & @xmath1298 & @xmath600 & @xmath785 & @xmath1144 + + ska2 ( @xmath769 ) & @xmath1332 & @xmath1333 & @xmath949 & @xmath1334 & @xmath353 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath633 & @xmath646 & @xmath1335 & @xmath1336 + + ska1 ( @xmath764 ) & @xmath725 & @xmath1337 & @xmath1338 & @xmath1339 + + ska1 ( @xmath769 ) & @xmath259 & @xmath390 & @xmath1340 & @xmath232 + + ska2 ( @xmath764 ) & @xmath648 & @xmath198 & @xmath1341 & @xmath495 + + ska2 ( @xmath769 ) & @xmath378 & @xmath440 & @xmath376 & @xmath1342 + planck + sa(0.65 ) + bao + @xmath131 & @xmath721 & @xmath181 & @xmath300 & @xmath1343 + + ska1 ( @xmath764 ) & @xmath291 & @xmath960 & @xmath896 & @xmath1344 + + ska1 ( @xmath769 ) & @xmath341 & @xmath282 & @xmath1345 & @xmath1346 + + ska2 ( @xmath764 ) & @xmath1212 & @xmath680 & @xmath1347 & @xmath455 + + ska2 ( @xmath769 ) & @xmath536 & @xmath1348 & @xmath1349 & @xmath1350 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath1061 & @xmath847 & @xmath405 & @xmath688 & @xmath489 + + ska1 ( @xmath764 ) & @xmath1351 & @xmath613 & @xmath341 & @xmath1352 & @xmath541 + + ska1 ( @xmath769 ) & @xmath1353 & @xmath1354 & @xmath422 & @xmath1180 & @xmath493 + + ska2 ( @xmath764 ) & @xmath1355 & @xmath1356 & @xmath1228 & @xmath1357 & @xmath909 + + ska2 ( @xmath769 ) & @xmath1358 & @xmath1359 & @xmath1266 & @xmath1230 & @xmath786 + planck + sa(0.65 ) + bao + @xmath131 & @xmath1360 & @xmath738 & @xmath291 & @xmath701 & @xmath709 + + ska1 ( @xmath764 ) & @xmath1361 & @xmath1362 & @xmath622 & @xmath1363 & @xmath432 + + ska1 ( @xmath769 ) & @xmath1364 & @xmath1365 & @xmath1030 & @xmath623 & @xmath881 + + ska2 ( @xmath764 ) & @xmath1366 & @xmath625 & @xmath744 & @xmath1367 & @xmath587 + + ska2 ( @xmath769 ) & @xmath1368 & @xmath1369 & @xmath566 & @xmath784 & @xmath1293 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath340 & @xmath1370 & @xmath1371 & @xmath1372 + + ska1 ( @xmath764 ) & @xmath393 & @xmath1373 & @xmath1153 & @xmath1374 + + ska1 ( @xmath769 ) & @xmath405 & @xmath1239 & @xmath1375 & @xmath1376 + + ska2 ( @xmath764 ) & @xmath488 & @xmath1377 & @xmath1378 & @xmath1379 + + ska2 ( @xmath769 ) & @xmath637 & @xmath286 & @xmath1380 & @xmath1381 + planck + sa(0.65 ) + bao + @xmath131 & @xmath259 & @xmath194 & @xmath1382 & @xmath1383 + + ska1 ( @xmath764 ) & @xmath301 & @xmath375 & @xmath762 & @xmath1384 + + ska1 ( @xmath769 ) & @xmath341 & @xmath1201 & @xmath1385 & @xmath1386 + + ska2 ( @xmath764 ) & @xmath648 & @xmath496 & @xmath1387 & @xmath1388 + + ska2 ( @xmath769 ) & @xmath1038 & @xmath1174 & @xmath1389 & @xmath1390 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath1391 & @xmath686 & @xmath456 & @xmath688 & @xmath489 + + ska1 ( @xmath764 ) & @xmath1351 & @xmath613 & @xmath460 & @xmath444 & @xmath541 + + ska1 ( @xmath769 ) & @xmath1392 & @xmath1354 & @xmath1212 & @xmath1393 & @xmath541 + + ska2 ( @xmath764 ) & @xmath1394 & @xmath1395 & @xmath1228 & @xmath1032 & @xmath909 + + ska2 ( @xmath769 ) & @xmath1396 & @xmath1397 & @xmath698 & @xmath1398 & @xmath450 + planck + sa(0.65 ) + bao + @xmath131 & @xmath1399 & @xmath699 & @xmath593 & @xmath701 & @xmath709 + + ska1 ( @xmath764 ) & @xmath1400 & @xmath1362 & @xmath669 & @xmath1401 & @xmath432 + + ska1 ( @xmath769 ) & @xmath482 & @xmath1365 & @xmath937 & @xmath623 & @xmath432 + + ska2 ( @xmath764 ) & @xmath1402 & @xmath625 & @xmath744 & @xmath380 & @xmath919 + + ska2 ( @xmath769 ) & @xmath1403 & @xmath1404 & @xmath610 & @xmath991 & @xmath1152 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath340 & @xmath1370 & @xmath1405 & @xmath1406 + + ska1 ( @xmath764 ) & @xmath357 & @xmath1407 & @xmath1408 & @xmath1279 + + ska1 ( @xmath769 ) & @xmath405 & @xmath1409 & @xmath1244 & @xmath1281 + + ska2 ( @xmath764 ) & @xmath291 & @xmath333 & @xmath1410 & @xmath348 + + ska2 ( @xmath769 ) & @xmath1411 & @xmath219 & @xmath462 & @xmath1125 + planck + sa(0.65 ) + bao + @xmath131 & @xmath245 & @xmath1412 & @xmath1413 & @xmath1414 + + ska1 ( @xmath764 ) & @xmath488 & @xmath1415 & @xmath262 & @xmath1416 + + ska1 ( @xmath769 ) & @xmath578 & @xmath1417 & @xmath1418 & @xmath1419 + + ska2 ( @xmath764 ) & @xmath622 & @xmath502 & @xmath1252 & @xmath1420 + + ska2 ( @xmath769 ) & @xmath902 & @xmath680 & @xmath1421 & @xmath548 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath1058 & @xmath1059 & @xmath233 & @xmath703 & @xmath619 + + ska1 ( @xmath764 ) & @xmath1422 & @xmath1423 & @xmath259 & @xmath436 & @xmath768 + + ska1 ( @xmath769 ) & @xmath1262 & @xmath222 & @xmath1212 & @xmath288 & @xmath220 + + ska2 ( @xmath764 ) & @xmath1424 & @xmath985 & @xmath223 & @xmath204 & @xmath493 + + ska2 ( @xmath769 ) & @xmath1425 & @xmath1354 & @xmath784 & @xmath1426 & @xmath225 + planck + sa(0.016 ) + bao + @xmath131 & @xmath1067 & @xmath993 & @xmath398 & @xmath214 & @xmath537 + + ska1 ( @xmath764 ) & @xmath1101 & @xmath278 & @xmath301 & @xmath288 & @xmath325 + + ska1 ( @xmath769 ) & @xmath1427 & @xmath841 & @xmath814 & @xmath637 & @xmath342 + + ska2 ( @xmath764 ) & @xmath997 & @xmath443 & @xmath279 & @xmath1104 & @xmath225 + + ska2 ( @xmath769 ) & @xmath998 & @xmath1428 & @xmath817 & @xmath844 & @xmath786 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath288 & @xmath696 & @xmath1074 & @xmath1075 + + ska1 ( @xmath764 ) & @xmath288 & @xmath1429 & @xmath1080 & @xmath1430 + + ska1 ( @xmath769 ) & @xmath456 & @xmath292 & @xmath1431 & @xmath1432 + + ska2 ( @xmath764 ) & @xmath480 & @xmath957 & @xmath952 & @xmath1433 + + ska2 ( @xmath769 ) & @xmath394 & @xmath767 & @xmath1434 & @xmath1435 + planck + sa(0.016 ) + bao + @xmath131 & @xmath186 & @xmath1087 & @xmath1088 & @xmath1089 + + ska1 ( @xmath764 ) & @xmath186 & @xmath1092 & @xmath244 & @xmath1019 + + ska1 ( @xmath769 ) & @xmath496 & @xmath1436 & @xmath1408 & @xmath1022 + + ska2 ( @xmath764 ) & @xmath460 & @xmath1437 & @xmath1438 & @xmath1024 + + ska2 ( @xmath769 ) & @xmath680 & @xmath233 & @xmath1439 & @xmath1013 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath1098 & @xmath1059 & @xmath199 & @xmath316 & @xmath619 + + ska1 ( @xmath764 ) & @xmath983 & @xmath1423 & @xmath301 & @xmath1440 & @xmath768 + + ska1 ( @xmath769 ) & @xmath1441 & @xmath222 & @xmath878 & @xmath241 & @xmath220 + + ska2 ( @xmath764 ) & @xmath1442 & @xmath985 & @xmath203 & @xmath685 & @xmath541 + + ska2 ( @xmath769 ) & @xmath989 & @xmath1354 & @xmath777 & @xmath627 & @xmath205 + planck + sa(0.016 ) + bao + @xmath131 & @xmath1105 & @xmath993 & @xmath1035 & @xmath333 & @xmath537 + + ska1 ( @xmath764 ) & @xmath1261 & @xmath278 & @xmath760 & @xmath288 & @xmath220 + + ska1 ( @xmath769 ) & @xmath1037 & @xmath841 & @xmath1211 & @xmath1411 & @xmath837 + + ska2 ( @xmath764 ) & @xmath775 & @xmath443 & @xmath279 & @xmath1357 & @xmath205 + + ska2 ( @xmath769 ) & @xmath1039 & @xmath999 & @xmath784 & @xmath535 & @xmath450 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.016 ) + bao + @xmath131 & @xmath288 & @xmath696 & @xmath1113 & @xmath1114 + + ska1 ( @xmath764 ) & @xmath288 & @xmath1443 & @xmath1444 & @xmath1445 + + ska1 ( @xmath769 ) & @xmath502 & @xmath1446 & @xmath1447 & @xmath1448 + + ska2 ( @xmath764 ) & @xmath593 & @xmath1409 & @xmath1449 & @xmath1048 + + ska2 ( @xmath769 ) & @xmath760 & @xmath1450 & @xmath1410 & @xmath1451 + planck + sa(0.016 ) + bao + @xmath131 & @xmath186 & @xmath1087 & @xmath1126 & @xmath1053 + + ska1 ( @xmath764 ) & @xmath186 & @xmath1092 & @xmath716 & @xmath1044 + + ska1 ( @xmath769 ) & @xmath496 & @xmath1452 & @xmath952 & @xmath1046 + + ska2 ( @xmath764 ) & @xmath578 & @xmath1453 & @xmath1454 & @xmath1048 + + ska2 ( @xmath769 ) & @xmath955 & @xmath1455 & @xmath1456 & @xmath1051 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath1215 & @xmath966 & @xmath255 & @xmath363 & @xmath837 + + ska1 ( @xmath764 ) & @xmath1457 & @xmath1103 & @xmath460 & @xmath662 & @xmath342 + + ska1 ( @xmath769 ) & @xmath1458 & @xmath987 & @xmath814 & @xmath569 & @xmath570 + + ska2 ( @xmath764 ) & @xmath1459 & @xmath613 & @xmath279 & @xmath490 & @xmath225 + + ska2 ( @xmath769 ) & @xmath1460 & @xmath1461 & @xmath817 & @xmath817 & @xmath432 + planck + sa(0.2 ) + bao + @xmath131 & @xmath1224 & @xmath438 & @xmath456 & @xmath533 & @xmath883 + + ska1 ( @xmath764 ) & @xmath1462 & @xmath447 & @xmath598 & @xmath1348 & @xmath493 + + ska1 ( @xmath769 ) & @xmath939 & @xmath1463 & @xmath378 & @xmath517 & @xmath453 + + ska2 ( @xmath764 ) & @xmath1147 & @xmath1464 & @xmath490 & @xmath1149 & @xmath786 + + ska2 ( @xmath769 ) & @xmath1465 & @xmath1466 & @xmath886 & @xmath1467 & @xmath1152 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath385 & @xmath1231 & @xmath1232 & @xmath1233 + + ska1 ( @xmath764 ) & @xmath385 & @xmath1237 & @xmath1468 & @xmath1469 + + ska1 ( @xmath769 ) & @xmath357 & @xmath1470 & @xmath1471 & @xmath1472 + + ska2 ( @xmath764 ) & @xmath301 & @xmath931 & @xmath508 & @xmath1473 + + ska2 ( @xmath769 ) & @xmath622 & @xmath295 & @xmath1250 & @xmath1346 + planck + sa(0.2 ) + bao + @xmath131 & @xmath270 & @xmath1243 & @xmath1244 & @xmath1245 + + ska1 ( @xmath764 ) & @xmath270 & @xmath1474 & @xmath498 & @xmath1475 + + ska1 ( @xmath769 ) & @xmath674 & @xmath1476 & @xmath1382 & @xmath1477 + + ska2 ( @xmath764 ) & @xmath637 & @xmath1035 & @xmath1439 & @xmath1119 + + ska2 ( @xmath769 ) & @xmath906 & @xmath680 & @xmath376 & @xmath1175 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath1254 & @xmath966 & @xmath219 & @xmath255 & @xmath837 + + ska1 ( @xmath764 ) & @xmath1351 & @xmath1103 & @xmath499 & @xmath637 & @xmath837 + + ska1 ( @xmath769 ) & @xmath482 & @xmath987 & @xmath1211 & @xmath527 & @xmath570 + + ska2 ( @xmath764 ) & @xmath885 & @xmath694 & @xmath279 & @xmath223 & @xmath205 + + ska2 ( @xmath769 ) & @xmath1478 & @xmath1479 & @xmath844 & @xmath991 & @xmath432 + planck + sa(0.2 ) + bao + @xmath131 & @xmath1260 & @xmath438 & @xmath288 & @xmath378 & @xmath493 + + ska1 ( @xmath764 ) & @xmath1225 & @xmath447 & @xmath662 & @xmath1348 & @xmath493 + + ska1 ( @xmath769 ) & @xmath1187 & @xmath1480 & @xmath313 & @xmath1481 & @xmath453 + + ska2 ( @xmath764 ) & @xmath1482 & @xmath1483 & @xmath490 & @xmath1484 & @xmath450 + + ska2 ( @xmath769 ) & @xmath1192 & @xmath1485 & @xmath852 & @xmath1194 & @xmath854 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.2 ) + bao + @xmath131 & @xmath385 & @xmath1268 & @xmath1269 & @xmath1270 + + ska1 ( @xmath764 ) & @xmath385 & @xmath892 & @xmath1486 & @xmath1379 + + ska1 ( @xmath769 ) & @xmath357 & @xmath1487 & @xmath1021 & @xmath1488 + + ska2 ( @xmath764 ) & @xmath488 & @xmath194 & @xmath679 & @xmath1048 + + ska2 ( @xmath769 ) & @xmath637 & @xmath398 & @xmath1489 & @xmath1051 + planck + sa(0.2 ) + bao + @xmath131 & @xmath405 & @xmath1243 & @xmath1276 & @xmath1277 + + ska1 ( @xmath764 ) & @xmath405 & @xmath1247 & @xmath262 & @xmath1207 + + ska1 ( @xmath769 ) & @xmath674 & @xmath1490 & @xmath958 & @xmath346 + + ska2 ( @xmath764 ) & @xmath598 & @xmath1027 & @xmath650 & @xmath495 + + ska2 ( @xmath769 ) & @xmath1211 & @xmath955 & @xmath1491 & @xmath1214 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath1061 & @xmath847 & @xmath405 & @xmath688 & @xmath489 + + ska1 ( @xmath764 ) & @xmath1262 & @xmath1227 & @xmath637 & @xmath1393 & @xmath445 + + ska1 ( @xmath769 ) & @xmath608 & @xmath1140 & @xmath906 & @xmath736 & @xmath541 + + ska2 ( @xmath764 ) & @xmath1482 & @xmath1492 & @xmath448 & @xmath1493 & @xmath746 + + ska2 ( @xmath769 ) & @xmath1494 & @xmath1485 & @xmath852 & @xmath1495 & @xmath611 + planck + sa(0.65 ) + bao + @xmath131 & @xmath1360 & @xmath738 & @xmath291 & @xmath701 & @xmath709 + + ska1 ( @xmath764 ) & @xmath271 & @xmath1496 & @xmath873 & @xmath1221 & @xmath709 + + ska1 ( @xmath769 ) & @xmath1497 & @xmath1498 & @xmath697 & @xmath1499 & @xmath432 + + ska2 ( @xmath764 ) & @xmath1325 & @xmath1500 & @xmath600 & @xmath1426 & @xmath505 + + ska2 ( @xmath769 ) & @xmath1299 & @xmath1501 & @xmath949 & @xmath1502 & @xmath1503 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath340 & @xmath1370 & @xmath1371 & @xmath1372 + + ska1 ( @xmath764 ) & @xmath456 & @xmath1239 & @xmath1504 & @xmath1304 + + ska1 ( @xmath769 ) & @xmath368 & @xmath1170 & @xmath1505 & @xmath1306 + + ska2 ( @xmath764 ) & @xmath460 & @xmath1209 & @xmath261 & @xmath1506 + + ska2 ( @xmath769 ) & @xmath680 & @xmath622 & @xmath1172 & @xmath307 + planck + sa(0.65 ) + bao + @xmath131 & @xmath259 & @xmath194 & @xmath1382 & @xmath1383 + + ska1 ( @xmath764 ) & @xmath259 & @xmath1062 & @xmath1507 & @xmath1508 + + ska1 ( @xmath769 ) & @xmath578 & @xmath1035 & @xmath1418 & @xmath1509 + + ska2 ( @xmath764 ) & @xmath1174 & @xmath955 & @xmath1456 & @xmath1510 + + ska2 ( @xmath769 ) & @xmath569 & @xmath444 & @xmath1511 & @xmath1512 + [ 0.77 ] l|ccccc + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath1391 & @xmath686 & @xmath456 & @xmath688 & @xmath489 + + ska1 ( @xmath764 ) & @xmath1513 & @xmath1227 & @xmath669 & @xmath211 & @xmath445 + + ska1 ( @xmath769 ) & @xmath1514 & @xmath1140 & @xmath440 & @xmath695 & @xmath541 + + ska2 ( @xmath764 ) & @xmath1515 & @xmath1516 & @xmath448 & @xmath1223 & @xmath746 + + ska2 ( @xmath769 ) & @xmath1517 & @xmath1518 & @xmath817 & @xmath1519 & @xmath911 + planck + sa(0.65 ) + bao + @xmath131 & @xmath1399 & @xmath699 & @xmath593 & @xmath701 & @xmath709 + + ska1 ( @xmath764 ) & @xmath201 & @xmath1496 & @xmath814 & @xmath945 & @xmath746 + + ska1 ( @xmath769 ) & @xmath1520 & @xmath1521 & @xmath614 & @xmath744 & @xmath709 + + ska2 ( @xmath764 ) & @xmath1331 & @xmath1500 & @xmath600 & @xmath1522 & @xmath1144 + + ska2 ( @xmath769 ) & @xmath1332 & @xmath1523 & @xmath886 & @xmath1524 & @xmath403 + + & @xmath226 & @xmath227 & @xmath228 & @xmath1000 + planck + pb-2(0.65 ) + bao + @xmath131 & @xmath340 & @xmath1370 & @xmath1405 & @xmath1406 + + ska1 ( @xmath764 ) & @xmath456 & @xmath1525 & @xmath1526 & @xmath1207 + + ska1 ( @xmath769 ) & @xmath368 & @xmath1527 & @xmath647 & @xmath346 + + ska2 ( @xmath764 ) & @xmath578 & @xmath1035 & @xmath1380 & @xmath495 + + ska2 ( @xmath769 ) & @xmath422 & @xmath637 & @xmath1456 & @xmath1203 + planck + sa(0.65 ) + bao + @xmath131 & @xmath245 & @xmath1412 & @xmath1413 & @xmath1414 + + ska1 ( @xmath764 ) & @xmath245 & @xmath1201 & @xmath1528 & @xmath1529 + + ska1 ( @xmath769 ) & @xmath578 & @xmath1035 & @xmath1275 & @xmath1530 + + ska2 ( @xmath764 ) & @xmath1531 & @xmath669 & @xmath1202 & @xmath455 + + ska2 ( @xmath769 ) & @xmath313 & @xmath845 & @xmath1532 & @xmath1350 +
in this paper , we have studied how well we can constrain the total neutrino mass @xmath1533 , the effective number of neutrino species @xmath13 and the neutrino mass hierarchy by using 21 cm line ( ska ) and cmb ( planck + polarbear-2 or simons array ) observations .
it is essential to combine the 21 cm line observation with the precise cmb polarization observation to break various degeneracies in cosmological parameters when we perform multiple - parameter fittings . about the constraints on the @xmath148@xmath13 plane
, we have found that there is a significant improvement in the sensitivity to @xmath148 and @xmath13 by adding the information of bao observation to that of cmb .
however , for a fiducial value @xmath169 ev , it is impossible to detect the nonzero neutrino mass at 2@xmath166 level even by using the combination of simons array and desi . on the other hand , by adding the information of 21 cm observation ( ska ) to that of cmb , we have found that there is a substantial improvement . by using planck + simons array + bao ( desi ) + ska phase 1
, we can detect the nonzero neutrino mass if the value satisfies @xmath1534 0.1 ev ( but it is necessary to remove foregrounds with high degree of accuracy ) . for a fiducial value @xmath171 ev , which corresponds to the lowest value in the normal hierarchy of the neutrino total mass , we need the sensitivity of ska phase 2 with relatively large sky coverage and very strong foreground removal in order to detect the nonzero neutrino mass at 2@xmath166 level . as for the determination of the neutrino mass hierarchy
, we have used the parameter @xmath1535 , and studied how to discriminate a true mass hierarchy from the other by constraining @xmath155 . as was clearly shown in fig .
[ fig : hie_ellipse_fsky0016]-[fig : hie_ellipse_m8 ] , by adopting the combination of planck + simons array + bao ( desi ) + ska phase 2 , we will be able to determine the hierarchy to be inverted or normal at 2@xmath166 unless the mass structure is degenerated . by using omniscope
, we can discriminate any mass hierarchies up to @xmath980 ev @xcite .
our results indicate that combining the 21 cm line observations with the cmb polarization observations has strong impacts on the determinations of the neutrino property and the origin of matter in the universe .
we thank kiyotomo ichiki for a useful correspondence about specications of ska , and maresuke shiraishi for useful comments about treatments of the residual foregrounds of cmb . this work is supported in part by mext kakenhi grant numbers 15h05889 ( k.k . )
, 15h05891 ( m.h . ) , jsps kakenhi grant numbers 26105520 ( k.k . ) , 26247042 ( k.k . ) and 26220709 ( m.h . ) .
the work of k.k . is also supported by the center for the promotion of integrated science ( cpis ) of sokendai ( 1hb5804100 ) .
while finalizing this manuscript , ref.@xcite appeared which has some overlaps with this work .
p. adamson _ et al .
_ [ minos collaboration ] , phys . rev .
* 101 * , 131802 ( 2008 ) [ arxiv:0806.2237 [ hep - ex ] ] .
t. thummler [ katrin collaboration ] , phys .
nucl . * 42 * , 590 ( 2011 ) ; m. beck [ katrin collaboration ] , j. phys .
ser . * 203 * , 012097 ( 2010 ) [ arxiv:0910.4862 [ nucl - ex ] ] .
j. j. gomez - cadenas , j. martin - albo , m. sorel , p. ferrario , f. monrabal , j. munoz - vidal , p. novella and a. poves , jcap * 1106 * , 007 ( 2011 ) [ arxiv:1010.5112 [ hep - ex ] ] .
ino , india based neutrino observatory , url http://www.ino.tifr.res.in/ino/ m. blennow and t. schwetz , arxiv:1203.3388 [ hep - ph ] .
e. k. akhmedov , s. razzaque and a. y. smirnov , arxiv:1205.7071 [ hep - ph ] .
d. s. ayres _
[ nova collaboration ] , hep - ex/0503053 .
m. ishitsuka , t. kajita , h. minakata and h. nunokawa , phys .
d * 72 * , 033003 ( 2005 ) [ hep - ph/0504026 ] .
k. hagiwara , n. okamura and k. -i .
senda , phys .
b * 637 * , 266 ( 2006 ) [ erratum - ibid .
b * 641 * , 491 ( 2006 ) ] [ hep - ph/0504061 ] .
a. badertscher , t. hasegawa , t. kobayashi , a. marchionni , a. meregaglia , t. maruyama , k. nishikawa and a. rubbia , arxiv:0804.2111 [ hep - ph ] .
u. seljak _ et al . _
[ sdss collaboration ] , phys .
d * 71 * , 103515 ( 2005 ) [ astro - ph/0407372 ] .
a. goobar , s. hannestad , e. mortsell and h. tu , jcap * 0606 * , 019 ( 2006 ) [ astro - ph/0602155 ] . k. ichiki , m. takada and t. takahashi , phys .
d * 79 * , 023520 ( 2009 ) .
s. a. thomas , f. b. abdalla and o. lahav , phys . rev .
lett . * 105 * , 031301 ( 2010 ) [ arxiv:0911.5291 [ astro-ph.co ] ] .
s. riemer - sorensen , c. blake , d. parkinson , t. m. davis , s. brough , m. colless , c. contreras and w. couch _ et al .
_ , phys .
d * 85 * , 081101 ( 2012 ) [ arxiv:1112.4940 [ astro-ph.co ] ] .
s. hannestad , a. mirizzi , g. g. raffelt and y. y. y. wong , jcap * 1008 * , 001 ( 2010 ) .
p. crotty , j. lesgourgues and s. pastor , phys .
d * 69 * , 123007 ( 2004 ) [ hep - ph/0402049 ] .
u. seljak , a. slosar and p. mcdonald , jcap * 0610 * , 014 ( 2006 ) [ astro - ph/0604335 ] .
m. fukugita , k. ichikawa , m. kawasaki and o. lahav , phys .
d * 74 * , 027302 ( 2006 ) .
e. komatsu _ et al .
_ [ wmap collaboration ] , astrophys . j. suppl . * 180 * , 330 ( 2009 ) [ arxiv:0803.0547 [ astro - ph ] ] .
b. a. reid , w. j. percival , d. j. eisenstein , l. verde , d. n. spergel , r. a. skibba , n. a. bahcall and t. budavari _ et al .
_ , mon .
soc . * 404 * , 60 ( 2010 ) [ arxiv:0907.1659 [ astro-ph.co ] ] .
b. a. reid , l. verde , r. jimenez and o. mena , jcap * 1001 * , 003 ( 2010 ) [ arxiv:0910.0008 [ astro-ph.co ] ] .
j. hamann , s. hannestad , j. lesgourgues , c. rampf and y. y. y. wong , jcap * 1007 * , 022 ( 2010 ) [ arxiv:1003.3999 [ astro-ph.co ] ] .
e. di valentino , s. galli , m. lattanzi , a. melchiorri , p. natoli , l. pagano and n. said , phys .
d * 88 * , no . 2 , 023501 ( 2013 ) [ arxiv:1301.7343 [ astro-ph.co ] ] .
p. a. r. ade _ et al .
_ [ planck collaboration ] , arxiv:1303.5076 [ astro-ph.co ] .
p. a. r. ade _ et al .
_ [ planck collaboration ] , arxiv:1502.01589 [ astro-ph.co ] .
e. giusarma , e. di valentino , m. lattanzi , a. melchiorri and o. mena , phys .
d * 90 * , no . 4 , 043507 ( 2014 ) [ arxiv:1403.4852 [ astro-ph.co ] ] .
e. di valentino , e. giusarma , m. lattanzi , o. mena , a. melchiorri and j. silk , arxiv:1507.08665 [ astro-ph.co ] . e. pierpaoli , mon . not .
342 * , l63 ( 2003 ) [ astro - ph/0302465 ] .
p. crotty , j. lesgourgues and s. pastor , phys .
d * 67 * , 123005 ( 2003 ) [ astro - ph/0302337 ] .
n. said , e. di valentino and m. gerbino , phys .
d * 88 * , no . 2 , 023513 ( 2013 ) [ arxiv:1304.6217 [ astro-ph.co ] ] .
e. di valentino , m. gerbino and a. melchiorri , phys .
d * 87 * , no .
10 , 103523 ( 2013 ) .
e. di valentino , m. lattanzi , g. mangano , a. melchiorri and p. serpico , curvaton scenario , phys .
d * 85 * , 043511 ( 2012 ) [ arxiv:1111.3810 [ astro-ph.co ] ] .
k. n. abazajian _ et al .
_ [ topical conveners : k.n .
abazajian , j.e .
carlstrom , a.t .
lee collaboration ] , astropart .
phys . * 63 * , 66 ( 2015 ) [ arxiv:1309.5383 [ astro-ph.co ] ]
. w. l. k. wu , j. errard , c. dvorkin , c. l. kuo , a. t. lee , p. mcdonald , a. slosar and o. zahn , astrophys .
j. * 788 * , 138 ( 2014 ) [ arxiv:1402.4108 [ astro-ph.co ] ] .
m. mcquinn , o. zahn , m. zaldarriaga , l. hernquist and s. r. furlanetto , astrophys .
j. * 653 * , 815 ( 2006 ) [ arxiv : astro - ph/0512263 ] .
a. loeb and s. wyithe , phys .
lett . * 100 * , 161301 ( 2008 ) .
j. r. pritchard and e. pierpaoli , phys .
d * 78 * , 065009 ( 2008 ) .
j. r. pritchard and e. pierpaoli , nucl .
phys . proc .
* 188 * , 31 ( 2009 ) .
k. n. abazajian , e. calabrese , a. cooray , f. de bernardis , s. dodelson , a. friedland , g. m. fuller and s. hannestad _ et al .
_ , astropart .
phys . * 35 * , 177 ( 2011 ) [ arxiv:1103.5083 [ astro-ph.co ] ] .
y. oyama , a. shimizu and k. kohri , phys .
b * 718 * , 1186 ( 2013 ) [ arxiv:1205.5223 [ astro-ph.co ] ] .
y. oyama , `` constraints on the neutrino parameters by future cosmological 21 cm line and precise cmb polarization observations ( phd thesis , the graduate university for advanced studies ( sokendai ) , 2014 ) '' , arxiv:1510.05161 [ astro-ph.co ] .
k. kohri , y. oyama , t. sekiguchi and t. takahashi , jcap * 1409 * , 014 ( 2014 ) [ arxiv:1404.4847 [ astro-ph.co ] ] .
m. tegmark and m. zaldarriaga , phys .
d * 79 * , 083530 ( 2009 ) [ arxiv:0805.4414 [ astro - ph ] ] .
m. tegmark and m. zaldarriaga , phys .
d * 82 * , 103501 ( 2010 ) [ arxiv:0909.0001 [ astro-ph.co ] ] .
k. kohri , y. oyama , t. sekiguchi and t. takahashi , jcap * 1310 * , 065 ( 2013 ) [ arxiv:1303.1688 [ astro-ph.co ] ] .
s. furlanetto , s. p. oh and f. briggs , phys . rept .
* 433 * , 181 ( 2006 ) .
j. r. pritchard and a. loeb , arxiv:1109.6012 [ astro-ph.co ] .
p. madau , a. meiksin and m. j. rees , astrophys . j. * 475 * , 429 ( 1997 ) [ arxiv : astro - ph/9608010 ] .
s. furlanetto , mon . not .
* 371 * , 867 ( 2006 ) [ arxiv : astro - ph/0604040 ] .
j. r. pritchard and a. loeb , phys .
d * 78 * , 103511 ( 2008 ) [ arxiv:0802.2102 [ astro - ph ] ] .
g. b. field , proc .
ire . * 46 * , 240 ( 1958 ) .
y. mao , m. tegmark , m. mcquinn , m. zaldarriaga and o. zahn , phys .
d * 78 * , 023529 ( 2008 ) [ arxiv:0802.1710 [ astro - ph ] ] .
m. mcquinn , a. lidz , o. zahn , s. dutta , l. hernquist and m. zaldarriaga , mon . not .
soc . * 377 * , 1043 ( 2007 ) [ astro - ph/0610094 ] .
m. mcquinn , l. hernquist , m. zaldarriaga and s. dutta , mon . not .
soc . * 381 * , 75 ( 2007 ) [ arxiv:0704.2239 [ astro - ph ] ] . m. tegmark , a. taylor and a. heavens , astrophys .
j. * 480 * , 22 ( 1997 ) [ arxiv : astro - ph/9603021 ] .
e. chapman , f. b. abdalla , g. harker , v. jelic , p. labropoulos , s. zaroubi , m. a. brentjens and a. g. de bruyn _ et al .
_ , arxiv:1201.2190 [ astro-ph.co ] . http://www.lofar.org/ m. takada , e. komatsu and t. futamase , phys . rev .
d * 73 * , 083520 ( 2006 ) [ astro - ph/0512374 ] .
a. slosar , phys .
d * 73 * , 123501 ( 2006 ) [ astro - ph/0602133 ] .
f. de bernardis , t. d. kitching , a. heavens and a. melchiorri , phys .
d * 80 * , 123509 ( 2009 ) [ arxiv:0907.1917 [ astro-ph.co ] ] .
r. jimenez , t. kitching , c. pena - garay and l. verde , jcap * 1005 * , 035 ( 2010 ) [ arxiv:1003.5918 [ astro-ph.co ] ] .
t. okamoto and w. hu , phys .
d * 67 * , 083002 ( 2003)[astro - ph/0301031 ] .
d. baumann _ et al .
_ [ cmbpol study team collaboration ] , aip conf .
proc . * 1141 * , 10 ( 2009 ) [ arxiv:0811.3919 [ astro - ph ] ] .
l. verde , h. peiris and r. jimenez , jcap * 0601 * , 019 ( 2006 ) [ astro - ph/0506036 ] .
n. katayama and e. komatsu , astrophys .
j. * 737 * , 78 ( 2011 ) [ arxiv:1101.5210 [ astro-ph.co ] ]
. l. page _ et al .
_ [ wmap collaboration ] , astrophys . j. suppl . * 170 * , 335 ( 2007 ) [ astro - ph/0603450 ] .
e. m. leitch , j. m. kovac , n. w. halverson , j. e. carlstrom , c. pryke and m. w. e. smith , astrophys . j. * 624 * , 10 ( 2005 ) [ astro - ph/0409357 ] .
d. p. finkbeiner , m. davis and d. j. schlegel , astrophys .
j. * 524 * , 867 ( 1999 ) [ astro - ph/9905128 ] . j. tauber _ et al .
_ [ planck collaboration ] , astro - ph/0604069 .
a. albrecht , g. bernstein , r. cahn , w. l. freedman , j. hewitt , w. hu , j. huth and m. kamionkowski _ et al .
_ , astro - ph/0609591 .
c. blake , d. parkinson , b. bassett , k. glazebrook , m. kunz and r. c. nichol , mon . not .
soc . * 365 * , 255 ( 2006 ) [ astro - ph/0510239 ] .
a. font - ribera , p. mcdonald , n. mostek , b. a. reid , h. j. seo and a. slosar , jcap * 1405 * , 023 ( 2014 ) [ arxiv:1308.4164 [ astro-ph.co ] ] .
a. lewis , a. challinor and a. lasenby , astrophys .
j. * 538 * , 473 ( 2000 ) [ astro - ph/9911177 ] .
a. liu , j. r. pritchard , r. allison , a. r. parsons , u. seljak and b. d. sherwin , arxiv:1509.08463 [ astro-ph.co ] . | observations of the 21 cm line radiation coming from the epoch of reionization have a great capacity to study the cosmological growth of the universe .
besides , cmb polarization produced by gravitational lensing has a large amount of information about the growth of matter fluctuations at late time . in this paper
, we investigate their sensitivities to the impact of neutrino property on the growth of density fluctuations , such as the total neutrino mass , the effective number of neutrino species ( extra radiation ) , and the neutrino mass hierarchy .
we show that by combining a precise cmb polarization observation such as simons array with a 21 cm line observation such as square kilometer array ( ska ) phase 1 and a baryon acoustic oscillation observation ( dark energy spectroscopic instrument : desi ) we can measure effects of non - zero neutrino mass on the growth of density fluctuation if the total neutrino mass is larger than @xmath0 ev .
additionally , the combinations can strongly improve errors of the bounds on the effective number of neutrino species @xmath1 at 95 % c.l .. finally , by using ska phase 2 , we can determine the neutrino mass hierarchy at 95 % c.l . if the total neutrino mass is similar to or smaller than 0.1 ev .
.5 in * constraints on the neutrino parameters by future cosmological 21 cm line and precise cmb polarization observations * .45 in yoshihiko oyama@xmath2 , kazunori kohri@xmath3 and masashi hazumi@xmath4 .45 in _
@xmath5 institute for cosmic ray research , the university of tokyo , 5 - 1 - 5 kashiwanoha , kashiwa , chiba 277 - 8582 , japan + @xmath6 institute of particle and nuclear studies , kek , 1 - 1 oho , tsukuba , ibaraki 305 - 0801 , japan + @xmath7 the graduate university for advanced studies ( sokendai ) , 1 - 1 oho , tsukuba , ibaraki 305 - 0801 , japan + @xmath8 kavli ipmu ( wpi ) , utias , the university of tokyo , 5 - 1 - 5 kashiwanoha , kashiwa , chiba 277 - 8583 , japan _
.4 in |
the spin - dependent transport through magnetic nanostructures has attracted much interest.@xcite in the early 1970s , meservey and tedrow have showed that tunneling electrons between a ferromagnetic metal ( fe , co , ni ) and a thin film of superconducting aluminium ( al ) are spin - polarized.@xcite the ferromagnet / insulator / superconductor ( fm / i / sc ) tunnel junctions are one of the most powerful tools to extract the spin polarization of the conduction electrons near the fermi level . in fm / i / sc and fm / i / sc / i / fm tunnel junctions , the suppression of superconducting gap due to spin accumulation by injection of spin polarized quasiparticles ( qp s ) has been shown.@xcite the qp s spin transport and relaxation in sc has been studied in detail .
@xcite in recent years , much attention has been focused on fm / sc metallic contacts both theoretically@xcite and experimentally @xcite since the spin polarization of conduction electrons is measured by using the andreev reflection@xcite : an electron injected from the fm into the sc is reflected as a hole at the fm / sc interface and a cooper pair is generated in the sc .
the andreev reflection includes a conversion process of the qp current to the supercurrent carried by cooper pairs in the range of the penetration depth , which is approximately equal to the ginzburg - landau ( gl ) coherence length,@xcite from the fm / sc interface.@xcite thus , a fm / sc / fm double junction is particularly interesting because the magnetoresistance is expected due to the overlap of the qp penetration in the sc by the andreev reflection .
recently , gu _ et al . _ have measured the magnetoresistance in a current perpendicular to plane ( cpp ) geometry consisting of a superconducting niobium ( nb ) thin film sandwiched by ferromagnetic permalloys ( py ) and proposed a method for estimating the penetration depth by measuring the magnetoresistance.@xcite in this paper , we present a theory of the andreev reflection in the fm / sc / fm double junction system and derive an expression of the current through the junction by extending the theory of blonder , tinkham and klapwijk ( btk).@xcite we numerically calculate the current for the parallel and anti - parallel alignments of magnetizations , and investigate the dependences of the magnetoresistance on the thickness of the sc , temperature , the exchange field of fm s and the height of the interfacial barriers .
it is shown that these dependences are understood by considering the penetration of quasiparticles to the sc by the andreev reflection process .
finally , we compare our results with the recent experimental results by gu _
is sandwiched by two semi - infinite ferromagnetic electrodes . the system is rectangular and the cross section is a square of side @xmath0 .
( b ) the current flows along the @xmath1-axis .
the interfaces between fm1/sc and sc / fm2 are located at @xmath2 and @xmath3 , respectively . ]
we consider a fm1/sc / fm2 double junction system consisting of three rectangular blocks as shown in figs .
1(a ) and 1(b ) .
the cross section of the system is a square of side @xmath0 and the thickness of the sc is @xmath4 .
the current flows along the @xmath1-axis and the interfaces between fm1/sc and sc / fm2 are located at @xmath2 and @xmath3 , respectively . for simplicity , we assume that the system is symmetric : fm1 and fm2 are made of the same ferromagnetic materials and the potentials for the left and right interfaces are the same .
the system we consider is described by the following bogoliubov - de gennes ( bdg ) equation @xcite : @xmath5 where @xmath6 is the single particle hamiltonian , @xmath7 is the qp energy measured from the fermi energy @xmath8 and @xmath9 is for the up-(down-)spin band . the exchange field @xmath10 is given by @xmath11 where @xmath12 and @xmath13 represent the exchange fields for the parallel and anti - parallel alignments , respectively .
the superconducting gap is expressed as @xmath14 we assume that the temperature dependence of the superconducting gap is given by @xmath15,@xcite where @xmath16 is the superconducting gap at @xmath17 and @xmath18 is the superconducting critical temperature . in order to capture the essential effect of the interfacial scattering ,
we employ the following @xmath19-function type potential at the interfaces : @xmath20 throughout this paper , we neglect the spin - flip scattering in the sc and the proximity effect near the interfaces.@xcite since the system is rectangular , the wave function in the transverse ( @xmath21 and @xmath22 ) directions is given by @xmath23 where @xmath24 and @xmath25 are the quantum numbers which define the channel .
the eigenvalue of the transverse mode for the channel ( @xmath24,@xmath25 ) is @xmath26 .
\end{aligned}\ ] ] the solution of the bdg equation ( [ eq : bdg ] ) in the sc region is given by @xmath27 where @xmath28 and @xmath29 are the coherence factors expressed as @xmath30 , \end{aligned}\ ] ] and @xmath31 is the @xmath1 component of the wave number of an electron-(hole-)like qp in the channel @xmath32 defined as @xmath33 in the fm region , the solutions are given by @xmath34 where @xmath35 is the @xmath1 component of the wave number of an electron ( hole ) with @xmath36-spin in the channel @xmath32 ; @xmath37 ) is denoted by 0 , along with the resulting scattering processes : the andreev reflection ( 1 ) , the normal reflection ( 2 ) at the interface of fm1/sc , the transmission to the sc ( 3 , 4 ) and the reflection at the interface of sc / fm2 ( 5 , 6 ) , the transmission as an electron to the fm2 ( 7 ) and the one as a hole ( 8) . ]
the wave function of the fm1/sc / fm2 double junction system is given by the linear combination of the solutions .
let us consider the scattering of an electron with up - spin in the channel @xmath32 injected into the sc from the fm1 ( 0 in fig.[fig : process ] ) .
there are the following eight processes : the andreev reflection ( 1 in fig.[fig : process ] ) , the normal reflection ( 2 in fig.[fig : process ] ) at the interface of fm1/sc , the transmission to the sc ( 3 , 4 in fig.[fig : process ] ) and the reflection at the interface of sc / fm2 ( 5 , 6 in fig.[fig : process ] ) , the transmission as an electron to the fm2 ( 7 in fig.[fig : process ] ) and the one as a hole ( 8 in fig.[fig : process ] ) .
therefore , the wave function in the fm1 @xmath38 is given by @xmath39 { \rm s}_{nl}(x , y ) .
\label{eq : wf2-fm1}\end{aligned}\ ] ] in the sc @xmath40 we have @xmath41 { \rm s}_{nl}(x , y ) , \label{eq : wf2-sc}\end{aligned}\ ] ] and in the fm2 @xmath42 @xmath43 { \rm s}_{nl}(x , y ) .
\end{split } \label{eq : wf2-fm2}\ ] ] here @xmath44 , @xmath45 and @xmath46 are the wave numbers in the fm1 , sc and fm2 , respectively .
the coefficients @xmath47 , @xmath48 , and @xmath49 are determined by matching the wave functions at the left and right interfaces .
the matching conditions for the wave functions ( [ eq : wf2-fm1 ] ) - ( [ eq : wf2-fm2 ] ) at the interfaces are @xmath50 solving eq.([eq : bound ] ) , the probabilities of transmission and reflection are calculated following the btk theory .
@xcite when an electron with @xmath36-spin is injected from the fm1 , the probability of the andreev reflection @xmath51 , the normal reflection @xmath52 and the transmission as an electron and the one as a hole to the fm2 , @xmath53 and @xmath54 are given by @xmath55 where the subscript @xmath56 in eq .
( [ eq : rt ] ) indicates the electron ( hole ) in the fm2 .
let us evaluate the current in the fm1 .
when the bias voltage @xmath57 is applied to the fm1/sc / fm2 system , the current carried by electrons with @xmath36-spin is written as @xmath58 } de , \end{aligned}\ ] ] where @xmath59 is planck constant and @xmath60 is the distribution function of an electron with a positive group velocity in the @xmath1 direction and expressed as @xmath61 where @xmath62 is the fermi distribution function .
the distribution function of the electron with a negative group velocity in the @xmath1 direction is written as @xmath63\nonumber\\ & & = r_{\sigma , nl}^{eh } f_0 { \left ( \mbox{$e + \frac{ev}{2}$}\right ) } + r_{\sigma , nl}^{ee } f_0 \left ( \mbox{$e - \frac{ev}{2}$ }
\right ) \nonumber\\ & & \ \ \ + t_{\sigma , nl}^{ee ' } f_0 \left ( \mbox{$e + \frac{{ev}}{2}$ }
\right ) + t_{\sigma , nl}^{eh ' } f_0 \left ( \mbox{$e - \frac{ev}{2}$ } \right),\end{aligned}\ ] ] where @xmath64 is the fermi velocity of an electron with @xmath36-spin in the channel @xmath32 of the fm1 ( fm2 ) and @xmath65 is the density of states of @xmath36-spin band in the channel @xmath32 of the fm1 ( fm2 ) . using the conservation of probability , @xmath66 , we have @xmath67 de . \end{split}\ ] ] the current carried by holes @xmath68
is calculated in the similar way .
the total current in the fm1/sc / fm2 double junction system is obtained as @xmath69\\ & = \frac{e}{h}\sum_{nl,\sigma } \int_0^\infty \left ( r_{nl,\sigma}^{eh } + r_{nl,\sigma}^{he } + t_{nl,\sigma}^{ee ' } + t_{nl,\sigma}^{hh ' } \right)\\ & \times \left [ { f_0 \left ( \mbox{$e - \frac{ev}{2}$ } \right ) - f_0 \left ( \mbox{$e + \frac{ev}{2}$ } \right ) } \right]de .
\end{split } \label{current}\ ] ] note that this expression of the current eq .
( [ current ] ) reduces to the one derived by lambert@xcite for the normal metal / superconductor / normal metal system when @xmath70 . the magnetoresistance ( mr )
is defined as @xmath71 where @xmath72 is the resistance in the parallel ( anti - parallel ) alignment .
. from top to bottom ,
temperature @xmath73 is 1 , 0.9 , 0.7 , 0.5 , 0.3 , and 0.1 .
we assume @xmath74 . ] in fig .
[ fig : thick ] the mr is plotted as a function of the thickness of the sc , @xmath4 , multiplied by the fermi wave number @xmath75 .
we assume that the strength of the interfacial barrier @xmath76 and the exchange field @xmath77 . the side length of the cross section is taken to be @xmath78 throughout this paper .
when the sc is in the normal conducting state ( @xmath79 ) , the mr is constant since we neglect the spin - flip scattering in the sc .
when the sc is in the superconducting state ( @xmath73 = 0.1 , 0.3 , 0.5 , 0.7 , and 0.9 ) , the mr decreases with increasing the thickness of the sc .
the mr at low temperatures @xmath80 shows an exponential decrease in a wide range of @xmath81 .
the decrease of the mr due to the superconductivity can be explained by considering the decay of the quasiparticle current in the sc . to obtain the charge and spin currents , we extend the btk theory of the andreev reflection @xcite to the fm1/sc / fm2 double junction system .
let us first consider the charge transport in the sc .
when the current flows in the positive @xmath1 direction , electrons and holes are injected into the sc from the fm1 and the fm2 , respectively .
the conservation law of the charge density @xmath82 in the sc , where @xmath83 and @xmath84 are electron- and hole - like components of the wave function in the channel ( @xmath85 ) , respectively , is derived from the bdg equation ( [ eq : bdg ] ) and obtained as @xmath86 where @xmath87 is the qp current density with @xmath36-spin and the @xmath1 component of @xmath87 per unit area @xmath88 is written as @xmath89 component of ( a ) the charge current density and ( b ) the spin current density in the sc with the thickness @xmath90 is shown .
@xmath91 , @xmath92 and @xmath93 are the qp current density , the supercurrent density and the spin current density in the parallel ( anti - parallel ) alignment , respectively .
we assume @xmath74.,title="fig : " ] component of ( a ) the charge current density and ( b ) the spin current density in the sc with the thickness @xmath90 is shown .
@xmath91 , @xmath92 and @xmath93 are the qp current density , the supercurrent density and the spin current density in the parallel ( anti - parallel ) alignment , respectively .
we assume @xmath74.,title="fig : " ] the right hand side of eq .
( [ contin ] ) corresponds to the gradient of supercurrent carried by cooper pair @xmath94 defined as @xmath95 from which the @xmath1 component of @xmath94 per area @xmath96 is obtained as @xmath97 the @xmath1 coordinate dependences of the qp current density @xmath98 and the supercurrent @xmath99 in the case that the thickness of the sc @xmath90 are shown in fig .
[ fig:3000](a ) . for @xmath100 .
we assume @xmath74 . ]
we find that @xmath101 decays from the interfaces of the fm1/sc and the sc / fm2 and becomes zero in the interior of the sc .
on the other hand , @xmath102 increases and becomes dominant in the interior of the sc to conserve the total current density . in the energy region below the superconducting gap ( @xmath103 )
where the energy of the transverse mode @xmath104 is smaller than fermi energy @xmath8 , the wave number @xmath105 is expanded as @xmath106 the imaginary part in eq .
( [ k - expansion ] ) gives the exponential decay term @xmath107 in @xmath101 , where @xmath108 is the penetration depth given by @xmath109 where @xmath110 is the fermi velocity .
@xmath108 has a strong temperature dependence shown in fig .
[ fig : xiq ] .
when the thickness of the sc is much larger than the penetration depth ( @xmath111 ) as in fig .
[ fig:3000](a ) , @xmath112 decays in the range of @xmath108 from the interfaces .
note that @xmath108 is approximately equal to the clean - limit gl coherence length @xmath113 in the low energy regime : @xmath114.@xcite next , we consider the spin transport in the sc .
the conservation law of the spin density @xmath115 , where @xmath116 , is derived from the bdg equation ( [ eq : bdg ] ) and expressed as @xmath117 where @xmath118 is the spin current density
. the @xmath1 component of @xmath119 per unit area @xmath120 is written as @xmath121 figure [ fig:3000](b ) shows the @xmath1 coordinate dependence of the spin current density @xmath122 in the sc with the thickness @xmath90 . as shown in fig .
[ fig:3000](b ) , no spin current flows through the sc both in the parallel and in the anti - parallel alignments because the qp current with spin changes to the supercurrent carried by cooper pairs with no spin .
this means that the spin injected from fm1 does not reach to the fm2 . as a result , the qp current density in the parallel alignment @xmath123 and that in the anti - parallel alignment @xmath124 are almost the same as shown in fig . [
. component of ( a ) the charge current density and ( b ) the spin current density in the sc with the thickness @xmath125 is shown .
the parameters are the same as in fig .
[ fig:3000].,title="fig : " ] component of ( a ) the charge current density and ( b ) the spin current density in the sc with the thickness @xmath125 is shown .
the parameters are the same as in fig .
[ fig:3000].,title="fig : " ] the @xmath1 coordinate dependence of the charge current in the case that the thickness of the sc is much smaller than the penetration depth ( @xmath126 ) is shown in fig .
[ fig:10](a ) , where the qp current density @xmath101 is almost constant and the supercurrent density @xmath102 is nearly zero in the sc . as shown in fig .
[ fig:10](b ) , the value of the spin current in the parallel alignment is larger than that in the anti - parallel alignment because the value of qp current with up - spin is much larger than that with down - spin in the parallel alignment , whereas the value of qp current with up - spin is equal to that with down - spin in the anti - parallel alignment .
this means that the spin injected from the fm1 is transferred to the fm2 and therefore the value of the qp current density strongly depends on the relative orientation of the fm s magnetizations shown in fig .
[ fig:10](a ) . from the above discussion , the result shown in fig .
[ fig : thick ] is understood as follows . in the sc ,
the qp current with spin decreases exponentially and changes to the supercurrent carried by cooper pairs with no spin in the range of @xmath108 from the interfaces . as a result
, it becomes difficult to transfer the spin from the fm1 to the fm2 and the mr decreases with increasing the thickness of the sc .
the finite mr in the region of large @xmath4 is due to the qp s with energy above the superconducting gap ( @xmath127 ) .
( @xmath128 ) is plotted as a function of temperature .
the thickness of the sc is taken to be @xmath129 and 300 in panels ( a ) and ( b ) , respectively . from bottom to top , the exchange field @xmath130 = 0.1 , 0.3 , 0.5 , 0.7 , and 0.9 .
the mr at @xmath131 is plotted as a function of the exchange field @xmath132 in the inset of panel ( a ) .
we assume @xmath74.,title="fig : " ] ( @xmath128 ) is plotted as a function of temperature .
the thickness of the sc is taken to be @xmath129 and 300 in panels ( a ) and ( b ) , respectively . from bottom to top , the exchange field @xmath130 = 0.1 , 0.3 , 0.5 , 0.7 , and 0.9 .
the mr at @xmath131 is plotted as a function of the exchange field @xmath132 in the inset of panel ( a ) .
we assume @xmath74.,title="fig : " ] the temperature dependences of the mr normalized by the value at @xmath18 ( @xmath128 ) for several values of the exchange field @xmath132 in the case of @xmath133 and @xmath134 are shown in figs .
[ fig : mr_temp_pol](a ) and [ fig : mr_temp_pol](b ) , respectively .
the @xmath128 decreases with decreasing temperature because the number of electrons and holes with the energy @xmath127 which contribute to the @xmath128 decreases with decreasing temperature . at low temperatures , electrons and holes mainly distribute in the energy region @xmath103 .
when the thickness of the sc is much larger than the penetration depth ( fig .
[ fig : mr_temp_pol](a ) ) , electrons and holes with energy @xmath103 injected to the sc do not contribute to the @xmath128 because the qp current changes to the supercurrent in the sc by the andreev reflection and the qp transmission from the fm1 to the fm2 does not occur . as a result ,
the @xmath128 becomes zero at low temperatures @xmath135 . on the other hand , when the thickness of the sc is comparable to the penetration depth ( fig .
[ fig : mr_temp_pol](b ) ) , the qp transmission by the andreev reflection in the energy region @xmath103 occurs and therefore the finite @xmath128 remains even at low temperatures . .
the strength of the interfacial barrier is @xmath76 , @xmath136 and @xmath137 for panels ( a ) , ( b ) and ( c ) respectively . from top
to bottom the thickness of the sc is @xmath138 300 , 600 , 900 , 1200 .
we assume @xmath74.,title="fig : " ] .
the strength of the interfacial barrier is @xmath76 , @xmath136 and @xmath137 for panels ( a ) , ( b ) and ( c ) respectively . from top
to bottom the thickness of the sc is @xmath138 300 , 600 , 900 , 1200 .
we assume @xmath74.,title="fig : " ] .
the strength of the interfacial barrier is @xmath76 , @xmath136 and @xmath137 for panels ( a ) , ( b ) and ( c ) respectively . from top
to bottom the thickness of the sc is @xmath138 300 , 600 , 900 , 1200 .
we assume @xmath74.,title="fig : " ] figures [ fig : mr_temp_finite_p05](a)-(c ) show the temperature dependence of the mr in the cases of @xmath76 , @xmath136 and @xmath137 , respectively , for the several values of @xmath4 . the temperature dependence of the mr in the case of the transparent interfacial barrier ( fig .
[ fig : mr_temp_finite_p05](a ) ) is explained by the same way as in fig . [
fig : mr_temp_pol ] . in the case of the finite interfacial barrier ( figs .
[ fig : mr_temp_finite_p05](b ) and [ fig : mr_temp_finite_p05](c ) ) , for all values of @xmath4 , the normal reflection mainly occurs especially in the energy region below the superconducting gap ( @xmath103 ) because of the scattering at the interfaces .
therefore , the main contribution to the mr at temperature @xmath139 comes from the qp s transmission in the energy region @xmath127 , whose probability is independent of @xmath4 . as a result ,
the differences in the magnitude of the mr for the different values of @xmath4 become smaller especially for temperature @xmath139 .
the solid curves show theoretical results for the thickness of the sc , @xmath14030 , 40 , 50 , 60 , 80 , and 100 nm from top to bottom , where @xmath141 is taken to be @xmath142 for nb .
the symbols show the experimental results by gu _
et al._@xcite for the thickness of the nb , @xmath143=30 , 40 , 50 , 60 , 80 , and 100 nm from top to bottom . ]
let us compare our theory with recent experimental results in py / nb / py structures measured by gu _
et al._.@xcite the mean free path in the nb film @xmath144 nm@xcite is much smaller than the clean - limit coherence length @xmath145 nm@xcite and therefore the nb film is in the diffusive regime . in order to analyze the experimental results in the dirty nb film , we need to extend the theory in the ballistic case to that in the diffusive case .
the diffusive effect on the andreev reflection is incorporated into our theory by replacing the penetration depth @xmath108 in the ballistic theory with the penetration depth in the dirty - limit @xmath146 .
thus , the value of @xmath147 at @xmath17 obtained by fitting the experimental data is interpreted as the dirty - limit penetration depth @xmath148@xcite at @xmath17 .
figure [ fig : pratt ] shows the excess resistance @xmath149 normalized by the value at ( in the experiment , @xmath150 slightly above ) @xmath18 ( @xmath151 ) as a function of temperature .
the solid curves indicate the calculated results and the symbols indicate the experimental ones.@xcite by fitting the calculated values to those of the experimental data , we obtain @xmath152 = 46 , 36 , 36 , 33 , 30 , and 27 nm for the curves of @xmath4=30 , 40 , 50 , 60 , 80 , and 100 nm , respectively , where @xmath141 is taken to be @xmath142 for nb.@xcite the values of the penetration depth @xmath152 estimated by our theory become larger for the smaller thickness of the nb film and are larger than the dirty - limit penetration depth in a bulk nb @xmath153 = 16.2 nm .
this indicates that @xmath154 in the nb film is reduced compared to that in a bulk nb .
the suppression of @xmath154 is due to the proximity effect . actually , the height of the realistic superconducting gap depends on the position @xmath1 in the nb film by the proximity effect . here
we interpret the value of @xmath154 as the averaged value of the realistic superconducting gap with respect to @xmath1 in the nb film .
gu _ et al . _ have obtained the dirty - limit penetration depth @xmath155 nm by assuming that @xmath151 is proportional to @xmath156 , where @xmath143 is the thickness of the nb .
this value of the penetration depth is comparable to those estimated by our theory .
although we neglect the effect of spin relaxation on the mr in our theory , this assumption is justified as follows . in the nb film in the py / nb / py structure , the spin diffusion length @xmath157 is about @xmath158 nm,@xcite while the penetration depth @xmath159 is about @xmath160 nm at low energy . as seen in fig .
[ fig : xiq ] , @xmath161 is almost constant at low temperatures and shows the divergent behavior only near @xmath18 , indicating that @xmath108 is smaller than @xmath157 in most of the temperature range below @xmath18 .
therefore , the effect of the qp current penetration is dominated for the mr and the spin relaxation effect on the mr is neglected except in the close vicinity to @xmath131 .
the magnetoresistance in the ferromagnet / superconductor / ferromagnet double junction system is studied theoretically .
the dependences of the magnetoresistance on the thickness of the superconductor , temperature , the exchange field of ferromagnets and the height of the interfacial barriers are understood by considering the andreev reflection of spin - polarized current .
our theory shows good agreement with recent experimental results .
the authors thank w. p. pratt , jr . for informing them their experimental data prior to publication .
this work is supported by a grant - in - aid from mext and crest of japan .
h. i. is supported by encouragement of young scientists for mext .
soulen , jr .
osofsky , b. nadgorny , t. ambrose , p. broussard , s.f .
cheng , j. byers , c.t .
tanaka , j. nowack , j.s .
moodera , g. laprade , a. barry , and m.d .
coey , j. appl . phys . * 85 * , 4589 ( 1999 ) .
we assume that the dirty - limit penetration depth is expressed as @xmath162 by analogy with the clean - limit case , where @xmath163 is the dirty - limit gl coherence length . using the relation @xmath164,@xcite the dirty - limit penetration depth
can be expressed as @xmath148 . | we present a theory of andreev reflection in a ferromagnet / superconductor / ferromagnet double junction system .
the spin polarized quasiparticles penetrate to the superconductor in the range of penetration depth from the interface by the andreev reflection .
when the thickness of the superconductor is comparable to or smaller than the penetration depth , the spin polarized quasiparticles pass through the superconductor and therefore the electric current depends on the relative orientation of magnetizations of the ferromagnets .
the dependences of the magnetoresistance on the thickness of the superconductor , temperature , the exchange field of the ferromagnets and the height of the interfacial barriers are analyzed .
our theory explains recent experimental results well . |
studies of * s*tar * c*lusters ( * sc*s ) are interesting for many reasons : not only to learn about sc formation , evolution and destruction , but also because scs are valuable benchmarks for stellar evolution models . for instance , effects and typical parameters of stellar rotation and binarity can well be studied on homogeneous ( one age , one metallicity ) , reasonably sized samples of stars in clusters .
pixel - by - pixel analyses of hst acs multi - band imaging data for the tadpole galaxy , an ongoing merger , and other systems , with our galev evolutionary synthesis models have shown that sc formation is a major if not the dominant mode of * s*tar * f*ormation ( * sf * ) in starbursts and galaxy mergers with , e.g. , 70 % of the total u - band light coming from young scs in the tadpole , not only in the main body of this galaxy , but all along its 180 kpc long tidal tail ( de grijs _ et al . _
2003a ) . the antennae galaxies , a beginning merger of 2 large gas - rich spiral galaxies similar to the milky way and m31 feature thousands of young scs , many of them with radii and masses in the range of galactic gcs ( fritze - v. a. 1998 , 1999 , anders & fritze - v. a. _ submitted _ ) .
the hst wfpc2 images we analysed , however , only cover the innermost region of this nearby system at @xmath1 mpc .
salt s large field of view will make a great difference here .
while we can derive ages , metallicities , @xmath2 , masses and radii individually for all the clusters with precise enough photometry in at least 4 passbands , we can not say which or how many of these young massive compact scs will survive long enough to be called * g*lobular * c*lusters ( * gc*s ) in the end . an older version of essentially the same phenomenon , the already elliptical - like spiral spiral merger remnant ngc 7252 , however , still features more than 150 bright and compact scs with ages @xmath3 myr , i.e. very probably young gcs , since they have already survived the critical times during the violent relaxation process in this system .
we had predicted the possiblity that a secondary population of gcs could have formed in this system from our finding that as derived from the strength of the balmer absorption lines in the integrated spectrum of ngc 7252 the starburst on a global scale in this merger had a tremendously high sf efficiency , @xmath4 orders of magnitude higher than in normal sf regimes and well in the range of sf efficiencies required for the formation of strongly bound , long - lived clusters by hydrodynamical cluster formation models .
we also had predicted the metallicity for these clusters formed from pre - enriched gas in the spirals to be around @xmath5 ( fritze - v. a. & gerhard 1994a , b , fritze - v. a. & burkert 1995 ) . soon after hst had detected the scs ,
the metallicities of two brightest ones were spectroscopically confirmed to be around @xmath6 ( whitmore _ et al .
_ 1993 , schweizer & seitzer 1993 ) .
this shows that scs forming in abundance in gas - rich mergers are long - lived tracers of their parent galaxy s violent ( star ) formation history .
luminous elliptical galaxies are generally observed to show bimodal distributions for the optical colors ( mostly v@xmath7i ) of their gcs ( e.g. gebhardt & kissler - patig 1999 , kundu & whitmore 2001 ) , consistent with an early major merger origin of their parent galaxies but difficult to reconcile with hierarchical formation scenarios .
while the blue peak is fairly universal and consistent with old and metal - poor gcs like the ones in the milky way halo , the red peak seems variable in position and relative height . with one color only , however , it is not possible to determine age and metallicity differences between the two populations .
first optical - nir color distributions seem to show more structure than the optical ones ( kissler - patig _ et al .
_ 2002 , puzia _ et al .
_ 2002 , hempel _
2003 ) .
using galev evolutionary synthesis models for scs of various metallicities , we showed that secondary gcs from spiral spiral mergers can well explain the red @xmath8 peak of the gc color distribution in e / s0s , but also that a wide range of combinations of age and metallicity can hide within 1 optical color peak but should split up visibly in an optical - nir color like @xmath9 ( fritze - v. alvensleben 2004 ) .
gc multi - color distributions hence provide a valuable key to the formation history of their parent galaxies .
salt s extraordinary u - band sensitivity and its large field of view offer a great potential here , in particular with its future nir arm .
therefore , we have settled out to study sc formation and evolution in a wide range of nearby galaxies with salt , from normal galaxies with quiescent sf all through the strongest starbursts in massive gas - rich mergers .
our evolutionary synthesis models galev for star clusters and galaxies calculate the time evolution of cmds , integrated spectra , luminosities , colors , emission and absorption line strengths from 4 myr all through 16 gyr ( http://www.astro.physik.uni-goettingen.de/@xmath10galev ) .
they clearly show the great importance of the u - band for age and metallicity determinations of young stellar populations ( cf .
1 ) . ( b@xmath7v ) and
v@xmath7(u@xmath7v ) comparing two scs of ages 500 and 600 myr and metallicities [email protected] and 0.004 , respectively . while both scs lie completely on top of each other in the v@xmath7(b@xmath7v )
diagram , their rgbs are very neatly separated in v@xmath7(u@xmath7v).,title="fig:",width=245](b@xmath7v ) and v@xmath7(u@xmath7v ) comparing two scs of ages 500 and 600 myr and metallicities [email protected] and 0.004 , respectively . while both scs lie completely on top of each other in the v@xmath7(b@xmath7v )
diagram , their rgbs are very neatly separated in v@xmath7(u@xmath7v).,title="fig:",width=245 ] scs are easy to model , they contain one stellar generation with one well - defined metallicity . for a given initial mass function of the stars , e.g. salpeter
, the initial mass of the cluster determines its initial luminosity .
soon after its formation a sc starts losing mass and fades due to stellar evolution in a way that depends on metallicity .
the rate of fading , however , is different in different wavelength bands .
fading in all bands is fastest in early evolutionary stages and slows down later on .
in addition to this stellar evolutionary mass loss and fading , real clusters also lose stars from the tail of their maxwellian velocity distribution due to two - body relaxation , eventually enhanced by external gravitational forces .
this mass loss and fading is not included in our modelling . at a given age ,
the broad band luminosities in filters ubvrijhk determine the * s*pectral * e*nergy * d*istribution ( * sed * ) of a sc .
we have calculated a grid of @xmath12 seds for scs with metallicities in the range @xmath13 \leq + 0.4}$ ] , ages in the range 4 myr @xmath14 14 gyr , and extinction values @xmath15 using calzetti _
_ s ( 2000 ) starburst extinction law .
5 examples of seds for scs of solar metallicity , @xmath16 , and ages 8 , 60 , 200 myr , 1 and 10 gyr are shown in fig .
2 the mass of a cluster shifts the seds up and down . the strongest changes in the course of evolution are seen in the u - band relative to the longer wavelength bands .
. observed sc multi - band photometric seds can now be compared to this grid of model seds by means of a @xmath17 algorithm analysed ( anders _ et al . _
2004a ) to obtain ages , metallicities , extinction values , and masses of all individual scs including their respective @xmath18 uncertainties .
from these the age , metallicity and mass distributions of sc systems can be constructed ( anders _ et al . _
2004b ) that then give valuable clues to their parent galaxy s sf history .
artificial star cluster analyses have identified the best optical passband combinations for sc younger ( older ) than a few gyr to be ubri and ubvi ( ubvi ) . in any case , the u - band is important for ages , metallicities and extinctions of young stellar populations , while the nir is important for metallicities of old stellar populations .
a long wavelength basis and good photometric accuracy ( @xmath19 mag ) are essential in any case ( anders _ et al . _
2004b ) .
we stress that at typical photometric accuracies ( @xmath20 mag ) , broad band photometry with useful passband combinations is as powerful in disentangling ages and metallicities ( and extinction ) as is spectroscopy with typical s / n ( cf . also cardiel _ et al .
ngc 1487 is an ongoing merger of two spiral galaxies at a distance of 11.8 mpc with two pronounced tidal tails .
we are interested in its sc population as described above ( ages , masses , metallicities , extinction ) , but also in closely examining the merger - induced starburst over the entire system to assess the amount of sf that goes into cluster vs. field star formation .
10 min u- and b - band exposures were taken with salticam very early in the performance verification phase ( @xmath11 pvp ) with a point spread function ( psf ) of 1.96 arcsec fwhm and shown in fig .
an overlay of b - band surface brightness contours on the @xmath21 color map clearly reveals that the blue color peaks coincide with the highest surface brightness regions . a flux - conserving unsharp masking technique , developed by papaderos ( 1998 ) applied to the u- and b - band images reveals more than 100 compact sources fainter than 25 mag in b , a very encouraging result at this early stage and with only part of the total observing time we requested .
an improvement in the psf by about a factor 2 is expected by the time salt / salticam starts regular operations .
comparison with hst wfpc2 bvi data on the pc chip shows that with unsharp masking salt clearly resolves a wealth of sfing complexes .
the full analysis in the way described above requires additional observations in v and i and longer exposure times in u and b to reach fainter limits .
once the photometry is performed , which for this pvp object can be tied to the available hst photometry for absolute calibration , our sed analysis tool analysed will immediately return ages , metallicities , extinction , and masses for all the individual scs and star - forming complexes .
salt s large field of view and its unique u - band sensitivity make salticam an ideal instrument for multi - band photometric analyses of young and old stellar systems , not only star clusters as shown here , but also for galaxies ( see fritze - v. alvensleben _
et al . , this volume _ ) .
accurate multi - band photometry allows to derive star cluster ages , metallicities , extinctions and masses including their respective @xmath22 uncertainties ( as well as galaxy types , redshifts , star formation histories , masses and metallicities ) with accuracies comparable to those achieved in spectroscopic studies , but reaching out to much larger distances , and provide a valuable key to empirically constrain galaxy formation histories .
we greatfully acknowledge the salt astronomers and salticam instrument team for their very efficient and successful work and for their help with our initial data reduction problems .
we also greatfully acknowledge travel support from the dfg under fr 916/15 - 1 ( ufva ) and pa 1228/4 - 1 ( pp ) , without which we could not have attended this conference . + ufva cordially thanks jay gallagher for providing her the invitation to the inauguration of salt .
anders , p. , fritze - v. alvensleben , 2003 , a&a * 401 * , 1063 anders , p. , bissantz , n. , fritze v. alvensleben , u. , de grijs , r. , 2004a , mn * 347 * , 196 anders , p. , de grijs , r. , fritze v. alvensleben , u. , bissantz , n. , 2004b , mn * 347 * , 17 fritze - v. alvensleben , u. , 1998 , a&a * 336 * , 83 fritze - v. alvensleben , u. , 1999 , a&a * 342 * , l25 fritze - v. alvensleben , u. , 2004 , a&a * 414 * , 515 fritze - v. alvensleben , u. , gerhard , o. e. , 1994a , a&a * 285 * , 751 fritze - v. alvensleben , u. , gerhard , o. e. , 1994b , a&a * 285 * , 775 fritze - v. alvensleben , u. , burkert , a. , 1995 , a&a * 300 * , 58 | conventionally , cmd analyses of nearby star clusters are based on observations in 2 passbands .
they are plagued by considerable degeneracies between age , metallicity , distance ( and extinction ) that can largely be resolved by including additional passbands with u being most appropriate for young scs and i or a nir band for old globular clusters .
for star clusters that can not be resolved , integrated photometry in suitably selected passbands was shown to be as accurate as spectroscopy in independently revealing ages , metallicities , internal extinction , and photometric masses and their respective 1@xmath0 uncertainties , when analysed with a dedicated analysis tool for their * s*pectral * e*nergy * d*istributions (= * sed*s ) ( cf .
anders _
et al .
_ 2004a , b , de grijs _ et al . _
2003b ) . for external galaxies , rich star cluster populations
can thus be efficiently analysed using deep exposures in 4 suitable filters . again
, the inclusion of the u - band significantly reduces the uncertainties in the cluster parameters .
the age and metallicity distributions of star cluster systems yield valuable information about the formation history of their parent galaxies ( fritze - v. a. 2004 ) . here
, we present our galev evolutionary synthesis models for star clusters of various metallicities ( anders & fritze - v. a. 2003 ) , recently extended to include the time evolution of cmds , the dedicated sed analysis tool analysed we developed , show results on the basis of hst data , and first results from our salt pvp project on young star clusters in starburst and interacting galaxies . |
the production of soft photons in hadron collisions is governed by the low theorem @xcite that is based on very general properties of qed as a vector gauge theory . in hadron collisions , the low theorem states that soft photons are produced by the bremsstrahlung off the charged hadrons @xcite , and relates the soft photon yield to the measured hadron spectrum .
this allows to make predictions that can be tested experimentally .
surprisingly , nearly every experiment that studied the production of soft photons in high energy hadron collisions found a dramatic ( by factor of @xmath3 ) enhancement above the low theorem s predictions , see e.g. @xcite this is the long - standing puzzle of the anomalous photon production " .
many theoretical models have been proposed to explain the anomalous photon production ( for a review , see e.g. @xcite ) ; some of them are based on collective effects in produced hadronic matter @xcite , including the effects of anomalies @xcite ; other invoke the synchrotoron radiation in the qcd vacuum @xcite , new light bound states @xcite or strong coupling phenomena treated within holography @xcite however none of them explains all features of the observed phenomenon .
0.3 cm a particularly striking recent result is the measurement of the direct soft photon yield in hadronic decays of the @xmath2 boson by the delphi collaboration @xcite .
the data recorded during the operation of the lep @xmath4 collider at cern show the photon spectrum similar to the one expected from bremsstrahlung , but with a magnitude about _ four times _ higher than the prediction of the low theorem based on the measured charged hadron yields .
contrary to processes with hadronic final states , the delphi measurement @xcite of photons produced in the @xmath5 channel is in good agreement with the theoretical expectation based on muon inner bremsstrahlung so the puzzle of the anomalous soft photons seems to be specific to the production of hadronic final states .
remarkably , the soft photon yield was found @xcite to be more sensitive to the neutral hadron multiplicity than to the charged one that is expected to govern the photon bremsstrahlung .
0.3 cm the anomalous soft photons present a challenge to the foundations of theory .
moreover , this puzzle is an obstacle to using soft photons for diagnosing quark - gluon plasma , as we need to understand the background " - the mechanism of photon production in elementary collisions . to be specific , in this paper we concentrate on the direct soft photon production in hadronic decays of @xmath2 that are dominated by the fragmentation of the quark jets produced in the @xmath6 process .
the direct ( i.e. not originating from hadron decays ) soft photon emission by the fragmenting quark and antiquark within the perturbative framework was found @xcite insufficient to explain the data . since the bremsstrahlung of final state charged hadrons also does not describe the yield , one is naturally led to a possible nonperturbative mechanism of photon emission in the quark fragmentation process . because of this , it is important to re - examine the problem by putting emphasis on the possible non - perturbative qcd effects on soft photon production .
0.3 cm the quark propagating through the confining qcd vacuum pulls from the dirac sea the quark - antiquark pairs that later form hadrons ; even if these hadrons are neutral , all quarks possess electric charge and can radiate photons .
it is clear that such a mechanism involves qcd dynamics at large distances , where the coupling is strong and we have to rely on an effective theory .
large energy of the fragmenting jets suggests using a dimensionally reduced theory describing the processes developing along the jet axis .
confinement is a crucial part of the mechanism that we study , and it has to be a part of the effective theory . based on the picture of confinement through the condensation of magnetic monopoles , which results in the quasi - abelian projection @xcite , we assume that the dynamics along the longitudinal direction is abelian .
finally , to account for the hadronization process , we also need a confining theory in the presence of light fermions that severely affect the mechanism of confinement @xcite ; for review , see @xcite . based on these assumptions , in @xcite
we described the longitudinal dynamics of the jet by massless @xmath7 , known as the schwinger model @xcite , which is exactly soluble .
originally , the schwinger model was used to discuss the qualitative features of jet fragmentation in @xcite .
this model was also previously used to describe high energy processes in refs .
an important property of this theory is that the particle production is determined by the axial anomaly as we will show below .
the production of quark - antiquark pairs in the fragmentation of the quark allows interpretation in terms of topology of the gauge field and chirality of the quarks . in this section
, we will illustrate this using the example of the @xmath0 dimensionally reduced theory . as explained in the introduction , we will assume that the dynamics of pair production along the jet axis can be modeled by massless quantum electrodynamics in @xmath0 dimensions or @xmath7 : @xmath8 where @xmath9 , @xmath10 is the gauge field , @xmath11 is the field strength and @xmath12 is an external current . as in @xcite
, the leading quarks of the jets are introduced through an external current composed by the fermion and the anti - fermion moving back - to - back with equal velocities @xmath13 ( see fig.[fig : btbchrg ] ) : @xmath14 where @xmath15 and @xmath16 where @xmath17 is the jet momentum and @xmath18 gev is the time - like virtuality scale at which the pqcd dglap cascade stops , and the effects of confinement described by our effective theory begin to operate .
it is well known that the schwinger model can be solved by bosonisation @xcite .
for the vector current , the bosonization relation is @xmath19 where @xmath20 is a real scalar field ; note that with ( [ eq : bc ] ) the conservation of vector current is automatic .
it is easy to express the axial current in terms of the scalar field as well in @xmath0 dimensions , @xmath21 , therefore @xmath22 using the bosonization relations above , one can see that the original @xmath7 acquires the form of a theory describing a real massive scalar field coupled to a classical source @xmath23 that is given by @xmath24 the equation of motion for the scalar field is @xcite @xmath25 where @xmath26 and @xmath27 ; note that the coupling constant @xmath28 has dimension of mass in @xmath0 dimensions .
0.3 cm the coupling to a classical source results in particle creation .
the produced particles ( that after bosonization become the quanta of the @xmath20 field ) can be interpreted as neutral mesons produced in the fragmentation of the string stretched between the original quark and antiquark .
let us for a moment consider the ultra - relativistic limit @xmath29 in which the problem becomes somewhat simpler . due to the lorentz invariance ,
we can write the equation of motion in terms of the proper time @xmath30 and rapidity @xmath31 : @xmath32 the equation of motion in the @xmath29 limit is independent of rapidity and can be written as @xmath33 which is an ordinary differential equation in the proper time : @xmath34 where the prime denotes differentiation with respect to @xmath35 .
on the other hand , @xmath36=\sqrt{\pi}\theta(t^2-z^2)=\sqrt{\pi}\theta(\tau^2 ) ; \label{eq:}\ ] ] we therefore have to solve @xmath37 the solution can be written as @xmath38 where @xmath39 is the bessel function . the equations and show that the evolution in proper time gives rise to oscillation in the vector ( electric ) charge density @xmath40 it is this oscillation of electric charge that will be responsible for the enhancement of the soft photon yield once we introduce the coupling to the dynamical @xmath41 dimensional electromagnetic field .
0.3 cm since the oscillation of electric charge is crucial for our interpretation of the anomalous soft photon production , it is worthwhile to discuss this phenomenon in more detail .
we will now show that the oscillation of electric charge is induced by the axial anomaly in the presence of chirality imbalance .
dimensions , these are the crucial ingredients of the chiral magnetic effect @xcite . ] in @xmath42 dimensions , the helicity of a fermion is determined simply by the direction of its motion the fermion moving to the right is right - handed , and the fermion moving to the left is left - handed . for an antifermion , just like in @xmath41 dimensions , chirality and helicity have the opposite signs - so the antifermion moving to the left has right - handed chirality .
our original configuration of a fermion - antifermion pair moving back - to - back therefore has two units of chirality , see fig.1 .
0.3 cm the index theorem in two dimensions is given by @xmath43 where @xmath44 is the number of left / right movers .
we can see that a nonzero axial charge gives rise to an electric field .
this has a simple physical interpretation
the presence of electric field due to the lorentz force creates an asymmetry between the left- and right - moving charged fermions . using bosonization relations and requiring that the fields vanish at infinity
, we can find the electric field induced by the pairs created by the original external source : @xmath45 0.3 cm let us now show that the axial charge @xmath46 oscillates as a function of time .
to do this , we will use , where @xmath47 will be the sum of the electric field ( [ eq : ind ] ) induced by the pair creation and the field created by the external source : @xmath48=-g\theta(t^2-z^2 ) .
\label{eq : ees}\ ] ] the total axial charge as a function of time is thus given by @xmath49 \label{eq:}\end{aligned}\ ] ] the axial charge thus indeed oscillates with the period @xmath50 the appearance of @xmath51 of course is not surprising since it is the only scale in the theory .
however the oscillation of the axial charge is a non - trivial consequence of i ) the periodicity of the @xmath1-vacuum of the schwinger model , and ii ) non - equilibrium nature of our process .
the chiral charge of the original quark - antiquark pair with time is screened by the electric field , which then consequently decays producing additional chiral quark - antiquark pairs
because the set - up describing the separating quark and antiquark jets is far from equlibrium , the axial charge keeps oscillating around its equilibrium value of @xmath52 . because the axial anomaly in @xmath42 dimensions couples the axial and vector currents , as indicated by ( [ eq : bc ] ) and ( [ eq : acurr ] ) , the oscillations of axial charge translate into the oscillations of electric current analog of the chiral magnetic wave @xcite a gapless collective mode induced by the axial anomaly in @xmath41 dimensional hydrodynamics . ] .
we will now couple our theory to the @xmath41 electromagnetic field , and show that the fluctuations of electric current induced by the anomaly indeed source the soft photon production .
in the previous section we have considered the sources moving along the lightcone , and have demonstrated that the axial anomaly leads to the undamped axial charge oscillations with frequency @xmath51 ( the mass of the scalar meson ) .
let us now consider a more realistic case when the sources move with velocity @xmath53 given by .
we will solve equation @xmath54 and using the bosonization relation get the total electric current induced by the quark - antiquark pairs .
we then couple this current to @xmath41 electromagnetic field and compute the rate of the soft photon bremsstrahlung .
0.3 cm the general solution to the equation of motion is given by @xmath55 where @xmath56 is the retarded scalar field propagator and @xmath57 satisfies the klein - gordon equation . taking the fourier transform of @xmath58
we can solve for @xmath59 @xmath60 let us recall that the direction of @xmath61 is along the jet axis , which we choose to be the @xmath62 direction , therefore @xmath63 .
we are only interested in the contribution to which is induced by the interaction with the external source ; therefore we can write the solution in momentum space as @xmath64 this is the scalar field induced by the external source
. we will need the fourier transform of and it is given by @xmath65 we also need @xmath66 , which can be computed using the conservation equation @xmath67 : @xmath68 the induced meson field in momentum space can now be writen down as @xmath69 by taking the inverse fourier transform of and using the bosonization relation @xmath70 , we can compute the induced vector current @xmath71 .
0.3 cm so far we have been considering only the strong interaction dynamics that within a jet was modeled by a @xmath42 dimensionally reduced effective theory . since we know that all quarks possess in addition to color also the electric charge , they couple to electromagnetic fields , with a different coupling constant @xmath72 .
the electromagnetic field of course is not confined within a string ( unlike the gauge field @xmath10 that we considered so far ) and so can propagate in @xmath41 dimensions .
therefore we have to couple the total current @xmath73 of the @xmath42 quarks to the @xmath41 dimensional electromagnetic gauge field @xmath74 .
our system thus resembles a quantum wire where the charges propagate only along the wire but can radiate photons in @xmath41 dimensions .
the resulting theory has a familiar form @xmath75 the photon bremsstrahlung spectrum can now be evaluated using the standard formula @xmath76 where @xmath77 is the fourier transform of the current . in @xmath78 dimensions ,
the photon distribution will depend on the @xmath41 fourier transform of the current , therefore we make the identification @xmath79 . the contribution from the sea quarks ( i.e. the quarks produced from the vacuum by the external source )
is given by @xmath80 where @xmath81 is the fraction of the electric charge for a given quark flavor .
note that the spatial component is now in the @xmath62 direction , so in we should take @xmath82=1 .
we should also add to the contribution from valence quarks
the original quark and antiquark moving back - to - back along the @xmath62 direction with velocities @xmath13 .
their current is given by @xmath83 \nonumber \\ j_{\mathrm{val}}^3(x)&=&eq_f\delta(x)\delta(y)[v\delta(z - vt)\theta(z)+v\delta(z+vt)\theta(-z ) ] \label{eq : j03}\end{aligned}\ ] ] we now take the fourier transform @xmath84 similarly ,
@xmath85 the total current that contributes to the photon production is given by the sum of the sea and valence contributions : @xmath86 from and , we can now compute the photon spectrum .
let us consider the case of @xmath87 decay to quarks ; to do so we have to introduce the probability for @xmath87 to decay to a certain flavor of quark .
the final formula for the photon spectrum is thus given by @xmath88 where @xmath89 is the decay width of @xmath87 to quark - antiquark of flavor @xmath90 and @xmath91 is the total decay width of @xmath87 to hadrons ; the particle data group @xcite gives the values @xmath92 and @xmath93 .
0.3 cm the formula @xmath94 is the central result of our paper . without the second term in the parenthesis ,
it is just the usual formula for the bremsstrahlung radiation off the original quark and antiquark produced by the @xmath87 decay .
the second term in the parenthesis is the contribution of the quantum back - reaction of the vacuum in other words , this term represents the photons produced by the transient quark - antiquarks pairs created in the fragmentation of the string .
this term originates from the scalar propagator of the field @xmath20 which in our model is stable , so we get a spectral density of an infinitely narrow resonance
so there is a sharp resonance in photon production at @xmath95 , the frequency of the vacuum current oscillation .
it is important to note that in the soft photon limit of @xmath96 , the two terms in the parenthesis cancel each other this is in accord with the low theorem stating that the very soft photons can be produced only by the asymptotic states the charged mesons . in our case , the transient quarks and antiquarks are ultimately bound into neutral mesons and so are not allowed to contribute to the photon spectrum in the @xmath96 limit .
it is clear that our model is unrealistic in assuming the zero width of the meson @xmath20 all mesons that exist in the hadron spectrum possess non - zero width .
moreover , the mass @xmath51 in reality can not be a fixed number , as the scalar meson of our effective theory represents the entire hadron spectrum . to make our model more realistic we thus have to i ) consider a distribution in @xmath51 and ii ) to account for a finite width of the mesons . to reach the objective i ) , let us first consider the potential acting between the static fermions in the schwinger model that we use . for a fermion - antifermion pair separated by the distance @xmath97 , the potential in this model is given by @xcite @xmath98 at large distances , the potential is screened by the produced pairs , but at short distances @xmath99 the potential is linear , @xmath100 with the string tension @xmath101 . to introduce a distribution in @xmath51 ,
let us assume that the string tension fluctuates with a gaussian probability distribution @xcite @xmath102 where @xmath103 we use the mean value of the string tension @xmath104 gev / fm suggested by the lattice studies and regge phenomenology .
0.3 cm to account for a finite decay width of the meson , we write the propagator as @xmath105 where @xmath106 is an effective width . from ( [ eq : tpp1 ] ) it is clear that the soft photon production will be dominated by the longest living resonances . using the pdg @xcite values for the masses and widths of the neutral isoscalar resonances , we find the values of @xmath107 in the range of @xmath108 gev for the @xmath109 meson , and @xmath110 gev for the @xmath111 meson .
we will see that the value of @xmath106 extracted from the fit to the delphi data on soft photon production falls in this range .
0.3 cm the delphi collaboration @xcite measured the photons with transverse momenta @xmath112 mev and total energies within @xmath113 gev .
we thus substitute in and compute the total number of photons in this kinematic domain @xmath114 by integrating over the appropriate range of transverse momentum and energy . since @xmath106 is small , we can use @xmath115 and write @xmath116 we use the delta function to eliminate the integral over @xmath117 in ( [ eq : total ] ) .
we use the standard value of the string tension @xmath118 gev / fm and extract the value of the parameter @xmath119 gev by fitting to the measured experimental photon yield . in fig .
[ fig : ngvspjet ] we show the result as compared to the delphi data on the total number of photons as a function of the jet momentum @xcite .
one can see that our mechanism describes the observed enhancement reasonably well , with the fitted value of the parameter @xmath119 gev within a reasonable range expected for neutral isoscalar resonances .
we have modeled the propagation of a high energy quark through the confining qcd vacuum by using an exactly soluble @xmath0 dimensional massless abelian gauge model . while this is of course a drastic oversimplification of the problem , it is important to keep in mind that this model shares with qcd all of the ingredients involved in the mechanism that we propose : confinement , chiral symmetry breaking , axial anomaly , and the periodic @xmath1-vacuum .
moreover , at high energies the factorization of the transverse and longitudinal degrees of freedom is natural , and so the use of a dimensionally reduced theory may be justified . in this theory , we have established the phenomenon of coherent oscillations of the axial and vector ( electric ) charges coupled by the axial anomaly and induced by the propagating high energy quark .
these oscillations originate from the continuous production of quark - antiquark pairs pulled from the vacuum to screen the axial and electric charges of the external source .
we have found that soft photons provide an important signature of this mechanism , as it leads to a strong enhancement of the soft photon yield . at the cost of introducing an adjustable parameter
, our model can then describe the delphi data on the soft photon production .
0.3 cm we readily admit that the use of a @xmath0 model , and the procedure we use to compare our results to the experimental data can be questioned .
moreover , our numerical result depends on an adjustable parameter that can not be determined within our model ( even though its value appears reasonable ) . nevertheless , we hope that the described mechanism of string fragmentation may be close to the one in real @xmath41 qcd
. it would be interesting to generalize our study to a @xmath41 model as it would allow , for example , to address the effects of spin and chirality on the fragmentation of a polarized quark ( see @xcite for an attempt in this direction )
. 0.2 cm * acknowledgements . *
we thank yuri dokshitzer for stimulating this study by bringing the soft photon puzzle in @xmath120 decays to our attention .
we are grateful to joshua ilany for participating in the early stage of this work , and to edward shuryak , alexander stoffers , and ismail zahed for useful discussions .
this work was supported in part by the u.s .
department of energy under contracts no .
de - fg-88er40388 and de - ac02 - 98ch10886 .
f. e. low , phys .
rev . * 110 * , 974 ( 1958 ) .
v. n. gribov , sov .
j. nucl .
* 5 * , 280 ( 1967 ) [ yad .
fiz . * 5 * , 399 ( 1967 ) ] .
v. chliapnikov _ et al . _ [ brussels - cern - genoa - mons - nijmegen - serpukhov collaboration ] , phys .
b * 141 * , 276 ( 1984 ) .
f. botterweck _ et al . _
[ ehs - na22 collaboration ] , z. phys .
c * 51 * , 541 ( 1991 ) .
s. banerjee _ et al .
_ [ sophie / wa83 collaboration ] , phys .
b * 305 * , 182 ( 1993 ) .
a. belogianni _ et al .
_ [ wa91 collaboration ] , phys .
b * 408 * , 487 ( 1997 ) [ hep - ex/9710006 ] .
a. belogianni , w. beusch , t. j. brodbeck , d. evans , b. r. french , p. ganoti , a. jacholkowski and j. b. kinson _ et al .
_ , phys .
b * 548 * , 122 ( 2002 ) .
a. belogianni , w. beusch , t. j. brodbeck , f. s. dzheparov , b. r. french , p. ganoti , j. b. kinson and a. kirk _ et al .
_ , phys .
b * 548 * , 129 ( 2002 ) .
v. balek , n. pisutova and j. pisut , acta phys .
b * 21 * , 149 ( 1990 ) .
j. abdallah _ et al .
_ [ delphi collaboration ] , eur .
j. c * 47 * , 273 ( 2006 ) [ hep - ex/0604038 ] .
j. abdallah _ et al .
_ [ delphi collaboration ] , eur .
j. c * 67 * , 343 ( 2010 ) [ arxiv:1004.1587 [ hep - ex ] ] .
j. abdallah _ et al .
_ [ delphi collaboration ] , eur .
j. c * 57 * , 499 ( 2008 ) [ arxiv:0901.4488 [ hep - ex ] ] .
e. v. shuryak , phys .
b * 231 * , 175 ( 1989 ) .
p. lichard and l. van hove , phys .
b * 245 * , 605 ( 1990 ) ; p. lichard , phys .
d * 50 * , 6824 ( 1994 ) [ hep - ph/9812206 ] .
g. basar , d. kharzeev , d. kharzeev and v. skokov , phys .
lett . * 109 * , 202303 ( 2012 ) [ arxiv:1206.1334 [ hep - ph ] ] .
a. a. andrianov , v. a. andrianov , d. espriu and x. planells , phys .
b * 710 * , 230 ( 2012 ) [ arxiv:1201.3485 [ hep - ph ] ] .
k. fukushima and k. mameda , phys .
d * 86 * , 071501 ( 2012 ) [ arxiv:1206.3128 [ hep - ph ] ] .
k. tuchin , phys .
c * 87 * , 024912 ( 2013 ) [ arxiv:1206.0485 [ hep - ph ] ] .
h. -u .
yee , arxiv:1303.3571 [ nucl - th ] ; k. a. mamo and h. -u .
yee , arxiv:1307.8099 [ nucl - th ] .
g. w. botz , p. haberl and o. nachtmann , z. phys .
c * 67 * , 143 ( 1995 ) [ hep - ph/9410392 ] . c. -y .
wong , phys .
c * 81 * , 064903 ( 2010 ) [ arxiv:1001.1691 [ hep - ph ] ] . y. hatta and t. ueda , nucl .
b * 837 * , 22 ( 2010 ) [ arxiv:1002.3452 [ hep - ph ] ] .
g. t hooft , nucl .
* b138 * , 1 ( 1978 ) . v. n. gribov , eur .
j. c * 10 * , 91 ( 1999 ) [ hep - ph/9902279 ] .
y. l. dokshitzer and d. e. kharzeev , ann .
nucl . part .
sci . * 54 * , 487 ( 2004 ) [ hep - ph/0404216 ] .
f. loshaj and d. e. kharzeev , int .
j. mod .
e * 21 * , 1250088 ( 2012 ) [ arxiv:1111.0493 [ hep - ph ] ] .
d. e. kharzeev and f. loshaj , phys
. rev . * d 87 * , 077501 ( 2013 ) [ arxiv:1212.5857 [ hep - ph ] ]
. j. s. schwinger , phys .
* 128 * , 2425 - 2429 ( 1962 ) .
s. r. coleman , phys .
d * 11 * , 2088 ( 1975 ) .
s. mandelstam , phys .
d * 11 * , 3026 ( 1975 ) .
d. kharzeev , phys .
b * 633 * , 260 ( 2006 ) [ hep - ph/0406125 ] ; d. kharzeev and a. zhitnitsky , nucl . phys .
a * 797 * , 67 ( 2007 ) [ arxiv:0706.1026 [ hep - ph ] ] ; d. e. kharzeev , l. d. mclerran and h. j. warringa , nucl .
a * 803 * , 227 ( 2008 ) [ arxiv:0711.0950 [ hep - ph ] ] ; k. fukushima , d. e. kharzeev and h. j. warringa , phys .
d * 78 * , 074033 ( 2008 ) [ arxiv:0808.3382 [ hep - ph ] ] .
d. e. kharzeev and h. -u .
yee , phys .
d * 83 * , 085007 ( 2011 ) [ arxiv:1012.6026 [ hep - th ] ] .
j. beringer _
[ particle data group collaboration ] , phys .
d * 86 * , 010001 ( 2012 ) . | the propagation of a high energy quark disturbs the confining qcd vacuum inducing the currents in dirac sea .
since quarks possess electric charge , these induced vacuum quark currents act as a source of soft photon radiation .
this can lead to the enhancement of the soft photon production above the expectations based on the charged hadron yields and the low theorem .
we illustrate the phenomenon by using the exactly soluble @xmath0 dimensional massless abelian gauge model that shares with qcd all of the ingredients involved in this mechanism : confinement , chiral symmetry breaking , axial anomaly , and the periodic @xmath1-vacuum .
we show that the propagating quark throughout the process of hadronization induces in the vacuum charged transition currents that lead to a strong resonant enhancement of the soft photon yield ; the low theorem however remains accurate in the limit of very soft momenta .
we then construct on the basis of our result a simple phenomenological model and apply it to the soft photon production in the fragmentation of jets produced in @xmath2 decays .
we find a qualitative agreement with the recent result from the delphi collaboration . |
charged mesons formed from a quark and anti - quark can decay to a charged lepton pair when these objects annihilate via a virtual @xmath12 boson .
[ ptoellnu ] illustrates this process for the purely leptonic decay of a @xmath13 meson .
leptonic decays in the standard model .
, title="fig:",width=288]-0.02 mm similar quark - antiquark annihilations via a virtual @xmath14 to the @xmath15 final states occur for the @xmath16 , @xmath17 , @xmath18 , and @xmath19 mesons .
( charge - conjugate particles and decays are implied . )
let @xmath20 be any of these pseudoscalar mesons .
to lowest order , the decay width is @xmath21 here @xmath22 is the @xmath20 mass , @xmath23 is the @xmath24 mass , @xmath25 is the cabibbo - kobayashi - maskawa ( ckm ) matrix element between the constituent quarks @xmath26 in @xmath20 , and @xmath27 is the fermi coupling constant .
the parameter @xmath5 is the decay constant , and is related to the wave - function overlap of the quark and antiquark .
the decay @xmath28 starts with a spin-0 meson , and ends up with a left - handed neutrino or right - handed antineutrino . by angular momentum conservation , the @xmath29
must then also be left - handed or right - handed , respectively . in the @xmath30 limit
, the decay is forbidden , and can only occur as a result of the finite @xmath24 mass .
this helicity suppression is the origin of the @xmath31 dependence of the decay width .
there is a complication in measuring purely leptonic decay rates .
the process @xmath32 is not simply a radiative correction , although radiative corrections contribute .
the @xmath20 can make a transition to a virtual @xmath33 , emitting a real photon , and the @xmath33 decays into @xmath34 , avoiding helicity suppression .
the importance of this amplitude depends on the decaying particle and the detection technique . the @xmath35 rate for a heavy particle such as @xmath7 decaying into a light particle such as a muon can be larger than the width without photon emission @xcite .
on the other hand , for decays into a @xmath36 , the helicity suppression is mostly broken and these effects appear to be small .
measurements of purely leptonic decay branching fractions and lifetimes allow an experimental determination of the product @xmath37 .
if the ckm element is well known from other measurements , then @xmath5 can be well measured .
if , on the other hand , the ckm element is not well measured , having theoretical input on @xmath5 can allow a determination of the ckm element .
the importance of measuring @xmath38 depends on the particle being considered . for the @xmath7 system , @xmath6 is crucial for using measurements of @xmath39-@xmath40 mixing to extract information on the fundamental ckm parameters .
knowledge of @xmath41 is also needed , but it can not be directly measured as the @xmath42 is neutral , so the violation of the su(3 ) relation @xmath43 must be estimated theoretically .
this difficulty does not occur for @xmath44 mesons as both the @xmath13 and @xmath18 are charged , allowing the direct measurement of su(3 ) breaking and a direct comparison with theory .
for @xmath45 and @xmath18 decays , the existence of a charged higgs boson ( or any other charged object beyond the standard model ) would modify the decay rates ; however , this would not necessarily be true for the @xmath13 @xcite .
more generally , the ratio of @xmath46 to @xmath47 decays can serve as one probe of lepton universality @xcite .
as @xmath48 has been quite accurately measured in super - allowed @xmath49 decays @xcite , with a value of 0.97425(22 ) @xcite , measurements of @xmath50 yield a value for @xmath51 .
similarly , @xmath52 has been well measured in semileptonic kaon decays , so a value for @xmath53 from @xmath54 can be compared to theoretical calculations .
lattice gauge theory calculations , however , have been claimed to be very accurate in determining @xmath55 , and these have been used to predict @xmath52 @xcite .
we review current measurements , starting with the charm system .
the cleo collaboration has performed the only measurement of the branching fraction for @xmath56 @xcite .
cleo uses @xmath57 collisions at the @xmath58 resonant energy where @xmath59 pairs are copiously produced .
they fully reconstruct one of the @xmath44 s , find a candidate muon track of opposite sign to the tag , and then use kinematical constraints to infer the existence of a missing neutrino and hence the @xmath60 decay of the other @xmath44 .
they find @xmath61 .
we use the well - measured @xmath13 lifetime of 1.040(7 ) ps , and assuming @xmath62 equals @xmath63 @xcite minus higher order correction terms @xcite , we find @xmath64 .
the cleo branching fraction result then translates into a value of @xmath65 this result includes a 1% correction ( lowering ) of the rate due to the presence of the radiative @xmath66 final state based on the estimate by dobrescu and kronfeld @xcite .
before we compare this result with theoretical predictions , we discuss the @xmath18 . measurements of @xmath9 have been made by several groups and are listed in table [ tab : fds ] @xcite .
we exclude values @xcite obtained by normalizing to @xmath18 decay modes ( mentioned in the 2008 version of this review @xcite ) that are not well defined .
many measurements , for example , used the @xmath67 mode .
this decay is a subset of the @xmath68 channel which has interferences from other modes populating the @xmath69 mass region near the @xmath70 , the most prominent of which is the @xmath71 .
thus the extraction of effective @xmath67 rate is sensitive to the mass resolution of the experiment and the cuts used to define the @xmath70 mass region @xcite@xcite .
the cleo and belle @xmath72 results rely on fully reconstructing all the final state particles except for the neutrino and using a missing - mass technique to infer the existence of the neutrino .
cleo uses @xmath73 collisions at 4170 mev , while belle uses @xmath74 collisions at energies near the @xmath75 .
when selecting the @xmath76 and @xmath77 decay modes , cleo uses both calculation of the missing - mass and the fact that there should be no extra energy in the event beyond that deposited by the measured tagged @xmath78 and the @xmath79 decay products .
the @xmath80 mode , however , uses only extra energy .
babar measures @xmath81 using the @xmath82 mode . here
the @xmath78 tag is formed similarly to belle by finding events with a @xmath44 a @xmath11 and pions opposite a single positron and little extra energy .
then the analysis is performed selecting modes with a @xmath83 consistent with arising from the decay of a @xmath18 .
( the fourth error in table [ tab : fds ] on their measurement reflects the uncertainty due to the pdg value of @xmath84 . ) .experimental results for @xmath85 , @xmath86 , and @xmath9 .
numbers for @xmath9 have been extracted using updated values for masses and @xmath87 ( see text ) ; radiative corrections have been included .
common systematic errors in the cleo results have been taken into account .
[ tab : fds ] [ cols="<,<,^,^,^",options="header " , ] there are large backgrounds under the signals in all cases .
the systematic errors are also quite large , on the order of 20% .
thus , the significances are not that large .
belle quotes 3.5@xmath88 and 3.8@xmath88 for their hadronic and semileptonic tags , respectively , while babar quotes 2.8@xmath88 for their combined result .
we note that the four central values are remarkably close to the average considering the large errors on on all the measurements
. more accuracy would be useful to investigate the effects of new physics . here
the effect of a charged higgs is different as it can either increase or decrease the expected sm branching ratio . the factor @xmath89 in the 2hdm that multiplies the right side of eq .
[ equ_rate ] is given in terms of the @xmath7 meson mass , @xmath90 , by @xcite @xmath91 we can derive limits in the 2hdm @xmath92@xmath93 plane .
again , we need to know the sm prediction of this decay rate .
we ascertain this value using eq .
[ equ_rate ] . here
theory provides a value of @xmath94 mev @xcite .
we also need a value for @xmath95 . here
significant differences arise between using inclusive charmless semileptonic decays and the exclusive decay @xmath96 @xcite .
we find that the inclusive decays give rise to a value of @xmath97 , while the @xmath98 measurements yield @xmath99 .
taking an average over inclusive and exclusive determinations , and enlarging the error using the pdg prescription because the results differ , we find @xmath100 , where the error is dominantly theoretical .
we thus arrive at the sm prediction for the @xmath101 branching fraction of @xmath102 . taking the ratio of the experimental value to the predicted branching ratio at its 90% c.l .
_ upper _ limit and using eq .
[ eq : hou ] with @xmath103 set to zero , we find that we can limit @xmath104 gev .
the 90% c.l .
_ lower _ limit also permits us to exclude the region 3.8 gev @xmath105 gev @xcite .
considering the large uncertainties on @xmath106 and the branching ratio measurements , this should be taken more as indication of what the data can eventually tell us when and if the situation improves .
we now discuss the determination of charged pion and kaon decay constants .
the sum of branching fractions for @xmath107 and @xmath108 is 99.98770(4)% .
the two modes are difficult to separate experimentally , so we use this sum , with eq .
[ equ_rate ] modified to include photon emission and radiative corrections @xcite . the branching fraction together with the lifetime 26.033(5 ) ns gives @xmath109 the first error is due to the error on @xmath48 , 0.97425(22 ) @xcite ; the second is due to the higher - order corrections , and is much larger . similarly , the sum of branching fractions for @xmath110 and @xmath111 is 63.55(11)% , and the lifetime is 12.3840(193 ) ns @xcite .
measurements of semileptonic kaon decays provide a value for the product @xmath112 , where @xmath113 is the form - factor at zero four - momentum transfer between the initial state kaon and the final state pion .
we use a value for @xmath112 of 0.21664(48 ) @xcite .
the @xmath113 must be determined theoretically .
we follow blucher and marciano @xcite in using the lattice calculation @xmath114 @xcite , since it appears to be more precise than the classic leutwyler - roos calculation @xmath115 @xcite .
the result is @xmath116 , which is consistent with the hyperon decay value of @xmath117 @xcite .
we derive @xmath118 the first error is due to the error on @xmath119 ; the second is due to the ckm factor @xmath52 , and the third is due to the higher - order corrections .
the largest source of error in these corrections depends on the qcd part , which is based on one calculation in the large @xmath120 framework .
we have doubled the quoted error here ; this would probably be unnecessary if other calculations were to come to similar conclusions .
a large part of the additional uncertainty vanishes in the ratio of the @xmath121 and @xmath122 decay constants , which is @xmath123 these measurements have been used in conjunction with calculations of @xmath124 in order to find a value for @xmath125 .
two recent lattice predictions of @xmath124 are @xmath126 @xcite and @xmath127 @xcite . together with the precisely measured @xmath48 , this gives an independent measure of @xmath52 @xcite .
most predictions for the rate for @xmath128 are in the range of 1 to 20 times the rate for @xmath129 .
see g. burdman , t. goldman , and d. wyler , phys .
d * 51 * , 111 ( 1995 ) ; p. colangelo , f. de fazio , and g. nardulli , phys .
b * 372 * , 331 ( 1996 ) ; _ ibid .
_ , * 386 * , 328 ( 1996 ) ; a. khodjamirian , g. stoll , and d. wyler , phys .
b * 358 * , 129 ( 1995 ) ; g. eilam , i. halperin , and r. mendel , phys .
b * 361 * , 137 ( 1995 ) ; d. atwood , g. eilam , and a. soni , mod .
a * 11 * , 1061 ( 1996 ) ; c.q .
geng and c.c .
lih , phys.rev .
d * 57 * , 5697 ( 1998 ) and mod .
a * 15 * , 2087 ( 2000 ) ; g.p .
korchemsky , d. pirjol , and t.m .
yang , phys .
d * 61 * , 114510 ( 2000 ) ; c.w .
hwang , eur .
j. c * 46 * , 379 ( 2006 ) .
b. a. dobrescu and a. s. kronfeld , phys .
lett . * 100 * , 241802 ( 2008 ) .
j. p. alexander _
et al . _ , ( cleo collab . ) , phys .
d * 79 * , 052001 ( 2009 ) ; m. artuso _
et al . _ , ( cleo collab . ) , phys .
* 99 * , 071802 ( 2007 ) .
j. dingfelder , `` hot topics from babar : leptonic and semileptonic @xmath7 and @xmath44 decays , '' presented at moriond qcd and high energy interactions , la thuile , italy march 2010 ; j. p. lee
_ et al._(babar collab . ) .
we have not included a babar result : @xmath132 and @xmath133 mev based on @xmath134 .
these measurements determined the ratio of the leptonic decay to a hadronic decay , usually @xmath135 .
see b. aubert _ et al . _ ( babar collab . ) , phys .
lett . * 98 * 141801 ( 2007 ) .
e. follana _
_ , ( hpqcd and ukqcd collabs . ) , phys . rev . lett .
* 100 * , 062002 ( 2008 ) .
the statistical error is given as 0.6% , with the other systematic errors added in quadrature yielding the 3 mev total error .
we use the updated numbers from a. bazavov _
et al . _
( fermilab lattice and milc collaborations ) , pos lattice 2009 249 , ( 2009 ) , an update that increases both @xmath8 and @xmath9 by 10 mev and 11 mev , respectively , from the values originally given in c. bernard _ et al .
_ , ( 2009 ) ; see also c. aubin _ et al .
_ , phys .
lett . * 95 * , 122002 ( 2005 ) .
i. dorner , s. fajfer , j. f. kamenik , and n. konik , phys .
lett.b * 682 * , 67 ( 2009 ) ; s. fajfer , i. dorner , j. f. kamenik , and n. konik , arxiv:0910.5407 , presented at low energy constraints on extensions of the standard model , kazimierz , poland , 2327 july 2009 . c. bernard _
et al . _
( fermilab , milc collabs . ) , pos lattice 2008 278 , ( 2008 ) calculate @xmath136 mev , while e. gamiz _
et al . _
( hpqcd collab . )
d * 80 * , 014503 ( 2009 ) find @xmath137 mev .
we average the two results assuming the errors are fully correlated .
in supersymmetric models there may be corrections which weaken our bound ; a slightly different formula for @xmath89 is given by isidori and paradisi , phys .
b * 639 * , 499 ( 2006 ) that depends on a model - dependent value of a parameter @xmath103 . | we review the physics of purely leptonic decays of @xmath0 , @xmath1 , @xmath2 , @xmath3 , and @xmath4 pseudoscalar mesons .
the measured decay rates are related to the product of the relevant weak - interaction - based ckm matrix element of the constituent quarks and a strong interaction parameter related to the overlap of the quark and anti - quark wave - functions in the meson , called the decay constant @xmath5 .
the interplay between theory and experiment is different for each particle .
theoretical predictions of @xmath6 that are needed in the @xmath7 sector can be tested by measuring @xmath8 and @xmath9 in the charm sector . currently , these tests are unsatisfactory .
the lighter @xmath10 and @xmath11 mesons provide stringent comparisons between experiment and theory due to the accuracy of both the measurements and the theoretical predictions . an abridged version of this review was prepared for the particle data group s 2010 edition @xcite . |
interstellar clouds are thought to be turbulent and supersonic .
their mach numbers range from a 1 ( @xmath2 70008000 k , h i clouds ) through 10 ( @xmath3 10 k , molecular clouds forming low - mass stars ) to 50 ( @xmath3 1050 k , molecular clouds forming high - mass stars ) .
since supersonic turbulent motions carry mass and produce large - amplitude density fluctuations , turbulent fragmentation is expected to occur in the interstellar medium .
a useful tool for describing the overall structure of interstellar clouds is the scalar virial theorem ( vt ) .
it is obtained by dotting the momentum equation by the position vector and integrating over the volume of interest ( [ sec : vt ] ) .
being directly derived from the momentum equation , the vt always holds for any parcel of fluid .
virial equilibrium ( ve ) is a restrictive condition of the vt , and it is defined by the condition that the parcel of fluid under study has a second time derivative of the moment of inertia equal to zero .
it has been invoked extensively to analyze the stability of interstellar clouds . however , since in a trans- or super - sonic turbulent interstellar medium , clouds should be redistributing their mass as a consequence of their own turbulent motions , it already seems difficult to achieve ve in a supersonic , turbulent cloud .
other simplifications , such as the assumption that the cloud is isolated , or that the surface terms in the vt are negligible , are frequently made in many ( if not in most ) astrophysical studies of the vt .
those simplifications have been thought to be applicable to molecular clouds and their substructure for nearly three decades ( e.g. , the textbooks by * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) , maybe as a consequence of the old idea that in the ism , `` all forces are in balance and the medium is motionless , with no net acceleration '' ( * ? ? ?
11 ) , in which `` the observational evidence '' seemed to be consistent with the expectation that interstellar `` clouds tend toward pressure equilibrium '' @xcite .
the possible inapplicability of these assumptions has been mentioned in passing in some previous papers ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) but no attention has been paid in general to its implications .
thus , in the present paper i discuss in detail the applicability of those assumptions for molecular clouds and their cores . in [ sec : vt ] i write explicitly the vt for fluids in its lagrangian and eulerian forms . in
[ sec : myths ] i discuss the six more common assumptions of the vt and their validity in a turbulent environment . in
[ sec : reloading ] i explain why even though clouds are not in ve , they appear to be in energy equipartition , and argue that asymmetries in the line profiles are the evidence for clouds out of ve . finally , in [ sec : conclusions ] i draw the main conclusions .
the virial theorem can be derived from the momentum equation , by dotting it by the position vector @xmath4 and integrating it over the volume of interest .
althoughit is usually written in its lagrangian form , i.e. , by following the mass ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , it can be also obtained in its eulerian form , i.e. , by fixing the volume in space ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) , obtaining @xmath5 @xmath6 where @xmath7 is the moment of inertia of the cloud ( subindexes @xmath8 and @xmath9 in eqs .
[ [ eq : evt ] ] and [ [ eq : lvt ] ] stand for eulerian and lagrangian , respectively ) , @xmath10 and @xmath11 are the kinetic energy of the cloud and the kinetic stresses evaluated at the surface of the cloud , respectively , @xmath12 is the internal energy , @xmath13 is the pressure surface term , @xmath14 is the magnetic energy , @xmath15 is the magnetic stress at the surface of the cloud , @xmath16 is the gravitational term , with @xmath17 being the gravitational potential , and @xmath18 is the flux of moment of inertia through the surface of the cloud . in the previous equations , @xmath19 , @xmath20 , @xmath21 , @xmath22 , and @xmath23 are the density , the @xmath24 component of the velocity @xmath25 , the @xmath24 component of the magnetic field @xmath26 , the pressure , and a unitary vector perpendicular to the surface @xmath27 that surrounds the volume @xmath28 , over which the integrals are performed , respectively . in the notation above it is used the einstein convention , where repeated indexes are summed .
various assumptions for the terms involved in eqs .
( [ eq : evt ] ) and ( [ eq : lvt ] ) have been made in the literature .
some of them , indeed , have been converted into myths , since their applicability to interstellar clouds not only is not demonstrated , but it is not even questioned , either in textbooks , or research papers .
[ sec : support ] it is almost universally considered in the literature that the turbulent ( or kinetic ) energy , @xmath29 , provides support to clouds against collapse .
while this is true for a system of particles , and partially valid if the kinetic energy is in the form of large - scale expansion and/or rotation , it is by no means certain that all the kinetic energy available will help against collapse in a system where turbulent fragmentation can occur , as discussed below .
this idea has its origins in @xcite , who proposed that in the analysis of the gravitational instability the turbulent velocity field should be included . in his description ,
an effective sound speed is introduced , given by @xmath30 where @xmath31 is the sound speed , and @xmath32 is the velocity dispersion of the turbulent motions ( see , e.g. , * ? ? ?
* ; * ? ? ?
* for a review ) .
this description is valid only if ( a ) turbulent motions are confined to scales much smaller than the size of the cloud @xcite , and ( b ) such motions do not produce new , smaller - scale jeans - unstable density enhancements .
the first hypothesis disregards one of the main features of turbulent flows in general ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and of interstellar clouds in particular @xcite , namely , that the largest velocities occur at the largest scales .
an attempt to include this fact has been proposed by @xcite , who suggested including the value of the rms velocity dispersion _ at each scale _ @xmath33 , i.e. , @xmath34 where @xmath35 the wavenumber corresponding to the scale @xmath36 , and is given by the energy spectrum @xmath37 as @xmath38 where @xmath39 is a constant and @xmath40 is the spectral index .
the second condition , i.e. , that turbulent motions do not produce jeans - unstable density enhancements , has the underlying complication that motions at scales larger than @xmath41 will be very anisotropic with respect to structures of size @xmath36 . those modes will produce shear ( through vortical modes ) or compressions ( through compressible modes ) to the structures .
compressions in particular reduce the local jeans mass @xcite , and can induce local collapse .
thus , a fraction of the turbulent kinetic energy is involved in promoting collapse , rather than opposing it . by decomposing the velocity field in its solenoidal and compressible components , the kinetic energy modes that provide support to the clouds are those having divergence larger or equal to zero ,
@xmath42 this includes the solenoidal modes ( @xmath43 ) , and the expansional component of the compressible modes ( @xmath44 ) . in other words , the precise result of collapse or support must then reflect the balance between all the agents that favor collapse against those agents that provide support . in the first group , not only the gravitational energy should be included , but also the kinetic energy involved in the compressible modes ( @xmath45 ) , versus the kinetic energy involved in the expansional and rotational modes ( @xmath46 ) .
it is frequently found in the literature that the surface terms are neglected altogether , especially in observational work @xcite , mainly because there is not a direct way of measuring them observationally , although it is also a common practice in theoretical studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and textbooks ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this assumption is based on the idea that self - gravitating clouds may be considered isolated because then their internal energies dominate the dynamics .
the most notable exception is the thermal pressure surface term , which is frequently invoked for `` pressure confinement '' ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
although some works have considered , by analogy , the possibility of turbulent pressure confinement by means of the term @xmath47 ( e.g. , * ? ? ?
* ) , such confinement of a cloud is difficult to achieve because the large - scale turbulent motions are anisotropic and will in general distort or even disrupt the cloud @xcite .
although we can not measure the surface terms from observations , the possibility that they are as important as their corresponding volumetric terms suggest to investigate into numerical simulations of the interstellar medium .
in fact , @xcite found that for an ensemble of clouds in two - dimensional simulations of the interstellar medium at a kiloparsec scale , the surface terms have magnitudes as large as those of the volumetric ones ( for three - dimensional simulations , see also * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this result suggests that , on one hand , either surface and volumetric terms are of comparable importance in shaping and supporting the clouds . on the other hand
, it suggests that clouds must be interchanging mass , momentum and energy with the surrounding medium .
in such an environment , the meaning of thermal or ram pressure confinement is not clear , since motions at all scales must morph and deform the cloud .
the gravitational term entering the vt is written as @xmath48 splitting - up the gravitational potential as the contribution from the cloud itself ( @xmath49 ) , plus the contribution from the outside ( @xmath50 ) , @xmath51 the gravitational term can be written as : @xmath52 where @xmath53 is the gravitational energy of the cloud alone , since the volume of integration of @xmath49 and the volume @xmath28 of the integral coincide ( e.g. , * ? ? ?
the second term in the right - hand side is usually either implicitly or explicitly assumed to be negligible compared to the first term , and the gravitational term @xmath54 is then assumed to equal the gravitational energy @xmath55 .
this is valid only if the distribution of mass is spheroidal , or if the medium outside the cloud is tenuous such that its contribution to the potential is negligible .
however , clouds are more similar to irregular fractals with arbitrary shapes ( frequently filamentary ) than to spheroids ( e.g. , * ? ? ?
* ) , and the contribution from the external gravitational field may not be negligible , giving rise to tidal torques .
although up to now there is no observational estimation of the contribution of the external mass to the gravitational term for any interstellar cloud , mass estimates for for h i `` envelopes '' around molecular clouds are of the same order of magnitude than the mass of the molecular clouds themselves ( e.g. , * ? ? ?
* ; * ? ? ?
thus , it is not difficult to realize that the contribution of the second term in the right - hand side of eq .
( [ eq : eg1_eg2 ] ) to the gravitational term @xmath54 can be of the same order of magnitude as the gravitational energy @xmath55 .
similar arguments can be made for the interiors of molecular clouds : even though we can approximate their shapes as ( triaxial ) spheroids @xcite , embedded molecular cloud cores are subject not only to their own self - gravity , but also to the tidal forces from their parental molecular cloud .
thus , it is not clear that the tidal forces represented in the second term of eq .
( [ eq : eg1_eg2 ] ) will be negligible , and the assumption that the gravitational term equals the gravitational energy of the cloud seems unjustified .
what is the meaning and the effect of the second term of eq .
( [ eq : eg1_eg2 ] ) on the energy budget of the clouds ?
it can be seen that it involves the gradient of the external potential .
for non - symmetrical distributions of mass , this term is out of balance even if the distribution of mass inside the cloud is symmetric .
thus , this term represents the tidal forces over the mass contained in the volume @xmath28 , and it can be split up into three terms : @xmath56 using the gauss theorem , the first term on the right - hand side of eq .
( [ eq : w - descomposicion ] ) can be interpreted as the gravitational pressure evaluated at the boundary of the cloud , @xmath57 the second term , by similarity with the gravitational energy , can be interpreted as three times the work done to assemble the density distribution of the cloud against the external mass , @xmath58 finally , the last term in eq ( [ eq : w ] ) , @xmath59 involves the gradient of the density field inside the cloud .
although there is no clear interpretation of this term , it is worthwhile noting that its contribution is null for a homogeneous distribution of mass inside the volume @xmath28 of integration . from the numerical point of view
, several studies @xcite have found that the gravitational term can be negative or positive . if negative
, it will be a confining agent . if positive , its overall action will be to contribute to the disruption of the cloud / core .
it is frequently argued in the literature that if @xmath61 , the cloud must be expanding , while @xmath62 implies that the cloud is contracting .
equilibrium is assumed to occur when @xmath63 .
this idea rests on the fact that the gravitational energy @xmath55 has a negative sign , and it is a confining agent , while the internal energies ( thermal , kinetic , or magnetic ) are positive , and they are assumed to act as supporting agents against collapse . by neglecting the other terms , it can be assumed that if the gravitational energy is larger than the sum of the internal energies , the sign of the right - hand side of eq .
( [ eq : lvt ] ) is negative .
physically , if gravity wins , the cloud collapses .
although energetically this is true , it is not hard to find an example in which an expanding cloud has a negative second time derivative of the moment of inertia .
assume a sphere with constant density and fixed mass @xmath64 .
its moment of inertia is @xmath65 if its size varies with time , for instance , as a power law , @xmath66 its second time derivative is given by @xmath67 which is negative if @xmath68 , even though it is expanding . in general , @xmath60 has been treated in the literature as if it were @xmath69 .
the definition of virial equilibrium is that the left - hand side of the lagrangian virial theorem ( eq . [ [ eq : lvt ] ] ) equals zero : @xmath70 ( see , e.g. , * ? ? ?
* ) . although there are some observational papers showing molecular clouds out of ve ( e.g. , * ? ?
* ; * ? ? ?
* ) , it is frequently encountered in the literature the ve assumption for mcs and their substructure ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as mentioned in the introduction , this statement is based on the ( old ) idea that all the forces ( in the ism ) should be in balance , and the medium should have no net acceleration .
however , both observational @xcite , and numerical @xcite studies have found that the ism is not in strict pressure balance , but exhibits strong pressure fluctuations , and in fact the turbulent ram pressure is significantly larger than the thermal one ( e.g. , * ? ? ?
from an observational point of view , it can not be demonstrated that clouds are in ve because neither the detailed three - dimensional density structure of molecular clouds , nor the time derivatives for interstellar clouds can be measured observationally . from a theoretical point of view , the ve assumption has strong implications : either supersonically turbulent clouds are not redistributing their mass inside , or the way in which the time - derivative of the moment of inertia ( @xmath69 ) vary ( which can be interpreted as the time - variation of their mass redistribution ) is constant .
both statements seem implausible in a highly dynamical , nonlinear ism .
@xcite , for instance , have recognized the difficulty of achieving ve in a turbulent medium , because turbulent motions carry fluid elements , redistributing their mass . from the numerical point of view , fortunately , it is possible to calculate all the terms of the vt for clouds , since we know all the variables involved in the numerical simulations . in order to test the virial theorem , ( * ? ? ?
* see also @xcite ) calculated all the integrals in the evt for clouds in numerical simulations of the interstellar medium at a kiloparsec scale by @xcite .
they found that the second time derivative of the moment of inertia never goes to zero , suggesting that clouds must be transient entities .
@xcite has recognized the difficulty to achieving ve for actual mcs .
he proposed two possibilities for assuming clouds in ve : the first one is to average in time the second time derivative of the moment of inertia .
he suggests that @xmath71 if the averaging time considered is much larger than the dynamical timescale of the cloud , i.e. , if @xmath72 .
the second one is that for an ensemble of clouds , some of them may have positive values of @xmath73 , and some others will have negative values , so that ve holds for the ensemble .
regarding the first assumption , there is a hidden assumption behind it : the cloud maintains its identity during several dynamical timescales .
a similar assumption is made by @xcite . ] .
however , from the numerical point of view , simulations of the ism suggest that the clouds are continually morphing and exchanging mass , energy and momentum with their surroundings , so that over @xmath74 they have changed significantly .
on the other hand , observational evidence @xcite also suggests that the lifetimes of the clouds are not significantly larger than their own dynamical times ( i.e. , they are transient ) . concerning the second assumption , the analysis by ( * ? ? ? * see also @xcite ) shows that the moment of inertia spans up to 7 orders of magnitude ( in absolute value ) between the largest clouds and the smallest .
thus , it is not clear that an average for such large scatter will be representative of the actual dynamics of the clouds . in other words , even if @xmath60 did average out to zero for the ensemble , this does not alter the fact that , individually , clouds are not in ve .
the mean density - size and velocity dispersion - size relations first discussed by @xcite are thought to be an observational demonstration of virial equilibrium
. however , as @xcite recognize , they are actually only compatible with equipartition .
for instance , assuming equipartition between kinetic ( turbulent ) and gravitational energy , @xmath75 thus , if there is a mean density - size power - law relationship @xmath76 , then , there should be a velocity dispersion - size relationship of the form @xmath77 with @xmath78 several caveats must be mentioned at this point .
first of all , this derivation does not mean ve , but energy equipartition , since the only assumption was @xmath79 ( see eq . [ [ eq : equipartition ] ] ) .
a similar result can be found if the equipartition assumed is valid between the gravitational and magnetic energy . in this case , @xmath80 is proportional to the alfvn speed ( see , e.g. , * ? ? ?
second , as discussed by @xcite , in the particular case of @xmath81 , @xmath82 .
however , the pair @xmath83 , @xmath84 is not unique . any pair of values satisfying the equation ( [ eq : beta - alpha ] ) will be consistent with energy equipartition ( again , not ve ) . finally , it is convenient to recall that the validity of larson s relations has been called into question , especially the mean density - size relationship @xcite .
it seems that this relationship is more a consequence of the observational process , in which the dynamical range of the observations is limited below by the minimum sensitivity of the telescopes , and above by saturation of the detectors , optically thick effects , and depletion .
as it is discussed , there are at least six common assumptions related to the virial theorem which seem to be unjustified for a turbulent ism . some questions , such as why virial masses are good order - of - magnitude estimations of the actual mass , or why in principle
, the sub- or super - criticality of a molecular cloud core is a good estimation of the dynamical state of such a core , are still valid to ask .
the answer is that those estimations are based on energy equipartition , not on ve ( i.e. , @xmath85 does not mean that @xmath86 is negligible when it is compared to the left - hand side of eq . [ [ eq : lvt ] ] ) .
molecular clouds seem to be in approximated equipartition between self - gravity , kinetic , and magnetic energy ( see , e.g. , * ? ? ?
* ) although the super - alfvnic nature of molecular clouds is still a matter of debate , ( e.g. , * ? ? ?
* see also @xcite for an observational review ; @xcite , for a theoretical one ) .
the question thus , is why mcs seem to be in energy equipartition ?
the answer is related to what we identify as a cloud , and to observational limitations : in the first case , clouds with a substantial excess of internal energy will rapidly expand and merge with the more diffuse medium .
this is the case of an h ii region or sn explosion expanding within its parental molecular cloud .
the amount of energy provided by those events is enough to disrupt their parental environment within a few myr ( e.g. , * ? ? ?
the excess of internal energy is mostly in the form of uv photons , which rapidly ionize the molecular gas , reducing even more the life timescale of what it is identified as the parental molecular cloud ( see , e.g. , * ? ? ?
* ) . in the second case ,
if the cloud has an excess of gravitational energy , its free - fall time velocity is at most a factor of @xmath87 the velocity dispersion needed for equipartition .
this means that the difference between a free - fall collapsing cloud and one in equilibrium is just a factor of two in the energy , and a factor of @xmath87 in velocity .
both systems will be , in principle , in order - of - magnitude energy equipartition .
note , furthermore , that in the free - fall collapsing case all the kinetic energy will be observationally identified as energy for support , where it is not actually providing any support .
this re - enforces the need to distinguish between the compressible , expansive and vortical modes of the kinetic energy ( see [ sec : support ] ) .
as discussed above , not all the terms entering the vt can be measured observationally .
however , all of them can be measured from numerical simulations .
the work by @xcite for turbulent realizations of mcs shows that all the terms entering the vt are of similar importance .
note that the same problems related to the identification of an expanding cloud are applicable to the numerical work . in this case , what we identify as a cloud are the density enhancements . by definition ,
vacuumed regions ( due either to modeled stellar activity , or by the turbulence itself ) are no longer considered clouds . as for the collapsing case
, strongly self - gravitating clouds will develop a velocity field that is a factor of @xmath87 the value of energy equipartition , if only gravity is acting , and of the order of magnitude if the turbulence is forced .
thus , order of magnitude equipartition between magnetic , kinetic , and gravitational energy is valid to assume to predict whether the clouds should be supercritical or subcritical @xcite , or to understand why subcritical cores are unlikely to survive long - lifetimes within mcs @xcite .
however , the order - of - magnitude coincidence of the involved energies does not mean that clouds are in equilibrium at all .
it should be stressed that the difference is not just semantic ( virial mass vs energy - equipartition mass ) .
the difference is conceptual : turbulent clouds can not be in equilibrium because turbulent fragmentation takes them out of equilibrium , and thus they do not last long . in order to achieve equilibrium , ( even if it is a time - averaged equilibrium , @xmath71 , as suggested by @xcite ) , it is necessary to allow the cloud to live for , at least , several crossing times . to give an example , the taurus molecular cloud has a size of @xmath88 pc , and a velocity dispersion of @xmath89 .
its lateral crossing time is @xmath90 . in order to achieve equilibrium
, taurus has to live 20 - 30 myr without a serious distortion or modification .
such a condition seems difficult to achieve if one realizes that taurus has a mach number of the order of 10 , and the associated h i gas has mach numbers of 20 - 30 in respect to the internal co gas ( see , e.g. , the velocity - position diagrams of h i and co in * ? ? ?
the fact that turbulent clouds can not be in ve and can not last several crossing times is in clear contradiction with @xcite , who made theoretical arguments that favor star formation occurring during several dynamical timescales .
however , in their theoretical derivation they applied the star formation rate per free - fall time expression given by eq .
( 30 ) of @xcite to a clump in hydrostatic equilibrium .
as @xcite point out , their eq .
( 30 ) is valid only for large mach numbers ( 20 - 40 ) , while by definition , a clump in hydrostatic equilibrium is subsonic . thus , the theoretical arguments in favor of slow star formation are wrong another point to be stressed concerns line profiles of h i , co and its isotopes , and even higher density tracers ( cs , nh@xmath91 , etc . ) .
if interstellar clouds and/or their cores are not in virial equilibrium , they must be distorted and disrupted within a dynamical timescale
. how will their line - profiles look , and how will they look if they were in equilibrium ?
certainly , if turbulence were microscopic ( necessary to achieve equilibrium ) , line profiles should be symmetric .
the fact that h i and co clouds ( e.g. , * ? ? ?
* ; * ? ? ?
* ) as well as molecular cloud cores ( e.g. , * ? ? ?
* ) exhibit non - symmetric line profiles is probably the best and the only evidence that the term @xmath86 is different from zero ) .
this system , however , is not in ve , as the definition of ve , eq .
( [ eq : ve ] ) involves the _ lagrangian _ frame of reference ( note that the difference between the eulerian and lagrangian vt is @xmath92 , as pointed out by * ? ? ? * ) . ] .
even a small degree of asymmetry in the line profiles suggests that large - scale motions are present in the observed system , which is a natural consequence of turbulence being a multiscale phenomenon .
this does not mean that such a system is collapsing or expanding as a whole , as formerly suggested by @xcite .
instead , it means that turbulent large - scale motions are present in the system , which should be evolving within a dynamical timescale . in this context
, it should be stressed that clouds presenting large - scale motions and evolving to form stars rapidly not necessarily will have a high star formation efficiency , as is the common belief since @xcite .
although former models of quasi - static evolution of molecular clouds were proposed to reduce the star formation efficiency , the turbulent models producing local collapse rapidly , have small efficiency because gravo - turbulent fragmentation involves only a small fraction of the mass of the system in collapsing regions @xcite . in numerical models , when energy feedback from stars is included , the cloud is blowed out rapidly ( see , e.g. , * ? ? ?
however , it should be recognized that a detailed quantification of the star formation efficiency in turbulent simulations with open boundary conditions and stellar feedback is needed .
the present contribution has discussed the applicability of the six more common assumptions on the virial theorem . specifically , 1 . it was shown that a decomposition of the velocity field into its vortical and compressible modes is necessary , since only modes satisfying the condition @xmath44 provide support , while modes satisfying @xmath93 foment collapse .
it was argued that for a supersonic , turbulent ism , surface terms should not be neglected .
it was shown that the gravitational term can be decomposed into a contribution from the cloud itself , and a contribution from the outside .
the first part is the well - known gravitational energy .
the second part is the sum of three terms : the gravitational pressure evaluated at the surface of the cloud plus three times the work done against the external mass to assemble the density distribution of the cloud , plus a term that depends on the gradient of the density distribution of the cloud .
these represent the tidal forces due to an external potential , as could be the case of a dense core within a giant molecular cloud , or a giant molecular cloud close to a spiral arm .
it is argued that this contribution can be as important as the self - gravitational energy .
4 . using a simple counter - example
, it was shown that the sign of the second - time derivative of the moment of inertia does not determine whether the cloud is contracting or expanding .
an expanding cloud may very well satisfy the condition @xmath0 , and a contracting one may satisfy @xmath1 , contrary to the common belief . in other words , @xmath60 has been treated in the literature as if it were @xmath69 .
it was argued that interstellar clouds are not likely to satisfy the virial equilibrium ( ve ) condition @xmath63 .
it was shown that larson s ( 1981 ) relations are not observational proof for clouds being in ve .
clouds seem to be in energy equipartition because of either observational limitations , as well as because of the intrinsic definition of a cloud .
turbulent fragmentation plays a crucial role for the inapplicability of the vt to interstellar clouds , since it will induce a flux of mass , moment and energy between the clouds and their environment , and will favor local collapse while disrupting the clouds within a dynamical timescale .
the common assumptions discussed in the present contribution drive our understanding of the dynamical state of the interstellar clouds toward a picture that favors a static ism .
however , they are highly difficult to fulfill if the ism is highly turbulent , as it was found to be many years ago ( e.g. , * ? ? ?
inferences of the star formation efficiency for supersonic ( mach numbers @xmath94 ) clouds in virial equilibrium living several dynamical timescales ( e.g. , * ? ? ?
* ; * ? ? ?
* ) should be taken with caution .
the lack of observational evidence for clouds being in ve , and the identificaion of asymmetrical line profiles observed toward interstellar clouds using different tracers are the best evidence of clouds being out - of - equilibrium systems .
these facts lead us to the conclusion that clouds should be transient structures , which exchange mass , momentum and energy with their environment .
this is precisely the opposite point of view to the old one in which clouds should be at rest , which still is present in textbooks and papers , but it is consistent with recent observationally - based works that favor a scenario of rapid cloud and star formation ( see , e.g. , the reviews by * ? ? ?
* ; * ? ? ?
* and references therein . ) .
i want to thank paola dalessio , and hebe vessuri who encouraged me to write this paper .
to paola dalessio , lee hartmann , ralf klessen , stan kurtz , mordecai - m .
mac low , padelis papadopoulos and enrique vzquez - semadeni , for the careful reading of this manuscript and their useful comments .
special thanks to lee and enrique for fruitful and enjoyable discussions on the subject , and to bruce elmegreen , the referee , for an encouraging referee report . thanks also to mary - ann hall ( [email protected] ) for proofreading this paper .
this work was supported by unam - papiit grant number 110606 , and made extensive use of the nasa - ads database . | the interstellar medium is highly dynamic and turbulent . however ,
little or no attention has been paid in the literature to the implications that this fact has on the validity of at least six common assumptions on the virial theorem ( vt ) , which are : ( i ) the only role of turbulent motions within a cloud is to provide support against collapse , ( ii ) the surface terms are negligible compared to the volumetric ones , ( iii ) the gravitational term is a binding source for the clouds since it can be approximated by the gravitational energy , ( iv ) the sign of the second - time derivative of the moment of inertia determines whether the cloud is contracting ( @xmath0 ) or expanding ( @xmath1 ) , ( v ) interstellar clouds are in virial equilibrium ( ve ) , and ( vi ) larson s ( 1981 ) relations ( mean density - size and velocity dispersion - size ) are the observational proof that clouds are in ve .
turbulent , supersonic interstellar clouds can not fulfill these assumptions , however , because turbulent fragmentation will induce flux of mass , moment and energy between the clouds and their environment , and will favor local collapse while may disrupt the clouds within a dynamical timescale .
it is argued that , although the observational and numerical evidence suggests that interstellar clouds are not in ve , the so - called `` virial mass '' estimations , which actually should be called `` energy - equipartition mass '' estimations , are good order - of magnitude estimations of the actual mass of the clouds just because observational surveys will tend to detect interstellar clouds appearing to be close to energy equipartition .
similarly , order of magnitude estimations of the energy content of the clouds are reasonable . however , since clouds are actually out of ve , as suggested by asymmetrical line profiles , they should be transient entities .
these results are compatible with observationally - based estimations for rapid star formation , and call into question the models for the star formation efficiency based on clouds being in ve .
[ firstpage ] ism : general clouds kinematics and dynamics turbulence stars : formation |
recently , the authors @xcite studied the inelastic light - scattering intensity of a semi - infinite superconducting superlattice with a bilayer basis . motivated by the cooper pair tunneling model proposed by chakravarty , anderson , and coworkers @xcite for high @xmath3 layered superconductors , we discussed the in - phase and out - of - phase phase modes ( corresponding to the phase fluctuations of the two superconducting order parameters in a bilayer ) which arise in the presence of interlayer cooper pair tunneling @xcite .
these modes couple into density fluctuations and , as a result , show up in the raman inelastic light scattering .
the intensity is weak because of screening associated with the coulomb interaction , and in fact is below the threshold of current raman experiments .
however , our results are of sufficient interest that such experiments should be attempted . in this paper , we extend our previous calculations @xcite and show how the raman spectrum for @xmath4 is significantly modified when we include the surface contribution .
it is also the first time surface plasmons ( which occurs above @xmath0 ) are included in the calculation of the raman intensity of layered superconductors .
the present paper is based on the approach of jain and allen @xcite , who considered normal layered electron gas ( leg ) . in their calculation for a semi - infinite superlattice , both the bulk and surface contributions were included .
they found that there were two effects of the surface : ( 1 ) van hove singularities at the upper and lower limits of the bulk plasmon band were completely canceled out by negative surface contributions .
( 2 ) depending on the background dielectric constants , surface plasmons @xcite can appear , either above or below the bulk plasmon band .
we show that the analogous effects arise in a semi - infinite _ superconducting _ superlattice , resulting in major modifications of the bulk contribution to the raman scattering spectrum given in ref .
apart from the out - of - phase phase mode contribution , the raman scattering intensity is found to be strongly dependent on the value of momentum transfer @xmath1 ( parallel to the layers ) .
for simplicity , we only discuss the surface effects for a superlattice with a single layer per unit cell .
this is sufficient to understand the essential physics .
the isotropic inelastic light - scattering cross section is given by @xcite @xmath5 where @xmath6 here @xmath7 represents the position of the @xmath8-th layer in the @xmath9-th unit cell .
the density response function @xmath10 in ( [ eq : intensity ] ) has been evaluated in ref @xcite and represents the correlation between the charge density on layer @xmath11 and the charge density on layer @xmath12 .
the incident photon has momentum @xmath13 , energy @xmath14 , and polarization @xmath15 and the scattered photon is similarly described by @xmath16 , @xmath17 , and @xmath18 .
we assume that the energy transfer @xmath19 in ( [ eq : intensity ] ) is very small compared to the photon frequencies , _
i.e. _ , @xmath20 .
the momentum transfer parallel to the interface is @xmath21 and the momentum perpendicular to the interface is @xmath22 . for small - angle scattering
, we have @xmath23 and @xmath24 where @xmath25 is the momentum carried by the incident photon and @xmath26 describes the damping of the photons in the medium . the result given in ( [ eq : intensity ] ) shows that as a result of the finite value of @xmath26 , the inelastic light - scattering cross section involves a _ weighted _ sum of the correlation functions for electronic densities in the different layers .
we are interested in the interplay between the bulk and surface contributions to the raman spectra . for this purpose
, we only keep the _ isotropic _ matrix element for the raman interaction given in ( [ eq : raman ] ) @xcite .
we consider superconductors with _
s_-wave layer pairing interaction but similar calculations could be done for _ d_-wave superconductors , as discussed in refs
. @xcite .
for later comparison , we first recall the results of ref .
@xcite for the raman spectra for a semi - infinite superconducting _ bilayer _ superlattice , ignoring the surface contributions .
one finds that ( [ eq : intensity ] ) reduces to @xmath27\nonumber\\ & + & { e_- \over 2d_{- } } \left(1+e^{-2d/\delta}-2e^{-d/\delta}\cos 2k_\perp d\right)\biggr\ } , \label{eq : i.2}\end{aligned}\ ] ] where we have introduced the functions @xmath28 the spacing of the bilayer is @xmath29 , the unit cell length is @xmath30 , @xmath31 is the 2d coulomb interaction , and @xmath32 is the superlattice background static dielectric constant . in the long - wavelength limit ( @xmath33 ) ,
the functions @xmath34 are given by @xcite @xmath35 , \label{eq : e+-.final}\ ] ] where we have defined @xmath36 , & ~~\mbox{$\bar{\omega}>1$~ , } \end{array}\right .
\label{eq : i.s}\end{aligned}\ ] ] and @xmath37 here @xmath38 , @xmath39 , @xmath40 is the 2d electronic density of states at the fermi level with @xmath41 being the effective electronic mass , @xmath42 is the in - layer pairing interaction , and @xmath43 is the interlayer cooper pair tunneling strength . replacing @xmath44 is a simple way of including finite energy - resolution . on the rhs of ( [ eq : i.2 ] ) , the first term ( @xmath45 ) gives the contribution from the in - phase phase fluctuations , while the second term ( @xmath46 ) is associated with the out - of - phase phase fluctuations .
one finds that the in - phase first term in ( [ eq : i.2 ] ) has _ three _ poles , given by @xmath47 @xmath48 gives an in - phase plasmon mode which raman scattering picks up ( in an approximate way @xcite , this mode is similar to the plasmon mode of an infinite superlattice , with @xmath49 ) .
the additional two ( van hove ) singularities given by the solutions of @xmath50 correspond to the upper ( @xmath51 ) and lower ( @xmath52 ) limits of the `` bulk plasmon '' band for an _ infinite _ superlattice @xcite .
in contrast , the second term in ( [ eq : i.2 ] ) only has a single pole given by @xmath53 corresponding to out - of - phase phase mode discussed in detail in ref . @xcite . because the unit cell summation is over many bilayers ,
the raman intensity is strongly enhanced in the superlattice case compared to the isolated bilayer case ( the latter is discussed in ref .
this is the origin of the prefactor @xmath54^{-1}\approx \delta/2c$ ] in ( [ eq : i.2 ] ) .
using @xmath55 and @xmath56 , this prefactor is @xmath57 .
we note in ( [ eq : i.2 ] ) that , in the limit of @xmath58 , we have @xmath59 and hence @xmath60 .
this implies that the in - phase phase modes given by @xmath48 and @xmath50 have less weight when @xmath1 is small , an expected consequence of the screening due to the coulomb interaction .
in contrast , even in the low-@xmath1 limit , @xmath61 is still finite , being proportional to the pair tunneling strength @xmath43 .
this means that the out - of - phase phase mode given by @xmath62 has a weight proportional to @xmath63 and is not too dependent on the value of @xmath1 ( in the range probed in raman scattering experiments ) .
in addition , one can see from ( [ eq : i.2 ] ) that the intensities of @xmath64 and @xmath65 are also dependent on the factors ( @xmath66 ) which arise from the lattice summation in ( [ eq : intensity ] ) . since @xmath67 and @xmath68
, we see that the intensity for the out - of - phase phase modes ( @xmath52 sign ) is greatly reduced compared to that of the in - phase phase modes ( @xmath51 sign ) as a result of these factors . in fig .
[ fig : raman.superlattice ] , we plot the raman light - scattering intensity based on ( [ eq : i.2 ] ) . in this and other figures ,
we use the parameters : bilayer spacing @xmath69 , the unit cell size @xmath70 , pairing strength @xmath71 , fermi momentum @xmath72 and hence a 2d hole density @xmath73 , layer effective mass @xmath74 , background static dielectric constant @xmath75 , photon momentum in the @xmath76-direction @xmath77 , the optical penetration depth @xmath78 , the superconducting energy gap @xmath79 , and the finite - energy resolution @xmath80 .
the momentum transfer parallel to the layers is @xmath81 in fig .
[ fig : raman.superlattice ] .
the out - of - phase phase mode is well - defined , as expected . while the raman intensity from the out - of - phase contribution shown in fig .
[ fig : raman.superlattice ] is not too dependent on the value of @xmath1 , roughly speaking , the intensity from the in - phase contribution is proportional to @xmath82 . as shown in fig .
[ fig : raman.superlattice ] for the in - phase contributions , one has a small peak at @xmath83 corresponding to the pair - breaking gap in an _ s_-wave superconductor ( which is identical to the upper limit of the bulk plasmon band , _ i.e. _ , the pole given by @xmath84 ) .
in addition , `` hidden '' in the low - frequency broadened peak is the in - phase phase mode contribution given by the solution of @xmath48 , which overlaps on the van hove singularity corresponding to the pole @xmath85 at the lower limit of the bulk superlattice plasmon band ( see fig . [
fig : b=-1 ] ) .
this will become more transparent when we discuss the surface contribution . in a normal metal superlattice @xcite ,
the lower limit of the bulk plasmon band is far away from the particle - hole continuum and , as a result , both the upper ( @xmath86 pole ) and lower ( @xmath85 pole ) limits of the bulk plasmon band are well - defined .
we recall that the bulk plasmon band refers to the plasmons labeled by @xmath87 in an _
infinite _ superconducting superlattice .
these give rise to the van hove singularities discussed in ref .
in contrast , in a superconducting superlattice , the particle - hole excitation spectrum ( which begins at the pair - breaking gap @xmath0 ) is strongly coupled into the superlattice bulk plasmon spectrum . as a consequence ,
the bulk plasmon band is split into two different regions above and below the pair - breaking gap ( @xmath0 ) .
for the plasmon band below @xmath0 ( which we are most interested in ) , one can find a well - defined line for @xmath85 corresponding to the lower band limit of the bulk plasmon band ( which is generally at low frequencies ) .
however , due to the strong coupling between the bulk plasmon and bcs particle - hole continuum , starting at @xmath0 , there is no well - defined solution for @xmath86 . nevertheless , the peak at @xmath0 in fig .
[ fig : raman.superlattice ] supports the argument that @xmath83 can be considered as the effective upper limit of a superconducting bulk plasmon band @xcite . as jain and
allen @xcite have pointed out for a normal superlattice with one layer per unit cell , the van hove singularity associated with @xmath86 corresponds to all the neighboring layers oscillating in - phase ; while the one associated with @xmath85 corresponds to all the neighboring layers oscillating out - of - phase with each other . in contrast , the out - of - phase phase mode in fig .
[ fig : raman.superlattice ] is a collective mode associated with the `` internal cooper pair dynamics '' exhibited by a bilayer via the interlayer cooper pair tunneling between the two layers .
the physics of this out - of - phase phase mode is completely different from the out - of - phase @xmath85 bulk plasmon in a superlattice .
the result in ( [ eq : i.2 ] ) does not include the surface contribution .
we now include it but only consider the case of a semi - infinite superlattice of _ single _ layer per unit cell since this already describes the interplay between bulk and surface contributions . in the case of a single layer per unit cell
, we need only to replace the usual 2d lindhard function in the formulas given by jain and allen @xcite by the appropriate density response function , [ _ i.e. _ , @xmath88 in ( [ eq : e+-.final ] ) ] for a _ neutral _
2d superconductor @xcite .
using eq .
( 50 ) in ref .
@xcite , we find that the resulting raman intensity is given by @xmath89 \biggr\ } , \label{eq : i.2.surface}\end{aligned}\ ] ] where we have defined @xmath90 with @xmath91
/\sinh q_\parallel c,\nonumber\\ h&\equiv&{1\over 2}[u^{-1}(b^2 - 1)^{-{1\over 2}}-e^{-q_\parallel c}/ \sinh q_\parallel c ] /\sinh
q_\parallel c,\nonumber\\ q&\equiv&{1\over 2}[1-(b^2 - 1)^{-{1\over 2}}(1-b\cosh q_\parallel c)/ \sinh q_\parallel c]\nonumber\\ & & -{1\over 2}\alpha e^{q_\parallel c } ( b^2 - 1)^{-{1\over 2 } } ( \cosh q_\parallel c - b)/\sinh q_\parallel c , \label{eq : ghq}\end{aligned}\ ] ] where @xmath92 , and @xmath93 are defined in ( [ eq : abu ] ) and now @xmath94 .
the parameter @xmath95 depends on the optical dielectric constants ( @xmath32 ) inside and ( @xmath96 ) outside the superlattice .
it plays a key role in determining the surface contributions as well as the appearance and energy of surface plasmons .
the formula for @xmath97 can be rewritten in the useful form @xmath98 .
we call attention to the similarity between the first term in ( [ eq : i.2.surface ] ) and the first term in ( [ eq : i.2 ] ) . in the rhs of ( [ eq : i.2.surface ] ) ,
the first term ( @xmath99 ) gives the _ bulk _ contribution , while the second term ( @xmath100 ) is associated with the _ surface _ contribution .
the three poles mentioned earlier are exhibited by both contributions : @xmath48 corresponding to an in - phase plasmon which raman scattering picks up and @xmath50 corresponding to the upper and lower limits of the bulk plasmon band .
there is a new pole of the surface contribution @xmath101 given by @xmath102 which can be shown to correspond to a surface plasmon .
one has a nontrivial solution of @xmath103 only when @xmath104 ( _ i.e. , _
@xmath105 ) , that is , the surface of the superconducting superlattice must separate regions with different dielectric constants to give rise to surface plasmons . in fig .
[ fig : b=-1 ] , we show the dispersion relation of surface plasmons in a superconducting superlattice , as given by the solutions of @xmath103 , for various values of @xmath97 . the shaded area between the line denoted by @xmath85 to the line @xmath83 represents the bulk plasmon band of an infinite superconducting superlattice .
we find that the surface plasmon appears only above the upper limit ( _ i.e. _ , @xmath83 ) of this bulk plasmon band , where bcs particle - hole damping of collective modes can occur . for positive values of @xmath97 ( @xmath106 )
, a surface plasmon appears at very large energies @xmath107 , in which case it is essentially identical to that in a normal superlattice .
we remark that for a high-@xmath3 material with a dielectric constant @xmath108 in a vacuum ( @xmath109 ) , one has @xmath110 . for @xmath97 increasingly negative ( @xmath111 ) , which requires @xmath112 , the surface plasmon energy slowly
_ decreases _ toward @xmath0 ( see fig . [
fig : b=-1 ] ) . in fig .
[ fig : b=-1 ] , the dispersion relation denoted by @xmath48 represents a ( bulk ) plasmon , which is a pole of the raman scattering intensity in ( [ eq : i.2.surface ] ) .
one sees that @xmath48 mode is well - defined only when @xmath113 ( solid line ) .
the critical value of @xmath1 ( @xmath114 ) changes for different choices of @xmath26 and @xmath25 .
when @xmath115 , we find no solution for @xmath48 .
the dashed line represents the minima of @xmath93 ( _ i.e. _ , an over - damped or relaxational mode ) .
this broad resonance is always peaked at @xmath116 , whatever the values chosen for @xmath26 and @xmath25 . in fig . [
fig : cancel_1 ] , we show the net raman intensity based on ( [ eq : i.2.surface ] ) for a semi - infinite superconducting single - layer superlattice , showing the surface and the bulk components . comparing fig .
[ fig : cancel_1 ] with fig .
[ fig : raman.superlattice ] , one sees that the low - weight van hove singularity at @xmath85 from the bulk contribution is canceled by the negative surface contribution .
in contrast , the _ in - phase _ phase bulk plasmon mode ( @xmath48 ) shows up as a sharp peak in the low - frequency region .
as shown in fig .
[ fig : cancel_1 ] , a peak associated with the @xmath84 van hove singularity can still appear at @xmath83 since the bulk and surface contributions do not completely cancel each other . fig .
[ fig : cancel_2 ] is similar to fig .
[ fig : cancel_1 ] using the same parameters , but at a much higher momentum transfer @xmath117 .
the _ cancellation _ between the bulk and surface contributions is clearly shown not only for the two boundaries of the bulk plasmon band at @xmath85 and @xmath83 , but almost for the entire region below @xmath0 .
we also see that at @xmath117 , the broad relaxational mode [ always peaked at @xmath118 corresponding to the minimum of @xmath93 ( see fig .
[ fig : b=-1 ] ) has only low weight in the raman scattering spectrum . to see how the interplay between the bulk and surface contribution depends on @xmath1 , we plot in fig .
[ fig : qinverse ] the _ total _ raman intensities for various values of @xmath1 .
comparing fig .
[ fig : qinverse ] with fig .
[ fig : b=-1 ] , one finds that for @xmath113 , a well - defined bulk plasmon mode ( @xmath48 ) gives rise to a sharp peak even after the cancellation of the bulk and surface contributions .
a peak associated with @xmath83 still shows up . as mentioned above , for @xmath115 , the minimum of @xmath93 results in a very broad maximum at @xmath119 .
we also note that for @xmath115 , while both the bulk and surface contributions are roughly proportional to @xmath82 , the _ net _ spectrum after the mutual cancellation of bulk and surface contribution effectively _ decreases _ as @xmath1 increases . for @xmath120 , @xmath121 , and @xmath122 appropriate for layered superconductors , the total raman intensity given in ( [ eq : i.2.surface ] )
can be reduced after some calculation to @xmath123 , \label{eq : i.tot.approx}\ ] ] which is valid for @xmath124 and for all values of @xmath97 .
we note that in ( [ eq : i.tot.approx ] ) , the function @xmath93 given in ( [ eq : abu ] ) is a very sensitive function of @xmath125 and @xmath126 and therefore we can not approximate it by @xmath127 in the limit of @xmath125 , @xmath128 .
we can see directly from ( [ eq : i.tot.approx ] ) that the poles of @xmath50 are removed as a consequence of the cancellation between bulk and surface contributions .
the only pole now is given by @xmath48 .
one has a broad spectrum without any sharp peak for @xmath115 . in this region ,
one may verify that @xmath129 . as a result ,
the net raman intensity @xmath130 decreases as @xmath1 increases , roughly proportional to @xmath131 ( see fig .
[ fig : qinverse ] ) .
in addition to the surface contribution discussed above , there may be a surface plasmon in the region @xmath132 ( see fig . [
fig : b=-1 ] ) but this only arises when we have different dielectric constants inside and outside the superlattice ( @xmath133 ) . in fig .
[ fig : cancel_3 ] , we show the contribution of a surface plasmon to the raman intensity .
the parameters used are as in fig .
[ fig : cancel_1 ] but at a much higher momentum transfer @xmath134 ( which is probably the upper limit for raman scattering in high-@xmath3 superconductors ) . in order to have a surface plasmon energy fairly close to @xmath0 ,
the dielectric constants are taken to give @xmath135 .
the latter value requires @xmath136 , _ i.e. _ , a layered superconductor with _ much lower _ dielectric constant compared to the overlay material . as shown in fig .
[ fig : cancel_3 ] , this gives a raman spectrum with lots of structure , with a broad surface plasmon peak at an energy above @xmath0 .
this region is the pair - breaking region , where there is strong bcs particle - hole damping of collective modes .
the intensity of the surface plasmon is roughly proportional to @xmath82 . once again , fig .
[ fig : cancel_3 ] shows the almost complete cancellation between the bulk and surface contributions in the region of @xmath137 .
this is because at this relatively large value of @xmath1 , there is no well - defined solution of @xmath48 ( see fig .
[ fig : b=-1 ] ) .
one might worry that the bilayer out - of - phase phase mode shown in fig .
[ fig : raman.superlattice ] might also be strongly modified due to the surface contributions at higher values of @xmath1 . however , while the intensity ( and the dispersion relation ) of the out - of - phase phase mode is very sensitive to the pair tunneling strength @xmath43 , it is not strongly dependent on the value of @xmath1 .
it is associated with out - of - phase oscillation of order parameters in a single bilayer , with no _ net _ charge fluctuation @xcite .
therefore , we have no reason to expect any strong modification of the out - of - phase phase mode at larger values of @xmath1 .
we have shown that the surface contribution plays a major role in determining the final raman - spectrum line shape from semi - infinite superconducting superlattices .
we find that , as in the case of normal superlattices discussed by jain and allen @xcite , the proper inclusion of the surface contribution cancels `` spurious '' bulk contributions associated with van hove singularities ( see ref .
@xcite ) of the upper and lower limits of the bulk plasmon band of an infinite superconducting superlattice .
we have also found that the bulk plasmon mode in the region @xmath138 ceases to be well - defined when @xmath1 reaches a critical value ( @xmath139 for the parameters we have used ) .
in addition , the surface plasmon usually contributes to the raman intensity in the region well _ above _ @xmath0 and only approaches @xmath0 if the superconductor is overlayed by a transparent material with a dielectric constant @xmath96 much larger than that of the superlattice @xmath32 .
the surface - plasmon mode intensity increases as @xmath82 and thus one wants @xmath1 as large as possible if one wants to study it ( see fig . [ fig : cancel_3 ] ) .
as mentioned above , for relatively large values of @xmath1 , the ( negative ) surface contribution tends to completely cancel out the bulk contribution in the entire frequency region _ below _ @xmath0 . because the out - of - phase phase mode discussed in ref .
@xcite is not expected to be strongly affected by the surface contribution , at relatively large values of @xmath1 , this mode ( see fig . [
fig : raman.superlattice ] ) should be the only structure remaining in the raman spectrum below @xmath0 .
the absolute intensity of the raman spectra we discuss in this paper is somewhat below current experimental sensitivity .
however we hope that the interesting predictions we make concerning the out - of - phase phase modes as well as surface plasmons will encourage future experimental efforts . | generalizing recent work , the raman scattering intensity from a semi - infinite superconducting superlattice is calculated taking into account the surface contribution to the density response functions .
our work makes use of the formalism of jain and allen developed for normal superlattices .
the surface contributions are shown to strongly modify the bulk contribution to the raman - spectrum line shape below @xmath0 , and also may give rise to additional surface plasmon modes above @xmath0 .
the interplay between the bulk and surface contribution is strongly dependent on the momentum transfer @xmath1 parallel to layers .
however , we argue that the scattering cross - section for the out - of - phase phase modes ( which arise from interlayer cooper pair tunneling ) will not be affected and thus should be the only structure exhibited in the raman spectrum below @xmath0 for relatively large @xmath2 .
the intensity is small but perhaps observable .
epsf # 1 0.1 true in |
protein structure prediction plays an important role in the fields of bioinformatics and biology .
traditional protein structure prediction approaches include template - based modeling ( tbm , including homology modeling , and threading ) , and free modeling ( fm ) .
in particular , a threading algorithm takes a query protein sequence as input , recognizes the most likely fold , and finally reports the alignments of the query sequence to structure - known templates as output .
the existing threading approaches mainly utilizes the information of protein sequence profile , solvent accessibility , contact probability , etc .
the threading strategy has been shown to be successful in structure prediction of a great amount of proteins ; however , the existing threading approaches show poorly performance for remote homology proteins .
how to improve the fold recognition for remote homology proteins remains a challenge to protein structure prediction .
the sequences of proteins in remote homology generally show relatively weak signal of structure .
however , this does not mean that there is no sequence conservation hints for structure . the success of multiple - templates strategy implies the existence of common frameworks , i.e. some regions of proteins are conservative both in the structure and sequence .
such common frameworks should be responsible to the structural stability and then conservative in the evolution .
based on this we proposed a novel threading approach in three steps .
first , for each template , the common structural frameworks shared by its homologous proteins were calculated .
second , unlike in traditional threading methods where the alignment is made against the whole template , we aligned the query protein sequence against a common framework first .
this strategy avoids the drawback of the traditional threading approach , i.e. the alignment of variable regions beyond conserved motifs is prone to bringing in error .
third , the final alignments were generated via aligning query sequence against candidate full - length templates in the family .
briefly speaking , we run treethreader@xcite to build alignments of query against the new template database , and ranked alignments by e - value for model generation . finally , we generated models by modeller based on candidate alignments .
the generated models are ranked according to ddfire@xcite energy function .
for each template with known structure , all of its remote homology proteins are first identified based on structure alignment .
then , a linear programming was designed to identify the common framework shared by these remote homology proteins .
the common framework identification problem can be described as : given a collection of homologous proteins @xmath0 with length @xmath1 , the objective is to find @xmath2 segments with length @xmath3 with high sequence conservation and structural similarity . as an example , fig .
[ ilpobjective ] shows the common frameworks shared by protein 3gxr_a and its homologous proteins .
the common framework poses double - fold requirements , i.e. , significantly high sequence conservation and structural similarity . in the linear program ,
the objective function was designed to describe structural similarity , and the constraints were designed to describe sequence similarities .
specifically , the linear program utilizes a set of boolean variables to represent the location of conserved segments , i.e. , @xmath4 denotes that in the @xmath5th protein , the @xmath6th segment is located at the @xmath7-th residue .
then , the structural similarity objective and sequence similarity constraints can be described using @xmath8 .
the constraints were designed to represent the following requirements .
* for any sequence , the @xmath6th segment in common framework is unique ; * no segment in a common framework overlaps nor crosses .
* the segments should have significantly high sequence similarity .
the integer linear programming model can be described as : + @xmath9 @xmath10 where @xmath11 denotes the length of the @xmath5th protein , and @xmath12 denotes the pre - calculated sequence similarity matrix . in particular , the cell @xmath13 denotes the sequence similarity between of the segment starting from @xmath14 in the @xmath15th protein and the segment starting from @xmath16 in the @xmath17th protein . in our model
, the structure similarity is described using dscore @xcite .
@xmath18 where @xmath19 and @xmath20 denote the c@xmath21 distance of residue @xmath5 and residue @xmath7 in protein @xmath22 and @xmath23 respectively .
the final integer linear programming can be formulated as : @xmath24 @xmath25 where the indicator @xmath26 equals to 1 iff all the four item @xmath27 , @xmath28 , @xmath29 , @xmath30 equal to 1 , and 0 otherwise .
the indicator @xmath31 equals to 1 iff both @xmath32 and @xmath33 equal to 1 .
the @xmath34 and @xmath12 matrix are calculated in advance .
the cell @xmath35 denotes the approximated dscore of segments start from @xmath36 and @xmath37 in the @xmath15th protein and segments start from @xmath37 and @xmath38 in the @xmath17th protein .
[ c37111 ] shows the common frameworks shared by two domains in scop family c.37.1.11 .
the common frameworks has a dscore of 8.35 and an rmsd of 1.9 , implying a significantly high structure conservation .
for a total of over 27,000 proteins in pdb70 , updated at apr . 19 , 2014 , the common frameworks were identified to yield a database called topo .
the test set consists of 142 pairs of protein structures similar in structure but with low sequence identity .
traditional threading approaches , say hhpred , fail to build an accurate alignment between such protein pairs . in contrast , our alignment method successfully build accruate alignment ( tmscore@xmath39 ) for seven protein pairs , and generate accurate contact information for 45 protein pairs .
take a pair of protein 3dz1_a vs. 1twd_a as an example .
the two proteins share similar protein structure ( tmscore=0.56 ) ; however , the alignment generated by hhpred has a tmscore of only 0.22 .
in contrast , our alignment method generates an alignment with tmscore=0.43 .
unlike close homology proteins , remote homology proteins show weakly overall sequence signals of structure similarity . however , they still share common frameworks which carry strong sequence signals of structure similarity . aligning against the common frameworks instead of whole protein sequences
improves the fold recognition .
the study was funded by the national basic research program of china ( 973 program ) under grant 2012cb316502 , the national nature science foundation of china under grants 11175224 and 11121403 , 31270834 , 61272318 , 30870572 , and 61303161 and the open project program of state key laboratory of theoretical physics ( no.y4kf171cj1 ) .
this work made use of the einfrastructure provided by the european commission co - funded project chain - reds ( ga no 306819 ) . | ( extended abstract ) = 13.5pt = 13.5pt |
studies of galaxy kinematics in clusters remain a major tool in determining the mass distribution in these objects , complemented by methods based on the analysis of the hot x - ray gas and gravitational lensing . due to a limited number of measured galaxy redshifts per cluster such analyses
have been usually performed on composite clusters by combining data from many objects ( e.g. carlberg et al .
1997 ; van der marel et al .
2000 ; biviano & girardi 2003 ; mahdavi & geller 2004 ; katgert , biviano & mazure 2004 ; biviano & katgert 2004 ; goto 2005 ; biviano & salucci 2005 ) .
the normalizations needed in stacking the clusters together reduce however the number of parameters that can be estimated from the analysis . besides
, such studies are usually restricted to the analysis of velocity dispersion profile with the simplifying assumption of isotropic galactic orbits . [ cols="^,^,^,^,^,^,^,^,^,^ " , ] in this study we attempt a kinematical analysis of six individual clusters which we supplement in the end by a similar procedure performed on a composite cluster created from the galaxies belonging to the six clusters .
our method relies on an extension of the usual jeans formalism beyond the lowest - order velocity moment and including also the kurtosis of the line - of - sight velocity distribution ( okas 2002 ; okas & mamon 2003 ) .
the formalism has been successfully applied to study the dark matter distribution in the coma cluster of galaxies by okas & mamon ( 2003 ) .
it has been shown that , for a restricted class of dark matter distributions motivated by the results of cosmological @xmath1-body simulations , the joint analysis of velocity dispersion and kurtosis allows us to break the usual degeneracy between the mass distribution and velocity anisotropy and constrain the parameters of the dark matter profile .
recently we have tested the reliability of this approach against a series of @xmath1-body simulations ( sanchis , okas & mamon 2004 ) and also applied it to constrain the dark matter distribution in the draco dwarf spheroidal galaxy ( okas , mamon & prada 2005a ) . here
we further test the method on a different set of cosmological @xmath1-body simulations by studying in detail the errors in the estimated parameters following from the sampling errors of velocity moments .
we also introduce and test a new procedure of interloper removal which we then apply to the clusters . with a number of available galaxy redshifts per cluster much smaller than for coma we had however to introduce a number of simplifications in our modelling compared to okas & mamon ( 2003 ) .
we model the total mass distribution instead of only the dark matter component , we assume that galaxies trace the total mass distribution and we use the velocity data of all galaxies , not only ellipticals .
on the other hand , we believe that the carefully selected clusters studied here are much more relaxed compared to coma and therefore we avoid any uncertainties due to departures from dynamical equilibrium . the paper is organized as follows . in section 2
we briefly describe our data set . in section 3
we summarize our method of data modelling including the removal of interlopers and fitting of velocity moments and test it on a sample of dark matter haloes extracted from a cosmological @xmath1-body simulation .
the results for the six galaxy clusters and the composite cluster are presented in section 4 .
the discussion follows in section 5 .
we have searched the nasa / ipac extragalactic database ( ned ) for nearby ( @xmath4 ) well - studied galaxy clusters with at least 120 galaxies within projected distance of about 2 mpc from the cluster centre and with @xmath5 velocities differing from the cluster mean by less than @xmath6 km s@xmath7 . among a few tens of clusters selected in this way we have chosen those which are likely
to be relaxed judging by the regularity of their x - ray surface brightness maps obtained with einstein ( jones & forman 1999 ) and asca ( horner et al .
2000 ) satellites .
next we have looked at the regularity of the diagrams showing the line - of - sight velocities of galaxies as a function of projected distance from the cluster centre . we have rejected those with irregular diagrams which may indicate merging or presence of neighbouring structures ( okas et al .
2005b ) for which our method of interloper removal does not work .
table [ properties ] lists the clusters we chose for the analysis together with their positions and redshifts as given by ned .
we also give the number of galaxies per cluster used for the calculation of the velocity moments ( the numbers are for the final samples , after removal of interlopers ) .
the last two columns give the values of the line - of - sight velocity moments , dispersion and kurtosis calculated for all galaxies .
the method of estimating these values is discussed in the next section .
the list includes some very well known clusters .
a0496 was studied in detail by durret et al .
( 2000 ) who called it a prototype of a relaxed cluster .
indeed the cluster seems to have the most regular x - ray luminosity distribution of those in our sample and a single central galaxy whose position coincides with the centre of the gas distribution .
the cluster a1060 ( hydra i , fitchett & merritt 1988 ) , although quite similar to a0496 in many aspects , has two central galaxies and somewhat less uniform x - ray distribution so it may have just reached equilibrium after a major merger . none of the clusters , however , is completely free of substructure ( see e.g. escalera et al .
1994 ) or neighbours ( even a0496 ) .
a3558 ( shapley 8) is the richest member of the shapley supercluster ( dantas et al .
1997 ; bardelli et al .
1998 ) including a much smaller neighbour a3556 and a bigger , more distant cluster a3562 . for this cluster
we have restricted the analysis to distances where members of a3556 and a3562 are not likely to contaminate the sample .
a2199 seems to be quite a relaxed cluster but has a less massive neighbour a2197 and other groups ( rines et al .
2001 , 2002 ) which made us restrict the analysis to distances much smaller than the estimated virial radius .
also a0262 , a member of the perseus - pisces supercluster , can be affected by neighbouring structures .
it is generally believed that elliptical galaxies in clusters comprise a virialized , more relaxed subsample compared to ( late ) spirals which might be infalling into the cluster for the first time .
it could therefore be desirable to restrict the analysis to ellipticals as was done in the case of coma cluster by okas & mamon ( 2003 ) .
however , for the present sample of clusters the morphological information is available only for a small fraction of galaxies making such an analysis impossible . given a larger number of galaxies per cluster one could also attempt to measure their surface density distribution reliably and use it as an input in the kinematical analysis . with scarce samples presently available
we are forced to assume that the number density of galaxies follows the total density distribution . for the kinematical analysis we have chosen as the centres of the clusters their central cd galaxies which coincide with the centre of the x - ray surface brightness distribution . in the case of the presence of two central galaxies ( in a1060 and a3158 ) we have chosen as a centre the position of the one which is closer to the centre of the x - ray surface brightness distribution .
the galaxy velocities have been transformed to the reference frame of the cluster and in order to calculate the distances within clusters we have transformed the cluster velocities to the reference frame of the cosmic microwave background . the concordance cosmological model ( @xmath8cdm ) with parameters @xmath9 , @xmath10 and @xmath11 is assumed throughout the paper .
in this section we summarize our method of determining the mass distribution from velocity moments , as developed in okas ( 2002 ) and okas & mamon ( 2003 ) , and test it against mock data obtained from @xmath1-body simulations .
the method relies on fitting the solutions of the jeans equations for the second and fourth velocity moments to the profiles of the moments determined from the data . for the velocity dispersion projected along the line of sight , @xmath12 , the jeans formalism gives ( binney & mamon 1982 ) @xmath13 where @xmath14 and @xmath15 are the 3d and the surface distribution of the tracer as a function of a true ( @xmath16 ) and projected ( @xmath17 ) distance from the object centre respectively .
the parameter @xmath18 describes a relation between the angular @xmath19 and radial @xmath20 velocity dispersions and characterizes the anisotropy of the tracer orbits .
we will assume it here to be constant with radius and consider @xmath21 which covers all interesting possibilities from radial orbits ( @xmath22 ) to isotropy ( @xmath23 ) and circular orbits ( @xmath24 ) .
for constant @xmath25 the radial velocity dispersion , @xmath26 in equation ( [ proj1 ] ) is @xmath27 where @xmath28 is the gravitational potential . for the fourth projected velocity moment
we have @xmath29 where @xmath30 and @xmath31 in the following we will rescale the fourth moment to obtain the line - of - sight or projected kurtosis @xmath32 whose value is 3 for a gaussian distribution . having measured the line - of - sight velocities and projected positions for a number of tracer particles or galaxies one can estimate the profiles of velocity dispersion and kurtosis by taking @xmath33 tracer particles per bin and using the following estimators of variance and kurtosis @xmath34 and @xmath35 where @xmath36 is the mean of velocities in a bin . since in the case of study of galaxy kinematics in clusters the number of galaxies usually does not exceed a few hundred , and for our least sampled cluster in table [ properties ] ( a0262 ) is as low as 120 , in order to have at least 4 data points in each of the velocity dispersion or kurtosis profiles we need to adopt a rather low number of @xmath37 objects per bin . using the monte carlo method described in the appendix of okas & mamon ( 2003 ) one can then construct unbiased and gaussian - distributed estimators of line - of - sight velocity dispersion @xmath38 and kurtosis - like variable @xmath39 @xmath40 @xmath41^{1/10}\ ] ] where @xmath42 and @xmath43 are given by equations ( [ app1])-([app3 ] ) .
the factor @xmath44 in equation ( [ app4 ] ) is the well known correction for bias when estimating the sample variance , valid independently of the underlying distribution . in ( [ app5 ] )
the factor @xmath45 corrects for the bias in the kurtosis estimate , i.e. unbiased estimate of kurtosis is @xmath46 , while the rather complicated function of @xmath47 assures that the sampling distribution of @xmath39 is approximately gaussian .
we have also checked that defined in this way the measured velocity dispersion and kurtosis in a given bin are very weakly correlated so the data points can be fitted as independent .
the standard errors in the case of @xmath38 are @xmath48 while in the case of @xmath39 are approximately 0.02 ( for @xmath49 ) . in the following
we assign these sampling errors to our mock and real data points .
we will further assume that the total density distribution in the studied objects ( simulated haloes or galaxy clusters ) and the distribution of the tracer are well approximated up to the virial radius @xmath50 by the nfw formula ( navarro , frenk & white 1997 ) @xmath51 where @xmath52 , @xmath53 is the present critical density , @xmath54 is the characteristic density parameter , @xmath55 is the concentration parameter and @xmath56^{-1}$ ] .
we define the virial mass and radius as those with mean density @xmath54 times the critical density according to the spherical collapse model for the standard @xmath8cdm cosmology ( see okas & hoffman 2001 ) .
the surface distribution of the tracer @xmath15 following from 3d density profile ( [ ro ] ) can be found , together with other properties of the nfw haloes , in okas & mamon ( 2001 ) .
having estimated the velocity dispersion and kurtosis profiles from the measured positions and velocities we can fit these data with the solutions ( [ proj1])-([proj2 ] ) estimating three free parameters : the virial mass @xmath57 , concentration @xmath55 and anisotropy @xmath25 . in the remaining part of this section we apply the method to a set of mock data in order to assess its viability in reproducing the properties of real galaxy clusters .
our approach here is similar to that of sanchis et al .
( 2004 ) , but in addition we address the problem of interlopers and we estimate the errors in the parameters due to sampling errors . for this work we used the results of a cosmological dark matter simulation described by wojtak et al .
the simulation was performed within a box of size 150 @xmath58 mpc assuming the concordance cosmological model ( @xmath8cdm ) with parameters @xmath9 , @xmath10 , @xmath11 and @xmath59 .
we focused on 10 massive and isolated haloes extracted from the final output of the simulation whose properties are listed in table 1 of wojtak et al .
( 2005 ) . in order to emulate the observations we place an observer at a distance of 100 mpc from a centre of a given halo so that he will be able to see it receding with a velocity of around 7000 km s@xmath7 .
we assume that the line of sight is parallel to the x , y or z axis of the simulation box , with respect to which the haloes should be oriented randomly .
the observer is located far enough from the halo so that the cone of observation can be approximated by a cylinder .
we then project all particle velocities along the line of sight and the distances on the surface of the sky restricting the observations to the circle of projected radius @xmath60 on the sky .
next we reject all particles with velocities differing from the mean velocity of the halo by more than @xmath6 km s@xmath7 , as we did for the real clusters .
this cut - off corresponds to at least @xmath61 ( with @xmath12 calculated for all particles in the halo ) so it is not very restrictive . from the obtained sample of particles
we randomly draw 300 out of about @xmath62 per halo .
= 8.3 cm the velocities and positions of 300 particles obtained in this way for halo 1 ( of mass @xmath63 and virial radius @xmath64 mpc ) observed along x , y and z axis ( upper , middle and lower row respectively ) are shown in fig .
[ vrhalo1 ] . in the plots shown in the left column of the figure the particles that happen to lie inside the virial radius of the halo ( as verified using 3d information )
were coded with filled circles , while those outside the virial radius are shown with open symbols . in the right panel
the coding is similar but the particles were divided into those which are bound ( have velocities smaller that the escape velocity , @xmath65 ) or unbound to the halo .
the axes along which the particles were observed are marked in the lower left corner of each panel together with the labels showing whether the distinction between the particles was made with the criterion of @xmath50 or @xmath66 . in the upper left corners of the plots
we give the numbers of particles fulfilling or not the criterion which sum to the total number of 300 . averaging over 10 haloes and 3 directions of observation
we find that of the total number of particles ( 300 ) , 76 percent reside inside @xmath50 , 87 percent inside 2 @xmath50 , while 92 percent are actually bound to its halo .
in addition , the unbound particles are always a subsample of those with @xmath67 , i.e. there are no unbound particles inside @xmath68 . the particles that do not fulfill any of the criteria are more common at larger projected distances from the centre of the halo .
the particles from outside @xmath50 are obviously candidates for interlopers since they would not be used to estimate the density profile from the 3d information . however , as the cited numbers show , about half of them are actually close to the halo ( within 2 @xmath50 ) and most of them are bound to the halo and therefore probably reasonably good tracers of the potential .
this agrees with the recent studies based on @xmath1-body simulations ( e.g. klypin et al .
2003 ; prada et al .
2005 ; betancort - rijo et al . 2005 ) which demonstrate that the virialized region extends somewhat beyond the virial radius , as we define it . besides ,
the simulated dark matter haloes are usually not spherical and by imposing spherical symmetry in our definition of virialized region we may in fact cut out particles which are actually members of the halo . anyway , most of the candidate interlopers reside close to the mean velocity of the halo and only a few of them are true outliers which could significantly alter our estimates of the velocity moments .
we proceed to remove these outliers in the following way .
first we calculate the velocity dispersion profiles by binning the data ( with 30 particles per bin ) and assigning them sampling errors , as described in the previous subsection .
we then fit the data with solutions ( [ proj1 ] ) assuming @xmath23 and adjusting @xmath57 and @xmath55 .
although the @xmath25 values of dark matter haloes are mildly radial ( with mean @xmath69 , see e.g. wojtak et al . 2005 ; mamon & okas 2005 ) the actual values of the parameters are not very important at this stage since our purpose now is only to reproduce the shape of the dispersion profile and this can be done well with 2 parameters instead of 3 .
if the velocity dispersion increases strongly and the fit goes to values of @xmath70 ( the nfw profile would not make sense since the scale radius would be larger than the virial radius ) we keep @xmath71 and adjust @xmath57 and @xmath25 instead .
we then reject all particles lying outside the mean velocity @xmath72 where @xmath73 is the velocity dispersion profile obtained with our best - fitting parameters .
the procedure is repeated until no more particles are removed . in each iteration
we also calculate new estimate of the mean line - of - sight velocity of the particles with respect to which the rejection is performed .
the @xmath72 profiles from the last iterations are shown as solid lines in fig .
[ vrhalo1 ] .
as we can see , the procedure removes the most obvious interlopers which affect the velocity moments most strongly .
the mean number of particles removed in all 30 experiments is 14 which corresponds to 61 percent of the mean number of unbound particles per halo . in rare cases
where the initial velocity dispersion profile from the data is strongly increasing the first fit may not remove any particles and it is necessary to repeat the fitting for a smaller number of data points .
in such cases it may also happen that a member particle from the centre of the halo is removed .
the effectiveness of the method will be compared to other methods of interloper removal used in the literature in a forthcoming paper ( wojtak et al .
, in preparation ) .
= 7.2 cm we now assess the viability of our method by jointly fitting both velocity dispersion and kurtosis profiles calculated from our mock data for the simulated dark matter haloes . after removal of interlopers we are left with 8 - 9 data points for both dispersion and kurtosis which is of the same order as what we have for real galaxy clusters .
the quality of the obtained fits and magnitude of errors in the parameters should therefore also be similar .
we jointly fit the mock data for velocity dispersion and kurtosis with the predictions from the jeans formulae ( [ proj1])-([kurt ] ) adjusting the three parameters @xmath57 , @xmath55 and @xmath25 .
the parameters which minimize @xmath74 are shown in fig .
[ fittedh ] as filled dots for 10 haloes observed along x axis of the simulation box ( for the observations along the y and z axis the results are similar ) .
the @xmath75 errors in the parameters due to the sampling errors , were found by exploring the 3-dimensional confidence region in the @xmath76 , @xmath77 and @xmath78 parameter planes and finding probability contours corresponding to @xmath75 , @xmath79 and @xmath80 i.e. @xmath81 .
the parameters of the haloes obtained from the full 3d information are marked as crosses .
the masses and concentrations were taken from table 1 of wojtak et al .
( 2005 ) and the @xmath25 parameters were calculated from the @xmath25 profiles of the haloes by taking averages in 10 radial bins inside the virial radius .
the discrepancies between the fitted and true values of the parameters may be due to non - sphericity , presence of substructure , projected outliers and departures from equilibrium e.g. in the form of streaming motions .
however , as we can see , the true values of the parameters are almost always within the estimated @xmath75 error bars of the fitted values ( except for the mass estimates for haloes 5 , 7 and 9 where the fitted values are somewhat lower ) .
we conclude from this analysis that the sampling errors are the main source of error in this method .
we also note that the method might possibly work even better for real galaxies in clusters than for our randomly selected particles .
although it would be worthwhile to test the method on simulated galaxies , at the present stage of the simulations this would be reduced to using subhaloes detected with standard halo - finding techniques .
the distributions of subhaloes both in space and velocity are known to be biased with respect to those of dark matter particles ( diemand et al .
2004 ) and probably still suffer from overmerging problem .
the density distribution of subhaloes is flat in the centre while both the particles in simulated haloes and galaxies in clusters have cuspy profiles .
it would be very difficult to disentangle the effects mentioned above from the uncertainties due to the use of subhaloes .
besides , simulations including baryons suggest that clusters tend to be more spherical than pure dark matter haloes ( kazantzidis et al .
2004 ; basilakos et al .
2005 ) which would reduce the projection effects due to non - sphericity of the systems .
we also believe that cluster - cluster mergers can be more easily detected in real clusters than in a single final output of an @xmath1-body simulation of dark matter haloes by studies of the x - ray emitting gas and so our sample of clusters is probably more relaxed than the sample of dark matter haloes we studied . in the application of the method to real galaxy clusters
we will therefore neglect other sources of errors and estimate the uncertainties in the parameters only from the sampling errors of the velocity moments .
= 8.3 cm we now proceed to apply the method of joint fitting of velocity moments to the sample of six clusters listed in table [ properties ] .
we start by plotting in fig .
[ vr ] the line - of - sight velocities of galaxies with respect to the cluster mean velocity as a function of projected distance from the cluster centre . in order to separate the galaxies which will be used in the calculation of velocity moments from the supposed interlopers we apply the procedure for interloper removal described in the previous section , exactly as we did for simulated data with the additional assumption that the tracer density is proportional to the total mass density ( we do not infer the tracer density from the observed surface number density of galaxies , because the latter would be very uncertain for such small samples and probably suffers from incompleteness that varies with projected radius ) . since we do
not a priori know the virial radius of the cluster , after estimating the virial mass in each iteration we check whether all fitted data points lie inside the virial radius or an additional data point could be included and adjust the number of points accordingly .
the @xmath72 profiles obtained in the final iteration of the procedure are shown as solid lines in fig .
as can be read from the figure the number of rejected galaxies is between 0 ( for a2199 ) and 4 ( for a3558 ) , much lower than the number of particles rejected during the application of the method to simulated haloes .
the rejected galaxy in the centre of a3558 lies inside the final @xmath72 lines but was removed in the earlier iteration of the procedure .
having found the final sample of galaxies we divide the data into radial bins of 30 galaxies ( except for a3158 which has 29 galaxies per bin ) and calculate the velocity dispersion and kurtosis profiles which are shown in fig .
[ dispersion ] and [ kurtosis ] with @xmath75 sampling errors .
we see that the profiles are similar to those characteristic for objects with nfw - like density profiles and orbits close to isotropic : the global trend is that both profiles slightly decrease with radius ( see okas & mamon 2001 ; sanchis et al .
a3558 shows the most variable velocity dispersion profile with one discrepant point at 0.7 mpc ( which is due to the single galaxy with discrepant velocity present in this bin
see fig .
[ vr ] ) and a secondary increase at about 1.8 mpc while its kurtosis profile remains rather uniform . on the other hand a0496 , believed to be a very relaxed cluster , has the most variable kurtosis profile . in an attempt to verify whether this variability may be due to departures from equilibrium e.g. in the form of the presence of infalling groups of galaxies
we have also looked at the mean line - of - sight velocity profiles with respect to the cluster mean .
they do not however depart strongly from zero , typically remaining in each bin within @xmath82 ( the circular velocity at the virial radius as determined in the next subsection ) .
= 8.3 cm = 8.3 cm = 7.2 cm cccccc no . & cluster & @xmath83 & @xmath55 & @xmath25 & @xmath84 + & & @xmath85 $ ] & & & + 1 & a0262 & @xmath86 & @xmath87 & @xmath88 & 1.1/5 + + 2 & a0496 & @xmath89 & @xmath90 & @xmath91 & 18.5/15 + + 3 & a1060 & @xmath92 & @xmath93 & @xmath94 & 8.6/19 + + 4 &
a2199 & @xmath95 & @xmath96 & @xmath97 & 3.8/9 + + 5 & a3158 & @xmath98 & @xmath99 & @xmath100 & 5.0/7 + + 6 & a3558 & @xmath101 & @xmath102 & @xmath103 & 27.7/25 + [ fittedpar ] in the following analysis we assume that the galaxy distribution follows that of the total mass distribution .
although this assumption is far from obvious there is evidence from studies of clusters ( e.g. carlberg et al .
1997 ; okas & mamon 2003 ; biviano & girardi 2003 ) that both number density and luminosity density of galaxies in clusters are cuspy and therefore can be quite well approximated by the nfw profile ( contrary to the distribution of subhaloes in the dark matter simulations , which probably have a shallow core , see diemand et al .
the concentration of these distributions does not have to be the same as that of the total mass , but with numbers of galaxies as low as 120 per cluster we are not able to reliably estimate the distribution of the tracer .
we jointly fitted the data for velocity dispersion and kurtosis with the predictions from the jeans formulae ( [ proj1])-([kurt ] ) by minimizing @xmath74 and adjusting the three parameters @xmath57 , @xmath55 and @xmath25 .
the best - fitting parameters for the six clusters are shown in fig .
[ fitted ] as filled dots together with @xmath75 error bars .
the exact values are listed in table [ fittedpar ] where the last column shows also the goodness of fit measure , @xmath84 .
( the virial radii corresponding to the virial masses are @xmath104^{1/3}$ ] mpc for our adopted cosmological model . ) the errors in the parameters due to the sampling errors were found by exploring the 3-dimensional confidence region in the @xmath76 , @xmath77 and @xmath78 parameter planes and finding probability contours corresponding to @xmath75 , @xmath79 and @xmath80 i.e. @xmath81 .
the best - fitting velocity dispersion and kurtosis profiles are shown in fig .
[ dispersion ] and [ kurtosis ] as solid lines .
interestingly , a3558 has the lowest concentration among all studied clusters .
although the rather large error bars prevent us from concluding too much , this is expected since a3558 is in the shapley supercluster and therefore probably in the early stage of cluster - cluster mergers which destroy the inner cusp .
it is also worth noting that the highest value of anisotropy ( @xmath105 ) is obtained for a0496 which has the highest value of line - of - sight kurtosis ( see table 1 ) as expected for non - rotating systems ( merritt 1987 ) .
= 7.2 cm the relation between the virial mass and concentration of dark matter haloes is a well established result of @xmath1-body simulations and has been studied by many authors ( e.g. nfw ; bullock et al .
2001 ; dolag et al .
2004 ) . in this subsection
we address a question whether our best - fitting parameters of the clusters agree with the trend of concentration decreasing with mass as found in @xmath1-body simulations
. we will compare our results to those of bullock et al .
( 2001 ) because they used the same definition of the virial radius . in fig .
[ cm ] we plot estimated concentration @xmath55 as a function of cluster virial mass @xmath57 ( in units of @xmath106 ) for our six clusters .
the @xmath75 errors following from sampling errors of velocity moments were assigned to the points .
the solid line shows the approximation of the @xmath107 relation calculated from the toy model proposed by bullock et al .
( 2001 ) which reproduces well the properties of a large sample of haloes found in their @xmath1-body simulations .
the predictions were obtained for the standard cosmological model as in our simulations and @xmath108 from their equations ( 9)-(13 ) with parameters @xmath109 and @xmath110 as advertised for masses @xmath111 .
although the errors in the estimated parameters are quite large it is clear from fig .
[ cm ] , that the data for clusters are consistent with the @xmath112 relation found in @xmath1-body simulations . to make this statement more quantitative
we have fitted to the data ( neglecting the errors in mass ) a linear relation of the form @xmath113 + b $ ] and found the best - fitting parameters @xmath114 and @xmath115 ( at 68 percent confidence level ) .
this best - fitting relation is shown in fig .
[ cm ] as a dashed line .
therefore , although the best - fitting slope is negative , the data are also consistent at 1@xmath116 level with constant concentration or even concentration increasing with mass .
= 7.1 cm = 7.2 cm the low number of galaxies per cluster in our sample results in rather large sampling errors in the measured velocity moments and therefore also large errors in the estimated parameters of the clusters . in order to reduce the uncertainties and study a typical cluster we have combined the position and velocity data for our six objects to create a composite cluster .
we have normalized the galaxy distances by the estimated virial radius of their cluster , and the velocities by the circular velocity at the virial radius @xmath117 . in this way
we make the velocity moments independent of the virial mass .
the normalized velocities and velocity moments ( calculated with 121 galaxies per bin ) for the composite cluster made of 1465 galaxies are shown in fig . [ comp ] .
the kurtosis - like variable is now @xmath118^{1/10}$ ] where the coefficient 2.90 was adjusted to the number of galaxies per bin .
the combined samples of galaxies were those after removal of interlopers in each cluster , but we perform the procedure ( fitting velocity dispersion profile and adjusting concentration while keeping @xmath23 ) again on the total sample and it turns out that two more interlopers have to be removed .
the two galaxies have been marked as before with open circles in the upper panel of fig .
the two solid lines in this panel plot the last iteration of @xmath72 profiles separating the interlopers from the galaxies included as members of the cluster .
with the mass dependence factored out , the velocity moments of the composite cluster depend only on two parameters : concentration @xmath55 and anisotropy @xmath25 .
we adjust the parameters first by fitting only the velocity dispersion profile .
the best fit is then at @xmath119 and @xmath120 with @xmath121 .
the @xmath75 , @xmath79 and @xmath80 probability contours corresponding to @xmath122 are shown in the upper panel of fig .
[ betac ] with the dot marking the pair of best - fitting parameters .
we can see that the @xmath75 confidence region is not restricted to the vicinity of the best fit but appears also at more negative @xmath25 and larger @xmath55 , towards the second local minimum at @xmath123 and @xmath124 with @xmath125 .
fitting both the velocity dispersion and kurtosis breaks this degeneracy but only to some extent : the best fit is now at @xmath126 and @xmath127 with @xmath128 , but another local minimum of @xmath129 is found at @xmath130 and @xmath131 .
the presence of the two minima can be traced to the parameters estimated for each cluster separately in the last section . in our sample
we had three clusters with concentration around 3 and three clusters with concentrations of the order of 10 .
those more concentrated also had lower mean value of @xmath25 .
indeed , performing similar fitting for composite clusters made separately of these two groups we find single minima at @xmath132 and @xmath133 for low - concentration clusters while @xmath134 and @xmath135 for high - concentration clusters .
the profiles of the normalized moments for the best - fitting parameters are plotted as solid lines in the middle and lower panel of fig .
the solutions corresponding to the second local @xmath74 minimum are plotted as dashed lines .
recall , that these are best - fitting profiles obtained from the joint fit of dispersion and kurtosis and not to each of them separately this is why the fitted kurtosis profile does not match the corresponding data perfectly .
actually the best fit to the kurtosis data alone would have @xmath25 closer to zero ( isotropic orbits ) , but the velocity dispersion forces @xmath25 towards more radial or more tangential values .
therefore , to summarize our results for the composite cluster , we find @xmath2 and @xmath3 at the 95 percent confidence level .
we performed a joint analysis of velocity dispersion and kurtosis profiles of galaxies in six nearby relaxed galaxy clusters estimating two global parameters of their mass distribution and the anisotropy of galaxy orbits .
the method of joint fitting the moments was tested on simulated dark matter haloes and a new procedure for interloper removal was applied . with the presently available number of galaxy redshifts per cluster the errors in the estimated parameters are still large .
the estimates of mass have an error from 20 percent in the case of a0496 to 50 percent in the case of a3158 and are in rough agreement with previous estimates ( e.g. girardi et al .
1998 ; rines et al .
the uncertainties in the estimates of concentration are even larger .
the fitted parameters are consistent with the mass - concentration relation found in @xmath1-body simulations . the method has been first applied a few years ago to the coma cluster ( okas & mamon 2003 ) for which the largest number of galaxy redshifts is available .
this larger amount of data allowed for a more detailed analysis : the dark matter distribution could be modelled separately from stars and gas , the luminosity density profile could be used as a tracer distribution without assuming that galaxies trace the total or dark matter distribution , galaxies identified as members of binaries could be removed from the sample and the analysis could be restricted to early type galaxies which are believed to be more relaxed .
for the present sample of galaxy clusters we had smaller numbers of galaxies and had to simplify the analysis . however , we believe that the clusters discussed here are really relaxed which appears not to be the case for coma ( which is probably the product of a recent merger as suggested by its perturbed gas distribution and the presence of two cd galaxies ) . in addition , in the present study we have improved the method by introducing a reliable new procedure for the removal of interlopers
. inclusion of kurtosis in the analysis allowed us to constrain the anisotropy and thereby break the degeneracy between the mass distribution and anisotropy present in the analysis of velocity dispersion .
this led us to expect that very tight constraints on the concentration parameter and anisotropy could be obtained if the analysis was performed on a composite cluster made of galaxies belonging to all six clusters .
although the anisotropy was indeed constrained also in this case some degeneracy still persists between @xmath55 and @xmath25 and almost equally good fits can be found for very different pairs of the two parameters .
the degeneracy may however also be due to the rather simple form of anisotropy which we assumed to be constant with radius .
although isotropic orbits ( @xmath23 ) can very well be a realistic case for virialized , mostly elliptical galaxies in clusters , our samples contained also spirals which probably are on more radial orbits ( biviano & katgert 2004 ) .
while isotropy might be preserved in the centre , our composite cluster may be better described by an anisotropy parameter increasing with radius . this possibility will be studied elsewhere .
the critical factor in studies of galaxy kinematics in clusters is the number of available redshift measurements . since the number of such measurements
typically does not exceed a few hundred , while the errors in the estimated parameters are mainly due to sampling errors , as we have shown , the future of such studies will probably still be in analyzing composite clusters .
this does not allow however to study relations between individual objects , like the mass - concentration or mass - temperature relation .
those have been reserved , till recently , to studies based on x - ray observations . although much tighter constraints e.g. on the concentration parameter can be obtained in this case ( pointecouteau , arnaud & pratt 2005 ) , these studies rely on assumptions like the hydrostatic equilibrium and rather uncertain temperature profiles of the x - ray emitting gas . they should therefore be complemented by extensive analysis of galaxy kinematics in clusters . this may soon become possible with systematic photometric and spectroscopic surveys like wings ( fasano et al .
we wish to thank m. moles , k. rines and the referee , a. biviano , for their comments on this work .
computer simulations used in this paper were performed at the leibnizrechenzentrum ( lrz ) in munich .
e and rw are grateful for the hospitality of astrophysikalisches institut potsdam , institut dastrophysique de paris and instituto de astrofsica de andalucia where part of this work was done . rw acknowledges the summer student program at copernicus center . 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 .
we have also used the leda database ( http://leda.univ-lyon1.fr ) .
this work was partially supported by the polish ministry of scientific research and information technology under grant 1p03d02726 as well as the jumelage program astronomie france pologne of cnrs / pan and the exchange program of csic / pan .
bardelli s. , pisani a. , ramella m. , zucca e. , zamorani g. , 1998 , mnras , 300 , 589 basilakos s. , plionis m. , yepes g. , gottlber s. , turchaninov v. , 2005 , mnras , in press , astro - ph/0505620 betancort - rijo j. e. , sanchez - conde m. a. , prada f. , patiri s. g. , 2005 , submitted to apj , astro - ph/0509897 binney j. , mamon g. a. , 1982 , mnras , 200 , 361 biviano a. , girardi m. , 2003 , apj , 585 , 205 biviano a. , katgert p. , 2004 ,
a&a , 424 , 779 biviano a. , salucci p. , 2005
, submitted to a&a , astro - ph/0511309 bullock j. s. , kolatt t. s. , sigad y. , somerville r. s. , kravtsov a. v. , klypin a. a. , primack j. r. , dekel a. , 2001 , mnras , 321 , 559 carlberg r. g. et al .
1997 , apjl , 485 , 13 dantas c. c. , de carvalho r. r. , capelato h. v. , mazure a. , 1997 apj , 485 , 447 diemand j. , moore b. , stadel j. , 2004 , mnras , 352 , 535 dolag k. , bartelmann m. , perrotta f. , baccigalupi c. , moscardini l. , meneghetti m. , tormen g. , 2004 , a&a , 416 , 853 durret f. , adami c. , gerbal d. , pislar v. , 2000 , a&a , 356 , 815 escalera e. , biviano a. , girardi m. , giuricin g. , mardirossian f. , mazure a. , mezzetti m. , 1994 , apj , 423 , 539 fasano g. , et al . , 2005 ,
a&a , in press , astro - ph/0507247 fitchett m. , merritt d. , 1988 , apj , 335 , 18 girardi m. , giuricin g. , mardirossian f. , mezzetti m. , boschin w. , 1998 , apj , 505 , 74 goto t. , 2005 , mnras , 359 , 1415 horner d. j. , baumgartner w. h. , gendreau k. c. , mushotzky r. f. , loewenstein m. , molnar s. m. , 2000 , 197th aas meeting , bulletin of the american astronomical society , 32 , 1581 jones c. , forman w. , 1999 , apj , 511 , 65 katgert p. , biviano a. , mazure a. , 2004 , apj , 600 , 657 kazantzidis s. , kravtsov a. v. , zentner a. r. , allgood b. , nagai d. , moore b. , 2004 , apjl , 611 , 73 klypin a. , hoffman y. , kravtsov a. v. , gottlber s. , 2003 , apj , 596 , 19 okas e. l. , 2002 , mnras , 333 , 697 okas e. l. , hoffman y. , 2001 , in spooner n. j. c. , kudryavtsev v. , eds , proc .
3rd international workshop , the identification of dark matter .
world scientific , singapore , p. 121 okas e. l. ,
mamon g. a. , 2001 , mnras , 321 , 155 okas e. l. , mamon g. a. , 2003 , mnras , 343 , 401 okas e. l. , mamon g. a. , prada f. , 2005a , mnras , 363 , 918 okas e. l. , prada f. , wojtak r. , moles m. , gottloeber s. , 2005b , mnras letters , in press , astro - ph/0507508 mahdavi a. , geller m. j. , 2004 , apj , 607 , 202 mamon g. a. , okas e. l. , 2005 , mnras , 363 , 705 merritt d. , 1987 , apj , 313 , 121 navarro j. f. , frenk c. s. , white s. d. m. , 1997 , apj , 490 , 493 pointecouteau e. , arnaud m. , pratt g. w. , 2005 , a&a , 435 , 1 prada f. , klypin a. a. , simonneau e. , betancort - rijo j. , patiri s. , gottloeber s. , sanchez - conde m. a. , 2005 , submitted to apj , astro - ph/0506432 rines k. , mahdavi a. , geller m. j. , diaferio a. , mohr j. j. , wegner g. , 2001 , apj , 555 , 558 rines k. , geller m. j. , diaferio a. , mahdavi a. , mohr j. j. , wegner g. , 2002 , aj , 124 , 1266 rines k. , geller m. j. , kurtz m. j. , diaferio a. , 2003 , aj , 126 , 2152 sanchis t. , okas e. l. , mamon g. a. , 2004 , mnras , 347 , 1198 van der marel r. p. , magorrian j. , carlberg r. g. , yee h. k. c. , ellingson e. , 2000 , aj , 119 , 2038 wojtak r. , okas e. l. , gottlber s. , mamon g. a. , 2005 , mnras , 361 , l1 | we study the mass distribution in six nearby ( @xmath0 ) relaxed abell clusters of galaxies a0262 , a0496 , a1060 , a2199 , a3158 and a3558 . given the dominance of dark matter in galaxy clusters we approximate their total density distribution by the nfw formula characterized by virial mass and concentration .
we also assume that the anisotropy of galactic orbits is reasonably well described by a constant and that galaxy distribution traces that of the total density . using the velocity and position data for 120 - 420 galaxies per cluster we calculate , after removal of interlopers , the profiles of the lowest - order even velocity moments , dispersion and kurtosis .
we then reproduce the velocity moments by jointly fitting the moments to the solutions of the jeans equations .
including the kurtosis in the analysis allows us to break the degeneracy between the mass distribution and anisotropy and constrain the anisotropy as well as the virial mass and concentration .
the method is tested in detail on mock data extracted from @xmath1-body simulations of dark matter haloes .
we find that the best - fitting galactic orbits are remarkably close to isotropic in most clusters . using the fitted pairs of mass and concentration parameters for the six clusters we conclude that the trend of decreasing concentration for higher masses found in cosmological @xmath1-body simulations is consistent with the data .
by scaling the individual cluster data by mass we combine them to create a composite cluster with 1465 galaxies and perform a similar analysis on such sample .
the estimated concentration parameter then lies in the range @xmath2 and the anisotropy parameter in the range @xmath3 at the 95 percent confidence level .
galaxies : clusters : general galaxies : clusters : individual : a0262 , a0496 , a1060 , a2199 , a3158 , a3558 galaxies : kinematics and dynamics cosmology : dark matter |
drell - yan(dy ) production of lepton pairs @xcite from nucleons and nuclear targets is an important tool to study the quark structure of nucleons and its modification in the nuclear medium .
in particular , the proton induced dy production of muon pairs on nucleons and nuclei provides a direct probe to investigate the quark parton distribution functions(pdfs ) . in a dy process(shown in fig.[fig0 ] ) , a quark of beam(target ) hadron gets annihilated from the antiquark of target(beam ) hadron and gives rise to a photon which in turn gives lepton pairs of opposite charge . the basic process is @xmath2 , where b and t indicate the beam proton and the target nucleon / hadron .
a quark(antiquark ) in the beam carrying a longitudinal momentum fraction @xmath3 interacts with an antiquark(quark ) in the target carrying longitudinal momentum fraction @xmath4 of the target momentum per nucleon to produce a virtual photon .
the cross section per target nucleon @xmath5 in the leading order is given by @xcite : @xmath6 where @xmath7 is the fine structure constant , @xmath8 is the charge of quark / antiquark of flavor f , @xmath9 is the photon virtuality and @xmath10 and @xmath11 are the beam(target ) quark / antiquark pdfs of flavour f. this process is directly sensitive to the antiquark parton distribution functions @xmath12 in target nuclei which has also been studied by dis experiments through the observation of emc effect .
quantitatively the emc effect describes the nuclear modification of nucleon structure function @xmath13 for the bound nucleon defined as @xmath14 $ ] and gives information about the modification of the sum of quark and antiquark pdfs @xcite which is dominated by the valence quarks in the high @xmath4 region ( @xmath15 ) . in the low @xmath4 region ( @xmath16 ) , where sea quarks are expected to give dominant contribution , the study of @xmath13 gives information about sea quark and antiquark pdfs .
thus , nuclear modifications are phenomenologically incorporated in @xmath17 and @xmath18 using the experimental data on @xmath13 and are used to analyze the dy yields from nuclear targets .
some authors succeed in giving a satisfactory description of dis and dy data on nuclear targets using same set of nuclear @xmath19 and @xmath12 @xcite , while some others find it difficult to provide a consistent description of dis and dy data using the same set of nuclear pdfs @xcite . on the other hand
, there are many theoretical attempts to describe the nuclear modifications of quark and antiquark pdfs to explain dis which have also been used to understand the dy process on nuclear targets @xcite-@xcite .
the known nuclear modifications discussed in literature in the case of dis are ( a ) modification of nucleon structure inside the nuclear medium , ( b ) a significantly enhanced contribution of subnucleonic degrees of freedom like pions or quark clusters in nuclei and ( c ) nuclear shadowing .
vs @xmath3 at e=800gev(@xmath20=38.8gev ) .
spectral function : dashed line , including mesonic contribution : dashed - dotted line and the results obtained using the full model i.e. spectral function+meson cloud contributions+shadowing effects+energy loss : solid line .
the results in the different columns are obtained at different values of m(=@xmath21 ) .
experimental points are data of e772 experiment @xcite . ] however , in the case of dy processes there is an additional nuclear effect due to initial state interaction of beam partons with the target partons which may be present before the hard collisions of these partons giving rise to lepton pairs . as the initial beam traverses the nuclear medium it loses energy due to interaction of beam partons with nuclear constituents of the target .
this can be visualized in terms of the interaction of hadrons or its constituents with the constituents of the target nucleus through various inelastic processes leading to energy loss of the interacting beam partons .
this has been studied phenomenologically using available parameterization of nuclear pdfs or theoretically in models based on qcd or glauber approaches taking into account the effect of shadowing which also plays an important role in the low @xmath4 region , however , any consensus in the understanding of physics behind the beam energy loss has been lacking . in this scenario
most of the calculations incorporate a phenomenological description of beam energy loss to explain the experimental data on dy yields @xcite . in this region of @xmath4 the nuclear modification of sea quark pdf and mesonic contributions also become important . thus in this process ,
main nuclear effects are due to nuclear structure , mesonic contributions and shadowing ( as in the case of dis ) with additional effect of parton energy loss in the beam parton energy due to the presence of nuclear targets .
vs @xmath3 at e=800gev(@xmath20=38.8gev ) .
lines and points have the same meaning as in fig.[fig1 ] ] in this paper , we present the results of nuclear medium effects on dy production of lepton pairs calculated in a microscopic nuclear model which has been successfully used to describe the dis of charged leptons and @xmath22/@xmath23 from various nuclei @xcite .
the model uses a relativistic nucleon spectral function to describe target nucleon momentum distribution incorporating fermi motion , binding energy effects and nucleon correlations in a field theoretical model .
the model has also been used to include the mesonic contributions from @xmath0 and @xmath1 mesons .
the beam energy loss has been calculated using some phenomenological models discussed in the literature @xcite-@xcite .
the results have been presented for the kinematic region of experiments e772 @xcite and e866 @xcite or proton induced dy processes in nuclear targets like @xmath24 , @xmath25 , @xmath26 , @xmath27 and @xmath28 in the region of @xmath29 .
the numerical results extended up to @xmath30 , should be useful in analyzing the forthcoming experimental results from the seaquest e906 experiment being done at fermi lab @xcite . in section-[sec :
ne ] , we present the formalism in brief ; in section-[sec : rd ] , the results are presented and discussed ; and finally in section-[sec : cc ] , we summarize the results and conclude our findings .
vs @xmath3 at e=800gev(@xmath20=38.8gev ) .
lines and points have the same meaning as in fig.[fig1 ] ]
when dy process takes place in nuclei , nuclear effects appear which are generally believed to be due to + ( a ) nuclear structure effects arising from fermi motion , binding energy and nucleon correlations , + ( b ) additional contribution due to subnucleonic degrees of freedom like mesons and/or quark clusters in the nuclei , + ( c ) shadowing effect , and + ( d ) energy loss of the beam proton as it traverses the nuclear medium before producing lepton pairs . in the case of proton
induced dy processes in nuclei , the target nucleon has a fermi momentum described by a momentum distribution .
the target bjorken variable @xmath4 is defined for a free nucleon as @xmath31 , where q is the four momentum of @xmath32 pair , @xmath33 and @xmath34 are respectively the beam and target four momenta in the nuclear medium . moreover ,
the projectile bjorken variable @xmath3 expressed covariantly as @xmath35 also changes due to the energy loss of the beam particle caused by the initial state interactions with the nuclear constituents as it travels through the nuclear medium before producing lepton pairs .
these nuclear modifications are incorporated while evaluating eq.([dy1 ] ) . furthermore , there are additional contributions from the pion and rho mesons which are also taken into account . in the following ,
we briefly outline the model and refer to earlier work @xcite for details . in a nucleus
, scattering is assumed to take place from partons inside the individual nucleons which are bound and moving with a momentum @xmath36 within a limit given by the fermi momentum .
the target bjorken variable @xmath4 becomes fermi momentum dependent and pdf for quarks and antiquarks in the nucleus i.e. @xmath37 and @xmath38 are calculated as a convolution of the pdfs in bound nucleon and a momentum distribution function of the nucleon inside the nucleus .
the parameters of the momentum distribution are adjusted to correctly incorporate nuclear properties like binding energy , fermi motion and the nucleon correlation effects in the nuclear medium .
we use the lehmann representation of the relativistic dirac propagator for an interacting fermi sea in nuclear matter to derive such a momentum distribution and local density approximation to translate these results for a finite nucleus @xcite .
the free relativistic propagator for a nucleon of mass @xmath39 is written in terms of positive and negative energy components as @xmath40 for a noninteracting fermi sea where only positive energy solutions are considered the relevant propagator is rewritten in terms of occupation number @xmath41 for p@xmath42 while @xmath43=0 for p@xmath44 : @xmath45\right\}\end{aligned}\ ] ] the nucleon propagator in an interacting fermi sea is then calculated by making a perturbative expansion of @xmath46 in terms of free nucleon propagator @xmath47 given in eq .
( [ prop2 ] ) by retaining the positive energy contributions only ( the negative energy components are suppressed ) .
vs @xmath3 at e=800gev(@xmath20=38.8gev ) .
lines and points have the same meaning as in fig.[fig1 ] ] this perturbative expansion is then summed in ladder approximation to give dressed nucleon propagator @xmath46 @xcite @xmath48 where @xmath49 is the nucleon self energy .
this allows us to write the relativistic nucleon propagator in a nuclear medium in terms of the spectral functions of hole and particle as @xcite @xmath50\,\end{aligned}\ ] ] where @xmath51 and @xmath52 being the hole and particle spectral functions respectively , which are derived in ref .
@xcite , and @xmath53 is the chemical potential .
we use : @xmath54 for @xmath55 @xmath56 for @xmath57 .
the normalization of this spectral function is obtained by imposing the baryon number conservation following the method of frankfurt and strikman @xcite . in the present paper ,
we use local density approximation ( lda ) where we do not have a box of constant density , and the reaction takes place at a point @xmath58 , lying inside a volume element @xmath59 with local density @xmath60 and @xmath61 corresponding to the proton and neutron densities at the point @xmath58 .
this leads to the spectral functions for the protons and neutrons to be the function of local fermi momentum given by @xmath62^{1/3},~~k_{f_n}({\bf r})= \left [ 3\pi^{2 } \rho_{n}({\bf r})\right]^{1/3}\end{aligned}\ ] ] and therefore the normalization condition may be imposed as @xmath63 where the factor of two is to take into account spin degrees of freedom of proton and neutron , and @xmath64 and @xmath65 are the chemical potentials for proton and neutron respectively .
this further leads to the normalization condition given by @xmath66 the average kinetic and total nucleon energy in a nucleus with the same number of protons and neutrons are given by : @xmath67 @xmath68 where @xmath69 is the baryon density for the nucleus which is normalized to a and is taken from the electron nucleus scattering experiments . the binding energy per nucleon
is given by @xcite : @xmath70 for an isospin symmetric nucleus is derived to be @xcite : @xmath71 using this , the electromagnetic structure function @xmath72 for a non - symmetric ( n@xmath73z ) nucleus in dis is obtained as @xcite , @xmath74 \nonumber\\\end{aligned}\ ] ] for the numerical calculations , we have used cteq6.6 @xcite nucleon parton distribution functions(pdfs ) for quark(@xmath75 ) and antiquark(@xmath76 ) of flavor f. following the same procedure as taken for the evaluation of nuclear structure function , we incorporate the nuclear medium effects like fermi motion , binding energy and nucleon correlations in the evaluation of bound quarks in nucleons of a nucleus . @xmath77 and @xmath78 are expressed in terms of spectral function as @xcite : @xmath79 where @xmath80 is the quark(antiquark ) pdfs for flavor f inside a nucleon of kind i and the factor of 2 is because of quark(antiquark ) spin degrees of freedom . @xmath81 which is obtained from the covariant expression of @xmath82 with @xmath83 direction .
vs @xmath3 at e=800gev(@xmath20=38.8gev ) .
the results are obtained using the full model at m=4.5gev .
a stands for several nuclei like @xmath25 , @xmath26 , @xmath27 and @xmath28 .
these results are obtained using different models for the energy loss viz .
@xmath84 in eq.([alpha ] ) shown by the solid line , @xmath85 in eq.([aalpha ] ) shown by the solid line with stars and @xmath86 in eq.([beta ] ) shown by the dashed line .
experimental points are data of e772 experiment @xcite . ]
as the nucleons are strongly interacting particles and inside the nucleus continuous exchange of virtual mesons take place , therefore , we have also taken into account the probability of interaction of virtual photons with the meson clouds .
in the present work , we have considered @xmath0 and @xmath1 mesons . for this
the imaginary part of the meson propagators are introduced instead of spectral function which were derived from the imaginary part of the nucleon propagator .
therefore , in the case of pion , we replace in eq.([f2anuclei ] ) @xcite : @xmath87 where @xmath88 is the pion propagator in the nuclear medium given by @xmath89^{- 1}\,,\ ] ] with @xmath90 here , @xmath91 is the @xmath92 form factor , @xmath93=1gev , @xmath94 , @xmath95 is the longitudinal part of the spin - isospin interaction and @xmath96 is the irreducible pion self energy that contains the contribution of particle - hole and delta - hole excitations . following a similar procedure , as done in the case of nucleon , the contribution of the pions to hadronic tensor in the nuclear medium may be written as @xcite @xmath97 however , eq.([w2pion ] ) also contains the contribution of the pionic contents of the nucleon , which are already contained in the sea contribution of nucleon through eq.([qbara ] ) , therefore , the pionic contribution of the nucleon is to be subtracted from eq.([w2pion ] ) , in order to calculate the contribution from the excess pions in the nuclear medium .
this is obtained by replacing @xmath98 by @xmath99 @xcite as @xmath100 using eq.([w2pion ] ) , pion structure function @xmath101 in a nucleus is derived as @xmath102 \ ; \theta ( x_\pi - x ) \ ; \theta ( 1 - x_\pi),\ ] ] where @xmath103 .
this in turn leads to the expression for the pion quark pdf in the nuclear medium .
for example , @xmath104 is derived as @xcite : @xmath105 and a similar expression for @xmath106 .
similarly , the contribution of the @xmath1-meson cloud to the structure function is taken into account in analogy with the above prescription and the rho structure function is written as @xcite @xmath107 \theta ( 1 - x_{\rho})\theta ( x_\rho - x)\ ] ] and the expression for the rho pdf @xmath108 is derived as @xcite : @xmath109 where @xmath110 is now the @xmath1-meson propagator in the nuclear medium given by : @xmath111^{- 1}\,,\ ] ] where @xmath112 here , @xmath113 is the transverse part of the spin - isospin interaction , @xmath114 , @xmath115 is the @xmath116 form factor , @xmath117=1gev , @xmath94 , and @xmath96 is the irreducible rho self energy that contains the contribution of particle - hole and delta - hole excitations and @xmath118 .
quark and antiquark pdfs for pions have been taken from the parameterization given by gluck et al.@xcite and for the rho mesons we have taken the same pdfs as for the pions .
it must be pointed out that the choice of @xmath93 and @xmath117(=1gev ) in @xmath0nn and @xmath1nn form factors have been fixed in our earlier works@xcite while describing nuclear medium effects in electromagnetic structure function @xmath119 to explain the latest data from jlab and other experiments performed using charged lepton beams on several nuclear targets .
we have also taken into account shadowing effect which arises due to coherent multiple scattering interactions of the intermediate states , which is important in dy production at small @xmath4 .
various theoretical calculations have indicated that shadowing in dis as well as in dy processes has a common origin . for the shadowing effect
we have followed the model of kulagin and petti @xcite .
the incident proton beam traverses the nuclear medium before the beam parton undergoes a hard collision with the target parton . the incident proton may lose energy due to soft inelastic collisions as it might scatter on its way within the nucleus before producing a lepton pair .
there are many papers in literature @xcite where the effect of energy loss on dy process is discussed and models are given to incorporate them in the calculation of dy yields . however , there is no model which has the preference over the others .
most of them perform phenomenological fits and the best value of the parameters are those which have been obtained in the independent analysis of the experimental data .
the present situation is summarized by accardi et al.@xcite .
for example duan et al .
@xcite have used two different kinds of quark energy loss expression , in which the fractional parton energy @xmath3 is modified to @xmath120 , where in the linear fit @xmath121 are given by @xmath122 and by @xmath123 where @xmath124$]fm is the average path length of the incident quark in the nucleus a , @xmath125 is the energy of the incident proton .
the constants @xmath7 and @xmath126 are varied to get a good fit with the experimental data which were found to be in the range of @xmath127 gev / fm and @xmath128 @xmath129 @xcite .
gavin and milana @xcite have parameterized the energy loss effect as @xmath130 where @xmath86 . however , in some recent work of johnson et al .
@xcite , garvey and peng @xcite , and kulagin and petti @xcite , it has been pointed out that a quantitative estimate of energy loss effect in dy processes depends upon how the shadowing effect is treated . in the presence of shadowing effect
, the fitted parameter for energy loss alpha in equation [ alpha ] is found to be somewhat smaller in the range of @xmath131 to @xmath132 .
we have taken into account the various nuclear effects discussed above in this section and write the cross section for the dy process as @xmath133 where @xmath134 is the dy cross section from the nucleons in the nucleus after incorporating the nuclear medium effects like fermi motion , binding energy , nucleon correlations through the use of spectral function .
furthermore , we have also incorporated shadowing effect following kulagin and petti @xcite and energy loss effect following the phenomenological model given in eq .
[ alpha ] with @xmath7=1 .
the expression for @xmath134 is given by : @xmath135 \theta ( x_t^\prime ) \theta ( 1 - x_t^\prime)~\theta ( 1 - x_b)\end{aligned}\ ] ] where @xmath136 is the hole spectral function for the nucleon in the nucleus .
@xmath137 and @xmath138 are the nucleon pdfs of flavor f averaged over proton and neutron in the cases of quarks and antiquarks , respectively .
vs @xmath3 at @xmath139 for @xmath140 , @xmath26 , @xmath27 and @xmath28 . for the beam energy e=120gev(@xmath20=15gev ) the results
are obtained with spectral function : dotted line , including the mesonic contribution : dashed line , and for the full calculation : solid line . ]
vs @xmath4 at e=800gev(@xmath20=38.8gev ) , @xmath141 , @xmath142 , with @xmath84 in eq.([alpha ] ) .
spectral function : dashed line , including the mesonic contribution : dashed - dotted line , results of the full calculation : solid line .
experimental points are of e772 experiment @xcite .
right panel : @xmath143 vs @xmath4 , lines have same meaning as in the left panel . ] vs @xmath4 at e=120gev(@xmath20=15gev ) , @xmath141 , @xmath142 .
the results are obtained with spectral function : dotted line , spectral function+mesonic contribution : dashed line .
the results of our full calculations are obtained with energy loss using eq.([alpha ] ) with @xmath84 ( solid line ) and @xmath85 in eq.([aalpha ] ) ( solid line with stars ) .
right panel : @xmath143 vs @xmath4 , lines have same meaning as in the left panel . ]
similarly to include the pionic contribution @xmath144 and the rho contribution @xmath145 , the dy cross sections are respectively written as @xcite : @xmath146 \theta ( x_{\pi } - x_t ) \ ; \theta ( 1 - x _ { \pi } ) \ ; \theta ( 1 - x_b)\end{aligned}\ ] ] and @xmath147 \theta ( x _ { \rho } - x_t ) \ ; \theta ( 1 - x _ { \rho } ) \ ; \theta ( 1 - x_b ) \end{aligned}\ ] ] since in the various experiments the dy cross sections are also obtained in terms of other variables like @xmath148 , @xmath149 , @xmath150 , etc , where , @xmath151 , @xmath152 , @xmath153 , therefore , we have also obtained dy cross sections in terms of some of these variables .
for example , using jacobian transformation eq.([spec ] ) may be written as : @xmath154 \theta ( x_t^\prime ) \theta ( 1 - x_t^\prime)~\theta ( 1 - x_b)\end{aligned}\ ] ] most of the experimental results for the dy process have been presented in the form of @xmath155 i.e. the ratio of dy cross section in a nuclear target ( @xmath156 ) to the dy cross section in deuteron ( @xmath157 ) . therefore , to evaluate proton - deuteron dy cross section , we write @xmath158 and to take into account the deuteron effect , the quark / antiquark distribution function inside the deuteron target have been calculated using the same formula as for the nuclear structure function but performing the convolution with the deuteron wave function squared instead of using the spectral function .
the deuteron wave function has been taken from the works of lacombe et al .
@xcite . in terms of the deuteron wave function
, one may write @xmath159 where the four momentum of the proton inside the deuteron is described by @xmath160 with @xmath161 as the energy of the off shell proton inside the deuteron and @xmath162 is the deuteron mass .
a similar expression has been used for the antiquarks @xmath78 .
vs @xmath163 ,
right panel : @xmath164 vs @xmath165gev , at e=800gev(@xmath20=38.8gev ) , with @xmath84 in eq.([alpha ] ) .
experimental points are of e866 experiment @xcite with @xmath166 , @xmath167 and @xmath168 .
spectral function : dashed line , including the mesonic contribution : dashed - dotted line , results of the full calculation : solid line . ]
the results presented here are based on the following calculations : \(1 ) dy cross section for proton - nucleus scattering i.e. @xmath169 , where a stands for a nuclear target , has been obtained by using the spectral function @xmath136 which takes into account fermi motion , nucleon correlations and binding energy . the spectral function with parameters fixed by eqs.([norm2 ] ) and
( [ norm4 ] ) has been used to calculate the nucleon contribution which reproduce mass number of the nucleon , the binding energy per nucleon for a given nucleus .
( 2 ) we add contributions obtained from the pion cloud using eq.([spec1 ] ) and eq.([specrho ] ) for the rho meson to nucleon contributions . for evaluating the mesonic contributions the parameters of @xmath88 in eq.([dpi ] ) and
@xmath110 in eq .
( [ dro ] ) are fixed by fitting experimental data on @xmath170 in dis on various nuclei @xcite .
\(3 ) we have also included shadowing effect following the works of kulagin and petti @xcite . with the inclusion of shadowing effect along with the spectral function and meson cloud contributions , the numerical results have been presented .
vs @xmath3 at different @xmath165 with e=800gev(@xmath20=38.8gev ) , and @xmath84 in eq.([alpha ] ) .
experimental points are of e866 experiment @xcite .
spectral function : dashed line , including the mesonic contribution : dashed - dotted line , results of the full calculation : solid line ] \(4 ) for the energy loss we have used eq.([alpha ] ) with @xmath84 .
there are other phenomenological models available in the literature to take into account the energy loss effect .
we have , therefore , studied the dependence of dy cross sections on energy loss if one uses other phenomenological parameterizatios .
some of the expressions are given in eq.([aalpha ] ) and eq.([beta ] ) .
for our numerical calculations we have taken @xmath126=0.2 .
also we have performed calculations using eq.([beta ] ) with @xmath86 .
\(5 ) dy cross section for proton - deuteron scattering has been obtained using eq.([f2a_deut1 ] ) and no energy loss effect has been taken in deuteron .
we have obtained the results by using eq.([f2a_deut ] ) with deuteron effect which has found to be small .
\(6 ) for nucleon quark / antiquark pdfs cteq6.6 @xcite has been used and for pion quark / antiquark pdfs parameterization of gluck et al .
@xcite has been used .
vs @xmath3 at different @xmath165 with @xmath20=38.8gev , with @xmath84 in eq.([alpha ] ) .
experimental points are of e866 experiment @xcite .
spectral function : dashed line , including the mesonic contribution : dashed - dotted line , results of the full calculation : solid line . ] in figs.[fig1]-[fig4 ] , we present the results for the ratio @xmath171 vs @xmath3 for m=4.5 , 5.5 , 6.5 and 7.5gev .
the center of mass energy ( @xmath20 ) is 38.8gev . here
a is @xmath25 in fig.[fig1 ] , @xmath26 in fig.[fig2 ] , @xmath27 in fig.[fig3 ] and @xmath28 in fig.[fig4 ] .
these results are presented for the numerator obtained using the spectral function , including the mesonic effect , and also including shadowing and energy loss effect which is our full model .
we find that the nuclear structure effects due to bound nucleon lead to a suppression in the dy yield of about @xmath172 in @xmath25 in the region of @xmath173 .
this suppression increases with the increase in mass number of the nuclear target .
for example , in @xmath28 it becomes @xmath174 for @xmath173 .
furthermore , we find that there is a significant contribution of mesons which increases the dy ratio i.e. its effect is opposite to the effect of spectral function .
for example , the dy yield increases by around @xmath174 for @xmath173 in @xmath25 .
moreover , we observe that the effect is more at low @xmath3(@xmath175 ) than at high @xmath3 .
this increase in the dy yield from meson cloud contribution also increases with the mass number a , for example in @xmath28 it is around @xmath176 for @xmath177 .
we find the contribution from rho meson cloud to be much smaller than the contribution from pion cloud . when the shadowing corrections are included there is further suppression in the dy yield and it is effective in the low region of @xmath3 .
the effect of beam energy loss is also to reduce the dy yield .
both effects further adds to the suppression obtained using spectral function , where as mesonic effects lead to an enhancement .
the net effect of shadowing and the energy loss effect is @xmath178 at @xmath179 in @xmath25 which becomes @xmath180 at @xmath181 for @xmath139 .
the shadowing effect as well as energy loss effect are more pronounced in heavier nuclei and suppresses the dy ratio considerably .
for example , in @xmath28 at low @xmath3 this suppression is about @xmath182 at @xmath183 which becomes @xmath184 at @xmath185 .
moreover , the shadowing effect is @xmath148 dependent , like in @xmath28 the total effect becomes about @xmath186 at @xmath183 for @xmath187 .
it is observed that the suppression in the dy ratio due to energy loss effect ( not shown in these figures ) is @xmath188 in the case of @xmath25 which increases to @xmath189 in @xmath26 and @xmath27 , and becomes around @xmath190 in @xmath28 in the region of @xmath191 . to observe the effect of energy loss using the various approaches , in fig.[fig66 ] , we present the results of full calculation for @xmath192 at m=4.5gev and @xmath20=38.8gev for @xmath7=1 in eq .
( [ alpha ] ) , @xmath126=0.2 in eq .
( [ aalpha ] ) and @xmath86 in eq.([beta ] ) .
we find that there is hardly any difference in the results obtained with @xmath126=0.2 in eq .
( [ aalpha ] ) in comparison to the results obtained with @xmath7=1 in eq .
( [ alpha ] ) . when we obtain the results using @xmath86 in eq.([beta ] ) , it is found that there is less reduction in the dy ratio for heavier nuclear targets like @xmath28 as compared to the results obtained using @xmath7=1 in eq .
( [ alpha ] ) , while for the light nuclear targets the results are comparable . we have also studied(not shown here ) the dependence of the parameter @xmath7 used in the expression given in eq.([alpha ] ) for energy loss , we have varied @xmath7 in the range @xmath193 , corresponding to the range of values used in literature @xcite , for @xmath192 at m=4.5gev and @xmath20=38.8gev .
we find that in the case of p-@xmath25 dy process , there is less(more ) suppression when @xmath7 is taken as 0.5(2 ) from the results obtained at @xmath7=1 , our reference value .
for example , this is around @xmath194 lesser(larger ) at @xmath195 0.1 - 0.2 at @xmath7=0.5(2 ) . with the increase in mass number the difference in the results increases .
for example , in @xmath28 this suppression is around @xmath196 lesser(larger ) at @xmath195 0.1 - 0.2 at @xmath7=0.5(2 ) .
thus we observe that with the increase in @xmath7 , dy ratio decreases considerably and there is an a dependence on the dy ratio .
the numerical results are compared with the e772 experimental data and found to be in fair agreement . to explicitly compare the effect of nuclear medium as well as energy loss in dy production cross section at different center of mass energies , in fig .
[ fig13 ] , we have presented the results for @xmath197 vs @xmath3 at @xmath139 for @xmath140 , @xmath26 , @xmath27 and @xmath28 at e=120gev(@xmath20=15gev ) .
these results depict how the dy ratio vary at the different center of mass energies when the results are obtained with spectral function , including mesonic contribution and the full calculation without energy loss and with energy loss for @xmath84 in eq.([alpha ] ) . for e=120gev , we observe that the effect of spectral function is to reduce dy yield by about @xmath178 for @xmath25 and the reduction increases with the increase in mass number like for @xmath28 it is @xmath198 .
when the meson cloud contributions are included the dy yield get enhanced by @xmath199 at low @xmath3 and @xmath186 at mid and high values of @xmath3 and is found almost independent of @xmath200 . when shadowing and energy loss effects are added , the results get reduced by @xmath201 in the range of @xmath202 and show significant reduction at high values of @xmath3 which is around @xmath203 in @xmath25 .
furthermore , we observe a strong nuclear mass dependence on the dy yield due to energy loss and shadowing effects , for example it is @xmath204 for @xmath28 .
when we compare the present results obtained using full model with the results obtained at e=800gev for m=4.5gev(figs.[fig1]-[fig4 ] ) , it may be observed that the decrease in dy yield is mainly due to the large energy loss at low center of mass energy . in fig.[fig7 ] , we present the results for @xmath205 vs @xmath4 at @xmath20=38.8gev by integrating over @xmath3 . the integration over @xmath3
is done by putting the constraints as @xmath206 and @xmath142 .
we find the effect of spectral function to be around @xmath174 in @xmath25 which increases with mass number and becomes around @xmath207 in @xmath28 , and is found to be almost independent of @xmath4 .
the addition of the mesonic contribution enhances the dy yield , for example , in @xmath25 this increase is @xmath186 which further get enhances with increase in mass number and becomes around @xmath208 in @xmath28 .
when the results are obtained with shadowing and the beam energy loss effect the suppression is the numerical results in @xmath25 is around @xmath209 , and it increases with mass number and becomes around @xmath207 in @xmath28 .
the present theoretical results are also compared with e772 @xcite experimental data . in view of the e906 seaquest experiment
being done at fermi lab , we have presented the results in fig .
[ fig77 ] , for @xmath205 vs @xmath4 , ( @xmath200=@xmath25 , @xmath26 , @xmath27 and @xmath28 ) , at @xmath20=15gev corresponding to the energy of the incident proton e=120gev .
we find that the effect of spectral function to be around @xmath210 in @xmath25 which increases with mass number and becomes around @xmath207 in @xmath28 .
when mesonic effects are included the rise in the dy ratio is around @xmath211 in @xmath25 which increases with mass number and becomes around @xmath212 in @xmath28 . when shadowing and energy loss effects are further added , then there is large reduction which is mainly due to loss of beam energy at low @xmath213 , which increases considerably with the increase in mass number of target nuclei .
for example , for @xmath7=1 in eq .
( [ alpha ] ) the results are reduced by @xmath214 in @xmath25 , @xmath215 in @xmath26 and @xmath27 , and @xmath216 in @xmath28 . to see the dependence of the different approaches of energy loss effect on the dy ratio
, we have also obtained the results using @xmath85 in eq .
( [ aalpha ] ) , and find that for low mass nuclei there is difference in the results of about 8 - 10@xmath217 in @xmath25 , which becomes negligible with the increase in mass number of nuclear target . in e866 experiment @xcite ,
the results were obtained for @xmath218 vs @xmath163 , where @xmath219 and @xmath220 vs @xmath148 , with @xmath151 . using eqs.[spec ] and [ duan1 ] ,
we have obtained the results respectively for @xmath218 vs @xmath163 and @xmath220 vs @xmath148 and shown these results in fig.[fig8 ] . for @xmath218
vs @xmath163 , we have integrated over @xmath3 between the limits @xmath221 and followed the kinematical cuts of @xmath222 gev used in the analysis of e866 @xcite experiment . in the case of @xmath220 vs @xmath148 , we have integrated over @xmath3 between the limits @xmath221 and put the kinematical constraint @xmath223 as used in e866 @xcite experiment .
these results are shown for the dy ratio for @xmath224 vs @xmath163(left panel ) and @xmath225 vs @xmath148(right panel ) , where i stands for iron(top panel ) and tungsten(bottom panel ) nuclei .
these results are obtained with spectral function , including mesonic effect , and using our full model with energy loss effect . for the energy loss ,
we have used eq.([alpha ] ) with @xmath84 .
we observe that the effect of spectral function(@xmath226 ) and meson cloud contributions(@xmath227 ) are small . when shadowing and energy loss effects are included there is a significant reduction in dy yield which is around @xmath228 in @xmath229 and the reduction increases to @xmath176 in @xmath230 .
we find a good agreement with the experimental results for the various dy ratios available from e866 @xcite experiment . in fig .
[ fig9 ] , we present the results of dy ratio for @xmath231 vs @xmath3 for different values of @xmath165 , between the kinematic limits @xmath223 and @xmath221 as used in e866 experiment @xcite .
the results of this ratio for @xmath232 are shown in fig .
[ fig10 ] .
the results are presented to observe the effect for spectral function , meson cloud contributions , and with shadowing effect and energy loss effect on the dy production .
we have studied nuclear medium effects in dy process using quark parton distribution functions and nucleon structure functions for a bound nucleon .
we have used a microscopic nuclear model which takes into account the effect of fermi motion , nuclear binding and nucleon correlations through a relativistic spectral function of bound nucleon .
the contributions of @xmath0 and @xmath1 mesons are also included .
furthermore , shadowing corrections are taken into account .
we have also included the beam energy loss effect due to initial state interactions of protons with nuclear constituents before they suffer hard collisions to produce lepton pair .
we find a reduction in the dy yield due to nuclear structure effects and an enhancement due to mesonic contribution .
the effect of shadowing is to reduce the dy yield in the region of very low @xmath4(@xmath233 ) .
both the reduction as well as the enhancement in the case of dy yields are found to be of similar magnitude as in the case of dis of charged leptons . in the case of dy yields
there is a further reduction due to beam energy loss effect in the nuclear medium which has been treated phenomenologically using a parameter describing the beam energy loss .
the numerical results are compared with the experimental results from e772 @xcite and e866 @xcite experiments .
a reasonable agreement with the experimental results presently available for @xmath25 , @xmath26 , @xmath27 , and @xmath28 has been found .
we have also presented in this paper , results for @xmath5 vs @xmath3 for various values of @xmath165 and the results for @xmath234 vs @xmath4 relevant to the forthcoming e906 seaquest @xcite experiment at fermi lab .
our results show that the model for describing the nuclear medium effects in the dis of charged leptons and neutrino and antineutrino with nuclear targets is able to explain the experimental results in the case of dy yield in the region @xmath235 .
high statistics , high precision data from e906 seaquest @xcite experiment on @xmath5 in various regions of @xmath3 and @xmath4 will provide important information about the modification of quark pdfs and nucleon structure function in the nuclear medium .
m. s. a. is thankful to department of science and technology(dst ) , government of india for providing financial assistance under grant no .
sr / s2/hep-18/2012 .
i.r.s . thanks fis2014 - 59386-p spanish project for financial support and juan de la cierva - incorporacion contract from spanish mineco .
d. f. geesaman , k. saito and a. w. thomas , ann .
nucl . part .
sci . * 45 * , 337 ( 1995 ) .
k. j. eskola , h. paukkunen , c. a. salgado , jhep * 0904 * ( 2009 ) , 065 ; k. j. eskola , v. j. kolhinen , h. paukkunen and c. a. salgado , jhep * 0705 * , 002 ( 2007 ) , s. kumano , phys .
d * 43 * , 59 ( 1991 ) ; * 43 * , 3067 ( 1991 ) , m. hirai , s. kumano and t. -h .
nagai , phys .
c * 76 * , 065207 ( 2007 ) , m. hirai , s. kumano and t. -h .
nagai , phys .
c * 70 * , 044905 ( 2004 ) .
i. schienbein , j.y .
yu , c. keppel , j.g .
morfin , f. olness and j.f .
owens , phys .
d * 77 * , 054013 ( 2008 ) , i. schienbein , j.y .
yu , k. kovarik , c. keppel , j.g .
morfin , f. olness , j.f .
owens , phys .
d * 80 * 094004 ( 2009 ) . c. h. llewellyn smith , phys .
b * 128 * , 107 ( 1983 ) . c. e. carlson and t. j. havens phys .
lett . * 51 * , 261 ( 1983 ) .
f. e. close , r. g. roberts and g. g. ross , phys .
b * 129 * , 346 ( 1983 ) .
m. ericson and a. w. thomas , phys .
b * 148 * , 191 ( 1984 ) .
e. l. berger and f. coester , phys .
d * 32 * , 1071 ( 1985 ) .
h. jung and g. a. miller , phys .
c * 41 * , 659 ( 1990 ) .
f. e. close , r. l. jaffe , r. g. roberts and g. g. ross , phys .
d * 31 * , 1004 ( 1985 ) .
g. e. brown , m. buballa , z. b. li and j. wambach , nucl .
a * 593 * , 295 ( 1995 ) .
e. l. berger , f. coester and r. b. wiringa , phys .
d * 29 * , 398 ( 1984 ) .
r. l. jaffe and x. -d .
ji , phys .
lett . * 67 * , 552 ( 1991 ) .
a. e. l. dieperink and c. l. korpa , phys .
c * 55 * , 2665 ( 1997 ) . c. l. korpa and a. e. l. dieperink , phys .
c 87 , 014616 ( 2013 ) . f. eichstaedt , s. leupold and u. mosel , phys . rev .
d * 81 * , 034002 ( 2010 ) .
e. marco and e. oset , nucl .
phys . a * 645 * , 303 ( 1999 ) .
m. b. johnson et al .
c * 65 * , ( 2002 ) 025203 .
m. b. johnson _ et al .
_ [ fnal e772 collaboration ] , phys . rev .
lett . * 86 * , 4483 ( 2001 ) .
f. arleo , phys .
b * 532 * , 231 ( 2002 ) g. t. garvey and j. c. peng , phys .
* 90 * ( 2003 ) 092302 .
h. haider , f. zaidi , m. sajjad athar , s. k. singh and i. ruiz simo , nucl .
a * 943 * , 58 ( 2015 ) .
h. haider , f. zaidi , m. s. athar , s. k. singh and i. r. simo , nucl .
a(in press ) .
d. m. alde et al .
lett . * 64 * , 2479 ( 1990 ) .
e. a. hawker et al .
( fnal e866/nusea ) , phys .
lett . * 80 * , 3715 ( 1998 ) .
l. l. frankfurt and m. strikman phys . lett.b 183 1987 254 .
pavel m. nadolsky et al . , phys .
d * 78 * , 013004 ( 2008 ) ; http://hep.pa.msu.edu/cteq/public m. gluck , e. reya and a. vogt , z. phys .
c * 53 * , 651 ( 1992 ) . c. g. duan , n. liu and z. y. yan , eur .
j. c * 50 * , 585 ( 2007 ) . c. g. duan , n. liu and g. l. li , phys .
c * 79 * , 048201 ( 2009 ) . | we study the nuclear medium effects in drell - yan process using quark parton distribution functions calculated in a microscopic nuclear model which takes into account the effects of fermi motion , nuclear binding and nucleon correlations through a relativistic nucleon spectral function .
the contributions of @xmath0 and @xmath1 mesons as well as shadowing effects are also included .
the beam energy loss is calculated using a phenomenological approach .
the present theoretical results are compared with the experimental results of e772 and e886 experiments .
these results are applicable to the forthcoming experimental analysis of e906 sea quest experiment at fermi lab . |
we consider finite undirected graphs that do not contain loops . for a graph @xmath2 and a positive integer @xmath3 define @xmath4 are pairwise edge - disjoint matchings of @xmath5 , and let @xmath6 . a subgraph @xmath7 of @xmath2
is called maximum @xmath3-edge - colorable , if it is @xmath3-edge - colorable and contains exactly @xmath8 edges .
terms and concepts that we do not define can be found in @xcite .
the first theorem allows us to `` embed '' maximum @xmath1edge - colorable subgraphs of a cubic graphs into minimum @xmath9edge - colorings .
having done this , we can apply the results of @xcite for the investigation of our problems .
first of all let us show that the graph @xmath11 contains no path of length three . on the opposite assumption ,
consider a path @xmath12 of the graph @xmath13 , and let us assume that @xmath7 is properly colored with colors @xmath14 .
define the sets @xmath15 as the colors of the edges incident to the vertices @xmath16 , respectively . if @xmath18 , then there is @xmath19 with @xmath20 . now
, if we color the edge @xmath21 with color @xmath22 , then we would get a proper @xmath1edge - coloring of the subgraph @xmath23 , contradicting the maximality of @xmath7 .
thus : @xmath24 now , let us show that @xmath25 assume the contrary , that is @xmath26 .
there is @xmath19 with @xmath27 . now ,
if we color the edge @xmath21 with color @xmath22 , then we would get a proper @xmath1edge - coloring of the subgraph @xmath23 , contradicting the maximality of @xmath7 .
thus , ( [ unioncolors ] ) must be true .
since @xmath28 , then there is @xmath19 with @xmath29 . now ,
if we color the edge @xmath30 with color @xmath22 , then we would get a proper @xmath1edge - coloring of the subgraph @xmath31 , contradicting the maximality of @xmath7 .
case 1 : @xmath33 , that is , @xmath21 is a multiple edge .
note that @xmath34 , thus there is @xmath35 with @xmath36 .
now , if we color one of edges connecting @xmath37 and @xmath38 with color @xmath22 , then we would get a proper @xmath1edge - coloring of the subgraph @xmath39 , contradicting the maximality of @xmath7 .
case 2 : @xmath40 .
note that @xmath41 .
like in the proof of ( [ unioncolors ] ) , one can easily show that @xmath42 thus @xmath43 .
suppose that @xmath44 and @xmath45 .
consider the maximal @xmath46 alternating paths @xmath47 , starting from vertices @xmath16 , respectively .
note that there is @xmath48 such that @xmath49 .
now , shift the colors on the path @xmath50 to obtain a new coloring of the maximum @xmath1edge - colorable subgraph @xmath7 , where the color @xmath22 is absent in both of vertices @xmath51 and @xmath38 .
now , if we color the edge @xmath52 with color @xmath22 , then we would get a proper @xmath1edge - coloring of the subgraph @xmath53 , contradicting the maximality of @xmath7 .
the proof of the theorem [ complementmatching ] is completed . for a @xmath56factor @xmath57 of @xmath2 , choose a maximum @xmath1edge - colorable subgraph @xmath7 of @xmath2 with @xmath58 let us show that @xmath59 . on the opposite assumption ,
consider an edge @xmath60 and let us assume that @xmath7 is properly colored with colors @xmath14 .
due to theorem [ complementmatching ] , the edges adjacent to @xmath61 are belong to @xmath7 .
let @xmath62 and @xmath63 denote the colors of edges that are incident to @xmath64 and @xmath65 , respectively .
note that the maximality of @xmath7 implies that @xmath66 choose @xmath67 .
consider the subgraph @xmath68 , where @xmath69 is the edge that is incident to @xmath64 and is colored by @xmath22 .
note that @xmath70 is a maximum @xmath1edge - colorable subgraph of @xmath2 with @xmath71contradicting the choice of @xmath7 .
the proof of the theorem [ 1factorextension ] is completed .
next , we prove a result which claims that the uncolored edges with respect to a maximum @xmath1edge - colorable subgraph of @xmath2 always can be `` left '' in a given @xmath56factor , or , equivalently , any @xmath0factor of a cubic graph @xmath2 can also be extended to a maximum @xmath1edge - colorable subgraph of @xmath2 .
[ 2factorextension]let @xmath57 be any @xmath56factor of a cubic graph @xmath2 , and let @xmath72 be the complementary @xmath0factor of @xmath57 .
then there is a maximum @xmath1edge - colorable subgraph @xmath7 of @xmath2 , such that : for a given @xmath56factor @xmath57 of a cubic graph @xmath2 , consider a maximum @xmath1edge - colorable subgraph @xmath7 of @xmath2 with @xmath75 let us show that @xmath76 .
we only need to verify that @xmath77 .
assume that there is @xmath78 such that @xmath61 belongs to neither of @xmath57 and @xmath7 , and let us assume that @xmath7 is properly colored with colors @xmath14 .
consider the edges adjacent to @xmath61 .
due to theorem [ complementmatching ] , these edges belong to @xmath7 .
if @xmath62 and @xmath63 denote the colors of edges that are incident to @xmath64 and @xmath65 , respectively , then the maximality of @xmath7 implies that @xmath66 suppose that @xmath79 due to @xcite , there is an odd cycle @xmath80 containing the edge @xmath61 , whose all other edges are colored @xmath81 and @xmath82 , alternatively , while the edges that are incident to a vertex of @xmath80 and do not lie on @xmath80 , are colored by @xmath22 . we need to consider two cases : let @xmath84 .
consider a proper partial @xmath1edge - coloring of the graph @xmath2 obtained from the coloring of @xmath7 as follows : @xmath85 is left uncolored , the edges of the even path @xmath86 are colored @xmath81 and @xmath82 , alternatively , the colors of the rest of edges are left unchanged .
note that the new partial @xmath1edge - coloring corresponds to a maximum @xmath1edge - colorable subgraph @xmath70 of @xmath2 with @xmath87contradicting the choice of @xmath7 . 1 .
the edges that are incident to a vertex of @xmath80 , and do not lie on @xmath80 , are colored by @xmath22 and belong to @xmath57 , which and @xmath88 imply that : 2 .
all maximum @xmath1edge - colorable subgraphs @xmath70 of @xmath2 , which can be obtained from the coloring of @xmath7 , by leaving any edge @xmath89 uncolored , by coloring the edges of the even path @xmath90 @xmath81 and @xmath82 , alternatively , and leaving the colors of the rest of edges unchanged , satisfy the condition @xmath91 now , we consider a proper partial @xmath1edge - coloring @xmath92 of the graph @xmath2 obtained from the coloring of @xmath7 by deleting the colors of the all edges lying on @xmath80 . since @xmath80 is an odd cycle , we imply that there is an @xmath93 alternating path @xmath94 in the @xmath1edge - coloring @xmath92 that starts from a vertex @xmath95 and does not end on @xmath80 .
choose an edge @xmath96 , and let @xmath97 be the other edge of @xmath80 that is incident to @xmath98 .
* shift the colors on the path @xmath94 , and clear the color of the edge of @xmath99 that is incident to @xmath98 ; * color @xmath100 by @xmath22 , and color the edges of the even path @xmath101 by @xmath102 alternatively , starting from the edge @xmath97 .
it is not hard to see that the new partial @xmath1edge - coloring of @xmath2 corresponds to a maximum @xmath1edge - colorable subgraph @xmath70 of @xmath2 , which due to ( ii ) satisfies @xmath87contradicting the choice of @xmath7 .
the proof of the theorem [ 2factorextension ] is completed .
there are graphs attaining bounds of the theorem [ maintheoremcubics ] .
the graph from figure [ examples attaining the bounds]a attains the first bound and the graph from figure [ examples attaining the bounds]b the second bound .
note that if there were a cubic graph @xmath2 attaining the two bounds at the same time , then we would have : @xmath105 thus , to show the absence of a cubic graph attaining all the bounds of theorem [ maintheoremcubics ] at the same time , it suffices to show the following suppose that @xmath2 contains a triangle @xmath109 that has no multiple edge .
consider a cubic graph @xmath110 obtained from @xmath2 by contracting the triangle @xmath109 to a new vertex . since @xmath111 , we imply @xmath112 .
on the other hand , it can be easily checked that @xmath113 ( in fact , with a bit more work it can be shown that one always has equalities here ) , therefore @xmath114contradicting the choice of @xmath2 .
the proof of the claim [ trianglemultiple ] is completed . before me
move to the proof of the next claim , let us also note that @xmath2 contains no edge that connects two triangles , since if there were such an edge , then taking into account that @xmath2 is connected , we would have that @xmath2 is isomorphic to the second graph from figure [ examples attaining the bounds ] , which clearly satisfies the statement of the theorem .
[ triangleneighbourmultiple ] let @xmath115 be vertices of a triangle of @xmath2 , and let @xmath116 be its multiple edge ( claim [ trianglemultiple ] ) .
assume that @xmath64 is the neighbour of @xmath117 lying outside the triangle .
then @xmath64 is not incident to a multiple edge .
[ triangleadjacent ] let @xmath121 and @xmath122 be vertices of two triangles of @xmath2 , and let @xmath123 be their multiple edges ( claim [ trianglemultiple ] ) . then , there is no vertex @xmath64 that is adjacent to both @xmath124 and @xmath125 .
[ triangleadjacentneighbours ] let @xmath121 and @xmath122 be vertices of two triangles of @xmath2 , and let @xmath123 be their multiple edges ( claim [ trianglemultiple ] ) .
let @xmath38 and @xmath129 be the neighbours of @xmath124 and @xmath125 , respectively , that lie outside the triangles
. then @xmath38 and @xmath129 are not connected by an edge .
first of all note that , due to claim [ triangleadjacent ] , @xmath130 .
suppose that @xmath38 and @xmath129 are adjacent , and let us consider the cubic graph @xmath110(figure [ trianglesadjacentneighbours ] ) .
[ triangleneighbourscommonneighbour ] let @xmath121 and @xmath122 be vertices of two triangles of @xmath2 , and let @xmath123 be their multiple edges ( claim [ trianglemultiple ] ) .
let @xmath38 and @xmath129 be the neighbours of @xmath124 and @xmath125 , respectively , that lie outside the triangles .
then , there is no vertex @xmath98 , that is adjacent to both @xmath38 and @xmath129 .
note that @xmath131 , hence @xmath112 .
it can be easily seen that @xmath132 therefore @xmath134contradicting the choice of @xmath2 .
the proof of the claim [ triangleneighbourscommonneighbour ] is completed .
the proved claims imply that the triangles of @xmath2 are placed `` sufficiently far '' , and this allows us to complete the proof by using the odd cycles corresponding to edges that do not lie in some maximum @xmath135-edge - colorable subgraph @xcite .
let @xmath7 be any maximum @xmath1edge - colorable subgraph of @xmath2 , and let @xmath136 be its color classes , which are pairwise disjoint matchings with @xmath137 .
consider the @xmath138tuple @xmath139 , and for any @xmath140 define @xmath141 as the number of edges that @xmath65 is incident from this @xmath138tuple ( edges of @xmath142 and @xmath143 contribute to @xmath141 by two ) .
note that @xmath144=\sum_{v\in v}{\delta(v)}.\end{gathered}\ ] ] thus , to get the desired contradiction , it suffices to show , that @xmath145 * if a vertex @xmath140 lies outside these cycles , then it contributes to the sum ( [ maininequality ] ) by five ; * the total contribution of vertices of triangles from figure [ magictriangles ] ( a ) and ( b ) to the sum ( [ maininequality ] ) is twelve ; * the total contribution of vertices of triangles from figure [ magictriangles ] ( c ) to the sum ( [ maininequality ] ) is eleven ; * the total contribution of vertices of cycles from figure [ magicoddcycles ] ( a ) and ( b ) to the sum ( [ maininequality ] ) is @xmath146 , where the cycle is assumed to be of length @xmath147 ; * the total contribution of vertices of cycles from figure [ magicoddcycles ] ( c ) to the sum ( [ maininequality ] ) is @xmath148 , where the cycle is assumed to be of length @xmath147 .
thus , if we have a triangle with vertices @xmath115 , whose multiple edge is @xmath116 ( claim [ trianglemultiple ] ) , and @xmath64 is the neighbour of @xmath117 , lying outside this triangle , then if @xmath64 does not lie on either of these odd cycles , then @xmath149 on the other hand , if a cycle of length @xmath147 from figure [ magicoddcycles ] contains @xmath150 vertices , which are adjacent to a vertex of a triangle from figure [ magictriangles ] ( c ) , then the proved claims imply that @xmath151therefore the total contribution of vertices of these @xmath150 triangles and the odd cycle of length @xmath152 to the sum from ( [ maininequality ] ) is at least @xmath153as @xmath154 . | we show that any @xmath0factor of a cubic graph can be extended to a maximum @xmath1edge - colorable subgraph .
we also show that the sum of sizes of maximum @xmath0 and @xmath1edge - colorable subgraphs of a cubic graph is at least twice of its number of vertices . |
in spite of a huge experimental effort , the higgs particle , the last missing ingredient of the standard model ( sm ) of electroweak interactions has not been discovered yet . for a higgs - boson mass @xmath4
the most promising production channel has been the radiation off a @xmath5-boson at lep2 : @xmath6 ; using this reaction the lep data provided a limit @xcite on the sm higgs - boson mass : @xmath7 .
the higgs particle also contributes radiatively to several well measured quantities , this can be used to derive an upper bound @xcite on @xmath8 : @xmath9 at 95 % c.l .. however , one should be aware that both limits are highly model - dependent .
there exist other theoretical restrictions of @xmath10 based on the so - called triviality and vacuum stability arguments . as it is well know @xcite the renormalized @xmath11 theory can not contain an interaction term ( @xmath12 ) for any non - zero scalar mass : the theory must be trivial . within a perturbative approach the statement corresponds to the fact that for any non - zero scalar mass ( since the mass is @xmath13 this condition corresponds to a non - vanishing initial value for the renormalization group ( rg ) evolution of @xmath14 ) there exists a finite energy scale at which @xmath14 diverges ( the landau pole ) .
consequently , only for zero scalar mass the theory can be consistent for all energy scales .
an analogous effect occurs in the scalar sector of the sm ( modified to some extend by the presence of gauge and yukawa interactions ) .
this , however , does not necessarily implies zero higgs - boson mass since there is no reason to believe the sm is valid at arbitrarily high energy scale .
for example , it is often assumed that the sm represents the low energy limit of some underlying more fundamental theory whose heavy excitations decouple at low energy , but become manifest at a scale @xmath2 . within that scenario
the sm is an effective theory valid possibly only at the energy scale of the order of the fermi scale : @xmath15 .
if the sm is to be accurate for energies below @xmath16 the landau pole should occur at scale @xmath16 or above , and this condition gives a ( @xmath2-dependent ) upper bound on @xmath17 @xcite . on the other hand , for sufficiently small @xmath17 radiative corrections can destabilize the ground state .
this occurs if the scalar self coupling constant @xmath18 becomes negative at some scale that can be identified with the scale of new physics @xmath2 . alternatively requiring the sm vacuum to be stable for scales below @xmath19 implies a lower bound on @xmath17 @xcite .
the consequences of the above arguments ( triviality and vacuum stability ) are usually discussed assuming sm interactions . however
, if the scale of new physics is sufficiently low ( of the order of a few tev ) one could expect that the non - standard interactions would modify the electroweak theory at the lower scale and influence the scalar potential in such a way that the above bounds on the higgs - boson mass are changed .
the problem deserves a special attention in the context of possible higgs - boson discovery @xcite at lep2 at the mass @xmath20 since in this case the sm constraint from vacuum stability requires @xmath21 @xcite ( the precise number depends on the top quark mass ) with the attractive possibility that @xmath19 is actually much lower .
it then becomes interesting to determine the manner in which heavy physics with scales in the @xmath22 region modify the stability and triviality bounds on the higgs - boson mass . in this lecture
we address this question in a model - independent way .
we parameterize the heavy physics effects using an effective lagrangian satisfying the sm gauge symmetries . since lhc , the future proton - proton collider ,
is expected to be sensitive to scales @xmath16 of the order of a few tev , the results will be presented for scales between @xmath23 and @xmath24 .
the paper is organized as follows . in section [ non_stand ] , we define the lagrangian relevant for our discussion
. section [ triv_bound ] is devoted to the derivation of the triviality bound including effects of non - standard interactions . in section [ vac_bound ] ,
we calculate the effective potential with one insertion of an effective operator and discuss its consequences for the vacuum stability bounds .
concluding remarks are given in section [ summary ] .
our study of the stability and triviality constraints on the higgs - boson mass will be based on the sm lagrangian modified by the addition of a series of effective operators whose coefficients parameterize the low - energy effects of the heavy physics @xcite . assuming that these non - standard effects decouple implies @xcite that the operators appear multiplied by appropriate inverse powers of @xmath16 .
the leading effects are then generated by operators of mass - dimension 6 ( dimension 5 operators necessarily violate lepton number @xcite and are associated with new physics at very large scales ; so we can safely ignore their effects ) . given our emphasis on higgs - boson physics the effects of all fermions excepting the top - quark
can be ignored .
we then have _ tree & = & -14 f_^if^i -14 b_b^+ | d |^2 - ( -v^2 + ||^2 ) ^2 + + & & i |q d q
+ i |t d t + f ( |q t + ) + _ i_i , [ lagrangian ] where @xmath25 ( @xmath26 ) , @xmath27 and @xmath28 are the scalar doublet , third generation left - handed quark doublet and the right - handed top singlet , respectively .
@xmath29 , @xmath30 and @xmath31 denote a covariant derivative and @xmath32 , @xmath33 field strength whose couplings we denote by @xmath34 and @xmath35 .
the factors @xmath36 are unknown coefficients that parameterize the low - energy effects of the non - standard interactions and we have neglected contributions @xmath37 . in addition , for weakly coupled theories , the @xmath38 that can be generated only through loop effects are sub - dominant as they are suppressed by numerical factors @xmath39 @xcite ; hence we will consider only those operators that can be generated at tree - level by the heavy physics .
even with all the above restrictions there remain 16 operators that involve exclusively the fields in ( [ lagrangian ] ) . of these only 5 contribute directly to the effective potential , the remaining 11 affect the results only through their rg mixing which , being suppressed by a factor @xmath40 are expected to play a sub - dominant role . in the calculations below
we will include only one of these operators ; our results do justify the claim that the corresponding effects are small .
the following set of operators will be considered : = 3 | |^6 & _ = ( | @xmath41 , @xmath42 , @xmath43 , @xmath44 are the 5 operators contributing to the effective potential , while @xmath45 is included to estimate the effects of rg mixing . of the first five operators only @xmath46 contributes at the tree level to the scalar potential : v= - ^2 ||^2 + ||^4 - _ 3 ^2 | |^6 [ tree_pot ] where we have used the notation : @xmath47 .
in order to test the high energy behavior of the scalar potential one has to derive the rg running equations for @xmath14 , @xmath48 and @xmath49 .
the @xmath50 functions for these parameters are influenced by all the operators in ( [ operators ] ) and by the gauge and yukawa interactions , so the full rg evolution also require the @xmath51 function for the corresponding couplings . both for the @xmath50 functions and
then for the effective potential we will adopt dimensional regularization and @xmath52 renormalization scheme .
we will restrict ourself to the one - loop approximation keeping sm contributions and terms linear in the effective operators , defined by eq.([operators ] ) .
the evolution equations for the running coupling constants are the following : # 1 d # 1 d t @xmath53 \cr & & + { 3\over 2 } \left[-\lambda \left(3 g^2 + g'{}^2 \right ) + { g'{}^4 + 2 g^2 g'{}^2 + 3 g^4\over8 } \right ] \cr 8 \pi^2 \dert \eta & = & \eta\left[6 \lambda + 3 f^2 -2 \eta \left ( \alpha_{\partial\phi } + 2\alpha_\phi\up1 + \alpha_\phi\up3 \right ) -{3\over 4 } \left ( 3g^2 + g'{}^2 \right)\right ] \cr 8 \pi^2 \dert f & = & { 9\over4 } f^3 + { 1\over2 } \eta \left [ 6 \alpha_{t\phi } - f \left ( \alpha_{\partial\phi } + 2 \alpha_\phi\up1 + \alpha_\phi\up3 + 3 \alpha_{qt}\up1 \right ) \right ] \cr & & -{f\over 8 } \left ( 9 g^2 + { 17\over3}g'{}^2 \right ) \cr 8 \pi^2 \dert{\alpha_\phi}&= & 9 \alpha_\phi \left(6 \lambda + f^2 \right ) + 12 \lambda^2 ( 9\alpha_{\partial\phi}+6 \alpha_\phi\up 1 + 5\alpha_\phi\up 3 ) + 36 \alpha_{t\phi } f^3
\cr & & - { 9 \over 4 } ( 3 g^2 + g'{}^2 ) \alpha_\phi\cr & & - { 9 \over 8 } \left [ \alpha_\phi\up1 ( 3g^4 + 2 g^2 g'{}^2 + g'{}^4 ) + \alpha_\phi\up3 ( g^2 + g'{}^2 ) ^2 \right ] \cr 8 \pi^2 \dert{\alpha_{\partial\phi}}&= & 2 \lambda \left ( 7 \alpha_{\partial\phi } - \alpha_\phi\up1 + \alpha_\phi\up3 + 3 { \alpha_{\partial\phi } f^2 \over \lambda } - 3 { \alpha_{t\phi } f \over \lambda } \right ) \cr 8 \pi^2 \dert{\alpha_\phi\up1}&= & 2 \lambda \left(\alpha_{\partial\phi}+5\alpha_\phi\up1 + \alpha_\phi\up3 + 3 { \alpha_\phi\up1 f^2 \over \lambda } - 3 { \alpha_{t\phi } f \over \lambda } \right ) \cr 8 \pi^2 \dert{\alpha_\phi\up3}&= & 6 ( \lambda + f^2 ) \alpha_\phi\up3 \cr 8 \pi^2 \dert{\alpha_{t\phi } } & = & -3 f ( f^2+\lambda ) \alpha_{qt}\up1 + { 3\over4 } ( 5f^2 - 16\lambda ) \alpha_{t\phi } \cr & & - \inv2 f^3 \left ( 2 \alpha_{\partial\phi } + \alpha_\phi\up1 + \alpha_\phi\up3 \right ) \cr 8 \pi^2 \dert{\alpha_{qt}\up1 } & = & { 3\over2 } \alpha_{qt}\up1 f^2 \label{beta_fun}\end{aligned}\ ] ] where @xmath54 and @xmath55 denotes the renormalization scale . from this set of equations
it is straightforward to obtain the triviality constraints on @xmath10 as a function of @xmath56 requiring that the position of the landau pole is beyond the scale @xmath2 .
there is a comment here in order , namely , in actual calculations the position of the landau pole can not be accurately determined to any finite order in perturbation theory . therefore the triviality bound on @xmath10 will be obtained by requiring @xmath14 and @xmath57 to become smaller than specified values ( as opposed from requiring an actual divergence ) up to the scale @xmath2 : ( t)_max |_i(t)|1.5 0 t<(/ ) .
[ triv_con ] where we considered @xmath58 and @xmath59 .
we have verified that our results are quite insensitive to the values chosen as upper limits for the @xmath38 . in order to solve the equations ( [ beta_fun ] )
we have to specify appropriate boundary conditions . for the sm parameters
these are determined by requiring that the correct physical parameters ( such as the higgs - boson and top - quark masses ) are obtained at the electroweak scale .
these initial conditions should also insure that the correct sm ground state is realized , in which the scalar field has the expectation value @xmath60 .
although we will discuss the effective potential in more detail later , it will be useful to provide here the general 1-loop relation between the sm tree - level vacuum @xmath61 and the physical electroweak vacuum in the theory defined by the equation ( [ lagrangian ] ) @xmath62 : = v + v v - 1 4 ( 0 ) v ^2 .(v_eff - v_sm^(tree ) ) ( / ) |_=v / , [ vev ] where @xmath63 is the tree - level sm potential and @xmath64 is the 1-loop effective potential that includes effective operator contributions ; @xmath65 denotes the running coupling constant evaluated at the scale @xmath66 .
having the vacuum determined by the above equation , the following low - scale relations will be adopted to fix initial conditions at @xmath66 for the rg equations for @xmath14 , @xmath48 and @xmath67 .
^2 & = & 2 ^2 + ^(1),&= & ( f+_t^2 ^ 2 ) + ^(1 ) [ higgs_mass ] where @xmath68 denote the 1-loop radiative corrections to the corresponding masses . in the calculations
bellow we use the expression for @xmath69 of ref .
for the top - quark the deviations from the tree - level value are smaller than the experimental error and
so , for simplicity we will use the expression @xmath70 .
the initial conditions are non - linear functions of the higgs - boson mass , and so the solutions to ( [ beta_fun ] ) will depend on both @xmath56 and @xmath71 .
the boundary conditions for @xmath36 are naturally specified at the scale @xmath72 since below this scale it is appropriate to describe the effects of the heavy excitations in terms of the coefficients @xmath36 . according to ref .
@xcite it is natural to assume that @xmath73 .
the triviality bound is obtained by solving the equations ( [ beta_fun ] ) with the mixed ( defined in part at the electroweak scale @xmath74 and at the new - physics scale @xmath2 ) boundary conditions described above and requiring that at least one of the inequalities in eq.([triv_con ] ) is saturated .
this provides a relationship between @xmath17 and @xmath16 that we plot in fig.[bounds ] ( a ) for two values of @xmath75 .
= 4.5 in in order to understand qualitatively the corrections to the triviality bound we have obtained , is useful to switch off all @xmath36 but @xmath57 .
then , as it is seen from eq.([beta_fun ] ) a landau pole in the evolution of @xmath76 causes a singularity in evolution of @xmath57 at _ this same _ energy scale .
however , as we have just mentioned it is natural to assume @xmath77 , it is clear that strictly speaking it is impossible to satisfy that condition .
nevertheless , since we are using a perturbation expansion , we must stop the evolution at a scale that corresponds to a large but finite value @xmath78 , therefore we can satisfy @xmath77 .
however , since @xmath79 is positive remains positive in the whole integration region , it addition @xmath80 what guarantees that @xmath81 .
] therefore in the evolution from the scale @xmath2 down , @xmath82 decreases reaching typically @xmath83 at the scale @xmath66 . that explains screening of the effects generated by operator @xmath84 : even if @xmath77 it can not grow any larger . so concluding , the corrections to the sm triviality bound from the non - standard physics ( embedded in the coefficients @xmath36 ) are negligible .
in order to investigate the vacuum structure of the effective theory we will first calculate the effective potential : @xmath85 where @xmath86 are n - point one - particle - irreducible greens functions with zero external momenta and @xmath87 is the classical scalar field .
adopting the landau gauge functions and the tree - level potential @xmath88 are gauge - independent , a consistent rg improved tree - level effective potential is in fact gauge independent . for the one - loop sm
rg improved effective potential , the error caused by the gauge dependence has been estimated in ref.@xcite at @xmath89 . ]
we obtained : @xmath90 , \non \label{effpot}\end{aligned}\ ] ] where @xmath91 { |\pb|^2\over \lambda^2 } \cr g & = & \lambda ( - v^2 + 2 |\pb|^2 ) - \left[\lambda ( - v^2 + 2 |\pb|^2 ) \inv3 ( 3 \alpha_\phi\up1 + \alpha_\phi\up3 ) + \alpha_\phi |\pb|^2 \right ] { |\pb|^2\over \lambda^2 } \cr w & = & { g^2 \over2}|\pb|^2 \left ( 1 + { |\pb|^2 \alpha_\phi\up1 \over \lambda^2 } \right ) \cr z & = & { g^2 + g'{}^2 \over2 } |\pb|^2 \left ( 1 + { |\pb|^2 ( \alpha_\phi\up1 + \alpha_\phi\up3 ) \over \lambda^2 } \right ) \cr t & = & f^2
|\pb|^2 \left(1 + { 2 \alpha_{t \phi } | \pb|^2 \over f \lambda^2 } \right),\non\end{aligned}\ ] ] where @xmath34 and @xmath35 denote the @xmath32 and @xmath33 running gauge coupling constants , respectively .
the _ form _ of the effective potential is precisely the same as the one in the pure sm , the whole effect of the effective operators can be absorbed in a re - definition of the sm quantities @xmath92 , @xmath93 , etc . ) for the effective potential have been obtained adopting the diagrammatic approach ( with one insertion of an effective operator ) according to eq.([eff_pot_def ] ) and also using the functional definition of the effective potential proposed by jackiw @xcite . ]
it should be noticed here that the last term in eq.([effpot ] ) is a constant that is needed to construct a scale invariant effective potential , for details see ref .
the constant term chosen here is consistent with the diagrammatic definition of the effective potential eq.([eff_pot_def ] ) , that implies @xmath94 .
since we will consider values of @xmath87 substantially larger then the electroweak scale @xmath62 , we shall chose an appropriate renormalization scale @xmath95 in order to moderate the logarithms that appear in the effective potential . as in the previous section
we shall use the rg running equations to relate the coupling constants renormalized at the high scale @xmath87 to the low - scale parameters @xmath62 , @xmath96 and @xmath10 . finally , ( and unlike the pure @xmath11 )
the interaction of the scalars with the fermions and gauge bosons , generate a non - trivial scalar field anomalous dimension @xmath97 .
we therefore also include the corresponding scale dependence of @xmath98 : ( t)=\{-_t_z^t dt } ( t_z ) , [ phi_scaling ] where = 116 ^2 hereafter we will consider the rg improved effective potential @xmath99 .
we note that the rg improved effective potential given by eq.([effpot ] ) is scale invariant .
that is , to one loop and ignoring terms quadratic in the @xmath100 , @xmath101 obeys the renormalization group equation : @xmath102 where @xmath88 and @xmath103 denote , respectively , the tree , eq.([tree_pot ] ) , and 1-loop , eq.([effpot ] ) , contributions to @xmath64 , and @xmath104 are defined in ( [ beta_fun ] ) .
we note that terms quadratic ( and higher ) in the @xmath38 are associated with contributions of order @xmath105 to the effective lagrangian and are sub - dominant .
fig.[potential ] ( a ) illustrates the behavior of the effective potential renormalized at the scale @xmath106 . since the minimum at @xmath107 is very shallow , in order to make it visible we plot the following function of the effective potential : @xmath108 $ ] . to show the relevance of rg running of effective - potential parameters we also plot in fig.[potential ] the evolution of @xmath14 ( b ) and @xmath57 ( c ) .
the curves contained in the figure correspond to two sets of initial conditions ( 1 ) and ( 2 ) that lead to the higgs - boson mass and the new - physics scale marked in fig.[bounds ] by @xmath109 s . as it is seen from the figure effects of the running are substantial , e.g. for the set ( 2 ) @xmath14 changes by almost 100% while @xmath57 by more than 200% . at the electroweak scale , @xmath57 s start with positive values , however then , through the evolution they switch signs and eventually reach @xmath110 .
that should illustrate the fact that the rg running of the coefficients @xmath36 is crucial for the stability of the system has been previously discussed in ref .
@xcite . however , there the authors did not consider one - loop contributions to the effective potential that are generated by insertions of effective operators .
rg running of @xmath57 has also been neglected . ] .
the initial conditions for the running couplings guarantee that the electroweak vacuum is at @xmath111 . however if @xmath112 at some large value of the field @xmath113 is smaller than @xmath114 this vacuum becomes unstable ( as there would be a possibility of tunneling towards the region of lower energy ) .
this will occur when the higgs - boson mass is sufficiently small ( corresponding to a small value of @xmath115 ) , and will provide a lower bound on @xmath116 . in this case @xmath117 defines a scale at which the theory breaks down , so that @xmath118 . in actual calculations we took @xmath119 since
( [ lagrangian ] ) is valid for scales below @xmath120 , hence the stability bound on @xmath116 is determined by the condition ( = 0.75)|_=0.75 = (= /)|_=/ [ stab_con ] where , as mentioned previously , we have chosen the renormalization scale @xmath121 to tame the effects of the logarithmic contributions to @xmath122 .
the resulting bound on @xmath17 as a function of @xmath19 for various choices of @xmath123 is plotted in fig .
[ bounds ] ( b ) . in obtaining the stability bounds of fig .
[ bounds ] ( b ) we assumed all couplings @xmath38 had the same magnitude at the high scale @xmath124 , and @xmath125 ( the results are insensitive to the sign of the other @xmath38 except @xmath126 ) . for other values of @xmath36
we found that when @xmath127 there is a curve in the @xmath128 plane below which either @xmath129 is not a minimum or , if it is , then there is another deeper minimum at a scale @xmath130 ; we can roughly say that this unphysical scenario can be avoided if @xmath131 are included : a contribution @xmath132 can balance the destabilizing effect of @xmath133 only when @xmath134 which again leads to @xmath135 . ] .
there is an important remark here in order .
if the higgs boson mass , as suggested by lep data , is indeed @xcite @xmath136 , then the sm vacuum stability bound implies @xmath21 . as it is seen in fig .
[ bounds ] ( b ) presence of effective operators could dramatically change the sm picture .
even for the modest values of the coefficient @xmath137 the upper bound on @xmath19 is significantly reduced to @xmath138 , respectively !
other limits on the scale @xmath2 could be obtained form so called precision observables .
the most elegant approach is to calculate the oblique parameters @xmath139 , @xmath140 and @xmath141 @xcite within the effective theory contributes to the oblique parameters ( @xmath140 ) and therefore is constrained by the precision data , however as it has been shown here the operator that is most relevant for the triviality and vacuum stability bound is @xmath84 and contributions from @xmath142 are much less important . ] and then fit their experimental values @xcite .
the limits obtained that way depend also on the higgs - boson mass @xmath10 therefore it would be interesting to superimpose precision - measurement limits , the direct lep limit and those obtained here , consistently taking into account higher dimensional operators , that is however beyond the scope of this paper .
we have considered restrictions on the higgs - boson mass that emerge form requirement of perturbative behavior of the quartic coupling constant ( the triviality bound ) and from the condition of stable electroweak vacuum taking into account possible non - standard interactions described by effective operators of dimension @xmath143 .
it was shown that for the scale of new physics in the region @xmath144 the standard model triviality upper bound remains unmodified whereas the lower bound from requirement of vacuum stability is naturally increased by @xmath1 depending on the scale @xmath2 and strength of coefficients of effective operators .
therefore the allowed region of the higgs - boson mass is reduced substantially . if the higgs - boson mass is close to its lower lep limit then the upper bound on the scale of new physics that follows from the vacuum stability requirement
could be decreased dramatically even for modest values of coefficients of effective operators implying new physics already at the scale of @xmath145 .
this work is supported in part by the state committee for scientific research ( poland ) under grant 5 p03b 121 20 and funds provided by the u.s .
department of energy under grant no .
de - fg03 - 94er40837 .
one of the authors ( bg ) is indebted to cern , slac and u.c .
riverside for the warm hospitality extended to him while this work was being performed .
99 t. junk , the lep higgs working group , at lep fest october 10th 2000 , http://lephiggs.web.cern.ch/lephiggs/talks/index.html .
e. tournefier , the lep electroweak working group , talk presented at the 36th rencontres de moriond on electroweak interactions and unified theories , 2001 , les arcs , france , hep - ex/0105091 .
k. wilson , .
l. maiani , g. parisi , and r. petronzio , ; m. lindner , .
n. cabibbo , ; for a review see m. sher , and references therein .
igo - kemenes , the lep higgs working group , at lepc 3 november 2000 , http://lephiggs.web.cern.ch/lephiggs/talks/index.html .
j. a. casas _ et al .
b * 436 * , 3 ( 1995 ) [ erratum - ibid .
b * 439 * , 466 ( 1995 ) ] [ hep - ph/9407389 ] ; m. quiros , iem - ft-153 - 97 , hep - ph/9703412 . s. weinberg , hep - th/9702027 .
h. georgi , ann .
nucl . part .
sci . * 43 * , 209 ( 1993 ) .
t. appelquist and j. carazzone , phys .
d * 11 * , 2856 ( 1975 ) .
j. collins , f. wilczek and a. zee , phys .
d * 18 * , 242 ( 1978 ) .
w. buchmeller and d. wyler , .
g. t hooft , _ lecture given at cargese summer inst . , cargese , france , aug 26 - sep 8 , 1979 _ , in _ c79 - 08 - 26.4 _ print-80 - 0083 ( utrecht ) . c. arzt , m.b . einhorn and j. wudka , , hep - ph/9405214 .
m. chanowitz , .
l. dolan and r. jackiw , .
w. loinaz and r.s .
willey , , hep - ph/9702321 .
r. jackiw , . c. ford , , hep - lat/9210033 .
a. datta , b.l .
young and x. zhang , , hep - ph/9604312 . g. isidori _ et al .
_ , , hep - ph/0104016 . m.e . peskin and t. takeuchi , , , see also g. altarelli , r. barbieri , .
g. sanchez - colon and j. wudka , , hep - ph/9805366 ; r. barbieri and a. strumia , , hep - ph/9905281 ; l. hall and c. kolda , , hep - ph/9904236 ; j.a .
bagger , a.f .
falk and m. swartz , , hep - ph/9908327 .
chivukula and n. evans , , hep - ph/9907414 ; c. kolda and h. murayama , , hep - ph/0003170 . | the triviality and vacuum stability bounds on the higgs - boson mass were revisited in presence of weakly - coupled new interactions parameterized in a model - independent way by effective operators of dimension 6 .
it was shown that for the scale of new physics in the region @xmath0 the standard model triviality upper bound remains unmodified whereas it is natural to expect that the lower bound derived from the requirement of vacuum stability is increased by @xmath1 depending on the scale @xmath2 and strength of coefficients of effective operators .
it turns out that if the higgs - boson mass is close to its lower lep limit then the scale of new physics that follows from the vacuum stability requirement would be decreased dramatically even for modest values of coefficients of effective operators implying new physics already at the scale of a few @xmath3 .
= by -1 |
the asymptotic freedom of qcd allows to calculate short distance quantities perturbatively form the first principles .
together with the use of factorization theorems , the cross sections for ( semi-)inclusive reactions are predicted at a large scale as a convolution of perturbative parton cross sections and long distance parton distributions . in the case of exclusive reactions , such as two - photon reactions in the light - cone
dominated region or diffractive processes , the amplitudes factorizes in hard - scattering amplitudes and hadron distribution amplitudes and/or skewed parton distributions ( spd ) .
such distributions were already proposed in the past @xcite and they become more popular nowadays @xcite since it has been realized their vital rle for the proper description of the mentioned reactions . moreover , as it was pointed out by j. xi , their measurement can clarify the issue of how the hadron total angular momentum is distributed among its constiutents .
one can expect that the study of the spds will be one of the central issues of hadronic physics in nearest future . a promising way to access
the spds is to probe the nucleon with a highly virtual photon via the compton scattering process , which is called deeply virtual compton scattering ( dvcs ) .
indeed , it is auspicious that a first evidence of experimental feasibility to detect such a signal has been revealed by the zeus collaboration @xcite . on the theoretical side the factorization of the collinear singularities has been proved at leading twist - two level to all orders in coupling @xcite .
the local version of the ope does not converge in the considered kinematical region . fortunately
, this problem can be overcome by a non - local light - cone expansion in terms of light - ray operators @xcite .
moreover , from the renormalization group equation of these operators one can derive in a straightforward manner the evolution equations for spds . here
we shortly review our approach to complete next - to - leading ( nlo ) calculation , which is accomplished by a heavy use of the conformal symmetry to map the known results from forward deep inelastic scattering ( dis ) to the non - forward kinematics of dvcs and simple one - loop addenda .
the problems within our modus operandi have been resolved by employing conformal ward identities and constraints arising from the algebra of the collinear conformal group @xcite .
the virtual compton scattering ( vcs ) process on a nucleon can be measured in the reaction @xmath0 however , as shown in fig .
[ dvcs&bh ] , there is contamination from the so - called bethe - heitler ( bh ) process in which the final real photon is radiated from the lepton beam .
thus , the scattering amplitude square @xmath1 in the cross section , @xmath2 contains beside the vcs and bh parts also the interference term : @xmath3 the bh - amplitude is purely real and is given as a contraction of the leptonic tensor @xmath4 u(k , \lambda)\ ] ] with the hadronic current @xmath5 : @xmath6 since dirac and pauli form factors @xmath7 and @xmath8 are ( partially ) known from experimental measurements and can be parametrized to a certain accuracy by double dipole formulae , one can disentangle the content of the hadronic tensor @xmath9 : @xmath10 this tensor , defined by the time ordered product of two electromagnetic currents @xmath11 is a result of parity invariance . ] independent kinematical structures @xcite . measuring the cross section ( [ cross - section - def ] ) in quite different experiments : with negatively or positively charged lepton beam , with ( un)polarized beam or ( un)polarized target , the rich structure of the hadronic tensor can be explored in different kinematical regions . in the following we are interested in dvcs , where @xmath12 is large and
@xmath13 is small .
unfortunately , since the dvcs amplitude ( [ vcs - amplitude ] ) decreases with increasing @xmath14 while the bh amplitude ( [ bh - amplitude ] ) grows up with decreasing @xmath13 , in this kinematics it will be a delicate task to extract the vcs amplitude for large virtualities @xmath15 and especially low @xmath13 from experimental data .
on the other hand it is promising to look for asymmetries which allow one to measure the interference term directly and get in this way access to the real and imaginary part of the dvcs amplitude @xcite : 1 .
@xmath16 2 .
( either the lepton - beam or the nucleon is polarized ) @xmath17 3 .
@xmath18 where @xmath19 with @xmath20 is the azimuthal angle of the final nucleon ( lepton ) .
to describe different `` two - photon '' processes in the light - cone dominated region in a unique way , we generalize the bjorken limit and introduce the scaling variables : @xmath21 where @xmath22 is related to the bjorken variable @xmath23 and @xmath24 is the so - called skewedness parameter .
formally , this parameter interpolates between different processes as given in table [ tab - kinlcd ] .
.[tab - kinlcd ] the value of the scaling variable @xmath24 for different two - photon processes .
[ cols="<,^ , < " , ] since in the case of dvcs the longitudinal part of the hadronic tensor vanishes , there are only two kinematical structures left to leading twist - two accuracy : @xmath25 here the transverse part of the metric tensor , denoted by @xmath26 , and the @xmath27-tensor are contracted with the projection operators @xmath28 : @xmath29 , @xmath30 .
the appearing amplitudes can be calculated by means of the ope .
in contrast to the case of dis , a larger number of ( local ) operators and their wilson coefficients contribute in the non - forward case .
fortunately , we can appreciate the conformal ope ( cope ) @xcite , which tells us that all wilson coefficients are fixed up to an overall normalization , which can be taken from known results in dis .
the conformal operators appearing in the cope are : @xmath31 where @xmath32 and @xmath33 .
the @xmath34 and @xmath35 components are obtained by contraction with the two light - like vectors @xmath36 and @xmath37 , such that @xmath38 and @xmath39 .
they carry the spin quantum number @xmath40 and the conformal spin @xmath41 . in the conformal subtraction ( cs )
scheme their mixing is only induced by the trace anomaly and proportional to @xmath42 @xcite : @xmath43 the anomalous dimension matrix @xmath44 is , up to a normalization factor in the mixed channels , the same as in dis .
unfortunately , for the case of dvcs kinematics the cope in its local version does not converge .
thus , it is necessary to resum the local operators into so - called light - ray ones .
this provide us a definition for the spds : @xmath45 the momentum fractions of the incoming [ outgoing ] parton is @xmath46 [ @xmath47 . a form factor decomposition would give us the functions @xmath48 , @xmath49 and @xmath50 , @xmath51 for the parity even and odd sectors , respectively , as introduced by ji @xcite .
next we have to reconstruct the coefficient functions appearing in @xmath52,\end{aligned}\ ] ] where @xmath53 , @xmath54 and the sum runs over quark species @xmath55 with electrical charge @xmath56 .
the convolution is defined by @xmath57 .
this task can be achieved by some manipulation of the cope and , e.g. for the flavour non - singlet case , can be cast into the form @xmath58 the evolution kernel @xmath59 and the coefficient function @xmath60 are diagonal w.r.t .
the conformal waves the gegenbauer polynomials : @xmath61 and their eigenvalues @xmath62 and @xmath63 coincide with anomalous dimensions and wilson coefficients , respectively , which appear in dis .
a closed form of the function @xmath64 can be deduced order by order in @xmath65 from the sum @xmath66 conformal covariance does not hold true in the minimal subtraction ( ms ) scheme .
the transformation , which rotates the cs result ( [ nonloc - t ] -[def - f ] ) to the minimal one , is induced by the conformal anomalies appearing in the ms scheme
. the rotation matrix can be found by study of the conformal ward identities for green functions with operator insertions and calculating the anomalous contributions via ( modified ) feynman rules . indeed , the off - diagonal piece of the anomalous dimension matrix to nlo is induced by the special conformal anomaly matrix @xmath67 and the first coefficient of the @xmath42-function @xcite : @xmath68_{jk } \over 2 ( j - k)(j + k + 3 ) } , \
\mbox{\boldmath $ b$}_{jk } = - \left [ 1 + ( -1)^{j - k } \right]\theta_{jk } ( 2 k + 3 ) \mbox{\boldmath $ 1 $ } .\end{aligned}\ ] ] the remaining task to transform the coefficient functions to the ms scheme have been solved already in @xcite . in the case of dvcs , they are nothing else as the analytical continuation of the hard - scattering amplitudes for the meson production by two photon - collision . in the case of the axial - vector channel they read : @xmath69 , \nonumber \\ { ^{g}\!t^{a(1 ) } } \!\!\!&=&\!\!\ !
\frac{n_f}{2 } \left [ \left ( \frac{1}{1 - t^2 } + \frac{\ln\frac{1 - t}{2}}{(1 + t)^2 } \right ) \left ( \ln\frac{1 - t}{2 } - 2 \right ) -
\frac{\ln^2 \frac{1 - t}{2}}{2(1 + t)^2 } \right ] .
\nonumber\end{aligned}\ ] ]
there are different proposals for modeling the spd . here
we will deal only with two of them , equating the spd equal to the forward parton distributions ( fpd - model ) and using the proposal of radyushkin for the ansatz of the double distribution ( dd ) functions @xcite @xmath70 , where the profile function @xmath71 for quarks and gluons are given by @xmath72 ^ 2 - z_-^2 } { [ 1 - |z_+|]^3 } , \qquad { ^{g}\!\pi } ( z_- , z_+ ) = \frac{15}{16 } \frac { \left\ { [ 1 - |z_+|]^2 - z_-^2 \right\}^2}{[1 - |z_+|]^5},\ ] ] the spd is obtained by an integration @xcite : @xmath73 we apply the dd model at different scales . to evolve the spd ,
we expand them : @xmath74 in terms of legendre polynomials and express the expansion coefficients @xmath75 in terms of conformal moments which evolution is governed by the anomalous dimension matrix ( [ adcodt ] ) .
this method is suited for not too small @xmath24 . in fig .
[ figinp ] we show our input distributions at a scale of @xmath76 .
let us also mention , that the relative nlo corrections due to the evolution are as expected not very large .
starting at an input scale of 0.7 gev , we found that the nlo corrections do not exceed 10% to 30% @xcite .
for the numerical estimation of the hard - scattering amplitudes ( [ def - t ] ) we perform also an expansion with respect to legendre polynomials which allows us to include easily evolution effects . however , this method is not appropriate to study the low @xmath77 region . in general
we found that the nlo corrections for @xmath78 are typically of the order of 20 - 50% ( see fig .
[ pred1]a ) . in special cases ,
e.g. for the dd model with mrs parameterization , it can even be much larger .
fortunately , there is partial cancellation of perturbative corrections in the ratio of real to imaginary part . as shown in fig .
[ pred1]b , this ratio is model dependent and may its measurement can be used to get a deeper insight in the spd .
99 d. mller et al . , fortschr .
42 ( 1994 ) 101 .
x. ji , phys .
d 55 ( 1997 ) 7114 ; a.v .
radyushkin , phys .
b 380 ( 1997 ) 417 ; j.c .
collins , l. frankfurt , m. strikman , phys .
d 56 ( 1997 ) 2982 .
saul ( for zeus collaboration ) , prompt photon production and deeply virtual compton scattering , talk at eps99 , tampere , finland ( july , 1999 ) .
radyushkin , phys .
d 56 ( 1997 ) 5524 ; j.c .
collins , a. freund , phys .
d 59 ( 1999 ) 074009 ; x. ji , j. osborne , phys . rev .
d 58 ( 1998 ) 094018 . s.a .
anikin , o.i .
zavialov , ann .
( 1978 ) 135 ; b. geyer et al .
, z. phys .
c 26 ( 1985 ) 591 ; i.i .
balitsky , v.m .
braun , nucl .
b 311 ( 1989 ) 541 .
d. mller , phys . rev .
d 49 ( 1994 ) 2525 ; a.v .
belitsky , d. mller , nucl .
b 527 ( 1998 ) 207 , nucl . phys .
b 537 ( 1999 ) 397 .
p. kroll , m. schrmann , p.a.m .
guichon , nucl .
phys . a 598 ( 1996 ) 435 .
frankfurt , a. freund , m. strikman , phys .
d 58 ( 1998 ) 114001 ; p.m. guichon , m. vanderhaeghen , prog . part .
41 ( 1998 ) 125 ; a. freund , m. strikman , single spin asymmetry in dvcs , hep - ph/9906205 . s. ferrara , r. gatto , a.f .
grillo , phys .
b 36 ( 1971 ) 124 , phys .
d 5 ( 1972 ) 3102 .
d. mller , phys . rev .
d 58 ( 1998 ) 054005 ; a.v .
belitsky , d. mller , phys .
b 417 ( 1997 ) 129 .
i.v . musatov and a.v .
radyushkin , evolution and models for skewed parton distributions , hep - ph/9905376 .
belitsky , et al .
b 437 ( 1998 ) 160 , nucl . phys .
b 546 ( 1999 ) 279 . a.d .
martin , r.g .
roberts , w.j .
stirling , phys .
d 47 ( 1993 ) 867 .
m. glck , e. reya , a. vogt , eur .
j. c 5 ( 1998 ) 461 . | we discuss possibilities of measurement of deeply virtual compton scattering amplitudes via different asymmetries in order to access the underlying skewed parton distributions .
perturbative one - loop coefficient functions and two - loop evolution kernels , calculated recently by a tentative use of residual conformal symmetry of qcd , are used for a model dependent numerical estimation of scattering amplitudes . |
the multi- and many - core ( mc ) era we have reached was triggered after the beginning of the century by the stalling of single - processor performance .
technology allowed more transistors to be placed on a die , but they could not reasonably be utilized to increase single - processor performance .
predictions about the number of cores has only partly been fulfilled : today s processors have dozens rather than the predicted hundreds of cores ( although the chinese supercomputer @xcite announced in the middle of 2016 comprises 260 cores on a die ) . despite this , the big players are optimistic .
they expect that moore - law persists , though based on presently unknown technologies
. the effect of the stalled clock frequency is mitigated , and it is even predicted @xcite that `` _ _ now that there are multicore processors , there is no reason why computers should nt begin to work faster , whether due to higher frequency or because of parallel task execution . and with parallel task execution it provides even greater functionality and flexibility!. _ _ ''
parallelism is usually considered in many forums @xcite to be the future , usually as the only hope , rather than as a panacea .
people dealing with parallelism are less optimistic . in general
, the technical development tends to reduce the human effort , but `` _ _ parallel programs ... are notoriously difficult to write , test , analyze , debug , and verify , much more so than the sequential versions _ _ ''
the problems have led researchers to the _ viewpoint _
@xcite , that _ multicore hardware for general - purpose parallel processing is broken_.
the essence of the present viewpoint is that multicore hardware can perhaps be mended .
although one can profoundly agree with the arguments @xcite that using manycore chips can not contribute much to using parallelism in general , and especially not in executing irregular programs , one has to realize also that this is not the optimal battlefield for the manycore chips , at least not in their present architecture .
present manycore systems comprise many segregated processors , which make no distinction between two processing units that are neighbours within the same chip or are located in the next rack .
the close physical proximity of the processing units offers additional possibilities , and provides a chance to implement amdahl s dream @xcite of cooperating processors .
paradigms used presently , however , assume a private processor and a private address space for a running process , and no external world . in many - core systems , it is relatively simple to introduce signals , storages , communication , etc . , and deploy them in reasonable times .
they can not , however , be utilized in a reasonable way , if one can not provide compatibiliy facades providing the illusion of the private world .
cooperation must be implemented in a way which provides complete ( upward ) compatibility with the presently exclusively used single - processor approach ( spa ) @xcite .
it means that on the one hand that new functionality must be formulated using the terms of conventional computing , while on the other , it provides considerably enhanced computing throughput and other advantages .
it is well known , that general purpose processors have a huge handicap in performance when compared to special purpose chips , and that the presently used computing stack is the source of further serious inefficiencies .
proper utilization of available manycore processors can eliminate a lot of these performance losses , and in this way ( keeping the same electronic and programming technology ) can considerably enhance ( apparently ) the performance of the processor .
of course , there is no free lunch .
making these changes requires a _ simultanous _ change in nearly all elements of the present computing stack . before making these changes , one should scrutinize the promised gain , and whether the required efforts will pay off .
cc = [ scale=0.75 ] ( l1 ) @xmath0 ; ( e1 ) [ above of = l1 ] @xmath1 ; ( e2 m ) [ above of = e1 ] ; ( e2 ) at ( @xmath2 ) @xmath3 ; ( l2 ) [ right of = l1 ] @xmath4 ; ( l3 ) [ right of = l2 ] @xmath5 ; ( l4 ) [ right of = l3 ] @xmath6 ; ( x1f ) at ( @xmath7 ) ; ( x1 ) [ below of = x1f ] @xmath8 ; ( x2f ) at ( @xmath9 ) ; ( x2 ) [ below of = x2f ] @xmath10 ; ( plus ) [ below of = x1 ] @xmath11 ; ( minus ) [ below of = x2 ] @xmath12 ; \(a ) [ below of = plus ] @xmath13 ; ( b ) [ below of = minus ] @xmath14 ; ( l1 ) edge ( x1 ) ( l2 ) edge ( x1 ) ( l3 ) edge ( x2 ) ( l4 ) edge ( x2 ) ; ( x1 ) edge ( plus ) ( x1 ) edge ( minus ) ( x2 ) edge ( plus ) ( x2 ) edge ( minus ) ; ( plus ) edge ( a ) ( minus ) edge ( b ) ; ( c1 ) [ left of = l1 ] @xmath15 ; @scan@one@point@firstofone(c1 ) @scan@one@point@firstofone(x1 ) @scan@one@point@firstofone(plus ) ( c2 ) at ( @xmath16 ) @xmath17 ; ( c3 ) at ( @xmath18 ) @xmath19 ; & = [ scale=0.75 ] ( o1 ) @xmath20 ; ( h1 ) at ( @xmath21 ) @xmath22 ; ( h2 ) at ( @xmath23 ) @xmath24 ; ( o1 ) edge [ bend left ] node ( h1 ) ( o1 ) edge [ bend left ] node ( h2 ) ; ( l11 ) at ( @xmath25 ) @xmath26 ; ( l12 ) at ( @xmath27 ) @xmath28 ; ( h1 ) edge [ bend left ] node ( l11 ) ( h1 ) edge [ bend left ] node ( l12 ) ; ( l21 ) at ( @xmath29 ) @xmath30 ; ( l22 ) at ( @xmath31 ) @xmath32 ; ( h2 ) edge [ bend left ] node ( l21 ) ( h2 ) edge [ bend left ] node ( l22 ) ; ( x1 ) at ( @xmath33 ) @xmath8 ; ( l11 ) edge [ bend right ] node ( x1 ) ( l12 ) edge node ( x1 ) ; ( x2 ) at ( @xmath34 ) @xmath10 ; ( l21 ) edge [ bend right ] node ( x2 ) ( l22 ) edge node ( x2 ) ; @scan@one@point@firstofone(o1 ) @scan@one@point@firstofone(x2 ) ( o2 ) at ( @xmath35 ) @xmath36 ; ( x1 ) edge node ( o2 ) ( x2 ) edge [ bend right ] node ( o2 ) ; ( plus ) at ( @xmath37 ) @xmath11 ; ( minus ) at ( @xmath38 ) @xmath12 ; ( o2 ) edge [ bend left ] node ( plus ) ( o2 ) edge [ bend left ] node ( minus ) ; @scan@one@point@firstofone(minus ) ( o3 ) at ( @xmath39 ) @xmath40 ; ( plus ) edge node ( o3 ) ( minus ) edge node ( o3 ) ; ( c1 ) [ left of = o1 ] @xmath15 ; @scan@one@point@firstofone(c1 ) @scan@one@point@firstofone(x1 ) ( c2 ) at ( @xmath16 ) @xmath17 ; @scan@one@point@firstofone(o2 ) ( c3 ) at ( @xmath18 ) @xmath19 ; + below , some easy - to follow case studies are presented , all of which lead to the same conclusion : we need a cooperative and flexible rather than rigid architecture comprising segregated mcs , and the 70-years - old von neumann computing paradigms should be extended . at the end , the feasibility of implementing such an architecture is discussed .
the recently introduced explicitly many - processor approach @xcite seems to be quite promising : it not only provides higher computing throughput , but also offers advantageous changes in the behavior of computing systems .
todays computing utilizes many forms of parallelism @xcite , both hardware ( hw ) and software ( sw ) facilities .
the software is systematically discussed in @xcite and hardware methods are scrutinized in @xcite . a remarkable difference between the two approaches
is , that while the sw methods tend to handle the parallel execution explicitly , the hw methods tend to create the illusion that only one processing unit can cope with the task , although some ( from outside invisible ) helper units are utilized in addition to the visible processing unit .
interestingly enough , both approaches arise from the von neumann paradigms : the abstractions _ process _ and the _ processor _ require so .
the inefficiency of using several processing units is nicely illustrated with a simple example in @xcite ( see also fig [ fig : flexibleproc ] , left side ) . a simple calculation comprising 4 operand loadings and 4 aritmetic operations , i.e. altogether 8 machine instructions ,
could be theoretically carried out in 3 clock cycles , provided that only dependencies restrict the execution of the instructions and an unlimited number of processing units ( or at least 4 such units in the example ) are available .
it is shown that a single - issue processor needs 8 clock cycles to carry out the calculation example .
provided that memory access and instruction latency time can not be further reduced , the only possibility to shorten execution time is to use more than one processing unit during the calculation .
obviously , a fixed architecture can only provide a fixed number of processing units . in the example @xcite
two such ideas are scrutinized : a dual - issue single processor , and a two - core single issue processor .
the hw investment in both cases increases by a factor of two ( not considering the shared memory here ) , while the performance increases only moderately : 7 clock cycles for the dual - issue processor and 6 clock cycles for the dual - core processor , versus the 8 clock cycles of the single - issue single core processor .
the _ obvious reasons here are the rigid architecture and the lack of communication possibilities _ , respectively .
consider now a processor with flexible architecture , where the processor can outsource part of its job : it can rent processing units from a chip - level pool just in the time it takes to execute a few instructions .
the cores are smart : they can communicate with each other , and even they know the task to be solved and are able to organize their own work while outsourcing part of the work to the rented cores .
the sample calculation , borrowed from @xcite as shown in fig .
[ fig : flexibleproc ] , left side , can then be solved as shown on the right side of the figure .
the core ; originally receives the complete task to make the calculation , as it would be calculated by a conventional single - issue , single core system , in 8 clock cycles .
however , ; is more intelligent . using the hints hidden in the object code
, it notices that the task can be outsourced to another cores . for this purpose
it rents , one by one , cores ; and ; to execute two multiplications .
the rented ; cores are also intelligent , so they also outsource loading the operands to cores ; and ; .
they execute the outsourced job : load the operands and return them to the requesting cores ; , which then can execute the multiplications ( denoted by ;) and return the result to the requesting core , which can then rent another two cores ; and ; for the final operations . two results
are thus produced .
this unusual kind of architecture must respond to some unusual requirements .
first of all , the architecture must be able to organize itself as the received task requires it , and build the corresponding `` processing graph '' , see fig .
[ fig : dynpar ] , for legend see @xcite .
furthermore , it must provide a mechanism for mapping the virtually infinite number of processing nodes to the finite number of cores .
cores ; must receive the address of the operand , i.e. at least some information must be passed to the rented core .
similarly , the loaded operand must be returned to the renting core in a synchronized way . in the first case synchronization
is not a problem : the rented core begins its independent life when it receives its operands . in the second case
the rented core finishes its assigned operation and sends the result asyncronously , independently of the needs of the renting core .
this means that the architecture must provide a mechanism for transferring some ( limited amount of ) data between cores , a signalization mechanism for renting and returning cores , as well as a latched intermediate data storage for passing data in a synchronized way .
the empty circles are the theoretically needed operations , and the shaded ones are additional operations of the `` smart '' cores .
the number of the cores being used changes continuously as they are rented and returned .
although _ physically _ they may be the same core , _ logically _ they are brand new .
note that the `` smart '' operations are much shorter they comprise simple bit manipulations and multiplexing , than the conventional ones that comprise complex machine instructions , and since the rented cores work in parallel ( or at least mostly overlap ) , the calculation is carried out in 3 clock periods .
the cycle period is somewhat longer , but the attainable parallelism approaches the theoretically possible one , and is more than twice as high as the one attainable using either two - issue or dual - core processors .
although the average need of cores is about 3 , these cores can be the simplest processors , i.e. the decreasing complexity of the cores ( over)compensates for the increasing complexity of the processor .
in addition , as the control part of the processors increases , the need for the hidden parallelization ( like out - of - order and speculation ) can be replaced by the functionality of the flexible architecture , the calculational complexity can be decreased , and as a result , the clock speed can be increased .
a processor with such an internal architecture appears to the external world as a `` superprocessor '' , having several times greater performance than could be extracted from a single - threaded processor . that processor can adapt itself to the task : unlike in the two issue processor , all ( rented ) units are permanently used .
_ the many - core systems with flexible architecture comprising cooperating cores can approach the theoretically possible maximum parallelism . _ in addition , the number of the cores can be kept at a strict minimum , allowing reduction of the power consumption .
in the many - many processor ( supercomputer ) systems the processing units are assembled using the spa @xcite , and so their maximum performance is bounded by amdahl s law . although amdahl s original model @xcite is pretty outdated , its simple and clean interpretation allows us to derive meaningful results even for today s computing systems .
amdahl assumed that in some @xmath41 part of the total time the computing system engages in parallelized activity , in the remaining ( @xmath42 ) part it performs some ( from the point of view of parallelization ) non - payload activity , like sequential processing , networking delay , control or organizational operation , etc .
the essential point here is that all these latter activities behave _ as if they were sequential processing_. under such conditions , the efficiency @xmath43 is calculated as the ratio of the total speedup @xmath44 and the number of processors @xmath45 : @xmath46 although in the case of supercomputers ( @xmath42 ) comprises contributions of a technically different nature ( it can be considered as the `` imperfectness '' of implementation of the supercomputer ) , it also behaves as if it were a sequentially processed code .
[ supercomputertimeline ] shows how this `` imperfectness '' was decreased during the development of supercomputers , calculated from the actual data of the first three supercomputers in the year in question over a quarter of a century .
as the figure shows , this parameter behaves similarly to the moore - observation , but it is independent of that one ( because the parameter is calculated from @xmath47 , any technology dependence is removed ) . at first glance , it seems to be at least surprising to look for any dependence in function of `` imperfectness '' .
the key is equ .
( [ eq : soverk ] ) .
since the @xmath41 approaches unity , the term @xmath48 determines the overall efficiency of the computing system . to _ increase _ @xmath45 by an order or magnitude alone is useless if not accompanied by an order of magnitude _ decrease _ in the value of ( @xmath42 ) .
however , while increasing @xmath45 is simply a linear function , decreasing ( @xmath42 ) as any kind of increasing perfectness , is exponentially more difficult .
[ supercomputertimeline ] proves that today s supercomputers are built in spa , and makes it questionable whether further significant decrease of value ( @xmath42 ) could be reached at reasonable cost .
this means that it is hopeless _ to build exa - scale computers , using the principles drawn from the spa_. looking carefully at @xmath48 , one can notice that the two terms describe two important behavioral features of the computing system
. as already discussed , @xmath49 decribes , how much the work of the many - processor system is _
coordinated_. the factor @xmath45 , on the other hand , describes , how much the processing units _
cooperate_. in the case of using the spa , the processing units are segregated entities , i.e. they do not cooperate at all .
if we could make a system where the processing units behave differently in the presence of another processors , we could write @xmath50 in equ .
( [ eq : soverk ] ) . depending on how cores behave together in the presence of another cores when solving a computing task
, the @xmath50 , the cooperation of the processing units can drastically increase the efficiency of the many - processor systems .
in other words , to increase the performance of many - many - processor computers , _ the cores must cooperate _
( at least with some ) other cores . _ using cooperating cores is inevitable for building supercomputers at a reasonable cost . _
a computer computes everything , because it can not do any other type of operations .
computational density has reached its upper bound , so no further performance increase in that direction is possible .
in addition to introducing different forms of hw and sw parallelism , it is possible to omit some non - payload , do - not - care calculations , through providing and utilizing special hw signals instead .
the signals can be provided for the participating cores , and can be used to replace typical calculational instruction sequences by using special hardware signals .
the compilation is simple : where the compiler should generate non - payload loop organization commands , it should give a hint about renting a core for executing non - payload instructions and providing external synchronization signals . a simple example : when summing up elements of a vector , the only payload instruction is the respective ` add ` .
one has , however , to address the operand ( which includes handling the index , calculating the offset and adding it to the base address ) , to advance the loop counter , to compare it to the loop bound , and to jump back conditionally .
all those non - payload operations can be replaced by handling hw signals , if the cores can cooperate , resulting in a speed gain of about 3 , using an extra core only .
even , since the intermediate sum is also a do - not - care value until the summing is finished , a different sumup method can be used , which may utilize dozens of cores and result in a speed gain of dozens .
when organizing a loop , the partial sum is one of the operands , so it must be read before adding a new summand , and must be written back to its temporary storage , wasting instructions and memory cycles ; in addition it excludes the possibility of parallelizing the sumup operation . for details and examples
see @xcite .
this latter example also demonstrates that _ the machine instruction is a too rigid atomic unit of processing_. _ utilizing hw signals from cooperating cores rather than providing some conditions through ( otherwise don - not - care ) calculations , allows us to eliminate obsolete calculational instructions , and thus apparently accelerate the computation by a factor of about ten . _
the need for multi - processing ( among others ) forced to use exceptional instruction execution .
i.e. , a running process is _ _
interrupt__ed , its hw and sw state is saved and restored , because the hard and soft parts of the _ only _ processor must be lent to another process .
the code of the interrupting process is effectively inserted in the flow of executing the interrupted code .
this maneuver causes an indeterministic behavior of the processor : the time when two consecutive machine instructions in a code flow are executed , becoming indeterminate .
the above is due to the fact that during development , some of the really successful accelerators , like the internal registers and the highest level cache , became part of the architecture : the soft part of the processor . in order to change to a new thread
, the current soft part must be saved in ( and later restored from ) the memory .
utilizing asynchronous interrupts as well as operating system services , implies a transition to new operating mode , which is a complex and very time - consuming process .
all these extensions were first developed when the computer systems had only one processor , and the only way to provide the illusion of running several processes , each having its own processor , was to detach the soft part from the hard one . because of the lack of proper hardware support , this illusion depended on using sw services and on the architectures being constructed with a spa in mind , conditions that require rather expensive execution time : in modern systems a context change may require several thousands of clock cycles . as the hyper - threading proved , detaching soft and hard part of the processors results in considerable performance enhancement . by having more than one processor and the explicitly many - processor approach @xcite , the context change can be greatly symplified .
for the new task , such as providing operating system services and servicing external interrupts a dedicated core can be reserved .
the dedicated core can be prepared and held in supervisor mode .
when the execution of the instruction flow follows , it is enough to clone the relevant part of the soft part : for interrupt servicing nothing is needed , for using os services only the relevant registers and maybe cache .
( the idea is somewhat similar to utilizing shadow registers for servicing an asynchronous interrupt . )
if the processors can communicate among each other using hw signals rather than os actions , and some communication mechanism , different from using ( shared ) memory is employed , the apparent performance of the computing systems becomes much faster .
_ for cooperating cores no machine instructions ( that waste real time , machine and memory cycles ) are needed for a context change , allowing for a several hundredfold more rapid execution in these spots .
_ the application can even run parallel with the system code , allowing further ( apparent ) speedup . using the many - processor approach
creates many advantageous changes in the real - time behavior of the computing systems .
since the processing units do not need to save or restore anything , the servicing can start immediately and is restricted to the actual payload instructions .
the dedicated processing units can not be addressed by non - legal processing units , so issues like exluding priority inversion are handled at hw level .
and so on .
from all points of view ( the just - a - few and many - many processors , as well as utilizing kernel - mode or real - time services ) we arrive at the same conclusion : segregated processors in the many - processor systems do not allow a greater increase in the performance of our computing systems , while cooperating processors can increase the attainable single - threaded performance .
amdahl contented this by a half century ago : `` _ the organization of a single computer has reached its limits and that truly significant advances can be made only by interconnection of a multiplicity of computers in such a manner as to permit cooperative solution .
@xcite at this point the many - core architectures have the advantage that they are in the close proximity to one another : there is no essential difference between that a core needing to reach its own register ( or signal ) or that of another core .
the obstacle is actually the spa : for a core and a process , there exists no other core . in the suggested new approach , which can be called _ explicitly many - processor approach _
( empa ) , the cores ( through their supervisor ) can know about their neighbours . today
, radical departures from conventional approaches ( including rethinking the complete computing stack ) are advanced @xcite , but at the same time a smooth transition must be provided to that radically new technology . _ to preserve compatibility with conventional computing , the empa approach @xcite is phrased using the terms of conventional computing _
( i.e. it contains spa as a subset ) .
some of the above - mentioned boosting principles are already implemented in the system . from the statistics
one can see that in some spots , performance gain in the range 3 - 30 can be reached .
the different algorithms need different new accelerator building stone solutions in frame of empa .
for example , the gain 3 in an executing loop , when used in an image processing task where for edge detection a 2-dimensional matrix is utilized , means nearly an order of magnitude performance gain , using the same calculational architecture in calculating a new point . and , to consider all points of the picture another double loop is used .
this means , that a 4-core empa processor can produce nearly 100 times more rapid processing ( not considering that several points can be processed in parallel on processors with more cores ) .
this is achieved not by increasing computing density , but by replacing certain non - payload calculations with hw signals , and so executing 100 times less machine instructions .
the mc architecture comprising segregated cores is indeed broken .
it can , however , be mended , if the manycore chips are manufactured in the form using cooperating cores . as the first step toward implementing such a system , for simulating its sophisticated internal operation and providing tools for understanding and validating it , an empa development system @xcite has been prepared .
an extended assembler prepares empa - aware object code , while the simulator allows us to watch the internal operation of the empa processor . to illustrate the execution of programs using the empa method ,
a processing diagram is automatically prepared by the system , and different statistics are assembled .
[ fig : dynpar ] shows the equivalent of fig .
[ fig : flexibleproc ] , running on an 8-core and a 4-core processor , respectively ( for legend see @xcite ) .
the left hand figure depicts the case when `` unlimited '' number of processing units are available , the right hand one shows the case when the processor has a limited number of computing resources to implement the maximum possible parallelism .
the code assembled by the compiler is the same in both cases .
the supervisor logic detects if not enough cores are available ( see right side ) , and delays the execution ( outsourcing more code ) of the program fragments until some cores are free again .
the execution time gets longer if the processor can not rent enough cores for the processing , but the same code will run in both cases , without deadlock and violating dependencies . for electronic implementation
, some ideas may be borrowed from the technology of reconfigurable systems . there , in order to minimize the need for transferring data , some local storage ( block - ram ) is located between the logical blocks , and a lot of wires is available for connecting them . in analogy also with fpgas
, the cores can be implemented as mostly fixed functionality processing units , having multiplexed connecting wires to their supervisor with fixed routing .
some latch registers and non - stored program functionality gates can be placed near those blocks , which can be accessed by both cores and supervisor .
the inter - core latch data can be reached from the cores using pseudo - registers ( i.e. they have a register address , but are not part of the register file ) and the functionality of the cores also depends on the inter - core signals . in the prefetch stage the cores can inform the supervisor about the presence of metainstruction in their object code , and in this way the mixed code instructions can be directed to the right destination . in order to be able to organize execution graphs , the cores ( after renting ) are in parent - child relation to unlimited depth .
as was very correctly stated @xcite , `` due to its high level of risk , prototype development fits best within the research community . ''
the principles and practice of empa differ radically from those of spa . to compare the performance of both
, empa needs a range of development .
many of the present components , accelerators , compilers , etc .
, with spa in mind , do not fit empa
. the research community can accept ( or reject ) the idea , but it definitely warrants some cooperative work .
h. esmaeilzadeh .
approximate acceleration : a path through the era of dark silicon and big data . in _ proceedings of the 2015 international conference on compilers , architecture and synthesis for embedded systems _ , cases 15 , pages 3132 , piscataway , nj , usa , 2015 .
ieee press . | a suggestion is made for mending multicore hardware , which has been diagnosed as broken . |
galaxy clustering as measured in redshift - space contains the imprint of the linear growth rate of structure @xmath5 , in the form of a measurable large - scale anisotropy @xcite .
this is produced by the coherent peculiar velocity flows towards overdensities , which add an angle - dependent contribution to the measured redshift . in linear theory , these _ redshift - space distortions _
( rsd ) in the clustering pattern can be quantified in terms of the ratio @xmath6 ( where @xmath7 is the linear bias of the sample of galaxies considered ) .
a value for @xmath4 can be obtained by modeling the anisotropy of the redshift - space two - point correlation function @xmath8 ( where @xmath9 and @xmath10 are the separations perpendicular and parallel to the line of sight ) or , equivalently , of the power spectrum ( see @xcite for a review ) . since @xmath7 can be defined as the ratio of the _ rms _ galaxy clustering amplitude to that of the underlying matter , @xmath11 , the measured product @xmath12 is equivalent to the predicted combination @xmath13 @xcite .
the latter is a prediction depending on the gravity theory , once normalized to the amplitude of matter fluctuations at the given epoch , e.g. using cmb measurements .
measurements of the growth rate @xmath5 are crucial to pinpoint the origin of cosmic acceleration , distinguishing whether it requires the addition of `` dark energy '' in the cosmic budget , or rather a modification of general relativity .
these two radically alternative scenarios are degenerate when considering the expansion rate @xmath14 alone , as yielded , e.g. , by the hubble diagram of type ia supernova ( e.g. @xcite ) or baryonic acoustic oscillations ( bao , e.g @xcite ) .
although the rsd effect is well known since long , its important potential in the context of dark energy studies has been fully appreciated only recently @xcite .
this led to a true renaissance of interest in this technique @xcite , such that rsd have quickly become one of the most promising probes for future large dark energy surveys .
this is the case of the recently approved esa euclid mission @xcite , which is expected to reach statistical errors of a few percent on measurements of @xmath5 in several redshift bins out to @xmath15 using this technique ( coupled to similar precisions with the complementary weak - lensing experiment ) . in general , forecasts of the statistical precision reachable by future projects on the measurements of different cosmological parameters have been produced through widespread application of the so - called fisher information matrix technique @xcite .
this has also been done specifically for rsd estimates of the growth rate and related quantities @xcite .
one limitation of these forecasts is that they necessarily imply some idealized assumptions ( e.g. on the gaussian nature of errors ) and have not been verified , in general , against systematic numerical tests .
this is not easily doable in general , given the large size of planned surveys .
a first attempt to produce general forecasts based on numerical experiments was presented by @xcite , who used mock surveys built from the millennium simulation @xcite to numerically estimate the random and systematic errors affecting their measurement of the growth rate from the vimos vlt deep survey . using a grid of reference survey configurations , they calibrated an approximated scaling relation for the relative error on @xmath4 as a function of survey volume and mean density .
the range of parameters explored in this case was however limited , and one specific class of galaxies only ( i.e. bias ) was analyzed .
the second crucial aspect to be taken into consideration when evaluating fisher matrix predictions , is that they only consider statistical errors and can not say anything about the importance of systematic effects , i.e. on the _ accuracy _ of the expected estimates .
this is clearly a key issue for projects aiming at percent or sub - percent precisions , for which systematic errors will be the dominant source of uncertainty .
in fact , a number of works in recent years suggest that the standard linear kaiser description of rsd is not sufficiently accurate on quasi - linear scales ( @xmath16 ) where it is routinely applied ( @xcite ) .
various non - linear corrections are proposed in these papers , the difficulty often being their practical implementation in the analysis of real data , in particular in configuration space @xcite .
one may hope that in the future , with surveys covering much larger volumes , it will be possible to limit the analysis to very large scales , where the simple linear description should be adequate .
still , ongoing surveys like wigglez @xcite , boss @xcite and vipers ( guzzo et al .
, in preparation ) , will still need to rely on the clustering signal at intermediate scales to model rsd . here
, we shall address in a more systematic and extended way the impact of random and systematic errors on growth rate measurements using rsd in future surveys .
we shall compare the results directly to fisher matrix predictions , thoroughly exploring the dependence of statistical errors on the survey parameters , including also , in addition to volume and density , the bias parameter of the galaxies used .
this is also relevant , as one could wonder which kind of objects would be best suited to measure rsd in a future project .
these will include using halos of different mass ( i.e. bias ) , up to those traced by groups and clusters of galaxies .
potentially , using groups and clusters to measure rsd could be particularly interesting in view of massive galaxy redshift surveys as that expected from euclid @xcite , which can be used to build large catalogues of optically - selected clusters with measured redshifts . a similar opportunity will be offered by future x - ray surveys , such as those expected from the e - rosita mission @xcite , although in that case , mean cluster redshifts will have to be measured first .
this paper is complementary to the parallel work of @xcite , where we investigate the impact on rds of redshift errors and explore how to disentangle geometrical distortions introduced by the uncertainty of the underlying geometry of the universe
i.e. the alcock - paczynski effect @xcite on measurements of rsd .
also , while we were completing our work , independent important contributions in the same direction appeared in the literature by @xcite and @xcite .
the paper is organized as follows . in
[ sec sims ] we describe the simulations used and the mass - selected subsamples we defined ; in [ sec rsd ] we discuss the technical tools used to estimate and model the two - point correlation function in redshift space , @xmath8 , and to estimate the intrinsic values of bias and distortion to be used as reference ; in [ sec err ] we present the measured @xmath8 and show the resulting statistical and systematic errors on @xmath4 , as a function of the halo bias ; here we discuss in detail how well objects related to high - bias halos , as groups and clusters , can be used to measure rsd ; in [ sec form ]
we organise all our results into a compact analytic formula as a function of galaxy density , bias and survey volume ; we then directly compare these results to the predictions of a fisher matrix code ; finally we summarize our results in
[ sec concl ] .
the core of this study is based on the high - resolution baryonic acoustic - oscillation simulations at the institute for computational cosmology ( basicc ) of @xcite , which used @xmath17 particles of mass @xmath18 to follow the growth of structure in dark matter in a periodic box of side @xmath19 .
the simulation volume was chosen to allow for growth of fluctuations to be modelled accurately on a wide range of scales including those of bao .
the very large volume of the box also allows us to extract accurate measurements of the clustering of massive halos .
the mass resolution of the simulation is high enough to resolve halos that should host the galaxies expected to be seen in forthcoming high - redhift galaxy surveys ( as e.g. luminous red galaxies in the case of sdss - iii boss ) .
the cosmological parameters adopted are broadly consistent with recent data from the cosmic microwave background and the power spectrum of galaxy clustering ( @xcite ) : the matter density parameter is @xmath20 , the cosmological constant density parameter @xmath21 , the normalization of density fluctuations , expressed in terms of their linear amplitude in spheres of radius @xmath22 at the present day @xmath23 , the primordial spectral index @xmath24 , the dark energy equation of state @xmath25 , and the reduced hubble constant @xmath26 .
we note the high value of normalization of the power spectrum @xmath27 , with respect to more recent wmap estimates ( @xmath28 , @xcite ) .
this has no effect on the results discussed here ( but see @xcite for a method to scale self - consistently the output of a simulation to a different background cosmology ) .
outputs of the particle positions and velocities are stored from the simulations at selected redshifts .
dark matter halos are identified using a friends - of - friends ( fof ) percolation algorithm @xcite with a linking length of @xmath29 times the mean particle separation .
position and velocity are given by the values of the center of mass . in this paper , only groups with al least @xmath30 particles are considered ( i.e only halos with mass @xmath31 ) .
this limit provides reliable samples in term of their abundance and clustering , which we checked by comparing the halo mass function and correlation function against @xcite and @xcite respectively .
we use the complete catalogue of halos of the simulation at @xmath32 , from which we select sub - samples with different mass thresholds ( i.e. number of particles ) .
this corresponds to samples with different bias values .
table [ tab halo masses ] reports the main features of these catalogues .
.properties of the halo catalogues used in the analysis .
@xmath33 is the threshold value of @xmath34 , e.g. the catalogue @xmath35 is the set of groups ( i.e. halos ) with at least @xmath36 dm particles ; @xmath37 is the corresponding threshold mass ; @xmath38 is the total number of halos ( i.e. the number of halos with @xmath39 ) ; @xmath40 is the number density ( i.e. @xmath41 , where @xmath42 is the simulation volume ) .
[ cols="^,^,^,^",options="header " , ] in doing this exercise , a specific problem is that , as shown in table [ tab halo masses ] , catalogues with larger mass ( i.e. higher bias ) are also less dense .
our aim is to separate the dependence of the errors on these two variables .
to do so , once a population of a given bias is defined by choosing a given mass threshold , we construct a series of diluted samples obtained by randomly removing objects .
the process is repeated down to a minimum density of @xmath43 , at which shot noise dominates and for the least massive halos the recovered @xmath4 is consistent with zero . in this way
, we obtain a series of sub - samples of varying density for fixed bias , as reported in table [ tab dilution ] .
the full samples are the same used to build , e.g. , figure [ fig beta lin - exp ] . in figure [ fig surface ]
we plot the relative errors on @xmath4 measured from each catalogue of table [ tab dilution ] , as a function of the bias factor and the number density .
these 3d plots are meant to provide an overview of the global behavior of the errors ; a more detailed description is provided in figures [ fig fisher n]-[fig fisher b ] , where 2d sections along @xmath40 and @xmath7 are reported .
for all the samples considered , the volume is held fixed . as shown by the figure ,
the bias dependence is weak and approximately described by @xmath44 , i.e. the error is slightly larger for higher - bias objects .
this indicates that the gain of a stronger clustering signal is more than cancelled by the reduction of the distortion signal , when higher bias objects are considered .
this is however fully true only for samples which are not too sparse intrinsically .
we see in fact that at extremely low densities , the relationship is inverted , with high - bias objects becoming favoured . at the same time , there is a clear general flattening of the dependence of the error on the mean density @xmath40 . the relation is not a simple power - law , but becomes constant at high values of @xmath40 . in comparison , over the density range considered here , the old scaling formula of guzzo et al . would overestimate the error significantly .
this behaviour is easily interpreted as showing the transition from a shot - noise dominated regime at low densities to a cosmic - variance dominated one , in which there is no gain in further increasing the sampling .
such behaviour is clear for low - mass halos ( i.e. low bias ) but is much weaker for more massive , intrinsically rare objects .
we can now try to model an improved empirical relation to reproduce quantitatively these observed dependences .
let us first consider the general trend , @xmath45 , which describes well the trend of @xmath46 in the cosmic variance dominated region ( i.e. at high density ) . in figure [ fig surface ] such a power - law
is represented by a plane .
we then need a function capable to warp the plane in the low density region , where the relative error becomes shot - noise dominated .
the best choice seems to be an exponential : @xmath47 , where , by construction , @xmath48 roughly corresponds to the threshold density above which cosmic variance dominates .
finally , we need to add an exponential dependence on the bias so that at low density the relative error decreases with @xmath7 , such that the full expression becomes @xmath49 $ ] .
the grid shown in figure [ fig surface ] represents the result of a direct fit of this functional form to the data , showing that it is indeed well suited to describe the overall behaviour . in the right panel we have oriented the axes as to highlight the goodness of the fit : the _ rms _ of the residual between model and data is @xmath50 , which is an order of magnitude smaller than the smallest measured values of @xmath46 .
this gives our equation the predictive power we were looking for : if we use it to produce forecasts of the precision of @xmath4 for a given survey , we shall commit a negligible error ( figure [ fig fisher n ] ) , with the _ rms _ of the residuals , @xmath50 . ]
( @xmath51 ) on @xmath46 ( at least for values of bias and volume within the ranges tested here ) . to fully complete the relation , we only need to add the dependence on the volume , which is in principle the easiest .
to this end , we split the whole simulation cube into @xmath52 sub - cubes , with @xmath53 . by applying this procedure to 5 samples with different bias and number density ( see table [ tab dilution ] ) we make sure that our results do not depend on the particular choice of bias and density .
figure [ fig rbt vs vol ] shows that @xmath54 independently of @xmath40 and @xmath7 , confirming the dependence found by @xcite .
we can thus finally write the full scaling formula for the relative error of @xmath4 we were seeking for @xmath55 where @xmath56 and @xmath57 .
clearly , by construction , this scaling formula quantifies random errors , not the systematic ones .
the fisher information matrix provides a method for determining the sensitivity of a particular experiment to a set of parameters and has been widely used in cosmology . in particular , @xcite introduced an implementation of the fisher matrix aimed at forecasting errors on cosmological parameters derived from the galaxy power spectrum @xmath58 , based on its expected observational uncertainty , as described by ( * ? ? ?
this was adapted by @xcite to the measurements of distances using the baryonic acoustic oscillations in @xmath58 . following the renewed interest in rsd , over the past few years
the fisher matrix technique has also been applied to predict the errors expected on @xmath4 , @xmath59 and related parameters ( e.g * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the extensive simulations performed here provides us with a natural opportunity to perform a first simple and direct test of these predictions .
given the number of details that enter in the fisher matrix implementation , this can not be considered as exhaustive . yet , a number of interesting indications emerge , as we shall see .
we have computed fisher matrices for all catalogues in table [ tab dilution ] , using a code following @xcite . in particular , our fisher matrix predicts errors on @xmath4 and @xmath7 , given the errors on the linear redshift space power spectrum modeled as in eq .
( [ eq kaiser spectrum beta ] ) @xcite .
we first limit the computations to linear scales , applying the standard cut - off @xmath60 .
we also explore the possibility of including wavenumbers as large as @xmath61 ( that should better match the typical scales we fit in the correlation functions from the simulations ) , accounting for non - linearity through a conventional small - scale lorentzian damping term .
our fiducial cosmology corresponds to that used in the simulation , i.e. @xmath20 , @xmath21 , @xmath62 and @xmath23 today .
we also choose @xmath63 km s@xmath64 as reference value for the pairwise dispersion .
we do not consider geometric distortions @xcite , whose impact on rsd is addressed in the parallel paper by @xcite . to obtain the fisher predictions on @xmath4
, we marginalize over the bias , to account for the uncertainty on its precise value , and on the pairwise velocity in the damping term ( when present ) .
figure [ fig fisher n ] shows the measured relative errors on @xmath4 as a function of the number density , compared to the fisher forecasts for the two choices of @xmath65 .
we also plot the scaling relation from eq .
( [ eq err mine general ] ) , which best represents the simulation results .
we see that the simulation results are in in fairly good agreement with the fisher predictions , when we limit the computation to very linear scales in the power spectrum ( solid line ) .
the inclusion of higher wavenumbers produces unrealistically small errors and with a wrong dependence on the number density .
both the solid lines and points reproduce the observed flattening at high number densities , which corresponds to the transition between a shot - noise and a cosmic - variance dominated regime , respectively . similarly ,
figure [ fig fisher b ] looks at the dependence of the error on the linear bias parameter , comparing the simulation results ( points and scaling formula best - fit ) to the fisher forecasts .
the behaviour is similar to that observed for the number density : there is a a fairly good agreement when the fisher predictions are computed using @xmath66 , except for very low values of the number density and the bias .
again , when non - linear scales are included , the fisher predictions become too optimistic by a large factor .
we have performed an extensive investigation of statistical and systematic errors in measurements of the redshift - distortion parameter @xmath4 from future surveys . we have considered tracers of the large - scale distribution of mass with varying levels of bias , corresponding to objects like galaxies , groups and clusters . to this purpose
, we have analyzed large catalogues of dark - matter halos extracted from a snapshot of the basicc simulation at @xmath32 .
our results clearly evidence the limitations of the linear description of redshift - space distortions , showing how errors depend on the typical survey properties ( volume and number density ) and the properties of the tracers ( bias , i.e. typical mass ) .
let us recap them and discuss their main implications . * estimating @xmath4 using the hamilton / kaiser harmonic expansion of the redshift - space correlation function @xmath1 extended to typical scales , leads to a systematic error of up to @xmath67 .
this is much larger than the statistical error of a few percent reachable by next - generation surveys .
the larger systematic error is found for small bias objects , and decreases reaching a minimum for halos of @xmath68 .
this reinforces the trend observed by @xcite . *
additional analysis of mock surveys from the millennium run confirm that the observed systematic errors are not the result of potentially missing sub - structure in the basicc halo catalogues .
* the use of the deprojected correlation function increases the statistical error , inducing also some additional systematic effects ( details are given in appendix [ sec deprojected ] and also in the companion paper by @xcite ) . * for highly biased objects , which are sparser and whose
surveys typically cover larger , more linear scales , the simple kaiser model describes fairly well the simulated data , without the need of the empirical damping term with one extra parameter accounting for non - linear motions .
this results in smaller statistical errors .
* we have derived a comprehensive scaling formula , eq .
( [ eq err mine general ] ) , to predict the precision ( i.e. relative statistical error ) reachable on @xmath4 as a function of survey parameters .
this expression improves on a previous attempt @xcite , generalizing the prediction to a population of arbitrary bias and properly describing the dependence on the number density . +
this formula can be useful to produce quite general and reliable forecasts for future surveys ] .
one should in any case consider that there are a few implementation - specific factors that can modify the absolute values of the recovered _ rms _ errors .
for example , these would depend on the range of scales over which @xmath1 is fitted .
the values obtained here refer to fits performed between @xmath69 and @xmath70 .
this has been identified through several experiments as an optimal range to minimize statistical and systematic errors for surveys this size @xcite .
theoretically , one may find natural to push @xmath71 , or both @xmath72 and @xmath71 to larger scales , as to ( supposedly ) reduce the weight of nonlinear scales . in practice , however , in both cases we see that random errors increase in amplitude ( while the systematic error is not appreciably reduced ) .
+ similarly , one should also keep in mind that the formula is strictly valid for @xmath32 , i.e. the redshift where it has been calibrated .
there is no obvious reason to expect the scaling laws among the different quantities ( density , volume , bias ) to depend significantly on the redshift .
this is confirmed by a few preliminary measurements we performed on halo catalogues from the @xmath73 snapshot of the basicc .
conversely , the magnitude of the errors may change , as shown , e.g. , in @xcite .
we expect these effects to be described by a simple renormalization of the constant @xmath74 .
+ finally , one may also consider that the standard deviations measured using the 27 sub - cubes could be underestimated , if these are not fully independent .
we minimize this by maximizing the size of each sub - cube , while having enough of them as to build a meaningful statistics .
the side of each of the 27 sub - cubes used is in fact close to @xmath75 , benefiting of the large size of the basicc simulation .
* we have compared the error estimations from our simulations with idealized predictions based on the fisher matrix approach , customarily implemented in fourier space .
we find a good agreement , but only when the fisher computation is limited to significantly large scales , i.e. @xmath76 .
when more non - linear scales are included ( as an attempt to roughly match those actually involved in the fitting of @xmath77 in configuration space ) , then the predicted errors become unrealistically small .
this indicates that the usual convention of adopting @xmath78 for these kind of studies is well posed . on the other hand
, it seems paradoxical that in this way with the two methods we are looking at different ranges of scales .
the critical point clearly lies in the idealized nature of the fisher matrix technique .
when moving up with @xmath65 and thus adding more and more nonlinear scales , the fisher technique simply accumulates signal and dramatically improves the predicted error , clearly unaware of the additional `` noise '' introduced by the breakdown of linearity . on the other hand , if in the direct fit of @xmath79 ( or @xmath80 ) one conversely considers a corresponding very linear range @xmath81 , a poor fit is obtained , with much larger statistical errors than shown , e.g. , in fig .
[ fig beta lin - exp ] .
there is no doubt that smaller , mildly nonlinear scales at intermediate separations have necessarily to be included in the modelling if one aims at reaching percent statistical errors on measurements of @xmath4 ( or @xmath59 ) .
if one does this in the fisher matrix , then the predicted errors are too small .
the need to push our estimates to scales which are not fully linear will remain true even with surveys of the next generation , including tens of millions of galaxies over gpc volumes , because that is where the clustering and distortion signals are ( and will still be ) the strongest .
of course , our parallel results on the amount of systematic errors that plague estimates based on the standard dispersion model also reinforce the evidence that better modelling of nonlinear effects is needed on these scales .
the strong effort being spent in this direction gives some confidence that significant technical progress will happen in the coming years ( see e.g. * ? ? ?
* ; * ? ? ?
* and references therein ) .
+ in any case , this limited exploration suggests once more that forecasts based on the fisher matrix approach , while giving useful guidelines evidence the error dependences , have to be treated with significant caution and possibly verified with more direct methods .
similar tension between fisher and monte carlo forecasts has been recently noticed by @xcite . *
finally , in appendix [ sec likelihood ] we have also clarified which is the most unbiased form to be adopted for the likelihood when fitting models to the observed redshift - space correlation function , proposing a slightly different form with respect to previous works . with
redshift - space distortions having emerged as probe of primary interest in current and future dark - energy - oriented galaxy surveys , the results presented here further stress the need for improved descriptions of non - linear effects in clustering and dynamical analyses . on the other hand , they also indicate the importance of building surveys for which multiple tracers of rsd ( with different bias values ) can be identified and used in combination to help understanding and minimizing systematic errors .
we warmly thank m. bersanelli for discussions and constant support and c. baugh for his invaluable contribution to the basicc simulations project .
db acknowledges support by the universit degli studi di milano through a phd fellowship .
em is supported by the spanish micinns juan de la cierva programme ( jci-2010 - 08112 ) , by cicyt through the project fpa-2009 09017 and by the community of madrid through the project hephacos ( s2009/esp-1473 ) under grant p - esp-00346
. financial support of prin - inaf 2007 , prin - miur 2008 and asi contracts i/023/05/0 , i/088/06/0 , i/016/07/0 , i/064/08/0 and i/039/10/0 is gratefully acknowledged .
lg is partly supported by erc advanced grant # 291521 ` darklight ' .
99 acquaviva v. , hajian a. , spergel d. n. , das s. , 2008 , phrvd , 78 , 043514 alcock c. , paczynski b. , 1979 , nature , 281 , 358 angulo r. e. , baugh c. m. , frenk c. s. , lacey c. g. , 2008 , mnras , 383 , 755 angulo , r.e . , white , s.d.m .
, 2010 , mnras , 405 , 143 bianchi d. , 2010 , master laurea thesis , unversity of milan blake c. , et al . , 2011 , mnras , 415 , 2876 bueno belloso a. , garca - bellido j. , sapone d. , 2011 , jcap , 10 , 10 cabr a. , gaztaaga e. , 2009 , mnras , 393 , 1183 cappelluti n. , et al . , 2011 , msais , 17 , 159 davis m. , peebles p. j. e. , 1983 , apj , 267 , 465 davis m. , efstathiou g. , frenk c. s. , white s. d. m. , 1985 , apj , 292 , 371 de la torre s. , guzzo l. , 2012 , arxiv , arxiv:1202.5559 de lucia g. , blaizot j. , 2007 , mnras , 375 , 2 di porto c. , amendola l. , branchini e. , 2012 , mnras , 419 , 985 eisenstein d. j. , et al . , 2011 ,
aj , 142 , 72 feldman h. a. , kaiser n. , peacock j. a. , 1994 , apj , 426 , 23 fisher k. b. , davis m. , strauss m. a. , yahil a. , huchra j. , 1994 , mnras , 266 , 50 fisher k. b. , davis m. , strauss m. a. , yahil a. , huchra j. p. , 1994 ,
mnras , 267 , 927 fry j. n. , 1985 , phlb , 158 , 211 guzzo l. , et al . , 2008 , nature , 451 , 541 hamilton a. j. s. , 1992 , apj , 385 , l5 hamilton a. j. s. , 1993 , apj , 417 , 19 ha hamilton , a. j. s. 1998 , in d. hamilton , ed , the evolving universe .
kluwer , dordrecht , p. 185
hawken2012 hawken a. j. , abdalla f. b. , htsi g. , lahav o. , 2012 , mnras , 424 , 2 hawkins e. , et al .
, 2003 , mnras , 346 , 78 hewett p. c. , 1982 , mnras , 201 , 867 jenkins a. , frenk c. s. , white s. d. m. , colberg j. m. , cole s. , evrard a. e. , couchman h. m. p. , yoshida n. , 2001 , mnras , 321 , 372 jennings e. , baugh c. m. , pascoli s. , 2011 , mnras , 410 , 2081 kaiser n. , 1987 , mnras , 227 , 1 2011arxiv1105.1194k kwan j. , lewis g. f. , linder e. v. , 2012 , apj , 748 , 78 landy s. d. , szalay a. s. , 1993 , apj , 412 , 64 larson d. , et al . , 2011 , apjs , 192 , 16 laureijs r. , et al
. , 2011 , arxiv , arxiv:1110.3193 lightman a. p. , schechter p. l. , 1990 , apjs , 74 , 831 linder e. v. , 2008 , aph , 29 , 336 marulli2012 marulli f. , bianchi d. , branchini e. , guzzo l. , moscardini l. , angulo r. e. , 2012 , mnras , 426 , 2566 mcdonald p. , seljak u. , 2009 , jcap , 10 , 7 nesseris s. , perivolaropoulos l. , 2008 , phrvd , 77 , 023504 okumura t. , jing y. p. , 2011 ,
apj , 726 , 5 peacock j. a. , 1999 , cosmological physics , cambridge univ . press , cambridge peebles p. j. e. , 1980 , lssu.book , percival w. j. , white m. , 2009 , mnras , 393 , 297 percival w. j. , et al . , 2010 , mnras , 401 , 2148 perlmutter s. , et al .
, 1999 , apj , 517 , 565 riess a. g. , et al . , 1998 , aj , 116 , 1009 samushia l. , et al . , 2011 , mnras , 410 , 1993 2011arxiv1102.1014s samushia l. , percival w. j. , raccanelli a. , 2012 , mnras , 420 , 2102 snchez a. g. , baugh c. m. , percival w. j. , peacock j. a. , padilla n. d. , cole s. , frenk c. s. , norberg p. , 2006 , mnras , 366 , 189 saunders w. , rowan - robinson m. , lawrence a. , 1992 , mnras , 258 , 134 scoccimarro r. , 2004 , phrvd , 70 , 083007 seo h .-
j . , eisenstein d. j. , 2003 , apj , 598 , 720 sheth r. k. , mo h. j. , tormen g. , 2001 , mnras , 323 , 1 simpson f. , peacock j. a. , 2010 , phrvd , 81 , 043512 song y .- s .
, percival w. j. , 2009 , jcap , 10 , 4 springel v. , et al . , 2005 , natur , 435 , 629 2010arxiv1006.0699 t taruya a. , nishimichi t. , saito s. , 2010 , phrvd , 82 , 063522 tegmark m. , 1997 , phrvl , 79 , 3806 tinker j. l. , robertson b. e. , kravtsov a. v. , klypin a. , warren m. s. , yepes g. , gottlber s. , 2010 , apj , 724 , 878 tinker j. l. , weinberg d. h. , zheng z. , 2006 , mnras , 368 , 85 wang l. , steinhardt p. j. , 1998 , apj , 508 , 483 wang y. , 2008 , jcap , 5 , 21 wang y. , et al . , 2010 , mnras , 409 , 737 white m. , song y .- s . , percival w. j. , 2009 , mnras , 397 , 1348 zhangnew zhang , p. , liguori , m. , bean , r. , & dodelson , s. 2007 , physical review letters , 99 , 141302 zurek w. h. , quinn p. j. , salmon j. k. , warren m. s. , 1994 , apj , 431 , 559
to estimate @xmath4 , in section [ sec fitting procedure ] we defined a likelihood function comparing the measured correlation function @xmath1 and the corresponding parameterized models .
our likelihood is simply given by the standard @xmath82 expression @xmath83 where however the stochastic variable considered is not just the value of @xmath84 at each separation @xmath85 , but the expression @xmath86 \ ; , \ ] ] which has the desirable property of placing more weight on large , more linear scales .
this was first proposed by @xcite , who correspondingly adopt the following expression for the expectation value of the variance @xmath87 - \log [ 1 + \xi_{ij } - \delta(\xi_{ij})]\}}^2 \ ; .\ ] ] this simply maps onto the new variables @xmath88 , the interval including 68% of the distribution in the original variables @xmath89 , i.e. twice the standard deviation if this were gaussian distributed . strictly speaking , here an extra factor 1/2 would be formally required if one aims at defining the equivalent of a standard deviation , but this is in the end uneffective in the minimization and thus in finding the best - fitting parameters .
, as recovered from catalogues with varying density ( but same volume and bias ) , using the two different definitions of the variance of each data point of eqs .
[ eq scatter hawkins ] ( open blue squares ) and [ eq delta y ] ( open red circles ) .
the dashed line shows as reference the asymptotic common value of @xmath4 that both methods identically recover at high densities .
note how using eq .
[ eq delta y ] yields an unbiased estimate down to significantly smaller densities , whereas the estimator based on eq .
( [ eq scatter hawkins ] ) becomes rapidly more and more biased below @xmath90 .
the intrinsic scatter of the measurements , as usual obtained from the 27 sub - cubes of this specific catalogue , also follows a similar trend . , width=302 ] however , the weighting factors @xmath91 in the likelihood definition depend explicitly on @xmath89 , which may result in an improper weighting of the data when the correlation signal fluctuates near zero .
we have directly verified that when the estimate is noisy , it is preferable to use a smooth weighting scheme rather than one that is sensitive to local random oscillations of @xmath92 , which is more likely to yield biased estimates .
this supported our choice of adopting the usual sample - variance expression @xmath93 estimated over @xmath94 realizations of the survey .
this can be done using mock realizations @xcite , or , alternatively , through appropriate jack - knife or booststrap resamplings of the data . specifically , we find a significant advantage of the weighting scheme based on sample variance when dealing with low - density samples .
this is shown in figure [ fig bt vs n 20 ] , where @xmath4 is estimated on the catalogue with @xmath95 using the two likelihoods and gradually diluting the sample ( note that all computations in this section use the linear - exponential model , with @xmath96 directly measured in real - space ) . in order to understand the reasons behind this behaviour
, we have studied independently the various terms composing the likelihood .
we use one single sub - cube ( i.e. 1/27 of the total volume ) , from the catalogue with @xmath97 , and consider two extreme values of the mean density .
first , we consider the case of the highest density achievable by this halo catalogue , @xmath98 . in the upper panel of figure [ fig enzo perc100 pi9.75 refhawkins ] we plot a section of @xmath1 at constant @xmath99 , together with the model @xmath100 corresponding to the best - fit @xmath4 and @xmath101 parameters . in this density regime
the values of the recovered best - fit parameters are essentially independent of the form chosen for @xmath102 ( as shown by the coincident values of @xmath4 on the right side of figure [ fig bt vs n 20 ] ) .
the match of the model to the data is very good . in the central panel ,
we plot instead , for each bin @xmath103 along @xmath9 , the absolute value of the difference between model and observation , @xmath104 , together with the corresponding standard deviations in the two cases , which are virtually indistinguishable from each other .
finally , the lower panel shows the full values of the terms contributing to the @xmath82 sum , again showing the equivalence of the two choices in this density regime . .
top panel : cut - through @xmath1 at fixed @xmath105 ( broken line ) , and corresponding best fit model @xmath100 using the hawkins et al .
form for the scatter of each data point ( continuous line ) .
central panel : residual values @xmath106 between the data and model values ( light grey line ) and values for the scatter of each point , according to the two definitions of eqs .
[ eq delta y ] ( solid red line ) and [ eq scatter hawkins ] ( dashed blue line ) .
bottom panel : corresponding terms in the @xmath82 sum ( see eq .
( [ eq fit app ] ) ) .
the two definitions for the scatter , as expected , produce virtually identical values for the likelihood . , width=302 ] however ,
when we sparsely sample the catalogue , as to reach a mean density of @xmath107 @xmath108 ( leaving all other parameters unchanged ) , a very different behaviour emerges ( figure [ fig enzo perc3.08 pi9.75 refhawkins ] ) ( upper panel , second blue square from the left ) we show the same behaviour when averaged over 27 sub - samples . ] .
, but now in the low - density regime ( @xmath109 ) .
again , the model curve in the top panel corresponds to the best - fit parameters obtained using the hawkins et al .
form of the scatter of each measurements .
the fit is very unsatisfactory .
the bottom panel shows how the likelihood expression based instead on the standard deviation of @xmath110 as from eq .
( [ eq delta y ] ) rejects these parameter values , giving high @xmath82 values ( red solid curve ) .
note the different scale on the ordinate , with respect to previous figure .
, width=302 ] using the hawkins et al . definition for the variance yields a best - fit model that overestimates the data on almost all scales ( top panel ) , corresponding to unphysical values of @xmath111 and @xmath112 .
the central panel now shows how in this regime the two definitions of the scatter , ( which weigh the data - model difference ) , behave in a significantly different way , with the hawkins et al .
definition being much less stable than the one used here , and in general anti - correlated with the values of @xmath1 in the upper panel . in the lower panel ,
the dashed line shows how this anti - correlation smooths down the @xmath113 peaks resulting in erroneously low values for the @xmath82 that drive the fit to a wrong region of the parameter space . in the same panel ,
the solid line shows how the likelihood computed with our definition for these same parameters gives high @xmath82 values , thus correctly rejecting the model ( and @xmath114 ) we find @xmath115 .
consequently , @xmath116 is not well defined ( figure [ fig enzo perc3.08 pi9.75 refhawkins ] , central panel ) resulting in a zero weight for the corresponding @xmath82 summand ( lower panel ) . ] .
in a real survey , the direct measurement of @xmath96 is not possible . a way around this obstacle is to project @xmath1 along the line of sight , i.e. along the direction affected by redshift distortions .
we hence define the projected correlation function as @xmath117 inverting the integral we recover @xmath96 .
more precisely , following @xcite , we have @xmath118 where @xmath10 is the usual mathematical constant , not to be confused with the line - of - sight separation @xmath10 in eq .
( [ eq projected ] ) . a more extended investigation of the effects arising when using the deprojected @xmath96 instead of that directly measured ( hereafter @xmath119 and @xmath120 respectively )
is carried out in @xcite . here
we limit the discussion to the impact of the deprojection technique on the estimate of @xmath4 , as a function of the mass ( i.e. the bias ) of the adopted tracers , focussing on the systematic effects ( figure [ fig beta dep ] ) .
obtained when the real - space correlation function @xmath96 is directly measured from the simulation ( triangles ) or deprojected as in real surveys ( rhombs and inverted triangles ) .
the latter correspond to two different integration limits @xmath121 in the projection .
the two lower panels give ths systematic and statistical error as in figure [ fig lin vs exp].,width=302 ] one possible source of systematic error in performing the de - projection is the necessity of defining a finite integration limit @xmath121 in eq .
( [ eq deprojected ] ) . in figure [ fig beta dep ]
two different choices of @xmath121 are considered .
we notice that these choices ( purple inverted triangles and yellow rhombs ) result in different slopes of @xmath4 as a function of bias , which differ from the slope obtained using @xmath120 ( green triangles ) .
this is plausibly due to the fact that using a limiting @xmath121 we are underestimating the integral ( consider that @xmath122 for @xmath123 ) .
this effect grows when the bias increases , because of the corresponding growth of @xmath92 which leads to a larger `` loss of power '' in @xmath124 .
however , we can not use arbitrarily large values of @xmath121 because the statistical error increases for larger @xmath121 ( see lowest panel of figure [ fig beta dep ] ) .
this may be due to the increase of the shot noise at large separations .
similarly , the drop of correlation signal at small separations due to the finite size of the dark matter halos produces an impact on @xmath4 which grows with bias .
finally , as suggested previously @xcite and discussed extensively in @xcite , figure [ fig beta dep ] shows how using @xmath119 in modelling rsd , produces a statistical error about twice as large as that obtained using @xmath120 ( lower panel ) . | we investigate the impact of statistical and systematic errors on measurements of linear redshift - space distortions ( rsd ) in future cosmological surveys by analysing large catalogues of dark - matter halos from the basicc simulation .
these allow us to estimate the dependence of errors on typical survey properties , as volume , galaxy density and mass ( i.e. bias factor ) of the adopted tracer .
we find that measures of the specific growth rate @xmath0 using the hamilton / kaiser harmonic expansion of the redshift - space correlation function @xmath1 on scales larger than @xmath2 are typically under - estimated by up to 10% for galaxy sized halos .
this is significantly larger than the corresponding statistical errors , which amount to a few percent , indicating the importance of non - linear improvements to the kaiser model , to obtain accurate measurements of the growth rate .
the systematic error shows a diminishing trend with increasing bias value ( i.e. mass ) of the halos considered .
we compare the amplitude and trends of statistical errors as a function of survey parameters to predictions obtained with the fisher information matrix technique .
this is what is usually adopted to produce rsd forecasts , based on the fkp prescription for the errors on the power spectrum .
we show that this produces parameter errors fairly similar to the standard deviations from the halo catalogues , provided it is applied to strictly linear scales in fourier space ( @xmath3 ) . finally , we combine our measurements to define and calibrate an accurate scaling formula for the relative error on @xmath4 as a function of the same parameters , which closely matches the simulation results in all explored regimes .
this provides a handy and plausibly more realistic alternative to the fisher matrix approach , to quickly and accurately predict statistical errors on rsd expected from future surveys .
[ firstpage ] cosmological parameters dark energy large - scale structure of the universe . |
one unavoidable ingredient in the process of formation of intermediate / high - mass stars is that they are usually found associated with clusters of lower - mass stars ( e.g. , lada & lada 2003 ; rivilla et al .
2013 ) . however , the process yielding the formation of these clusters is not clear , as it remains ambiguous how a cloud core fragments to finally form a cluster .
the observed star formation efficiency is by far much smaller than the predicted if gravity dominated the dynamics of the collapse ( see e.g. , kruijssen 2013 for a review ) , and the number of subcondensations typically embedded in massive dense cores is again smaller by one order of magnitude than the predicted from a pure jeans fragmentation analysis ( e.g. , hennemann et al . 2009 ; zhang et al .
2009 , 2011 ; bontemps et al . 2010 ;
longmore et al . 2011 ;
pillai et al . 2011 ; wang et al . 2011 ; lee et al . 2013 ) .
thus , a broad observational evidence indicates that additional physical ingredients competing with gravity must play an important role in the fragmentation process , such as angular momentum , stellar feedback , magnetic fields and turbulence .
numerical simulations are becoming crucial to establish the relative importance of each of these ingredients in the fragmentation process .
simulations of fragmentation of rotating cores with different equations of state show that cores with some angular momentum easily fragment into multiple objects ( e.g. , boss & bodenheimer 1979 ; bate & burkert 1997 ) and thus need of additional mechanisms to decrease the fragmentation down to the observed levels ( see goodwin et al .
2007 and references therein ) . a crucial ingredient in this sense
is the presence of a magnetic field .
studies on the effects of magnetic fields in the evolution of a collapsing and rotating core indicate that a strong magnetic field can suppress fragmentation as it constitutes an additional form of support and also via the so - called ` magnetic braking ' ( e.g. , hosking & whitworth 2004 ; mellon & li 2008 ; hennebelle & teyssier 2008 ) , which is very effective in removing a significant part of the initial angular momentum through torsional alfvn waves .
similarly , radiation magnetohydrodynamic simulations of turbulent massive dense cores show that the combined effects of radiative feedback ( e.g. , krumholz et al . 2007 ; peters et al .
2011 ) and magnetic field can sum to efficiently suppress fragmentation ( e.g. , hennebelle et al . 2011 , commeron et al .
2011 ; myers et al .
2013 ) , yielding a better match with the observations ( e.g. , palau et al .
in addition , submillimeter observations show that the magnetic field is important in the dynamical evolution and fragmentation of massive dense cores ( e.g. , girart et al .
2009 , 2013 ; hezareh et al .
another ingredient which is probably playing a role in the fragmentation process is supersonic turbulence .
theoretical studies show that turbulence generates density structures due to supersonic shocks that compress and fragment the gas efficiently , especially for large mach numbers ( e.g. , padoan et al .
2001 , schmeja & klessen 2004 ; vzquez - semadeni et al .
more recently it has been shown that the turbulence mode , solenoidal or compressive ( e.g. , federrath et al .
2008 ) , affects differently the cloud structure and fragmentation .
solenoidal turbulence corresponds to a null divergence of the driving force ( divergence - free ) and arises from galactic rotation and magneto - rotational instabilities , and seems to be at work in the rosette and polaris flare clouds ( e.g. , hily - blant et al .
2008 ; federrath et al .
compressive turbulence corresponds to a null curl of the driving force ( curl - free ) and can be produced by expanding supernova shells or hii regions ( e.g. , federrath et al .
2010 ) , as seen in the pipe nebula and the orionb cloud ( peretto et al . 2012 ; schneider et al .
numerical simulations comparing both types of turbulence indicate that compressive turbulence promotes fragmentation and generates high density contrast structures , while solenoidal turbulence slows fragmentation down and keeps a smooth density structure in the cloud ( e.g. , federrath et al .
2010 ; girichidis et al .
2011 ) .
in addition to these physical agents opposing to gravity , the outcome of numerical simulations also depends on the geometry and physical properties of the collapsing core , in particular the density profile .
hydrodynamical simulations show that highly concentrated cores ( e.g. , @xmath1 ) lead to the formation of one single massive star , while flat density profiles ( @xmath2 or flatter ) produce an important number of low - mass stars ( e.g. , myhill & kaula 1992 ; burkert et al .
1997 ; girichidis et al .
this trend appears to be seen in high resolution observations of irdcg28.34 + 0.06 , where three massive dense cores with different density profiles show varying degrees of fragmentation ( zhang et al .
2009 ; wang et al .
finally , the substructure of molecular gas within a dense core is decreasing with time as the core turns gas into stars , as shown for example by smith et al .
( 2009 ) .
from an observational point of view , most of the studies focusing on fragmentation of massive dense cores deal with single regions and/or do not reach high enough spatial scales to study individual fragments in the core ( @xmath32000 au ; e.g. , longmore et al . 2011 ; pillai et al . 2011 ; miettinen 2012 ; beuther et al . 2013 ; kainulainen et al . 2013 ; takahashi et al . 2013 ) .
recently , bontemps et al .
( 2010 ) study a sample of six massive dense cores in cygnus x and conducted a consistent analysis for all the regions , finding cores with a variety of fragmentation levels .
a subsequent study by palau et al .
( 2013 ) focused on a list of 18 massive dense cores ( including bontemps et al .
2010 sources ) , and observed with similar spatial resolution ( @xmath3 au ) , sensitivity ( down to @xmath4 @xmath5 ) , and field of view ( @xmath6 pc of diameter ) , again reveals different fragmentation levels , with about @xmath7% of the regions showing no fragmentation at all .
palau et al .
( 2013 ) study possible correlations between the fragmentation level and properties of the cores such as mass , average density , evolutionary indicators , rotational - to - gravitational energy and turbulent velocity dispersion ( although these last two properties should be studied in a larger sample ) , and find no clear trend of fragmentation level with any of these core properties . in that work
, two main properties of the cores remained to be studied in relation to fragmentation , mainly the magnetic field and the density structure .
while the magnetic field can not currently be assessed in this sample in a uniform way , the density structure has been studied in roughly half of the sample ( e.g. , beuther et al .
2002a ; mueller et al .
2002 ; crimier et al . 2010 ; van der tak et al .
2013 ) , but the methodology used is different for each work , casting some doubt on the validity of any comparison . here
we present a uniform study of the density structure of each core in the sample of palau et al .
( 2013 ) , aiming at accurately assessing its impact on the fragmentation level . in section 2 we present the sample of cores and the data used to extract the submillimeter intensity profiles and the spectral energy distributions ( seds ) , and we also present the envelope model used to fit these data simultaneously for each core ; in section 3
, we present the results of our fits ; in section 4 , we discuss the possible trend between fragmentation level and density structure .
finally , in section 5 we summarize our main conclusions .
we aim at studying the density structure of the 18 massive dense cores presented in palau et al .
( 2013 ) plus dr21-oh , a massive dense core for which fragmentation has been recently studied at similar mass sensitivities and spatial resolution by girart et al .
( 2013 ) . among the total number of 19 massive dense cores , 16 of them are identified as submillimeter sources in the catalog of di francesco et al .
( 2008 ) and thus have james clerk maxwell telescope ( jcmt ) 850 and 450 @xmath0 m data available ( see di francesco et al .
2008 for a detailed description on the observations ) .
for the three cores not included in di francesco et al . catalog , iras22198 + 6336 , cygx - n12 and cygx - n63 ( following labeling of bontemps et al .
2010 ) , we used the submillimeter data of jenness et al .
( 1995 ) , who observed also the 450 and 850 @xmath0 m emission with the jcmt , and the iram30 m 1.2 mm data from motte et al .
the beam ( main + error ) of the jcmt data was taken from di francesco et al .
( 2008 ) , and the beam ( main + error ) of the iram30 m telescope at 1.2 mm was taken from greve et al .
three regions for which the 450 @xmath0 m data from di francesco et al .
( 2008 ) were not usable
were downloaded from the jcmt archive ( ic1396n : m98bu24 ; afgl5142 : m96bc57 ; w3irs5 : m02ac32 ) . the single - dish images at 450 and 850 @xmath0 m ( 1.2 mm ) for the 19 cores , together with the interferometric images showing the fragmentation in each core , are shown in the appendix ( figs .
[ fobs1]13 ) .
in order to characterize the spatial structure of the massive dense cores , we computed the circularly averaged radial intensity profile , in rings of @xmath8 width , as a function of the projected distance from the core center at both 850 and 450 @xmath0 m when possible , and applied the model described in fig .
[ fgeom ] and section 2.2 .
the intensity profiles at 850 and 450 @xmath0 m are shown in figs .
[ ffit1][ffit4 ] .
as can be seen in the figures , the emission is partially resolved for all the cores , revealing the existence of an envelope of decreasing intensity with radius , even at small distances ( @xmath9 ) from the core center . in some cases ,
the radial profile was calculated in particular angular ` sectors ' centered on the core peak , to avoid strong negative emission in the maps ( ngc7129-firs2 at 450 @xmath0 m ) , or contamination from a nearby strong core ( w3irs5 ) or from the dr21 ridge ( cygx - n48 , cygx - n53 , dr21-oh ) . for these cases we did not use the values of the flux density at 850 and 450 @xmath0
m listed in di francesco s catalog , but remeasured the flux densities to assure that we were not including emission from the nearby cores / ridge . for the case of w3irs5 ,
the sector was taken in the north - south direction also to avoid possible contamination of free - free emission from strong centimeter sources% at 1.1 mm , and only @xmath10% at 800 @xmath0 m , the centimeter sources a and b reported by tieftrunk et al .
( 1997 ) are spatially overlapping the core emission and we prefer to avoid them . ] .
for the case of cygx - n48 , we extracted the intensity profile avoiding the contribution from the north , where the strong dr21-oh core lies ; for dr21-oh and cygx - n53 we extracted the profile in a sector with no restrictions in the east - west direction , but allowing it to extend @xmath11 in the north - south direction , which corresponds to the approximate width of the dr21 filament in the east - west direction .
the sed of each massive dense core was built from centimeter wavelengths up to 5060 @xmath0 m , as this is approximately the wavelength where envelope and disk contributions start to be comparable ( e.g. , crimier et al .
2010 ; commeron et al .
2012 ) , and here we only model the envelope contribution .
to build the seds we used the data compiled in palau et al .
( 2013 ) , which included flux densities at different frequencies from iras , jcmt / scuba , akari and iram30 m data when available .
in addition , we measured the flux densities at 70 @xmath0 m of cores with available spitzer / mips2 data .
flux densities obtained in heavily saturated spitzer / mips2 data were considered as lower limits .
we also considered as lower limits the millimeter / submillimeter flux densities measured with an interferometer , as part of the core emission is filtered out and we did not take this into account in our model .
for omc-1s we recompiled the sed because the sed in palau et al .
( 2013 ) was built for the source 136@xmath12 ( about @xmath13 north of the submillimeter peak ) , while now we have considered the fluxes associated with the submillimeter peak , as we are modeling the submillimeter profiles .
in addition , we used the fir / submillimeter / centimeter fluxes from the literature .
the apertures where fluxes were assessed were figured out from the literature or adopted as reported in the used catalogs and references ( see above ) , and for peak intensities we took the beam as aperture radius .
c + we modeled the thermal emission of the cores at millimeter and submillimeter wavelengths with a spherically symmetric envelope of gas and dust , with density and temperature decreasing with radius .
the density and temperature were assumed to be power - laws of radius , with indices @xmath14 and @xmath15 , @xmath16 and @xmath17 . for the dust opacity law we assumed that the dust opacity is a power - law of frequency with index @xmath18 , @xmath19 , where @xmath20 is an arbitrary reference frequency .
the values used for the dust opacity were those of ossenkopf & henning ( 1994 ) , for thin ice mantles at a gas density of @xmath21 @xmath22 , i.e. @xmath23 @xmath24 g@xmath25 for @xmath26 ghz , being @xmath18 a free parameter of the model .
cc + cc + cc + cc + in the case of optically - thin dust emission , in the rayleigh - jeans approximation , and for an infinite radius envelope , the intensity as a function of the projected distance to the source center , @xmath27 , can be derived analytically ( see , for instance , beltrn et al .
2002a ) , and is a power - law of the projected distance to the envelope center @xmath28 , @xmath29 , with index @xmath30 . in our model , however , none of the above approximations ( optically thin , rayleigh - jeans ) were made , and there is no analytical expression for the intensity profile .
in general , the contribution to the radiated intensity from an element of the envelope of geometrical depth @xmath31 , at a radius @xmath32 from the center ( see fig . [ fgeom ] ) is given by @xmath33\,e^{-\tau_\nu(b , z ) } \rho(r)\,\kappa_\nu\,dz,\ ] ] where @xmath34 with @xmath35 , is the optical depth from the observer to the element of the envelope at a projected distance @xmath28 and position @xmath36 along the line of sight ( see fig .
[ fgeom ] ) , and @xmath37 is the planck function in units of temperature , @xmath38 then , the resulting intensity as a function of the projected distance to the source , @xmath27 , has to be evaluated as @xmath39 where @xmath40 and @xmath41 is the maximum radius where the model is calculated .
the computational burden is low , since both integrals , those of eqs.[eqtau ] and [ eqib ] , can be evaluated simultaneously in a single run .
regarding the temperature distribution of the envelope , chandler et al .
( 1998 ) derive from the expressions of scoville et al .
( 1976 ) that , for the radiative heating of a dust cloud by a central luminous source , the temperature power - law index @xmath15 is a function of the dust emissivity index @xmath18 , @xmath42 .
however , for the outer part of the envelope the external heating can be important , and a constant temperature of 10 k was adopted for radii larger than the radius where the temperature distribution of index @xmath43 drops to 10 k. regarding the density distribution , we adopted , as maximum radius of the envelope @xmath41 , the radius for which the envelope density dropped to a value comparable to that of the ambient gas of the intercore medium , taken to be 5000 @xmath22 . in summary ,
the four free parameters of the model are the dust emissivity index , @xmath18 ; the envelope temperature at the reference radius @xmath44 , @xmath45 ; the envelope density at the reference radius @xmath44 , @xmath46 ; and the density power - law index , @xmath14 .
the reference radius @xmath44 is arbitrary , and has been taken as @xmath47 au . with this model
we fitted the observed radial intensity profiles at 450 and 850 @xmath0 m or 1.2 mm , and the observed sed from centimeter wavelengths up to 60 @xmath0 m simultaneously .
the uncertainty adopted for the intensity profiles was @xmath48% at 850 @xmath0 m and 1.2 mm , and @xmath49% at 450 @xmath0 m ( di francesco et al .
the uncertainty of the flux densities used to build the sed was adopted to be @xmath49% . in order to compare the model with the observed intensity profile at each wavelength , we computed the intensity map from the model , and we convolved it with the beam , given as the sum of two circular gaussians ( the main and error beams ) .
finally , the circularly averaged profile was computed from the convolved map .
we also simulated the effect of the finite chop throw used in the observations to cancel the sky emission .
thus , for each map position we estimated the off - source intensity as the average intensity at a chop - throw distance of @xmath50 from the map ( on - source ) position , and we computed the difference between the on- and off - source intensities . from this , we computed the radial profile corrected for chopping . in order to compare the model with the observed sed , for each data point with wavelength higher than 60 @xmath0 m we computed the model flux density by integration of the model intensity profile inside the reported aperture used for each measurement .
in addition , for points with frequency below 1000 ghz , we added the free - free emission of an unresolved source with a flux density extrapolated from the data points at centimeter wavelengths .
the fitting procedure was the sampling of the four - dimensional parameter space , using the same procedure as that described in estalella et al .
( 2012 ) and snchez - monge et al .
for every set of parameters , we computed the residual @xmath51 , @xmath52 ^ 2,\ ] ] where the sum encompasses the points of the intensity profiles at 850 and 450 @xmath0 m , and the points of the sed fitted by the model .
the parameter space was searched for the minimum value of @xmath51 . for each source ,
the initial search ranges for the four parameters was @xmath53 , @xmath54 k , @xmath55 g @xmath22 , and @xmath56 .
the search range was reduced a factor of 0.8 around the best fit value of the parameters found for each loop , consisting in 2000 samples of the parameter space .
the final best fit values were taken after 10 loops .
once the best fit parameters were found , their uncertainties were estimated through the increase in @xmath51 ( see estalella et al .
2012 , snchez - monge et al .
we assessed the quality of the fits through the value of the reduced chi - square , @xmath57 in general for all the sources we found fits with @xmath58 , or even less , indicating that the errors of the data were slightly overestimated . in fig .
14 we show as example , the parameter space around the best fit solution for the source ic1396n .
[ cols="<,^,>,^,^,^,^,^,^,^,^ " , ] parameters inferred ( not fitted ) from the modeling .
@xmath15 is the temperature power - law index , and @xmath59 is the radius of the core where the temperature has dropped down to @xmath60 k ; @xmath41 is the radius at the assumed ` ambient ' density of 5000 @xmath22 ; @xmath61 is the mass computed analytically from the model integrating until the radius where the density profile could be measured for each source . @xmath62 and
@xmath63 are the surface density and density inside a region of 0.05 pc of radius .
@xmath64 is the rotational - to - gravitational energy ratio as in palau et al .
( 2013 ) and adding the new measured values from nh@xmath65(1,1 ) for omc-1s ( wiseman & ho 1998 ) and dr21-oh ( mangum et al . 1992 ) .
@xmath66 is the non - thermal velocity dispersion for a quiescent nearby core ( thus a first approach to the initial velocity dispersion ) , as in palau et al .
( 2013 ) and adding a new value for cygx - n48 and dr21-oh ( mangum et al .
@xmath66 was estimated assuming a temperature of 20 k. @xmath67 is the non - thermal velocity dispersion for the studied core ( thus including outflow feedback from the nascent protostars ) from data in the literature ( see section [ sdaffectingdens ] for references ) .
@xmath67 was estimated using the temperature at radius 0.05 pc as inferred from the modeled temperature profile ( tables 1 and 2 ) . @xmath68 and @xmath69
( with @xmath70 being the mass within a radius of 0.05 pc ) are assessed using the non - thermal velocity dispersion including feedback from outflows . `
mode ' refers to tentative turbulence mode according to works in the literature : ` c ' indicates that a detailed study has been done comparing the properties of the region with simulations or studying the probability density function ; ` c ? ' indicates that the only hints of possible compressive turbulence are that the region lies at the border of an expanding bubble / hii region .
references about turbulence mode : 1 : beltrn et al .
( 2002b ) ; 2 ; schneider et al .
( 2013 ) ; 3 : sakai et al .
( 2013 ) ; 4 : liu et al .
( 2012 ) ; 5 : lpez - sepulcre et al .
( 2013 ) ; 6 : hunter et al .
( 1995 ) ; 7 : leurini et al .
( 2007 ) ; 8 : schneider et al .
( 2011 ) ; 9 : schneider et al .
( 2006 ) .
we compared our inferred density power - law indices to the results of previous works where a similar methodology was applied for some of the cores of our sample .
in the two last columns of table [ tfit1 ] we list @xmath71 , the density power - law index from the literature , together with the corresponding references .
crimier et al .
( 2010 ) , van der tak et al .
( 2000 , 2013 ) and mueller et al .
( 2002 ) fit one single power - law to the radial intensity profiles , being comparable to the work presented here , while beuther et al .
( 2002 ) used two power - laws with a break at @xmath72 . within the regions fitted by the same authors
the tendency of flat and concentrated profiles is consistent with the results found in the present work ( for example , ic1396n is flatter than ngc7129-firs2 and cb3-mm as in crimier et al . 2010 ; or i20126 is more concentrated than i22134 and i05358ne as in beuther et al .
however , the absolute value of the density power - law indices can not be directly compared between different works because of the different methods applied , which yield differences of 0.20.4 in the density power - law index .
the most important difference between previous works and this work is the way of determining the temperature profile . while mueller et al .
( 2002 ) , crimier et al .
( 2010 ) , and van der tak et al .
( 2013 ) calculate the temperature profile self - consistently through radiative transfer codes , we assume that the temperature profile is related to the dust emissivity index ( section 2.2 ) and leave the temperature at a given radius as a free parameter .
this is also different to beuther et al .
( 2002 ) work , who assumed a fixed temperature power - law index of 0.4 .
our method finds in general flatter temperature profiles compared to crimier et al .
( 2010 ) and van der tak et al .
( 2013 ) , yielding higher density power - law indices . on the other hand ,
our method is perfectly consistent with the results of mueller et al .
( 2002 ) for afgl2591 .
therefore , once the different methodologies are taken into account , our methodology yields consistent results with previous works , and allows to fairly compare the fitted parameters from one region to the other .
c + before comparing the inferred density structure for each massive dense core ( estimated through the density power - law index @xmath14 ) , to the fragmentation level ( estimated through the @xmath73 parameter , see table [ tfit1 ] and palau et al .
2013 ) , we discuss here the main uncertainties in the assessment of these properties and possible biases .
first of all we stress that the determination of both @xmath14 and @xmath73 has been carried out applying the same methodology and criteria to all the cores and thus , although the absolute values might be subject to discussion , the relative value between the regions is meaningful and valid for a comparison .
in particular , @xmath73 is a lower limit to the total number of fragments because of flux missed by the interferometers , and because we are missing objects emitting at wavelengths different to millimeter .
however , the sample presented in palau et al .
( 2013 ) was selected applying strict criteria for the filtered emission and @xmath74-coverage ( see column 8 of table 3 of palau et al .
2013 ) , constraining this study to the most compact fragments ( @xmath75 au ) .
since larger non - centrally peaked structures might be transient , this is not likely affecting our estimate of the fragmentation level .
in addition , the number of infrared sources were compared to @xmath73 and were found to be correlated , indicating that @xmath73 is systematically underestimating the ` real ' fragmentation level of the core by only a factor of @xmath76 .
finally , palau et al .
( 2013 ) also used radiation mhd simulations to show that @xmath73 is sensitive to the original fragmentation of the core .
altogether suggests that @xmath73 is already a good indicator of the fragmentation level if used to compare the different regions and not as an absolute value .
a possible bias could come from the different distances in the sample .
however , for 16 out of 19 cores distances range only from 0.8 to 2.5 kpc , and no trend of @xmath73 , @xmath14 , or any of the fitted parameters vs distance is found in our data ( table 1 ) , indicating that this is not affecting our conclusions .
another bias could arise from different evolutionary stages of the cores of our sample , as simulations show that the degree of substructure in a cloud decreases with age ( e.g. , smith et al .
our sample is not likely dramatically affected by this because the targets were selected to be in similar evolutionary stages : all targets harbor an intermediate / high - mass protostar detected in the far - infrared , are associated with a strong and compact millimeter core , are associated with molecular outflows , and have not developed an ( ultra-)compact hii region yet .
the small differences in evolutionary stage of the cores of this sample were also studied in palau et al .
( 2013 ) by estimating evolutionary indicators , and no relation between @xmath73 and these indicators was found ( see e.g. , fig .
8b of palau et al . 2013 ) .
therefore , no biases in distance , evolutionary stage or missing fragments in @xmath73 should be strongly affecting the results presented in this work .
in this section we study the relation between the core parameters inferred from the model and the fragmentation level , quantified through the number of millimeter sources within a field of view of 0.1 pc ( in diameter ) , @xmath73 , as in palau et al .
we find no trend of @xmath18 , @xmath45 and total mass of the core with the fragmentation level ( tables [ tfit1 ] and [ tfit2 ] ) .
conversely , fig .
[ fcorr ] presents the relation between the density power - law index @xmath14 , the central density ( at 1000 au ) @xmath46 , and the density within a radius of 0.05 pc , @xmath63 , with the fragmentation level .
[ fcorr]-a reveals a possible weak trend of @xmath73 with the density power - law index , with low fragmentation levels found mainly in cores with steeper density profiles . concerning @xmath73 vs @xmath46 ,
our sample splits up into two groups , cores with high fragmentation and high @xmath46 , and cores with lower fragmentation and low @xmath46 , suggesting that high central densities ( in opposition to average density , as assessed in palau et al .
2013 ) , could be related to the fragmentation level .
finally , there is a clear trend of high fragmentation levels in cores with high @xmath63 ( fig .
[ fcorr]-c )
. a closer inspection of the data using statistical tests , like the spearman s @xmath77 and kendall s @xmath78 rank correlation coefficients , indicates that @xmath79 and @xmath14 show a faint anti - correlation ( @xmath80 , @xmath81 , pearson s linear correlation coefficient @xmath82 ) .
as for @xmath63 , it is strongly correlated with @xmath73 ( @xmath83 , @xmath84 ) , and the two parameters are well described by a linear relationship ( @xmath85 ) .
these relations indicate that the high @xmath63 likely arise in cores with rather flat @xmath14 and also high @xmath46 . a trend of high fragmentation level for flat density profiles was already suggested theoretically by myhill & kaula ( 1992 ) . more recently , girichidis et al . ( 2011 )
show , by studying the evolution of collapsing non - magnetized turbulent massive cores for different initial density profiles , that cores with a flat density profile produce a high number of protostars , while concentrated density profiles rather yield one massive protostar in the core center .
this can be understood in the framework of jeans fragmentation , as the larger the density , the smaller the jeans mass ( for a given temperature ) and the larger the number of expected fragments @xmath86 . in fig .
[ fcorr]-c we overplot the expected number of fragments estimated as @xmath87 , with @xmath88^{3/2}[n_\mathrm{h_2}/10 ^ 5\mathrm{cm}^{-3}]^{-1/2}$ ] @xmath5 , for which we used @xmath89 k , and @xmath70 being the mass within a radius of 0.05 pc ( calculated directly from the density law inferred from our model ) . the core formation efficiency ( cfe ) was adopted to be @xmath90% , the average of the cfes measured in palau et al .
thus , the trend of higher fragmentation for higher densities within a radius of 0.05 pc is consistent with what is expected for jeans fragmentation .
the relation found in this work between fragmentation level ( which could be regarded as a proxy to cluster richness ) and gas density is reminiscent of the relation found on larger spatial scales between cluster mass or protostellar density and gas surface density ( e.g. , gonzlez - lpezlira et al . 2012 ; lombardi et al . 2013 ;
pflamm - altenburg et al .
2013 ) and further studies should be conducted in this direction . with the aim of shedding light on which physical conditions promote fragmentation in massive dense cores , in this section we study if there is any relation between different physical processes and the density structure ( density profile and central density ) of massive dense cores .
[ [ time - evolution ] ] time evolution + + + + + + + + + + + + + + it has been shown both from observations and numerical simulations that structures within collapsing molecular clouds become more centrally peaked with time ( e.g. , smith et al .
2009 ; giannetti et al . 2013 ) .
therefore , the cores for which we measure higher values in the density power - law index @xmath14 could just correspond to the most evolved ones . however , we searched for any relation between @xmath14 and the ratio of the bolometric luminosity to the single - dish mass parameter ( an evolutionary indicator used in palau et al .
2013 ) , and found no trend in our data ( fig .
[ fcorr - p]-a ) . this lack of correlation of @xmath14 with time probably results from the fact that our sample consists of cores at very similar evolutionary stages ( section 4.2 ) , not properly sampling broad enough timescales , and suggests that other processes in addition to time evolution are probably also playing a role in determining the density structure of a dense core .
[ [ rotational - to - gravitational - energy - beta_mathrmrot ] ] rotational - to - gravitational energy ( @xmath64 ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the rotation of a massive dense core is expected to promote flattening of its density profile . in order to test this in our sample
, we completed the values of rotational - to - gravitational energy ratios already estimated in palau et al .
( 2013 ) with the value for the core added in this work ( dr21-oh ) , observed by girart et al .
( 2013 ) with similar conditions of sensitivity and spatial resolution as the sample of palau et al .
here we estimated @xmath64 for dr21-oh following exactly the same method as in palau et al .
( 2013 ) , i.e. , using nh@xmath65(1,1 ) data ( mangum et al .
1992 ) , and obtained a value of 0.24 .
we also estimated @xmath64 for omc-1s from the nh@xmath65(1,1 ) data by wiseman & ho ( 1998 ) .
it is worth mentioning that these measurements of @xmath64 must be regarded with caution as the velocity gradients observed might not be necessarily associated with rotation , but could arise from multiple velocity components or infall motions ( e.g. , lee et al .
2013 ) . keeping this in mind , in fig .
[ fcorr - p]-b we present the relation between @xmath64 and @xmath14 , showing a weak inverse trend .
this suggests that the initial angular momentum alone might not be the only agent shaping the density profile of a core .
we also present an updated version of the plot of @xmath73 vs @xmath64 ( after palau et al .
2013 ) including these new measurements , and the updated plot , shown in fig .
[ fnmmbetarot ] , indicates a possible trend of @xmath64 with fragmentation level , which deserves further studies .
c + c + [ [ non - thermal - velocity - dispersion - sigma_mathrmnon - th ] ] non - thermal velocity dispersion ( @xmath91 ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + palau et al .
( 2013 ) estimate the non - thermal velocity dispersion from the linewidth measured in nh@xmath65(1,1 ) of a quiescent massive dense core in the immediate surroundings of the core where fragmentation was assessed . this can be regarded as a first approach to the initial turbulent dispersion .
we completed the values of non - thermal ` initial ' velocity dispersions by adding two new measurements ( cygx - n48 and dr21-oh by mangum et al .
1992 , table [ tfit2 ] ) .
however , the non - thermal velocity dispersions cover a small range of values , from 0.33 to 0.67 km s@xmath25 , and did not show any relation with the fragmentation level .
we additionally study the relation of non - thermal velocity dispersion of the fragmenting cores ( thus including feedback from outflows ) , covering a broader range from 0.41 to 2.16 km s@xmath25 , with the density profile ( table [ tfit2 ] , where we used the values from nh@xmath65(1,1 ) observed with the vla : torrelles et al .
1989 ; zhou et al . 1990 ; mangum et al . 1992 ; tieftrunk et al . 1998 ; wiseman & ho 1998 ; gmez et al . 2003 ; snchez - monge et al .
2013 ) . since no clear trend is found ( see fig .
[ fcorr - p]-c and table [ tfit2 ] ) , this indicates that probably the density profile and the turbulent velocity dispersion are independent quantities . finally , we estimated the virial masses and virial parameter ( @xmath92 ) using the non - thermal velocity dispersions of the fragmenting cores ( i.e. , including feedback ) , and found no relation of these quantities with the density power - law index or the fragmentation level .
[ [ turbulence - mode - compressive - vs - solenoidal ] ] turbulence mode : compressive vs solenoidal + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in order to study if different turbulence modes could be affecting the density structure of massive dense cores , we searched the literature for any evidence of our massive dense cores being affected by expanding shells / bubbles created by nearby o - type stars , and our findings and references are listed in the two last columns of table [ tfit2 ] .
regions with hints in the literature of being associated with expanding shells or hii regions , such as bright - rimmed clouds , etc . , or located at the meeting point of converging flows , are classified as ` possible compressive ' , while regions for which the probability density function has been found to be consistent with compressive turbulence are classified as ` confirmed compressive ' . both ` possible ' and ` confirmed ' compressive - turbulent cores ( see table [ tfit2 ] to distinguish among both types ) are assigned a ` red ' color in the figures , while the ` blue ' color corresponds to cores where the turbulence mode can not be assessed with the current information ( thus a solenoidal mode can not be ruled out for ` blue ' cores ) .
[ fcorr ] shows that ` compressive - turbulent ' cores tend to have large @xmath73 , and cover a wide range of density power - law indices ( from 1.5 to 2 ) .
in addition , cores with @xmath93g@xmath22 and @xmath63 @xmath94@xmath22 fall all within the ` compressive - turbulent ' group ( see fig .
[ fcorr]-b , c ) .
this could be consistent with compressive turbulence because this mode yields higher density contrast structures than the solenoidal mode ( e.g. , federrath et al .
2008 , 2010 ) .
recently , tremblin et al . ( 2013 ) find that expanding hii regions or bubbles tend to induce steep density profiles in the cores at the bubble edges .
thus , the steep density profiles obtained in dr21-oh , afgl5142 , cb3 ( @xmath95 ) could have been produced by the expanding ionization fronts from the massive stars in these regions ( hunter et al .
1995 ; schneider et al . 2011 ; sakai et al .
2013 ) . a more detailed and uniform study of the turbulence mode in each core would be required to definitely assess the role of the different turbulence modes in setting the density structure of massive dense cores .
c + [ [ magnetic - field ] ] magnetic field + + + + + + + + + + + + + + magnetic field could affect the density structure of a core because it helps removing angular momentum yielding rather steep density profiles .
we tested this by using the magnetohydrodynamic simulations from commeron et al .
( 2011 ) of strongly ( @xmath96 , with @xmath0 being the mass - to - flux over the critical mass - to - flux ratio ) and weakly ( @xmath97 ) magnetized cores .
these simulations were found to reproduce remarkably well our data ( palau et al .
it is important to note that in these simulations the initial density profile is the same and rather flat for both @xmath96 and @xmath97 cases .
we extracted the density profiles of commeron et al .
( 2011 ) simulations for @xmath96 and @xmath97 cases , by making an average of the density on concentric spherical shells weighted by the mass for each cell in the shell .
the result is displayed in fig .
[ fsim ] , and shows that the @xmath96 case evolved into a more concentrated profile than the @xmath97 case , for radii up to @xmath98 pc or 10000 au .
this results from the fact that for the low magnetic field case ( @xmath97 ) , the fragmented area is large because of conservation of angular momentum , which gives a flatter density distribution at the center .
and the opposite , strong magnetic braking ( in the @xmath96 case ) helps concentrate matter in the center and gives a steeper profile .
in addition , previous theoretical work including ambipolar diffusion ( e.g. , lizano & shu 1989 ) also predict steep density profiles ( density power - law indices of @xmath99 ) .
thus , the current theoretical and numerical works suggest a relation between the inferred density power - law index of massive dense cores and the magnetic field properties . finally , we also used the simulations from commeron et al .
( 2011 ) to estimate the density within a radius of 0.05 pc for both @xmath96 and @xmath97 , and find @xmath63 @xmath100 @xmath22 and @xmath101 @xmath22 , respectively .
this , together with the number of millimeter sources that a typical interferometer would detect in each simulation ( estimated in palau et al .
2013 ) , allowed us to compare the results of these simulations to our @xmath73 vs @xmath63 plot showing observational data ( see fig .
[ fcorr]-c ) .
the values predicted by the simulations match remarkably well the trend seen in the observational data , suggesting that the magnetic field strength might be related to @xmath63 of massive dense cores , with small densities found in the strongly magnetized cores and large densities found in weakly magnetized cores .
overall , the agreement between commeron et al .
( 2011 ) simulations and our data is very good and suggests that magnetic field might be affecting the shape and density of massive dense cores , which in turn seems to be related to the fragmentation level .
in order to study the relation between the density profile of massive dense cores and the fragmentation level , we have fitted simultaneously the radial intensity profiles at 450 and 850 @xmath0 m and the sed of 19 massive dense cores for which their relative fragmentation level was known from previous works .
although the parameter used to estimate the fragmentation level is a lower limit , its relative value for the different cores is significant because it is assessed in a uniform way for all of them .
the sample consists of massive dense cores undergoing active star formation and being at similar evolutionary stages .
we modeled the cores as spherical envelopes with density and temperature decreasing with radius following power - laws , and found a weak trend of lower fragmentation for cores with steeper ( @xmath102 ) density profiles .
a correlation is found between the fragmentation level and the density within a radius of 0.05 pc ( the radius where fragmentation was assessed ) , suggesting that high fragmentation levels might be related with flat cores and cores with high central densities .
this correlation is consistent with jeans fragmentation , as high densities yield lower jeans masses . while we find no clear relation between velocity gradients or velocity dispersion and the density power law index of our massive dense cores , cores with compressive turbulence seem to be associated with high central densities , thus promoting fragmentation . on the other hand
, a comparison with radiation magnetohydrodynamical simulations suggests that a strong magnetic field could help shape steep density profiles in massive dense cores , for which low fragmentation levels are found .
thus , it is of crucial importance to study the relation of density profile , turbulence mode and magnetic field from both an observational an numerical point of view .
the authors are grateful to the anonymous referee for a critical reading of the manuscript providing key comments that improved the paper .
is grateful to tim jenness , luis chavarria , and floris van der tak for prompt support regarding the jcmt submillimeter continuum data , to ana duarte - cabral for sharing the herschel photometry of cygnus - x cores , to christoph federrath , susana lizano , enrique vzquez - semadeni , javier ballesteros - paredes , and andrew myers for insightful discussions , and carmen jurez and jennifer wiseman for sharing nh@xmath65 vla data .
a.p . and j.m.g
are supported by the spanish micinn grant aya2011 - 30228-c03 - 02 ( co - funded with feder funds ) , and by the agaur grant 2009sgr1172 ( catalonia ) .
thanks the spanish mineco for funding support from grants csd2009 - 00038 and aya2012 - 32032 .
is supported by an italian space agency ( asi ) fellowship under contract number i/005/11/0 .
acknowledge the financial support from dgapa , unam , and conacyt , mxico .
this research used the facilities of the canadian astronomy data centre operated by the national research council of canada with the support of the canadian space agency . | in order to shed light on the main physical processes controlling fragmentation of massive dense cores , we present a uniform study of the density structure of 19 massive dense cores , selected to be at similar evolutionary stages , for which their relative fragmentation level was assessed in a previous work .
we inferred the density structure of the 19 cores through a simultaneous fit of the radial intensity profiles at 450 and 850 @xmath0 m ( or 1.2 mm in two cases ) and the spectral energy distribution , assuming spherical symmetry and that the density and temperature of the cores decrease with radius following power - laws .
even though the estimated fragmentation level is strictly speaking a lower limit , its relative value is significant and several trends could be explored with our data .
we find a weak ( inverse ) trend of fragmentation level and density power - law index , with steeper density profiles tending to show lower fragmentation , and vice versa .
in addition , we find a trend of fragmentation increasing with density within a given radius , which arises from a combination of flat density profile and high central density and is consistent with jeans fragmentation .
we considered the effects of rotational - to - gravitational energy ratio , non - thermal velocity dispersion , and turbulence mode on the density structure of the cores , and found that compressive turbulence seems to yield higher central densities .
finally , a possible explanation for the origin of cores with concentrated density profiles , which are the cores showing no fragmentation , could be related with a strong magnetic field , consistent with the outcome of radiation magnetohydrodynamic simulations . |
there is convincing evidence for the existence of an unseen non - baryonic component in the energy - density of the universe .
the most promising dark matter candidates appear to be weakly interacting massive particles ( wimps ) and in particular the so - called neutralinos , arising in supersymmetric scenarios as well as much heavier particles ( wimpzillas ) which could have been produced non - thermally in the early universe ( for a review of particle candidates for dark matter see e.g. ellis ( 1998 ) ) .
the annihilations of these x - particles would generate quarks , leptons , gauge and higgs bosons and gluons .
consequently @xmath0 pairs would be produced either directly or by successive decays of mesons , and they are expected to lose their energy through synchrotron radiation as they propagate in the galactic magnetic field .
this radiation is expected to be greatly enhanced in the proximity of the galactic centre , where the existence of a massive black hole creates a region of very high dark matter particle density and consequently a great increase in the annihilation rate and synchrotron radiation .
we discuss in section 2 the distribution of dark matter particles in our galaxy and in particular around the central black hole , following gondolo & silk ( 1999 ) ( from now on , paper i ) . in section 3
we evaluate the annihilation @xmath0 spectrum and their synchrotron emission . finally in section 4
we discuss our results and determine the regions of the _ mass - annihilation cross section _ plane which give predictions compatible with experimental data .
there is strong evidence for the existence of a massive compact object lying within the inner @xmath1 of the galactic centre ( see yusef - zadeh , melia & wardle ( 2000 ) and references therein ) .
this object is a compelling candidate for a massive black hole , with mass @xmath2 .
the galactic halo density profile is modified in the neighborood of the galactic centre from the adiabatic process of accretion towards the central black hole . if we consider an initially power - law type profile of index @xmath3 , similar to those predicted by high resolution n - body simulations ( navarro , frenk & white 1997 ; ghigna et al .
2000 ) , the corresponding dark matter profile after accretion is , from paper i @xmath4^{\gamma_{sp}-\gamma } \rho_d \ ; g(r ) \left ( \frac{d}{r } \right)^{\gamma_{sp } } \label{modaccre}\ ] ] where @xmath5 , @xmath6 is the solar distance from the galactic centre and @xmath7 is the density in the solar neighbourhood . the factors @xmath8 and
@xmath9 can not be determined analytically ( for approximate expressions and numerical values see paper i ) .
the expression ( [ modaccre ] ) is only valid in a central region of size @xmath10 where the central black hole dominates the gravitational potential .
if we take into account the annihilation of dark matter particles , the density can not grow to arbitrarily high values , the maximal density being fixed by the value @xmath11 where @xmath12 is the age of the central black hole .
the final profile , resulting from the adiabatic accretion of annihilating dark matter on a massive black hole is @xmath13 following a power law for large values of r , and with a flat core of density @xmath14 and dimension @xmath15
among the products of annihilation of dark matter particles , there will be electrons and positrons , which are expected to produce synchrotron radiation in the magnetic field around the galactic centre .
the @xmath0 component produced by the hadronic jets has been computed using the mlla approximation .
the galactic magnetic field can be described using the equipartition assumption , where the magnetic , kinetic and gravitational energy of the matter accreting on the central black hole are in approximate equipartition ( see melia ( 1992 ) ) . in this case
the magnetic field can be expressed as @xmath16 [ [ energy - loss - length - scale . ] ] energy - loss length scale .
+ + + + + + + + + + + + + + + + + + + + + + + + + + as we shall see , most of the annihilations happen at very small distances from the centre , typically at @xmath17 , i.e. in a region with magnetic fields of the order of @xmath18 . under these conditions ,
comparable to the size of the region where most of the annihilations occur , the electrons lose their energy almost in place . to see this , consider the critical synchrotron frequency @xmath19 i.e. the frequency around which the synchrotron emission of an electron of energy e , in a magnetic field of strength b , peaks , namely @xmath20 inverting this relation , we find the energy of the electrons which give the maximum contribution at that frequency , @xmath21 the typical synchrotron loss length for an energy corresponding to @xmath22 can be expressed as a function of the distance @xmath23 from the central black hole for the magnetic field profile in eq .
( [ mag ] ) : @xmath24 for a frequency of 408 mhz , which produces , as we shall see , the most stringent upper bound on the galactic centre emission , we get @xmath25 the diffusion length @xmath26 , which can be approximated as a third of the radius of gyration of the electron , is given by @xmath27 since @xmath28 for all practically relevant parameters , an electron will diffuse a distance @xmath29 before it loses most of its energy .
numerically , @xmath30 or , expressing the distance as a function of the schwarzschild radius @xmath31 we can thus assume that the electrons lose their energy practically in place .
[ [ electron - production - spectrum . ] ] electron production spectrum .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + to compute the synchrotron luminosity produced by the propagation of @xmath32 in the galactic magnetic field , we need to evaluate their energy distribution in the magnetic field which is given by ( see gondolo ( 2000 ) ) @xmath33 where @xmath34 is the annihilation rate @xmath35 the function @xmath36 is given by @xmath37 and @xmath38 is the total synchrotron power spectrum .
note that the general expression for @xmath36 would have to take into account spatial redistribution by diffusion ( see e.g. gondolo ( 2000 ) ) , which we demonstrated to be negligible in our model .
the quantity @xmath39 is the number of @xmath0 with energy above e produced per annihilation , which depends on the annihilation modes .
( [ em ] ) shows that for the frequencies we are interested in , @xmath40 , and thus the energy dependence of @xmath39 can be neglected .
we will estimate @xmath39 by the number of charged particles produced in quark fragmentation ( see footnote ) ; in figure [ y ] we show the values of y as a function of the particle mass .
[ [ synchrotron - luminosity . ] ] synchrotron luminosity .
+ + + + + + + + + + + + + + + + + + + + + + + + for each electron the total power radiated in the frequency interval between @xmath41 and @xmath42 is given by @xmath43 where we introduced @xmath44 integrating this formula we obtain the total synchrotron luminosity @xmath45 which by substitution becomes @xmath46 it is possible to simplify this formula by introducing the following approximation for the function @xmath47 ( see rybicki & lightman ( 1979 ) ) @xmath48 the evaluation of the integral then gives @xmath49 where @xmath50 [ [ synchrotron - self - absorption . ] ] synchrotron self - absorption .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + the synchrotron self - absorption coefficent is by definition ( see rybicki & lightman ( 1979 ) ) @xmath51 where @xmath52 is the optical depth as a function of the cylindrical coordinate b @xmath53 and the coefficent @xmath54 is given by @xmath55 f\left ( \frac{\nu}{\nu_c } \right ) \;\ ; de\ ] ] the final luminosity is obtained by multiplying eq .
( [ lum ] ) with @xmath56 given by eq .
( [ anu ] ) .
it is evident that in the limit of small optical depths the coefficent @xmath57 , as can be seen by expanding the exponential .
furthermore the lower limit of integration of expression [ la ] is @xmath58 @xmath59 using the approximation introduced in eq .
( [ approx ] ) we find for @xmath54 the following expression @xmath60 which can in turn be used to evaluate @xmath52 in eq .
( [ la ] ) and @xmath56 in eq .
( [ anu ] ) .
note that the synchrotron loss time - scale is proportional to @xmath61 we compare this with the annihilation time , which determines where the inner density drops drastically : @xmath62 in other words , the annihilation time goes to zero more rapidly than the synchrotron loss time as @xmath23 goes to 0 .
hence synchrotron losses are important throughout the annihilation region .
extension to the wimpzilla mass range introduces far more uncertain physics .
the wimpzilla cross sections strongly depend on new physics beyond the electroweak scale . similarly to the case of electroweak gauge bosons , one can expect a general scaling @xmath63 for the self - annihilation cross section where @xmath64 is some dimensionless gauge coupling . in four dimensional field theory
, the conservation of probability ( unitarity ) roughly corresponds to this scaling for @xmath65 ( see , e.g. , weinberg ( 1995 ) ) . in order to retain generality
, we will therefore consider wimpzilla annihilation cross sections in the range @xmath66 .
we note , however , that beyond four dimensional field theory , such as in theories with extra dimensions and in string theory , larger cross sections may be possible .
we first evaluate the self - absorption coefficent for selected values of the mass , as a function of frequency .
the result , shown in fig.[absnu ] , is a coefficent which grows from very low values ( showing that absorption is important at 408 mhz ) and then reaches the value 1 , around a frequency which is strongly dependent on the cross section and the mass @xmath67 , but not so much on the profile power law index @xmath3 . on the left
the above coefficent is evaluated for two different values of the cross section , the first one corresponding to the maximum possible value ( the so called unitary bound , see griest & kamionkowski ( 1990 ) ) and the second one corresponding to typical cross sections one can find in supersymmetric scenarios ( see , e.g. , bergstrom , ullio & buckley ( 1997 ) ) . the right part of fig.[absnu ] shows the self - absorption coefficent at the fixed frequency of 408mhz as a function of the neutralino mass .
the behaviour shown is qualitatively the same for any value of the cross - section and for different @xmath3 . in figure [ predi1 ]
we compare the predicted spectrum with the observations ; we choose a set of parameters @xmath67 , @xmath3 and @xmath68 in order to reproduce the observed normalisation .
it is remarkable that in that way one can reproduce the observe d spectrum over a significant range of frequencies .
the set of dark matter parameters for which the fluxes predicted in our model are consistent with observation is shown in the exclusion plot of fig .
[ excludo ] .
the boundaries of the excluded range represent the parameter values for which the observed flux is explained by the dark matter scenario discussed here .
we have shown that present data on the emission from sgr a * are compatible with quite a wide set of dark matter parameters .
the evaluation of synchrotron self - absorption has enabled us to reach an alternative conclusion from an earlier claim of incompatibility of cuspy halos with the existence of annihilating wimp dark matter .
our final result is that the experimental data on sgr a * spectrum at radio wavelengths could be explained by synchrotron emission of electrons produced in the annihilation of relatively massive dark matter particles , extending from tev masses to @xmath69 .
the former is relevant to recent studies of coannihilation ( boehm , djouadi & drees ( 2000 ) ) , that suggest that wimps with @xmath70 can extend up to several tev , and the latter is relevant for particles ( wimpzillas ) that are produced non - thermally in the primordial universe .
we have found that with the current data situation , the synchrotron emission tends to give somewhat sharper constraints on masses and cross sections than the gamma - ray fluxes ( cf .
baltz et al .
( 2000 ) ) .
this situation could be reversed by more sensitive gamma - ray observations anticipated from upcoming experiments .
however the synchrotron predictions are uncertain because of our relative ignorance about the magnetic field strength near the central black hole , and gamma ray fluxes are subject to similarly uncertain amounts of self-absorption.the antares neutrino experiment will eventually be able to set a relatively model - independent limit on the annihilation flux from the galactic centre .
200 baltz e.a .
, briot c. , salati p. , taillet r. , silk j. , 2000 , prd , 61 , 023514 bergstrom l. , ullio p. , buckley j.h . , 1997 , astro - ph/9712318 bhattacharjee p. & sigl g. , 2000 , phys . rep .
, 327 , 109 , and references therein boehm c. , djouadi a. , drees m. , 2000 , prd , 62 , 035012 dokshitzer yu .
l. , khoze v. a. , mueller a. h. , & troyan s. i. , 1991 , basics of perturbative qcd , editions frontiers , saclay ellis j. , 1998,astro - ph/9812211 ellis r. k. , stirling w. j. , & webber b. r. , 1996 , qcd and collider physics , cambridge univ .
press , cambridge , england ghigna , s. , moore , b. , governato , f. , lake , g. , quinn , t. and stadel , j. 2000 , apj in press .
gondolo p. & silk j. , 1999 , prl , 83 , 1719 gondolo p. , 2000 , astro - ph/0002226 k.griest & m.kamionkowski , 1990 , prl , 64 , 615 khoze v. a. & ochs w. , int .
j. mod .
phys . , 1997 ,
a12 , 2949 melia f. , 1992 , apj , 387 , l25 navarro j. , frenk c.s . & white s.d.m . , 1997 , apj , 490 , 493 narayan et al . , 1998 , apj , 492 , 554 rybicki g.b . ,
lightman a.p .
, radiative processes in astrophysics , 1979 , john wiley & sons s. weinberg , 1995 , the quantum theory of fields vol 1 : foundations , cambridge university press , cambridge yusef - zadeh f. , melia f. & wardle m. , 2000 , astro - ph/0002376 | annihilations of weakly interacting dark matter particles provide an important signature for the possibility of indirect detection of dark matter in galaxy halos .
these self - annihilations can be greatly enhanced in the vicinity of a massive black hole .
we show that the massive black hole present at the centre of our galaxy accretes dark matter particles , creating a region of very high particle density .
consequently the annihilation rate is considerably increased , with a large number of @xmath0 pairs being produced either directly or by successive decays of mesons .
we evaluate the synchrotron emission ( and self - absorption ) associated with the propagation of these particles through the galactic magnetic field , and are able to constrain the allowed values of masses and cross sections of dark matter particles .
[ firstpage ] |
the interplay of magnetic impurities with strongly correlated electron systems has attracted considerable attention over the past decade . the discovery of high - temperature superconductors stimulated studies of impurities in two - dimensional mott - insulators with long - range antiferromagnetic order @xcite , where the copper - oxide parent compounds are driven to superconductivity by doping with holes or electrons . at low doping ,
the dopants are localized and it is therefore insightful to study the limit of isolated static holes , which have been realized experimentally @xcite and extensively studied theoretically @xcite .
there is a considerable body of work devoted to impurity bonds and added spins @xcite , a generalization of the static hole case , and an unexpected behavior of the impurity magnetic susceptibility at low temperature has been revealed quite recently @xcite .
there is also a separate and very interesting kondo - like problem of an impurity in a magnetic system close to an @xmath9 quantum critical point @xcite .
however , in this work , we consider impurities ( added spins ) in a system with long - range antiferromagnetic order .
the very unusual behavior we find , closely related to that observed for the magnetic susceptibility and other quantities in the papers @xcite , is due to gapless goldstone spin - wave excitations .
the rest of the paper is organized as follows . in section [ sec : model ]
we formulate the model , derive the spin - wave vertices and introduce an effective hamiltonian describing the interaction between impurities , which is explicitly calculated in sections [ sec : interaction ] an [ sec : rabi ] as a function of the distance between impurities .
a similar problem has been considered earlier in ref . , but only the case of small separations has been addressed , here we focus on the long - range behavior .
the magnetization disturbance in the host antiferromagnet induced by a single impurity is then calculated in section [ sec : mag ] , and finally section vi presents our conclusions .
we consider one and two spin @xmath10 impurities coupled to an isotropic heisenberg antiferromagnet ( @xmath11 ) on a square lattice with lattice spacing @xmath12 .
the impurities are connected to the origin and site @xmath13 of the antiferromagnet .
the hamiltonian of this system reads @xmath14 here @xmath15 denote nearest - neighbors , @xmath16 is a spin-@xmath17 operator at site @xmath18 and @xmath19 are spin-@xmath20 operators describing the impurities , which are either both ferromagnetic ( @xmath21 ) or antiferromagnetic ( @xmath22 ) .
after integration over quantum fluctuations of the system , one obtains the effective hamiltonian for the interaction between impurities @xmath23\nonumber\\ & + & \epsilon\left({\bf r}\right ) \sigma_1^z
\sigma_2^z+\left(m\left({\bf r}\right ) \sigma_1^+ \sigma_2 ^ - + h.c.\right)\ .\end{aligned}\ ] ] the calculation of this effective hamiltonian is the goal of the present work .
the diagonal interaction term @xmath24 , as well as the off - diagonal term @xmath25 are different , depending on whether the impurities are coupled to the same or different sublattices , see fig .
[ fig : lattice](a ) and ( b ) respectively . without loss of generality , we place the origin on the up sublattice ( sublattice `` a '' ) and indicate the sublattice the second impurity is coupled to by the corresponding letter , see figs .
[ fig : lattice ] . in this way , `` a '' ( `` b '' ) refers to impurities coupled to the same ( different ) sublattice .
, scaledwidth=30.0% ] we study the weak coupling regime @xmath26 , so that the dimensionless coupling constant is small @xmath27 to account for quantum fluctuations , we treat the host antiferromagnet in the linear spin - wave approximation where excitations are described by operators @xmath28 and @xmath29 creating spin - waves with @xmath30 and @xmath31 respectively , see ref . for a review . in this approximation ,
the hamiltonian ( [ eq : hamiltonian ] ) can be decomposed as @xmath32 , where @xmath33 \label{eq : hamup } & h_i^{(a)}&=\nonumber\\ & -&j ' \sigma_i^z \frac{2}{n}\sum_{{\bf p } , { \bf q } } e^{i\left({\bf p}-{\bf q}\right ) \cdot \bf{r_i } } \left ( u_{\bf p } \alpha_{\bf p}^\dag + v_{\bf p}\beta_{\bf p } \right ) \left ( u_{\bf q } \alpha_{\bf q } + v_{\bf q } \beta_{\bf q}^\dag\right)\nonumber\\ & + & j ' \sqrt{\frac{s}{n } } \left ( \sigma_i^+ \sum_{\bf q } e^{i{\bf q } \cdot \bf{r_i } } \left ( u_{\bf q } \alpha_{\bf q}^\dag + v_{\bf q}\beta_{\bf q}\right ) + \text{h.c .
} \right)\ , \\ \label{eq : hamdown } & h_i^{(b ) } & = \nonumber\\ & & j ' \sigma_i^z \frac{2}{n}\sum_{{\bf p } , { \bf q } } e^{i\left({\bf p}- { \bf q}\right ) \cdot \bf{r_i } } \left ( u_{\bf p } \beta_{\bf p}^\dag + v_{\bf p } \alpha_{\bf p } \right ) \left ( u_{\bf q } \beta_{\bf q } + v_{\bf q } \alpha_{\bf q}^\dag \right)\nonumber\\ & + & j ' \sqrt{\frac{s}{n } } \left ( \sigma_i^+ \sum_{\bf q } e^{i { \bf q } \cdot \bf{r_i } } \left ( u_{\bf q } \beta_{\bf q } + v_{\bf q } \alpha_{\bf q}^\dag \right ) + \text{h.c . }
\right ) \ .\end{aligned}\ ] ] here @xmath34 is the ground state energy of the antiferromagnetic host and @xmath35 the number of sites .
the upper ( lower ) sign in the @xmath36 expression refers to the situation where the second impurity is coupled to a spin on sublattice `` a '' ( sublattice `` b '' ) .
this convention is used throughout the whole paper .
the bogoliubov parameters @xmath37 and @xmath38 are given by @xmath39 with @xmath40 and @xmath41 , see ref . .
in this notation , the spin - wave dispersion is @xmath42
. the interaction hamiltonians ( [ eq : hamup ] ) and ( [ eq : hamdown ] ) generate one and two spin - wave vertices summarized in table [ tab : vertices ] .
cccc symbol & operator & factor for @xmath43 on `` a '' & or @xmath44 on `` b '' + + @xmath45 & @xmath46 & @xmath47 & @xmath48 + @xmath49 & @xmath50 & @xmath51 & @xmath52 + @xmath53 & @xmath54 & @xmath55 & @xmath56 + @xmath57 & @xmath58 & @xmath59 & @xmath60 + + @xmath61 & @xmath62 & @xmath63 & @xmath64 + @xmath65 & @xmath66 & @xmath67 & @xmath68 + @xmath69 & @xmath70 & @xmath71 & @xmath72 + @xmath73 & @xmath74 & @xmath75 & @xmath76 + let us return to the effective hamiltonian ( [ eq : effham ] ) .
the first two terms are obvious and do not require calculations .
they simply generate three energy levels , @xmath77 , @xmath78 , and @xmath79 . to be specific ,
let us consider the case shown in fig .
[ fig : lattice](a ) , with impurities coupled to the same sublattice , then @xmath80 is the true stationary quantum state .
the states @xmath81 , @xmath82 , and @xmath83 decay to the ground state with emission of spin - waves . using the fermi golden rule and the decay matrix elements presented in tab .
[ tab : vertices ] , one finds the following widths of excited states with respect to the emission of magnons @xmath84 the three - level system is well defined , since @xmath85 .
but the diagonal and off - diagonal interaction energies @xmath24 and @xmath25 in the effective hamiltonian ( [ eq : effham ] ) have limited meaning because of the finite lifetime . only in the ground state @xmath86 , the diagonal interaction energy is well defined . in the spin flip
states @xmath81 , @xmath82 , and @xmath83 the diagonal interaction energy does not make much sense , because we will see that it is always much smaller than the corresponding decay width .
however , there is a regime where the off - diagonal interaction @xmath25 is larger than the decay width and hence leads to rabi oscillations between states @xmath81 and @xmath82 .
in the leading order of the @xmath89-expansion , the interaction energy @xmath24 arises in second , third and fourth orders of perturbation theory , describing the exchange of two spin - waves between impurities .
the corresponding contributions in usual rayleigh - schrdinger perturbation theory are @xcite @xmath90 where @xmath91 is the perturbation ( [ hamv ] ) . in general , the expressions for third and fourth order energy corrections are more complex than those in ( [ 234 ] ) .
the complication is due to contributions similar to ( [ 234 ] ) , but in which intermediate states @xmath92 coincide with the initial state @xmath86 .
fortunately , the perturbation ( [ hamv ] ) does not allow such intermediate states and therefore eqs .
( [ 234 ] ) are valid .
-corrections for antiferromagnetic impurities coupled to sites on the same sublattice ( case `` a '' ) .
the spin - wave vertices are summarized in tab .
[ tab : vertices ] , dashed and dotted lines represent @xmath93- and @xmath94-spin - waves respectively . _
[ fig : diagramssame],scaledwidth=25.0% ] the matrix elements @xmath95 are given in table [ tab : vertices ] and it is convenient to represent the corrections ( [ 234 ] ) by diagrams where each vertex corresponds to some particular matrix element @xmath95 .
-corrections for antiferromagnetic impurities coupled to sites on different sublattices ( case `` b '' ) .
the notation is the same as in fig .
[ fig : diagramssame ] . _
[ fig : diagramsdifferent],scaledwidth=25.0% ] the diagrams describing the @xmath89-corrections are shown in figs .
[ fig : diagramssame ] and [ fig : diagramsdifferent ] for antiferromagnetic impurities ( @xmath96 ) coupled to the same and different sublattices respectively .
ferromagnetic impurities ( @xmath97 ) generate similar diagrams with interchanged @xmath93- and @xmath94-spin - waves .
the first diagram in fig .
[ fig : diagramssame ] for instance represents the expression @xmath98 with a factor @xmath99 taking into account the similar process with exchanged impurities . summing these contributions
, we obtain the interaction energy in the leading @xmath89-order @xmath100 with @xmath101 the sum extends over momenta in the magnetic brillouin zone . in the case of ferromagnetic impurities ( @xmath21 )
, the bogoliubov parameters in ( [ eq : int ] ) have to be interchanged . for large distances , @xmath102 ,
the interaction energy comes from small momenta , @xmath103 .
one can therefore approximate the bogoliubov parameters and the dispersion by @xmath104 and hence simplify the interactions ( [ eq : int ] ) @xmath105 here @xmath106 is the dimensionless coupling constant ( [ gc ] ) . since both integrals are ultraviolet convergent , we extend the integration domain to infinity .
there is no difference between ferro- and antiferromagnetic ( @xmath107 ) impurities in this approximation .
we first consider _ very large _
distances @xmath108 . in this case ,
momenta in ( [ e12 ] ) are limited by @xmath109 and the corrections to the ground state energy are equal to @xmath110 where according to our convention the plus sign corresponds to @xmath111 ( same sublattice ) and the minus sign corresponds to @xmath112 ( different sublattices ) .
a first integration over angles gives two bessel functions @xmath113 , which can be integrated using @xmath114 , yielding @xmath115 interestingly , in this limit , the interaction between impurities is independent of their coupling to the antiferromagnetic host @xmath116 . in the case of
_ intermediate _ distances between impurities , @xmath117 , a similar calculation using @xmath118 , leads to @xmath119 if the separation between the impurities is small , @xmath120 , the approximation ( [ eq : approximation ] ) for the bogoliubov parameters and the dispersion is no longer valid and the integrals in ( [ eq : int ] ) have to be calculated numerically . in order to circumvent finite - size effects
, we extrapolate the results obtained from lattices with up to @xmath121 sites by a polynomial @xmath122 .
even for very small distances , the interaction between additional spins is surprisingly well fitted by the intermediate distance attractive behavior ( [ dei ] ) .
the situation is analogous to the magnetization disturbance , see section v , for which fig .
[ fig : mag ] provides an illustration of qualitatively similar fits .
the attractive interaction ( [ dei ] ) found in the weak coupling regime is independent of the sublattice . in comparison , in the strong coupling limit @xmath123
an attractive interaction is found for nearest neighbors , but the interaction between impurities on next - nearest sites is repulsive @xcite . in the above calculations
, we assumed that the impurities are in the ground state , see eq .
( [ 012 ] ) . to check the kinematic structure of the diagonal interaction in the effective hamiltonian ( [ eq : effham ] ) one has to perform similar calculations for states @xmath81 , @xmath82 , and @xmath83 , see ( [ 012 ] ) . in the case
where both impurities are flipped with respect to their ground state configuration , the interaction is the same as in the ground state , since the situation corresponds to ferromagnetic coupling with @xmath22 . for an excited state with only one flipped impurity , a calculation analogous to ( [ eq : int ] ) shows that the interaction has opposite sign and hence justifies the kinematic structure of the @xmath124 term in ( [ eq : effham ] ) .
we emphasize ( see also end of section [ sec : model ] ) , that because of the finite lifetime , the diagonal interaction in the spin - flipped states has limited meaning .
if a magnetic impurity is flipped with respect to the ground state configuration , e.g. in a nuclear magnetic resonance ( nmr ) or electron paramagnetic resonance ( epr ) experiment , then the excited state has a finite lifetime due to emission of magnons , see first of eqs .
( [ gsw ] ) .
however , there is another mechanism of spin relaxation , due to presence of distant similar magnetic impurities : a spin - wave exchange leads to rabi oscillations between two impurities . in the effective hamiltonian ( [ eq : effham ] )
this process is described by the off - diagonal term @xmath25 .
diagrams for @xmath25 are shown in fig .
[ fig : diagramssrabi ] .
rabi oscillations can only be observed between impurities coupled to spins on the _ same _ sublattice , so @xmath125 , scaledwidth=25.0% ] the oscillation frequency of this two - level system is proportional to the real part of the mixing matrix element @xmath126 , since the probability @xmath127 to find the system in a state with flipped impurities after the time @xmath128 is equal to @xmath129 . using tab . [
tab : vertices ] to evaluate the diagrams shown in fig .
[ fig : diagramssrabi ] , we find @xmath130 for large distances @xmath102 we use the approximate bogoliubov parameters ( [ eq : approximation ] ) .
the mixing element becomes @xmath131 where @xmath132 is the neumann function . for _
very large _ distances between impurities , @xmath5 , the real part of the mixing matrix element is comparable to the imaginary one and both are much smaller than the width ( [ gsw ] ) . in this case , there are no rabi oscillations .
however , in the _ intermediate _ regime , @xmath117 , the real part @xmath133 is logarithmically enhanced compared to the imaginary part and compared to the spin - wave width ( [ gsw ] ) .
thus , in this regime , rabi oscillations between impurities are well pronounced and this mechanism gives the main contribution to the effective width of the magnetic resonance line @xmath134
an interesting question is how an additional spin influences the magnetic order in the host antiferromagnet , described by the staggered magnetization . let us consider an impurity @xmath135 at the origin and a local magnetic field @xmath136 on site @xmath1 .
the hamiltonian reads @xmath137 and the variation of the expectation value of the local magnetization @xmath138 in the ground state is given by @xmath139 where @xmath140 is the part of the energy dependent on @xmath116 .
now we can consider the same framework as for the calculation of the interaction between impurities .
( [ sh ] ) together with the rayleigh - schrdinger perturbation theory , one can see that @xmath141 where @xmath142 is the energy correction described by those diagrams in figs .
[ fig : diagramssame ] and [ fig : diagramsdifferent ] without spin flip of the second impurity .
these diagrams give the following explicit expressions for the magnetization variation @xmath143 with @xmath144assuming @xmath96 .
the expressions for the ferromagnetic case ( @xmath145 ) are obtained by a change of the global sign and interchanged bogoliubov parameters in ( [ eq : mag ] ) . using the approximations ( [ eq : approximation ] ) for @xmath146
, we find that the magnetization disturbance is equal to @xmath147 in the _ very large _ distance limit , and @xmath148 for the _ intermediate _ region . in the case of small separation from the impurity
, one has to calculate the integrals in ( [ eq : mag ] ) numerically . using the same finite - size extrapolation scheme as in section [ sec : interaction ]
we find that the variation of the magnetization in the vicinity of the impurity is described by eq .
( [ dsi ] ) down to @xmath149 .
[ fig : mag ] displays the variation of the magnetization calculated numerically at @xmath150 .
the results are very close to the analytical expression ( [ dsi ] ) . ) at site @xmath1 induced by a ferromagnetic impurity coupled to the origin by @xmath150 .
the curves show fits of these points according to @xmath151 , where @xmath152 is the fitting coefficient .
fitted values of @xmath152 ( see inset ) are very close to the value @xmath153 that follows from eq .
( [ dsi ] ) .
[ fig : mag]_,scaledwidth=45.0% ] in agreement with ref .
an added spin always enhances the nel order in the host magnet , independent of the sign of the exchange coupling .
in contrast to a vacancy , this enhancement is not limited to nearest neighbor sites @xcite , but extends over the whole magnet .
it is also interesting to compare our result for an added spin to an in - plane impurity considered in ref . .
if the impurity is placed inside the host , it weakens the nel order of the surrounding spins , but the magnetization disturbance also decreases as @xmath154 for @xmath146 @xcite .
to conclude , we have studied the long - range dynamics of one and two spin-@xmath20 impurities in a two - dimensional heisenberg antiferromagnet with on site spin-@xmath17 treated in the linear spin - wave approximation .
the impurities are assumed to be weakly coupled to the host magnet by a small dimensionless coupling constant @xmath106 . a systematical treatment of the corrections contributing to the leading order of the @xmath89-expansion leads to non - trivial long - range dynamics .
the interaction between two impurities can be separated into two regimes : for very large separations ( @xmath5 ) it is universal ( independent of @xmath116 ) and decreases as @xmath154 .
the interaction is repulsive ( attractive ) for impurities coupled to the same ( different ) sublattices . in an intermediate region @xmath117 ,
the interaction decreases only as @xmath155 and is attractive , independent of the sign of the exchange couplings .
it is shown that rabi oscillations between impurities coupled to spins on the same sublattice are possible and well pronounced in the intermediate regime .
the effective hamiltonian decsribing the interaction in terms of the impurity spins is derived .
it exhibits an xyz anisotropy which leads to nmr and epr line broadening .
the magnetization disturbance in the host magnet induced by a single impurity is analyzed in the same framework .
it is shown that the disturbance exhibits behaviors similar to the interaction energy , always enhancing the magnetic order in the antiferromagnetic host .
numerical results indicate that the intermediate regimes can be extended down to @xmath156 .
99 e. manousakis , rev .
phys . * 63 * , 1 ( 1991 ) .
e. dagotto , rev .
phys . * 66 * , 763 ( 1994 ) .
cheong , a. s. cooper , l. w. b. batlogg , j. d. thompson , and z. fisk , , 9739 ( 1991 ) .
s. t. ting , p. pernambuco - wise , j. e. crow , e. manousakis , and j. weaver , , 11772 ( 1992 ) .
m. corti , a. rigamonti , f. tabak , p. caretta , f. licci , and l. raffo , , 4226 ( 1995 ) .
p. carretta , a. rigamonti , and r. sala , , 3734 ( 1997 )
. o. p. vajk , p. k. mang , m. greven , p. m. gehring , and j. w. lynn , science * 295 * , 1691 ( 2002 ) .
n. nagaosa , y. hatsugai , and m. imada , j. phys .
. jpn . * 58 * , 978 ( 1989 ) .
n. bulut , d. hone and d. j. scalapino , and e. y. loh , , 2192 ( 1989 ) .
d. poilblanc , d. j. scalapino , and w. hanke , , 884 ( 1994 ) .
s. a. krivenko and g. g. khaliullin , jetp lett . * 62 * , 723 ( 1995 ) .
n. nagaosa and t .- k .
ng , , 15588 ( 1995 ) . a. w. sandvik , e. dagotto , and d. j. scalapino , , 11701 ( 1997 ) .
v. yu . irkhin , a. a. katanin , and m. i. katsnelson , , 14779 ( 1999 ) .
j. igarashi , k. murayama , and p. fulde , , 15966 ( 1995 ) .
k. murayama and j. igarashi , j. phys .
. jpn . * 66 * , 1157 ( 1997 ) .
v. n. kotov , j. oitmaa , and o. p. sushkov , , 8495 ( 1998 ) ; _ ibid_. * 58 * , 8500 ( 1998 ) .
s. sachdev , c. buragohain , and m. vojta , science * 286 * , 2479 ( 1999 ) .
m. vojta , c. buragohain , and s. sachdev , , 15152 ( 2000 ) .
k. h. hglund and a. w. sandvik , , 077204 ( 2003 ) . s. sachdev and m. vojta , , 064419 ( 2003 ) .
o. p. sushkov , , 094426 ( 2003 ) .
o. p. sushkov , , 12135 ( 2000 ) . m. troyer , suppl .
theor . phys . *
145 * , 326 ( 2002 ) .
a. l. chernyshev , y. c. chen , and a. h. castro neto , , 104407 ( 2002 ) .
l. d. landau and e. m. lifshitz .
quantum mechanics : non - relativistic theory .
oxford , new york , pergamon press , 1977 . | we consider a two - dimensional heisenberg antiferromagnet on a square lattice with weakly coupled impurities , i.e. additional spins interacting with the host magnet by a small dimensionless coupling constant @xmath0 .
using linear spin - wave theory , we find that the magnetization disturbance at distance @xmath1 from a single impurity behaves as @xmath2 for @xmath3 and as @xmath4 for @xmath5 .
surprisingly the magnetization disturbance is inversely proportional to the coupling constant !
the interaction between two impurities separated by a distance @xmath1 is @xmath6 for @xmath3 and @xmath7 for @xmath5 . for large distances ,
the interaction is therefore universal and independent of the coupling constant .
we have also found that the frequency of rabi oscillations between two impurities is logarithmically enhanced compared to the decay width @xmath8 at @xmath3 .
this leads to a logarithmic enhancement for nmr and epr line broadening .
all these astonishing results are due to the gapless spectrum of magnetic excitations in the quantum antiferromagnet . |
for comparison with ce@xmath2nd@xmath3coin@xmath0 , we grew crystals of ce@xmath2nd@xmath3rhin@xmath0 by an in - flux technique @xcite and studied its pressure and field dependence by electrical resistivity and ac calorimetry measurements ( see methods for details ) . figure [ fig : fig1]a shows the low - temperature electrical resistivity , @xmath33 , on sample s1 at representative pressures , and the inset displays @xmath33 in the whole @xmath12-range .
although nd - substitution reduces @xmath34 from 3.8 k to 3.4 k and slightly increases the residual resistivity @xmath35 to 0.2 @xmath36 , the @xmath37-dependence reported in fig .
1a is essentially identical to that of cerhin@xmath0 below @xmath38 gpa where @xmath5 equals @xmath25 .
we also note that , at zero pressure , the @xmath39 phase diagram of ce@xmath2nd@xmath3rhin@xmath0 closely resembles the one found in cerhin@xmath0 .
these results indicate that @xmath1 nd does not change drastically the local afm character of @xmath25 below @xmath40 .
once @xmath5 exceeds @xmath25 at @xmath40 , there is no evidence for magnetism in @xmath33 , and the possibility of nd - induced magnetism is obscured by the zero - resistance state below @xmath5 . to investigate whether there is afm order in the sc state , heat capacity measurements are necessary .
figure [ fig : fig1]b shows the @xmath12-dependence of heat capacity divided by temperature ( @xmath41 ) for sample s2 at various pressures . at ambient pressure , @xmath41 peaks at @xmath25 as in cerhin@xmath0 . as @xmath12 is lowered further
, however , @xmath42 turns up below @xmath43 k , which was not observed in cerhin@xmath0 @xcite .
this upturn is presumably associated with the nuclear moment of nd ions , and it can be fit well by a sum of electronic ( @xmath44 ) and nuclear ( @xmath45 ) terms @xcite .
the inset of fig .
1b shows that @xmath46 is linear in @xmath47 , consistent with the presence of a nuclear schottky contribution .
we note , however , that an upturn also is observed at @xmath48 gpa ( not shown ) where magnetic order is absent , suggesting that the hyperfine field may not be solely responsible for splitting the nuclear levels .
although nd nuclei have large zero - field quadrupole moments , kondo - hole physics can not be ruled out as a possible source of the upturn @xcite .
dependence of the in - plane electrical resistivity , @xmath33 , of ce@xmath2nd@xmath3rhin@xmath0 ( s1 ) under pressure .
arrows mark @xmath25 determined by peaks in the first derivative .
the inset shows @xmath33 over the entire @xmath12-range .
b ) @xmath42 @xmath49 @xmath12 for ce@xmath2nd@xmath3rhin@xmath0 ( s2 ) under pressure .
a vertical offset of 2.5 units is added for clarity .
arrows ( asterisks ) denote @xmath25 ( @xmath5 ) .
inset shows a linear fit in a @xmath46 @xmath49 @xmath47 plot . ] for pressures below @xmath40 , @xmath25 evolves with @xmath37 as it does in transport data .
evidence for bulk sc ( marked by asterisks ) , however , is observed at lower temperatures relative to the zero - resistance state in @xmath33 .
a difference between zero - resistance and bulk sc transitions also appears in cerhin@xmath0 and has been attributed to filamentary sc due to the presence of long - range afm order @xcite . unlike cerhin@xmath0 at pressures greater than @xmath26
, however , there is evidence for a phase transition in the sc state of ce@xmath2nd@xmath3rhin@xmath0 without an applied field . at @xmath50 gpa , an anomaly in @xmath42
is observed at @xmath51 k ( arrow in fig .
[ fig : fig1]b ) , below the sc transition at @xmath5 = @xmath52 k. for reasons discussed below , this anomaly stems from a magnetic order , and it is fundamentally different from the afm order displayed by the system for pressures smaller than @xmath26 .
the shape and magnitude of the anomaly relative to that at @xmath5 are very similar to those at @xmath25 in ce@xmath2nd@xmath3coin@xmath0 ( see supplemental materials ) , and the small entropy associated with it suggests that the magnetic order is most likely a small - moment density wave . as we will come to later , this evidence is most obvious in data shown in fig .
[ fig : fig3 ] .
we summarize the zero - field results discussed above in the @xmath12-@xmath37 phase diagram shown in fig .
[ fig : fig2 ] .
local - moment - like afm order coexists with bulk sc in a narrow pressure range below @xmath40 . from just below to just above @xmath40 , @xmath29 changes discontinuously , in contrast to field - induced magnetism in cerhin@xmath0 , which is a smooth continuation of @xmath29 from below @xmath26 @xcite .
this supports the interpretation that @xmath9- and nd - induced magnetic orders have different origins .
therefore , the nd - induced transition is labeled @xmath53 to distinguish it from @xmath25 in pure cerhin@xmath0 . above @xmath40
, @xmath53 is suppressed at a rate of @xmath54 k / gpa and extrapolates to zero temperature , i.e. a quantum - critical point , at @xmath55 gpa inside the superconducting phase . whether the coincidence of @xmath56 and @xmath57 is accidental or not requires further investigation beyond the scope of our work . as shown in the supplemental materials ,
the rate of suppression of this nd - induced transition is the same rate found in ce@xmath2nd@xmath3coin@xmath0 within experimental uncertainty , strongly indicating a common origin . because entropy associated with the zero - field transition is rather small , as found in ce@xmath31nd@xmath32coin@xmath0 , the typical signature of quantum criticality ( i.e. , divergence of @xmath58 at low-@xmath12 ) is likely hidden by sc and by the upturn in @xmath58 .
we also note that the highest @xmath5 achieved in ce@xmath2nd@xmath3rhin@xmath0 , @xmath59 k , is @xmath60 k lower than @xmath61 of cerhin@xmath0 .
this same suppression of @xmath61 is observed in ce@xmath2nd@xmath3coin@xmath0 and indicates that nd ions act similarly as magnetic pair - breaking impurities .
phase diagram of ce@xmath2nd@xmath3rhin@xmath0 ( s2 ) obtained from ac calorimetry measurements . here
@xmath62 gpa and @xmath55 gpa . ] to further investigate the nature of the nd - induced magnetism , we turn to the field - dependent heat capacity data .
figure [ fig : fig3]a shows @xmath41
@xmath49 @xmath12 at @xmath50 gpa ( @xmath63 ) and low magnetic fields .
the zero - field transition at @xmath64 k remains unchanged in a field of @xmath65 koe .
as the field is increased further , however , the specific heat anomaly splits , one anomaly moving to lower temperatures and the other to higher temperatures . at @xmath66 koe ,
our data show features at @xmath51 k and @xmath67 k. previous reports on cerhin@xmath0 at similar pressures ( @xmath68 gpa ) show that the field - induced transition emerges at @xmath69 k when @xmath70 koe @xcite .
it is thus reasonable to associate the anomaly we observe at @xmath51 k and 11 koe with the field - induced magnetism in cerhin@xmath0 .
further , fig .
[ fig : fig3]b shows the high-@xmath9 evolution of @xmath25 with a field dependence very similar to that of cerhin@xmath0 : @xmath25 first increases with @xmath9 and then remains constant above the upper critical field @xmath28 . as shown in the @xmath39 phase diagram ( fig .
3c ) , this field - induced @xmath25 clearly competes with @xmath53 , as reflected in the rapid suppression of @xmath53 as a function of @xmath9 .
in fact , no evidence for @xmath53 is observed at fields @xmath71 koe , implying a field - induced quantum - critical point in addition to the pressure - induced , zero - field critical point of this order . due to the reasons explained above
, our results strongly point to two distinct types of magnetism emerging in ce@xmath2nd@xmath3rhin@xmath0 .
the first one is due to nd ions and it has the hallmarks of that in ce@xmath2nd@xmath3coin@xmath0 . the second is the @xmath9-induced magnetism that appears in pure cerhin@xmath0 @xcite .
what is the role of nd and why is it special ? at a concentration of @xmath1 , average spacing of @xmath72@xmath73 and non - periodic distribution on ce - sites , nd is too dilute to induce magnetic order by dipole or indirect ruderman - kittel - kasuya - yosida interactions .
its role , then , must be more subtle . using the bulk modulus of cerhin@xmath0 and the unit - cell volume variation in ce@xmath31nd@xmath32rhin@xmath0
, we estimate that ce@xmath2nd@xmath3rhin@xmath0 experiences an effective chemical pressure of @xmath74 gpa relative to cerhin@xmath0 @xcite . from the phase diagram of cerhin@xmath0
, this @xmath75 would increase @xmath25 by @xmath76 k instead of producing the observed reduction .
hence , we conclude that chemical pressure @xmath77 is not the dominant tuning parameter .
the disruption of translational periodicity of the ce lattice by nd substitution creates a kondo hole " that contributes to reducing @xmath25 at zero pressure @xcite .
the latter conclusion is supported by the observation that non - magnetic la substitution for ce in cerhin@xmath0 also depresses @xmath25 similarly @xcite .
neodymium , however , carries an additional magnetic moment that is unlikely to be quenched by kondo screening . in the context of cecoin@xmath0 ,
michal and mineev @xcite proposed that the @xmath13-phase observed in the presence of an in - plane magnetic field is the consequence of the condensation of the spin - exciton collective mode found in the sc phase .
thus , it is natural to consider whether the nd magnetic moments immersed in cerhin@xmath0 at zero field could also promote a similar behavior . as discussed in detail in the supplemental materials , condensation of spin excitons takes place when the spin - resonance - mode frequency @xmath78 vanishes . within an random phase approximation ( rpa )
approach , the latter is given by the pole of the renormalized magnetic susceptibility , i.e. when @xmath79 , where @xmath80 is the effective electronic interaction projected in the sdw channel and @xmath81 is the non - interacting magnetic susceptibility inside the sc state .
when the ordering vector @xmath16 connects points of the fermi surface with different signs of the sc gap , * @xmath82 * , @xmath83 generically diverges when @xmath84 and remains non - zero when @xmath85 .
thus , even a very weak @xmath80 can in principle induce a resonance mode with frequency near @xmath86 . once the interaction increases , @xmath78 moves to lower frequencies
. when the interaction overcomes a critical value , @xmath87 , the resonance mode vanishes and sdw order is established inside the sc dome . as function of the total impurity scattering rate for the cases of non - magnetic ( red curve ) and paramagnetic impurities ( blue curve ) .
b ) the suppression of the effective sc gap @xmath4 by both magnetic and non - magnetic impurities . in these plots
, we considered point - like impurities .
here , @xmath88 is the gap of the clean system whereas @xmath89 is the landau damping ( see supplemental materials for details ) . ] in our case , the interaction @xmath80 is presumably independent of pressure .
thus , in order for magnetic impurities to promote spin - exciton condensation , @xmath90 must increase ( i.e. the critical interaction value @xmath91 must decrease ) as function of the impurity potential . to investigate whether this is a sensible scenario , we considered a toy model consisting of two `` hot spots '' located at momenta @xmath92 and @xmath93 at the fermi surface such that * @xmath82*. note that such a hot - spots model has been previously employed to study the effects of disorder on sc @xcite . to focus on the general properties of the model , we linearize the band dispersion around the hot spots and compute both @xmath90 and the effective gap amplitude @xmath4 at @xmath94 as function of the total impurity scattering rate @xmath95 within the self - consistent born approximation ( similarly to what was done in ref .
@xcite for @xmath96 sc and perfectly nested bands ) .
as shown in figure 4 , whereas both magnetic and non - magnetic impurity scattering suppress @xmath4 at the same rate ( panel 4b ) , we find that @xmath90 is suppressed for non - magnetic impurity but enhanced by magnetic impurity scattering ( panel 4a ) .
thus , in the case of non - magnetic impurities , athough the resonance mode frequency may decrease as compared to the clean case , it never collapses to zero .
because magnetic impurity scattering enhances @xmath90 but not necessarily destroys sc , the system may undergo an sdw transition inside the sc dome .
although the fate of the system will depend on microscopic details beyond those captured by the toy model considered here , our model nicely illustrates that it is plausible for magnetic impurity scattering to drive spin - exciton condensation in the sc phase . in this regard
, we note that previous investigations of a microscopically - motivated theoretical model also found that the @xmath13-phase may be stabilized by magnetic impurities even at zero external field @xcite .
these results suggest that other magnetic impurities could induce the same type of sdw order in both cecoin@xmath0 and pressurized cerhin@xmath0 .
in fact , we show in the supplemental materials that easy - plane gd@xmath97 ions ( @xmath98 ) also induce a transition in the heat capacity data of both co and rh members .
this anomaly is similar to the one induced by easy - axis ( @xmath8-axis ) nd@xmath97 ions ( @xmath99 , @xmath100 , @xmath101 ) discussed above .
hence , nd is not `` special '' in inducing magnetic order in the superconducting states of ce@xmath2nd@xmath3coin@xmath0 and ce@xmath2nd@xmath3rhin@xmath0 , and other unconventional superconductors that host a spin resonance mode may also display zero - field magnetism via the same mechanism .
our results also imply that non - magnetic impurities will not induce condensation of excitations , in agreement with experimental data on la - substituted cerhin@xmath0 @xcite .
finally , we note that the sdw ordering vector in nd - doped cecoin@xmath0 corresponds closely to the nodal structure @xcite that is also found in the superconducting state of cerhin@xmath0 @xcite .
hence , the magnetic wave - vector of zero - field order above @xmath40 in ce@xmath2nd@xmath3rhin@xmath0 should also be close to * * q**@xmath102 , and we expect neutron scattering experiments to find a spin resonance of @xmath8-axis character below @xmath5 in cerhin@xmath0 at @xmath30 .
in summary , we generalize the observation of nd - induced magnetism in cecoin@xmath0 to pressurized cerhin@xmath0 and gd - substituted members . this spin - density wave order , which reflects the nodal - gap symmetry , is argued to be a consequence of the condensation of spin excitations that arise inside the sc state . given the several similarities between cecoin@xmath0 and ce@xmath103pdin@xmath104 @xcite , nd substitution might nucleate afm order in its superconducting state .
appropriate substitutions in other unconventional superconductors that host a spin resonance also should induce zero - field magnetism by the same mechanism , and magnetism should be tunable to a quantum - critical point inside their sc phase .
we thank a. v. chubukov , h. l@xmath105hneysen , s. maiti , and p. g. pagliuso for useful discussions .
work at los alamos by y.l .
, f.r . , and j.d.t .
was performed under the auspices of the u.s .
department of energy , office of basic energy sciences , division of materials science and engineering .
f. s. r. acknowledges a director s postdoctoral fellowship through the lanl ldrd program . the theoretical work ( j.k . and r.m.f .
) was supported by the u.s .
department of energy , office of science , basic energy sciences , under award number de - sc0012336 .
the series ce@xmath31nd@xmath32rhin@xmath0 was grown by the in - flux technique and its properties are reported elsewhere @xcite .
crystals with _
_ x__=0.05 and free of unreacted in were mounted in a hybrid piston - cylinder pressure cell , filled with silicone fluid as the pressure medium , and a piece of pb whose change in _ _
t__@xmath106 served as a manometer .
gd - substituted crystals were grown using the same method .
electrical resistivity was measured by a four - probe method with current flow in the _
semi - quantitative heat capacity was obtained by an ac calorimetry technique described elsewhere @xcite .
magnetic fields to 9 t were applied parallel to the _ ab_-plane .
the results above have been reproduced in different crystals and , for clarity , we show resistivity and calorimetry data for two representative samples labeled sample 1 ( s1 ) and sample 2 ( s2 ) , respectively . 99 monthoux p , pines d & lonzarich gg ( 2007 ) superconductivity without phonons .
nature 450 , 1177 - 1183 .
scalapino dj ( 2012 ) a common thread : the pairing interaction for unconventional superconductors .
84 , 1383 .
schilling a , cantoni m , guo jd & ott hr ( 1993 ) superconductivity above @xmath107 k in the hg - ba - ca - cu - o system .
_ nature _ 363 , 56 .
stewart gr .
superconductivity in iron compounds ( 2011 ) _ rev .
phys . _ * 83 * , 1589 .
white bd , thompson jd , maple mb ( 2015 ) unconventional superconductivity in heavy - fermion compounds . _ physica c _ * 514 * , 246 . for a review ,
see m. eschrig , adv .
55 , 47 ( 2006 ) .
yu g , li y , motoyama em & greven , m ( 2009 ) a universal relationship between magnetic resonance and superconducting gap in unconventional superconductors . _
phys . _ * 5 * , 873 .
park wk , sarrao jl , thompson jd , and greene lh ( 2008 ) andreev reflection in heavy - fermion superconductors and order parameter symmetry in cecoin@xmath0 .
100 , 177001 .
zhou bb , _ et al . _
( 2013 ) visualizing nodal heavy fermion superconductivity in cecoin@xmath0 .
nature phys . 9 , 474 . stock c , broholm c , hudis j , kang h j & petrovic c ( 2008 ) spin resonance in the @xmath108-wave superconductor cecoin@xmath0 .
lett . _ * 100 * , 087001 .
raymond s & lapertot g ( 2015 ) ising incommensurate spin resonance of cecoin@xmath0 : a dynamical precursor of the @xmath13 phase .
_ * 115 * , 037001 .
panarin j , raymond s , lapertot g & flouquet j ( 2009 ) evolution of the spin resonance in cecoin@xmath0 under magnetic field .
_ j. phys .
jpn . _ * 78 * , 113706 .
stock c , broholm c , zhao y , demmel f , kang hj , rule kc & petrovic c ( 2012 ) magnetic field splitting of the spin resonance in cecoin@xmath0 .
_ * 109 * , 167207 .
young bl , _ et al . _
( 2007 ) microscopic evidence for field - induced magnetism in cecoin@xmath0 . _
_ * 98 * , 036402 .
kenzelmann m , _ et al . _
( 2008 ) coupled superconducting and magnetic order in cecoin@xmath0 .
_ science _ * 321 * , 1652 .
kenzelmann m , _ et al . _
( 2010 ) evidence for a magnetically driven superconducting @xmath13 phase of cecoin@xmath0 .
lett . _ * 104 * , 127001 .
michal vp & mineev vp ( 2011 ) field - induced spin - exciton condensation in the @xmath109-wave superconductor cecoin@xmath0 .
b _ * 84 * , 052508 .
hu r , lee y , hudis j , mitrovic vf & petrovic c ( 2008 ) composition and field - tuned magnetism and superconductivity in nd@xmath31ce@xmath32coin@xmath0 .
b _ * 77 * , 165129 .
raymond s , ramos sm , aoki d , knebel g , mineev vp & lapertot g ( 2014 ) magnetic order in ce@xmath2nd@xmath3coin@xmath0 : the @xmath13-phase at zero magnetic field . _ j. phys .
jpn . _ * 83 * , 013707 .
lake b , _ et al . _
( 2002 ) antiferromagnetic order induced by an applied magnetic field in a high - temperature superconductor .
_ nature _ * 415 * , 299 .
demler e , sachdev s & zhang y ( 2011 ) spin - ordering quantum transitions of superconductors in a magnetic field .
lett . _ * 87 * , 067202 .
park t , _ et al_. ( 2006 ) hidden magnetism and quantum criticality in the heavy fermion superconductor cerhin@xmath0 .
_ nature _ * 440 * , 65 - 68 .
bao w , _ et al .
_ ( 2000 ) incommensurate magnetic structure of cerhin@xmath0 .
b 62 , r14621(r ) .
hegger h , _ et al . _ ( 2000 ) pressure - induced superconductivity in quasi-2d cerhin@xmath0 .
lett . _ * 84 * , 4986 .
mito t , _ et al_. ( 2003 ) coexistence of antiferromagnetism and superconductivity near the quantum criticality of the heavy - fermion compound cerhin@xmath0 .
lett . _ * 90 * , 077004 .
knebel g , aoki d , braithwaite d , salce b & flouquet j ( 2006 ) coexistence of antiferromagnetism and superconductivity in cerhin@xmath0 under high pressure and magnetic field .
b _ * 74 * , 020501(r ) .
pham ld , park t , maquilon s , thompson jd & fisk z ( 2006 ) reversible tuning of the heavy - fermion ground state in cecoin@xmath0 .
97 , 056404 .
rosa pfs , oostra a , thompson jd , pagliuso pg & fisk z ( 2016 ) unusual kondo - hole effect and crystal - field frustration in nd - doped cerhin@xmath0 .
b _ 94 , 045101 .
sidorov va , thompson jd & fisk z ( 2010 ) magnetic transitions in a heavy - fermion antiferromagnet u@xmath103zn@xmath110 at high pressure .
_ j. phys . : condens
* 22 * , 406002 .
we note that an upturn in @xmath58 also has been observed in a commercial quantum design ppms by means of quasiadiabatic thermal relaxation technique .
anderson ac , holmstrom b , krusius m & pickett gr ( 1969 ) calorimetric investigation of the hyperfine interactions in metallic nd , sm , and dy .
rev . _ * 183 * , 546 .
pagliuso pg , _ et al . _
( 2002 ) ce - site dilution studies in the antiferromagnetic heavy fermions ce@xmath111rh@xmath112in@xmath113 ( @xmath114 ) .
b _ * 66 * , 054433 .
bauer ed , _ et al . _
( 2011 ) electronic inhomogeneity in a kondo lattice .
usa _ * 108 * , 6857 park t & thompson jd ( 2009 ) magnetism and superconductivity in strongly correlated cerhin@xmath0 .
_ new j. phys . _ * 11 * , 055062 .
mendona - fereira l , _ et al_. ( 2008 ) tuning the pressure - induced superconducting phase in doped cerhin@xmath0 .
* 101 * , 017005 .
pagliuso pg , _ et al . _
( 2002 ) structurally tuned superconductivity in heavy - fermion ce@xmath17in@xmath0 ( @xmath18 co , ir , rh ) . _
physica b _ * 320 * , 370 - 375 .
park t , bauer ed , & thompson jd ( 2008 ) probing the nodal gap in the pressure - induced heavy fermion superconductor cerhin@xmath0 .
lett . _ * 101 * 177002 .
kang j and fernandes rm ( 2016)robustness of quantum critical pairing against disorder .
b * 93 * , 224514 .
maiti s , knolle j , eremin i , and chubukov av ( 2011 ) effect of nodes , ellipticity , and impurities on the spin resonance in iron - based superconductors .
b * 84 * , 144524 .
martiny jhj , gastiasoro mn , vekhter i , and andersen bm ( 2015 ) impurity - induced antiferromagnetic order in pauli - limited nodal superconductors : application to heavy - fermion cecoin@xmath0 .
phys . rev .
b * 92 * , 224510 .
panarin j , raymond s , lapertot g , flouquet j , and mignot jm ( 2011 ) effects of nonmagnetic la impurities on the spin resonance of ce@xmath31la@xmath32coin@xmath0 single crystals as seen via inelastic neutron scattering .
b _ * 84 * , 052505 . nakatsuji s , _ et al . _
( 2002 ) intersite coupling effects in a kondo lattice .
lett . * 89 * 106402 .
dong jk , _ et al_. ( 2011 )
field - induced quantum critical point and nodal superconductivity in the heavy - fermion superconductor ce@xmath103pdin@xmath104 .
x _ * 1 * , 011010 .
* + 0.3 cm competing magnetic orders in the superconducting state of heavy - fermion cerhin@xmath0 *
0.25 cm single crystals of nominally ce@xmath2nd@xmath3coin@xmath0 were grown from an excess in flux and characterized by pressure - dependent ac calorimetry , as described in the main text .
these measurements were carried out in a pressure cell under conditions also discussed in the main text .
figure s1 shows @xmath41 as a function of @xmath12 at two pressures . at atmospheric pressure ( top curve ) , a pronounced peak in @xmath41 defines the superconducting transition temperature @xmath5 = 1.63 k that is followed at lower temperatures by an anomaly peaked near 0.7 k. these features in data at atmospheric pressure coincide well with those reported in ref .
[ 15 ] ( main text ) and indicate that the nominal nd content is very close to the actual content .
consequently , we associate the lower temperature anomaly with antiferromagnetic order at @xmath53 .
we note that the shapes of the two anomalies resemble those found in ce@xmath2nd@xmath3rhin@xmath0 . as shown in fig .
s1a , an applied pressure of only 0.13 gpa reduces @xmath53 to @xmath115 k. assuming that @xmath53 decreases linearly with @xmath37 ( fig .
s1a inset ) gives d@xmath53/dp @xmath116 k / gpa .
this rate of decrease is very close to that found in fig . 2 of the main text where , for ce@xmath2nd@xmath3rhin@xmath0 above @xmath26 , d@xmath117dp @xmath118 k / gpa .
the fact that nd - induced zero - field antiferromagnetic order in similarly doped cecoin@xmath0 and cerhin@xmath0 displays the same sensitivity to pressure strongly suggests that the zero - field magnetism has a common origin in both materials .
@xmath49 @xmath12 of ce@xmath2nd@xmath3coin@xmath0 at two pressures .
arrows show @xmath25 and inset shows its extrapolation to @xmath94 .
( b ) @xmath42 @xmath49 @xmath12 of ce@xmath2gd@xmath3coin@xmath0 at 2.05 gpa .
( c ) @xmath42 @xmath49 @xmath12 of ce@xmath31gd@xmath32coin@xmath0 ( @xmath119 ) at ambient pressure .
the curves are shifted by @xmath120 j / mol.k@xmath121 for clarity .
arrows show @xmath25 determined as the minimum in the first derivative of the data . ]
single crystals of nominally ce@xmath2gd@xmath3rhin@xmath0 were grown from an excess in flux and characterized by pressure - dependent ac calorimetry , as described in the main text .
figure s1b shows @xmath41 as a function of @xmath12 at 2.05 gpa .
a pronounced peak in @xmath41 defines the superconducting transition temperature @xmath5 = 1.95 k that is followed at lower temperatures by an anomaly peaked near 0.7 k determined by the minimum in derivative of @xmath58 .
finally , single crystals of nominally ce@xmath31gd@xmath32coin@xmath0 ( @xmath119 ) were grown from an excess in flux and characterized by specific heat at ambient pressure in a commercial ppms .
figure s1c shows @xmath41 as a function of @xmath12 for all concentrations .
a pronounced peak in @xmath58 defines the superconducting transition temperature @xmath5 = @xmath122 k , @xmath123 k , and @xmath124 k that is followed at lower temperatures by an anomaly peaked near @xmath125 k , @xmath67 k , and @xmath126 k.
0.25 cm to gain insight into the problem , and keep the calculation analytically tractable , we employ the hot spots approximation . in particular , we consider two points on the fermi surface , labeled @xmath8 and @xmath108 , separared by the sdw vector @xmath127 , such that @xmath128 .
next , we linearize the dispersions around the hot spots @xmath129 where the momentum @xmath130 is measured with respect to @xmath131 and @xmath132 are the fermi velocities . here
, we consider @xmath133 and let the relative angle between them be arbitrary but non - zero .
the case @xmath134 corresponds to perfect nesting , which was studied in ref .
@xcite . here
, we focus on the case where nesting is not perfect .
as long as the sdw instability is driven by the low - energy electronic states , the linearized approximation can be used to compute the leading contribution to the magnetic susceptibility .
there are two different types of impurity potentials : the non - magnetic potential @xmath135 , which couples to the charge degrees of freedom , and the paramagnetic potential @xmath136 , which couples to the spin degrees of freedom .
each impurity potential can be split into small - momentum scattering , @xmath137 and @xmath138 ( which does not couple the fermions on the two hot spots ) , and large - momentum scattering @xmath139 and @xmath140 ( which couples the fermions on the two hot spots ) .
first , we study the impact of the various types of disorder on the pairing gap . here
, we assume the existence of a sc state , without discussing its origin .
thus , we consider a static pairing interaction @xmath141 corresponding to a repulsive interaction coupling fermions from different hot spots , such that @xmath142 .
to proceed , it is convenient to define nambu spinors @xmath143 .
the green s function in nambu space is then given by : @xmath144 where @xmath145 is the imaginary part of the normal component of the self - energy and @xmath146 ( the gap normalized by impurities ) is the real part of the anomalous component of the self - energy .
as shown in fig [ fig : fermionse ] , the fermionic self - energy in the self - consistent approximation becomes , at @xmath94 @xmath147 to solve for @xmath145 and @xmath146 , we average over impurities and introduce the scattering rates @xmath148 for non - magnetic impurities and @xmath149 for paramagnetic impurities . in the hot - spots model ,
the density of states is @xmath150 , with @xmath151 denoting the momentum cutoff parallel to the fermi surface .
we obtain : @xmath152 from the last equation , it is clear that only the inter - hot - spot non - magnetic impurity and the intra - hot - spot paramagnetic impurity suppress the sc gap .
the last equation can be solved self - consistently to find @xmath146 ; it is convenient then to define an effective sc order parameter given by : @xmath153 @xmath158\left(-\gamma_{\mathrm{sdw}}^{\alpha}f_{\delta^{2}}+\gamma_{\mathrm{sdw}}^{\beta}f_{\omega\delta}\right)\\ \gamma_{\mathrm{sdw}}^{\beta } & = \tau_{\mathrm{damping}}\left[\left(\frac{1}{2\tau_{0}}-\frac{1}{2\tau_{{\boldsymbol{q}}}}\right)+\frac{1}{3}\left(\frac{1}{2\tau_{0}^{p}}-\frac{1}{2\tau_{{\boldsymbol{q}}}^{p}}\right)\right]\left(\gamma_{\mathrm{sdw}}^{\alpha}f_{\omega\delta}+\gamma_{\mathrm{sdw}}^{\beta}f_{\delta^{2}}\right ) \end{split}\ ] ] computing the static magnetic susceptibility ( @xmath160 ) gives : @xmath161^{-1}\ ] ] where we defined the total non - magnetic scattering rate @xmath162 and the total paramagnetic scattering rate @xmath163 .
note that the dressed sdw vertex becomes a constant , independent of the frequency @xmath164 .
clearly , while non - magnetic scattering always reduces @xmath165 , paramagnetic impurity scattering enhances it . in fig .
4 of the main text , we considered the case of point - like impurities , in which @xmath166 and @xmath167 . | applied pressure drives the heavy - fermion antiferromagnet cerhin@xmath0 towards a quantum critical point that becomes hidden by a dome of unconventional superconductivity .
magnetic fields suppress this superconducting dome , unveiling the quantum phase transition of local character . here
, we show that @xmath1 magnetic substitution at the ce site in cerhin@xmath0 , either by nd or gd , induces a zero - field magnetic instability inside the superconducting state .
this magnetic state not only should have a different ordering vector than the high - field local - moment magnetic state , but it also competes with the latter , suggesting that a spin - density - wave phase is stabilized in zero field by nd and gd impurities similarly to the case of ce@xmath2nd@xmath3coin@xmath0 . supported by model calculations , we attribute this spin - density wave instability to a magnetic - impurity driven condensation of the spin excitons that form inside the unconventional superconducting state .
unconventional superconductivity ( sc ) frequently is found as an antiferromagnetic ( afm ) transition is tuned by chemical substitution or pressure toward a zero - temperature phase transition , a magnetic quantum - critical point .
this observation has a qualitative explanation : the proliferation of quantum fluctuations of magnetic origin at low temperatures can trigger the formation of a new ordered state .
unconventional superconductivity is a natural candidate state because it can be induced by an attractive cooper - pair interaction provided by the fluctuating magnetism @xcite .
typical examples include copper - oxides , which without chemical substitution are afm mott insulators @xcite , metallic iron - based antiferromagnets that superconduct under pressure or with chemical substitutions @xcite , and rare - earth heavy - fermion compounds with large effective electronic masses @xcite .
a characteristic manifestation of the unconventional nature of the superconducting state is the momentum dependence of the sc gap @xmath4 that develops below the superconducting transition temperature ( @xmath5 ) .
in contrast to conventional superconductors , @xmath4 is not uniform but instead has different signs in different regions of the fermi surface . despite the distinct chemical and electronic properties of these materials ,
the interplay between magnetism and sc is common among them , calling for a deep understanding of this relationship . in this
regard , heavy - fermion materials offer an ideal platform to explore the relationship between these two phases .
an additional common feature among these different classes of superconductors is the emergence of a collective magnetic excitation below @xmath5 often attributed to the formation of a spin exciton @xcite .
this collective mode , whose energy has been shown to scale with @xmath4 across different materials @xcite , is a direct consequence of the sign - changing nature of @xmath4 .
an example is the heavy - fermion superconductor cecoin@xmath0 , known to be very close to an afm quantum - critical point without tuning @xcite .
indeed , its sc gap is sign - changing @xcite and a spin resonance mode is observed below @xmath5 @xcite . the energy of this mode in cecoin@xmath0 scales with its sc gap with the same proportionality found in copper - oxide and iron - based systems .
recent inelastic neutron scattering experiments find that the resonance mode in cecoin@xmath0 is incommensurate at the wavevector @xmath6 @xcite .
due to ce s @xmath7 crystal - field environment , this mode is a doublet and the corresponding fluctuations are polarized along the @xmath8-axis .
when a magnetic field @xmath9 is applied in the tetragonal @xmath10-plane , this mode splits into two well - defined branches @xcite .
the field dependence of the zeeman - split lower energy mode extrapolates to zero energy at @xmath11 koe , which is remarkably close to the field where long - range afm order develops inside the low-@xmath12 , high-@xmath9 sc state @xcite .
spin - density wave ( sdw ) order in this so - called @xmath13-phase has a small @xmath8-axis ordered moment of @xmath14 @xmath15 , which corresponds closely to the spectral weight of the low - energy resonance mode .
moreover , the sdw displays the same incommensurate wave - vector @xmath16 as the spin resonance mode @xcite .
these observations suggest that the @xmath13-phase is the result of a condensation of spin excitations @xcite .
in addition to the field - induced @xmath13-phase , afm order is found in ce@xmath2nd@xmath3coin@xmath0 below @xmath5 , in this case at zero field @xcite .
the wave - vector and moment size of the nd - induced magnetism are the same as those observed in the @xmath13-phase of cecoin@xmath0 @xcite .
although the sign - changing @xmath4 , with its nodes on the fermi surface , plays a non - trivial role in enabling these orders , the obvious similarity between @xmath9- and nd - induced magnetism strongly suggests that they have a common origin , namely condensation of the spin excitations that give rise to the resonance mode .
no evidence for the @xmath13-phase has been found in other ce@xmath17in@xmath0 members ( @xmath18 rh , ir ) or , for that matter , in any other superconductor .
it is uncommon to find a magnetic transition below @xmath5 when both superconducting and magnetic states arise from the same electrons . besides the example of cecoin@xmath0
, field - induced magnetism has been observed in la@xmath19sr@xmath20cuo@xmath21 @xcite .
this afm order , however , is distinct from a @xmath13-like phase and is closely related to the field - induced magnetism in the sc state of pressurized cerhin@xmath0 @xcite . at zero pressure , cerhin@xmath0 displays afm order at @xmath22 k and @xmath23 @xcite .
pressurizing cerhin@xmath0 tunes its magnetic transition toward a quantum - critical point and induces sc that coexists with afm order for pressures up to @xmath24 gpa , where @xmath5 equals @xmath25 . above @xmath26 ,
evidence for @xmath25 is absent and only sc is observed @xcite .
application of a magnetic field , however , induces magnetism in the sc state between @xmath26 and the quantum - critical point at @xmath27 gpa @xcite . unlike magnetic order in the @xmath13-phase , which exists only inside the sc state , field - induced magnetism in cerhin@xmath0
persists into the normal state above the pauli - limited @xmath28 and is a smooth continuation of the zero - field @xmath29 boundary @xcite .
this magnetism may obscure or preempt the formation of a @xmath13-like phase , but strong similarities of cecoin@xmath0 to cerhin@xmath0 at @xmath30 @xcite suggest the possibility that afm order might develop in the high pressure sc state of ce@xmath31nd@xmath32rhin@xmath0 in zero field . in this paper
, we show that nd induces a zero - field phase transition in the high - pressure sc phase of ce@xmath2nd@xmath3rhin@xmath0 and present evidence that the phase transition is due to magnetic order .
this result generalizes the observation of magnetic order below @xmath5 in ce@xmath2nd@xmath3coin@xmath0 because pressure suppresses the magnetic order at the same rate in both compounds .
our model calculations support our conclusion that the magnetism in nd - doped cerhin@xmath0 is due to the condensation of spin excitations promoted by magnetic impurity scattering , and is thus distinct from the local - moment magnetism in pure cerhin@xmath0 promoted by the application of magnetic fields . in agreement with this proposal , we observe a competition between the field - induced magnetism , which displays the same behavior as in cerhin@xmath0 , and the nd - induced magnetism in zero field . hence , we expect a spin resonance with @xmath8-axis character below @xmath5 in cerhin@xmath0 at pressures greater than @xmath26 .
more generally , our work reveals a route to induce zero - field magnetic order via chemical substitution of magnetic impurities in other unconventional superconductors that host spin resonance modes . |
the @xmath2 was discovered about 20 years ago in @xmath3 collisions at cleo @xcite and cusb @xcite .
its mass is @xmath4 mev , about 300 mev higher than the @xmath5 resonance , and its width of @xmath6 mev is about five times larger than the @xmath5 resonance .
its production cross - section was measured to be about 0.3 nb , about ten times lower than the @xmath5 cross - section .
owing to its mass , the @xmath2 can decay to a variety of final states containing @xmath0 and @xmath1 mesons .
the accessible @xmath0-meson final states include : @xmath7 , @xmath8 , @xmath9 , @xmath10 , @xmath11 , @xmath12 and @xmath13 ; final states with @xmath1 are : @xmath14 , @xmath15 , and @xmath16 .
further study of the @xmath2 is of great interest because of it being a potentially large source of @xmath1 mesons at an @xmath3 machine .
if suitably large samples can be collected , their study would provide a complementary probe of cp violation and rare decays to lhcb .
studies involving rare decays , particularly those with soft photons , and untagged time - dependent and time - independent measurements , which provide access to the width difference between the @xmath1 mass eigenstates , @xmath17 , offer complementary information to measurements at hadron machines @xcite . because of the small lorentz boost at the @xmath2 and the large value of the mixing frequency , @xmath18 @xcite , time - dependent cp violation measurements in mixing are not possible . until recently
, there existed only weak limits on the production rate of @xmath1 at the @xmath2 .
several models attempt to describe the hadronic cross - section in the 10 - 11 gev region .
the simplest model , the quark - pair creation model ( qpc ) @xcite , allows for quark pairs to be ` popped ' out of the vacuum .
this model by itself is insufficient to describe the hadronic cross - section .
et al . _
@xcite are able to reproduce the spectrum by introducing a @xmath19 hybrid at 10683 mev ( the qhm model ) . a third model , the unitarized quark model ( uqm ) @xcite includes coupled channel effects between @xmath20 states via real ( two - meson ) intermediate states .
the qhm predicts @xmath8 is dominant at 10865 mev , with relative rates of @xmath9:@xmath8:@xmath7 @xmath21 1:3:6 . on the other hand
, the uqm also reproduces the hadronic cross - section , but predicts that @xmath9 is dominant , _
e.g. , _ @xmath9:@xmath8:@xmath7
2:1:0.3 . a more thorough experimental study of the @xmath2 has been undertaken by cleo , followed by belle , to better understand the hadronic cross - section in this region as well as assess the potential for a future high luminosity @xmath3 machine operating at the @xmath2 . in 2003 , cleo collected 0.42 @xmath22 of data at the @xmath2 ( 10871 mev ) in 2005 , belle collected about 1.9 @xmath22 on the @xmath2 resonance , followed by a larger sample of 22 @xmath22 a year later . here
we report on results from cleo and the first belle data set .
the hadronic resonance cross section is measured by counting hadronic events recorded at the @xmath2 using various selection criteria employed to reduce the continuum background .
the continuum background is estimated using data taken below the @xmath5 , and is scaled by : @xmath23 , where @xmath24 and @xmath25 are the integrated luminosities and center - of - mass energies squared at the @xmath5 and @xmath2 .
the @xmath2 cross - sections are measured to be @xmath26 nb @xcite and @xmath27 nb @xcite , by cleo and belle , respectively .
many of the same techniques of exclusive reconstruction that are employed at the @xmath5 are used to reconstruct @xmath0 and @xmath1 mesons at the @xmath2 .
the kinematic variables @xmath28 and the beam - energy constrained mass , @xmath29 are used . here
, @xmath30 is the beam energy and @xmath31 ( @xmath32 ) is the energy ( momentum ) of the @xmath0 candidate .
because of the proximity of the @xmath1 mass to the beam energy , @xmath14 , @xmath15 , and @xmath16 are kinematically well - separated . for @xmath0 mesons , @xmath7 , @xmath8 , @xmath9 and @xmath13
are also kinematically separated .
@xmath10 , @xmath11 and @xmath12 occupy a broad , but distinct region in @xmath33 and overlap with one another , and partially with @xmath13 . because @xmath34 candidates are reconstructed ( not @xmath35 ) , only for @xmath7 and @xmath14 does @xmath36 peak at zero and @xmath37 peak at the physical @xmath0 and @xmath1 masses .
for the other processes , @xmath36 and @xmath37 are offset by a predictable amount . by measuring the yields in each of these kinematically defined regions , one can determine the @xmath0 and @xmath1 fractions at the @xmath2 as well as the contributing subprocesses .
a second approach for obtaining the @xmath0 and @xmath1 yields at the @xmath2 is to use inclusive particle yields .
this technique exploits the large difference in inclusive rates for @xmath0 and @xmath1 to various particles , such as @xmath38 , @xmath39 , and @xmath40 .
cleo reconstructs @xmath0 mesons via their decays @xmath41 and @xmath42 ( 25 channels in total ) @xcite .
first , the inclusive @xmath0 yield is determined by computing the invariant mass of all @xmath0 meson candidates with @xmath43 mev and @xmath44 mev . unlike @xmath37 , invariant mass
does not depend on the particular @xmath2 decay mode .
a fit to the invariant mass spectrum yields a signal of @xmath45 events . using efficiencies ( @xmath46 ) from monte carlo ( mc ) simulation and branching fractions ( @xmath47 ) from the pdg @xcite , it is found @xmath48 , where the sum is over the 25 decay modes . from this
it is concluded that @xmath49 nb . using this cross - section and the measured total hadronic cross - section , the @xmath1 fraction in @xmath2 decays , @xmath50 ,
is computed to be @xmath51 . to determine the contributing subprocesses ,
events are selected in the expected signal band in the @xmath33 plane and projected onto the @xmath37 axis .
the resulting distribution of events is shown in fig .
[ fig : mbc_b ] . a dominant @xmath9 peak is observed , along with a smaller contribution from @xmath8 , and an insignificant yield from @xmath7 .
the distributions are fit to the sum of 3 gaussians , with resolutions determined from mc and the mass difference between the @xmath9 and @xmath7 peaks constrained to the expected value of 47.5 mev .
the fitted yields are @xmath52 @xmath7 , @xmath53 @xmath8 and @xmath54 @xmath9 events .
insignificant yields in @xmath55 and @xmath13 lead to 90% confidence level upper limits of 13.1 and 6.4 events , respectively .
thus , it is found that @xmath56 of the @xmath0 production is through @xmath9 and @xmath57 is from @xmath8 . for @xmath7 , @xmath55 and @xmath13 90% upper limits of 22% , 32% and 14% ,
respectively , are obtained .
these results strongly disfavor the qhm and are consistent with the uqm .
the @xmath37 peak position for @xmath9 and the corresponding peak position from @xmath16 @xcite can be used to make a precise measurement of the @xmath58 mass . noting that @xmath59 mev , and using the well - measured @xmath60 mass @xcite , gives @xmath61 mev , which provides a significant improvement over the previous measurement .
the m1 mass splitting for @xmath1 is thus found to be @xmath62 mev , which is consistent with the value in the @xmath0 system of @xmath63 mev @xcite , as expected from heavy quark symmetry . for @xmath0 signal candidates in cleo .
the histogram is the data and the line is a fit as described in the text .
the shaded regions indicate the shapes expected from @xmath55 and @xmath13 .
the inset shows the invariant mass for these particular final states , as described in the text .
[ fig : mbc_b ] ]
both cleo and belle reconstruct @xmath64 @xcite and @xmath65 candidates . in fig .
[ fig : bs - excl ] we show the @xmath36 versus @xmath37 and @xmath37 distribution for @xmath65 candidates from the cleo @xmath2 data sample .
the three boxes , from left to right , indicate signal regions for @xmath14 , @xmath15 , and @xmath16 .
all 10 signal candidates are @xmath16 .
an additional 4 events are observed in the @xmath66 modes , all of which are also @xmath16 . using efficiencies from simulation and branching fractions from the pdg @xcite , yields @xmath67 , where @xmath68 neglects the evidently small @xmath14 and @xmath15 contributions .
belle observes a total of 23 events ( 21 @xmath16 , 2 @xmath15 , 0 @xmath14 ) and determines @xmath69 of @xmath1 are from @xmath16 @xcite . versus @xmath37 ( left ) and @xmath37 ( right ) for @xmath65 candidates in the cleo @xmath2 data sample .
the superimposed boxes are discussed in the text .
[ fig : bs - excl ] ]
inclusive analyses are senstive to @xmath50 , provided the _ probe _
particle chosen has sufficiently different production rate for @xmath0 and @xmath1 decays .
cleo has used both @xmath38 and @xmath40 mesons , both of which are expected to have substantially higher rates from @xmath1 than @xmath0 mesons @xcite .
belle followed up with a similar pair of analyses that use @xmath38 and @xmath39 mesons @xcite .
measurement of the @xmath1 fraction , @xmath50 , is obtained using an expression of the form : @xmath70 where @xmath71 is the number of @xmath2 decays , and we have inserted @xmath38 as the probe particle as an example . here
one measures @xmath72 and @xmath73 , and makes a model - dependent estimate of @xmath74 @xcite , leaving only @xmath50 unknown . using similar arguments
, it is estimated that @xmath75 @xcite .
the first measurement of @xmath50 using this technique was reported by cleo , where they measure : @xmath76 @xcite using @xmath77 mesons .
the @xmath39 and @xmath38 scaled momentum spectra from belle using @xmath2 data ( points ) and the below-@xmath5 continuum ( histogram ) are shown in fig .
[ fig : dsd0inc ] .
the excess above continuum corresponds to the left - hand side of eqn .
[ eq : eq1 ] . using @xmath78 and @xmath79 @xcite ,
belle obtains : @xmath80 from @xmath38 and @xmath81 from @xmath39 , which are consistent with each other and with the previous cleo result .
, for ( a ) @xmath82 and ( b ) @xmath83 from belle .
the points are data taken at the @xmath2 resonance and the histogram is the continuum obtained from @xmath5 data , normalized by the ratio of luminosities and @xmath84 .
[ fig : dsd0inc],title="fig : " ] , for ( a ) @xmath82 and ( b ) @xmath83 from belle .
the points are data taken at the @xmath2 resonance and the histogram is the continuum obtained from @xmath5 data , normalized by the ratio of luminosities and @xmath84 . [
fig : dsd0inc],title="fig : " ] cleo has also performed an analysis using inclusive @xmath40 production @xcite , where @xmath85 .
the measured yields as a function of scaled momentum are shown in fig .
[ fig : phi-5s ] . the first bin @xmath86 is obtained from mc simulation , because in this bin the kaons generally have too low momentum to be reconstructed .
the uncertainty is taken to be 100% of its value . from the measured @xmath87 , @xmath88 and @xmath89 branching fractions @xcite , it is deduced that most @xmath40 mesons in @xmath34 decay come from the cascade @xmath90 or @xmath91 , which allows one to make a model - dependent estimate @xmath92 .
one then uses eqn .
[ eq : eq1 ] to obtain @xmath93 .
the combined cleo average , using exclusive @xmath0 mesons , and inclusive @xmath38 and @xmath40 mesons , is @xmath94 .
( filled circles ) and @xmath95 ( triangles ) .
[ fig : phi-5s ] ] these studies firmly establish the @xmath1 production rate to be about 25% , with roughly 90% of it produced as @xmath96 .
it is also found that @xmath9 is dominant , with @xmath8 about ( 1/3)@xmath9 .
other processes are less than about 25% of @xmath9
. these results are consistent with the uqm , and inconsistent with the qhm .
with a large sample of data collected on the @xmath2 , one can begin to search for rare @xmath1 decays .
while many of the decays can be searched for in @xmath97 ( cdf , d0 ) and @xmath98 ( lhcb at cern , starting 2008 ) , the @xmath3 environment enjoys low particle multiplicity and good signal - to - background for low energy photons .
final states with soft photons are particularly difficult for detectors at hadron machines , and so this is an area where machines operating at the @xmath2 can make significant contributions . taking @xmath99 nb ( 25% of the resonance cross - section ) , a 50 @xmath22 dataset , which could be collected by belle in a couple of months , yields about 7.5 million @xmath1 decays . in the standard model , the width difference , @xmath100 between the cp eigenstates , @xmath101 and @xmath102
is given by @xmath103 , where @xmath104 is the off - diagonal matrix element in the @xmath1 decay matrix , which connects @xmath1 and @xmath105 via on - shell intermediate states .
these matrix elements are saturated by tree - diagrams , such as @xmath106 , and therefore @xmath107 is expected to be insensitive to new physics . because there are more cp even than cp odd final states in @xmath1 decays , one expects @xmath108@xcite . in the presence of cp violation ,
one finds a width difference between the mass eigenstates of @xmath109 , where @xmath110 , is the difference between the @xmath1 mixing phase and the phase of @xmath111 . in the standard model , @xmath40
is expected to be small , of order 0.01 .
therefore , one expects @xmath112 . since new physics could affect @xmath1 mixing , _
i.e. , _ the phase of @xmath113 , new physics could drive @xmath17 away from @xmath100 .
precise measurements of @xmath100 can be obtained by measuring the branching fractions and cp content of final states accessible to @xmath1 and @xmath105 .
such final states are dominated by @xmath114 , but also include @xmath115 , @xmath116 @xmath117 , etc . under certain assumptions @xcite , it can be shown that @xmath106 is pure cp even , and neglecting the charmonium states , one finds @xmath118 . with a 50 @xmath22 @xmath2 data sample , one would expect @xmath2110 events in each decay mode @xcite , yielding @xmath2125% accuracy on @xmath119 .
if these assumptions are relaxed , one must measure the cp fraction of the various vector - vector final states , which would require substantially larger statistics .
various rare loop - induced decay modes have been searched for in the belle data sample , including @xmath120 , @xmath121 , and @xmath122 @xcite .
the limit @xmath123 , is already more stringent than the current world average .
in summary , we report on measurements of the @xmath1 fraction , @xmath50 at cleo and belle , and find about 25% of the @xmath2 decays produce @xmath1 , predominantly in the form of @xmath16 .
we also find a dominance of @xmath9 among ordinary @xmath0 decays . both of these observations are consistent with the uqm and inconsistent with the qhm .
an @xmath3 machine operating at the @xmath2 produces about 15 million @xmath1 per 100 @xmath22 .
studies of decay channels that produce soft photons could provide complementary information to @xmath1 studies at hadron machines .
s. ono , a. i. sanda , and n. a. tornqvist , phys . rev . *
d 34 * ( 1986 ) 186 ; n. tornqvist , phys .
lett * 53 * ( 1984 ) 87 ; k. heikkila , n. a. tornqvist and s. ono , phys . rev . * d 29 * ( 1984 ) 110 ; n. byers and e. eichten , nucl .
suppl . * 16 * ( 1990 ) 281 . | we present recent measurements of @xmath0 and @xmath1 production using data collected on the @xmath2 resonance at cleo and belle .
we also briefly discuss what can be learned using sufficiently larger data samples in the future . |
the quenching of large-@xmath1 hadrons in heavy - ion collisions is undoubtedly one of the most important discoveries obtained so far at rhic energies ( @xmath2 gev ) , suggesting significant parton energy loss in the dense qcd medium produced in those reactions @xcite .
it was therefore quite natural to wonder whether a similar phenomenon would be observed , either qualitatively or quantitatively , in the first pb
pb collisions obtained at lhc at a center - of - mass energy of @xmath3 tev . even though data at this energy were then desperately lacking , the first alice measurements @xcite soon indicated that the quenching of inclusive charged hadrons is indeed similar to what has been reported at rhic , at least in a large @xmath1 domain ( @xmath4 gev ) and within the experimental uncertainties .
these first lhc data have later been confirmed by the atlas @xcite , alice @xcite , cms @xcite preliminary measurements shown at this conference , which now extend up to @xmath5 gev .
the results obtained at the lhc together with the rhic measurements thus open up hope for a better understanding of energy loss processes through detailed phenomenological studies .
the quenching factor is defined as the normalized yield of particles in heavy - ion collisions ( in a given centrality class @xmath7 ) over the production cross section measured in collisions , @xmath8 where the thickness function @xmath9 is determined from a monte carlo glauber model .
the uncertainty on @xmath9 therefore translates into an overall normalization error on @xmath6 , which is of the order of 8% in central pb
pb collisions @xcite . assuming that the production of _ hard _ processes , i.e. with a large energy scale @xmath10 , scales with the number of binary nucleon - nucleon collisions @xmath11 , the quenching factor ( [ eq : defraa ] ) becomes @xmath12 in the absence of cold and hot medium effects . on the contrary
the yield of _ soft _ particles ( e.g. light hadron production at small @xmath1 ) is expected to be proportional to the number of participating nucleons in the collision , @xmath13 , leading to @xmath14 in central a a collisions .
this conventional assumption , not obvious _ a priori _ for soft radiation and proportional to @xmath13 for hard radiation , due to the lpm coherence effect @xcite , offering a counter - example .
] , remains to be checked experimentally for a variety of `` baseline '' processes insensitive to medium effects .
even though observing binary scaling for one hard process does not prove that this assumption is valid for another , at least this would give some confidence that this is indeed the case .
the quenching factor also requires the knowledge of the production cross section at the same collision energy , as is the case at rhic . at lhc
, various interpolation procedures using data at @xmath15 and 7 tev have been performed tev p
p run in 2011 . ] assuming a power - law behavior for the @xmath16-dependence of the cross section at fixed @xmath1@xcite or fixed @xmath17@xcite .
the latter is probably more appropriate @xcite since at large @xmath1 the invariant production cross section is expected to scale like @xmath18 .
moreover , the @xmath1 interpolation is problematic from a statistical point of view since this procedure is limited to the maximal @xmath1-bin available at the _ lower _ center - of - mass energy , for which phase - space is more restricted . in order to check how these two interpolation prescriptions differ , i compared the yields at @xmath19 gev at @xmath20 tev using the cms data within those two procedures .
remarkably , the yields differ by as much as @xmath21 , the @xmath1-interpolation leading to a yield _ smaller _ than the @xmath17-interpolation procedure . as a consequence , if data indeed follow @xmath17 scaling in the @xmath157 tev range , the measured yield at @xmath20 tev should be larger than the alice ( @xmath1-interpolated ) yield by @xmath22 , thus leading to @xmath6 values _ smaller _ by this amount .
this illustrates well the need for and a a measurements at the same energy .
the abundant production of hard processes in collisions at the lhc offers an outstanding opportunity for qcd studies .
processes with many particles in the final state allow for probing multi - leg processes , while more inclusive observables set tight constraints on parton distribution functions ( pdf ) and fragmentation functions ( ff ) but also serve as _ reference processes _ to which heavy - ion measurements can be compared .
remarkably , the first lhc measurements of prompt photons @xcite and jets @xcite proved in excellent agreement with qcd expectations . on the contrary , discrepancy between data and nlo calculations were reported in the hadron production channel @xcite .
the origin of this discrepancy is not clear yet , although it is worth stressing that ff are rather poorly constrained , especially in the gluon sector and at large momentum fraction @xmath23 .
higher - twist processes might also contribute substantially to hadron production at not too large @xmath1 @xcite .
it might be tempting to assume that the present uncertainties in the absolute cross sections are not an issue in heavy - ion collisions since these most likely cancel out in eq .
( [ eq : defraa ] ) . as we shall see later
, this is not necessarily the case , since the medium effects which affect @xmath6 actually depend on the dynamics of the hard process itself .
it is thus important to keep in mind such uncertainties when discussing the various phenomenological studies and their assumptions . in heavy - ion collisions ,
measurements are made more delicate because of the larger underlying event .
thanks to recent jet reconstruction techniques in high - multiplicity environments @xcite , it has become possible to measure the @xmath6 of jets , at rhic and now at lhc @xcite .
these results nicely supplement hadron production data , as they allow for tracing more efficiently the dynamics of parton energy loss and induced gluon radiation in a medium .
although energy loss processes are expected to play an important role in the quenching of hadrons and jets , other effects could actually affect @xmath6 : nuclear pdf ( npdf ) and saturation effects , multiple scattering in nuclei and recombination processes in qgp , to cite a few examples .
hopefully , most of these processes die out at transverse momenta at which the sole effects of parton energy loss remain visible .
it is e.g. the case of npdf effects which become small at large @xmath1 20 gev where @xmath24 gev at lhc .
] , below 20% above @xmath25 gev according to eps09 @xcite .
in order to disentangle these mechanisms , it is essential to compare systematically hadron / jet data to baseline measurements : * _ baseline experiments _ , typically deep inelastic scattering in nuclei ( @xmath26a ) or proton nucleus ( @xmath0a ) collisions .
these reactions allow for probing initial - state effects , such as npdf effects or parton propagation in a cold medium ; * _ baseline observables _ in heavy - ion collisions , such as prompt photon , drell - yan pair ( @xmath27 ) and @xmath28 production , which are expected to be insensitive to medium effects .
yet , they are essential since they offer a `` normalization '' of @xmath6 in central heavy - ion collisions and serve as tests of the binary collision scaling assumption . the need for baseline measurements is best exemplified at rhic , where @xmath6 of hadrons in @xmath29au collisions and that of photons in au au collisions strongly suggest that the hadron quenching at rhic is a genuine phenomenon ( i.e. not a normalization artifact ) due to a final - state effect . at lhc ,
the situation is extremely encouraging regarding the production of electroweak observables , as shown by the first @xmath30 and isolated photon measurements by cms @xcite . as discussed in @xcite , a dedicated @xmath0pb run at the lhc
is needed and hopefully will be carried out before shutdown in 2013 .
before discussing more specific models , i discuss in this section qualitative but hopefully model - independent features of parton energy loss on @xmath6 .
let us assume for simplicity that the rate of @xmath31 hadron production ) would be more appropriate for jets than for single - inclusive hadron production , but this is unimportant in this discussion .
] in a a collisions is given by @xmath32 where @xmath33 is the corresponding rate in collisions and @xmath34 the energy loss probability distribution .
taylor expanding @xmath33 in @xmath35 , the quenching factor reads to first order @xmath36 where @xmath37 is the probability for no energy loss .
the index @xmath38 is the power - law exponent of the rate , which is a scaling function of @xmath17 ( up to qcd scaling violations ) , almost independent of @xmath17 at small @xmath17 and getting larger as @xmath17 goes to 1 ( see fig . 3 of @xcite ) .
its value depends on the slope of the ff into hadrons , consequently ( i ) the uncertainties on the ff at large @xmath23 ( thus on @xmath38 ) will translate into an uncertainty on @xmath6 , ( ii ) a stronger suppression in the baryon channel , e.g. @xmath39 , is in principle expected since the spectra of baryons are softer than that of mesons . apart from the dependence of @xmath38 on @xmath17 , @xmath6 depends on the _ fractional _ energy loss , @xmath40 . the important lever arm in @xmath1 at the lhc
therefore allows for probing the energy dependence of @xmath41 .
let us illustrate this by assuming that @xmath42 and fixing @xmath43 gev@xmath44 , as suggested by the first lhc measurements . from eq .
( [ eq : raamodel ] ) and neglecting the @xmath1 dependence of @xmath38 , the `` predicted '' @xmath6 at @xmath45 gev would range from @xmath46 to @xmath47 when the parameter @xmath48 is varied from 0 to 1 leading to an effective exponent @xmath49 . ] .
provided @xmath50 , note that @xmath51 at asymptotic energies @xmath52 with @xmath53 fixed ,
i.e. the edge of phase - space is not reached . ] .
it would therefore be interesting to measure the values of @xmath54 for which energy loss effects are negligible , i.e. @xmath55 , for various centralities in order to constrain the amount of energy lost in the medium .
the large phase - space available at the lhc is also very satisfactory from a theoretical point of view as it allows for a better separation of the different energy scales involved in the process .
the `` ideal playground '' for energy loss studies could be summarized as @xmath56 .
the first inequality is likely to be fulfilled if the medium is dense or thick enough ; phenomenological estimates on the transport coefficient @xmath57 extracted at rhic indicate that it is actually the case .
more problematic is the fact that the typical energy loss is of the order of the energy of the hard parton , @xmath58 , leading on average to a full stopping of the propagating parton and therefore to strong surface bias effects @xcite .
this is a severe limitation in particular at rhic for which transverse momenta are presently restricted to @xmath59 gev .
the rightmost inequality is perhaps less crucial , yet it is likely to be responsible for the observed flatness of @xmath6 as a function of @xmath1 at rhic energies : as @xmath60 , the index @xmath61 gets larger and compensates the decrease of @xmath62 at large @xmath1 . at the lhc , this region is not probed unless hadrons are measured at forward rapidities , for which @xmath63 .
the hadron quenching in heavy - ion collisions has now been observed at two different colliding energies , offering a promising possibility to understand further the underlying phenomena from a systematic comparison of rhic and lhc .
one natural possibility is to compare @xmath6 at fixed @xmath1 and in the same centrality class , as the propagating partons would have typically the same energy and would cover the same medium length @xmath64 , despite the larger @xmath57 at lhc is expected to be proportional to the particle density , and @xmath65 @xcite . ] .
similarly one could also compare @xmath6 at fixed @xmath1 and particle density , leading to the same @xmath57 but different @xmath64 . finally ,
comparing data at fixed @xmath17 would be interesting since the partonic slopes and npdf / saturation effects would be identical , yet parton energies would be rather different .
there is a plethora of phenomenological studies on @xmath6 at rhic / lhc , mostly based on parton energy loss .
the strategy usually followed can be summarized as follows : 1 .
a general energy loss framework is chosen . to be schematic
, one usually distinguishes the approaches in which propagating partons experience soft multiple scattering ( bdmps - asw & amy ) to the ones in which only few hard scatterings take place ( dglv & higher - twist ( ht ) ) .
the bdmps - asw and dglv set - ups assume static scattering centers , while amy applies to thermal media .
in the ht approach the amount of energy loss is related to the strength of quark - gluon correlations in the medium .
additional work is then needed to treat _ multiple _ gluon emission , assuming independent ( poisson ) gluon emission ( bdmps - asw & dglv ) , solving rate equations ( amy ) or virtuality - ordered , dglap - like , evolution equations ( ht ) .
comparing these frameworks quantitatively has been the goal of the techqm collaboration over the last few years . for more details , see @xcite ; 2 . apart from radiative energy loss , other medium effects are sometimes implemented , e.g. collisional energy loss which can be important at not too large @xmath1 and for heavy - quarks or npdf effects ; 3 .
additional assumptions are made regarding the produced medium and its hydrodynamical evolution . for a long time
the medium expansion was described according to the simple bjorken model , i.e. a pure longitudinal expansion .
more recently , energy loss models have been implemented into more sophisticated 2d ( w / o viscous corrections ) or 3d hydro evolution . in most studies ,
the amount of energy loss is fixed at rhic and scaled up at the lhc according to the particle density . in fig .
[ fig : raadata ] are compared the preliminary cms data @xcite with various model predictions , either based on bdmps - asw ( pqm @xcite & renk @xcite ) or dglv ( glv @xcite & whdg @xcite ) frameworks .
the bands indicate the uncertainties coming from the initial particle density assumed in each model . apart from pqm ,
the various models predict a rising @xmath6 with @xmath1 in qualitative agreement with data . even though glv & whdg calculations best reproduce the magnitude of the data , no premature conclusion should be drawn at this stage regarding the other two predictions .
extracted from cms data.,height=257 ] extracted from cms data.,height=257 ] as mentioned previously , the @xmath1 _ slope _ of @xmath6 at the lhc offers an interesting possibility to probe energy loss processes . on the contrary ,
the _ magnitude _ of @xmath6 is useful in order to calibrate the strength of energy loss in the medium .
for this reason , it might be worth considering also the logarithmic derivative of @xmath6 , @xmath66 which quantifies the slope of @xmath6 with @xmath1 and which is free of normalization uncertainties .
this is illustrated in fig .
[ fig : lograadata ] , where @xmath67 has been extracted from both data and theory . ] .
remarkably , the theoretical uncertainties shrink since the bands in fig .
[ fig : raadata ] essentially reflect a normalization uncertainty on @xmath6 .
this variable is thus very much independent on the _ amount _ of energy loss but tests the dynamics of energy loss models . at large @xmath68 gev , data and models ( including pqm ) agree .
note also that @xmath67 is consistent with zero , indicating a flat behavior of @xmath6 in this @xmath1 domain . at lower @xmath1 , the model which describes best the data is whdg , while renk , glv and pqm tend to underpredict the slope of @xmath6 .
this discussion is not meant to be comprehensive , but is rather an illustration of the potential usefulness of quantifying the slope of @xmath6 through eq .
( [ eq : raaprime ] ) .
since @xmath67 is dominated by statistical uncertainties ( unlike @xmath6 at small @xmath1 ) , @xmath67 is expected to become more precise as more data are collected .
prompt photons have long been thought to be _ the _ ideal baseline measurement in heavy - ion collisions to which hadron production @xmath6 should be contrasted .
they can be used to probe nuclear pdf @xcite at not too large @xmath1 while at larger @xmath1 ( say , @xmath69 gev ) they offer a good test of binary collision scaling .
however , it soon became clear that various hot medium effects could actually affect photon production in nuclear collisions @xcite : * jet - to - photon conversion , in which a produced high - energy parton converts all of its momentum to a photon by a rescattering in the medium ; * photon bremsstrahlung induced by parton multiple scattering in the medium ; * quenching of the fragmentation component , similar to the quenching of hadrons due to energy loss effects .
both jet conversion and photon bremsstrahlung should _ enhance _ photon production leading to @xmath70 , @xmath6 should not be compared to unity since the neutrons in nuclei are less efficient to produce photons . ] and to negative second - fourier component , @xmath71 , in non - central collisions . on the contrary , one would expect @xmath72 and @xmath73 if fragmentation photons are quenched . also note that the last two effects would not show up when triggering on _ isolated _ photons , i.e. with little hadronic background in their vicinity . on the experimental side ,
phenix has reported long ago on preliminary measurements of inclusive photon production au
au collisions @xcite .
these data are consistent with no medium effects within the uncertainties , yet they can not be excluded .
hopefully the current situation will soon be clarified when final data come out .
star also reported on photon anisotropies @xmath74 at large @xmath1 @xcite . at the lhc ,
cms reported in this conference the first measurement of isolated photon production in heavy - ion collisions @xcite , which is a remarkable achievement .
the data are also consistent with @xmath75 although the present uncertainties are naturally rather large at this stage .
instead of a conclusion , let me mention some open questions related to `` jet quenching '' studies and energy loss processes .
regarding the theoretical frameworks themselves , efforts have been made to improve the current approximations , in particular including higher - order corrections and going beyond the soft and collinear limit in the calculation of medium - induced gluon spectrum @xcite ( see also @xcite and references therein ) . on a more phenomenological side
, it would be important to clarify the role of collisional ( in contrast to radiative ) energy loss effects , in light of the recent lhc data .
one surprise of the rhic data was the absence of proton quenching at rather small @xmath1 , whereas in pure energy loss scenarios one would expect a similar ( if not stronger ) quenching as that of mesons .
it would be particularly interesting to observe the quenching of hadrons according to their ff slopes , a phenomenon observed e.g. in dis on nuclei @xcite .
the heavy - quark sector should also allow for a critical test of energy loss processes and the possible dead cone hierarchy , @xmath76 @xcite , which seems _ a priori _ challenged by the alice @xmath77-meson @xmath6 @xcite and by theoretical arguments @xcite .
the double ratio @xmath78 vs. @xmath1 should e.g. set stringent constraints on energy loss models .
another challenge will be disentangle energy loss from possible saturation effects .
the fact that the hadron quenching at mid - rapidity is similar at lhc than at rhic , despite @xmath17 values smaller by an order of magnitude , questions the strength of such saturation effects .
hopefully future @xmath0pb data at lhc may soon clarify this question . without a doubt , the large-@xmath1 hadron and photon measurements at the lhc presented at this conference as well as that of rhic with unprecedented luminosities are the first fruits of what will be another exciting decade of tomographic studies .
10 denterria d 2009 _ springer verlag . landolt - boernstein vol . 1 - 23a _
{ http://arxiv.org/abs/}0902.2011[`0902.2011 ` ] arleo f and gousset t 2008 _ phys .
lett . _ * b660 * 181 { http://arxiv.org/abs/}0707.2944[`0707.2944 ` ] ; arleo f , eskola k j , paukkunen h et al .
2011 _ jhep _ * 04 * 055 { http://arxiv.org/abs/}1103.1471[`1103.1471 ` ] | the generic features of parton energy loss effects on the quenching of single hadron spectra in heavy - ion collisions are discussed , paying attention to the expected differences from rhic to lhc .
the need for precise baseline measurements in @xmath0pb collisions at the lhc is also emphasized .
finally i briefly mention the production of prompt photons in heavy - ion collisions as well as some open questions . |
at a distance of @xmath4 kpc @xcite , the center of the milky way provides an excellent laboratory for high resolution studies of galactic nuclei . images of radio continuum emission from the region are dominated by a strong point source , sgr a * , which is most likely a @xmath5 m@xmath6 black hole located at the dynamical center of the galaxy @xcite .
sgr a * is surrounded by a `` mini - spiral '' composed of arcs of ionized gas that are falling towards and orbiting the black hole @xcite .
a ring of molecular material at a radius of 2 pc from the galactic center surrounds the mini - spiral and is referred to as the circumnuclear disk ( cnd ) @xcite .
the name was inspired by early observations of hcn and hco+ which showed signatures of a single rotating disk with the center swept clear of material @xcite .
the western arc of the mini - spiral traces the ionized inner edge of the cnd , while the northern and eastern arms , which extend towards sgr a * , may be the result of cloud - cloud collisions in the cnd @xcite .
however , with the recent increase in sensitivity and resolution of instruments , the morphology and kinematics of the gas near the galactic center now appear to be much more complicated than originally thought .
the cnd is composed of many clumps with an average density of @xmath7 @xmath8 .
wide - field interferometric mosaics of the region in hcn(1 - 0 ) and hco+(1 - 0 ) @xcite show that the cnd has a distinct outer edge at @xmath9 from the nucleus , making it resemble a ring more than a disk .
@xcite observe that the northeast and southwest lobes of the cnd appear to have different angles of inclination .
while the northeastern sections of the cnd are best fit by an inclination of 70 - 75@xmath10 , the southwestern lobe of the cnd seems to be at an inclination of @xmath11 .
this observation led @xcite to the conclusion that the cnd is warped .
the southeastern side of the cnd is weak in many tracers ( including hcn and hco+ ) suggesting that this part of the `` ring '' may be missing altogether .
@xcite propose that the two bright lobes of the cnd are in fact _ distinct _ clouds on independent orbits around the nucleus and do not compose a single ring of gas .
@xcite find kinematic evidence that the cnd is composed of at least three distinct clouds which are orbiting the black hole .
in addition , the northern arm of the mini - spiral actually appears to originate outside the cnd and cross through on its way towards the galactic center @xcite .
the idea that the cnd and the mini - spiral are composed of multiple distinct clouds under the gravitational influence of the nucleus easily explains the observed `` warping '' as well as the complex kinematics and clumpy morphology observed in the region @xcite .
past observations of the central regions of the galaxy also provide insight into questions about the origin of the cnd and the clouds which compose it .
if the cnd is composed of many distinct clouds , then it should be possible to trace where the clouds come from and determine the origin of the cnd .
many attempts have been made to detect connections between two nearby giant molecular clouds ( gmcs ) and the cnd .
@xcite , @xcite , and @xcite detect a long filamentary `` streamer '' in nh@xmath0 ( 1,1 ) and ( 2,2 ) emission that connects the `` 20 km s@xmath2 cloud '' ( m-0.130.08 , @xcite ) to the southeastern edge of the cnd .
a small velocity gradient along this `` southern streamer '' as well as heating and increased line widths as the streamer approaches the galactic center indicate that gas is flowing from the 20 km s@xmath2 cloud towards the circumnuclear region .
this connection has also been observed in hcn(3 - 2 ) @xcite , @xmath12co(2 - 1 ) @xcite and 1.1 mm dust @xcite .
an extension of the 20 km s@xmath2 cloud towards the southwest lobe of the cnd was also seen in nh@xmath0 ( 1,1 ) and ( 2,2 ) . due to the lack of pointings to the west of the cnd ,
@xcite could not search for a connection between this extension and the cnd .
however , the morphological connection between the 20 km s@xmath2 cloud and the southwest lobe of the cnd is observed in dust @xcite .
other signs of connections between the gmcs and the cnd include a possible connection between the northeastern edge of the cnd and the `` 50 km s@xmath2 cloud '' ( m-0.030.07 , @xcite ) observed in hcn(1 - 0 ) @xcite .
we present maps of the central 10 parsecs of the galaxy in the ( 1,1 ) , ( 2,2 ) , and ( 3,3 ) transitions of nh@xmath0 .
unlike previous nh@xmath0 observations , our new data cover the entire velocity range seen in molecules at the galactic center and have improved spatial sampling which provide a more complete picture of nh@xmath0 in the central 10 pc of the galaxy and enable a more reliable comparison of nh@xmath0 emission to other tracers . with this data , we can probe the morphology and kinematics of all of the cnd as well as the surrounding molecular material .
paper i focuses on the general morphology of velocity integrated nh@xmath0 emission from the region , including comparisons to hcn(1 - 0 ) , hcn(3 - 2 ) , 1.2 mm dust , and 6 cm continuum emission .
new features to the north and west of the cnd are observed , producing a complete picture of nh@xmath0 ( 1,1 ) , ( 2,2 ) , and ( 3,3 ) emission in the central 10 parsecs .
at least three physical connections between the gmcs and the cnd are observed , with kinematic information presented for these features .
a detailed discussion of kinematics of the entire central 10 pc , including temperatures , opacities , and cloud masses , will be presented in a following paper .
we observed the ( 1,1 ) , ( 2,2 ) , and ( 3,3 ) rotation inversion transitions of nh@xmath0 , at 23.694495 , 23.722633 , and 23.870129 ghz , respectively , with the national radio astronomy observatory s very large array ( vla ) telescope in 1999 march .
observations were made in the d - northc array which provides the shortest baselines and the most circular synthesized beam possible for this low declination source ( @xmath13 ) .
a five - pointing mosaic was used to fully sample the central 4@xmath14 ( 10 pc ) of the galaxy .
the full width at half maximum ( fwhm ) of the primary beam is 125@xmath3 at 23 ghz for the 25 m vla antennas . to ensure the field
was properly sampled , one pointing was centered on sgr a * ( @xmath1517@xmath1645@xmath1740@xmath18 , @xmath19 ) while the other four pointings were offset by 1@xmath14 to the northeast , northwest , southeast , and southwest .
at least one hour on source was obtained for each pointing at each frequency .
the data consist of 31 channels with a width of 9.8 km s@xmath2 ( 0.78 mhz ) centered on a @xmath20 of 10.6 km s@xmath2 and covering @xmath21 km s@xmath2 .
coil & ho ( 1999,2000 ) imaged the circumnuclear region in nh@xmath0 ( 1,1 ) and ( 2,2 ) with a total velocity coverage of @xmath22 km s@xmath2 and a velocity resolution of 5 km s@xmath2 . in their mosaics ,
fields on the cnd were centered on @xmath23 km s@xmath2 while fields to the south and east of the cnd were centered on @xmath24 km s@xmath2 .
the narrow velocity coverage excluded high velocity emission which is seen up to @xmath25 110 km s@xmath2 in the cnd .
in addition , @xcite focused on the southern streamer and the 50 km s@xmath2 cloud and had no pointings to the north or west of the cnd .
the images presented in this paper represent the first time that the ( 1,1 ) , ( 2,2 ) , and ( 3,3 ) transitions of nh@xmath0 have been fully sampled over the central 4@xmath14 with a velocity coverage that includes all of the emission from the cnd .
each pointing was calibrated separately using nrao s astronomical imaging processing system ( @xmath26 ) .
anomalously high @xmath27 points and large closure errors were removed .
calibration of the amplitudes and phases were performed using 1331 + 305 ( 3c286 ) and 1733130 . since sgr a * is a strong source at 23 ghz and within the primary beam of each pointing , self - calibration was successfully applied to each of the fields .
the calibrator 1229 + 020 was used for bandpass calibration to correct instrumental phase offsets across the band and to remove the roll - off in sensitivity at the edge of the band .
channels 1 , 30 , and 31 could not be fully corrected and were removed from the dataset resulting in a velocity coverage of 140 to + 130 km s@xmath2 .
once the data were calibrated , line - free channels were used to model the continuum emission and remove it from the data .
the fully calibrated , continuum - free data cubes were imaged in miriad @xcite , which offers superior tools for making mosaics and analyzing the spectral information .
the five pointings were combined into a single mosaic in the @xmath27 plane . for the maps presented in this paper
, we used the maximum entropy method ( mem ) to deconvolve the images .
previous nh@xmath0 mosaics were deconvolved using the clean algorithm in which the field is assumed to be composed of points sources . as a result ,
negative bowls surround bright extended features , possibly obscuring faint connections between clouds ( see @xcite ) .
mem is better suited to handle fields of extended emission and does not produce these negative bowls .
however , mem does not conserve the integrated flux density in the map . to check our deconvolution , we deconvolved the data using both clean and mem .
the two images agree on every main feature with the only difference in the low - level emission surrounding bright features .
the integrated flux densities in both maps are also equal indicating that mem did not add substantial flux to the image . in order to recover the low - level emission in the map ,
the mosaic was imaged using a gaussian taper applied to the @xmath27 data resulting in a synthesized beam of @xmath28 at a position angle of @xmath29 in all three maps .
the nh@xmath0 ( 1,1 ) , ( 2,2 ) , and ( 3,3 ) velocity integrated ( 270 km s@xmath2 ) images are shown in figures [ 11.fig ] , [ 22.fig ] , and [ 33.fig ] . in all three images , the approximate location of the cnd @xcite is represented by two ellipses and the location of sgr a * is marked by a cross . if a straight summation of each channel map is used to create the velocity integrated image , emission from narrow - line features , which appear in only a small fraction of the channels , is diluted by the increased noise in the summed map . in order to avoid this problem and
present a map showing all of the emission in the region , each channel map was clipped below @xmath30 before the channels were summed .
this procedure was performed in aips using the iflux parameter in momnt .
the resulting image had features on the size scale of single pixels as a result of the blanking process . to remove these artifacts ,
the image was convolved with a gaussian beam with fwhm=@xmath31 .
since the gaussian was significantly narrower than the image beam , the final beam size for our images was only slightly increased to @xmath32 .
if the noise in one channel is known , then the noise in the velocity integrated image ( in jy beam@xmath2 km s@xmath2 ) is simply @xmath33 , where @xmath34 is the number of channels in the summation , @xmath35 is the channel width in km s@xmath2 , and @xmath36 is the noise in one channel . since the maps were clipped below @xmath37 in each channel , @xmath34 varies for each pixel in our maps .
however , we assume most features show emission in @xmath38 of the channels and adopt @xmath39 , where @xmath40 km s@xmath2 and @xmath36 is measured in the individual channels .
the resulting noise in our images is @xmath41 jy beam@xmath2 km s@xmath2 , @xmath42 jy beam@xmath2 km s@xmath2 , and @xmath43 jy beam@xmath2 km s@xmath2 for the nh@xmath0 ( 1,1 ) , ( 2,2 ) and ( 3,3 ) velocity integrated images , respectively . for features with emission in a very narrow velocity range ,
the signal - to - noise ( s / n ) will be _ underestimated _ in our map by a factor of @xmath44 where @xmath45 is the true number of channels in which the feature appeared .
thus , a feature with emission in only one channel will have @xmath46 and the s / n underestimated by a factor of 2.6 . the s / n of broad velocity features is _ overestimated _ in our images by a factor of @xmath47 and the s /
n for a feature with emission in all channels is overestimated by a factor of 2 .
there are few examples of these extremes in our data and the contours presented here are a reliable indicator of the significance of the detections .
the additional complication in the noise analysis is well worth the benefits of being able to see all of the emission from the galactic center without losing narrow - line features due to the increased noise of a straight summed map . for reference
, we also present the nh@xmath0 ( 3,3 ) channel maps in figure [ chan.fig ] .
no clipping has been performed on these maps and the noise should be constant at all pixels ( ignoring any effect of mem deconvolution ) .
the maps were not convolved by the @xmath48 gaussian beam and have a resolution of @xmath49 .
narrow - line features such as the northern ridge are clearly visible .
in addition to the velocity integrated images , a line ratio map of nh@xmath0 ( 2,2 ) to ( 1,1 ) emission is shown in figure [ temp.fig ] .
line ratios of the nh@xmath0 transitions are reliable indicators of gas temperatures @xcite and can be helpful in understanding conditions near the galactic center . only points with nh@xmath0(1,1 ) emission @xmath50 and nh@xmath0(2,2 ) emission @xmath51 are included in the map , and the single contour marks a @xmath52 detection of nh@xmath0(2,2 ) emission .
thus , regions _ inside _ this contour and with _ no _ greyscale were excluded on the basis of low nh@xmath0(1,1 ) brightness and may be very hot with high line ratios .
gas at the location of the cnd is seen in all three velocity integrated images although it is best traced by nh@xmath0 ( 3,3 ) .
nh@xmath0(3,3 ) is expected to be brighter than nh@xmath0(1,1 ) and ( 2,2 ) by a factor of 2 due to a quantum mechanical degeneracy of the states in the ortho species of nh@xmath0 @xcite .
the southern half of the cnd is much stronger in nh@xmath0 ( 2,2 ) than ( 1,1 ) and figure [ temp.fig ] shows the increased line ratio in the region .
assuming the cnd is optically thin in nh@xmath0 ( 1,1 ) and ( 2,2 ) ( consistent with the results of @xcite ) , the observed ratio of nh@xmath0 ( 2,2 ) to ( 1,1 ) emission of 1.5 implies a rotational temperature of @xmath5350 k and a kinetic temperature of 60 k in the southeastern half of the cnd @xcite .
this is hotter than material farther from the nucleus .
emission to the south and east of the cnd has a typical line ratio @xmath54 indicating a rotational temperature of no more than @xmath55 k for gas more than 2 pc to the south or east of the galactic center the nh@xmath0 images contain many long filamentary features which we refer to as streamers .
the `` southern streamer '' @xcite , labeled in figure [ 33.fig ] , can be seen beginning at @xmath56 moving to the northeast to @xmath57 , @xmath58 and intersecting the cnd at @xmath59 .
two clouds , previously observed by @xcite , are located to the east of the southern streamer and are labeled as `` se1 '' and `` se2 '' .
these clouds appear to connect to the southern streamer in all three images .
the bright , `` y''-shaped cloud in the northeast centered on @xmath60 is the `` 50 km s@xmath2 ''
giant molecular cloud which is being impacted on the west by sgr a east ( see section 4.2 ) .
emission from se1 , se2 , and the southern streamer connect the 20 km s@xmath2 cloud in the south to the 50 km s@xmath2 cloud in the east .
in fact , se2 is the southern half of the `` molecular ridge '' discussed by @xcite .
the detection of se1 is a good example of the improved quality of our new images compared to previous nh@xmath0 mosaics .
although the new data still show mainly filamentary structures , there are more faint connections between clouds than in previous images by @xcite .
only faint emission was seen from se1 in @xcite , while we find this feature to be as strong as se2 .
our ability to detect se1 reflects the lack of negative bowls near bright structures when mem is used to deconvolve the images , whereas only clean was used by @xcite . our improved spatial sampling and velocity coverage as well as
the use of mem deconvolution has resulted in more robust nh@xmath0 images . a new ridge of nh@xmath0(3,3 )
emission is seen in figure [ 33.fig ] to the north of the cnd from @xmath61 , @xmath62 to @xmath63 , @xmath64 and lies along the northern edge of sgr a east . in the nh@xmath0(1,1 ) and ( 3,3 ) images , this `` northern ridge '' is connected to the cnd at @xmath65 , @xmath66 by a narrow streamer .
the northern ridge was not in the field of view of the @xcite mosaic .
finally , there is a long `` western streamer '' running approximately north - south from @xmath67 to @xmath68 .
multiple faint connections between the western streamer and the cnd are seen at @xmath69 , @xmath70 , @xmath71 , @xmath72 , and @xmath73 , @xmath74 and are discussed in more detail in section 5 .
the large velocity coverage and uniform spatial sampling of our nh@xmath0 images present a unique opportunity to compare nh@xmath0 emission in the galactic center to emission from other tracers . in the following sections , nh@xmath0 ( 3,3 ) emission
is compared to hcn(1 - 0 ) , hcn(3 - 2 ) , 1.2 mm dust , and 6 cm continuum emission .
comparison of nh@xmath0(3,3 ) to hcn(@xmath751 - 0 ) has proved difficult with the two molecules apparently tracing different features at the galactic center @xcite .
however , the images that were compared were often unevenly sampled or had very different velocity coverage and resolution .
figure [ hcn.fig ] shows the velocity integrated nh@xmath0 ( 3,3 ) image in contours overlaid on a hcn(1 - 0 ) mosaic of the central 12.5 pc by @xcite .
the nh@xmath0(3,3 ) image has been convolved with a gaussian taper to a resolution of @xmath76 and the hcn(1 - 0 ) image has a resolution of @xmath77 .
although the nh@xmath0 mosaic was made from vla data and the hcn(1 - 0 ) image was made from a 19-pointing bima mosaic , these two images are similar in their velocity and spatial coverage .
the nh@xmath0(3,3 ) image is fully sampled out to 4@xmath14 with a velocity coverage of 140 km s@xmath2 to + 130 km s@xmath2 while the hcn(1 - 0 ) image is fully sampled out to 5@xmath14 and has a slightly broader velocity coverage of @xmath78 km s@xmath2 .
the velocity resolution of the hcn(1 - 0 ) image is 5 km s@xmath2 whereas the nh@xmath0 resolution is 10 km s@xmath2 .
the high quality of these two datasets enables the most robust comparison of emission from nh@xmath0(3,3 ) and hcn(1 - 0 ) .
the ring - like structure of the cnd is clearly seen as the backwards `` c '' in the center of the hcn(1 - 0 ) image .
both molecules show peaks in the northeast and southwest of the cnd , but the peaks are not in the same location . although the relative intensities differ , there is emission from both molecules around the entire ring except in the southeast where there is a conspicuous lack of emission .
a narrow , north - south gap in the northern part of the cnd is observable in hcn(1 - 0 ) , and there is a distinct edge in the nh@xmath0(3,3 ) emission coincident with the edge of the gap . while the hcn(1 - 0 ) emission appears to have been pulled into a long filament at the gap ( @xmath79 , @xmath80 to @xmath81 ) , the nh@xmath0 does not appear to be as extended towards the north .
however , the resolution of the nh@xmath0 maps makes it difficult to trace small scale features which can be seen in the hcn(1 - 0 ) image .
this gap may be the result of the northern arm crossing through the cnd ( see section 4.3 ) .
kinematic information must be used to determine whether nh@xmath0(3,3 ) is tracing the same features of the cnd observed in hcn(1 - 0 ) .
figure [ specpos ] shows the location of 12 spectra near the cnd .
the corresponding spectra are shown in figure [ spec ] .
spectra a - j are at the same locations as the hcn(1 - 0 ) spectra in @xcite ; spectra k and l are located at two bright nh@xmath0 features .
spectra for these two features were not shown in @xcite , but we include them here for comparison to nh@xmath0 .
the nh@xmath0(3,3 ) spectra are shown in thick lines with the hcn(1 - 0 ) spectra shown in thin lines , scaled down by a factor of 8 .
spectrum a is centered on sgr a*. hcn(1 - 0 ) shows significant absorption from known clouds along the line - of - sight ( @xcite and references therein ) .
nh@xmath0(3,3 ) shows some weak absorption at 0 km s@xmath2 , but also some emission near 40 km s@xmath2 .
less absorption along the line - of - sight in nh@xmath0(3,3 ) is due to the low optical depth @xmath82 of nh@xmath0(3,3 ) compared to hcn(1 - 0 ) which is optically thick over most of the region @xcite .
spectra e and g show features at the same velocity in nh@xmath0(3,3 ) and hcn(1 - 0 ) . in spectrum
e , both nh@xmath0(3,3 ) and hcn(1 - 0 ) show a feature at 70 km s@xmath2 , but hcn(1 - 0 ) shows an additional feature at 60 km s@xmath2 . in spectrum g , there are hcn(1 - 0 ) and nh@xmath0 features at both + 50 and + 80 km s@xmath2 .
although the dip in the profile looks like it could be the result of self - absorption , position - velocity cuts in nh@xmath0(3,3 ) through the western side of the cnd indicate that this spectrum is the result of the superposition of two clouds along the line - of - sight . in the spectra that agree kinematically , the relative intensities of the lines are often different .
nh@xmath0(3,3 ) is especially weak in high velocity gas . in hcn(1 - 0 ) , spectrum i is dominated by a single feature at 110 km s@xmath2 .
this feature is seen in nh@xmath0(3,3 ) , but it is weak and surrounded by faint features at 20 , 20 , and 50 km s@xmath2 . in spectrum c ,
there is bright hcn(1 - 0 ) emission at + 100 km s@xmath2 , but there is no significant detection of this feature in nh@xmath0(3,3 ) .
several nh@xmath0(3,3 ) spectra have a peak velocity 10 - 30 km s@xmath2 less than hcn(1 - 0 ) .
spectrum b peaks at 60 km s@xmath2 in nh@xmath0(3,3 ) and 75 km s@xmath2 in hcn(1 - 0 ) .
spectrum f shows a similar difference and peaks at 50 km s@xmath2 in nh@xmath0(3,3 ) and 60 km s@xmath2 in hcn(1 - 0 ) . in spectrum d ,
the nh@xmath0(3,3 ) emission is 30 km s@xmath2 lower than the hcn(1 - 0 ) emission .
while the hcn(1 - 0 ) emission peaks at 70 km s@xmath2 , we see the nh@xmath0 peak at @xmath83 km s@xmath2 . in @xcite , spectrum b
is nearly crossed by position - velocity cut @xmath84 ( see figure 13@xmath84 , position 25@xmath3 , @xcite ) . in this position - velocity diagram ,
absorption is seen at nearby positions from 60 to 80 km s@xmath2 .
this absorption may affect spectrum b , causing the profile of this spectrum to have a sharp cut - off below 80 km s@xmath2 .
if the low velocity emission has been absorbed , then the true distribution of hcn(1 - 0 ) emission may actually be centered at a lower velocity and agree with the peak velocity of 60 km s@xmath2 seen in nh@xmath0(3,3 ) emission .
similar indications of absorption are seen in position velocity cuts of hcn(1 - 0 ) emission near spectra d and f. in hcn(1 - 0 ) , spectrum d appears to be affected by absorption at 60 km s@xmath2 ( see figure 13@xmath85 , position 20@xmath3 , @xcite ) .
spectrum f has an asymmetric profile in hcn(1 - 0 ) .
although this may be the result of a shock , it could also be attenuated at low velocities by hcn(1 - 0 ) absorption .
there is absorption in hcn(1 - 0 ) from 60 to + 50 km s@xmath2 near the position of spectrum f ( see figure 13@xmath85 , position 50@xmath3 , @xcite ) .
spectra h and j indicate self - absorption of hcn(1 - 0 ) . in @xcite
, many hcn(1 - 0 ) features showed evidence for self - absorption , especially in the position - velocity diagrams . in spectrum h , nh@xmath0(3,3 ) appears to peak between the hcn(1 - 0 ) emission at + 20 and 40 km s@xmath2 indicating that the feature at 10 km s@xmath2 may have been self - absorbed in hcn(1 - 0 ) leaving only the wings of the emission . in spectrum
j , the two hcn(1 - 0 ) peaks at 40 and 70 km s@xmath2 could be the wings of a self - absorbed spectrum
. some of the brightest nh@xmath0(3,3 ) emission from the cnd is seen in spectra k and l. there is little hcn(1 - 0 ) emission in k , while in l the hcn(1 - 0 ) is shifted to a higher velocity .
the shift in l is similar to spectrum b which is near the same location in the cnd . in nh@xmath0(3,3 ) , both k and l have a very narrow profile in nh@xmath0 with a fwhm of @xmath86 km s@xmath2 .
these profiles are much narrower than all other spectra in the cnd .
the general agreement of the nh@xmath0(3,3 ) and hcn(1 - 0 ) spectra , especially when absorption along the line - of - sight and self - absorption are considered , shows that nh@xmath0(3,3 ) traces the main features of the cnd .
the complex spectra in both molecules are an indication of the complicated gas kinematics in the cnd .
when the nh@xmath0(3,3 ) data are considered , it becomes even more evident that the cnd is not a coherent ring , but rather is composed of many distinct features .
the difference in relative intensity of nh@xmath0(3,3 ) and hcn(1 - 0 ) may be explained by opacity or a variety of chemical effects including evaporation of nh@xmath0 off of dust grains ( see @xcite for a more detailed discussion ) .
it is also important to compare nh@xmath0(3,3 ) and hcn(1 - 0 ) emission outside the cnd .
overall , the hcn(1 - 0 ) appears to have a more diffuse distribution with faint emission seen over most of the field .
an exception is the northwest quadrant of the image which has little emission from either molecule .
this region is also empty in dust ( see figure [ pbcor ] ) .
there is emission from the 50 km s@xmath2 cloud in hcn(1 - 0 ) and nh@xmath0 although it is concentrated in different places .
hcn(1 - 0 ) peaks to the north of the nh@xmath0 cloud with the strongest emission at the connection between the 50 km s@xmath2 cloud and the northern ridge at @xmath87 , @xmath88 .
the most notable difference in this region of the 50 km s@xmath2 cloud is the `` hook '' of hcn(1 - 0 ) emission which extends past the edge of the nh@xmath0 emission to @xmath89 , @xmath90 .
this location is interior to the edge of sgr a east in projection and has a velocity of 80 km s@xmath2 in hcn(1 - 0 ) .
the southern streamer is weak in hcn(1 - 0 ) .
there is almost no hcn(1 - 0 ) emission at the peak of the nh@xmath0 emission at @xmath91 , @xmath92 .
some hcn(1 - 0 ) emission is coincident with se1 and se2 , but the features are not as connected as in nh@xmath0 . the southern third of the western streamer is seen in both hcn(1 - 0 ) and nh@xmath0 .
this feature is the high velocity ( 110 km s@xmath2 ) `` extension '' of the southwest lobe of the cnd seen by @xcite and @xcite .
the northern two - thirds of the streamer may show some faint hcn(1 - 0 ) emission , but they do not form a single feature .
the lack of hcn(1 - 0 ) emission from the brightest nh@xmath0(3,3 ) features is due in part to self - absorption of hcn(1 - 0 ) in the gmcs as well as absorption of hcn(1 - 0 ) along the line of sight @xcite .
position - velocity cuts by @xcite along the southern streamer and 50 km s@xmath2 cloud ( see figure 13 in @xcite ) show strong self - absorption of hcn(1 - 0 ) in these features .
self - absorption of hcn(1 - 0 ) could weaken the intensity of the gmcs and result in apparently enhanced emission from the cnd , where wide line widths make hcn(1 - 0 ) less susceptible to self - absorption .
the nh@xmath0 is less affected by absorption because the transitions are of higher energies and are more optically thin than the hcn(1 - 0 ) lines .
observations of the higher hcn ( @xmath753 - 2 ) transition by @xcite show good agreement with nh@xmath0(3,3 ) and support the idea that hcn(1 - 0 ) is highly affected by self - absorption and absorption along the line - of - sight .
channel maps by @xcite of hcn(3 - 2 ) emission show strong emission at the location of both the 20 and 50 km s@xmath2 clouds in addition to the cnd ( see figure 2 , @xcite ) .
the same figure also shows features to the west of the cnd that agree kinematically and spatially with nh@xmath0 emission from the western streamer .
emission to the southwest begins at 110 km s@xmath2 and continues to the north where it reaches positive velocities of + 50 km s@xmath2 .
the extension of the 20 km s@xmath2 cloud towards the southwestern lobe of the cnd , seen at @xmath71 , @xmath93 in figure [ 33.fig ] , is observed at 10 km s@xmath2 .
emission from the southern streamer is seen in the 30 km s@xmath2 channel and is just as strong as emission from the cnd .
the 50 km s@xmath2 and 20 km s@xmath2 gmcs also dominate the channel maps that correspond to their central velocities .
the northern ridge is easily observable in hcn(3 - 2 ) as the ridge to the north of the cnd in the 10 km s@xmath2 channel of @xcite . in conclusion , the nh@xmath0(3,3 ) and hcn(3 - 2 ) images agree on every main feature including the southern streamer , northern ridge , and western streamer which are completely absent in the hcn(1 - 0 ) image .
this indicates that images made with hcn(1 - 0 ) must be carefully studied to determine the effects of absorption and excitation .
the colorscale in figure [ pbcor ] shows a 1.2 mm continuum image taken by @xcite using the iram 30 m telescope .
this 1.2 mm image traces thermal dust emission in the region with additional free - free emission in the mini - spiral .
overlaid on the continuum emission is a primary beam corrected nh@xmath0 ( 3,3 ) velocity integrated map .
the primary beam correction was performed by dividing the velocity integrated ( 3,3 ) emission by the gain of the interferometer at each point in the map .
only points with a gain @xmath9415% are plotted in figure [ pbcor ] .
primary beam correction results in a map with varying noise characteristics across the field .
although the rms noise of this image near the field center is still @xmath530.33 jy beam@xmath2 km s@xmath2 , the noise is scaled up by the reciprocal of the gain and is @xmath532.2 jy beam@xmath2 km s@xmath2 at the map edge .
there is a striking correspondence between the thermal dust and nh@xmath0 ( 3,3 ) emission .
the southern streamer as well as se1 and se2 are clearly visible in both nh@xmath0 ( 3,3 ) and 1.2 mm dust .
the 20 km s@xmath2 gmc is the bright dust feature at @xmath95 , @xmath96 and is well traced by nh@xmath0 ( 3,3 ) to the edge of the mosaic .
the 50 km s@xmath2 cloud can be seen as the bright dust emission to the northeast from @xmath97 , @xmath90 to @xmath97 , @xmath98 .
this gmc is also close to the edge of the nh@xmath0 image and can be seen as the bright nh@xmath0(3,3 ) feature on the northeastern edge of the mosaic .
there is also striking agreement along the `` northern ridge . ''
this feature is clearly present in the dust image from @xmath63 , @xmath98 to @xmath87 , @xmath99 where it intersects the 50 km s@xmath2 cloud .
both tracers show little or no emission in the northwest quadrant of the image .
there is also a lack of nh@xmath0(3,3 ) and thermal dust emission in the cavity of sgr a east covering @xmath100 to @xmath63 and @xmath101 to @xmath102 .
the degree of correlation between nh@xmath0 and dust emission in the cnd is difficult to determine due to the free - free emission that dominates the dust image near sgr a west .
comparison of the 1.2 mm dust map to a 350 @xmath103 m map by @xcite shows that the dust emission is approximately constant over this wavelength range .
the similarity of the two maps implies that the 1.2 mm dust traces column density and is not strongly correlated with temperature in the region .
the dust image has a beam size of @xmath5311@xmath3 , and as a single dish map it detects highly extended emission . considering the lack of zero - spacing information in the nh@xmath0 data ( shortest baseline 35 m ) , the agreement between the two maps is compelling , suggesting that nh@xmath0 is a relatively unbiased tracer of column density in the galactic center region . a few features seen in the nh@xmath0(3,3 )
image are not observed in the thermal dust image .
the `` western streamer '' is not strongly correlated with the dust , and there is almost no dust emission in the upper two - thirds of this streamer .
line ratios of more than 1 in this region ( see figure [ temp.fig ] ) indicate that the western streamer is heated to @xmath104 k. the narrow stream of gas connecting the northern ridge to the cnd is also not strongly correlated with the dust and shows similar hints of high line ratios in figure [ temp.fig ] .
the heating and absence of dust in these features indicate that they may be located physically closer to sgr a * or may originate in a different way than the streamers that contain dust .
the velocity integrated nh@xmath0 ( 3,3 ) emission is overlaid on 6 cm continuum emission @xcite in figure [ sgeast.fig ] .
the arms of the sgr a west mini - spiral can be seen in blue in the center of the image .
the eastern edge of sgr a east extends to @xmath105 and the western edge is spatially coincident with the western edge of the cnd and mini - spiral .
sgr a east is expanding with an energy more than an order of magnitude greater than a typical supernova remnant @xcite . the nh@xmath0 ( 3,3 )
emission closely follows the outer edge of sgr a east .
the shell is impacting the 50 km s@xmath2 cloud in the east @xcite where material forms a `` ridge '' on the western edge of the cloud .
this ridge is seen in nh@xmath0(3,3 ) from @xmath97 , @xmath90 to @xmath87 , @xmath98 .
the northern ridge and western streamer also lie along the outside edge of sgr a east .
therefore , it appears that sgr a east is expanding into material on all sides of the cnd .
we observe multiple connections between the cnd and the ridges on the outside edge of sgr a east ( see section 5 ) .
these connections are a strong indication that the cnd and sgr a east are in close proximity to each other .
the southern edge of sgr a east is approximately coincident with the two southern clouds and the southern streamer . since the `` 20 km
s@xmath2 '' giant molecular cloud ( which appears to be the source of at least the southern streamer @xcite ) is thought to be located in front of sgr a east @xcite , sgr a east may be expanding into the back of these filaments .
in addition , there is evidence that a snr centered at @xmath106 ( * ? ? ? * and references therein ) may be impacting sgr a east along its southeastern edge .
this could account for the 1720 mhz oh masers detected by @xcite on the boundary between this snr and sgr a east .
the interaction of these two expanding shells could also produce filamentary structures like se1 , se2 , and the southern streamer .
figure [ sgeast.fig ] is also useful for studying the gap in nh@xmath0 ( 3,3 ) emission in the northern part of the cnd at the location of the northern arm of the mini - spiral .
as discussed in section 4.1 , this gap is also seen in hcn ( 1 - 0 ) emission . the lack of nh@xmath0 ( 3,3 ) emission in addition to the lack of hcn(1 - 0 ) emission and the observation of 1720 oh masers in this gap supports the idea that the northern arm of the mini - spiral originates outside the cnd and is crossing over or through the cnd on its way to the galactic center @xcite .
connections between the cnd and features at larger distances such as the 20 and 50 km s@xmath2 cloud can provide an explanation for the origin of the clouds that compose the cnd
. a connection must show more than morphological evidence to be considered a physical gas flow .
the presence of a velocity gradient along a streamer indicates that the gas is being gravitationally affected by the central gravitational field .
line ratios can also determine whether gas is heated as it approaches the nucleus , thus providing a way to discriminate against chance projections of clouds along the line - of - sight .
finally , as the gas intersects clouds in the cnd , the line widths should broaden .
streamers with these qualities are likely candidates for flows along which gas is transported towards the galactic center .
we detect at least three physical connections between the cnd and gas at larger distances .
these connections originate in the southern streamer , se1 , the northern ridge , and the western streamer and may represent inflow of material towards the nucleus .
figure [ pvpos ] shows the position of six position - velocity cuts along features which appear to be connecting the material to the cnd overlaid on nh@xmath0(3,3 ) emission .
the position velocity diagrams for nh@xmath0 ( 1,1 ) , ( 2,2 ) and ( 3,3 ) are shown in figures [ 11pv ] , [ 22pv ] , and [ pv ] . in each cut , position `` 0 '' corresponds to the labeled end of the cut . in the velocity integrated images ,
the brightest emission is at 30 - 50 km s@xmath2 mainly from the large gmcs . when performing mem deconvolution , channels with bright emission may have `` blooming '' where plateaus of low - level flux are added around strong features .
artifacts of blooming were seen in nh@xmath0(3,3 ) position - velocity diagrams as narrow lines centered at @xmath83 km s@xmath2 seen along most of the cut . to test the effects of blooming , we created cleaned data cubes .
the bright emission is identical in both diagrams , but the narrow ridges centered at 45 km s@xmath2 disappear in the cleaned diagrams . to be conservative ,
we present the cleaned diagrams in figures [ 11pv ] , [ 22pv ] , and [ pv ] .
the southern streamer has previously been observed to connect the 20 km s@xmath2 cloud to the southern lobe of the cnd @xcite .
position velocity cut @xmath107 in figures [ 11pv ] , [ 22pv ] , and [ pv ] show emission along the southern streamer . a small velocity gradient from 30 km s@xmath2 at a position of 40@xmath3 to 40 km
s@xmath2 at 70@xmath3 can be seen as gas approaches the cnd . with higher velocity resolution
, @xcite were able to detect a comparable shift of @xmath53 10 km s@xmath2 in this region of the southern streamer ( see fig .
8 , cut @xmath107 in @xcite ) .
in addition , the line is seen to broaden dramatically when the gas reaches the cnd . at a position of 90 - 100@xmath3 ,
nh@xmath0(3,3 ) emission is seen over a velocity range of almost 100 km s@xmath2 .
this broadening is also seen in nh@xmath0(2,2 ) and to a lesser degree in nh@xmath0(1,1 ) .
the peak of the emission is located at @xmath55 km s@xmath2 while broadening is seen in the lowest contours .
the broadening in all three transitions seems to be biased towards lower velocities and ranges from 70 to + 50 km s@xmath2 instead of being symmetric on both sides of the line center .
an increase in the nh@xmath0(2,2 ) to ( 1,1 ) line ratio as the streamer approaches the cnd can be seen in figure [ temp.fig ] and indicates that gas is being heated as it approaches the nucleus .
this heating could result from shocks as the gas merges with the cnd or from high - energy photons escaping from the nuclear region .
the nh@xmath0(3,3 ) hyperfine satellite components may be seen bracketing the strong central component .
asymmetric broadening of the main component may explain why the satellite lines do not appear to have equal intensity on both sides of the main component in position - velocity cut @xmath107 .
in figure [ pv]@xmath107 , nh@xmath0(3,3 ) emission at 30 km s@xmath2 and 110@xmath3 to 140@xmath3 appears to come from inside the cnd .
nh@xmath0(3,3 ) emission at 160@xmath3 is associated with the northwestern side of the cnd . in nh@xmath0(1,1 ) and ( 2,2 ) emission , however , the southern streamer stops sharply at the cnd .
this abrupt end is used by @xcite as evidence that the southern streamer is interacting with the cnd .
the small amount nh@xmath0(3,3 ) gas inside the cnd is probably hot since there is no emission in nh@xmath0(1,1 ) or ( 2,2 ) .
faint nh@xmath0 emission can be seen in figures [ 11.fig ] and [ 33.fig ] connecting the northern ridge to the northeastern lobe of the cnd .
position - velocity cut @xmath108 in figure [ pv ] shows the kinematic structure of this feature in nh@xmath0(3,3 ) .
gas in the northern ridge at a position of 30@xmath3 has a velocity of 10 km s@xmath2 and connects to the cnd with a smooth velocity gradient of 0.6 km s@xmath2arcsec@xmath2 .
this large velocity gradient covers 110@xmath3 ( 4 pc ) and reaches the cnd at 60 km s@xmath2 .
position velocity cuts in nh@xmath0(1,1 ) and ( 2,2 ) show the same kinematic signature for this streamer connecting the northern ridge to the cnd .
the positive velocity increase as the gas approaches the cnd places the northern ridge in front of the northeastern part of the cnd .
in addition , the low velocity of this streamer ( @xmath109 km s@xmath2 ) makes it likely to be gravitationally bound by the nucleus .
since the velocity increases as projected distance decreases , we can eliminate a circular orbit for this cloud .
the northern ridge has high nh@xmath0(2,2 ) to ( 1,1 ) line ratios and may have been impacted and disrupted by sgr a east ( see section 4.3 ) .
line ratios for most of the northern ridge and connecting stream are @xmath110 .
position - velocity cut @xmath108 can be used to better estimate the line ratios . at position 40@xmath3 ,
the ratio of nh@xmath0(2,2 ) to nh@xmath0(1,1 ) emission is @xmath111 .
this ratio increases to @xmath112 where the streamer intersects the cnd ( position 140@xmath3 ) . assuming that the streamer is optically thin along its entire length , this increase in nh@xmath0(2,2 ) to ( 1,1 ) line ratios corresponds to a temperature increase from @xmath113 k to @xmath55 k. although the upper part of the northern ridge is well traced in dust , there is little dust emission along the connection to the cnd ( see figure [ pbcor ] ) .
the northern ridge appears to have been swept up by the expansion of sgr a east .
some of the disrupted gas may be falling towards the cnd and it is possible that dust in the gas has been destroyed by photons from the nucleus .
although it is faint , se1 and the cnd are connected in all three velocity integrated images .
analysis of the kinematic data shows a strong connection between se1 and northeastern lobe of the cnd .
this connection can be easily seen in figure [ pv]@xmath85 . to the south
, se1 has a velocity of 40 km s@xmath2 at 50@xmath3 .
the velocity increases to the north until it reaches 50 km s@xmath2 at 90@xmath3 .
a connection to the cnd at 50 km s@xmath2 extends from 110@xmath3 to 140@xmath3 where it intersects the cnd at the same velocity .
se1 is also connected to the cnd by emission at 50 km s@xmath2 in nh@xmath0(1,1 ) and ( 2,2 ) .
the connection is narrow and weaker than emission from either se1 or the cnd
. it may be weak due to the intrinsic clumpiness of the clouds , or it also may have been disrupted by an interaction with another cloud .
kinematic evidence from hcn ( 1 - 0 ) indicates that the northern half of the cnd may be on a different orbit than the rest of the ring @xcite .
the independence of this feature from the rest of the cnd would also account for the different inclination angles observed for the two halves of the cnd .
given the strong connection between the cnd and se1 , it appears that some of the gas in the northern lobe originated in se1 and is not part of a coherent rotating ring .
position velocity cut @xmath114 in figure [ pv ] shows a striking velocity gradient of 1 km s@xmath2arcsec@xmath2 along the entire 150@xmath3 ( 6pc ) length of the western streamer .
the southern end of the streamer at 60@xmath3 has a velocity of 70 km s@xmath2 while the northern end has high velocity emission at + 90 km s@xmath2 at 220@xmath3 .
the same velocity gradient is observed in nh@xmath0(1,1 ) and ( 2,2 ) in position - velocity diagrams [ 11pv]@xmath114 and [ 22pv]@xmath114 .
the large velocity gradient along the length of the cloud could be due to intrinsic rotation or the cloud could be orbiting the nucleus .
if we assume a circular orbit for the western streamer at a distance of 2@xmath14 ( 4.6 pc ) the observed velocity gradient is consistent with a keplerian orbit around a central mass of 10@xmath115 m@xmath6 ( including sgr a * and the stellar population , see @xcite ) inclined by @xmath116 to the line of sight . additionally , the impact of sgr a east could enhance the gradient along this streamer .
the western streamer is not seen in 1.2 mm dust emission ( see figure [ pbcor ] ) .
the dust may have been destroyed by photons from the nucleus or by interactions with sgr a east . as seen in figure [ sgeast.fig ]
, the curve of the western streamer follows the edge of sgr a east .
it is possible that this gas originated closer to the nucleus and was pushed outward . in this scenario , any dust that was in the gas
was removed when it was close to the nucleus .
three narrow filaments appear to connect the western streamer to the cnd in the velocity integrated maps .
the southern - most projected connection , at @xmath71 , @xmath117 , is visible in position - velocity cut @xmath114 at 20@xmath3 with a velocity of 0 km s@xmath2 .
this cloud is the same extension of the 20 km s@xmath2 cloud towards the southwestern lobe of the cnd that was seen by @xcite in nh@xmath0(1,1 ) and ( 2,2 ) .
kinematically , it is associated with the 20 km s@xmath2 cloud and not the western streamer .
although the extension is spatially connected to the cnd in nh@xmath0(3,3 ) , hcn(1 - 0 ) and dust emission ( see figures [ hcn.fig ] and [ pbcor ] ) , there is no kinematic evidence for a physical connection between the 20 km s@xmath2 cloud and this south - western part of the cnd which has typical velocities of 110 km s@xmath2 .
the use of cleaned data for the position - velocity diagrams makes it difficult to see faint , extended connections between features .
it is possible that features which show morphological connections but no obvious kinematic associations with the cnd are connected to the cnd by faint and extended gas .
the second possible connection between the western streamer and the southwest lobe of the cnd is at @xmath71 , @xmath118 . in position velocity cut [ pv]@xmath84 the bright clump at 20@xmath3 , centered at -20
km s@xmath2 , is the western streamer .
emission from the cnd is is seen from 80@xmath3 to 180@xmath3 . in the western - most part of the cnd , there is bright emission at the same velocity as the western streamer ( 80@xmath3 ) , but any connection is tenuous .
the line widths are extremely high in this region ( fwhm up to 50 km s@xmath2 ) indicating turbulence .
nh@xmath0(1,1 ) and ( 2,2 ) show the same kinematics . in nh@xmath0(3,3 ) , the southwestern lobe of the cnd at 80 - 100@xmath3 shows two bright clumps centered at @xmath119 km s@xmath2 and @xmath120 km s@xmath2 . in nh@xmath0(1,1 ) there is prominent emission at @xmath121 km s@xmath2 at 110@xmath3 which is either part of the `` negative velocity lobe '' of the cnd observed in hcn(1 - 0 ) or high velocity gas at the nucleus .
the features at 10 and + 80 km s@xmath2 are more prominent in nh@xmath0 and do not fit a rotation pattern . in particular , the feature at 10 km s@xmath2 has a high line width which extends over 60 km s@xmath2 , especially in nh@xmath0(2,2 ) .
position - velocity cut @xmath122 follows northern - most possible connection to the cnd at @xmath73 , @xmath123 .
the western streamer is at a velocity of 30 km s@xmath2 at position of 40@xmath3 .
there is no obvious connection to the cnd which is at 70 km s@xmath2 and has a velocity gradient going to 30 km s@xmath2 towards the east .
the two northern projected connections between the cnd and the western streamer are seen as filaments in 6 cm continuum emission .
the filamentary structures in the continuum emission are intriguing and may be the result of supernova remnants or expansion of sgr a east through the cnd .
overall , we detect three physical connections to the cnd .
we confirm the presence of the southern streamer which connects the 20 km s@xmath2 cloud to the cnd .
the northern ridge is also connected to the cnd with a velocity gradient of 0.6 km s@xmath2 arcsec@xmath2 spanning 110@xmath3 ( 4 pc ) .
se1 extends northwards and kinematically connects to the eastern lobe of the cnd .
the western streamer shows a velocity gradient of 1 km s@xmath2 arcsec@xmath2 along a length of 150@xmath3 ( 6pc ) , but we do not see any definite physical connections to the cnd .
there is significant absorption in nh@xmath0(1,1 ) and ( 2,2 ) where cut @xmath84 passes close to sgr a * at 120@xmath3 .
the nh@xmath0(3,3 ) diagram , however , has little absorption and instead shows emission with a coherent velocity gradient from + 80 km s@xmath2 at 95@xmath3 to 0 km s@xmath2 at 145@xmath3 .
this feature is within 2 pc in projected distance from sgr a*. the positive velocity lobe of the cnd is seen at 170@xmath3 and is not a continuation of the gradient .
a similar gradient is seen as cut @xmath122 passes through the interior of the cnd ( 70 - 130@xmath3 ) . unlike cut @xmath84 ,
the velocity gradient of gas near sgr a * in cut @xmath122 is not constant .
the velocity decreases from 75 km s@xmath2 at 75@xmath3 to 50 km s@xmath2 at 90@xmath3 and then is approximately constant at 35 km s@xmath2 from 100 - 115@xmath3 . at 125@xmath3
, there is emission from 50 km s@xmath2 to + 40 km s@xmath2 .
although not shown in this paper , the same feature is seen over a range of position angles from 60 - 100@xmath10 .
the feature must contribute to nh@xmath0(3,3 ) emission seen close to sgr a * in the velocity integrated image ( see figure [ 33.fig ] ) , but the resolution of the image is too coarse to outline its morphology .
the observed velocity gradient may be the result of a cloud being tidally stripped by the central gravitational potential or may be a feature orbiting the black hole in the opposite sense as the cnd . at the position of sgr a * ( 125@xmath3 in @xmath84 and 110@xmath3 in @xmath122 ) , the velocity is centered at 30 km s@xmath2 making it unlikely that this feature is a counter - rotating inner disk around sgr a*. if we assume this feature is within 2 pc of the galactic center , then its maximum velocity ( 100 km s@xmath2 ) is well below the escape velocity for a gravitational potential of @xmath124m@xmath6 and the cloud is likely to be gravitationally bound to the nucleus .
however , considering projection effects and without the spatial morphology of this cloud , it is not possible to uniquely determine an orbit for this feature .
vla mosaics of the central 10 pc of the galaxy in nh@xmath0(1,1 ) , ( 2,2 ) , and ( 3,3 ) allow a detailed study of gas on all sides of the cnd .
the velocity integrated images show many filamentary features including the southern streamer , northern ridge , and western streamer as well as a high velocity cloud within 2 pc of the galactic center .
there is a high correlation between nh@xmath0(3,3 ) emission and emission from dust and hcn(3 - 2 ) , suggesting that nh@xmath0(3,3 ) traces column density . line ratios of nh@xmath0(2,2 ) to ( 1,1 ) emission indicate that the cnd and western streamer are hotter than gmcs which lie at a projected distance of @xmath125 pc .
the cnd is well - traced in nh@xmath0(3,3 ) although the relative intensity of features differs from hcn(1 - 0 ) . outside the cnd ,
the 50 km s@xmath2 cloud , northern ridge and western streamer are located around the edge of sgr a east and appear to have been swept up by the expanding shell .
many of the filamentary features including the northern ridge , southern streamer and se1 are kinematically connected to the cnd indicating that gmcs are feeding the nucleus along many paths .
genzel , r. , eckart , a. , ott , t. , & eisenhauer , f. 1997 , , 291 , 219 genzel , r. , stacey , g.j . ,
harris , a.i . ,
geis , n. , graf , u.u . ,
poglitsch , a. , & sutzki , j. 1990 , , 356 , 160 genzel , r. , watson , d.m . ,
crawford , m.k .
& townes , c.h .
1985 , , 297 , 766 ho , p.t.p .
1993 , in proc . of the 2@xmath126 cologne - zermatt symposium , the physics and chemistry of interstellar molecular clouds , ed .
g. winnewisser & g. pelz ( new york : springer ) , 33 ho , p.t.p .
, ho , l.c . , szczepanski , j.c . ,
jackson , j.m . ,
armstrong , j.t . ,
& barrett , a.h .
1991 , , 350 , 309 ho , p.t.p . &
townes , c.h .
1983 , , 21 , 239 zylka , r. 1998 , in iau symp . 184 , the central regions of the galaxy and galaxies , ed . y. sofue ( dordrecht : kluwer ) , 156 zylka , r. , gsten , r. , philipp , s. , ungerechts , h. , mezger , p.g . , & duschl , w.j .
1999 , in asp conf .
ser , 186 , the central parsecs of the galaxy , ed . h. falcke et al .
( san francisco : asp ) , 415 zylka , r. mezger , p.g . , & wink , p.e .
1990 , , 234 , 133 | new vla images of nh@xmath0 ( 1,1 ) , ( 2,2 ) , and ( 3,3 ) emission in the central 10 parsecs of the galaxy trace filamentary streams of gas , several of which appear to feed the circumnuclear disk ( cnd ) .
the nh@xmath0 images have a spatial resolution of @xmath1 and have better spatial sampling than previous nh@xmath0 observations .
the images show the `` southern streamer , '' `` 50 km s@xmath2 cloud , '' and new features including a `` western streamer '' , 6 parsecs in length , and a `` northern ridge '' which connects to the cnd .
nh@xmath0(3,3 ) emission is very similar to 1.2 mm dust emission indicating that nh@xmath0 traces column density well
. ratios of the nh@xmath0(2,2 ) to ( 1,1 ) line intensities give an estimate of the temperature of the gas and indicate high temperatures close to the nucleus and cnd .
the new data cover a velocity range of 270 km s@xmath2 , including all velocities observed in the cnd , with a resolution of 9.8 km s@xmath2 .
previous nh@xmath0 observations with higher resolution did not cover the entire range of velocities seen in the cnd .
the large - scale kinematics of the cnd do not resemble a coherent ring or disk .
we see evidence for a high velocity cloud within a projected distance of 50@xmath3 ( 2 pc ) which is only seen in nh@xmath0(3,3 ) and is likely to be hot .
comparison to 6 cm continuum emission reveals that much of the nh@xmath0 emission traces the outer edges of sgr a east and was probably pushed outward by this expanding shell .
the connection between the northern ridge ( which appears to be swept up by sgr a east ) and the cnd indicates that sgr a east and the cnd are in close proximity to each other .
kinematic evidence for these connections is presented in this paper , while the full kinematic analysis of the central 10 pc will be presented in paper ii . |
the central engine of gamma - ray bursts ( grbs ) remains an open question in grb physics ( zhang 2011 ) .
observations of grb prompt emission and early afterglow pose the following constraints on a successful central engine model : ( 1 ) the central engine must be able to power an outflow with an extremely high energy and luminosity ( e.g. zhang & mszros , 2004 ; meszaros 2006 ) ; ( 2 ) the ejecta must have a low baryon loading , with energy per baryon exceeding 100 ( e.g. lithwick & sari , 2001 ; liang et al .
2010 ) ; ( 3 ) the central engine should be intermittent in nature to account for the observed light curves with rapid variability ( fishman & meagan 1995 ) ; ( 4 ) the engine should last for an extended period of time to power delayed erratic x - ray flares ( burrows et al .
2005 ; zhang et al .
2006 ) or long - lasting x - ray emission followed by a sudden drop ( i.e. `` internal plateau '' , troja et al . 2007 ; liang et al .
2007 ; lyons et al .
2010 ) ; ( 5 ) finally , fermi observations require that the central engine should be strongly magnetized to launch a magnetically dominated outflow at least for some grbs ( zhang & peer 2009 )
. two types of grb central engine models have been discussed in the literature ( e.g. kumar & zhang 2014 for a review ) .
the leading type of models invokes a hyper - accreting stellar - mass black hole ( e.g. popham et al .
1999 ; narayan et al . 2001
; lei et al . 2013 ) , from which a relativistic jet is launched via neutrino - anti - neutrino annihilation ( ruffert et al . 1997 ; popham et al . 1999 ; chen & beloborodov 2007 ; lei et al . 2009 ) , blandford - znajek mechanism ( blandford & znajek 1997 ; lee et al . 2000 ; li 2000 ) , or episodic magnetic bubble ejection from the disk ( yuan & zhang 2012 ) .
the second type of models invokes a rapidly spinning , strongly magnetized neutron star dubbed a `` millisecond magnetar '' ( usov 1992 ; thompson 1994 ; dai & lu 1998a ; wheeler et al .
2000 ; zhang & mszros 2001 ; metzger et al . 2008 , 2011 ; bucciantini et al .
2012 ) . within this scenario ,
the energy reservoir is the total rotation energy of the millisecond magnetar , which reads @xmath2 where @xmath3 is the moment of inertia , @xmath4 is the initial angular frequency of the neutron star , @xmath5 , and the convention @xmath6 is adopted in cgs units for all other parameters throughout the paper . assuming that the magnetar with initial spin period @xmath1 is being spun down by a magnetic dipole with surface polar cap magnetic field @xmath0 , the spindown luminosity would evolve with time as ( zhang & mszros 2001 ) @xmath7 where @xmath8 is the characteristic spindown luminosity , and @xmath9 is the characteristic spindown time scale . the spin - down behavior of the magnetar can leave characteristic imprints in the observed grb emission .
dai & lu ( 1998a ) first proposed an energy injection model of millisecond pulsars to interpret a rebrightening feature of the first optical afterglow detected in grb 970228 .
the required @xmath0 is @xmath10 g , not quite a magnetar strength .
the prompt grb emission has to be attributed to additional physical processes , e.g. magnetic dissipation in a differentially rotating neutron star ( kluzniak & ruderman 1998 ) or strange quark star ( dai & lu 1998b ) .
zhang & mszros ( 2001 ) studied energy injection from a central engine with a general luminosity law @xmath11 ( the magnetar injection corresponds to @xmath12 for @xmath13 and @xmath14 for @xmath15 ) , and pointed out that besides the rebrightening feature discussed by dai & lu ( 1998a , b ) , for more typical magnetar parameters , one can have a shallow decay phase followed by a normal decay phase in the early afterglow of a grb .
such a shallow decay phase ( or plateau ) was later commonly observed in swift early xrt light curves ( zhang et al .
2006 ; nousek et al .
2006 ; obrien et al . 2006
; liang et al . 2007 ) .
it can be readily interpreted as energy injection from a millisecond magnetar central engine ( zhang et al .
an alternative energy injection model invokes a short - duration central engine , which ejects materials with a stratified lorentz factor ( @xmath16 ) profile .
energy is gradually added to the blastwave as the blastwave is gradually decelerated to progressively lower @xmath16 ( rees & mszros 1998 ; sari & mszros 2000 ; uhm et al .
both models can interpret the shallow decay phase of most x - ray light curves .
a tie - breaker grb was discovered in early 2007 .
grb 070110 ( troja et al . 2007 ) showed an extended plateau with a near flat light curve extending to over @xmath17 seconds before rapidly falling off with a decay index @xmath18 ( throughout the paper the convention @xmath19 is adopted ) .
such a rapid decay can not be accommodated in any external shock model , so that the entire x - ray plateau emission has to be attributed to internal dissipation of a central engine wind .
such an `` internal plateau '' was later discovered in several more grbs ( liang et al .
2007 ; lyons et al . 2010 ) .
the near steady x - ray emission observed in grb 070110 may not be easy to interpret within a black hole central engine model , but is a natural prediction of the magnetar central engine model ( eq.([lt ] ) when @xmath20 ) .
the rapid @xmath21 decay near the end is not predicted in the magnetic dipole radiation model .
troja et al . ( 2007 ) interpreted it as being due to collapse of the magnetar to a black hole after loosing centrifugal support .
interestingly , internal plateaus are also discovered in a good fraction of short grbs ( rowlinson et al .
2010 , 2013 ) .
modeling various afterglow features for both long and short grbs within the framework of the millisecond magnetar ( or pulsar with weaker magnetic field ) central engine model has gained growing attention ( dai et al .
2006 ; gao & fan 2006 ; fan & xu 2006 ; metzger et al . 2008 , 2011 ; dallosso et al .
2011 ; fan et al .
2011 ; bucciantini et al . 2012
; bernardini et al . 2013 ; gompertz et al .
2013 , 2014 ) .
numerical simulations of binary neutron star mergers indeed show that a stable magnetar can survive if the initial masses of the two neutron stars are small enough , which would power a short gamma - ray burst ( giacomazzo & perna .
2013 ) .
even though evidence of a magnetar central engine is mounting , it remains unclear whether the rich grb data accumulated over the years with the grb mission swift indeed statistically requires the existence of ( presumably ) two types of central engines .
if indeed magnetars are operating in some grbs while hyper - accreting black holes are operating in others , do the data show statistically significant differences between the two samples ?
do those grbs that seem to have a magnetar signature have physical parameters that are consistent with the predictions of the magnetar central engine model ? this paper is to address these interesting questions through a systematic analysis of the swift x
ray telescope ( xrt ) data .
the xrt data reduction details and criteria for sample selection are presented in 2 . in 3 , physical parameters of the grbs and the hypothetical magnetars are derived for all the samples .
a statistical comparison of the physical properties between the magnetar samples and the non - magnetar sample are presented in 4 , and conclusions are drawn in 5 with some discussion . throughout the paper , a concordance cosmology with parameters @xmath22 km s@xmath23
mpc @xmath23 , @xmath24 , and @xmath25 is adopted .
the xrt data are downloaded from the swift data archive .
we developed a script to automatically download and maintain all the xrt data on the local unlv machine .
the heasoft packages _ version 6.10 _ , including xspec , xselect , ximage , and the swift data analysis tools , are used for the data reduction .
an idl code was developed by the former group member b .- b .
zhang to automatically process the xrt data for a given burst in any user - specified time interval ( see zhang et al .
2007c for details ) .
we adopt this code with slight modifications to solve the problem designed for this paper .
the same idl code was used in several previous papers ( zhang et al .
2007c ; liang et al .
2007 , 2008 , 2009 ) of our group .
more details about the data reduction procedures can be found in zhang et al .
( 2007c ) and evans et al .
( 2009 ) .
our entire sample includes more than 750 grbs observed between 2005 january and 2013 august , whose xrt data are all processed with our data reduction tool . since the magnetar signature typically invokes a shallow decay phase ( or plateau ) followed by a steeper decay segment ( a normal decay for canonical light curves , or a very steep decay for internal plateaus ) , our attention is on those grbs that show such a transition in the x - ray light curves .
we first identify such bursts by inspecting their light curves . in order to grade their magnetar candidacy
, we next perform a temporal fit to the plateau behavior within a time interval @xmath26 , where @xmath27 is the beginning of the plateau , while @xmath28 is the end of the segment after the plateau break ( either last observed data point if there is no further break in the lightcurve , or the break time if a second break appears ) .
since we are mostly interested in the behavior around the break time @xmath29 , the exact positions of @xmath27 and @xmath28 do not matter much , so we pick them through visual inspection of the light curves .
we then fit the light curves with a smooth broken power law @xmath30^{-1/\omega},\ ] ] where @xmath31 is the break time , @xmath32 is the flux at the break time @xmath29 , @xmath33 and @xmath34 are decay indices before and after the break , respectively , and @xmath35 describes the sharpness of the break .
the larger the @xmath35 parameter , the sharper the break .
an idl routine named `` mpfitfun.pro '' is employed for our fitting ( mor 1977 ; markwardt 2009 ) .
this routine performs a levenberg - marquardt least - square fit to the data for a given model to optimize the model parameters . after processing all the data , we grade all long grbs in our sample into four groups ( `` gold '' , `` silver '' , `` aluminum '' , and `` non - magnetar '' ) according to their likelihood of being powered by a magnetar central engine . * gold : this sample is defined by those bursts that display an `` internal plateau '' .
these plateaus are followed by a decay slope steeper than 3 , which is essentially impossible to interpret within the external shock models ( gao et al .
2013b ) ( kumar & panaitescu 2000 ) , which is typically smaller than 3 , and is defined by the high - latitude `` curvature effect '' emission from a conical outflow , even if the emission abruptly ceases . ] .
it demands a long - lasting central engine , and a near steady flux is consistent with emission from a spinning down magnetar .
the rapid decay at the end of plateau may mark the implosion of the magnetar into a black hole ( troja et al .
2007 ; zhang 2014 ) .
there are altogether only 9 robust cases identified in this gold sample , 3 of which have redshift measurements . the light curves of these 9 grbs together with the broken power - law fittings ( red curves ) are shown in figure 1 , and the fitting parameters are summarized in table 1 . * silver : this sample includes grbs with a shallow decay phase followed by a normal decay phase , and the pre- and post - break temporal and spectral properties are well consistent with the external forward shock model with energy injection of a magnetar as defined in eq.([lt ] ) .
specifically , one requires two indepedent criteria to define this sample .
first , the temporal and spectral properties of the afterglow after the break ( the normal decay phase ) should satisfy the `` closure relation '' of the external shock model ( e.g. zhang & mszros 2004 ; gao et al .
2013b ) , i.e. @xmath36 here @xmath37 is the spectral index of the normal decay segment ( which is x - ray photon index minus 1 ) , and @xmath38 is the electron s spectral distribution index .
second , the pre - break slope @xmath33 should correspond to @xmath12 , while the post - break slope @xmath34 should correspond to @xmath39 ( for a constant energy fireball , the scaling law is the same as @xmath39 , zhang & mszros 2001 ) , so according to zhang et al .
( 2006 ) and gao et al .
( 2013b ) , one should have @xmath40 in our entire sample , 69 grbs can be grouped into this silver sample , with 33 having measured redshifts .
the light curves with fitting curves are presented online at http://grb.physics.unlv.edu/$\sim$lhj/silver/ , and the fitting results are reported in table 1 .
two examples ( grbs 060729 , see also grupe et al . 2007 , and 070306 ) are shown in figure 2 .
figure 3 shows all the grbs in the @xmath41 plane , with three theoretically favored lines of the magnetar models ( eq.([a1-a2 ] ) ) plotted .
those grbs falling onto these lines ( within error bars ) and also satisfy the closure relations are identified as silver sample grbs ( colored data points ) . in fig.4
we present the distribution of electron spectral index @xmath38 derived from the silver sample .
it has a gaussian distribution with a center value @xmath42 .
figure 5 shows the distribution of silver sample in the ( @xmath43)-plane combined with the closure relations for the models ( ism and wind medium ) . *
aluminum : other grbs with a shallow decay segment transiting to a steeper decay are included in the aluminum sample .
they either do not satisfy external shock closure relations in the post - break phase , or do not satisfy the @xmath44 relations predicted in the magnetar external shock models .
these are marked as grey points in fig.3 .
those grbs that fall onto the three magnetar model lines but are still denoted as aluminum are the ones that do not satisfy the closure relations in the post - break phase . on the other hand , since early magnetar spindown may not fully follow the simple dipole spindown law ( e.g. metzger et al . 2011 ) , and since the observed x - ray emission may not come from the external forward shock emission ( e.g. can be from external reverse shock , dai 2004 ; yu & dai 2007 , or from internal dissipation of the magnetar wind , yu et al .
2010 ) , these grbs could be still powered by magnetars .
we therefore still assign them as magnetar candidates , but with a lower grade .
there are 135 solid cases in the sample , 67 of which have redshift measurements .
the light curves with fitting curves are presented online at http://grb.physics.unlv.edu/$\sim$lhj / aluminum/. two examples ( grbs 070420 and 080430 ) are presented in fig.2 . * non - magnetar : all the other long grbs we have analyzed are included in the non - magnetar sample .
they either have a single power - law decay , or have erratic flares that prevent identifying a clear shallow decay phase , or present a rebrightening behavior , or the data are too poor to reach a robust conclusion .
there are more than 400 grbs in this group , 111 of which have redshift measurements . strictly speaking , some of these grbs
may still host a magnetar central engine .
we define these grbs as `` non - magnetar '' , simply because they do not present a clear magnetar signature .
two examples ( grbs 061007 , see also schady et al .
2007 , mundell et al .
2007 , and 081028 ) are presented in the fig .
finally , we also independently processed the x - ray data of short grbs that may harbor a magnetar central engine ( cf .
rowlinson et al 2013 ) .
we select the short grbs that have measured redshifts and high - quality x - ray data .
the light curves with fitting curves are presented online at http://grb.physics.unlv.edu/$\sim$lhj / sgrb/.
our purpose is to analyze and compare the physical properties of grbs with or without a magnetar signature . in this section ,
we use data to derive relevant physical parameters .
redshift measurements are crucial to derive the intrinsic parameters ( energy , luminosity , etc ) , so in the following we focus on those grbs with @xmath45 measurements only .
the isotropic prompt @xmath46-ray emission energy @xmath47 is usually derived from the observed fluence @xmath48 in the detector s energy band , and extrapolated to the rest - frame @xmath49 kev using spectral parameters ( the low- and high- energy spectral indices @xmath50 , @xmath51 , and the peak energy @xmath52 for a standard `` band - function '' fit , band et al .
1993 ) and through @xmath53-correction
. however , since the bat energy band is narrow ( 15 - 150 kev ) , for most grbs the spectra can be only fit by a cutoff power law or a single power law ( sakamoto et al .
2008 , 2011 ) .
we therefore apply the following procedure to estimate the band spectral parameters : ( 1 ) if a burst was also detected by _
gbm or _ konus _
wind , we adopt the spectral parameters measured by those instruments .
( 2 ) for those bursts that are not detected by other instruments but can be fit with a cutoff power law model , we adopt the derived @xmath50 and @xmath52 parameters and @xmath52 are slightly different for the cut - off power law and band - function models ( e.g. sakamoto et al . 2008 , 2011 ) , but the derived @xmath54 only shows a slight difference , which is ignored in our analysis . ] , and assume a typical value of @xmath55 .
( 3 ) for those grbs that can be only fit with a single power law , we have to a derive @xmath52 using an empirical correlation between the bat - band photon index @xmath56 and @xmath52 ( e.g. sakamoto et al .
2009 ; zhang et al . 2007b ; virgili et al . 2012 ; l et al .
the typical parameters @xmath57 , @xmath58 are adopted to perform the simulations .
we can then calculate the @xmath54 according to @xmath59 where @xmath45 is the redshift , @xmath60 is the luminosity distance , and @xmath53 is the @xmath53-correction factor from the observed band to @xmath49 kev in the burst rest frame ( e.g. bloom et al .
2001 ) .
another important parameter is the isotropic kinetic energy @xmath61 measured from the afterglow flux .
this value is increasing during the shallow decay phase , but becomes constant during the normal decay phase ( zhang et al .
we follow the method discussed in zhang et al .
( 2007a ) to calculate @xmath61 during the normal decay phase using the x - ray data .
noticing that fast - cooling is disfavored at this late epoch , we derive several relevant cases . for @xmath62 ,
the afterglow flux expression does not depend on the medium density , so the following expression ( zhang et al .
2007a ) applies to both ism and wind models @xmath63^{4/(p+2 ) } \nonumber \\ & \times & d_{28}^{8/(p+2)}(1+z)^{-1 } t_d^{(3p-2)/(p+2)}\nonumber \\ & \times & ( 1+y)^{4/(p+2 ) } f_p^{-4/(p+2)}\epsilon_{b,-2}^{(2-p)/(p+2 ) } \nonumber \\ & \times & \epsilon_{e,-1}^{4(1-p)/(p+2 ) } \nu_{18}{^{2(p-2)/(p+2)}}. \nonumber \\\end{aligned}\ ] ] for the @xmath64 ism model , one has ( zhang et al .
2007a ) @xmath65^{4/(p+3 ) } \nonumber \\ & \times & d_{28}^{8/(p+3)}(1+z)^{-1 } t_d^{3(p-1)/(p+3)}\nonumber \\ & \times & f_p^{-4/(p+3 ) } \epsilon_{b,-2}^{-(p+1)/(p+3 ) } \epsilon_{e,-1}^{4(1-p)/(p+3 ) } \nonumber \\ & \times & n^{-2/(p+3 ) } \nu_{18}{^{2(p-3)/(p+3)}}. \nonumber \\\end{aligned}\ ] ] for the @xmath64 wind model , one has ( gao et al . 2013b ) @xmath66 @xmath67 @xmath68 so that @xmath69 and @xmath70^{4/(p+1 ) } \nonumber \\ & \times & d_{28}^{8/(p+1)}(1+z)^{-(p+5)/(p+1 ) } t_d^{(3p-1)/(p+1)}\nonumber \\ & \times & f_p^{-4/(p+1 ) } \epsilon_{b,-2}^{-1 } \epsilon_{e,-1}^{4(1-p)/(p+1 ) } \nonumber \\ & \times & a_{\ast,-1}^{-4/(p+1 ) } \nu_{18}{^{2(p-3)/(p+1)}}.\end{aligned}\ ] ] here @xmath71 is the energy flux at @xmath72 hz ( in units of @xmath73 ) , @xmath74 is the density of the constant ambient medium , @xmath75 is the stellar wind parameter , @xmath76 is the time in the observer frame in days , and @xmath77 is the compton parameter .
the electron spectral index @xmath38 and the spectral index @xmath37 are connected through @xmath78 and @xmath79 is a function of the power law distribution index @xmath38 ( zhang et al .
2007a ) @xmath80 in our calculations , the microphysics parameters of the shock are assigned to standard values dervied from observations ( e.g. panaitescu & kumar 2002 ; yost et al .
2003 ) : @xmath81=0.1 and @xmath82 .
the compton parameter is assigned to a typical value @xmath83 . after deriving the break time @xmath29 through light curve fitting ,
we derive the break time luminosity as @xmath84 where @xmath85 is the x - ray flux at @xmath29 .
since the xrt band is narrow , no @xmath53-correction is possible to calculate @xmath86 .
a jet break was detected in some grbs in our sample .
for these grbs , we correct all the isotropic values to the beaming - corrected values by multiplying the values by the beaming correction factor ( frail et al . 2001 ) @xmath87 i.e. @xmath88 , and @xmath89 .
the jet angle information was searched from the literature ( e.g. liang et al .
2008 ; racusin et al . 2009 ; lu et al .
2012 ; nemmen et al . 2012 ) , which is collected in table 2 .
the grb radiation efficiency is defined as ( lloyd - ronning & zhang 2004 ) @xmath90 since @xmath61 ( and @xmath91 ) are increasing functions of time during the shallow decay phase , @xmath92 is different when @xmath61 ( @xmath91 ) at different epochs are adopted . following zhang et al .
( 2007a ) , we take a typical blastwave deceleration @xmath93 and the end of the shallow decay phase @xmath29 to calculate the radiative efficiencies . within the framework of the magnetar central engine model ,
the two efficiencies carry different physical meanings : @xmath94 denotes the efficiency of dissipating the magnetar wind energy during the prompt emission phase , while @xmath95 denotes the total efficiency of converting the spindown energy of a magnetar to @xmath46-ray radiation . for a magnetar undergoing dipolar spindown , two important magnetar parameters , i.e. the initial spin period @xmath1 and the surface polar cap magnetic field @xmath0 ,
can be solved by the characteristic luminosity @xmath96 ( eq.([l0 ] ) ) and the spindown time scale @xmath97 ( eq.([tau ] ) ) .
the spindown time scale can be generally identified as the observed break time , i.e. @xmath98 one caution is that @xmath97 can be shorter than @xmath99 if the magnetar is supra - massive , and collapses to a black hole before it is significantly spun down . on the other hand ,
the angular velocity of the magnetar does not change significantly until reaching the characteristic spindown time scale , so that the collapse of the supra - massive magnetar , if indeed happens , would likely happen at or after @xmath97 . in our analysis , we will adopt eq.([tau = tb ] ) throughout .
the characteristic spindown luminosity should generally include two terms : @xmath100 where @xmath101 is the x - ray luminosity due to internal dissipation of the magnetar wind , which is the observed x - ray luminosity of the internal plateau ( for external plateaus , one can only derive an upper limit ) , and @xmath102 is the kinetic luminosity that is injected into the blastwave during the energy injection phase .
it depends on the isotropic kinetic energy @xmath61 after the injection phase is over , which can be derived from afterglow modeling discussed above . for the gold sample ,
the @xmath101 component dominates , while for silver and aluminum samples , the @xmath103 component dominates . in any case , both components should exist and contribute to the observed flux ( zhang 2014 ) .
one can also define an x - ray efficiency to define the radiative efficiency for a magnetar to convert its spindown energy to radiation , i.e. @xmath104 in our analysis , we try to calculate both @xmath101 and @xmath105 from the data . for the gold sample grbs that show internal plateaus
, @xmath101 can be readily measured . for the cases where
the internal plateau lands on an external shock component ( e.g. troja et al . 2007 ) , @xmath103 can be also derived by modeling the late x - ray afterglow in the normal decay phase . for the gold sample cases where no late external shock component is available
, one can only set up an upper limit on @xmath105 . for silver and aluminum samples ,
the internal plateau component is not detectable . through simulations , we find that the external shock component would not be significantly modified if the internal plateau flux is below 50% of the observed external shock flux .
therefore for all the silver and aluminum sample grbs , we place an upper limit of @xmath101 as 50% of the observed x - ray flux .
we derive magnetar parameters ( @xmath1 and @xmath0 ) of the gold , silver and aluminum samples using eqs.([l0 ] ) , ( [ tau ] ) , and ( [ l0 - 2 ] ) and @xmath106 are used to derive them .
the grb prompt emission presumably also consumed spin energy and magnetic energy of the magnetar , so the true initial spin period can be somewhat smaller than @xmath1 , and the true initial ( effective ) dipole magnetic field at the pole can be somewhat larger than @xmath0 ) . ] .
first , we assume that the magnetar wind is isotropic , so that @xmath107 .
the derived @xmath1 , @xmath0 are presented in table 2 and fig.6a .
most `` magnetars '' have @xmath0 below @xmath108 g , some even have @xmath0 below @xmath109 g , which are not considered as magnetars .
more problematically , most derived @xmath1 s are much shorter than 1 ms .
this directly conflicts with the break - up limit of a neutron star , which is about 0.96 ms ( lattimer & prakash 2004 ) .
this suggests that the isotropic assumption for these long grb magnetar winds is not correct .
we then introduce the beaming factor @xmath110 for each grb .
if @xmath111 is measured , we simply adopt the value . otherwise , we choose @xmath112 , a typical jet opening angle for bright long grbs ( frail et al .
2001 ; liang et al . 2008 ) . very interestingly
, after such a correction , all the data points of gold and silver sample grbs fall into the expected region in the @xmath113 plot ( fig.6b ) .
also the additional conditions imposed by the causality argument ( i.e. that the speed of sound on the neutron star can not exceed the speed of light , lattimer et al . 1990 , and eqs.(9 ) and ( 10 ) of rowlinson et al .
( 2010 ) ) are satisfied for all grbs in all three ( gold , silver and aluminum ) magnetar samples , if one assumes @xmath114 .
all these suggest that the long grb magnetar winds are likely collimated .
some aluminum sample grbs are still to the left of the allowed region ( with @xmath1 shorter than the break - up limit ) .
this may suggest that those aluminum sample bursts are not powered by magnetars , or are powered by magnetars with even narrower jets . very interestingly
, the magnetar properties of short grbs derived under the isotropic assumption actually lie reasonably in the allowed region ( fig.6a , blue dots ) . after jet correction for long grb magnetars ( but keep short grb magnetar wind isotropic ) , the derived magnetar parameters are well mixed in the same region .
this suggests that the isotropic assumption for short grbs is reasonably good .
this is understandable within the framework of the progenitor models of grbs .
short grbs are believed to be powered by mergers of ns - ns or ns - bh systems ( paczynski 1986 ; eichler et al .
1989 ; paczynski 1991 ; narayan et al .
1992 ) . during the merger process ,
only a small amount of materials are launched ( freiburghaus et al .
1999 ; rezzolla et al .
2010 ; hotokezaka et al .
a millisecond magnetar is expected to launch a near isotropic wind .
this wind , instead of being collimated by the ejecta ( e.g. bucciantini et al . 2012 ) , would simply push the ejecta behind and accelerate the ejecta and make a bright electromagnetic signal in the equatoral directions ( fan & xu 2006 ; zhang 2013 ; gao et al . 2013 ; yu et al . 2013 ; metzger & piro 2013 ) . in the jet direction ,
the magnetar wind emission is not enhanced by the beaming effect , so that one can infer correct magnetar parameters assuming an isotropic wind . for long grbs , on the other hand ,
jets are believed to be launched from collapsing massive stars ( woosley 1993 ; macfadyen & woosley 1999 ) .
the initially near isotropic magnetar wind is expected to be soon collimated by the stellar envelope to a small solid angle ( bucciantini et al .
2008 ) .
figure 7 shows the correlations of @xmath115 and @xmath116 for the entire sample .
as shown in fig.7a , a higher isotropic @xmath46-ray energy generally has a higher x - ray break luminosity . for the gold and silver samples ,
a spearman correlation analysis gives a dependence @xmath117 with a correlation coefficient @xmath118 , and a chance probability @xmath119 .
adding the aluminum sample only slightly worsens the correlation ( @xmath120 , with @xmath121 and @xmath119 ) .
such a correlation is expected , which may be caused by a combination of intrinsic ( a more energetic magnetar gives more significant contribution to both prompt emission and afterglow ) and geometric effects ( a narrower jet would enhance both prompt emission and afterglow ) .
figure 7b presents an anti - correlation between @xmath86 and @xmath29 ( dainotti et al .
our gold + silver sample gives @xmath122 with @xmath123 and @xmath119 .
adding the aluminum sample only slightly worsens the correlation ( @xmath124 with @xmath125 and @xmath119 ) .
such an anti - correlation is consistent with the prediction of the magnetar model : given a quasi - universal magnetar total spin energy , a higher magnetic field would power a brighter plateau with a shorter duration , or vice versa ( see also xu & huang 2012 ) . in fig.8 , we compare the inferred @xmath126 with the total rotation energy @xmath127 ( eq.([erot ] ) ) of the millisecond magnetar .
it is found that the grbs are generally above and not too far above the @xmath128 line .
this is consistent with the magnetar hypothesis , namely , all the emission energy ultimately comes from the spin energy of the magnetar .
figure 8a includes all the grbs in the gold / silver / aluminum samples , with @xmath129 assumed if the jet angle is not measured .
figure 8b presents those grbs with jet measurements only .
essentially the same conclusion is reached .
a very interesting question is whether there are noticeable differences between the magnetar and non - magnetar samples .
one potential discriminator would be the total energetics of the grbs . while the magnetar model predicts a maximum value of the total energy ( eq.([erot ] ) ) , the black hole model is not subject to such a limit .
in fig.9 we make some comparisons .
the first three panels compare the histograms of the isotropic energies ( @xmath54 , @xmath61 , and @xmath130 ) of the magnetar and non - magnetar samples . for the magnetar sample , we in one case includes the most secure ( gold + silver ) sample only ( blue hatched ) , and in another case
includes all magnetar candidates ( gold + silver + aluminum ) ( red solid ) .
the non - magnetar sample is marked in grey .
the best gaussian fits to the three samples are presented as blue , red , and black dotted curves , respectively .
the center values of all the fits are presented in table 3 .
it is found that without jet correction , the isotropic values of the magnetar and non - magnetar samples are not significantly different .
next , we introduce beaming correction , and replot the histograms of the jet - corrected energies of the magnetar and non - magnetar samples . the results are presented in the later three panels in fig.9 .
one can see a clear distinction between the robust magnetar sample ( gold + silver ) and the non - magnetar sample . for the total energy ( @xmath131 ) , while the former peaks around 50.62 erg , the latter peaks around 51.81 erg .
more interestingly , all the gold+silver magnetar sample grbs have a total energy smaller than the limit set by the spin energy ( eq.([erot ] ) ) , while for some non - magnetar sample grbs , this upper limit is exceeded .
the results are generally consistent with the hypothesis that two types of grb central engines can both power grbs . in fig.10a and fig.10b , we compare @xmath54 and @xmath61 for the magnetar and non - magnetar samples . the kinetic energy of the blastwave @xmath61
is evaluated at @xmath29 for fig.10a , and at @xmath93 for fig.10b ( similar to zhang et al .
it is interesting to see that at @xmath93 , the magnetar central engine tends to power more efficient grbs ( due to the initial small @xmath91 value ) than the black hole central engine .
it is interesting to see after the energy injection phase ( at @xmath29 ) , the @xmath46-ray efficiencies of magnetar and non - magnetar samples are no longer significantly different . the same conclusion is also manifested in fig.11a and 11b , where we plot the histograms of @xmath92 for different samples .
if one accepts that millisecond magnetars are powering some grbs , it would be interesting to constrain the internal energy dissipation efficiency @xmath132 ( eq.([etax ] ) ) from the data .
in both fig.7a and fig.7b , it is found that the gold sample grbs have a relatively large @xmath86 value .
this is generally consistent with the expectation that a larger @xmath132 would give rise to an internal plateau ( zhang 2014 ) . in fig.10c
, we compare @xmath105 and @xmath133 .
it indeed shows that the gold sample grbs have a much higher @xmath132 than other grbs .
on the other hand , it is curious to ask why there is a gap in this phase space .
it appears that some magnetars are particularly efficient to dissipate the magnetar wind energy , while most magnetars are not .
plotting the histograms of @xmath132 ( fig.11c ) , it looks indeed like a bimodal distribution of @xmath132 , even though this second high @xmath132 component is not significant enough . in fig.12 ,
we present the scatter plots of @xmath132 against other parameters , including @xmath134 , @xmath54 , @xmath61 , and @xmath135 . in all cases , the gold sample ( the ones with very high @xmath132 ) tend to stick out and emerge as a separate population .
in order to address whether ( at least ) some grbs might have a magnetar central engine , we have systematically analysed the x - ray data of all the _ swift _ grbs ( @xmath136 ) detected before august 2013 . by applying some criteria to judge how likely a grb might harbor a millisecond magnetar central engine , we characterized long grbs into several samples : gold , silver , and aluminum magnetar samples , as well as the non - magnetar samples . for comparison , we also independently processed the data of short grbs that might have a magnetar central engine ( rowlinson et al .
2010 , 2013 ) . by deriving the basic magnetar parameters @xmath1 and @xmath0 from the data , we are able to reach two interesting conclusions . first , it seems that at least for the gold and silver sample grbs , the derived properties seem to be consistent with the expectations of the magnetar central engine model .
the consistency includes the following : 1 . after beaming correction ,
the derived @xmath1 and @xmath0 seem to fall into the reasonable range expected in the magnetar central engine model ; 2 .
the @xmath137 anti - correlation seems to be consistent with the hypothesis that there is a quasi - universal energy budget defined by the spin energy of the magnetars ( eq.([erot ] ) ) ; 3 .
the sum of @xmath138 and @xmath91 is generally smaller than @xmath127 , the total energy budget of a magnetar ; 4 .
most importantly , it seems that the magnetar and non - magnetar samples are different .
the robust magnetar sample ( gold + silver ) grbs all have a beaming - corrected energy smaller than the maximum energy allowed by a magnetar , i.e. @xmath139 erg
. the non - magnetar sample , on the other hand , can exceed this limit .
the two samples have two distinct distributions in @xmath140 , @xmath91 , and @xmath141 , suggesting that they may be powered by different central engines .
second , both long and short grbs can be powered by a millisecond magnetar central engine .
the characteristic magnetar signature , an internal plateau , is found in both long and short grbs , suggesting that different progenitors ( both massive star core collapses and compact star mergers ) can produce a millisecond , probably supra - massive magnetar as the central engine .
the data is consistent that a long grb magnetar wind is collimated , while a short grb magnetar wind is essentially isotropic .
all these have profound implications in several related fields in high - energy , transient astronomy .
for example , if the recently discovered fast radio bursts ( frbs , lorimer et al . 2007 ; thornton et al .
2013 ) are indeed produced
when a supra - massive neutron star collapses into a black hole ( falcke & rezzolla 2014 ; zhang 2014 ) , our analysis suggests that such supra - massive neutron stars very likely do exist in grbs , and that the frb / grb association suggested by zhang ( 2014 ) should be quite common , probably up to near half of the entire grb population .
this is higher than the rate of plausible detections made by bannister et al .
( 2012 ) , but that low detection rate ( 2 out of 9 grbs , bannister et al .
2012 ) may be due to the sensitivity limit of the parkes 12 m telescope they have used
. a rapid - slewing larger radio telescope would be able to detect more frb / grb associations , which would open a new window to study cosmology ( deng & zhang 2014 ) and conduct cosmography ( gao et al . 2014 ) .
for another example , the conclusion that short grbs can be powered by a millisecond magnetar with a near isotropic magnetar wind would give rise to relatively bright , early electromagnetic counterparts of gravitational wave bursts due to ns - ns mergers ( zhang 2013 ; gao et al .
2013 ; yu et al . 2013 ; metzger & piro 2013 ; fan et al .
2013 ) , which gives promising prospects of detecting electromagnetic counterparts of gravitational wave signals in the advanced ligo / virgo era .
our analysis also poses some curious questions .
one is regarding the magnetar dissipation efficiency @xmath132 .
the results seem to suggest that some magnetars are efficient in dissipating their magnetar wind energy to x - ray radiation , while most others are not .
a straightforward inference would be that there might be a dichotomy within the magnetar central engines .
a more plausible scenario would be that some ( or probably ) most normal plateaus ( those followed by normal decays ) could be also dominated by internal dissipation emission ( e.g. ghisellini et al . 2007 ;
kumar et al .
2008a , b ) .
they are not identified as internal plateaus because their post - break decay is not steep enough .
physically they may be stable magnetars or supra - massive magnetars with a much later collapsing time , so that the collapsing signature ( very steep decay ) is not detected .
if so , the @xmath132 distribution may be more spread out , without a clear bimodal distribution .
this possibility is worth exploring in the future .
another mystery is regarding collimation of magnetar wind in short grbs .
our analysis suggests that at late times the magnetar wind is essentially isotropic . on the other hand , during the prompt emission phase ,
at least some short grbs show evidence of collimation ( e.g. burrows et al . 2006 ; soderberg et al .
2006 ; berger 2013 for a review ) .
there is no well studied short grb prompt emission model within the magnetar central engine scenario . suggested scenarios
invoke an early brief accretion phase ( metzger et al .
2008 ) , an early brief differential rotation phase ( fan et al . 2013 ) , or an early brief phase - transition phase ( e.g. cheng & dai 1996 ; chen & labun 2013 ) .
the short grb could be collimated by the torus within the accretion scenario ( bucciantini et al .
2012 ) . upon finishing this paper , we were drawn attention to yi et al .
( 2014 ) , who performed an independent analysis on a sub - sample of grb magnetar candidates ( essentially our gold sample ) .
they assumed that the long grb magnetar winds are isotropic and used the data to constrain magnetar wind dissipation efficiencies . through a comparison with x - ray emission efficiency of spin - down powered pulsars , they offer support to the millisecond magnetar central engine model from a different point of view .
we acknowledge the use of the public data from the swift data archive and the uk swift science data center .
we thank an anonymous referee for helpful comments , he gao and en - wei liang for helpful discussion , and zi - gao dai , xue - feng wu , and shuang - xi yi for sharing their paper with us and related discussion .
this work is supported by the nasa adap program under grant nnx10ad48 g .
060202 & 198.9 & [email protected] & [email protected] & [email protected] & ( 0.28,1.7 ) & [email protected] & [email protected]&[email protected] & 563/521& + 060413 & 147.7 & [email protected] & [email protected] & [email protected] & ( 1.2,253.52)&[email protected] & [email protected] & [email protected] & 79/71& + 060522 & 71.1 & [email protected] & [email protected] & [email protected] & ( 0.2,0.9 ) & [email protected] & [email protected] & [email protected] & 12/11&@xmath143 + 060607a & 102.2&[email protected]&[email protected]&[email protected]&(1.52,39.52)&[email protected]&[email protected] & [email protected] & 132/139&@xmath144 + 070110 & 88.4 & [email protected] & 18@xmath1422 & [email protected] & ( 4.1,28.72 ) & [email protected] & [email protected] & [email protected]&44/46&@xmath145 + 070616 & 402.4 & [email protected] & [email protected] & [email protected] & ( 0.13,2.01 ) & [email protected] & [email protected] & [email protected] & 224/241& + 090419 & 450 & [email protected] & 25@xmath1422 & [email protected] & ( 0.12,1.72 ) & [email protected] & [email protected] & [email protected] & 77/72& + 120213a & 49 & [email protected] & 19@xmath1421 & [email protected] & ( 1.04,12.84)&[email protected] & [email protected] & [email protected]&49/53& + 130102a & 77.5 & [email protected] & [email protected] & [email protected] & ( 0.18,10 ) & [email protected] & [email protected] & [email protected] & 12/10& + silver&(those&with&measured & redshifts & only ) + 050401 & 33.3 & [email protected] & [email protected] & [email protected] & ( 0.13,548 ) & [email protected] & [email protected] & [email protected] & 107/92&@xmath146 + 050505 & 58.9 & [email protected] & [email protected] & [email protected] & ( 2.88,133 ) & [email protected] & [email protected] & [email protected] & 27/45&@xmath147 + 050803 & 87.9 & [email protected] & [email protected] & [email protected] & ( 0.32,1330)&[email protected]&[email protected] & [email protected] & 95/75&@xmath148 + 060108 & 14.3 & [email protected] & [email protected] & [email protected] & ( 0.75,368 ) & [email protected] & [email protected] & [email protected] & 7/7&@xmath149 + 060526 & 298.2 & [email protected] & [email protected] & [email protected] & ( 1.02,314 ) & [email protected] & [email protected] & [email protected] & 34/48&@xmath150 + 060604 & 95 & [email protected] & [email protected] & [email protected] & ( 1.23,824 ) & [email protected] & [email protected] & [email protected] & 35/41&@xmath151 + 060605 & 79.1 & [email protected] & [email protected] & [email protected] & ( 0.15,103 ) & [email protected] & [email protected] & [email protected] & 16/21&@xmath152 + 060614 & 108.7 & [email protected] & [email protected] & [email protected] & ( 4.54,1795)&[email protected] & [email protected] & [email protected] & 70/54&@xmath153 + 060729 & 115.3 & [email protected] & [email protected] & [email protected] & ( 0.52,8968)&[email protected] & [email protected] & [email protected] & 160/459&@xmath154 + 060906 & 43.5 & [email protected] & [email protected] & [email protected] & ( 0.42,258 ) & [email protected]&[email protected] & [email protected] & 5/7&@xmath155 + 060908 & 19.3 & [email protected] & [email protected] & [email protected] & ( 0.08,14.8)&[email protected]&[email protected] & [email protected] & 98/59&@xmath156 + 061110a & 40.7 & [email protected] & [email protected] & [email protected] & ( 3.08,756 ) & [email protected]&[email protected] & [email protected] & 7/5&@xmath157 + 070306 & 209.5 & [email protected] & [email protected] & [email protected] & ( 0.48,819 ) & [email protected] & [email protected] & [email protected] & 153/132&@xmath158 + 070529 & 109.2 & [email protected] & [email protected] & [email protected] & ( 0.17,445 ) & [email protected] & [email protected] & [email protected] & 23/19&@xmath159 + 080605 & 20 & [email protected] & 133@xmath1422 & [email protected] & ( 0.09,101 ) & [email protected]&[email protected] & [email protected] & 330/289&@xmath160 + 080607 & 79 & [email protected] & 240@xmath1429 & [email protected] & ( 0.62,401 ) & [email protected]&[email protected] & [email protected] & 103/98&@xmath161 + 080721 & 16.2 & [email protected] & 120@xmath14210 & [email protected] & ( 0.11,2011)&[email protected] & [email protected] & [email protected] & 54/49&@xmath162 + 080905b & 128 & [email protected] & 18@xmath1422 & [email protected] & ( 0.22,988 ) & [email protected] & [email protected] & [email protected] & 94/98&@xmath163 + 081008 & 185.5 & [email protected] & 43@xmath1422 & [email protected] & ( 0.71,502 ) & [email protected]&[email protected] & [email protected] & 33/38&@xmath164 + 081203a & 294 & [email protected] & 77@xmath1423 & [email protected] & ( 0.2,506 ) & [email protected]&[email protected] & [email protected] & 191/163&@xmath165 + 081221 & 34 & [email protected] & 181@xmath1423 & [email protected] & ( 0.25,498 ) & [email protected] & [email protected] & [email protected] & 285/312&@xmath166 + 090423 & 10.3 & [email protected] & [email protected] & [email protected] & ( 0.39,501 ) & [email protected]&[email protected] & [email protected] & 27/33&@xmath167 + 090618 & 113.2 & [email protected] & 1050@xmath14210 & [email protected] & ( 0.58,1998)&[email protected] & [email protected] & [email protected] & 128/132&@xmath168 + 090927 & 2.2 & [email protected] & 2@xmath142 0.3 & [email protected] & ( 2.52,1003)&[email protected] & [email protected] & [email protected] & 19/15&@xmath169 + 091208b & 14.9 & [email protected] & 33@xmath1422 & [email protected] & ( 0.14,969 ) & [email protected] & [email protected] & [email protected] & 79/68&@xmath170 + 100418a & 7 & [email protected] & [email protected] & [email protected] & ( 0.51,2002)&[email protected]&[email protected]&[email protected] & 44/49&@xmath171 + 111008a & 63.5 & [email protected] & 53@xmath1423 & [email protected] & ( 0.31,987 ) & [email protected] & [email protected] & [email protected] & 143/167&@xmath172 + 111228a & 101.2 & [email protected] & 85@xmath1422 & [email protected] & ( 0.42,2990 ) & [email protected] & [email protected] & [email protected] & 202/187&@xmath173 + 120422a & 5.35 & [email protected] & [email protected] & [email protected] & ( 0.49,2011)&[email protected]&[email protected]&[email protected] & 4/6&@xmath174 + 121024a & 69 & [email protected] & 11@xmath1421 & [email protected] & ( 2.01,504 ) & [email protected]&[email protected] & [email protected] & 47/52&@xmath175 + 121027a & 62.6 & [email protected] & 20@xmath1421 & [email protected] & ( 40.1,3019)&[email protected]&[email protected]&[email protected] & 54/46&@xmath176 + 121128a & 23.3 & [email protected] & 69@xmath1424 & [email protected] & ( 0.21,98.7)&[email protected]&[email protected] & [email protected] & 81/78&@xmath177 + 121229a & 100 & [email protected] & [email protected] & [email protected] & ( 2.04,205 ) & [email protected]&[email protected] & [email protected] & 3/5&@xmath178 + sgrbs&(those&with&measured & redshifts & only ) + 051221a & 1.4 & [email protected] & [email protected] & [email protected] & ( 6.02,655 ) & [email protected] & [email protected] & [email protected] & 41/44&@xmath179 + 060801 & 0.49 & [email protected] & [email protected] & [email protected] & ( 0.08,0.73)&[email protected] & [email protected] & [email protected] & 22/18&@xmath180 + 061201 & 0.6 & [email protected] & [email protected] & [email protected] & ( 0.11,30.9)&[email protected] & [email protected] & [email protected] & 16/18&@xmath181 + 070809 & 1.3 & [email protected] & [email protected] & [email protected] & ( 0.53,67.4)&[email protected] & [email protected]&[email protected] & 33/26&@xmath182 + 090426 & 1.2 & [email protected] & [email protected] & [email protected] & ( 0.13,17.6)&[email protected] & [email protected]&[email protected] & 25/19&@xmath183 + 090510 & 0.3 & [email protected] & [email protected] & [email protected] & ( 0.11,20.7)&[email protected] & [email protected] & [email protected] & 76/68&@xmath184 + 100724a & 1.4 & [email protected] & [email protected] & [email protected] & ( 0.38,0.89)&[email protected] & [email protected] & [email protected] & 45/33&@xmath185 + 101219a & 0.6 & [email protected] & [email protected] & [email protected] & ( 0.05,0.27)&[email protected] & [email protected] & [email protected] & 38/29&@xmath186 + 130603b & 0.18 & [email protected] & [email protected] & [email protected] & ( 0.07,48.1)&[email protected] & [email protected] & [email protected] & 111/98&@xmath187 + gold + 060522 & 5 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 060607a & 5 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 070110 & 5 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] + silver + 050401 & 5 & @xmath188&[email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 050505 & @xmath189 & @xmath190 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] + 050803 & 5 & @xmath191 & ( [email protected])e-4 & [email protected] & [email protected] & [email protected]&[email protected] & [email protected]&[email protected] + 060108 & 5 & @xmath192 & ( [email protected])e-3 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 060526 & @xmath193 & @xmath194&([email protected])e-2 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 060604 & 5 & @xmath195 & ( [email protected])e-2 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 060605 & @xmath196 & @xmath197 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 060614 & @xmath198 & @xmath199 & ( [email protected])e-5 & [email protected] & [email protected] & [email protected]&[email protected]&[email protected]&[email protected] + 060729 & @xmath200 & @xmath201&([email protected])e-3 & [email protected] & [email protected] & [email protected]&[email protected]&[email protected]&[email protected] + 060906 & @xmath202 & @xmath203 & ( [email protected])e-2 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 060908 & @xmath204 & @xmath205 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 061110a & 5 & @xmath206 & ( [email protected])e-5 & [email protected] & [email protected] & [email protected]&[email protected]&[email protected]&[email protected] + 070306 & @xmath207 & @xmath208 & ( [email protected])e-2 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 070529 & 5 & @xmath209 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 080605 & 5 & @xmath210 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 080607 & 5 & @xmath211&[email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 080721 & 5 & @xmath212&[email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] + 080905b & 5 & @xmath213 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 081008 & 5 & @xmath214 & ( [email protected])e-2 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 081203a & 5 & @xmath215 & ( [email protected])e-2 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 081221 & 5 & [email protected]&[email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 090423 & @xmath216 & @xmath217 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 090618 & @xmath218 & @xmath219 & ( [email protected])e-2 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 090927 & 5 & [email protected] & ( [email protected])e-3 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 091208b & @xmath220 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] + 100418a & 5 & [email protected] & ( [email protected])e-4 & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] & [email protected] + 111008a & 5 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 111228a & 5 & [email protected] & ( [email protected])e-3 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 120422a & 5 & [email protected] & ( [email protected])e-6 & [email protected] & [email protected] & [email protected]&[email protected]&[email protected] & [email protected] + 121024a & 5 & [email protected] & ( [email protected])e-3 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 121027a & 5 & [email protected] & ( [email protected])e-3 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 121128a & 5 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + 121229a & 5 & [email protected] & ( [email protected])e-3 & [email protected] & [email protected] & [email protected] & [email protected]&[email protected] & [email protected] + sgrbs + 051221a & - & @xmath221 & [email protected]&([email protected])e-3 & [email protected] & [email protected] & -&- & [email protected] + 060801 & - & @xmath222 & [email protected] & [email protected] & [email protected] & [email protected] & -&- & [email protected] + 061201 & - & @xmath223 & [email protected] & [email protected] & [email protected] & [email protected] & -&- & [email protected] + 070809 & - & @xmath224 & [email protected] & ( [email protected])e-3 & [email protected] & [email protected] & -&- & [email protected] + 090426 & - & @xmath225 & [email protected] & [email protected] & [email protected] & [email protected] & -&- & [email protected] + 090510 & - & @xmath226 & [email protected] & [email protected] & [email protected] & [email protected] & -&- & [email protected] + 100724a & - & @xmath227 & [email protected] & [email protected] & [email protected] & [email protected] & -&- & [email protected] + 101219a & - & @xmath228 & [email protected] & [email protected] & [email protected] & [email protected] & -&- & [email protected] + 130603b & - & @xmath229 & [email protected] & [email protected] & [email protected] & [email protected] & -&- & [email protected] + @xmath54 & ( @xmath230 ) erg & ( @xmath231 ) erg & ( @xmath232 ) erg + @xmath61 & ( @xmath233 ) erg & ( @xmath234 ) erg & ( @xmath235 ) erg + @xmath236 & ( @xmath237 ) erg & ( @xmath238 ) erg & ( @xmath239 ) erg + & silver & silver+aluminum & non - magnetar + @xmath240 & ( @xmath241 ) erg & ( @xmath242 ) erg & ( @xmath243 ) erg + @xmath91 & ( @xmath244 ) erg & ( @xmath245 ) erg & ( @xmath246 ) erg + @xmath247 & ( @xmath248 ) erg & ( @xmath249 ) erg & ( @xmath250 ) erg + vs. @xmath34 for the `` silver '' and `` aluminum '' samples .
the three solid lines indicate the closure relations of three specific external shock models invoking energy injection with the parameter @xmath12 , as is expected in the millisecond magnetar central engine model .
the colored data points belong to the silver sample , while grey data points belong to the aluminum sample . ] against spectral index @xmath37 along with the closure relations of the external shock models for the `` silver '' sample .
( a ) the case of the ism model : the solid line ( pre- jet break ) and the shaded region ( post jet break ) are for the spectral regime i ( @xmath252 ) , while the dashed line ( pre- jet break ) and hatched region ( post jet break ) are for the spectral regime ii ( @xmath253 ) .
half - solid ( black ) dots and solid ( red ) dots are for regime i and ii , respectively .
( b ) the case of the wind medium case .
same conventions , except that triangles ( blue ) denote the spectral regime ii.,title="fig : " ] against spectral index @xmath37 along with the closure relations of the external shock models for the `` silver '' sample .
( a ) the case of the ism model : the solid line ( pre- jet break ) and the shaded region ( post jet break ) are for the spectral regime i ( @xmath252 ) , while the dashed line ( pre- jet break ) and hatched region ( post jet break ) are for the spectral regime ii ( @xmath253 ) .
half - solid ( black ) dots and solid ( red ) dots are for regime i and ii , respectively .
( b ) the case of the wind medium case .
same conventions , except that triangles ( blue ) denote the spectral regime ii.,title="fig : " ] vs. surface polar cap magnetic field strength @xmath0 derived for different magnetar samples : gold ( red hexagons ) , silver ( green diamonds ) , aluminum ( grey ) , and short grbs ( blue ) .
( a ) the case of isotropic winds ; ( b ) the case with beaming corrections .
the vertical solid line is the breakup spin - period for a neutron star ( lattimer & prakash 2004 ) .
, title="fig : " ] vs. surface polar cap magnetic field strength @xmath0 derived for different magnetar samples : gold ( red hexagons ) , silver ( green diamonds ) , aluminum ( grey ) , and short grbs ( blue ) .
( a ) the case of isotropic winds ; ( b ) the case with beaming corrections .
the vertical solid line is the breakup spin - period for a neutron star ( lattimer & prakash 2004 ) .
, title="fig : " ] and @xmath116 correlations for the grbs in various magnetar samples .
the color convention is the same as fig.5 .
the solid line is a power - law fitting to the gold and silver sample grbs , and the two dashed lines denote the 2@xmath254 region of the fits.,title="fig : " ] and @xmath116 correlations for the grbs in various magnetar samples .
the color convention is the same as fig.5 .
the solid line is a power - law fitting to the gold and silver sample grbs , and the two dashed lines denote the 2@xmath254 region of the fits.,title="fig : " ] , @xmath61 , @xmath255 , @xmath256 , @xmath91 , and @xmath257 , respectively.,title="fig : " ] , @xmath61 , @xmath255 , @xmath256 , @xmath91 , and @xmath257 , respectively.,title="fig : " ] , @xmath61 , @xmath255 , @xmath256 , @xmath91 , and @xmath257 , respectively.,title="fig : " ] , @xmath61 , @xmath255 , @xmath256 , @xmath91 , and @xmath257 , respectively.,title="fig : " ] , @xmath61 , @xmath255 , @xmath256 , @xmath91 , and @xmath257 , respectively.,title="fig : " ] , @xmath61 , @xmath255 , @xmath256 , @xmath91 , and @xmath257 , respectively.,title="fig : " ] scattered plot for all the grbs with redshift measurements in our samples : gold ( red ) , silver ( green ) , aluminum ( grey ) , and non - magnetar ( black ) .
slanted dashed lines mark the constant @xmath46-ray efficiency ( @xmath258 ) lines .
@xmath61 is calculated at @xmath29 ; ( b ) same as ( a ) , but with @xmath259 calculated at @xmath93 ; ( c ) the @xmath260 scattered plot for the magnetar samples .
gold ( red ) , silver ( green ) , and aluminum ( grey ) .
the constant x - ray efficiency @xmath132 lines are over plotted . the @xmath133 value of silver and aluminum sample grbs are all upper limits . for one gold sample grb
, @xmath105 is an upper limit ( denoted in the figure).,title="fig : " ] scattered plot for all the grbs with redshift measurements in our samples : gold ( red ) , silver ( green ) , aluminum ( grey ) , and non - magnetar ( black ) .
slanted dashed lines mark the constant @xmath46-ray efficiency ( @xmath258 ) lines .
@xmath61 is calculated at @xmath29 ; ( b ) same as ( a ) , but with @xmath259 calculated at @xmath93 ; ( c ) the @xmath260 scattered plot for the magnetar samples .
gold ( red ) , silver ( green ) , and aluminum ( grey ) .
the constant x - ray efficiency @xmath132 lines are over plotted .
the @xmath133 value of silver and aluminum sample grbs are all upper limits . for one gold sample grb , @xmath105 is an upper limit ( denoted in the figure).,title="fig : " ] scattered plot for all the grbs with redshift measurements in our samples : gold ( red ) , silver ( green ) , aluminum ( grey ) , and non - magnetar ( black ) . slanted dashed lines mark the constant @xmath46-ray efficiency ( @xmath258 ) lines .
@xmath61 is calculated at @xmath29 ; ( b ) same as ( a ) , but with @xmath259 calculated at @xmath93 ; ( c ) the @xmath260 scattered plot for the magnetar samples .
gold ( red ) , silver ( green ) , and aluminum ( grey ) .
the constant x - ray efficiency @xmath132 lines are over plotted . the @xmath133 value of silver and aluminum sample grbs are all upper limits . for one gold sample grb
, @xmath105 is an upper limit ( denoted in the figure).,title="fig : " ] , @xmath94 and @xmath132 of our samples . for @xmath132 ,
the silver and aluminum samples only give upper limits.,title="fig : " ] , @xmath94 and @xmath132 of our samples . for @xmath132 ,
the silver and aluminum samples only give upper limits.,title="fig : " ] , @xmath94 and @xmath132 of our samples . for @xmath132 ,
the silver and aluminum samples only give upper limits.,title="fig : " ] vs. several parameters : @xmath261 , @xmath47 , @xmath61 , and @xmath127 .
color conventions are the same as fig.5 .
the @xmath132 values of all silver and aluminum sample grbs are all upper limits .
the blue arrow shows the lower limit of one grb in the gold sample.,title="fig : " ] vs. several parameters : @xmath261 , @xmath47 , @xmath61 , and @xmath127 .
color conventions are the same as fig.5 .
the @xmath132 values of all silver and aluminum sample grbs are all upper limits .
the blue arrow shows the lower limit of one grb in the gold sample.,title="fig : " ] vs. several parameters : @xmath261 , @xmath47 , @xmath61 , and @xmath127 .
color conventions are the same as fig.5 .
the @xmath132 values of all silver and aluminum sample grbs are all upper limits .
the blue arrow shows the lower limit of one grb in the gold sample.,title="fig : " ] vs. several parameters : @xmath261 , @xmath47 , @xmath61 , and @xmath127 .
color conventions are the same as fig.5 .
the @xmath132 values of all silver and aluminum sample grbs are all upper limits .
the blue arrow shows the lower limit of one grb in the gold sample.,title="fig : " ] | a rapidly spinning , strongly magnetized neutron star ( magnetar ) has been proposed as one possible candidate of the central engine of gamma - ray bursts ( grbs ) .
we systematically analyze the swift / xrt light curves of long grbs detected before 2013 august , and characterize them into four categories based on how likely they may harbor a magnetar central engine : gold , silver , aluminum , and non - magnetar .
we also independently analyze the data of short grbs with a putative magnetar central engine .
we then perform a statistical study of various properties of the magnetar samples and the non - magnetar sample , and investigate whether the data are consistent with the hypothesis that there exist two types of central engines . by deriving the physical parameters of the putative magnetars
, we find that the observations of the gold and silver samples are generally consistent with the predictions of the magnetar model .
for a reasonable beaming factor for long grbs , the derived magnetar surface magnetic field @xmath0 and initial spin period @xmath1 fall into the reasonable range .
magnetar winds in short grbs , on the other hand , are consistent with being isotropic .
no grb in the magnetar sample has a beam - corrected total energy exceeding the maximum energy budget defined by the initial spin energy of the magnetar , while some non - magnetar grbs do violate such a limit . with beaming correction , on average
the non - magnetar sample is more energetic and luminous than the magnetar samples .
our analysis hints that millisecond magnetars are likely operating in a good fraction , but probably not all , grbs . |
the phenix experiment at the relativistic heavy ion collider ( rhic ) reported that the transverse single spin asymmetry , denoted @xmath0 , for forward neutrons measured in transversely polarized proton
nucleus ( @xmath1 ) collisions at @xmath2 is far different from that in polarized proton
proton ( @xmath6 ) collisions at @xmath7 @xcite .
that is , the measured @xmath0 for @xmath3 and @xmath4 collisions are @xmath8 and @xmath9 , respectively . on the other hand , @xmath0 is @xmath10 in @xmath6 collisions @xcite . in these measurements , the neutrons produced in @xmath6 and @xmath1
collisions are detected by a zero - degree calorimeter @xcite in the polarized proton remnant side that is defined as the positive rapidity region . as studied in ref .
@xcite , the interference of pion ( spin - flip ) and @xmath11-reggeon ( spin nonflip ) exchanges successfully explains @xmath0 for forward neutrons in @xmath6 collisions at rhic .
we would expect that this mechanism can be extended to predict @xmath0 for @xmath1 collisions .
the authors of ref .
@xcite incorporate the pion@xmath11-reggeon interference with strong nuclear absorptive corrections and a nuclear breakup @xcite .
however , the predicted @xmath0 in @xmath1 collisions remains negative and the magnitude of @xmath0 is too small to explain the phenix results . in this paper
, we propose an alternative mechanism : ultraperipheral proton nucleus collisions ( upcs ) @xcite .
upcs occur when the impact parameter @xmath12 is larger than the sum of the radii of each colliding particle , namely , @xmath13 , where @xmath14 and @xmath15 are the radius of the proton and nucleus , respectively . as we explored in ref .
@xcite , upcs have a comparable cross section with the hadronic interactions in the very forward rapidity region . in ultraperipheral @xmath1 collisions , virtual photons ( @xmath16 ) emitted from the relativistic nucleus
interact with polarized protons .
thus the number of neutrons produced via the @xmath17 interaction depends on the azimuthal angle of the scattered neutrons relative to the proton polarization .
this may finally contribute to the large @xmath0 for forward neutrons in @xmath1 collisions .
we first develop the monte carlo ( mc ) simulation framework for upcs ( sec .
[ sec : upcsim ] ) and hadronic interactions ( sec . [ sec : qcdsim ] ) . in sec .
[ sec : results ] , using these mc simulations we show that upcs in @xmath1 collisions have sizable cross sections compared with hadronic interactions and the yield of forward neutrons in upcs certainly depend on the scattering azimuthal angle relative to the proton polarization axis . finally , in sec .
[ sec : discussions ] , we compare the simulated @xmath0 with the phenix results .
conclusions are drawn in sec .
[ sec : conclusions ] .
natural units @xmath18 are used throughout .
the mc simulation for upcs in this study comprises two steps .
first , we simulate the virtual photon flux as a function of the photon energy and impact parameter by using starlight @xcite ( sec .
[ sec : photon ] ) .
second , the simulation of the @xmath17 interaction and particle production is performed following the differential cross sections that are predicted by the @xmath19 model @xcite ( sec .
[ sec : photon+p ] ) .
the differential cross section for single neutron production in @xmath1 upcs is given by @xmath20 where @xmath21 is the number of the emitted photons , @xmath22 is the @xmath17 center - of - mass energy , @xmath23 with the neutron scattering polar angle @xmath24 and azimuthal angle @xmath25 in the @xmath17 center - of - mass frame , @xmath26 is the total cross section for a single photon interaction with a proton leading to single neutron production , and @xmath27 is the probability of having no hadronic interactions in @xmath1 collisions at given @xmath12 .
we calculate @xmath28 and @xmath29 in sec .
[ sec : photon ] and sec .
[ sec : photon+p ] , respectively .
a finite probability for having no hadronic interactions @xmath27 is introduced in order to account for a smooth cut off around the impact parameter @xmath30 @xcite . @xmath14 and @xmath15 , where @xmath31 and @xmath32 , are the radius of the proton and nucleus , respectively .
the range of the impact parameter considered in the simulation extends from @xmath33 to @xmath34 .
the value of @xmath35 is well below the sum of the effective radii of colliding particles , and @xmath27 rapidly approaches zero below these nuclear radii . for comparisons with the simulation results of hadronic interactions in sec .
[ sec : results ] and the experimental results from phenix in sec .
[ sec : discussions ] , we will numerically transform the differential cross section in eq .
( [ eq : dsigupc ] ) based on the @xmath17 center - of - mass frame to that in the detector reference frame .
both frames are defined in fig .
[ fig : frame ] . for simplicity ( unless otherwise noted ) ,
the discussion in this subsection is based on the proton rest frame .
the virtual photons flux emitted by the relativistic nucleus is simulated using starlight , which follows the weizscker - williams approximation @xcite .
the double differential photon flux due to the fast moving nucleus with velocity @xmath36 is written as @xmath37 where @xmath38 is the photon energy , @xmath39 is the atomic number ( @xmath40 and @xmath41 for al and au , respectively ) , @xmath42 is the fine structure constant , @xmath43 ( @xmath44 is the lorentz factor ) , and @xmath45 and @xmath46 are the modified bessel functions . in the case of a relativistic nucleus ( @xmath47 ) , we safely disregard the contribution of the term @xmath48 in eq .
( [ eq : gflux ] ) .
the photon energy @xmath38 in the proton rest frame is properly transformed to the @xmath17 center - of - mass energy @xmath22 to allow substitution of eq .
( [ eq : gflux ] ) for eq .
( [ eq : dsigupc ] ) .
the kinematics of the @xmath17 interaction is shown in fig .
[ fig : frame](b ) and is defined as @xmath49 where the variables in brackets indicate the four - momenta of each particle .
we use the following notations for these four - momenta : @xmath50 where the @xmath17 center - of - mass energy is given by @xmath51 . figure [ fig : frame](b ) also introduces the polar angle @xmath52 and azimuthal angle @xmath53 of @xmath54 , with reference to a coordinate system with @xmath55 , 1 , and 2 axes such that @xmath55 lies in the 1 - 3 plane .
the proton is transversely polarized along the 2 axis .
the frame @xmath56 is the scattering plane . the frame @xmath57 is defined such that the @xmath58 axis is directed into the @xmath55 direction ; the @xmath59 axis is perpendicular to the @xmath60@xmath58 reaction plane ; and the @xmath60 axis is given by @xmath61 .
single neutron and pion production from the @xmath17 interaction are simulated following the differential cross sections predicted by the @xmath19 model .
the cross section of the @xmath62 interaction is formed as in ref .
@xcite : @xmath63 where @xmath64 and @xmath65 are the response functions for pion photoproduction , and @xmath66 and @xmath67 are the proton polarization along @xmath59 and @xmath68 axes , respectively . in the third equation , @xmath69 and @xmath66
respectively are replaced with target asymmetry @xmath70 and @xmath71 .
we assume @xmath72 and @xmath73 in this study .
note that we can numerically obtain @xmath74 in eq .
( [ eq : dsigupc ] ) from @xmath75 in eq .
( [ eq : gpxsec ] ) with the relation @xmath76 and @xmath77 .
the photon virtuality is limited by @xmath78 , thus @xmath79 in @xmath3 collisions and @xmath80 in @xmath4 collisions . in the following simulations , we fix @xmath81 as default and take into account effects of nonzero @xmath82 as a model uncertainty .
the virtual photons are assumed to be fully unpolarized in this study .
both baryon resonance and nonresonance contributions are taken into account for neutron and pion productions in @xmath19 .
the model contains all four - star resonances with masses below and follows the breit - wigner forms for the resonance shape .
nonresonance background contribution contains the born terms as detailed in ref .
@xcite . due to the energy range guaranteed by @xmath19 ,
the minimum and maximum @xmath22 values in the upc mc simulation are set as @xmath83 and @xmath84 , respectively .
we will discuss cross section contribution from outside the above energy range in sec .
[ sec:@xmath85@xmath86au ] .
figure [ fig : t ] shows @xmath70 of the @xmath62 interaction as a function of @xmath22 . in the detector reference frame ,
the thick and thin curves correspond to the rapidity of produced neutrons of @xmath87 and 8.0 , respectively . because the neutrons produced along the negative @xmath58 axis , namely , @xmath88 ,
are more likely to have large positive rapidity in the detector reference frame , the thick and thin curves shift to the large @xmath52 region . of the @xmath62 interaction as function of @xmath22 .
in the detector reference frame , the thick and thin curves correspond to the rapidity of produced neutrons @xmath87 and 8.0 , respectively . ]
throughout this section , we use the detector reference frame defined in fig . [
fig : frame](a ) .
we effectively obtain the differential cross section of forward neutron production in @xmath1 hadronic interactions as follows .
first ( sec . [ sec:@xmath85@xmath86@xmath85 ] ) , we calculate the cross section of inclusive neutrons in @xmath90 collisions , @xmath91 , using a simple one - pion exchange model .
note that this calculation is performed for unpolarized protons .
the one - pion exchange model has well described forward neutron production in @xmath90 collisions at the intersecting storage rings @xcite and rhic @xcite and in @xmath92 collisions at the hadron - electron ring accelerator ( hera ) @xcite .
second ( sec .
[ sec : glauber ] ) , the cross section of the @xmath93 interaction , @xmath94 , can be calculated by @xmath91 and the gribov - glauber model @xcite . here
we avoid an implementation of multiple scattering of a projectile proton with a nucleus , and instead we multiply the @xmath90 cross section @xmath91 with the inelastic cross section ratio @xmath95 obtained from ref .
@xcite . for the @xmath96-dependence of the differential cross section ,
we multiply @xmath97 with @xmath98 in order to effectively take into account the single spin asymmetry @xmath99 ( sec . [ sec : glauber ] ) .
the differential cross section for inclusive neutrons in @xmath90 collisions at the center - of - mass energy @xmath100 as a function of the longitudianl momentum fraction @xmath101 and the transverse momentum @xmath102 is formed in terms of the pion - exchange model @xcite as @xmath103 where @xmath104 is the rapidity gap survival factor , @xmath105 is the pion trajectory with the slope @xmath106 and the pion mass @xmath107 , @xmath108 is the four - momentum transfer squared , @xmath109 is the effective vertex function , @xmath110 is the phase factor @xcite , and @xmath111 is the total cross section of the @xmath112 interactions at the @xmath113 center - of - mass energy @xmath114 . the effective vertex function is parameterized as @xmath115 using the pion nucleon coupling @xmath116 and the @xmath108-slope parameter @xmath117 . in this study
, we fix @xmath118 and @xmath119 which are consistent with the results at hera @xcite , and follow the best compete fit results @xcite for @xmath111 .
because the parameters @xmath104 and @xmath117 have been poorly determined to date , we use @xmath120 and @xmath121 that derive the best agreement with the forward neutron @xmath122 distribution measured at the phenix experiment @xcite .
these best - fit values are compatible with other experimental results @xcite . as introduced in the third paragraph of sec .
[ sec : qcdsim ] , we avoid an implementation of multiple scattering of a projectile proton with a nucleus . on the other hand
, we effectively obtain the @xmath123 cross sections @xmath94 by multiplying @xmath91 in eq .
( [ eq : dsighadpp ] ) with the inelastic cross section ratio @xmath124 that is calculated in ref .
thus we obtain : @xmath125 the single spin asymmetry for forward neutrons in @xmath6 interactions originate in the interference of pion ( spin - flip ) and @xmath11-reggeon ( spin nonflip ) exchanges @xcite that well reproduces the result from the phenix experiment : @xmath126 @xcite .
preliminary results in ref .
@xcite based on the same approach as ref .
@xcite state that single spin asymmetry for forward neutrons in hadronic @xmath1 collisions is also described by the pion@xmath11-reggeon interference followed by a nuclear breakup . here
, we do not implement the pion@xmath11 interference in the simulation .
instead , we multiply the @xmath123 differential cross section in eq .
( [ eq : dsighadpa ] ) by @xmath127 , where we take @xmath128 and @xmath129 from ref . @xcite . finally , we obtain using eq .
( [ eq : dsighadpa ] ) : @xmath130
first , we calculate the total cross section of the @xmath131 interaction at @xmath2 and compare it with upcs and hadronic interactions .
the total cross section for upcs is calculated by integrating eq .
( [ eq : dsigupc ] ) over @xmath132 , @xmath12 , and @xmath133 : @xmath134 where we require a single neutron scattered at @xmath135 and @xmath136 .
the rapidity limit corresponds to the acceptance of a zero - degree calorimeter at rhic and the @xmath101 limit is introduced to remove the contribution of low - energy forward neutrons .
these cuts are consistent with the rhic measurements @xcite . as addressed in sec .
[ sec : upc ] , we fix @xmath33 , @xmath137 , @xmath83 , and @xmath84 .
we then obtain @xmath138 .
for the discussions in sec .
[ sec:@xmath85@xmath86au_z ] and [ sec:@xmath85@xmath86au_sys ] , here we show the differential upc cross sections at @xmath2 as a function of @xmath22 in fig . [
fig : dsdw ] .
the @xmath139 values are calculated by integrating eq .
( [ eq : dsigupc ] ) over @xmath12 and @xmath133 . for simplicity ,
no kinematical limit is applied to such integration .
thick black curve indicates the @xmath140 interaction and thin blue curve indicates the two - pion production @xmath141 .
the total cross section for hadronic interaction is calculated by integrating eq .
( [ eq : dsighadpolpa ] ) over @xmath101 and @xmath102 : @xmath142 we obtain @xmath143 by requiring a single neutron emitted into @xmath135 and @xmath136 . according to the comparison of these two cross sections , we find that upcs lead to significant background contribution to the investigations of single spin asymmetry in terms of hadronic interactions . table [ tbl : singlexsec ] summarizes the calculated cross sections .
.cross sections for neutron production in ultra - peripheral collisions and hadronic interactions at @xmath2 .
cross sections in parentheses are calculated without @xmath144 and @xmath101 limits . [ cols="^,^,^,^ " , ] [ tbl : sys ] total cross sections for upcs and hadronic interactions in @xmath3 collisions are summarized in table [ tbl : singlexsec ] .
the upc cross section is @xmath145 which is @xmath146 of @xmath147 , where upcs in @xmath3 collisions are highly suppressed compared with those in @xmath4 collisions due to @xmath148 .
figure [ fig : dsdxf](c ) shows @xmath149 for upcs ( dashed [ red ] line ) and hadronic interactions ( solid [ black ] line ) .
we find upcs leading to subdominant contribution to the @xmath149 distribution at @xmath150 . finally , fig .
[ fig : dsdxf](d ) compares @xmath151 between upcs ( dashed [ red ] line ) and hadronic interactions ( solid [ black ] line ) .
although the upc cross section is about @xmath152 of the hadronic - interactions cross section , the large positive asymmetry of upcs eventually compensates the small negative asymmetry of hadronic interactions .
we compare the simulation results with the observed @xmath0 values in @xmath3 and @xmath4 collisions at @xmath2 .
figure [ fig : an ] shows @xmath0 as a function of the atomic number @xmath39 in @xmath6 ( for reference ) , @xmath3 and @xmath4 collisions .
filled [ black ] circles indicate the @xmath0 values inclusively measured by the phenix zero - degree calorimeter @xcite , where the neutron rapidity and @xmath101 ranges are limited by @xmath153 and @xmath136 , respectively .
these values can be compared with open [ red ] circles indicating the sum of upcs and hadronic interactions mc simulations , denoted @xmath154 .
these are written as @xmath155 since @xmath156 for the mc simulation results ( open [ red ] circles and open [ blue ] squares ) , the neutron rapidity and @xmath101 region limits , @xmath153 and @xmath136 , are also taken into account to be consistent with the phenix measurements . in @xmath3 collisions , we obtain @xmath157 which is consistent with the phenix result @xmath158 . in @xmath4 collisions ,
we have @xmath159 that can be understood by that upcs , having large positive @xmath160 and a cross section @xmath161 , significantly contribute to the inclusive @xmath154 value that are evident in fig .
[ fig : dsdxf ] and table [ tbl : singlexsec ] .
note that a model uncertainty in @xmath154 , estimated by taking account of nonzero @xmath82 discussed in sec .
[ sec:@xmath85@xmath86au_sys ] , amounts .
filled [ black ] squares in fig .
[ fig : an ] are the @xmath0 values measured by the phenix zero - degree calorimeter requiring a veto on the beam - beam counters ( bbcs ) covering @xmath162 @xcite . because a nucleus in upcs coherently scatters with a proton and thus does not generate underlying particles , such a bbc veto effectively selects upc - rich events . in @xmath4 collisions ,
the phenix data with the bbc veto has larger @xmath0 than the inclusive phenix data ( filled [ black ] circle ) .
this indicates that the fraction of upcs in the phenix data is enhanced at a certain level by the bbc veto , although the actual fraction is not presently reported .
conversely in @xmath3 collisions , the phenix data with the bbc veto provides @xmath163 which is far smaller than @xmath164 in @xmath4 collisions .
a possible inference for the difference is that the fraction of hadronic interaction is still sizable in the phenix data in @xmath3 collisions even though upc - rich events are preferentially selected by the bbc veto .
if pure upc data is experimentally available , the @xmath0 values may approach @xmath160 in both @xmath3 and @xmath4 collisions .
note that @xmath160 is same between @xmath3 and @xmath4 collisions , since @xmath160 depends only on the @xmath17 interactions which are common between @xmath3 and @xmath4 collisions . of forward neutron .
filled black marker indicates the phenix results .
open [ red ] circle and open [ blue ] squares indicate the asymmetry obtained by the sum of upcs and hadronic interactions and by only hadronic interactions , respectively . ]
it is demonstrated that ultraperipheral @xmath1 collisions have large @xmath0 for forward neutrons using the mc simulation framework developed for this study .
the present upc simulation comprised the following two parts ; first , the simulation of the virtual photon flux was performed by the starlight event generator and , second , the simulation of the @xmath62 interaction followed the differential cross sections predicted by @xmath19 unitary isobar model . in the @xmath17 interaction ,
the target asymmetry @xmath70 was appropriately treated . according to the mc simulations , we found upcs in @xmath1 collisions leading to @xmath165 .
concerning forward neutron production of @xmath1 hadronic interaction , the simulation model used an one - pion exchange model and the glauber model .
the single spin asymmetry was effectively taken in account by multiplying @xmath97 with @xmath166 where @xmath167 . combining the differential cross sections of upcs and hadronic interactions
, we simulated the @xmath101 and @xmath96 distributions for inclusive forward neutrons .
the @xmath0 values for inclusive neutrons at @xmath153 and @xmath136 were predicted as @xmath168 and @xmath169 in @xmath3 and @xmath4 collisions , respectively . these were consistent with the recently reported phenix results .
the phenix data with the bbc veto in @xmath1 collisions had larger @xmath0 than the inclusive phenix data , but were smaller than @xmath160 . this indicated that requiring the bbc veto enhanced the fraction of upcs in the phenix data at a certain level , although the actual fraction was presently unreported . for future analyses , we plan to extend the present simulation framework to include the contribution of the two - pion production @xmath141 .
this would provide a more accurate description of @xmath0 of forward neutrons .
another extension would be to investigate the possible interference between electromagnetic and hadronic interactions , which is known as coulomb - nuclear interference .
the author appreciates fruitful discussions with y. akiba , y. goto and i. nakagawa .
99 i. nakagawa _
et al_. , ( phenix collaboration ) , the first transverse single spin measurement in high energy polarized proton - nucleus collision at the phenix experiment at rhic , j. of phys .
* 736 * , 012017 ( 2016 ) .
et al_. , ( phenix collaboration ) , inclusive cross section and single transverse spin asymmetry for very forward neutron production in polarized @xmath85@xmath86@xmath85 collisions at @xmath7 , phys .
d * 88 * , 032006 ( 2013 ) . c. adler _
et al_. , the rhic zero degree calorimeters , nucl . instrum . meth . * a 470 * ( 2001 ) 488 .
b. z. kopeliovich , i. k. potashnikova , and ivn schmidt , single transverse spin asymmetry of forward neutrons , phys .
d * 84 * , 114012 ( 2011 ) .
b. z. kopeliovich , i. k. potashnikova , and ivn schmidt , leading neutrons from polarized proton
nucleus collisions , arxiv:1611.07365 . c. a. bertulani and g. baur , electromagnetic processes in relativistic heavy ion collisions , phys .
* 163 * ( 1988 ) 299 . c. a. bertulani , s. r. klein , and j. nystrand , physics of ultra - peripheral nuclear collisions , ann .
nucl . part .
* 55 * ( 2005 ) 271 .
g. mitsuka , forward hadron production in ultra - peripheral proton heavy - ion collisions at the lhc and rhic , eur .
j. c * 75 * , 614 ( 2015 ) .
s. r. klein , j. nystrand , j. seger , y. gorbunov , and j. butterworth , starlight : a monte carlo simulation program for ultra - peripheral collisions of relativistic ions , comput .
* 212 * ( 2017 ) 258 .
starlight webpage , https://starlight.hepforge.org/. d. drechsel , s. s. kamalov , and l. tiator , unitary isobar model maid2007 , eur .
j. a * 34 * , 64 . c. von weizscker , radiation emitted in collisions of very fast electrons , z. physik * 88 * ( 1934 ) 612 .
e. j. williams , nature of the high - energy particles of penetrating radiation and status of ionization and radiation formulae , phys .
* 45 * ( 1934 ) 729 .
d. drechsel and l. tiator , threshold pion photoproduction on nucleons , j. phys .
g : nucl . phys . * 18 * , 449 ( 1992 ) . w. flauger and f. mnnig , measurement of inclusive zero - angle neutron spectra at the cern isr , nucl .
b * 109 * , ( 1976 ) 347 .
s. chekanov _
et al_. , ( zeus collaboration ) , study of the pion trajectory in the photoproduction of leading neutrons at hera , phys .
b * 610 * 199 ( 2005 ) . v. n. gribov , interaction of gamma quanta and electrons with nuclei at high - energies , sov .
jetp * 30 * ( 1970 ) 709 .
r. j. glauber and g. matthiae , high - energy scattering of protons by nuclei , nucl .
* b 21 * ( 1970 ) 135 .
v. guzey and m. strikman , proton nucleus scattering and cross section fluctuations at rhic and lhc , phys . lett .
b * 633 * ( 2006 ) 245 .
b. z. kopeliovich , i. k. potashnikova , ivan schmidt , and j. soffer , damping of forward neutrons in@xmath85@xmath86@xmath85collisions , phys .
d * 78 * , 014031 ( 2008 ) .
f.d . aaron _
et al_. , ( h1 collaboration ) , measurement of leading neutron production in deep - inelastic scattering at hera , eur .
j. c * 68 * ( 2010 ) 381 .
k. a. olive _
et al_. ( particle data group ) , chin .
* c 38 * ( 2014 ) 090001 . s. chekanov _
et al_. , ( zeus collaboration ) , leading neutron production in @xmath170 collisions at hera , nucl .
b * 637 * , 3 ( 2002 ) .
a. fix and h. arenhvel , double pion photoproduction on nucleon and deuteron , eur . phys .
j. a * 25 * , 115 ( 2005 ) .
et al_. , phenix inner detectors , nucl . instrum
. meth . * a 499 * ( 2003 ) 549 . | the phenix experiment at the relativistic heavy ion collider recently reported that transverse single spin asymmetry , @xmath0 , for forward neutrons in @xmath1 collisions at @xmath2 .
@xmath0 in @xmath3 and @xmath4 collisions were measured as -0.015 and 0.18 , respectively .
these values are clearly different from the measured @xmath5 in @xmath6 collisions . in this paper
, we propose that a large @xmath0 for forward neutrons in ultraperipheral @xmath1 collisions may explain the phenix measurements .
the proposed model is demonstrated using two monte carlo simulations . in the ultraperipheral collision simulation
, we use the starlight event generator for the simulation of the virtual photon flux and then use the maid unitary isobar model for the simulation of the neutron production in the interactions of a virtual photon with a polarized proton . in the @xmath1 hadronic interaction simulation ,
the differential cross sections for forward neutron production are predicted by a simple one - pion exchange model and the glauber model .
the simulated @xmath0 values for both the contribution of ultraperipheral collisions and the hadronic interactions are in good agreement with the phenix results . |
ce based heavy fermion compounds cemin@xmath2 ( m = co , rh , ir ) , known as the 115 family , show a variety of ground states at ambient pressure . ceirin@xmath2 and cecoin@xmath2 are bulk superconductors with the transition temperatures @xmath20 = 0.4 k and 2.3 k , respectively @xcite .
filamentary superconductivity is observed in ceirin@xmath2 below @xmath21 1.2 k @xcite .
cerhin@xmath2 , an antiferromagnet with neel temperature @xmath22 = 3.8 k , shows ambient pressure superconductivity at very low temperatures , @xmath23 0.1 k @xcite and with the same value of @xmath23 2.5 k as cecoin@xmath2 under pressure of 2 gpa @xcite .
magnetic order can be induced in the superconducting members by cd and hg substitutions of in @xcite , so that phase diagrams with rich interplay of magnetism , quantum critical points ( qcp ) at @xmath24 @xcite and superconductivity are found as functions of pressure and composition ( see refs . for reviews ) .
doping studies found that ceirin@xmath2 @xcite , similar to cecoin@xmath2 @xcite , also lies in the vicinity of the magnetic quantum critical point , and thus its superconductivity is most likely magnetically mediated .
our studies of thermal conductivity in both cecoin@xmath2 @xcite and ceirin@xmath2 @xcite , though , found significant deviations from simple @xmath25-wave scenario , supported by many experiments @xcite . magnetic field can also be used as the non - thermal tuning parameter for quantum critical point , and in cecoin@xmath2 field - tuned qcp is mysteriously coinciding with the superconducting upper critical field for orientations of magnetic field along both @xmath3-axis @xcite and @xmath26-plane @xcite of the tetragonal crystals .
the existence of qcp is reflected in anomalous temperature dependences of electrical resistivity @xcite closely following @xmath27-linear dependence , of the specific heat showing logarithmic divergence , in band - dependent divergence of cyclotron effective masses @xcite and directional violation of wiedemann - franz law @xcite . in cerhin@xmath2 field - induced magnetic order
is found in the pressure range below @xmath28=2.5 gpa , corresponding to maximum superconducting @xmath20 @xcite .
interestingly , similar phenomenology of qcp coinciding with the first order transition at @xmath4 is found in closely related heavy fermion ce@xmath29pdin@xmath30 @xcite , as well as in some other unconventional superconductors , including cuprates @xcite and iron - based material kfe@xmath29as@xmath29 @xcite . of great interest
is the question if field tuned quantum criticality can be found in ceirin@xmath2 .
the two materials , cecoin@xmath2 and ceirin@xmath2 , have identical fermi surfaces @xcite and show very similar normal state properties @xcite , so it is natural to expect similarity in response to magnetic field .
despite this apparent similarity , heat capacity measurements @xcite do not indicate critical behavior near bulk upper critical field , finding nearly temperature - independent @xmath31 at the lowest temperatures . instead
, heat capacity , torque and resistivity studies suggest that ceirin@xmath2 shows metamagnetic transition with critical field of 25 t , far above superconducting @xmath4 @xcite .
these studies , though , reveal several unusual features . while the electronic term @xmath32 in heat capacity saturates at constant value below @xmath33 0.8 k in magnetic field of 1 t and
remains nearly constant with magnetic field @xmath34 in 1 t to 17 t range , with @xmath35 0.7 to 0.8 j / mole k@xmath36 @xcite , deviations from this value at high fields start at progressively lower @xmath37 @xcite .
another study suggested that the superconducting dome of ceirin@xmath2 is located inside the dome of precursor phase , similar to pseudogap state in the high @xmath20 cuprates @xcite , suppressed at a field of 6 t. we should keep in mind though that resistivity studies in ceirin@xmath2 can not probe the region close to bulk @xmath4 , hidden inside the domain of filamentary superconductivity @xcite .
subtraction of schottky anomaly in heat capacity measurements always leaves some uncertainty in the behavior at the lowest temperatures . to get an additional insight into response of ceirin@xmath2 to application of magnetic field ,
here we report field evolution of the electronic transport properties of ceirin@xmath2 using simultaneous electrical resistivity and thermal conductivity measurements , with the latter allowing to probe the range of filamentary superconductivity close to bulk @xmath4 .
we find that the wiedemann - franz law is universally obeyed in @xmath38 0 limit for all fields where resistivity measurements are not affected by filamentary superconductivity , proving true bulk nature of both electrical and heat transport .
for all fields above bulk @xmath4 we observe a fermi - liquid temperature dependence of thermal resistivity , @xmath39 , slightly decreasing with field @xmath16 coefficient and nearly field - independent @xmath37 .
notably smaller values of @xmath16 in ceirin@xmath2 when compared to cecoin@xmath2 , as well as their very mild dependence on magnetic field , suggest that ceirin@xmath2 is significantly further away from magnetic instability at @xmath4 and the field tuned criticality is significantly different in the two materials .
we link this difference to anomalously low value of kadowaki - woods ratio of the @xmath40 coefficient in electrical resistivity versus @xmath41 coefficient of the specific heat .
this suggests more localized character of magnetism in ceirin@xmath2 than in cecoin@xmath2 .
single crystals of ceirin@xmath2 were grown using self - flux method @xcite .
as - grown crystals were of thin plate - like shape , with large surfaces corresponding to the ( 001 ) basal plane of the tetragonal lattice .
two samples used in this study were from two different batches , but revealed nearly identical results .
samples were cut into rectangular shape with dimensions of @xmath42 mm@xmath43 and @xmath44 mm@xmath43 .
four contacts to the samples were made by soldering 50 @xmath45 silver wires with indium solder , which resulted in the contact resistance typically @xmath46 m@xmath47 at low temperature .
same contacts were used to measure electrical resistivity and thermal conductivity in a dilution refrigerator , thus essentially eliminating uncertainty of the geometric factor determination .
the thermal conductivity was measured using a standard four - probe steady - state method with two ruo@xmath29 chip thermometers calibrated _ in situ _ against a reference ge thermometer .
the electrical resistivity was measured using lr700 ac resistance bridge , operating at a frequency of 16 hz , by applying 0.1 ma excitation currents . to enable low - noise measurements of low resistance samples in identical conditions with thermal conductivity studies ,
measurements were performed in a separate dilution refrigerator run , during which the high - resistance wire coils of the thermal conductivity apparatus were temporary shorted with silver wires , see ref . for details . the phonon contribution to thermal conductivity , @xmath48 , was determined from electrical resistivity and thermal conductivity measurements in heavily disordered samples of ce@xmath49la@xmath50irin@xmath2 with @xmath51=0.2 and residual resistivity @xmath52 cm , see ref . for details . in these
heavily disordered samples temperature - dependent inelastic scattering contribution to resistivity can be neglected , and electronic contribution can be directly determined from wiedemann - franz law as @xmath53 , which gives @xmath54 .
this contribution represents minor correction in the temperature range below 1 k , the focus of this study .
in fig . [ fig : resistivity ] we plot temperature - dependent resistivity of ceirin@xmath2 measured in magnetic fields of 0 , 3.15 , 3.6 , 4 , 8 , 11 t applied along the tetragonal @xmath3-axis ( top left panel ) . in zero field , the filamentary superconductivity is observed below @xmath55 1.2 k , the resistive transition is completely suppressed in field of 4 t. the filamentary superconductivity was shown to have the same angular dependence of @xmath4 as the bulk one @xcite and suggested to originate from strain in the samples , likely induced by inclusions of different phase @xcite .
similar phenomenon is known in sr@xmath29ruo@xmath56 in which superconductivity with higher @xmath20 is induced by inclusions of ru - metal @xcite and in some other materials including chain nb@xmath29se@xmath57 @xcite .
the @xmath58 dependences at temperatures above 1 k are very similar for all magnetic fields .
they clearly deviate from the functional form expected in fermi - liquid theory , @xmath59 .
the residual electrical resistivity , @xmath60 , of the samples was determined from measurements down to 40 mk in the magnetic field of 4 t , which was sufficient to completely suppress filamentary superconductivity .
it was about @xmath61 cm , showing high purity of the samples .
this value is notably lower than @xmath62 cm in previous studies of nair et al . @xcite and @xmath63 cm in magnetic field of 12 t of capan et al .
top right panel shows the data in zero field and at 4 t plotted vs. @xmath64 with @xmath65=1.21 , best fit to power - law function for @xmath66 .
this value is very close to previously reported @xmath65=1.27 @xcite .
similar power - law dependence with @xmath65=1.21 is found for fields up to 11 t , with the range of power - law dependence continuing down to @xmath67 1 k. at lower temperatures , all @xmath58 curves show clear signs of saturation when @xmath27 tends to zero , and at the lowest temperatures follow expectations of the fermi - liquid behavior .
this is illustrated in the bottom left panel of fig .
[ fig : resistivity ] in which resistivity data are plotted vs. @xmath40 , linearizing the dependence with the slope proportional to @xmath17 coefficient .
slight upward deviations at the lowest temperatures for 8 t and 11 t curves are similar to those observed in cecoin@xmath2 @xcite and are due to orbital magnetoresistance . for all magnetic fields
@xmath68 , the @xmath69 plots reveal a range of linear dependence at the lowest temperatures .
deviations from @xmath40 dependence at high temperatures mark a characteristic temperature @xmath37 , denoted by arrows .
note that @xmath37 has values in the range 0.7 to 1 k , with resistivity crossing - over to @xmath64 at slightly higher temperatures .
@xmath37 , as determined from resistivity measurements , increases slightly with magnetic field , as shown in the phase diagram ( bottom right panel of fig .
[ fig : resistivity ] ) . due to a crossover character of deviations , definitions of @xmath37
have somewhat high errors , so it is difficult to make reliable extrapolation of @xmath70 towards bulk @xmath71 0.49 t , determined from both thermal conductivity measurements @xcite and heat capacity measurements @xcite . the line in the bottom right panel of fig .
[ fig : resistivity ] suggests extrapolation of @xmath37 to finite values in @xmath72 limit , suggesting non - existence of the field - tuned quantum critical point at @xmath4 or even in zero field . however , significant error bars of @xmath37 determination make this extrapolation from field 4 t unreliable towards bulk @xmath4 . to get additional insight into the magnetic field - induced evolution of the normal state in ceirin@xmath2
, we performed thermal conductivity measurements .
these measurements are not affected by filamentary superconductivity , and they can be performed in the very proximity of bulk @xmath4 . in fig .
[ fig : heat ] we show the thermal conductivity of ceirin@xmath2 measured in three magnetic fields : 0.52 t ( slightly above the bulk @xmath73 t ) , 4 t ( above @xmath74 ) , and 10.7 t ( far away from superconductivity ) .
as can be seen from 0.52 t curve , the @xmath75 vs @xmath27 plot does not show any features on entering filamentary zero resistance state below approximately 1.0 k ( not shown ) , leading to artificial violation of wiedemann - franz law . as long as filamentary superconductivity is suppressed ,
thermal conductivity and electrical resistivity perfectly obey wiedemann - franz law in the @xmath5 0 limit , with @xmath76 , where electrical conductivity @xmath77 and @xmath78)=2.45@xmath810@xmath9w@xmath47k@xmath79 is sommerfeld value of the lorenz number , expected for fermionic systems .
the validity of this ratio in @xmath5 0 limit can be seen by direct comparison of the temperature - dependent thermal analog of electrical resistivity , @xmath80 , and resistivity ( right bottom panel in fig .
[ fig : heat ] ) , which meet at @xmath81 0 .
ceirin@xmath2 is a good metal , and therefore phonon contribution to thermal conductivity can be completely neglected in the temperature interval of interest , @xmath821 k. in fig .
[ fig : heat ] we show this explicitly by subtracting phonon contribution @xmath48 as determined from measurements in disordered samples , from measured @xmath1 of pure ceirin@xmath2 .
the resultant electronic part of thermal resistivity @xmath83 is plotted in bottom right panel of fig .
[ fig : heat ] with solid line .
close match is observed between @xmath84 and @xmath83 below 1 k , validating our conclusion about negligible phonon correction .
as can be seen from top panels of fig . [
fig : resistivity ] , @xmath58 curves in ceirin@xmath2 strongly deviate from expectations of the fermi - liquid functional dependence above @xmath85 1 k and are well described by a @xmath86 dependence with @xmath65=1.21 .
observation of very strong @xmath58 dependence at so low temperatures ( where phonon scattering becomes negligible but the strength of inelastic scattering , @xmath87 @xmath88 cm , is significantly larger than @xmath89 @xmath90 cm ) suggests strong magnetic scattering in the compound , similar to cecoin@xmath2 @xcite .
the power - law non - fermi - liquid temperature dependent resistivity with similar exponents @xmath65 are not unusual in the heavy fermion systems close to quantum critical points .
for example similar values of the power - law exponents are found in cepd@xmath29si@xmath29 ( @xmath91 ) @xcite .
the low exponent @xmath65=1.21 finds explanation within spin fluctuation theory @xcite for nearly antiferromagnetic metals in the framework of a spin density wave ( sdw ) scenario , though significantly lower @xmath92 are usually expected in this case @xcite . in three - dimensional sdw scenario the exponents
are usually expected to be close to @xmath65=1.5 as found in 3d cein@xmath57 ( @xmath93 ) @xcite . for the case of two - dimensional magnetic fluctuations the exponents are usually expected to be close to @xmath65=1 @xcite , as found in ybrh@xmath29si@xmath29 @xcite and cecu@xmath94ag@xmath95 ( @xmath96 ) @xcite . in cecoin@xmath2
the @xmath58 in zero field is close to @xmath27-linear above @xmath20 @xcite , but turns to @xmath65=1.5 at field - tuned qcp @xcite . in this compound exponent
@xmath65 reveals significant anisotropy with @xmath65=1 in @xmath97 @xcite . in ceirin@xmath2 the same power law behavior of resistivity up to 3.5 k
is found for all magnetic fields .
this observation is suggestive , that despite magnetic character of scattering , dominant scattering is not tuned by magnetic field .
this is in stark contrast to cecoin@xmath2 which shows different power law resistivity for low and high magnetic fields in the normal state @xcite .
temperature - dependent part of electronic thermal resistivity obeys the same functional dependence as electrical resistivity , however , because of the different fermi surface integrals of heat and charge scatterings , the @xmath40 coefficients @xmath17 for resistivity and @xmath16 for thermal resistivity are not the same .
typically , for fermi surfaces without sharp corners , the two coefficient are different by a factor of about two @xcite . in the left panel of fig .
[ fig : avsb ] we plot magnetic field dependence of @xmath17 and @xmath16 coefficients of the @xmath40 terms in the in - plane electrical ( left axis ) and thermal ( right axis ) resistivity of ceirin@xmath2 .
the @xmath17 and @xmath16 coefficients in ceirin@xmath2 are indeed close to obeying @xmath98 relation , similar to cecoin@xmath2 and most other cases @xcite .
the @xmath17 and @xmath16 coefficients in ceirin@xmath2 closely follow each other as functions of magnetic fields , with @xmath16 decreasing from approximately 1.2 to 0.8 on field increase from 0.52 t to 11 t. for reference in the right panel we show similar data for cecoin@xmath2 @xcite .
note ten - fold difference in the scales for the two compounds .
in sharp contrast to cecoin@xmath2 , @xmath17 and @xmath16 coefficients in ceirin@xmath2 do not reveal any tendency for divergence as functions of magnetic fields .
the nearly ten - fold difference in @xmath17 and @xmath16 coefficients between cecoin@xmath2 and ceirin@xmath2 suggest notable difference in the effective masses , reflecting close proximity to quantum critical point in case of cecoin@xmath2 and relatively large distance from it for ceirin@xmath2 .
this very weak field dependence of @xmath17 and @xmath16 coefficients in ceirin@xmath2 is consistent with specific heat @xcite , finding weak field dependence of the electronic specific heat coefficient @xmath99 , and the dhva experiments @xcite and the cyclotron effective mass , @xmath100 , ( for field between 6 to 17 t ) at low temperature @xcite . in cecoin@xmath2 @xmath17 coefficient changes from 7.5 to 0.5 @xmath101 cm k@xmath79 for fields @xmath102 to 16 t ) @xcite , and significant field dependence is found for effective masses of carriers , particularly strong for @xmath103 and @xmath104 sheets of the fermi surface @xcite . in our study of field - tuned quantum criticality in cecoin@xmath2 on approaching from both inside the superconducting dome from below @xmath4 and from outside the superconducting dome from above @xmath4
@xcite we noticed that the most heavy electrons are strongly bound in the superconducting condensate . in this case the big difference in the strength of electron - electron renormalization between cecoin@xmath2 and ceirin@xmath2 , can be directly linked to five times difference in the value of the superconducting @xmath20 .
cecoin@xmath2 and ceirin@xmath2 have identical fermi surfaces @xcite , yet interestingly demonstrate very different electron renormalization and @xmath20 .
another interesting observation is that the transition at @xmath4 in ceirin@xmath2 is second order @xcite , while in cecoin@xmath2 it is first order .
first order character of the transition is taken as arising from paramagnetic limiting of the upper critical field .
we can speculate that orbital limiting of @xmath4 observed in ceirin@xmath2 in configuration @xmath34 can be responsible for the lack of coincident quantum critical point .
summarizing , we neither find field - tuned quantum critical point at bulk @xmath4 in ceirin@xmath2 , nor see indications of its existence in all range of magnetic fields studied ( up to 11 t ) .
this is in sharp contrast to cecoin@xmath2 , suggesting that ceirin@xmath2 is significantly further away from qcp .
this finds support in a ten - fold difference of the fermi liquid @xmath40 coefficients between the two compounds .
it is of interest to note that very big difference in @xmath17 and @xmath16 coefficients between cecoin@xmath2 and ceirin@xmath2 is not expected based on previous specific heat studies @xcite . in zero field
both materials have similar @xmath105 values above @xmath20 , somewhat higher in ceinin@xmath2 . big difference in @xmath17 coefficients
suggests that the kadowaki - woods ratio , defined as @xmath106 , is anomalously low in ceirin@xmath2 .
indeed , as shown in fig .
[ fig : avsb ] near @xmath4 in ceirin@xmath2 we find @xmath107 ( @xmath108=10@xmath101 cm mol@xmath109 k@xmath109/ j@xmath109 ) , and the ratio does not show significant field dependence . in cecoin@xmath2
, for instance , this ratio is about 0.52 @xmath108 at @xmath110 that falls on the universal line for some other heavy fermions @xcite .
the kadowaki - woods ratio cancels out mass renormalization in electrical resistivity and heat capacity measurements and leads to scaling of @xmath17 and @xmath111 proportional to level degeneracy of @xmath19-electrons @xcite .
ce ions in cecoin@xmath2 and ceirin@xmath2 are in the same crystal field environment , so it is natural to take the same value @xmath65=2 . while the @xmath112 for cecoin@xmath2 has a value typical for heavy fermion materials , it is anomalously low in ceirin@xmath2 .
anomalously low value of kadowaki - woods ratio in ceirin@xmath2 suggests that not the same groups of carriers are determining the electrical resistivity and the heat capacity of the samples .
two scenarios can be at play . in the first one , conduction electrons are strongly disconnected from @xmath19-electrons due to localized character of the latter . in this case
similar values of heat capacity in cecoin@xmath2 and ceirin@xmath2 reflect nearly identical @xmath19-electron contribution , but their significantly more localized character in ceirin@xmath2 . in the second scenario ,
we need to assume significant difference of properties of different groups of carriers on the multi - band fermi surface of the materials
. this scenario would require the presence of light and mobile carriers dominating in the electrical transport and heavy carriers dominating the heat capacity .
one possible way to distinguish these two scenarios is to study magnetoresistance . in the latter scenario
the magnetoresistance of ceirin@xmath2 should be significantly larger than in cecoin@xmath2 .
comparison of the resistivity data for ceirin@xmath2 in magnetic field of 11 t in fig .
[ fig : resistivity ] bottom left panel with similar data in cecoin@xmath2 @xcite we find that orbital magnetoresistance at the lowest temperatures is stronger in cecoin@xmath2 which is at odds with this scenario .
in conclusion , temperature - dependent electrical resistivity and thermal conductivity measurements in ceirin@xmath2 as functions of magnetic fields applied along @xmath3-axis revealed the validity of the wiedemann - franz law in @xmath113 limit as long as resistivity measurements are not affected by filamentary superconductivity . at temperatures above 1
k the temperature - dependent resistivity is well represented by a power - law function @xmath86 , with nearly field - independent @xmath65=1.21 . at the lowest temperatures the temperature dependences of both electrical resistivity , @xmath58 and of thermal resistivity , @xmath114 , follow expectations of the landau fermi - liquid theory with @xmath15 , as expected for fermi surfaces without sharp features in the angular fermi velocity distributions
. the coefficient @xmath16 does not show significant field dependence even on approaching @xmath4=0.4 t of bulk superconducting state .
weak response to magnetic field is in stark contrast with the behavior found in the closely related cecoin@xmath2 , in which field - tuned quantum critical point coincides with @xmath4 .
the value of the electron - electron mass enhancement , as judged by @xmath17 and @xmath16 coefficients , is about one order of magnitude reduced in ceirin@xmath2 as compared to cecoin@xmath2 , suggesting that the material is significantly further away from magnetic quantum critical point .
suppressed kadowaki - woods ratio in ceirin@xmath2 compared to cecoin@xmath2 suggests notably more localized nature of @xmath19-electrons in the compound .
we thank j. corbin for his assistance with the experiment and r. flint for useful discussions .
l.t . acknowledges support from the canadian institute for advanced research and funding from nserc , frqnt , cfi , and the canada research chairs program .
part of the work was carried out at the brookhaven national laboratory , which is operated for the u.s .
department of energy by brookhaven science associates ( no .
deac02 - 98ch10886 ) .
would like to acknowledges iran national science foundation ( insf ) for supporting this project . c. petrovic , r. movshovich , m. jaime , p. g. pagliuso , m. f. hundley , j. l. sarrao , z. fisk , and j. d. thompson , europhys .
lett . * 53 * , 354 ( 2001 ) . c. petrovic , p. pagliuso , m. hundley , r. movshovich , j. sarrao , j. d. thompson , z. fisk , and p. j. monthoux , phys .
matter * 13 * , l337 ( 2001 ) .
l. pham , t. park , s. maquilon , j. thompson , and z. fisk , phys .
lett . * 97 * , 056404 ( 2006 ) .
n. d. mathur , f. m. grosche , s. r. julian , i. r. walker , d. m. freye , r. k. w. haselwimmer and g.g .
lonzarich , nature * 394 * , 39 ( 1998 ) .
t. shang , r. e. baumbach , k. gofryk , f. ronning , z. f. weng , j. l. zhang , x. lu , e. d. bauer , j. d. thompson , h. q. yuan , phys .
b * 89 * , 041101(r ) ( 2014 ) .
p. pagliuso , c. petrovic , r. movshovich , d. hall , m. hundley , j. sarrao , j. thompson , and z. fisk , phys .
b * 64 * , 100503 ( 2001 ) .
k. an , t. sakakibara , r. settai , y. onuki , m. hiragi , m. ichioka , k. machida , phys . rev . lett . * 104 * , 037002 ( 2010 ) ; h. aoki , t. sakakibara , h. shishido , r. settai , y. onuki , p. miranovic , k. machida , j. phys .
condensed matter * 16 * , l13 ( 2004 ) . c. capan , l. balicas , t. p. murphy , e. c. palm , r. movshovich , d. hall , s. w. tozer , m. f. hundley , e. d. bauer , j. d. thompson , j. l. sarrao , j. f. ditusa , r. g. goodrich , and z. fisk , phys
b * 80 * , 094518 ( 2009 ) .
p. coleman , heavy fermions : electrons at the edge of magnetism , handbook of magnetism and advanced magnetic materials . edited by helmut kronmuller and stuart parkin .
vol 1 : fundamentals and theory .
( john willey and sons , 2007 ) , pp.95 - 148 . | the in - plane electrical resistivity , @xmath0 , and thermal conductivity , @xmath1 , of heavy fermion superconductor ceirin@xmath2 were measured down to 40 mk in the magnetic fields up to 11 t applied along the @xmath3-axis . for all fields above @xmath4=4 t of filamentary superconductivity , we find that the ratio of heat and charge conductivities in @xmath5 0 limit obeys the wiedemann - franz law , @xmath6 , where @xmath7=2.45@xmath810@xmath9 @xmath10 is sommerfeld value of lorenz number
. the temperature - dependent parts of both the electrical resistivity and of the thermal resistivity , @xmath11 , follow functional dependence expected for fermi liquid theory of metals with @xmath12 , @xmath13 , with @xmath14 and @xmath15
. the coefficient @xmath16 does not show significant field dependence even on approaching @xmath4=0.4 t of bulk superconducting state .
weak response to magnetic field is in stark contrast with the behavior found in a closely related cecoin@xmath2 , in which field - tuned quantum critical point coincides with @xmath4 .
the value of the electron - electron mass enhancement , as judged by @xmath17 and @xmath16 coefficients , is about one order of magnitude reduced in ceirin@xmath2 as compared to cecoin@xmath2 ( in spite of two times larger zero - field @xmath18 in ceirin@xmath2 than @xmath18 in cecoin@xmath2 ) , suggesting that the material is significantly further away from magnetic quantum critical point at bulk @xmath4 and at all fields studied .
suppressed kadowaki - woods ratio in ceirin@xmath2 compared to cecoin@xmath2 suggests notably more localized nature of @xmath19-electrons in the compound . |
interactions of very high energy ( vhe ) gamma - rays with the diffuse extragalactic background light ( ebl ) , as well as the the possible effects caused by propagation of secondary ( pair - produced ) electrons in the intergalactic medium ( see e.g. ref . ) , make the high energy gamma - rays unique carries of cosmological information about the epochs of formation of galaxies and their evolution in the past ( see e.g. ref . ) .
the method is based on the distinct features in the spectra of high energy gamma - rays arriving from distant extragalactic objects caused by interactions with ebl .
a serious obstacle in practical realization of this method is our poor knowledge concerning the primary ( intrinsic ) gamma - ray spectra produced in the source @xcite .
it is believed that the well coordinated observations of gamma - ray blazars in different energy bands should allow gamma - ray astronomers to identify the radiation mechanisms , derive the principal model parameters , and thus reconstruct the intrinsic gamma - ray spectra based on spectral and temporal properties of blazars ( see e.g. ref . ) .
even so , for reliable estimates of the intergalactic absorption effect , and ultimately for derivation of the flux and spectrum of ebl , one should take into account possible integral absorption of gamma - rays .
the importance of the internal photon - photon absorption in blazars has been discussed by many authors , in particular just after the discovery by cos b of the first gamma - ray blazar - 3c273 @xcite .
the complex behavior of internal photon - photon absorption may result in a rather `` irregular '' deformation of the spectrum of primary gamma - rays ( from _ steepening _ or even sharp cutoffs to _ hardening_)@xcite .
this , as well as the effects related to the redshift of blazrs@xcite , challenge the feasibility of reliable derivation of ebl based on absorption features in the gamma - ray spectra of extragalactic objects .
moreover , strictly speaking , the intergalactic absorption features contain information about the product of the diffuse extragalactic background radiation density @xmath0 and the hubble constant @xmath1 .
fortunately , it is possible to decouple @xmath0 and @xmath1 by studying the spectral and angular characteristics of vhe @xmath2 radiation from hypothetical electron - positron pair halos surrounding powerful nonthermal extragalactic objects .
these giant structures are unavoidably formed around any extragalactic vhe source due to development of pair cascades initiated by interactions of primary multi - tev photons with ebl@xcite .
for formation of electron - positron pair halos around extragalactic objects , the photon - photon pair production ( pp ) and inverse compton scattering ( ic ) are two principal processes which initiate and support the electromagnetic cascade development@xcite .
the mean free pathlengths and the energy spectra of secondary particles produced at these interactions are described below . for the pair production at a collision of a gamma - ray photon of energy @xmath3 with a target photon of energy @xmath4 the following threshold conditions
is required : @xmath5 where @xmath6 is a velocity of the electron ( positron ) @xmath7 in the cm frame and @xmath8 is an angle between the momenta of two photons in lab frame ( below we will use the energy of @xmath7 pairs in units of the electron rest mass , @xmath9 ) .
the mean free path length @xmath10 of a gamma - ray of energy @xmath3 traveling through the isotropically - distributed photon gas of differential energy density @xmath11 is @xmath12 where @xmath13 is the cross - section averaged over directions of interacting photons : @xmath14 here @xmath15 , @xmath16 and @xmath17 is the total cross section of the process ( see e.g. ref . )
@xmath18 , \end{aligned}\ ] ] where @xmath19 is the thomson cross section . for integration of eq.([eq
: pp - ang - ave - cross ] ) we use the following approximation@xcite , @xmath20 , \end{aligned}\ ] ] which provides an accuracy better than 3% as demonstrated in fig.[fig : pp - compare ] .
the resulting @xmath7 energy distribution is described as @xcite @xmath21 , \end{aligned}\ ] ] where @xmath22 is the energy of the electron ( positron ) ( in @xmath23 units ) and @xmath24 is the photon energy in the lab - frame .
the range of @xmath22 is @xmath25 figure [ fig : pp - diff - cross ] shows the distributions of electrons ( positrons ) for different @xmath26 .
the mean free path of an electron due to the inverse compton scattering , @xmath27 , for the electron propagating through the isotropically distributed photon gas density @xmath28 is , @xmath29 where the averaged over the angles cross section is @xmath30 the total ic cross - section in the lab frame ( see e.g. ref . ) can be presented in the form @xmath31 , \end{aligned}\ ] ] where @xmath32 .
the exact solution of eq.([eq : ic - int - cross ] ) can be presented in an analytical form @xcite , but the resulting expression contains a dilogarithm .
a quite good approximation , with an accuracy better than 10% has been suggested in ref . .
the spectral distribution of the scattered photons via ic for the isotropic photon gas is @xmath33 , \end{aligned}\ ] ] @xmath34 the spectra of upscattered photons depend only on the parameter @xmath35 , as illustrated in fig .
[ fig : ic_spectrum ] . if the value of @xmath35 is small , as in the thomson regime , the upscattered photon population is predominantly in the low energy regime .
if the value of @xmath35 is large , corresponding to the klein - nishina regime , the population is shifted towards high energies .
in order to study of formation of pair halos and to calculate the characteristics of their radiation , in this work we used the method of monte carlo simulations of interactions high energy electrons and photons with intergalactic radiation fields .
this method provides an effective tool for study of development of electromagnetic cascades , and , in particular , allows precise calculations of both energy and spatial distributions of cascade particles for an arbitrary background photon field .
the method of calculations presented below is described in ref . .
for the treatment of interaction lengths and energy distributions of electrons and photons produced at the photon - photon and electron - photon interactions we used the corresponding total and differential cross - sections of these processes .
the hit - or - miss method@xcite and the rejection method@xcite have been used for simulations of the interaction lengths and the energy distributions of cascade particles .
we assumed that the magnetic field in the intergalactic medium is sufficiently small , so the synchrotron cooling of electrons can be neglected . on the other hand , we assumed that the magnetic field is sufficiently large , thus the secondary electrons of all relevant energies are isotropised before they interact with ambient photons .
these two conditions are safely satisfied for a very broad range of the intergalactic magnetic fields between @xmath36 g to @xmath37 g. the background photon fields used in our calculation consist of cosmic microwave background ( cmb ) and the extragalactic background light ( ebl ) .
both the cmb and ebl are functions of redshift . since the background photon field at large cosmological redshifts significantly differ from that at the present epoch , the assumption of a consistent background during the development of the cascade can not be held for the large redshift sources . on the other hand ,
the cascade simulation with evolving background photon field are computationally expensive . to make calculations faster , the simulation scheme was adjusted by simulating the cascade only at a certain epoch or in the limited region around a source of redshift @xmath38 , when one can assume constant background radiation with the ebl flux corresponding the cosmological epoch of @xmath39 .
obviously , the simulation region should be big enough to guarantee an adequate accuracy of calculations of the spatial distributions of particles around the central source . at the same time
the region should be relatively compact so the constant background approximation can be applied .
for example , for a source at a distance of 1 gpc , the gamma - ray production region around the source with an angular size @xmath40 corresponds to the linear size @xmath41 mpc , so we can safely assume that the ebl is not noticeably changed across this size or over the corresponding time of @xmath42 yr . beyond the `` cascade region '' ,
we take into account the interactions with ebl only in terms of absorption ( i.e. without of the cascade treatment ) .
thus the results presented in this paper are relevant only for the core of halos within angular size of a few degree .
although pair halos do extend beyond @xmath40 , their cores , in fact , present the most interest , at least from the point view of detectability of these huge structures . therefore in most of cases , when we are interested in the compact cores of halos ,
the calculations are done using this approximate procedure . within this approach , to calculate the observed energy distribution @xmath43
, we modify the distribution @xmath44 calculated in the `` cascade region '' at the epoch @xmath38 , by multiplying to the attenuation factor , @xmath45 $ ] : @xmath46,\ ] ] where @xmath47 is the pp optical depth , @xmath48 here @xmath49 and @xmath50 are the @xmath2-ray and background photon energies at the redshift @xmath51 , @xmath1 is the hubble constant ; in this paper we use @xmath1= 65 km / s / mpc .
the attenuation factors at different redshifts are shown in fig .
[ fig : attenuation factor ] .
below we present results of calculations of the spectral energy distributions ( seds ) and angular distributions of gamma - rays from pair halos around sources located at different redshifts .
the main uncertainty of predicted energy spectra and morphology of pair halos is related to the uncertainties of ebl , especially in the mid and far infrared wavelengths .
here we use three different ebl models proposed by malkan & stecker ( 1998 , hereafter m98 ) , primack et al .
( 2000 , hereafter p00)@xcite and kneiske et al .
( 2002 , hereafter k02)@xcite .
one should note the recent developments related to the ebl at near infrared and optical wavelengths ( see e.g. ref . ) .
however this wavelength band is not critical for formation of pair halos . on the other hand
the above models represent quite broadly the possible range of ebl at most principal ( from mid- to far- infrared ) bands responsible for formation of pair halos around sources with redshifts up to @xmath52 . in fig .
[ fig : preliminary result2a ] we present the spectral energy distribution ( sed ) of a pair halo within different angles formed around a source with redshift @xmath53 .
this corresponds to the redshift of the blazar 1es 1426 + 482 , but is also representative for many tev blazars reported by the hess , magic and veritas collaborations .
for the ebl we adopted the p00 model .
the calculations are performed for a monoenergetic spectrum of primary gamma - rays with @xmath54 and luminosity @xmath55 .
the differential angular distribution of gamma - ray flux of the pair halo in different energy bands is shown in fig .
[ fig : preliminary result3a ] ( left panel ) . the angular distribution of the cumulative flux , @xmath56 is shown in fig .
[ fig : preliminary result3a ] ( right panel ) .
the initial parameters are the same as in fig .
[ fig : preliminary result2a ] .
as expected at higher energies = the halo looks more compact than at lower energies .
this is clearly seen also in fig .
[ fig : h1426halo ] , where the two - dimensional projections of pair halos at different energies are shown . in order to study the impact of the initial gamma - ray spectrum on the basic characteristics of pair halos
, we calculated the angular and spectral distributions of gamma - rays for a halo initiated by monoenergetic gamma - rays .
the results are shown in fig .
[ fig : compare monoenergetic1 ] for three different energies of primary gamma - rays : e@xmath57 = 20 , 100 , and 500 tev .
one can see that at the intermediate ( around the maximum of sed ) and low energies both the energy and angular distributions are quite insensitive to the primary energy .
the latter becomes important at higher energies , namely between 1 tev and 10 tev .
in particular , with an increase of @xmath58 the angular distribution becomes sharper , especially below @xmath59 .
also , with an increase of @xmath58 , the energy spectrum beyond the maximum around a 100 - 300 gev becomes significantly flatter .
this tendency continues , however , until @xmath60 tev . at higher energies of primary gamma - rays ,
the results of the spectral and angular distributions of the halo radiation becomes almost insensitive to @xmath58 . in fig .
[ fig : halo gamma ] we show the energy and angular distributions of of radiation of the pair halo for more realistic , power - law distribution of primary gamma - rays , @xmath61 between 100 gev and 100 tev , and @xmath62 outside this interval .
the results calculated for three different values of @xmath63=1 , 1.5 , and 2 , but for the fixed total gamma - ray luminosity @xmath64 are shown in fig .
[ fig : halo emax ] .
the result are quite similar which is explained by the fact that for the chosen hard energy spectra of primary gamma - ray , the main contributions come from the energy interval close to @xmath65 tev . in fig .
[ fig : halo emax ] we show the case of primary gamma - rays with fixed power - law photon index @xmath66 , and for different maximum energies , @xmath67 = 10 , 30 , 100 tev .
it is interesting that the pair halo seds are more sensitive to @xmath68 between 10 to 30 tev , whereas the angular distributions are more sensitive to @xmath68 between 30 to 100 tev .
when the redshift of the source exceeds 0.1 one has to take into account the evolution of ebl in the past .
the energy density of the cmb increases with @xmath51 proportional to @xmath69 , while the peak of distribution shifts to higher energies by a factor of @xmath70 ( see fig [ fig : background in many redshifts ] ) .
the ebl evolution is different from the cmb evolution since both the stellar - light and the dust component of ebl are produced at much later epochs with @xmath71 . in fig .
[ fig : background in many redshifts ] we show the evolution of ebl , i.e. the ebl flux at different epochs @xmath51 in accordance with the p00 model ( left panel ) .
the corresponding mean free paths are shown in the right panel of fig . [
fig : background in many redshifts ] .
one can see that at the epoch @xmath72 , the mean free paths of gamma - rays and electrons in the intergalactic medium were an order of magnitude shorter than at the present epoch .
this should obviously result in significant shift of the sed towards low energies and more compact halos ( in addition to the effects introduced by the redshift itself ) .
this is demonstrated in fig .
[ fig : compare redshifts halos ] , where the spectral luminosities and angular distributions of radiation of pair halos are shown for different redshifts of central sources .
the total luminosity of primary gamma - rays is assumed @xmath73 . for energy spectrum of primary gamma - rays extended to @xmath74 tev
, the @xmath7 pairs are produced mainly through interactions with the ebl photons .
therefore , the choice of the ebl model is principal for calculations of characteristics of pair halos . to study the impact of the ebl on the pair halos , we calculated the parameters of pair halos for two different ebl models .
the ebl from p00 is based on semi - analytic calculations .
all physical processes , including the formation and evolution of galaxies are included in the code .
the ebl model m98 is based on empirical assumptions .
it adopts an ebl flux at the present epoch ( @xmath75 , and assumes time evolution proportional to @xmath76 .
these two ebl models for @xmath77 , as well as the corresponding mean free paths of gamma - rays are shown in fig . [
fig : diff ebls ] . in fig .
[ fig : ebl pair halos ] we show the seds and angular distributions of a pair halo located at @xmath78 calculated for two ebl models shown in fig .
[ fig : diff ebls ] .
one can see that while the seds are quite similar at low energies , the energy spectrum corresponding to the p00 model is significantly shifted compared to the sed predicted by the m98 ebl model . also , the p00 ebl model predicts sharper angular distribution , especially at @xmath79 .
this is explained by the higher ebl flux at mid and far - infrared wavelengths assumed in the p00 model . for sources with large redshifts , @xmath80 ,
the energy and angular distributions of pair halos depends on the time - evolution of ebl .
in fact , this dependence is stronger than the dependence on the ebl flux at the present epoch . to demonstrate this effect , in fig .
[ fig : ebl pvsk ] we show the spectral and angular distributions of a halo at @xmath72 calculated for two ebl models - p00 and k02 @xcite , which are characaterised by significantly different cosmological evolution of the ebl flux .
these two models predict relatively similar fluxes at present epoch ( or at low redshifts , @xmath81 ) .
however , at large cosmological epochs , @xmath82 , the difference of two models becomes as large as an order of magnitude .
this is seen in fig .
[ fig : ebl pvsk ] , where the ebl fluxes predicted by the p00 and k02 model are shown for six different cosmological epochs .
the corresponding seds and differential angular distributions are shown fig .
[ fig : sld pvsk ] and [ fig : dad pvsk ] .
one can see that while at low redshifts two ebl models give relatively similar spectral and angular distributions of gamma - rays , at @xmath72 the difference in predictions of two ebl models becomes significant .
interactions of high energy gamma - rays with ebl initiate electroimagnetic cascades which lead to formation of electron - positron halos around powerful nonthermal extragalactic sources .
the pair halos contain unique information about the flux of ebl and its time - evolution in the past .
we studied the dependence of radiation of pair halos on several key parameters , in particular on the energy spectrum of the primary gamma - rays , the source redshift , and the ebl flux .
the results presented in this paper can be used for search for these giant structures which can serve as unique cosmological candles .
f. a. aharonian , tev blazars and cosmic infrared background radiation , in _ invited , rapporteur , and highlight papers .
cosmic ray conf .
highlight papers _ , eds r. schickeiser ( 2001 ) , hamburg , p. 250 | in this paper we study the formation of giant electrons - positron pair halos around the powerful high energy extragalactic sources .
we investigate the dependence of radiation of pair halos , in particular the spectral and angular distributions on the energy spectrum of the primary gamma - rays , the redshift of the source , and the flux of the extragalactic background light . |
in the cores of the majority of rich galaxy clusters the radiative cooling time of the hot , x - ray emitting intracluster medium ( icm ) is significantly shorter than the hubble time . in the absence of heat sources ,
the gas will cool and flow towards the bottom of the cluster potential well : hence the name ` cooling - flow ' clusters .
in fact , recent x - ray observations in particular _ xmm - newton _ high - resolution spectroscopy have ruled out the presence of significant cooling flows ( kaastra et al .
2001 ; molendi & pizzolato 2001 ; peterson et al . 2001 ; tamura et al .
2001 ; kaastra et al .
2004 , and references therein ) : the radiating gas in the cores of these clusters is heated somehow .
the most promising heating mechanism is feedback from the active galactic nucleus ( agn ) hosted by the central galaxy of the cluster .
the basic idea of this agn / cooling - flow scenario is that the cooling icm in the cluster core is heated by outbursts from the central agn : an unsteady equilibrium may be achieved , in the sense that the energy input from the agn balances the radiative losses of the icm over the lifetime of the cluster .
different models have been proposed in which the agn heats the thermal plasma with either mechanical ( binney & tabor 1993 ; tabor & binney 1995 ) or radiative feedback ( ciotti & ostriker 1997 , 2001 ; ostriker & ciotti 2005 ) .
although both processes might be effective to some extent and possibly cooperate , here we focus on mechanical heating , whose relevance to many observed systems is apparent ( e.g. blanton et al . 2004 and
omma & binney 2004 and references therein ) . the cooling luminosity @xmath2 can be determined directly from x - ray observations ( e.g. peres et al .
by contrast , the mechanical power @xmath10 of the radio sources is not only much harder to determine observationally ( e.g. bicknell 1995 ; eilek 2004 ) , but it is expected to be time - variable ( e.g. ulrich , maraschi & urry 1997 ) . if @xmath10 varies with time , how can it be compared with the cooling luminosity in observed systems ? this paper addresses this question .
the mechanical luminosity of an agn is expected to fluctuate on time - scales that range from months to several gigayears , but from the point of view of the symbiosis between the agn and its cooling flow , only time - scales in excess of @xmath11 are of interest .
it is clearly impossible to probe such time - scales by monitoring individual agn .
instead we must draw inferences by studying the demography of agn and cooling flows .
such studies are still in their infancy , but it is nevertheless worthwhile to develop a methodology that they can employ , and to see what conclusions this methodology and currently available data lead us to . cooling flows vary in size from those associated with large clusters of galaxies , such as abell 1795 , to those associated with individual elliptical galaxies , such as ngc 4472 .
we start by scaling out the variations in the mechanical luminosities @xmath10 of agn that simply reflect variations in size rather than temporal variability , by focusing on the dimensionless variable @xmath12 where @xmath2 is the x - ray luminosity of the cooling flow , which is expected to vary at most slowly and weakly with time .
the logarithm of @xmath13 , @xmath14 is a continuous random variable defined on the interval @xmath15 . throughout physics one
models such a variable through the fourier representation @xmath16 the complex numbers @xmath17 are random variables .
the simplest conjecture is that these variables are statistically independent of one another , and have phases that are uniformly distributed in @xmath18 . in this case
@xmath4 is a gaussian random process , and it is completely characterized by the power spectrum @xmath19 . the probability density of the luminosity at any given time @xmath20 is given by @xmath21 } \d \ell,\ ] ] where @xmath22 is the variance of @xmath23 .
it is easy to show that with this model the expectation value of the @xmath24th power of @xmath25 is @xmath26 the appearance of @xmath27 in the exponential signals the skewness of the probability distribution ( [ eqln ] ) .
the larger the value of @xmath24 , the more important this skewness is in the sense that rare excursions to high @xmath25 more strongly dominate @xmath28 .
a measure of the importance of skewness is the probability that an individual measurement of @xmath29 will return a value smaller than @xmath30 .
one can readily show that @xmath31.\ ] ] figure [ plessfig ] plots this function of @xmath32 . when @xmath33 , there is a high probability of finding the luminosity to be smaller than its mean value because much of the total energy output comes during infrequent outbursts .
we are interested in the properties of the light - curve observed with a finite time - resolution , i.e. , averaged over a given time @xmath34 .
we define this ` coarse - grained ' light - curve as @xmath35 where the filter function @xmath36 satisfies @xmath37 and is non zero in a region of width @xmath34 around the origin .
@xmath38 has average @xmath39 and variance and the mass variance @xmath40 usually adopted in cosmology to describe mass density fluctuations on mass scales larger than @xmath41 . ]
@xmath42 where @xmath43 is the fourier transform of the filter @xmath36 . approximating the window function @xmath44 with the heaviside step function @xmath45 , where @xmath46 , we get @xmath47 in words , contributions to the variance on time - scales shorter than @xmath34 vanish in the averaging process .
thus , for any power spectrum @xmath3 , @xmath48 is a non increasing function of @xmath34 , with @xmath49 and @xmath50 : the variance is smaller when averaging over longer @xmath34 .
an agn involves an enormous dynamical range of time - scales , from the few days characteristic of the dynamical time at the event horizon to the hundreds of megayears required for the development of a cooling catastrophe in the core of the cooling flow . in view of this great span of time - scales
, one might expect the power spectrum of the agn to be a power law in frequency .
so we assume that @xmath4 has a power spectrum in the form @xmath51 where @xmath52 is the power at some reference frequency @xmath53 , @xmath54 and @xmath55 . @xmath56 and @xmath57 are , respectively , the maximum and minimum time - scale of variability .
there is no difficulty in principle in deriving equations for a general power - law index @xmath8 , but we shall argue that @xmath8 must lie close to unity , and the discussion is greatly simplified if we specialize to @xmath58 at this point . the case @xmath58 is associated with ` flicker noise ' , which empirically arises in diverse physical phenomena that range from electrical noise , to wind speeds , to earthquake magnitudes ( e.g.press 1978 ; milotti 2002 ) . in case of flicker noise the relation between the variance and the power spectrum is @xmath59 the amplitude of flicker noise
is conveniently characterized by the dimensionless quantity @xmath9 , which is independent of @xmath60 .
the variability of x - ray emission from agn has been studied on time - scales shorter than @xmath61 . when the x - ray power spectra normalized to @xmath62 , while here we refer to the power spectrum of @xmath63 : however , the two have the same slope and amplitude as long as the variance of @xmath64 is small with respect to @xmath62 . ] of seyfert galaxies at frequencies
@xmath65 are fitted with a power - law , one finds @xmath66 ( for instance , @xmath67 for the seyfert galaxy ngc4051 ) , with amplitudes @xmath68 ( uttley , mchardy & papadakis 2002 ; mchardy et al .
hence , the only power - law indices that we can consistently adopt are @xmath7 .
there is a vast difference between the longest time - scales on which the power spectra have been empirically determined , and the time - scales @xmath69 of interest here .
hence to extrapolate blindly from the observed @xmath70 fluctuations out to @xmath71 ones is a very hazardous proceeding .
fortunately , independent estimates of the duty cycle of quasars provides a degree of reassurance .
ciotti & ostriker ( 2001 ) define the duty cycle of a source with light - curve @xmath64 as @xmath72 where the second equality follows from equation ( [ eqavln ] ) .
various arguments indicate that quasars , over the lifetime of their host galaxies have duty cycles @xmath73 , with episodic lifetime ( i.e. , duration of individual major outbursts ) @xmath74 ( haehnelt , natarajan & rees 1998 ; yu & tremaine 2002 ; martini & schneider 2003 ; haiman , ciotti & ostriker 2004 ) . by equation ( [ eqduty ] ) , these duty cycles require variances in the range @xmath75 . inserting these variances and the time - scale range @xmath76 into equation ( [ eqoneoverf ] ) , we find that the associated flicker spectra have amplitudes @xmath77 .
these values are larger by a factor @xmath78 than the amplitudes measured by uttley et al .
( 2002 ) for seyfert galaxies at frequencies higher by @xmath79 .
if @xmath8 differed significantly from unity , we would expect the amplitudes to be many orders of magnitude different , so we must have @xmath7 .
microquasars are radio - jet emitting x - ray binaries made up by a stellar - mass black hole ( bh ) and a companion star , and seem in many respects small - scale versions of agn ( mirabel & rodriguez 1999 ) .
it is widely accepted that on small to intermediate scales these objects are identical to agn scaled down in both length- and time - scales by the ratio of the black holes schwarzschild radii @xmath80 .
hence , we can study the temporal variability of microquasars on human time - scales to get information about the variability of agn on time - scales much longer than those that can be directly probed .
it seems likely that the scaling holds at out to time - scales bh , the characteristic viscous time - scale at @xmath81 is roughly @xmath82 .
longer time - scales correspond to larger radii , at which agn accretion disks might be truncated by self - gravity ( e.g. burderi , king & szuszkiewicz 1998 ; goodman 2003 ) . ]
@xmath83 , which for an agn black - hole mass @xmath84 corresponds to a readily observed frequency @xmath85 in microquasars .
thus from studies of microquasars we can estimate the amplitude of the agn power spectrum at frequencies corresponding to @xmath86 , and then use our knowledge that @xmath8 can not significantly differ from unity to extrapolate it to frequencies lower by a factor @xmath87 .
most studies of the low - frequency power spectrum of microquasars focus on x - ray emission .
reig , papadakis & kylafis ( 2002 ) find that the power spectrum of the microquasar grs 1915 + 105 is well fitted by a power law ( equation [ eqps ] ) with index @xmath7 in the frequency range @xmath88 to @xmath89 .
a power spectrum with this slope , over the same frequency range , has been also found for the x - ray light curve of cygnus x-3 by choudhury et al .
( 2004 ) . however , we are interested in the variability of the kinetic power of the jet , which is expected to be traced by radio emission rather than by x - rays .
the radio variability of microquasars has not been investigated as much as their x - ray variability .
a spectral analysis was carried out by fiedler at al .
( 1987 ) for the radio ( 2.25 and 8.3 ghz ) light curve of the peculiar microquasar ss 433 : they found that its power spectrum is consistent with a power law with index @xmath7 at frequencies @xmath88 to @xmath90 ( see also revnivtsev et al .
nipoti , blundell & binney ( 2005 ) show that at these low frequencies the _ slopes _ of the radio power spectra of the microquasars cygnus x-3 and grs 1915 + 105 are similar to those found in x - rays .
the x - ray light - curve of grs 1915 + 105 has power spectrum with amplitude @xmath91 at frequencies @xmath88 to @xmath85 ( reig et al .
an amplitude smaller by a factor of @xmath92 has been measured for the x - ray power spectrum of cygnus x-3 in the same frequency range ( choudhury et al .
however , there is evidence that the _ amplitude _ of the variability of the radio emission of these powerful microquasars is larger than that of their x - ray emission ( nipoti et al .
the radio emission of cygnus x-3 ( at 2.25 and 8.3 ghz ) has been monitored by the gbi - nasa monitoring program ( waltman et al .
1995 , and references therein ) .
we find that the daily - averaged 2.25-ghz radio light - curve of cygnus x-3 has logarithmic variance @xmath93 over the period @xmath94 .
if we adopt @xmath95 , @xmath96 for the boundaries of the frequency range sampled , equation ( [ eqoneoverf ] ) implies that the power spectrum of the radio emission has amplitude @xmath97 , about two orders of magnitude larger than that found in x - rays .
similarly , we estimated the amplitude of the radio power spectrum of the microquasar grs 1915 + 105 , using its 2.25-ghz light - curve obtained by the gbi monitoring program over about @xmath98 ( see foster et al .
we found that grs 1915 + 105 has amplitude @xmath99 at @xmath100ghz , also much larger than in x - rays .
thus , observations of microquasars suggest @xmath101 as a plausible value of the amplitude of the very low - frequency power spectrum of radio sources .
given the substantial uncertainies involved , this estimate is in remarkably good agreement with the estimate @xmath102 that we obtained by considering the duty cycle of quasars .
these considerations suggest that we can obtain an impression of the long - term variability of agn by rescaling the radio - frequency light curve of a microquasar .
[ pseudofig ] is a plot of the daily - averaged 2.25 ghz light - curve of cygnus x-3 rescaled in time by @xmath103 .
the light curve is characterized by short powerful outbursts , alternating with periods of quiescence .
in general , it is not feasible to measure directly the mechanical power of radio sources .
an exception is m87 the central galaxy of the virgo cluster for which high - resolution observations of the jet and the very inner hot atmosphere are available ( owen , eilek & kassim 2000 ; di matteo et al .
2003 ) . in the case of the central radio sources in galaxy clusters
@xmath10 can be estimated from depressions in the x - ray emission of the surrounding icm .
these ` cavities ' are interpreted as bubbles produced in the interaction between the jets of the radio source and the thermal plasma .
the energy transferred to the icm by a bubble expanding adiabatically is ( e.g. churazov et al .
2000 , 2002 ) @xmath104 where @xmath105 is the volume of the bubble , @xmath106 is the pressure of the surrounding icm , and @xmath107 is the adiabatic index of the gas in the bubble .
thus , @xmath108 for a non - relativistic gas ( @xmath109 ) , and @xmath110 for a relativistic fluid ( @xmath111 ) .
given a characteristic time @xmath112 associated with the bubble , an estimate of the average kinetic luminosity of the agn over @xmath112 is then @xmath113 the definition of @xmath112 is controversial : if we consider @xmath112 as the _ age _ of the bubble , this is of order the time @xmath114 required for the cavity to rise at the speed of sound from the centre to its current location ( gull & northover 1973 ; churazov et al .
alternatively , the age of the bubble can be defined as the time required for the icm to refill the displaced volume ( mcnamara et al . 2000 ) .
however , @xmath112 is not necessarily the age of the bubble : it could be the _ recurrence time _
@xmath115 of the radio source , defined as the time elapsing between two subsequent outbursts in which the jet blows bubbles .
as we are looking for the characteristic time to be associated with this measure of @xmath0 ( i.e. , the time over which we average in measuring from the bubbles ) , it is natural to identify @xmath112 with @xmath115 , so in the rest of the paper we assume @xmath116 .
brzan et al .
( 2004 , hereafter b04 ) present radio and x - ray data for a sample of 18 systems with cooling flows that show evidence of cavities in their diffuse x - ray emission .
the sample consists of 16 galaxy clusters , the galaxy group hcg 62 , and the elliptical galaxy m84 .
the authors provide for each of these objects the cooling luminosity ( @xmath2 ) , the monochromatic radio luminosity at 1.4 ghz ( @xmath117 ) , the integrated radio luminosity between 10 mhz and 5 ghz ( @xmath118 ) , and measures of @xmath119 for each cavity .
this sample is suitable for comparison with the agn / cooling flow model , because for each source different measurements of the mechanical power ( averaged on different time - scales ) are available . from measures of @xmath119
we derive @xmath0 and @xmath120 using equation ( [ eqlb ] ) , given the time - scale @xmath121 .
in general , we do not know the recurrence time for individual sources . in the perseus cluster , in which two generations of bubbles are observed ( fabian et al .
2000 ) , the recurrence time - scale is of the order of @xmath122 . estimating @xmath115 for other sources , in which just one generation of cavities is observed , is problematic , but statistical arguments suggest @xmath123 as a typical value ( b04 ) .
so we assume that @xmath112 is the same for all sources , and we parameterize @xmath0 as @xmath124 where @xmath125 we then use the kolmogorov smirnov ( ks ) test to estimate how well the measured values of @xmath29 fit the predicted log - normal distribution ( [ eqln ] ) with @xmath126 .
[ fbfig ] plots the 90 and 95 per cent confidence levels , derived from the ks test , in the space of the parameters @xmath127 and @xmath27 .
for instance , if we fix @xmath128 , and allow @xmath112 in the range @xmath129 , the factor @xmath127 lies in the interval @xmath130 .
such a range in @xmath127 corresponds to values of the variance @xmath131 that give distributions consistent with the data at the 95 per cent confidence level .
the number of systems in which cavities have been accurately observed is small , while the radio luminosities of thousands of radio galaxies have been measured .
it is clear that the radio power represents a very small fraction of the kinetic power emitted by the agn ( e.g. eilek 2004 ) , but the radio luminosity may be a useful indicator of the mechanical power . hence , much larger samples could be assembled if it were possible to establish a relation between the mechanical and radio luminosities of sources .
cavities in the x - ray emitting icm are classified as radio - filled if non - thermal radio emission at a given frequency coincides with the x - ray - depressed area ; otherwise they are called ` ghost ' cavities radio data might appear as radio - filled at lower frequencies , depending on the break frequency of the radio spectrum .
this has been observed in perseus by fabian et al .
( 2002 ) . ] .
if reacceleration of the synchrotron electrons is negligible , the age of the radio - filled bubbles will be of the order of the synchrotron lifetime , or smaller , while the age of the ghosts will be longer .
thus , from radio - filled cavities we can independently measure @xmath132 and @xmath117 and thus determine the relation that holds between them .
the top panel of fig .
[ konefourfig ] plots data for the nine objects with _ radio - filled _ cavities presented by b04 .
the values of @xmath119 are from _ chandra _ , and @xmath114 is from b04 .
@xmath133 , where @xmath134 is the monochromatic @xmath135ghz power measured by the vla .
the straight line is the relation @xmath136 with the best
fitting dimensionless conversion factor @xmath137 .
the lower panel of fig .
[ konefourfig ] shows the result of repeating this exercise using the bolometric radio luminosity @xmath138 instead of the monochromatic luminosity .
@xmath118 is estimated by integrated between @xmath139mhz and @xmath140ghz , assuming constant spectral index for each source ( b04 ) .
the straight line is the relation @xmath141 with best fitting @xmath142 .
b04 found that a non - linear relation the mechanical luminosity scaling as @xmath143 provides a better fit to their data ( including abell 478 ) than the simple proportionality @xmath144 considered here .
however , the sample is very small , and one can show that , excluding abell 478 and the peculiar radio source cygnus a the only fr ii ( fanaroff & riley 1974 ) source in the sample the slope of the best fitting power - law is close to 1 . we think that a larger sample is needed to verify the existence of a non linear relation between kinetic and radio power .
we next investigate the constraints that can be placed on @xmath145 and @xmath146 by assuming that @xmath147 .
we use the whole sample presented by b04 , _ without any distinction between radio - filled and ghost cavities _ , and assume that the conversion coefficients @xmath145 and @xmath146 are the same for all sources .
we use the ks test to evaluate the consistency between the values of @xmath10 obtained using trial values of these coefficients and the log - normal probability distribution ( [ eqln ] ) with @xmath126 .
[ k14sigfig ] shows contours of 90 and 95 percent confidence levels in the planes of @xmath27 versus @xmath148 ( top ) or @xmath149 ( bottom ) .
the dotted line in each panel shows the value of @xmath150 determined above from fig .
[ konefourfig ] under the assumption that @xmath128 .
the data are consistent for a wide range of parameters .
larger variances @xmath27 are required for smaller coefficients @xmath150 because the current values of @xmath10 are proportional to @xmath150 , and if these are small the assumption @xmath126 requires that significant energy is released during excursions to high luminosity . for the values of @xmath150 that we derived from radio - filled bubbles , @xmath151 and @xmath152 are favoured variances .
these values are intermediate between the values @xmath153 that we inferred from microquasars and @xmath154 that we estimated from quasars .
as discussed in section 2 , the variance @xmath27 associated with any measurement of @xmath10 should decrease with increasing timespan over which that measurement effectively averages @xmath10 . using radio data we estimate the kinetic power averaged over the lifetime of synchrotron - emitting relativistic electrons , which depends on the radio frequency .
a reference time - scale for synchrotron radiation at frequency @xmath155 is the synchrotron cooling time @xmath156 where @xmath157 is the modulus of the magnetic field ( e.g. carilli et al .
@xmath158 depends strongly on @xmath157 , which is usually poorly constrained in observed radio sources , and for fixed strength of the magnetic field decreases by about one order of magnitude from @xmath159 to @xmath160 .
actually , @xmath158 represents an upper limit on the lifetime of the radio - emitting electrons because adiabatic expansion losses are expected to be dominant over radiative losses ( e.g. blundell & rawlings 2000 ) .
as the effects of adiabatic expansion losses are difficult to quantify in general , in our treatment we simply assume that the averaging time - scale is @xmath161 , when @xmath0 is measured from radio luminosity at frequency @xmath155 .
[ obsdtfig ] plots the three ranges of acceptable values of @xmath27 that we have derived above against the characteristic averaging time @xmath162 associated with each measurement . for the mechanical luminosity measured from the bubbles
, we have @xmath163 , and we adopt @xmath164 as the range for the recurrence time ( see section 3.1 ) . for the mechanical luminosity derived from 1.4 ghz radio data , @xmath165 and @xmath166 , where the range in @xmath167 has been computed from equation ( [ eqtsync ] ) for @xmath157 in the range @xmath168 consistent with minimum - energy magnetic fields measured in the radio lobes of 3c 218 ( hydra a ; taylor et .
al 1990 ) , 3c 84 ( perseus ; fabian et al . 2002 ) , and 3c 405 ( cygnus a ; carilli et al .
1991 ) . considering the bolometric radio luminosity
, we get more stringent limits on the variance , @xmath169 , while the corresponding @xmath170 spans the same range as @xmath171 , because in this case the bolometric luminosity is measured from @xmath172 radiation .
we assume that the relevant limits on the variance are those found for @xmath173 , because the spread in the slopes of the radio spectra is likely to contribute to @xmath174 .
the dashed curves in fig . [ obsdtfig ] show the dependence of @xmath27 upon @xmath162 that is expected for flicker noise spectra ( equation [ eqsigapp ] ) with two different amplitudes , @xmath175 and @xmath176 .
the model and the data are consistent for amplitudes that lie within this range , which overlaps the interval between the amplitudes of the radio spectra of microquasars and those inferred from the duty cycle of quasars .
note that for @xmath170 we used the synchrotron cooling time , which is an upper limit .
a shorter @xmath170 would result in lower value of @xmath177 : for instance , reducing @xmath170 by a factor of 10 we get @xmath178 .
in the previous section we have shown that , under plausible assumptions , observed data of central radio galaxies in cooling - flow clusters are consistent with the hypothesis that agn heating balances radiative cooling . here
we discuss the consequences of extending the considered picture to cooling - flow systems in general .
in particular , applying our model to luminous early - type galaxies leads to predictions on the local radio luminosity function .
radio galaxies , though spanning several orders of magnitude in radio power , represent a very homogeneous family of objects when the properties of their host stellar systems are considered , being early - type galaxies with optical luminosity larger than @xmath179 .
for instance , magliocchetti et al .
( 2002 ) derived the local radio luminosity function for agn and early - type galaxies of the first-2df galaxy redshift survey ( 2dfgrs ) , finding that the optical counterparts have b - band luminosity characterized by a very narrow distribution , with average @xmath180 and scatter @xmath181 mag [ @xmath182 , where @xmath183 is the hubble constant ] .
let us assume that all bright elliptical galaxies have a cooling core , a central supermassive bh , and that , over the galaxy lifetime , the radiative losses from their hot plasma are balanced by the mechanical energy emitted in radio outbursts from the central agn ( @xmath184 ) .
the radio luminosity of the agn will be strongly variable , and poorly correlated with the cooling luminosity of the host cooling core .
however , under the assumption that the monochromatic 1.4 ghz radio luminosity traces the kinetic luminosity ( @xmath185 ) , for a sample of galaxies with the same cooling luminosity @xmath2 , we would expect @xmath186 .
if @xmath187 is the distribution of the cooling luminosities of bright early - type galaxies , the radio luminosity function of early - type galaxies will be given by @xmath188 where @xmath189 is a log - normal distribution with @xmath186 and variance @xmath174 .
the luminosity function @xmath187 has yet to be determined observationally , because of the lack of complete x - ray samples of early - type galaxies .
as well known , the diffuse x - ray luminosity of early - type galaxies with similar optical luminosity is characterized by significant scatter ( fabbiano , gioia & trinchieri 1989 ; osullivan , forbes & ponman 2001 ) .
given that the optical luminosity is roughly the same for all sources , it is natural to model the distribution of @xmath2 as @xmath190}\d\ell_{\rm x},\ ] ] where @xmath191 , @xmath192 is the logarithmic variance and @xmath193 is the number density . substituting @xmath194 and @xmath189 in equation ( [ eqrlf ] )
we get @xmath195}\d\ell_{1.4},\ ] ] where @xmath196 , @xmath197 , and @xmath198 .
thus , our model predicts that the radio luminosity function of early - type galaxies is log - normal with variance @xmath199 , where @xmath174 is the variance due to radio - variability of each source and @xmath192 is the variance in the distribution of x - ray emission of bright ellipticals .
we test this prediction by fitting the local radio luminosity function of early - type galaxies of the first-2dfgrs ( magliocchetti et al .
2002 ) with equation ( [ eqrlfln ] ) , leaving @xmath200 , @xmath201 and @xmath193 as free parameters .
we found best fitting parameters ( with 1 @xmath32 uncertainties ) @xmath202 , @xmath203 , and @xmath204 . fig .
7 plots the data and error bars as reported by magliocchetti et al .
( 2002 ) together with our best
fitting log - normal function .
we note that the best fit gives @xmath205 , consistent with values of @xmath174 found analysing b04 s sample of central radio galaxies in cooling cores ( section 3.2 ) plus a contribution from intrinsic scatter in x - ray luminosity . we based our analysis of the radio luminosity function on the assumption that most bright elliptical and s0 galaxies have agn - powered radio emission .
we can check this assumption a posteriori by comparing the best fitting number density @xmath193 with the number density of luminous early - type galaxies .
we assume that the b - band optical luminosity function of early - type galaxies @xmath206 is a schechter function with slope @xmath207 , characteristic luminosity @xmath208 , and normalization @xmath209 .
then the number density of objects brighter than @xmath210 is @xmath211 . if we consider the best fitting b - band luminosity function determined by madgwick et al .
( 2002 ) for the early - type galaxies of the 2dfgrs [ @xmath212 , @xmath213 , and we assume that all early - type galaxies with @xmath214 are radio galaxies , we find a number density @xmath215 .
thus , @xmath216 is a factor of @xmath217 larger than @xmath193 found from our best
fitting log - normal distribution .
the discrepancy between the two number densities is easily explained by our arbitrary choice of @xmath208 as the minimum b - band luminosity of a radio galaxy , and by uncertainties on the faint - end of radio luminosity function .
we conclude that available radio data are consistent with the proposed scenario , in which each bright elliptical oscillates , over its lifetime , along the radio luminosity function , spanning several orders of magnitude in radio power .
interestingly , results of high - redshift radio surveys also suggest that most massive ellipticals have experienced one or more episodes of powerful radio activity during their lifetime ( rawlings 2003 ) .
finally , we stress that the proposed picture naturally explains the very large scatter in radio luminosity for radio galaxies with the same bh mass ( lacy et al .
2001 ; mclure et al .
2004 , and references therein ) as a consequence of the strong time - variability of their radio emission ( see also nipoti et al .
since the cosmic density of matter is very inhomogeneous on small scales , it is not possible to derive its global properties from individual measurements of the density in small regions : the characteristics of large - scale structure must be determined from a model of the fluctuating density field , and the values that one derives from it of statistical quantities , such as the mass variance .
similarly , the temporal variability of agn is too large to ignore . we can not hope to understand the radio luminosity function , or to test the agn / cooling - flow model , without a model of agn variability .
the normalized mechanical luminosities of a sample of agn , @xmath120 , should show scatter only because of temporal variability , and be free of differences in the scales of systems . since the variability of agn is known to be large and @xmath29
is an inherently positive variable , the natural variable to model is @xmath218 , and the default assumption to make is that @xmath4 is a gaussian random process .
this assumption implies that individual measures of @xmath29 should have a log - normal distribution , so the time - sequence of @xmath0 is completely characterized by the power spectrum @xmath3 of @xmath4 .
the enormous range of physical scales involved in agn suggests that the power spectrum will be a power law @xmath219 .
several considerations suggest that @xmath7 , the index characteristic of flicker noise .
first , on the short time - scales that are directly accessible , monitoring of agn yields @xmath7 .
second , a comparison of the amplitudes measured in these studies with that inferred from the short duty cycle of quasars requires @xmath7 .
third , microquasars , whose power spectra can be probed over a camparatively wide range of frequencies , are found to have @xmath7 .
finally , diverse physical phenomena are found to be characterized by flicker - noise power spectra , so it would not be surprising if agn conformed to this pattern .
we estimate the spectral amplitude of the mechanical luminosity of agn at the centres of cooling flows by assuming that energy radiated by the thermal plasma is on the average replaced by the mechanical input of the central agn , and that x - ray cavities probe the mechanical luminosity averaged over time - scales @xmath220 .
the amplitudes required are slightly larger than those that characterize the radio output of two microquasars , and of the order of those estimated from the duty cycle of quasars .
we estimate the coefficients @xmath150 in an assumed proportionality between the mechanical luminosity of an agn and its radio power , either at @xmath135ghz or bolometric .
we do this in two ways . in the first we directly fit a straight line to the correlation between radio and mechanical luminosities of sources that have radio - loud cavities .
alternatively , we fit our model of agn variability to the radio data of all sources with cavities , whether radio - loud or ghost , under the assumption that @xmath184 .
the two methods yield compatible ranges of values of @xmath150 .
the available observations fail to constrain the model s parameters strongly , both because the currently available sample of sources is small and because our understanding of the physics of these sources is poor . despite the significant advance that was made possible by the discovery of numbers of x - ray cavities , measurements of the ratio of radio to mechanical luminosity
are still extremely uncertain .
also , our analysis clearly shows the importance of determining the duty cycle of the radio sources in order to quantify their contribution in the energetics of cooling flows : in particular , it is crucial to estimate the amplitude of their power spectrum on time - scales @xmath221 . we conclude that larger samples of agn / cooling - flow pairs and better theoretical models of radio sources are needed to challenge the agn / cooling - flow model . in a scenario
in which agn radio activity is strongly variable with time , the local luminosity function of radio galaxies should be interpreted as the distribution of the time spent by a galaxy at a given radio luminosity rather than a distribution of sources with different intrinsic properties .
this picture is supported by the observational finding that early - type galaxies hosting supermassive bhs with the same mass span several orders of magnitude in radio power .
we showed that the first-2dfgrs luminosity function of radio galaxies can be fitted by a log - normal distribution , consistent with the hypothesis that in all early - type galaxies , the cooling flow is balanced by a strongly variable radio source .
our model predicts a turnover at luminosities similar to the lowest currently probed .
any turnover in the luminosity function of agn will be to a large extent masked by radio emission from nuclear starbursts .
nonetheless , the prediction of a turnover is interesting and will be tested by forthcoming radio surveys with instruments such as the square kilometer array .
we are grateful to katherine m. blundell and steve rawlings for helpful discussions .
we used data made publicly available by the gbi - nasa monitoring program .
the green bank interferometer ( gbi ) was a facility of the national science foundation operated by the national radio astronomy observatory in support of nasa high energy astrophysics programs .
bicknell g.v . , 1995 , apjs , 101 , 29 binney j. , tabor g. , 1995 , mnras , 276 , 663 brzan l. , rafferty d.a . , mcnamara b.r .
, wise m.w . ,
nulsen p.e.j . , 2004 ,
apj , 607 , 800 ( b04 ) blanton e.l .
, sarazin c.l . ,
mcnamara b.r . ,
clarke t.e . , 2004 ,
apj , 612 , 817 blundell k.m . , rawlings s. , 2000 , aj , 119 , 1111 burderi l. , king a.r . , szuszkiewicz e. , 1998 , apj , 509 , 85 carilli c.l .
, perley r.a . , dreher j.w . ,
leahy j.p . , 1991 , apj , 383 , 554 choudhury m. , rao a.r .
, vadawale s.v . ,
jain a.k . ,
singh n.s . , 2004 , a&a , 420 , 665 churazov e. , forman w. , jones c. , bhringer h. , 2000 , a&a , 356 , 788 churazov e. , brggen m. , kaiser c.r .
, bhringer h. , forman w. , 2001 , apj , 554 , 261 churazov e. , sunyaev r. , forman w. , bhringer h. , 2002 , mnras , 332 , 729 ciotti l. , ostriker j.p . , 1997 , apj , 487 , l105 ciotti l. , ostriker j.p . , 2001 , apj , 551 , 131 di matteo t. , allen s.w . , fabian a.c . ,
wilson a.s .
, young a.j . , 2003 ,
apj , 582 , 133 eilek j.a . , 2004 , in reiprich t. h. , kempner j. c. , soker n. , eds , the riddle of cooling flows in galaxies and clusters of galaxies .
available from http://www.astro.virginia.edu/coolflow/proc.php fabbiano g. , gioia i.m .
, trinchieri g. , 1989 , apj , 347 , 127 fabian a.c .
2000 , mnras , 318 , l65 fabian a.c .
, celotti a. , blundell k.m . ,
kassim n.e .
, perley r.a . , 2002 , mnras , 331 , 369 fanaroff b.l . , riley j.m . , 1974 , mnras , 167 , 31p fiedler r.l .
et al . , 1987 ,
aj , 94 , 1244 foster r.s .
, waltman e.b . , tavani m. , harmon b.a .
, zhang s.n . , paciesas w.s . ,
ghigo f.d . , 1996 , apj , 467 , l81 ; goodman j. , 2003 , mnras , 339 , 937 gull s.f . ,
northover k.j.e . , 1973 , nature , 244 , 80 haehnelt m.g .
, natarajan p. , rees m.j .
, 1998 , mnras , 300 , 817 haiman z. , ciotti l. , ostriker j.p . , 2004 , apj , 606 , 763 kaastra j.s . , ferrigno c. , tamura t. , paerels f.b.s . , peterson j.r .
, mittaz j.p.d . , 2001 , a&a , 365 , l99 kaastra j.s et al . , 2004 , a&a , 413 , 415 lacy m. , laurent - muehleisen s.a . ,
ridgway s.e . ,
becker r.h . ,
white r.l . , 2001 , apj , 551 , l17 madgwick d.s .
et al . , 2002 ,
mnras , 333 , 133 magliocchetti m. et al . , 2002 , mnras , 333 , 100 martini p. , schneider d.p . , 2003 , apj , 597 , l109 mchardy i.m .
, papadakis i.e. , uttley p. , page m.j . , mason k.o . , 2004 , mnras , 348 , 783 mclure r.j
. , willott c.j . ,
jarvis m.j . , rawlings s. , hill g.j .
, mitchell e. , dunlop j.s . ,
wold m. , 2004 , mnras , 351 , 347 mcnamara b.r .
et al . , 2000 , apj , 534 , l135 milotti e. , 2002 , preprint ( physics/0204033 ) mirabel i.f . ,
rodriguez l.f . , 1999 ,
ara&a , 37 , 409 molendi s. , pizzolato f. , 2001 , apj , 560 , 194 nipoti c. , blundell k.m . , binney j. , 2005 , submitted omma h. , binney j. , 2004 , mnras , 350 , l13 ostriker j. p. , ciotti l. , 2005 , philos .
a , 363 , 667 osullivan e. , forbes d.a .
, ponman t.j . , 2001 ,
mnras , 328 , 461 owen f.n . , eilek j.a . , kassim n.e . , 2000 , apj , 543 , 611 peres c.b . , fabian a.c . , edge a.c . , allen s.w . , johnstone r.m . , white d.a . , 1998 , mnras , 298 , 416 peterson j.r . et al . , 2001 , a&a , 365 , l104 press w.h . , 1978 , comments astrophys .
, 7 , 103 rawlings s. , 2003 , new astron .
, 47 , 397 reig p. , papadakis i. , kylafis n.d .
, 2002 , a&a , 383 , 202 revnivtsev m. et al . , 2005 ,
a&a , submitted ( astro - ph/0501516 ) sun m. , jones c. , murray s.s . , allen s.w . , fabian a.c . ,
edge a.c . , 2003 ,
apj , 587 , 619 tabor g. , binney j. , 1993 , mnras , 263 , 323 tamura t. et al . , 2001 , a&a , 365 , l87 taylor g.b . ,
perley r.a .
, inoue m. , kato t. , tabara h. , aizu k. , 1990 , apj , 360 , 41 ulrich m .- h .
, maraschi l. , urry c.m .
, 1997 , ara&a , 35 , 445 uttley p. , mchardy i.m .
, papadakis i.e. , 2002 , mnras , 332 , 231 waltman e.b . ,
ghigo f.d .
, johnston k.j . , foster r.s .
, fiedler r.l . , spencer j.h . , 1995 ,
aj , 110 , 290 yu q. , tremaine s. , 2002 , mnras , 335 , 965 | there is increasing evidence that agn feedback is important in the energetics of cooling flows in galaxies and galaxy clusters .
it is possible that in most cooling - flow clusters radiative losses from the thermal plasma are balanced , over the cluster lifetime , by mechanical heating from the central radio source .
we investigate the implications of the variability of agn mechanical luminosity @xmath0 on observations of cooling flows and radio galaxies in general .
it is natural to assume that @xmath1 is a gaussian process
. then @xmath0 will be log - normally distributed at fixed cooling luminosity @xmath2 , and the variance in a measure of @xmath0 will increase with the time - resolution of the measure .
we test the consistency of these predictions with existing data for cooling flows and radio galaxies .
these tests hinge on the power spectrum @xmath3 of the gaussian process @xmath4 .
general considerations suggest that @xmath5 is a power law @xmath6 .
long - term monitoring of seyfert galaxies combined with estimates of the duty cycle of quasars imply that @xmath7 , which corresponds to flicker noise .
the power spectra of microquasars have similar values of @xmath8 .
we combine a sample of sources in cooling flows that have cavities with the assumption that the average mechanical luminosity of the agn equals the cooling flow s x - ray luminosity .
given that the mechanical luminosities are characterized by flicker noise , we find that their spectral amplitudes @xmath9 lie between the estimated amplitudes of quasars and the measured values for the radio luminosities of microquasars .
the model together with the observation that powerful radio galaxies lie within a narrow range in optical luminosity , predicts the luminosity function of radio galaxies .
both the shape and the normalization of the predicted function are in agreement with observations .
forthcoming radio surveys will test the prediction that the luminosity function turns over at about the smallest luminosities so far probed .
[ firstpage ] galaxies : clusters : general cooling flows galaxies : jets radio continuum : galaxies galaxies : luminosity function , mass function |
the effective hamiltonian with qcd effects + @xmath2 is given by @xmath3 the operators @xmath4 can be grouped into two categories @xcite,@xcite : i=1,2-current - current operators;@xmath5 denotes the relevant ckm factors that are : @xmath6 we should take the variable@xmath7and@xmath8,or x and y as , @xmath9the partial decay rate in the c rest frame is,@xmath10 here @xmath11 @xmath12
@xmath13 @xmath14@xmath15 @xmath16^{1/2},h_{yc}=[1-(y^{2}/(y^{2}+c^{2}))]^{1/2},\]]@xmath17^{1/2 } , h_{ya}=[1-(y^{2}/(y^{2}+a^{2}))]^{1/2},\]]@xmath18@xmath19@xmath20 @xmath21@xmath22@xmath22
in the spectator model @xcite the spectator quark is given a non - zero momentum having in this work a gaussian distribution , represented by a free ( but adjustable ) parameter,@xmath23 @xmath24 the total meson decay rate through a particular mode is then assumed to be @xmath25 equal to the initiating decay rate .
we have @xmath26 here@xmath27@xmath28@xmath29 the integration range is restricted by @xmath30.we call this mode of quark and antiquark combination ( colour favoured ) .
it is also possible that the spectator antiquark combines with the quark@xmath31,for which we get @xmath32@xmath33 we call this * process(c - s ) * ( colour suppressed ) .
summing , the decay rates of b mesons for * process(c - f ) * and * process(c - s ) * are : @xmath34 @xmath35 where @xmath36 and so on .
the differential decay rates for two boson system in the spectator quark model for current - current plus penguin operators in the effective hamiltonian is given by , @xmath37@xmath38 where @xmath39 + the integration region is restricted by the condition @xmath40,thus @xmath41 where @xmath42^{2}-(q_{si}/m_{c})^{2}+\]]@xmath43 @xmath44 where @xmath45^{2}-(q_{si}/m_{c})^{2}+\]]@xmath46 therefore using eq.(6 - 11 ) the phase space parameters will be defined by , @xmath47@xmath48 now , we can integrate over the two mass cuts ( two boson systems ) , and obtain the hadronic decay rates as follows , @xmath49@xmath50@xmath51
now we want to calculate the decay rates of effective hamiltonian @xmath52 for f+s at quark - level and spectator model .
the effective hamiltonian for f+s , is given by @xmath53 where@xmath54is defined by equations mentioned during in the last page , so we can obtain @xmath55.the decay rates of current - current plus penguin for f+s is given by , @xmath56 @xmath57,\ ] ] @xmath58,\ ] ] @xmath59.\ ] ]
we use the standard particle data group @xcite parameterization of the ckm matrix . following ali and greub
@xcite we treat internal quark masses in tree - level loops with the values ( gev)@xmath60and @xmath61.following g.buccella @xcite we choose the effective wilson coefficients @xmath62for the various @xmath63 transitions.we have used in spectator quark model the value@[email protected] the maximum mass of the quark - antiquark systems @xmath65we take a value midway between the lowest mass@xmath66state and the next most massive meson .
the decay rates of c quark for effective hamiltonian and effective hamiltonian of f+s shown in the table.1 . also the decay rates of c quark for f+s is given by @xmath67@xmath68 @xmath69@xmath70 @xmath71@xmath72
we used effective hamiltonian theory and spectator quark model for c quark and calculated hadronic decays of d mesons . in this model we added decays of channel hadronic decays of d mesons .
for colour favoured and suppressed we consider the channel @xmath73and achieved theoretical values very close to experimental ones .
finally it has been shown the case , in which the theoretical values are better than the amplitude of all the decay rates have been calculated . in table 1 ( below ) it must be noted that columns 2 and 4 have to be multiplied by @xmath74and columns 3 and 5 should be multiplied by @xmath75 .
[ cols="^,^,^,^,^",options="header " , ] +
99 a. j. buras , nucl instrum meth . a368 ( 1995 ) | in this research we describe effective hamiltonian theory and apply this theory to the calculation of current - current @xmath0 and qcd penguin @xmath1 decay rates we calculate the decay rates of semileptonic and hadronic of charm quark in effective hamiltonian theory .
we investigate the decay rates of * d meson decays * according to * spectator quark model(sqm ) * for the calculation of * d * meson decays .
we obtain the total decay rates of semileptonic and hadronic of charm quark in effective hamiltonian according to colour favoured ( c - f ) and colour suppressed ( c - s ) , and then to added amplitude of processes colour favoured and colour suppressed ( f - s ) and obtain the decay rates of them . |
although random matrices were originally introduced by john wishart @xcite in 1928 as a tool to study population dynamics in biology through principal component analysis , they became very popular much later in 1951 when wigner @xcite postulated that the fluctuations in positions of the energy levels of heavy nuclei are well described ( in terms of statistical properties ) by the eigenvalues of a very large hermitian random matrix .
random matrix theory ( rmt ) is now an active research area in mathematics and theoretical physics with applications in statistics , biology , financial mathematics , engineering and telecommunications , number theory etc .
( see @xcite for a state of the art ) .
the classical models of hermitian random matrices are the gaussian orthogonal , unitary and symplectic ensembles .
it is well known that the joint law of the eigenvalues of the matrices in those gaussian ensembles is the boltzmann - gibbs equilibrium measure of a one - dimensional repulsive coulomb gas confined in a harmonic well .
more precisely , this joint law has a probability density @xmath5 on @xmath6 ( @xmath7 is the dimension of the square matrices ) given by @xmath8 where the inverse temperature @xmath9 for the gaussian orthogonal ensemble , respectively @xmath10 for the unitary and symplectic ensembles .
the linear statistics of the point processes with joint probability density functions ( jpdf ) @xmath11 have been extensively studied in the literature with different methods @xcite . in 2002 ,
dumitriu and edelman @xcite came up with a new explicit ensemble of random tri - diagonal matrices whose eigenvalues are distributed according to the jpdf @xmath5 for any @xmath12 ( see also @xcite where invariant ensembles associated to general @xmath1 were constructed ) .
those tri - diagonal matrices have been very useful in the last decade , leading to important progress on the understanding of the local eigenvalues statistics in the limit of large dimension @xmath7 for general @xmath12 . at the edge of the spectrum ,
it was first proved @xcite that the largest eigenvalues converge jointly ( when zooming in the edge - scaling region of width @xmath13 around @xmath14 ) to the low lying eigenvalues of a random schrdinger operator called the stochastic airy operator ( see also @xcite ) .
similar results were proved for the bulk in @xcite by valk and virg . for @xmath15 belonging to the wigner sea @xmath16 , the authors of @xcite consider the point process @xmath17 where @xmath18 is distributed according to @xmath5 and @xmath19 is the wigner semi - circle density . indeed , the mean level spacing around level @xmath15 for the points with law @xmath5 is approximately @xmath20 when @xmath21 . the mean point spacing of @xmath22 defined in is therefore of order @xmath23 and in this scaling , one can now investigate the limiting statistics of this point process when @xmath24 .
the authors of @xcite precisely answer this question proving that the point process @xmath22 converges in law to a point process sine@xmath0 on @xmath3 first introduced in @xcite and characterized in terms of a family @xmath25 of coupled one - dimensional diffusion processes , the _
stochastic sine equations_. as expected for the eigenvalues statistics in the bulk , the point process sine@xmath0 is translation - invariant in law .
the family of diffusions @xmath25 can be interpreted as the hyperbolic angle of the brownian carousel with parameter @xmath15 and its law is characterized as follows : given a ( driving ) complex brownian motion @xmath26 , the diffusions @xmath27 satisfy @xmath28 note that all the diffusions @xmath27 are adapted to the filtration of the brownian motion @xmath29 .
this coupling induces a strong interaction between the diffusions which makes the joint law difficult to analyse , as we shall see . a key feature shared by the processes
@xmath30 is that they all converge almost surely as @xmath31 to a limit @xmath32 which is an integer multiple of @xmath23 .
the characterization of the law of the sine@xmath0 point process is a borel set of @xmath3 , sine@xmath33 inside @xmath34 denotes the number of points inside @xmath34 . in other words , sine@xmath0 is the counting measure of the point process .
] can now be enunciated as follows @xmath35)\right)_{\lambda < \lambda ' } \stackrel{(d)}{= } \left ( \frac{\alpha_{\lambda'}(\infty)-\alpha_\lambda(\infty)}{2\pi } \right)_{\lambda<\lambda ' } .
\end{aligned}\ ] ] ( blue curve ) and @xmath36 ( red curve ) for @xmath37 and for a small value of @xmath1 .
we see that whenever @xmath38 jumps , @xmath39 jumps at the same time in agreement with lemma [ inclusion - croissante ] .
, width=453 ] in this paper , we are interested in the limiting law of the sine@xmath0 process when the inverse temperature @xmath1 goes to @xmath2 .
theorem [ main ] is the main result of this paper and gives the convergence as @xmath40 of the sine@xmath0 process towards a poisson point process on @xmath3 .
this convergence at the continuous ( @xmath41 ) level seems rather natural since taking @xmath40 amounts to decreasing the electrostatic repulsion ( and hence the correlation ) at the discrete level , i.e. between the @xmath7 points distributed according to @xmath5 .
if one takes @xmath42 abruptly for fixed @xmath7 , the probability density @xmath43 corresponds ( up to a rescaling depending on @xmath1 ) to the joint law of independent gaussian variables and it is then straightforward to check the convergence ( as @xmath44 ) of the point process with law @xmath43 towards a poisson process .
theorem [ main ] exchanges the order of the limits @xmath45 and @xmath44 , describing the statistics when @xmath44 _ first _ and then @xmath40 . it also gives the precise rate of the convergence .
[ main ] as @xmath45 , the sine@xmath0 point process converges weakly in the space of radon measure ( equipped with the topology of vague convergence @xcite ) to a poisson point process on @xmath3 with intensity @xmath46 .
in particular , we have , for any @xmath47 and @xmath48 , @xmath49=k ] \to_{\beta\to 0 } \exp\left(\frac{\lambda-\lambda'}{2\pi } \right ) \frac { ( \lambda'-\lambda)^k } { ( 2\pi ) ^k k!}\ , , \end{aligned}\ ] ] and the numbers of points of sine@xmath0 inside two disjoint intervals are asymptotically independent .
let us briefly discuss some implications of theorem [ main ] and mention a few related questions on the spectral statistics of random matrices and random schrdinger operators . in @xcite
, the authors have shown that the circular @xmath1-ensemble , which was later shown to be sine@xmath0 in @xcite , interpolates between poisson and clock distributions on the circle ( point process with rigid spacings like the numerals on a clock ) by considering random cmv matrices .
theorem [ main ] provides a more precise description of this interpolating process on the poisson process side . in our study ,
we are led to examine a time homogeneous family of diffusions @xmath50 defined as @xmath51 this family of coupled diffusions also appears in @xcite to describe the law of the limiting point process of a certain critical random discrete schrdinger operator .
our result can be extended in this context to prove that this critical schrdinger operator continuously interpolates between the extended ( clock / picket fence ) and localized ( poisson ) regimes .
more precisely , one could prove using our ideas that the random spectrum has poissonian statistics in the limit of large temperature .
let us also compare the results stated in theorem [ main ] with those of a previous work @xcite where we consider the stochastic airy ensemble , airy@xmath0 , obtained in the scaling limit of @xmath1-ensembles at the edge of the spectrum . in this context
, we proved that the number of points airy@xmath52-\infty,\lambda ] ) $ ] inside the interval @xmath53-\infty,\lambda]$ ] displays poisson statistics in the small @xmath1 limit .
this permitted us to obtain the limiting distributions as @xmath45 of each of the lowest eigenvalues ( individually ) of the airy@xmath0 ensemble .
in particular , we obtained the weak convergence of the @xmath54 distribution towards the gumbel distribution .
although the sine@xmath0 and airy@xmath0 characterizations in law ( in terms of a family of coupled diffusions ) look very similar , the analysis of the limiting marginal statistics of the number of points inside a finite closed interval airy@xmath55 $ ] for @xmath48 and the asymptotic independence when @xmath45 of the respective numbers of points of the airy@xmath0 point process into two disjoint intervals remain open even after our study @xcite . in this aspect ,
theorem [ main ] gives a much more powerful and complete description of the sine@xmath0 process in the small @xmath1 limit . in this case
, we are able to prove the asymptotic independence between the number of points of the sine@xmath0 process in two disjoint intervals .
this part of the proof requires new ideas in order to obtain estimates on the relative positions between two coupled diffusions @xmath56 and @xmath57 .
the nice feature of the sine@xmath0 process is its translation invariance in law .
this property makes the analysis of sine@xmath0 easier than the one of the airy@xmath0 process .
the non - homogeneous intensity of the airy@xmath0 process is governed by the edge - scaling crossover spectral density of @xmath1-ensembles computed explicitly in @xcite for @xmath58 ( see also @xcite ) .
other spectral statistics of random matrices at high temperature , i.e. when @xmath40 , have been investigated in @xcite . in @xcite ,
the authors study the empirical eigenvalue density in the limit of large dimension @xmath7 for @xmath1-ensembles when @xmath1 tends to @xmath2 with @xmath7 as @xmath59 where @xmath60 is a constant .
the authors compute the limiting spectral density @xmath61 explicitly in terms of parabolic cylinder function and establish a gauss - wigner crossover , in the sense that the family @xmath62 interpolate between the gaussian probability distribution ( @xmath63 ) and the wigner semi - circle ( @xmath64 ) .
the case of gaussian wishart matrices has also been studied in @xcite .
it would be interesting to have a description of the crossover statistics of the @xmath1-ensembles obtained in the double scaling limits when @xmath1 tends to @xmath2 with @xmath44 ( this question is briefly discussed in @xcite for the statistics at the edge of the spectrum ) .
_ organization of the paper .
_ we start in section [ marginal - distributions ] by looking at the limiting marginal distributions of the random variables sine@xmath55 $ ] for @xmath65 .
we first study a classical problem on the exit time of a diffusion trapped in the well of a stationary potential ( obtained by neglecting the slow evolution with time ) .
then , we prove that the jump process of @xmath66 converges weakly to an ( inhomogeneous ) poisson point process by first approximating @xmath56 with diffusions processes with piecewise constant drifts on a subdivision of small intervals and then by using the convergence of the exit time of the stationary well established previously .
this requires estimates on the sample paths of a single diffusion , in a spirit similar to @xcite . in section
[ spatial - independence ] , we investigate the asymptotic independence of the numbers of points of sine@xmath0 in two disjoint intervals .
we prove a crucial estimate regarding the typical relative positions of two diffusions @xmath56 and @xmath67 for @xmath48 .
loosely speaking , the main point is to use this estimate to prove that , in the limit @xmath45 , the jumps of the process @xmath68 immediately follow those of @xmath69 ( see fig . [ without - diff ] ) while the processes @xmath68 and @xmath70 never jump at the same time ( see fig . [ with - diff ] ) .
the asymptotic independence follows essentially from the fact that two poisson point processes adapted to the same filtration are independent if and only if they never jump simultaneously .
we gather in the next paragraph important properties of the family @xmath71 already established in [ section 2.2 , @xcite ] that we will use throughout the paper . _
first properties of the coupling of the diffusions @xmath72 : _ 1 . for all @xmath48
, @xmath73 has the same distribution as @xmath74 ; 2 . `` increasing property ''
: @xmath75 is increasing in @xmath15 ; 3 .
@xmath76 is non - decreasing in @xmath77 ; 4 .
@xmath78 = \lambda \frac{\beta}{4}\int_0^t e^{-\beta s/4 } ds $ ] ; 5 .
@xmath79 exists and is an integer a.s .
we will also use the following notation : @xmath80 * acknowledgments * special thanks are addressed to chris janjigian and benedek valk .
we have benefited from insightful and precise comments from them .
their detailed feedback has helped us to improve the second version of this manuscript , especially the proofs of lemmas [ def - poissons ] and [ inclusion - croissante ] .
we are also grateful to them for pointing out references @xcite and the connections with our work .
we thank stphane benoist and antoine dahlqvist for useful comments and discussions .
r. a. received funding from the european research council under the european union s seventh framework programme ( fp7/2007 - 2013 ) / erc grant agreement nr .
258237 and thanks the statslab in dpmms , cambridge for its hospitality at the time this work was finished .
the work of l. d. was supported by the engineering and physical sciences research council under grant ep/103372x/1 and l.d .
thanks the hospitality of the maths department of tu and the weierstrass institute in berlin .
we are first interested in the limiting law of the random variable sine@xmath81 $ ] when @xmath45 for a single fixed @xmath15 . in this case , we re - write the diffusion @xmath56 in a more convenient way : @xmath82 where @xmath83 is a real standard brownian motion ( which depends on @xmath15 ) .
let us introduce the change of variable @xmath84 .
a straightforward computation ( see @xcite ) shows that : @xmath85 in this subsection , we study an exit time problem for a langevin diffusion @xmath86 evolving in a stationary potential @xmath87 defined for @xmath88 as @xmath89 this problem is relevant to our study thanks to the slow variation with time of the non - stationary potential as @xmath45 in which the diffusion @xmath90 evolves .
the diffusion @xmath86 satisfies the following stochastic differential equation @xmath91 in the whole paper , the diffusions @xmath90 and @xmath86 defined respectively in and are coupled , driven by the same brownian motion @xmath83 . under this coupling , we have almost surely @xmath92 for all @xmath93 where @xmath94 is the first explosion ( stopping ) time @xmath95 as a function of @xmath96 .
, width=377 ] in this paragraph we investigate the limiting law of the stopping time @xmath94 through its laplace transform defined for @xmath97 as @xmath98\ , , \end{aligned}\ ] ] where @xmath96 is the initial position of the diffusion @xmath86 . we know from classical diffusion theory ( see also @xcite ) that @xmath99 satisfies @xmath100 let us first examine the expectation of the first explosion time of @xmath101 . due to the strong barrier separating the local minimum and the local maximum ( see fig .
[ fig.potential ] ) , it is natural to expect the asymptotic of this mean exit time not to depend on the starting point of the diffusion , when @xmath45 , as long as it is located in the well ( `` memory - loss property '' ) .
this is the purpose of proposition [ mean.time.stat ] .
we then show that this first exit time properly rescaled by its mean value converges in law to an exponential distribution ( proposition [ conv - exp ] ) when the starting point is in the well .
[ mean.time.stat ] suppose that the diffusion @xmath86 starts from @xmath102 such that @xmath103 .
then its expected exit time denoted @xmath104 $ ] has the following equivalent when @xmath40 : @xmath105 _ proof of proposition [ mean.time.stat ] . _ from , it is easy to see that the expected exit time @xmath106 satisfies the boundary value problem @xmath107 solving explicitly we obtain the integral form @xmath108\right ) dy\ , .
\end{aligned}\ ] ] by extracting carefully the asymptotic behavior of this integral in the limit @xmath45 ( see appendix [ integrals ] ) , we obtained the desired result .
the laplace transform of the exit time satisfies the fixed point equation @xmath109 ) g_\xi(y ) dy .
\end{aligned}\ ] ] with classical arguments similar to those of @xcite , we prove the following proposition in appendix [ integrals ] .
[ conv - exp ] conditionally on @xmath110 such that @xmath111 when @xmath45 , the first explosion time @xmath94 of the diffusion @xmath112 converges weakly when rescaled by the expected exit time @xmath113 towards an exponential distribution with mean @xmath114 .
the process @xmath115 such that @xmath116 satisfies @xmath117 proposition translates into a convergence of the stopping time @xmath118 .
[ conv - exp - theta ] conditionally on @xmath119 such that @xmath120 when @xmath45 , the stopping time @xmath121 converges weakly when @xmath45 towards an exponential distribution with mean @xmath114 .
we consider the diffusion @xmath56 defined in or equivalently for a single @xmath15 , in . for @xmath122 ,
let @xmath123 note that those stopping times correspond to the jumps of the process @xmath66 .
we denote by @xmath124 the filtration associated to the diffusion process @xmath56 . in this paragraph
, we will sometimes omit the subscript @xmath15 in @xmath72 to simplify the notations .
we consider the rescaled empirical measure @xmath125 of the @xmath126 defined on @xmath127 @xmath128 = \sum_{k=1}^{+\infty } \delta_{\zeta_k}\left[0 , \frac{8 \pi t}{\beta } \right]\ , .
\end{aligned}\ ] ] we divide the time interval @xmath127 into random small intervals @xmath129,k\in { \mathbb{n}}$ ] , _ independent _ of the diffusion @xmath130 where @xmath131 and : @xmath132 where the @xmath133 are i.i.d .
random variables with mean @xmath114 uniformly distributed on @xmath134 $ ] . for each @xmath47 , conditionally on @xmath135 , we define the two diffusion processes @xmath136 and @xmath137 such that , for @xmath138 , @xmath139 by the increasing property , it follows that for any @xmath122 and @xmath138 , we have @xmath140 [ stopping - times - pp ] as @xmath45 , the empirical measure @xmath125 converges weakly in the space of radon measure ( equipped with the topology of vague convergence @xcite ) to a poisson point process on @xmath127 with inhomogeneous intensity @xmath141 .
in particular , we have , for any @xmath47 , @xmath142=k ] \to_{\beta\to 0 } \exp(-\frac{\lambda}{2\pi } ( 1-e^{-2\pi t } ) ) \frac{(\frac { \lambda}{2\pi})^k ( 1-e^{-2\pi t})^k } { k!}\ , .
\end{aligned}\ ] ] the proof of theorem [ stopping - times - pp ] is done in the next paragraph [ estimates - one - diff ] .
it uses a careful analysis of the behaviour of a single diffusion @xmath72 in the small @xmath1 limit .
thanks to the equality in law @xmath143\,,\end{aligned}\ ] ] we easily deduce from theorem [ stopping - times - pp ] the convergence of the marginals of the sine@xmath0 point process .
[ marginal - conv - main ] let @xmath48 .
the random variable @xmath144 $ ] converges weakly as @xmath45 to a poisson law with parameter @xmath145 .
we analyse in this paragraph the diffusion @xmath72 and the jumps of @xmath146 in the limit @xmath40 .
the results we obtain are derived using the diffusion @xmath147 satisfying the sde .
we defer the proof of those technical lemmas in appendix [ proof - auxiliary - results-1 ] .
lemma [ fast - explosion - diff ] first states that when the diffusion @xmath72 starts just below @xmath23 modulo @xmath23 , @xmath146 will jump in a short time with probability going to @xmath114 .
[ fast - explosion - diff ] let @xmath148 .
conditionally on @xmath149 , we define the first reaching time @xmath150 then , there exists a constant @xmath151 such that for all @xmath12 small enough , @xmath152 { \;\geqslant\;}1- \beta^c \ , .
\end{aligned}\ ] ] on the large scale - time of the order @xmath153 , we prove that the time spent by @xmath72 near @xmath2 modulo @xmath23 for @xmath72 is negligible .
this is the content of lemma [ expected - time - bad - region - alpha ] .
[ expected - time - bad - region - alpha ] let @xmath154 and @xmath155 \ , . \end{aligned}\ ] ] then , there exists @xmath156 independent of @xmath1 such that @xmath157 we can now prove theorem [ stopping - times - pp ] using the previous estimates : _ proof of theorem [ stopping - times - pp ] .
_ the proof follows the same lines as the proof of [ theorem 4.1 in @xcite ] .
the idea is to approximate the number of jumps of the diffusion @xmath130 by those of stationary diffusions and to use the increasing property . to this end
, we will use the subdivision of @xmath127 introduced above and the diffusions @xmath158 and @xmath159 . from kallenberg s theorem @xcite
, we just need to see that , for any finite union @xmath160 of disjoint and bounded intervals , we have when @xmath45 , @xmath161 & \longrightarrow \lambda \int_i e^{-2\pi t}\ , dt \ , , \label{cv1bis } \\
\p[\mu_\lambda^\beta(i ) = 0 ] & \longrightarrow \exp(- \lambda\int_i e^{-2\pi t}\ , dt ) \label{cv2bis } \ , .
\end{aligned}\ ] ] denote by @xmath162 $ ] the right most interval of @xmath160 , by @xmath163 the union of disjoint and bounded intervals such that @xmath164 $ ] and by @xmath165 the supremum of @xmath163 .
let us use the random subdivision introduced above .
it is crucial to control the position of the diffusion at the starting point of the sub - intervals . as a consequence of lemma [ expected - time - bad - region - alpha ] , with large probability ,
the diffusion @xmath130 is close to @xmath2 modulo @xmath23 in the beginning of each of the sub - intervals overlapping @xmath166 $ ] .
more precisely , denote by @xmath167 and consider : @xmath168 every sub - interval intersecting @xmath166 $ ] is taken care of in this event as by definition @xmath169 $ ] .
its probability is bounded from above by @xmath170 & { \;\leqslant\;}\sum_{k=0}^{\lfloor 2n t_2\rfloor + 1}\p [ \{\alpha((8\pi/\beta ) s_k / n)\}_{2\pi } > 4 \,\arctan(\beta^{1/4})]\\ & { \;\leqslant\;}2\ , \e[\int_{0}^{3n t_2 + 3 } 1_{\{\{\alpha((8\pi/\beta)u / n)\}_{2\pi } > 4 \,\arctan(\beta^{1/4})\ } } du]\\ & { \;\leqslant\;}\frac{n}{4\pi } \beta \ ; t_{\beta}(0,3t_2 + 1 ) \to_{\beta \to 0 } 0\end{aligned}\ ] ] where we used the result of lemma [ expected - time - bad - region - alpha ] to have the convergence towards @xmath2 in the last line .
we now turn to the proof of [ cv1bis ] .
note that thanks to the linearity of the expectation , we simply need to prove for intervals @xmath160 of the form @xmath171 $ ] .
the upper bound simply follows from the sde form : @xmath172 = \lambda \frac{\beta}{4 } \int_0^t e^{-\beta s/4 } ds \,.\end{aligned}\ ] ] we immediately derive : @xmath173 = \e[\lfloor \alpha((8 \pi/\beta)t_2)/(2\pi ) \rfloor ] { \;\leqslant\;}\lambda \int_0^{t_2 } e^{- 2 \pi s } ds \,.\end{aligned}\ ] ] for the lower bound , denote by @xmath174 , @xmath175 and @xmath176 the number of jumps of @xmath158 , @xmath159 and @xmath130 in the sub - interval @xmath177 $ ] .
@xmath178 & { \;\geqslant\;}\sum_{k=0}^{\lfloor 2n t_2\rfloor + 1}\e[n_k^-1_{\{(s_k / n ) < t_2\}}|
\mathcal{c}_k ] - \sum_{k=0}^{\lfloor 2n t_2\rfloor + 1 } \e[n_k^-1_{\{(s_k / n )
< t_2\}}| \mathcal{c}_k ] \p[\mathcal{c}_k^c]\,.\end{aligned}\ ] ] the second sum of the rhs can be simply bounded from above by : @xmath179 \times \sup_{k \in \{0 , \cdots , \lfloor 2n t_2\rfloor + 1\ } } \frac{\p[\mathcal{c}^c_k]}{\p[\mathcal{c}_k]}\,.\end{aligned}\ ] ] the first expectation is bounded from above by a number independent of @xmath1 and the second term ( bounded by @xmath180/(1-\p[\mathcal{c}^c])$ ] ) tends to @xmath2 as @xmath40 .
moreover , thanks to proposition [ conv - exp - theta ] , we have @xmath181 \to_{\beta \to 0 } \e[\lambda ( \tau_{k+1}/n ) \;\exp(-2 \pi
s_k / n ) ] .\end{aligned}\ ] ] using the convergence of the riemann sum when @xmath182 , it gives the desired lower bound .
let us now examine the convergence [ cv2bis ] for a single interval @xmath183 $ ] first . by the markov property : @xmath184 = 0 ] { \;\leqslant\;}\e\big[\prod_{k=0}^{\infty } \p\big[n_k^+ = 0 \,\big| \mathcal{c}_k , ( \tau_i)_i\big ] 1_{\{s_{k+1}/n { \;\geqslant\;}t_1,\;s_{k}/n { \;\leqslant\;}t_2\}}\big ] + \p[\mathcal{c}^c]\,.\end{aligned}\ ] ] using the convergence in each of the sub - intervals ( given by proposition [ conv - exp - theta ] ) @xmath185 \to_{\beta \to 0 } \e[\exp(- \lambda ( \tau_{k+1}/n ) \;\exp(-2 \pi
s_k / n ) ) ) ] , \end{aligned}\ ] ] we obtain : @xmath186 = 0 ] { \;\leqslant\;}\e[\prod_{k=0}^{\infty}\exp(- \lambda ( \tau_{k+1}/n ) \;\exp(-2 \pi
s_k / n ) ) ) 1_{\{s_{k+1}/n { \;\geqslant\;}t_1,\;s_{k}/n { \;\leqslant\;}t_2\}}]\,.\end{aligned}\ ] ] thanks to the convergence of the riemann sum when @xmath187 , we deduce the upper bound . the lower bound
can be done using similar techniques than above . to generalize the result
to @xmath160 finite union of interval , note that thanks to the simple markov property , we have @xmath188 = \e[1_{\{\mu_{\beta}(j ) = 0\}}\;\p[\mu_{\beta}([t_1,t_2 ] ) = 0| \alpha(t_0)]].\end{aligned}\ ] ] iterating the previous argument leads to the result .
for @xmath48 , we now control the expected time spent by the process @xmath189 below the process @xmath190 .
the following lemma is a crucial step towards the proof of the asymptotic independence of the limiting point process on disjoint intervals and will be used for the proof of lemmas [ inclusion - croissante ] and [ poissons - indep ] .
[ ordering - two - diffusions ] let @xmath191 and @xmath192 \ , .
\end{aligned}\ ] ] then , there exists two constants @xmath193 independent of @xmath1 such that @xmath194 _ proof . _
we set @xmath195 before evaluating the probability of the event @xmath196 , we need to introduce for @xmath197 the last jump of the process @xmath198 before time @xmath199 , i.e. @xmath200 the main idea is to prove that any @xmath201 is associated to a jump time of @xmath57 right before @xmath199 .
we set @xmath202 and bound from above the probability of the event @xmath196 by @xmath203 & \le \p\left [ \mathcal{e}_u \cap \left\{\zeta^{(u ) } \in \left[u_0 , u\right ] \right\ } \right ] + \p\left [ \mathcal{e}_u \cap\left\ { \zeta^{(u ) } < u_0 \right\}\right]\notag \\ & \le \p\left [ \mathcal{e}_u \cap \left\{\zeta^{(u ) } \in \left[u_0 , u\right ] \right\ } \right ] + \p\left[\bigcap_{s\in [ u_0,u ] } \mathcal{e}_s \right ] \label{good - shape1}\end{aligned}\ ] ] where we have noticed the inclusion @xmath204 } \mathcal{e}_s $ ] for the second line ( see fig .
[ fig - alpha - eu ] ) .
indeed the definition of @xmath205 implies that the process @xmath198 has not jumped during the time interval @xmath206 $ ] so that the relative ordering ( modulo @xmath23 ) @xmath207 at time @xmath208 has to be preserved for all @xmath209 .
.,width=340 ] we now tackle the second probability of using the fact that @xmath56 is close to @xmath2 modulo @xmath23 with large probability : @xmath210 } \mathcal{e}_s \right ] & \le \p\left [ \bigcap_{s\in [ u_0,u ] } \mathcal{e}_s \cap \left\ { \{\alpha_\lambda(u_0)\}_{2\pi } \le 4 \arctan(\beta^{1/4 } ) \right\ } \right]\\ & + \p\left[\{\alpha_\lambda(u_0)\}_{2\pi } \ge 4 \arctan(\beta^{1/4 } ) \right]\ , .
\end{aligned}\ ] ] we introduce the stopping time @xmath211 and notice that , conditionally on the event @xmath212 , we have almost surely for all @xmath213 $ ] , @xmath214 conditionally on the event @xmath215 , we define a diffusion process @xmath216 and its associated first reaching time of @xmath23 such that @xmath217 endowed with those definitions , we have the following upper - bound @xmath210 } \mathcal{e}_s \cap \mathcal{a}_{u_0 } \right ] & = \p\left[\left\ { \zeta_>(u_0 ) \ge u \right\ } \cap \mathcal{a}_{u_0 } \right]\\ & \le \p\left[{\widetilde}\zeta_{2\pi } \ge 9 \log\frac{1}{\beta } \right ] \le \beta^c \end{aligned}\ ] ] where we have used the equality in law @xmath218 for the second line and lemma [ fast - explosion - diff ] ( as well as the increasing property ) to obtain the last upper - bound . gathering the above inequalities
, we obtain @xmath203 & \le \p\left[\zeta^{(u ) } \in \left[u_0 , u\right ] \right ] + \beta^{c } + \p\left[\{\alpha_\lambda(u_0 ) \}_{2\pi } \ge 4 \arctan(\beta^{1/4 } ) \right]\,.\end{aligned}\ ] ] now , to conclude , we just have to integrate this latter inequality with respect to @xmath199 .
first notice that we have almost surely @xmath219 \right \ } } du \le 9 \log\frac{1}{\beta } \ , \mu_{\lambda'}^\beta[0,\frac{8\pi}{\beta}t]\ , .
\end{aligned}\ ] ] integrating the other terms as well with respect to @xmath220 $ ] and taking the expectation , we get @xmath221 \right]\right ) + \xi_\beta(t ) \\ & \le 9 \log \frac{1}{\beta}\left(1 + \frac{\lambda}{2\pi }
\right ) + 8\pi\ , t\ , \beta^{c-1 } + \xi_\beta(t ) \end{aligned}\ ] ] where @xmath222 is defined in .
the conclusion now follows from lemma .
lemma [ ordering - two - diffusions ] shall be an important tool to prove the asymptotic independence between @xmath32 and @xmath223 for @xmath48 .
theorem [ stopping - times - pp ] gives the weak convergence of the random measures @xmath125 and @xmath224 in the space of measures on @xmath127 equipped with the topology of vague convergence denoted @xmath225 . due to the equality in law @xmath226 theorem [ stopping - times - pp ] also implies the weak convergence of the ( positive ) random measure @xmath227 such that for all @xmath228 , @xmath229 : = \lfloor \frac{\alpha_{\lambda'}(t)-\alpha_\lambda(t)}{2\pi } \rfloor\end{aligned}\ ] ] towards a poisson measure @xmath230 with intensity @xmath231 .
we now work with the two diffusions @xmath56 and @xmath57 for @xmath48 _ coupled according to _ .
we are interested in the limiting joint distribution of the triplet of random measures @xmath232 according to this coupling . from the above convergences , it is straightforward to deduce the relative - compactness of the family of the triplets of ( random ) measures @xmath233 for the weak topology over @xmath234 equipped with the product topology of vague convergence .
let us take a sequence @xmath235 when @xmath236 such that the processes @xmath237 converge jointly weakly in the product space when @xmath238 to a triplet @xmath239 of point measures on @xmath127 whose marginal distributions are given respectively by the law of the poisson measures @xmath240 and @xmath241 and @xmath242 and whose joint law depends _ a priori _ on the chosen sub - sequence @xmath243 . in the following , we will study this triplet and we will drop the superscript @xmath244 to ease the notations .
we shall in fact see later that all the possible limit point have the same law .
therefore the law of the triplet does not depend on @xmath244 and the weak convergence of holds ( see remark [ remark - inclusion - facult ] ) .
in the next lemma , we regard the point measures @xmath245 and @xmath242 as poisson point processes and prove that they are indeed jointly poisson processes on a common filtration .
this is an important step for our needs .
[ def - poissons ] let @xmath246 and @xmath247 be the natural filtration associated to the process @xmath248 i.e. such that @xmath249 then , the processes @xmath240 and @xmath241 and @xmath242 are @xmath250-poisson point processes with respective intensities @xmath251 and @xmath252 and @xmath253 . the main point used to prove this lemma is that all the diffusions are measurable with respect to the same driving ( complex ) brownian motion @xmath29 as they are strong solutions of the sdes .
a proof can be found in appendix [ appendix - def - poissons ] . to fix notations
, we will denote by @xmath254 resp . @xmath255 the points associated to the poisson processes @xmath240 and @xmath241 and @xmath230 such that @xmath256 = \sum_i \delta_{\xi_i^\lambda}[0,t]\ , , \quad \mathcal{p}_{\lambda'}[0,t ] = \sum_i \delta_{\xi_i^{\lambda'}}[0,t ] \
, , \quad \mathcal{p}_{\lambda'-\lambda}[0,t ] = \sum_i \delta_{\xi_i^{\lambda'-\lambda}}[0,t ] \ , .
\end{aligned}\ ] ] for @xmath257 , we also recall the notations @xmath258 = \sum_i \delta_{\zeta_i^\lambda}[0,\frac{8 \pi t}{\beta}]\ , , \quad \mu^\beta_{\lambda'}[0,t ] = \sum_i \delta_{\zeta_i^{\lambda'}}[0,\frac{8 \pi t}{\beta } ] \ , , \quad \mu^\beta_{\lambda'-\lambda}[0,t ] = \sum_i \delta_{\zeta_i^{\lambda'-\lambda}}[0,\frac{8 \pi t}{\beta } ] \,.\end{aligned}\ ] ] [ inclusion - croissante ] let @xmath48 .
then , we have almost surely @xmath259 i.e. for all @xmath260 , there exists @xmath261 such that @xmath262 . _ proof .
_ we have to prove that for any @xmath154 , @xmath263 = 0\ , .
\end{aligned}\ ] ] the probability is the increasing limit of the sequence @xmath264 defined as @xmath265 \ , . \end{aligned}\ ] ] it suffices to prove that @xmath266 for any @xmath267 .
let us introduce the probability @xmath268\ , .
\end{aligned}\ ] ] denote by @xmath269 the space of point measures and recall that it is closed in the space @xmath270 for the vague convergence topology . notice that the set @xmath271 is open in @xmath272 equipped with the product topology of the vague convergence on the space of point measures .
it comes from the straightforward fact that if @xmath273 converges towards @xmath274 for the vague topology , the points of @xmath275 belonging to @xmath276 $ ] for all but finitely many @xmath277 converge to points of @xmath278 in @xmath276 $ ] .
if @xmath267 is fixed , we therefore have @xmath279 using the joint convergence of @xmath280 in the space @xmath272 ( along the subsequence @xmath281 ) and the portmanteau theorem .
it suffices to check that @xmath282 we need to work with a random subdivision of the interval @xmath283 $ ] .
as before , we consider a sequence @xmath284 of i.i.d .
positive random variables distributed uniformly in @xmath285 $ ] and form the sum @xmath286 . noting that for all @xmath287 such that @xmath288 , there exists @xmath47 such that @xmath289 $ ] , we obtain @xmath290 \mbox { but not } \lfloor \frac { \alpha_{\lambda'}}{2\pi } \rfloor \right]\ , .
\end{aligned}\ ] ] due to the increasing property , the event inside the probability can not happen if the process @xmath291 starts below @xmath292 at the beginning of the interval .
therefore , @xmath293 _ { 2\pi}\right ] \ , .
\end{aligned}\ ] ] which can in turn be upper - bounded as follows @xmath294 _ { 2\pi}\right ] & \le \frac{\beta n}{4\pi } \ , \e\left[\int_0^{\frac{8\pi}{\beta } ( 3 t+1 ) } 1_{\left\{\{\alpha_{\lambda ' }
( u ) \}_{2\pi } \le [ \alpha_{\lambda } ( u ) ] _ { 2\pi}\right\ } } du \right ] \notag\\ & \le \frac{\beta n}{4\pi } \theta_\beta(3t+1 ) \
, \le \beta^c \label{estimate1}\end{aligned}\ ] ] where we have used lemma [ ordering - two - diffusions ] to obtain the last inequality which holds for @xmath1 small enough ( @xmath295 is a constant which does not depend on @xmath1 ) .
[ poissons - indep ] let @xmath48 .
then the @xmath250-poisson point processes @xmath296 and @xmath297 are independent .
( blue curve ) and @xmath36 ( red curve ) together with @xmath298 ( green curve ) for @xmath37 and for a small value of @xmath1 .
again we see that @xmath38 and @xmath299 never jump simultaneously while @xmath39 and @xmath299 always jump at the same times in agreement with lemma [ poissons - indep ] , lemma [ inclusion - croissante ] and remark [ remark - inclusion - facult ] .
, width=453 ] _ proof of lemma [ poissons - indep ] . _ from a classical result ( see proposition ( 1.7 ) chapter xii , 1 , p.473 in @xcite ) on poisson processes , we know that it suffices to prove that the two @xmath250-poisson processes @xmath296 and @xmath297 do not jump simultaneously , i.e. that for @xmath154 , @xmath300 = 0\ , . \end{aligned}\ ] ] for @xmath301 , we consider the probability @xmath302\ , .
\end{aligned}\ ] ] if @xmath267 is fixed , then we have the convergence ( the studied set is open as in the proof lemma [ inclusion - croissante ] ) @xmath303\ , . \end{aligned}\ ] ] to prove , it therefore suffices to prove that @xmath304 for this , we need to work with a random subdivision of the interval @xmath283 $ ] . as before ,
we consider a sequence @xmath284 of i.i.d .
positive random variables uniformly distributed in @xmath285 $ ] and independent of the processes and form the sum @xmath286 ( @xmath305 ) . the probability can be upper - bounded as follows @xmath306 \right]\notag \\ & = \p\left[\exists k \le \lfloor 2 n t\rfloor+1 : \lfloor \frac{\alpha_{\lambda ' } - \alpha_\lambda}{2\pi } \rfloor \mbox { and } \lfloor \frac { \alpha_\lambda}{2\pi } \rfloor \mbox { both jump on the interval } \frac{8 \pi}{\beta } \left [ \frac{s_k}{n } , ( \frac{s_k}{n } + \frac{2}{n } ) \right ] \right]\notag\\ \label{sums - to - estimate } & \le \sum_{k=1}^ { \lfloor 2 n t\rfloor + 1 } \p\left [ \{\alpha_{\lambda ' } ( \frac{8 \pi}{\beta } \frac{s_k}{n } ) \}_{2\pi } \le [ \alpha_{\lambda } ( \frac{8 \pi}{\beta } \frac{s_k}{n } ) ] _ { 2\pi}\right ] \\ & + \sum_{k=1}^{\lfloor 2 n t\rfloor+1 } \p\left [ \lfloor \frac { \alpha_{\lambda'}}{2\pi } \rfloor \mbox { jumps two times during the interval } \frac{8 \pi}{\beta } \left [ \frac{s_k}{n } , ( \frac{s_k}{n } + \frac{2}{n } ) \right ] \right]\ , .
\end{aligned}\ ] ] for this bound , we have used the fact that , conditionally on @xmath307 _ { 2\pi}\,,\ ] ] the increasing property and the equality in law impose @xmath308 \right \ } \label{event1 } \\ & \subseteq \left\ { \lfloor \frac { \alpha_{\lambda'}}{2\pi } \rfloor \mbox { jumps two times during the interval } \frac{8 \pi}{\beta } \left [ \frac{s_k}{n } , ( \frac{s_k}{n } + \frac{2}{n } ) \right ] \right \ } \notag \ , .
\end{aligned}\ ] ] indeed , the equality in law implies that @xmath309 is increasing with respect to @xmath77 ( once the difference @xmath310 has reached the value @xmath311 where @xmath47 , it remains forever above this value ) .
this fact and the increasing property imply that under the event , the process @xmath312 has to jump two times on the interval @xmath313 $ ] ( see fig .
[ fig-2jumps ] ) .
we now estimate the two sums in .
the first one was already tackled in . for the other sum in , we write @xmath314 \right ] \\ & = \sum_{k=1}^{\lfloor 2 n t\rfloor+1 } \p\left [ \mu_{\lambda'}^\beta\left[\frac{8 \pi}{\beta } \left [ \frac{s_k}{n } , ( \frac{s_k}{n } + \frac{2}{n } ) \right]\right ] = 2 \right]\ , .
\end{aligned}\ ] ] if @xmath267 is fixed and @xmath45 , theorem [ stopping - times - pp ] gives the following convergence @xmath315 = 2 \right]\notag \\ & \rightarrow_{\beta\to 0 } \frac{\lambda^2}{16\pi^2}\sum_{k=1}^{\lfloor 2 n t\rfloor+1 } \exp\left(-\frac{\lambda}{2\pi } e^{-2\pi\frac{s_k}{n } } ( 1-e^{-2\pi\frac{2}{n } } ) \right ) e^{-4\pi\frac{s_k}{n } } ( 1-e^{-2\pi\frac{2}{n } } ) ^2 \notag \\ & = o(n^{-1 } ) \label{estimate2 } \ , .
\end{aligned}\ ] ] the convergence now follows from the two estimates and and the independence between the two poisson processes @xmath240 and @xmath230 is proved . [ remark - inclusion - facult ] using similar arguments as in the proof of lemma [ inclusion - croissante ] , one could also show that for all @xmath48 , @xmath316 .
using then lemma [ poissons - indep ] , it implies the almost - sure equality @xmath317 i.e we do nt have the equality between @xmath318 and @xmath319 . ] .
it therefore determines the law of the triplet @xmath320 and the weak convergence of the family when @xmath40 holds .
nevertheless , we do not need this stronger result as we only use for the @xmath321 case the previous equality for the total number of points in @xmath127 i.e. @xmath322 as an immediate consequence of lemma [ inclusion - croissante ] and lemma [ poissons - indep ] , we obtain the following result using the characterization of independent poisson processes mentioned above . [ final - independence ] let @xmath323 .
then , the @xmath250-poisson processes @xmath324 and @xmath325 are independent . from kallenberg
s theorem @xcite , we simply need to check that , for any finite union @xmath160 of disjoint and bounded intervals , we have when @xmath45 , @xmath326 & \longrightarrow \frac{1}{2\pi } \int_i d\lambda \ , , \label{cv1bis - main } \\
\p[{\rm sine}_\beta(i ) = 0 ] & \longrightarrow \exp(- \frac{1}{2\pi } \int_i d\lambda ) \label{cv2bis - main } \ , . \end{aligned}\ ] ] the convergence follows from corollary [ marginal - conv - main ] . towards
, we consider first a union of two disjoint interval of the form @xmath327 \cup [ \lambda_2,\lambda_3]\,,\end{aligned}\ ] ] where @xmath328 . by translation invariance of the sine@xmath0 point process
, we can suppose without loss of generality that @xmath329 . by definition of sine@xmath0
, we have the following equality in law @xmath330 , { \rm sine}_\beta[\lambda_2 , \lambda_3 ] , \right ) \stackrel{(d)}{= } \left ( \frac{\alpha_{\lambda_1}(\infty)}{2\pi } , \frac{\alpha_{\lambda_3 } ( \infty)- \alpha_{\lambda_2 } ( \infty ) } { 2\pi } \right ) \ , .
\end{aligned}\ ] ] using lemma [ final - independence ] , we deduce that @xmath331 = 0 , { \rm sine}_\beta[\lambda_2 , \lambda_3 ] = 0\right]\\ & \rightarrow_{\beta\to 0 } \p\left[\mathcal{p}_{\lambda_1}(\infty ) = 0 \right ] \p\left[\mathcal{p}_{\lambda_3-\lambda_2}(\infty ) = 0 \right ] = \exp\left(- \frac{\lambda_1}{2\pi } \right)\exp\left(- \frac{\lambda_3-\lambda_2}{2\pi } \right ) \ , .
\end{aligned}\ ] ] now , if @xmath332 and @xmath333 \cup \cdots \cup [ \lambda_{k-1},\lambda_k]$ ] where @xmath334 , the limiting poisson point processes @xmath335 obtained as the limit of the processes @xmath336 never jump simultaneously ( there are pairwise independent ) .
therefore , it is a family of independent processes ( see proposition ( 1.7 ) chapter xii , 1 , p.473 in @xcite ) .
theorem [ main ] is proved .
_ asymptotic of the integral . _
we denote by @xmath337 ( respectively @xmath338 ) the local minimum ( resp .
maximum ) of the potential @xmath339 and derive asymptotic expansions for @xmath337 and @xmath338 in the limit @xmath45 using @xmath340 we deduce that @xmath341 we consider the laplace transform @xmath346 of the rescaled exit time @xmath347 of the diffusion starting from @xmath96 . if suffices to prove that , if @xmath348 is such that @xmath103 as @xmath45 , we have for any @xmath349 , @xmath350 ( laplace transform of an exponential distribution with parameter @xmath114 ) . to simplify notations ,
we will just write @xmath351 for @xmath346 . using the asymptotic , we obtain a lower bound @xmath355 plugging into , we get an upper bound @xmath356\right ) t_\beta(y ) dy\ , .
\end{aligned}\ ] ] after the derivation of the asymptotic quadruple integral ( similar to the one of ) , we obtain @xmath357 iterating this argument ( the multiple integrals are always such that the integration ranges permit to catch the maximum value of @xmath358 inside the exponential as in ) , we get the result . recall the definition of @xmath359 and its differential equation : @xmath360 and denote by @xmath361 the first passage time to @xmath362 after time @xmath77 i.e. @xmath363 in this sub - section , we first study the diffusion @xmath364 and then translate the estimates to the diffusion @xmath72 .
we will denote by @xmath365 the law of the diffusion @xmath90 starting from position @xmath366 at time @xmath77 . to simplify notations ,
we omit the subscript @xmath77 in @xmath365 if @xmath367 and write @xmath368 instead of @xmath361 if it appears inside the probability @xmath365 .
we will also denote by @xmath369 the law of a brownian motion starting from @xmath370 .
_ proof of lemma [ fast - explosion]_. we have @xmath371 \ge \p_{\varepsilon
\log\frac{1}{\beta } } \left [ t_{2 \log\frac{1}{\beta } } < 8 \log \frac{1}{\beta } \wedge t_{\frac{\varepsilon}{2}\log\frac{1}{\beta } } \right ] \p_{2 \log\frac{1}{\beta},8\log\frac{1}{\beta } } \left[t_{+\infty } < \log\frac{1}{\beta } \right]\ , .
\end{aligned}\ ] ] if @xmath372 , the drift term @xmath373 for @xmath1 small enough .
thus , for @xmath374 , we have @xmath375 .
therefore , @xmath376
\ge p_{\varepsilon \log\frac{1}{\beta } } \left [ \inf_{0 \le t \le 8\log\frac{1}{\beta } } b_t >
\frac{\varepsilon}{2}\log\frac{1}{\beta } \right]\ , . \ ] ] this latter probability is easily computed with the reflection principle for brownian motion .
@xmath377 = p_0\left[\sup_{0 \le t \le 8\log\frac{1}{\beta } } b_t \le \frac{\varepsilon}{2}\log\frac{1}{\beta}\right ] \\ & = p_0 \left [ |b(1)| \le \frac{\varepsilon}{4\sqrt{2 } } \sqrt{\log\frac{1}{\beta } } \right ] = 1- o(\exp(-c\log\frac{1}{\beta}))\ , .
\end{aligned}\ ] ] we now consider the probability @xmath378\end{aligned}\ ] ] we define @xmath379 such that @xmath380 .
the function @xmath381 satisfies the following random _ ordinary _ differential equation @xmath382 on the event @xmath383 which occurs with probability bigger than @xmath384 where @xmath60 is a constant independent of @xmath1 , we can easily check that , for @xmath385 $ ] and @xmath12 small enough , @xmath386 this leads us to study the cauchy problem @xmath387 this ordinary differential equation can be solved explicitly .
we find @xmath388 it is easy to see that the exploding time of @xmath389 is of order @xmath390 as @xmath45 . on the event @xmath391
, we have @xmath392 for all @xmath228 .
_ proof of lemma [ lower - bound - proba - fast - explosion]_. the probability under consideration can be bounded below by @xmath395 \p_{\frac{1}{4 } \log\frac{1}{\beta } , \log\frac{1}{\beta } } \left [ t_{+\infty } < 9 \log \frac{1}{\beta } \right ] \,.\end{aligned}\ ] ] for the first probability , we note that @xmath396 and thus @xmath395 & { \;\geqslant\;}p_{0 } \left [ b ( \log\frac{1}{\beta } ) - \frac{1}{2 } \log \frac{1}{\beta } \ge \frac{1}{2 } \log \frac{1}{\beta } \right]\\ & = p_{0 } \left[b_1 { \;\geqslant\;}\sqrt { \log\frac{1}{\beta } } \right]\sim \frac{\sqrt{\beta}}{\sqrt{2\pi\log\frac{1}{\beta } } } \ , .
\end{aligned}\ ] ] on the large scale - time of the order @xmath153 , we prove that the time spent by @xmath364 in the well of the potential @xmath398 ( near @xmath2 modulo @xmath23 for @xmath72 ) is negligible , i.e. small compared to the typical time @xmath153 between two jumps .
this is lemma [ time - spent - bad - region ] : _ proof of lemma [ time - spent - bad - region ] .
_ the key estimate is the lower - bound given by lemma [ lower - bound - proba - fast - explosion ] on the probability for the diffusion @xmath90 starting from the position @xmath401 to blow up in a short time .
the idea is then to relate the time spent by the diffusion @xmath90 above the level @xmath401 with the number of explosions in the interval @xmath402 $ ] which is of order @xmath403 ( the typical time between two explosions is of order @xmath153 from lemma [ conv - exp ] ) . for any @xmath404 , we have @xmath405\\ & { \;\leqslant\;}\p_{-\infty}\left[r_\lambda ( u ) { \;\geqslant\;}- \frac{1}{4 } \log\frac{1}{\beta } , t_{+\infty } < 10 \log\frac{1}{\beta } \right ] + \p_{-\infty}\left[r_\lambda ( u ) { \;\geqslant\;}- \frac{1}{4 } \log\frac{1}{\beta } , t_{+\infty } \ge 10 \log\frac{1}{\beta } \right ] \\ & { \;\leqslant\;}\p_{-\infty}\left[r_\lambda ( u ) { \;\geqslant\;}- \frac{1}{4 } \log\frac{1}{\beta } , t_{+\infty } < 10 \log\frac{1}{\beta } \right ] + \left(1-\sqrt{\beta}\right)\ , \p_{-\infty}\left[r_\lambda ( u ) { \;\geqslant\;}- \frac{1}{4 } \log\frac{1}{\beta}\right]\,,\end{aligned}\ ] ] where we have used the simple markov property , the inequality as well as the increasing property in the third line .
we deduce that @xmath406 { \;\leqslant\;}\frac{1}{\sqrt{\beta } } \ , \p_{-\infty}\left[r_\lambda ( u ) { \;\geqslant\;}- \frac{1}{4 } \log\frac{1}{\beta } , t_{+\infty } < 10 \log\frac{1}{\beta } \right ] \notag \\ & { \;\leqslant\;}\frac{1}{\sqrt{\beta } } \ , \p_{-\infty}\left[\mbox{the interval $ [ u , u+ 10 \log\frac{1}{\beta } ] $ contains at least one explosion}\right ] \label{control - at - time - u } \ , .
\end{aligned}\ ] ] denoting by @xmath407 $ ] the total number of explosions of the diffusion @xmath90 in the interval @xmath402 $ ] and by @xmath408 the explosion times , we easily see that almost surely @xmath409 \ } } du \le 10 \log\frac{1}{\beta}\ , \left(1 + \mu_\lambda^\beta[0,t]\right ) \ , .
\end{aligned}\ ] ] using this inequality and integrating with respect to @xmath199 , we finally obtain @xmath410 \right ) \ , \frac{10}{\sqrt{\beta } } \log\frac{1}{\beta } \ , { \;\leqslant\;}\left(1+\frac{\lambda}{2\pi}\right ) \frac{10}{\sqrt{\beta } }
\log\frac{1}{\beta } \ , .
\end{aligned}\ ] ] denote @xmath411 the space of continuous functions @xmath412 such that @xmath413 as @xmath414 . to prove that @xmath296 is indeed a @xmath250-poisson process with the correct intensity
, it suffices to check that its laplace functional satisfies
@xmath415 = \exp\left(- \int_{s}^t ( 1-e^{-f(u ) } ) \lambda e^{-2 \pi u } du \right)\,.\end{aligned}\ ] ] we have to check that the filtration @xmath250 does not contain too much information compared to the natural filtration denoted @xmath416 associated to the process @xmath296 only , for to remain valid .
the conclusion of theorem [ stopping - times - pp ] is equivalent to the convergence of the laplace functional of the point process @xmath125 to the laplace functional of the poisson process @xmath296 ( see proposition 11.1.viii ( ii ) of @xcite ) .
let us fix @xmath420 .
we can deduce that for any @xmath421 and any @xmath422 , almost - surely : @xmath423 \rightarrow_{\beta\to 0 } \exp\left(- \int_{s+ { \varepsilon}}^t \left(1-e^{-\varphi(u)}\right ) \lambda
e^{-2 \pi u } du \right)\ , .
\end{aligned}\ ] ] the main point is that the sigma field @xmath424 already contains the information on the two diffusions @xmath425 and @xmath426 up to time @xmath427 as they are strong solutions to the stochastic differential equation system .
note also that we need to introduce a small gap between @xmath77 and @xmath428 so that [ cond - conv ] is valid : it is indeed possible that the position of @xmath429 induces a jump quickly after @xmath427 .
this issue can be circumvented by examining an independent random time @xmath430 distributed uniformly over the interval @xmath431 $ ] .
the position of @xmath432 belongs to an interval of the type @xmath433 $ ] with probability going to @xmath114 when @xmath40 ( this is lemma [ expected - time - bad - region - alpha ] ) .
a straightforward adaptation of the proof of theorem [ stopping - times - pp ] then shows that the conditional law of the measure @xmath434 over @xmath435 converges to the poisson measure of intensity @xmath436 and holds . from the joint convergence of the triplet along the sequence @xmath437
, we can deduce for @xmath438 , @xmath439 \notag \\ = \e\left[\exp\left(-\int_{s+{\varepsilon}}^t \varphi(u ) \mathcal{p}_\lambda(du)\right ) \exp\left(-\int_0^s ( f(u ) \mathcal{p}_\lambda(du ) + g(u)\mathcal{p}_{\lambda'}(du ) + h(u)\mathcal{p}_{\lambda''-\lambda'}(du ) ) \right)\right ] \label{with - joint - conv - mu } \ , .
\end{aligned}\ ] ] on the other hand , using in addition , we can check that @xmath440 \notag \\ & = \exp\left(- \int_{s+{\varepsilon}}^t \left(1-e^{-\varphi(u)}\right ) \lambda
e^{-2 \pi u } du \right ) \e\left[\exp\left(-\int_0^s ( f(u ) \mathcal{p}_\lambda(du ) + g(u)\mathcal{p}_{\lambda'}(du ) + h(u)\mathcal{p}_{\lambda''-\lambda'}(du ) ) \right)\right ] \label{with - joint - conv - cond } \ , .
\end{aligned}\ ] ] gathering and and taking the limit @xmath441 ( using that a.s .
@xmath296 does not jump on @xmath427 ) , we obtain .
20 j. wishart .
_ generalized product moment distribution in samples_. biometrika * 20 * a 425 ( 1928 ) .
_ on the statistical distribution of the widths and spacings of nuclear resonance levels
cambridge philos .
soc . , * 47 * , 790 - 798 xiii , 3 ( 1951 ) .
anderson , a. guionnet , and o. zeitouni .
an introduction to random matrices , cambridge studies in advanced mathematics ( cambridge university press , cambridge , 2009 ) .
z. bai and j. silverstein .
spectral analysis of large dimensional random matrices ( springer , new york , 2010 ) , 2nd ed .
, see theorem 9.2 . m.l .
random matrices ( elsevier , new york , 2004 ) .
p.j . forrester .
log gases and random matrices ( princeton university press , princeton , 2010 ) .
g. akemann , j. baik , and ph .
di francesco . the oxford handbook of random matrix theory ( oxford university press , new york , 2011 ) .
i. dumitriu and a. edelman . _ matrix models for beta ensembles _ j. math
* 43 * , 5830 - 5847 ( 2002 ) .
r. allez and a. guionnet .
_ a diffusive matrix model for invariant @xmath1-ensembles_. electron
. j. probab . * 18 * 62 , 1 - 30 ( 2013 ) .
r. allez , j .-
bouchaud and a. guionnet .
invariant @xmath1-ensembles and the gauss - wigner crossover .
lett . * 109 * , 094102 ( 2012 ) . j. a. ramrez , b. rider and b. virg . _ beta ensembles , stochastic airy spectrum , and a diffusion_. j. amer . math .
* 24 * 919 - 944 ( 2011 ) .
a. edelman and b. d. sutton .
_ from random matrices to stochastic operators_. j. stat
127 * 6 , 1121 - 1165 ( 2007 ) .
b. valk and b. virg .
_ continuum limits of random matrices and the brownian carousel_. invent . math .
* 177 * 463 - 508 ( 2009 ) .
r. allez and l. dumaz .
_ tracy - widom at high temperature_. j. stat .
( online first ) arxiv:1312.1283 ( 2014 ) .
m. j. bowick and e. brzin .
_ universal scaling of the tail of the density of eigenvalues in random matrix models_. phys .
b * 268 * 21 - 28 ( 1991 ) .
p. j. forrester .
_ spectral density asymptotics for gaussian and laguerre - ensembles in the exponentially small region_. j. phys .
a math . gen . * 45 * 075206 ( 2012 ) .
r. killip and m. stoiciu .
_ eigenvalue statistics for cmv matrices : from poisson to clock via random matrix ensembles_. duke math .
* 146 * 361 - 399 ( 2009 ) .
f. nakano .
_ level statistics for one - dimensional schrdinger operators and gaussian beta ensemble_. j. stat
* 156 * 66 - 93 ( 2014 ) .
e. kritchevski b. valk and b. virg . _ the scaling limit of the critical one - dimensional random schrdinger operator_. commun .
phys . * 314 * , 775 - 806 ( 2012 ) .
o. kallenberg .
_ random measures , 4th edition_. academic press , new york , london ; akademie - verlag , berlin ( 1986 ) .
r. allez , j .-
bouchaud , s. n. majumdar , p. vivo . _ invariant @xmath1-wishart ensembles , crossover densities and asymptotic corrections to the marchenko - pastur law_. j. phys .
a : math . theor . * 46 * 015001 ( 2013 ) .
l. dumaz and b. virg . _ the right tail exponent of the tracy - widom - beta distribution_. ann .
h. poincar probab .
statist . *
49 * , 4 , 915 - 933 , ( 2013 ) .
d. revuz , m. yor .
_ continuous martingales and brownian motion .
third edition_. springer ( 1999 ) .
d. j. daley and d. vere - jones .
_ an introduction to the theory of points processes .
volume ii : general theory and structure .
second edition . _ springer ( 2008 ) . | we study the sine@xmath0 process introduced in [ b. valk and b. virg .
invent . math .
* 177 * 463 - 508 ( 2009 ) ] when the inverse temperature @xmath1 tends to @xmath2 .
this point process has been shown to be the scaling limit of the eigenvalues point process in the bulk of @xmath1-ensembles and its law is characterized in terms of the winding numbers of the brownian carrousel at different angular speeds . after a careful analysis of this family of coupled diffusion processes , we prove that the sine@xmath0 point process converges weakly to a poisson point process on @xmath3 .
thus , the sine@xmath0 point processes establish a smooth crossover between the rigid clock ( or picket fence ) process ( corresponding to @xmath4 ) and the poisson process . |
discrete torsion was introduced many years ago by vafa in @xcite .
more recently douglas @xcite introduced an alternative picture in terms of d - branes .
the purpose of this note is to show the equivalence of these pictures . since both of these papers were very brief as regards to the general case we will review both constructions here . in general
a worldsheet @xmath0 of genus @xmath1 and no boundary will be associated to a given phase dictated by the @xmath2-field and the homology class of the worldsheet in the target space .
vafa wanted to generalize this notion of a @xmath2-field to the case of orbifolds with fixed points . in order to localize the picture
let us consider a discrete group @xmath3 acting on @xmath4 where the origin is fixed by all of @xmath3 .
the worldsheet @xmath0 on this orbifold may be pictured as a disk in the covering @xmath4 where the edges of the disk are identified by elements of the action of @xmath3 .
that is , the usual homology 1-cycles of the worldsheet @xmath0 are twisted by elements of @xmath3 .
the twisting of these cycles in @xmath0 may be viewed as a @xmath3-bundle on @xmath0 .
the holonomy of this bundle is given precisely by the twist associated to a given loop .
since @xmath3 is a discrete group , this bundle is flat .
such a bundle is classified by the homotopy class of a mapping @xmath5 , where @xmath6 is the classifying space for the group @xmath3 .
that is , @xmath7 and @xmath8 for @xmath9 .
in this way , @xmath6 in the case of an orbifold plays a rather similar role to the target space of the usual string theory .
vafa s idea was to define the discrete torsion version of the @xmath2-field as @xmath10 because of this analogy .
the connection between the @xmath2-field and group cohomology has been clarified recently by the work of sharpe @xcite .
now group ( co)homology , which is defined as the ( co)homology of @xmath6 , can be described algebraically directly in terms of @xmath3 as we now show . in order to complete the description of discrete torsion we need to describe the above picture in terms of this more intrinsic definition .
let @xmath11 be the _ group ring _ of @xmath3 .
that is , any element of @xmath11 may be written uniquely as @xmath12 for @xmath13 .
now let @xmath14 be a free resolution of @xmath15 as a @xmath11-module .
this is an exact sequence of @xmath11-modules where @xmath16 is a free @xmath11-module for any @xmath17 .
the @xmath3-action on @xmath15 is taken to be trivial .
now define @xmath18 as the @xmath3-coinvariant projection of @xmath16 .
that is , we divide @xmath16 by the equivalence @xmath19 for any @xmath20 . since @xmath16 is a free @xmath11-module , @xmath18 will be a free @xmath15-module , i.e. , a free abelian group .
the homology of the induced complex @xmath21 is then equal to @xmath22 , the homology of @xmath3 .
see section ii.4 of @xcite , for example , for a proof that this equals @xmath23 .
one way to explicitly compute @xmath22 is via the `` bar resolution '' as follows .
let @xmath16 be generated by @xmath24-tuples of the form @xmath25 where the @xmath3-action is defined as @xmath26 the boundary map @xmath27 is defined as @xmath28 where the hat indicates omission as usual . because of the @xmath3-action , we may use @xmath29 as a basis for @xmath16 or @xmath18 .
the `` bar '' notation is to write @xmath30 = ( 1,g_1,g_1g_2,\ldots , g_1g_2\cdots g_n).\ ] ] we will generally consider the homology of @xmath3 in terms of the generators @xmath31 $ ] .
one may show that a generator may be considered trivial if any of the entries @xmath32 are equal to 1 .
it is also useful to note the explicit form of the boundary map in ( [ eq : free ] ) of a 3-chain in the bar notation : @xmath33 = [ b|c]-[ab|c]+[a|bc]-[a|b ] .
\label{eq:3c}\ ] ] so how is the homology class of @xmath0 in @xmath6 described in terms of these bar chains ?
the free resolution ( [ eq : zgfree ] ) is actually an augmented simplicial chain complex , @xmath34 , as follows .
let the elements of @xmath3 be viewed as the vertices of @xmath34 .
now let @xmath25 be the @xmath17-dimensional simplex in @xmath34 with the corresponding vertices .
thus we have exactly one simplex in @xmath34 for every possible sequence of group elements .
the group @xmath3 acts on @xmath34 to give the complex ( [ eq : free ] ) as a quotient .
as far as homology is concerned , the abstract simplicial complex @xmath35 is a perfectly good representative for @xmath6 .
we may now take a simplicial decomposition of @xmath0 and explicitly map it into this simplicial picture of @xmath6 .
this will relate the homology of @xmath0 , and in particular its fundamental class , to the homology of @xmath3 . as a simple example let us consider a torus as shown in figure [ fig : torus ] .
this torus consists of one cycle twisted by the action of @xmath36 and another cycle twisted by the action of @xmath37 .
the topology of the torus dictates that @xmath38 .
the figure shows the labels of the vertices in terms of the group action and it also shows a simple simplicial decomposition .
we are free to label one vertex as `` 1 '' . in terms of the image of this torus in our peculiar simplicial model of @xmath6
we see that @xmath39 corresponds to a simplex @xmath40 and @xmath41 corresponds to a simplex @xmath42 . as usual in simplicial homology one needs to be careful about relative signs .
the fundamental class of @xmath0 is given by @xmath43 so that the diagonal line in figure [ fig : torus ] cancels for the boundary of @xmath43 .
this gives @xmath44 & = ( 1,a , ab)-(1,b , ab)\\ & = [ a|b]-[b|a ] , \label{eq : g1 } \end{split}\ ] ] in terms of the bar chains .
applying @xmath45 to this bar resolution we may go over to the group cohomology picture .
a group @xmath17-cochain can now be viewed as a map from @xmath46 to @xmath47 .
for example , if @xmath48 then @xmath49 for @xmath50 . from ( [ eq:3c ] )
we see that a 2-cochain is closed if @xmath51 if discrete torsion is written in terms of the @xmath52 cocycles then ( [ eq : g1 ] ) is translated in vafa s language into the statement that the phase associated to a genus one riemann surface is given by @xmath53 this is the result quoted in @xcite .
this method may be applied equally well to higher genus worldsheets .
the genus two case is shown in figure [ fig : g2 ] together with a specific simplicial decomposition .
it follows that @xmath54 = [ a_1|b_1]-[\gamma|b_1a_1]-[b_1|a_1]+[a_2|b_2 ] + [ \gamma|a_2b_2]-[b_2|a_2],\ ] ] where @xmath55 , which by the topology of a genus two surface must equal @xmath56 .
note that if @xmath57 , which would happen if @xmath3 were abelian for example , then the genus two homology class simply breaks into @xmath58-[b_1|a_1])+([a_2|b_2]-[b_2|a_2])$ ] which is the sum of two genus one classes .
this does _ not _ happen in general if @xmath59 .
one may view this statement in terms of a genus two surface degenerating into two genus one surfaces touching at a point .
such a degeneration is topologically obstructed if @xmath59 .
there are many ways of performing a simplicial decomposition of @xmath0 in general .
one may show that changing the decomposition will simply change @xmath60 $ ] in the computations above by the boundary of a group 3-chain and hence has no effect .
in @xcite douglas introduced another definition of discrete torsion associated to projective representations which we now review .
see @xcite for further discussion of this construction . consider the following central extension of @xmath3 : @xmath61 where @xmath62 is a set - theoretic map such that @xmath63 is the identity on @xmath3 .
such extensions are classified by @xmath64 .
the map @xmath62 defines a projective representation of @xmath3 .
given @xmath48 written in terms of the bar resolution of the previous section one may show that @xcite @xmath65 douglas considered @xmath62 as a lift of an orbifold action to the chan - paton factors on the end of an open string .
he then considered a riemann surface with a disk removed .
such a surface can represent either a multi - loop open string diagram or a multi - loop tadpole - like diagram for a closed string .
consider the genus two case shown in figure [ fig:2a ] .
when computing the amplitude of this diagram in the open string context one must include a trace of the group actions on the chan - paton elements along the boundary . as is clear from the figure , this boundary may be contracted to the outer polygon which is a sequence of 1-cycles which are twisted by elements of @xmath3 .
let @xmath66 be this weighting of the amplitude .
we then have @xmath67 where @xmath68 represent the associated twists . for a general genus @xmath1 surface
we clearly have @xmath69 let @xmath70 .
then the topology of the surface dictates that @xmath71 .
note that @xmath72 and thus @xmath73 .
it follows that @xmath74 .
repeated use of ( [ eq : proj ] ) and ( [ eq : coc2 ] ) then gives . the case @xmath75 is left as an easy exercise for the reader . ]
@xmath76 where @xmath77 .
this formula was also derived in @xcite .
now it is not hard to see that the final form of ( [ eq : big ] ) corresponds to a simplicial decomposition of a genus @xmath1 surface into @xmath78 triangles in the language of section [ s : vafa ] .
each @xmath52 factor represents one triangle oriented in just the right way to build up the complete surface .
we show the four triangles for a generic pair @xmath79 in figure [ fig : gen ] .
this factor @xmath80 is exactly the same factor as we obtain by vafa s method of the previous section applied to this simplicial decomposition .
thus we see that vafa s and douglas s definition of discrete torsion agree in general .
it is a pleasure to thank s. katz , r. plesser and e. sharpe for useful conversations .
the author is supported in part by nsf grant dms-0074072 and a research fellowship from the alfred p. sloan foundation . | for a general nonabelian group action and an arbitrary genus worldsheet we show that vafa s old definition of discrete torsion coincides with douglas s d - brane definition of discrete torsion associated to projective representations . 0 |
the lyman @xmath0 forest corresponds to the large number of absorption features located blueward of the lyman @xmath0 emission line in the spectra of quasi - stellar objects ( qsos ) . in the last decade , thanks to the power of 10-m class telescopes equipped with high - resolution spectrographs , considerable progress has been made toward understanding the nature of these absorption features .
this population of discrete lines is due to hi absorption arising from the filamentary structure of the cosmic - web , and in the approximation of a gaussian velocity dispersion can be well described by a series of voigt profiles ( rauch 1998 ) . a variable amount of absorption in the lyman @xmath0 forest is also due to ultraviolet ( uv ) transitions from heavy element ions which are usually associated with strong lyman @xmath0 lines .
quantities often studied in analyses of qso spectra include the power spectrum of the flux distribution , which is used to constrain cosmological parameters and probe the dark matter power spectrum on scales of order of 50 @xmath14 mpc ( e.g. croft et al .
2002 , viel et al .
2004 , mcdonald et al .
another quantity thoroughly studied in the last few years is the flux probability distribution function ( pdf ) , which is sensitive not only to the spatial distribution of dark matter , but also to the thermal state of the intergalactic medium ( igm ) .
previous studies of the lyman @xmath0 forest pdf have shed light on the physical state of the igm mostly at @xmath11 .
bolton et al .
( 2008 ) compared the flux pdf measured from a sample of qso spectra presented by kim et al .
( 2007 , hereafter k07 ) to a set of hydrodynamical simulations of the lyman @xmath0 forest , exploring different cosmological parameters and various thermal histories .
agreement between the data and simulations was obtained by adopting a power - law temperature - density relation , @xmath15 ( hui & gnedin 1997 ) , where the low density igm ( @xmath16 ) was close to isothermal ( i.e. with @xmath17 ) or inverted ( @xmath18 ) .
this result implied that the low - density regions of the igm may be considerably hotter and their thermal state more complicated than usually assumed at these redshifts .
radiative transfer effects during the epoch of heii reionization , which should occur at @xmath19 , are expected to play a non negligible role in setting this temperature - density relation . indeed ,
evidence that the tail - end of heii reionization is occurring at @xmath5 is supported by a variety of observations , such as the opacity of the heii lyman @xmath0 forest ( reimers et al .
1997 ; heap et al .
2000 ; kriss et al .
2001 , shull et al .
2010 , worseck et al .
2011 , syphers et al .
2011 ) , an increase in the temperature of the igm at mean density toward @xmath5 ( schaye et al .
2000 , theuns et al .
2002 , lidz et al .
2010 , becker et al .
2011 , but see mcdonald et al .
2001 , zaldarriaga , hui , & tegmark 2001 ) , the observed evolution of the civ / siiv metal - line ratio ( songaila & cowie 1996 ; songaila 1998 , but see kim , cristiani & dodorico 2002 ) and the evolution of the lyman @xmath0 forest effective optical depth ( bernardi et al .
2003 , faucher - giguere et al .
2008 , pris et al .
2011 , but see bolton , oh & furlanetto 2009b ) .
however , recent work suggests that photo - heating during heii reionization alone is not sufficient to achieve an inverted temperature - density relation , @xmath20 ( mcquinn et al .
2009 ; bolton , oh & furlanetto 2009a )
. a recently proposed alternative channel for heating , which naturally produces an inverted temperature - density relation , may be provided by very energetic gamma rays from blazars ( chang et al 2011 , puchwein et al .
2011 ) . in this scenario ,
very high energy gamma rays annihilate and pair produce on the extragalactic background light .
powerful plasma instabilities then dissipate the kinetic energy of the electron - positron pairs locally , volumetrically heating the igm . on the other hand
, continuum placement errors could partly explain the preference of the k07 data for an inverted temperature - density relation ( see lee 2012 for a recent discussion ) .
indeed , bolton et al . ( 2008 ) and more recently viel et al .
( 2009 ) found an isothermal temperature - density relation was consistent ( within @xmath21 ) with the k07 pdf constraints when treating continuum uncertainties conservatively .
ultimately , however , for further progress in understanding the origin of an inverted temperature - density relation , additional observational constraints on the igm thermal state are required .
an investigation of the flux probability distribution at redshifts around @xmath22 or higher may therefore shed more light on the possible inversion of the temperature - density relation .
such a measurement is also complementary to other constraints on the igm thermal state , such as those presented recently by becker et al .
( 2011 ) , where the temperature - density relation is not measured directly .
such an investigation is the main motivation of the present work . in this paper ,
we use eight high resolution qso spectra observed at @xmath23 to extend the study of the lyman @xmath0 forest pdf by k07 to higher redshifts .
the aim of this paper is two - fold .
first , we perform a detailed study of the redshift evolution of the flux pdf across a wide redshift range .
this enables us to study possible systematic trends in the pdf due to redshift - dependent effects , such as the presence of metal absorption lines . in order to corroborate our analysis
, we also measure the lyman @xmath0 effective optical depth and compare it to different estimates by other authors . in the second step , the measured flux pdf is compared to theoretical results obtained from cosmological hydrodynamical simulations .
this enables us to investigate the thermal state of the igm during this interesting redshift range .
this paper is organized as follows . in section 2
, we present the qso spectra and the data analysis , and in section 3 we present our measurements of the flux pdf and effective optical depth . in section 4 ,
we introduce the hydrodynamical simulations used to obtain constraints on the thermal state of the igm from the pdf and , in section 5 , we report the results from the analysis of the simulations .
finally , in section 6 , we present our conclusions .
the data set used in this paper consists of eight high - redshift qso spectra .
two of the spectra ( pks2126 - 158 and pks2000 - 330 ) were obtained with the ultraviolet and visual echelle spectrograph ( uves ) ( dekker et al .
2000 ) at the kueyen unit of the european southern observatory ( eso ) vlt ( cerro paranal , chile ) as part of the eso large programme ( lp ) : ` the cosmic evolution of the igm ' ( bergeron et al .
pks1937 - 101 is from the program carswell et al .
077.a-0166(a ) , q1317 - 0507 , sdssj16212 - 0042 and q1249 - 0159 are from kim et al .
075.a-0464(a ) , q1209 + 0919 is from the program dessauges - zavadsky et al .
073.b-0787(a ) and q0055 - 269 is from dodorico et al .
65.o-0296(a ) .
all the spectra have been reduced with the standard uves pipeline provided by eso .
wavelengths have been corrected to vacuum - heliocentric . [ cols= " < ,
< , < , < , < , < " , ] we next use the synthetic spectra to construct flux pdfs which may be compared to our observational measurements .
we explore the effect of varying the following cosmological and astrophysical parameters on the pdf : @xmath24 , @xmath25 , @xmath26 and @xmath27 for the cosmological parameters , and @xmath28 and @xmath29 for the igm thermal history .
note , however , that most of the cosmological parameters are only weakly constrained by the flux pdf .
the largest effect is induced by using a different value of @xmath30 ( e.g. bolton et al . 2008 ) .
we will therefore present marginalized cosmological constraints for the amplitude of the matter power spectrum only .
the astrophysical parameters @xmath28 and @xmath29 are parameterized as a broken power - law around @xmath6 , in order to investigate possible changes in the thermal state around heii reionization : @xmath31^{t_{0}^{\rm s}(z)}$ ] and @xmath32^{\gamma^s(z)}$ ] ( viel et al .
the effective optical depth evolution is parameterized @xmath33^{\tau_{\rm eff}^{\rm s}(z)}$ ] . in order to explore our parameter space efficiently with our limited number of simulations , we taylor
expand the flux pdf and compute derivatives to second order using two simulations ( plus the reference run ) for each parameter .
this closely follows the procedure described in viel & haehnelt ( 2006 ) , where this taylor expansion method was used in order to explore constraints derived from the sloan digital sky survey flux power spectrum ( mcdonald et al .
we start with the a1@xmath34 simulation and proceed to calculate the @xmath35 of our models by varying all the parameters that alter the flux pdf , enabling us to expand around the fiducial starting model .
the expansion method allows us to explore the parameter space close to the reference model with an accurate set of hydrodynamic simulations , even if the full parameter space can not be probed in this way with the same high accuracy .
we impose weak priors on the effective optical depth evolution , but both its amplitude and slope are constrained strongly by the data , so the priors do not affect our results . in order to present a conservative analysis , we also decide to consider only the values of the flux in the range @xmath36 $ ] , as this range should not be affected much by the errors made in the continuum fitting procedure ( e.g. k07 ; bolton et al .
2008 , and see fig .
[ contpdf ] ) . furthermore , when estimating the covariance between the pdf data points we use the covariance extracted from the simulations for the non - diagonal terms , which is less noisy ( e.g. lidz et al .
2006 ) .
the one dimensional marginalized constraints are shown in figure [ fig12 ] : the likelihoods obtained in the present work are represented by the black curves , while the results derived from @xmath11 k07 data set analysed by viel et al .
( 2009 ) are shown as dashed blue lines .
the joint constraints for both pdf measurements are plotted as the red lines . from this figure
we can see that overall there is agreement between the inferred parameter values for the high redshift ( this work ) and low redshift ( k07 ) measurements .
the @xmath37 parameter constraints from the sample of qsos analysed in this work and for the joint sample including the k07 data are summarized in tab .
the minimum @xmath35 for the best fitting model to the pdf measured in this work is 52 for 38 degrees of freedom ( d.o.f . ) which is likely to happen 7 per cent of the time .
in contrast , for the joint sample the minimum @xmath35 is 94 for 83 d.o.f .
with has a 21 per cent probability .
the probability is calculated using the full covariance matrix for all the parameters recovered .
note that the `` degrees of freedom '' value does not have a precise meaning if the error bars are heavily correlated .
however , this happens not to be the case in the flux ranges considered here , @xmath38@xmath39 $ ] , where correlations are weak .
we have also checked that there is good agreement between the derived constraints from the present data set when compared to constraints obtained by splitting the sample in two and three redshift bins .
finally , in figure [ fig_evol ] we display the redshift evolution of the temperature at the mean density , @xmath40 , and the slope of the temperature - density relation , @xmath29 , inferred from our analysis .
we use the broken power - law parameterization around @xmath6 introduced previously : @xmath41^{f^{\rm s}(z)}$ ] , where @xmath42 or @xmath29 , @xmath43 is the corresponding normalization factor at @xmath6 and @xmath44 is the slope of the relation ( see tab . 5 ) .
the black solid lines in figure [ fig_evol ] are the results obtained from this work ( @xmath19 ) while the shaded blue regions show their associated errors .
the red dot - dashed lines are the results obtained from the joint sample composed of this work and the k07 data .
the red dashed lines show the associated errors .
since the @xmath43 and @xmath45 parameters are correlated and the @xmath44 are in general poorly constrained , we define the error associated to @xmath46 , at a given redshift @xmath47 , as follows : @xmath48 where @xmath49 and @xmath50 are , respectively , the @xmath37 upper and lower limits on @xmath43 , and @xmath51 are the best fit values of the slope showed in tab .
the temperature at mean density and its redshift evolution , shown in the upper panel of figure [ fig_evol ] , are not strongly constrained by the flux pdf , although our measurements are consistent with other recent analyses presented in the literature which hint that heii reionization may be completing around @xmath52 ( lidz et al .
2010 , becker et al .
however , the pdf constraints alone are not precise enough to draw conclusions about the timing and extent of heii reionization ; the flux pdf is a relatively insensitive to @xmath40 ( see e.g fig .
[ fig_pdf_sim ] ) .
alternative measurements from doppler widths ( schaye et al 2000 ) wavelets ( lidz et al .
2010 ) or the curvature statistic ( becker et al .
2011 ) are typically better suited for this measurement . on the other hand ,
the flux pdf is sensitive to the slope of the temperature - density relation , @xmath29 .
the constraints displayed in fig .
[ fig_evol ] , based on the pdf measurements presented in this work , are formally consistent with an isothermal temperature - density relation at @xmath6 ( i.e. @xmath53 ) , although the uncertainties remain large .
a temperature - density relation which is close to isothermal may be expected if heii reionization was underway by @xmath6 ( e.g. schaye et al . 2000 , theuns et al .
when combining our data with the k07 measurements at lower redshift , however , the inferred @xmath29 at @xmath6 is shifted to lower , inverted @xmath54 values , consistent with previous results based on the k07 data ( e.g. bolton et al .
2008 , viel et al .
this is because the k07 data is more constraining at @xmath55 due to the larger number of qsos in this sample , and for the simple power - law parameterisation we have assumed this also pushes our joint constraints to smaller @xmath29 values at @xmath56 .
it has been suggested recently that an inverted temperature - density relation at may be explained by an igm which is volumetrically heated by tev emission from blazars ( e.g. chang et al . 2011 , puchwein et al 2011 ) .
this is in contrast to conventional photo - ionization heating models which predict a temperature - density relation which evolves from isothermal to @xmath57 following reionization ( hui & gnedin 1997 ) .
we caution , however , the uncertainties on the flux pdf measurements are still large ; the temperature - density relation suggested by the radiative transfer simulations of mcquinn et al . ( 2011 ) , @xmath58 at @xmath6 , is compatible at the @xmath21 level with the value derived from the sample considered in this work alone .
the k07 data alone are furthermore consistent with an isothermal igm , again to within @xmath21 ( viel et al .
we conclude that tighter constraints on @xmath29 from larger data sets , as well as a comparison of different measurement techniques to aid with the identification possible systematics , will be required for distinguishing between different igm heating scenarios .
in this paper , we have measured the lyman @xmath0 forest flux pdf from a set of eight qso spectra to investigate the thermal state of the igm at @xmath5 .
the absorption lines in the lyman @xmath0 forest have been fitted with voigt profiles , and absorption lines arising from heavy elements have been carefully identified .
the spectra have then subsequently had the metal absorption lines removed from within the lyman @xmath0 forest .
the resulting masked spectra have been used to study the redshift evolution of the lyman @xmath0 forest flux distribution and the evolution of the hi effective optical depth .
our results can be summarized as follows .
* by removing metals from the lyman @xmath0 forest , we attempt to eliminate pixels which will introduce a systematic uncertainty into our measurement of the flux pdf .
the effect of removing these metals on the pdf is to decrease the number of pixels in bins of low flux and increase the number of pixels in bins of high flux . as a result ,
the pdf measured after correcting for metal absorption is slightly steeper than the pdf measured without removing metals . in general , the effect of metal removal on the pdf is strongest for flux bins at @xmath59 , in agreement with previous results for other qso samples at lower redshift ( k07 ) . +
* we measure the pdf for our full sample in two redshift bins , both before and after metal decontamination , and find the effect of metal absorption on the lyman @xmath0 forest flux pdf is redshift dependent .
it is more significant at lower redshift , where the average opacity of the lyman @xmath0 forest is lower and there are fewer absorption lines .
this redshift dependence may be understood in terms of the larger number of metal lines which are blended with lyman @xmath0 absorption lines toward higher redshift , resulting in the identification of fewer isolated metal absorption lines . on the other hand ,
it is also possible that the metallicity of the intergalactic medium decreases with increasing lookback time . + * we present measurements of the effective optical depth , @xmath2 , from our qso sample , which provides a probe of the igm which is complementary to the pdf .
the values of @xmath2 measured from the qsos in our sample can be fitted with a power - law @xmath60 , in agreement with previous measurements by k07 after correcting for metal absorption .
we also find no substantial evidence for a deviation from a power law in the @xmath61 plane at @xmath62 .
+ there is a weak redshift dependence of the ratio of the hi effective optical depth measured before and after removing identified metal absorption lines .
the ratio increases toward lower redshift , and is consistent with our previous finding from the pdf that metal absorption plays a more significant role in the lyman @xmath0 forest toward lower redshift . + * we perform a set of high resolution hydrodynamical simulations with a range of cosmological and astrophysical parameters and use them to extract mock qso spectra with properties that closely resemble those of the observed spectra .
we compute the lyman @xmath0 flux pdf and compare it to the observed pdf in the range @xmath38@xmath39 $ ] via a taylor expansion at second order to find deviations from a best - guess case ( e.g. viel & haehnelt 2006 ) . by implementing this method within a monte carlo markov chain likelihood estimator , we present constraints for the igm thermal state . at @xmath6 ,
our pdf measurements are consistent with an isothermal temperature - density relation , @xmath9 , and a temperature at the mean density of @xmath63 k ( 1@xmath10 uncertainties ) .
+ * we also obtain joint constraints on the igm thermal state by analysing the lower redshift pdf measurements of k07 , which is a significantly larger data set than the one considered here . at @xmath6 , we find the joint data set is consistent with an inverted temperature - density relation , @xmath64 , with @xmath65 k ( 1@xmath10 uncertainties ) . + * in conclusion , the present work highlights the role of the flux pdf as a competitive and quantitative thermometer which may be used to constraint the igm temperature - density relation over a wide redshift range .
the constraints derived from the present work are complementary with respect to other approaches , such as the wavelets method ( lidz et al .
2010 ) , the cutoff in the doppler width - column density plane ( schaye et al .
2000 ) or the curvature ( becker et al . 2011 ) , which are primarily sensitive to higher density regions of the igm at @xmath5 . in the near future ,
the most promising route to making further progress on our knowledge of the thermal state of the igm will be to carefully compare and contrast results from these different probes of the igm temperature - density relation .
fc wishes to thank eros vanzella for several interesting discussions , cristian vignali for useful suggestions and jochen liske for support in the use of vpguess .
et acknowledges a fellowship from the european commission s framework programme 7 , through the marie curie initial training network cosmocomp pitn - ga-2009 - 238356 .
mv acknowledges support from infn / pd-51 , asi - aae grant a prin inaf 2009 , a prin miur and the european erc - starting grant `` cosmoigm '' .
jsb acknowledges the support of an arc australian postdoctoral fellowship ( dp0984947 ) .
the hydrodynamical simulations used in this work were performed using the darwin supercomputer of the university of cambridge high performance computing service ( http://www.hpc.cam.ac.uk/ ) , provided by dell inc . using strategic research infrastructure funding from the higher education funding council for england .
postprocessing of the simulations was performed at cineca ( italy ) and at the cosmos supercomputer in cambridge ( uk ) .
99 becker g. d. , bolton j. s. , haehnelt m. g. , sargent w. l. w. , 2011 , mnras , 410 , 1096 bergeron j. , et al .
, 2004 , msngr , 118 , 40 bernardi m. , et al . , 2003 , aj , 125 , 32 bolton j. s. , viel m. , kim t .- s . , haehnelt m. g. , carswell r. f. , 2008 , mnras , 386 , 1131 bolton j. s. , oh s. p. and furlanetto s. r , 2009a , mnras , 395 , 736 bolton j. s. , oh s. p. and furlanetto s. r , 2009b , mnras , 396 , 2405 haardt f. , madau p. , 2001
, cghr.conf , xxxvith rencontres de moriond , xxist moriond astrophysics meeting , march 10 - 17 , 2001 savoie , france .
edited by d.m .
neumann & j.t.t .
heap s. r. , williger g. m. , smette a. , hubeny i. , sahu m. s. , jenkins e. b. , tripp t. m. , winkler j. n. , 2000 , apj , 534 , 69 hui l. , gnedin n. y. , 1997 , mnras , 292 , 27 jarosik n. , et al . , 2011 ,
apjs , 192 , 14 kim t .- s .
, bolton j. s. , viel m. , haehnelt m. g. , carswell r. f. , 2007 , mnras , 382 , 1657 ( k07 ) kim t .- s . , cristiani s. , dodorico s. , 2002 , a&a , 383 , 747 komatsu e. , et al . , 2009 ,
apjs , 180 , 330 kriss g. a. , et al . , 2001 , sci , 293 , 1112 lidz a. , heitmann k. , hui l. , habib s. , rauch m. , sargent w. l. w. , 2006 , apj , 638 , , a. , faucher - gigure , c .- a . , dallaglio , a. , mcquinn , m. , fechner , c. , zaldarriaga , m. , hernquist , l. , & dutta , s. 2010 , apj , 718 , 19927 lee k .-
, 2012 , apj , submitted , ( arxiv:1103.2780 ) mcdonald p. , miralda - escud j. , rauch m. , sargent w. l. w. , barlow t. a. , cen r. , ostriker j. p. , 2000 ,
apj , 543 , 1 mcdonald p. , miralda - escud j. , rauch m. , sargent w. l. w. , barlow t. a. , cen r. , 2001 , apj , 562 , 52 mcdonald p. , et al .
, 2006 , apjs , 163 , 80 puchwein e. , pfrommer c. , springel v. , broderick a. e. , 2011 , mnras submitted , ( arxiv:1107.3837 ) rauch m. , 1998 , ara&a , 36 , 267 reichardt c. l. , et al . , 2009 , apj , 694 , 1200 reimers d. , kohler s. , wisotzki l. , groote d. , rodriguez - pascual p. , wamsteker w. , 1997 , a&a , 327 , 890 schaye j. , theuns t. , rauch m. , efstathiou g. , sargent w. l. w. , 2000 , mnras , 318 , 817 schaye j. , aguirre a. , kim t .- s . , theuns t. , rauch m. , sargent w. l. w. , 2003 , apj , 596 , 768 , j. m. , france , k. , danforth , c. w. , smith , b. , & tumlinson , j. 2010 , apj , 722 , 1312 tescari e. , viel m. , dodorico v. , cristiani s. , calura f. , borgani s. , tornatore l. , 2011 , mnras , 411 , 826 theuns t. , leonard a. , efstathiou g. , pearce f. r. , thomas p. a. , 1998 , mnras , 301 , 478 theuns t. , bernardi m. , frieman j. , hewett p. , schaye j. , sheth r. k. , subbarao m. , 2002 , apj , 574 , l111 viel m. , haehnelt m. g. , 2006 , mnras , 365 , 231 viel m. , bolton j. s. , haehnelt m. g. , 2009 , mnras , 399 , l39 | we present a measurement of the lyman @xmath0 flux probability distribution function ( pdf ) obtained from a set of eight high resolution quasar spectra with emission redshifts at @xmath1 .
we carefully study the effect of metal absorption lines on the shape of the pdf .
metals have a larger impact on the pdf measurements at lower redshift , where there are relatively fewer lyman @xmath0 absorption lines .
this may be explained by an increase in the number of metal lines which are blended with lyman @xmath0 absorption lines toward higher redshift , but may also be due to the presence of fewer metals in the intergalactic medium at earlier times .
we also provide a new measurement of the redshift evolution of the effective optical depth , @xmath2 , at @xmath3 , and find no evidence for a deviation from a power law evolution in the @xmath4 plane .
the flux pdf measurements are furthermore of interest for studies of the thermal state of the intergalactic medium ( igm ) at @xmath5 . by comparing the pdf to state - of - the - art cosmological hydrodynamical simulations , we place constraints on the temperature of the igm and
compare our results with previous measurements of the pdf at lower redshift . at redshift @xmath6 ,
our new pdf measurements are consistent with an isothermal temperature - density relation , @xmath7 , with a temperature at the mean density of @xmath8 k and a slope @xmath9(1@xmath10 uncertainties ) . in comparison ,
joint constraints with existing lower redshift pdf measurements at @xmath11 favour an inverted temperature - density relation with @xmath12 k and @xmath13 , in broad agreement with previous analyses .
= = = = = = = = # 1 # 1 # 1 # 1 @mathgroup@group @mathgroup@normal@groupeurmn @mathgroup@bold@groupeurbn @mathgroup@group @mathgroup@normal@groupmsamn @mathgroup@bold@groupmsamn = `` 019 = ' ' 016 = `` 040 = ' ' 336 = " 33e = = = = = = = = # 1 # 1 # 1 # 1 = = = = = = = = -0.25 in [ firstpage ] cosmology : intergalactic medium - methods : numerical - quasars : absorption lines . |
recall the famous property r theorem , proven in a somewhat stronger form by david gabai @xcite : [ thm : propr ] if @xmath2-framed surgery on a knot @xmath3 yields @xmath4 then @xmath5 is the unknot .
problem 1.82 in kirby s problem list @xcite conjectures a generalization to links : if surgery on an @xmath6-component link @xmath7 yields the connected sum @xmath8 then @xmath7 becomes the unlink after suitable handle slides .
( see the next section for further discussion . )
although this generalized property r conjecture can be traced back more than two decades , it seems that no progress has appeared in the literature until now .
the present paper studies the conjecture , focusing on the case of 2-component links .
we conclude that the conjecture is probably false , and analyze potential counterexamples .
these examples have relevance to the slice - ribbon conjecture .
we also propose a weaker version of the generalized property r conjecture that seems much more likely , and is related to the 4-dimensional smooth poincar conjecture .
we concentrate on links @xmath9 where @xmath10 is fibered .
critical to the argument are developments in sutured manifold theory , developments that go beyond those used by gabai in his original proof of property r. in particular we use sutured manifold theory to show ( theorem [ thm : main ] ) that any counterexample of this form to the generalized property r conjecture generates another 2-component counterexample for which one component has smaller genus than @xmath10 . with this motivation , we assume @xmath10 has genus 2 .
we then use heegaard theory to completely characterize such links @xmath7 satisfying the hypothesis of the generalized property r conjecture ( corollary [ cor : genustwo ] ) : after handle slides , @xmath11 is a special type of curve lying on the fiber of @xmath10 . for the simplest plausible @xmath10 , namely the square knot , this characterization becomes simple enough ( corollary [ cor : enumerate ] ) to allow enumeration of all examples up to handle slides .
@xmath12 at 95 200 @xmath12 at 338 203 @xmath13 at 25 112 @xmath14 at 410 115 we then use classical 4-manifold theory ( the 20-year old paper @xcite of the first author ) to argue that the square knot probably does lie in counterexamples to the generalized property r conjecture .
the link @xmath15 in figure [ fig : ln1 ] consists of a square knot interleaved with the connected sum of an @xmath16 torus knot with its mirror .
( each summand is depicted as an @xmath6-stranded spiral , with a full @xmath17-twist added relative to the plane of the paper .
the dotted arc connecting the components will be used momentarily . )
we show in section [ sect : nonstandard ] that @xmath15 is a counterexample to the generalized property r conjecture provided that the presentation @xmath18 of the trivial group is andrews - curtis nontrivial , as is deemed likely by group theorists when @xmath19 .
we will see ( section [ slice ] ) that this link and its generalizations @xmath20 ( figure [ fig : lnk ] ) are necessarily slice , so we can generate slice knots just by band - summing the two components together . for @xmath19 there is no apparent reason for the resulting knots to be ribbon .
thus , we obtain potential counterexamples to the slice - ribbon conjecture .
this appears to be the only known method for constructing such examples .
figure [ fig : sliceknot ] shows a simple potential counterexample obtained from @xmath21 by the band - move following the dotted arc in figure [ fig : ln1 ] .
@xmath13 at 22 112 @xmath14 at 407 115 the 4-dimensional viewpoint that discredits the generalized property r conjecture also shows how to rescue it by allowing more moves . in section [
sect : weakening ] we discuss this more plausible version of the conjecture , and its close connection with the 4-dimensional smooth poincar conjecture . since the key new move is somewhat mysterious , section [ sect : hopf ] shows explicitly how it works to trivialize our potential counterexamples @xmath20 .
the main result , proposition [ prop : ln ] , can in principle be used to construct other potential counterexamples . unless explicitly stated otherwise
, all manifolds throughout the paper will be smooth , compact and orientable .
to make sense of how property r might be generalized , recall a small bit of @xmath22-manifold handlebody theory @xcite .
suppose @xmath7 is a link in a @xmath23-manifold @xmath24 and each component of @xmath7 is assigned a framing , that is a preferred choice of cross section to the normal circle bundle of the component in @xmath24 .
for example , if @xmath25 , a framing on a knot is determined by a single integer , the algebraic intersection of the preferred cross - section with the longitude of the knot .
( in an arbitrary @xmath23-manifold @xmath24 a knot may not have a naturally defined longitude . ) surgery on the link via the framing is standard dehn surgery , though restricted to integral coefficients : a regular neighborhood of each component is removed and then reattached so that the meridian is identified with the cross - section given by the framing .
associated to this process is a certain @xmath22-manifold : attach @xmath22-dimensional @xmath0-handles to @xmath26 along @xmath27 , using the given framing of the link components .
the result is a @xmath22-dimensional cobordism , called the _ trace _ of the surgery , between @xmath24 and the @xmath23-manifold @xmath28 obtained by surgery on @xmath7 .
the collection of belt spheres of the @xmath0-handles constitute a link @xmath29 called the dual link ; the trace of the surgery on @xmath30 can also be viewed as the trace of a surgery on @xmath29 .
the @xmath22-manifold trace of the surgery on @xmath7 is unchanged if one @xmath0-handle is slid over another @xmath0-handle .
such a handle slide is one of several moves allowed in the kirby calculus @xcite . when the @xmath0-handle corresponding to the framed component @xmath10 of @xmath7 is slid over the framed component @xmath11 of @xmath7 the effect on the link is to replace @xmath10 by the band sum @xmath31 of @xmath10 with a certain copy of @xmath11 , namely the copy given by the preferred cross - section realizing the framing of @xmath11 .
if @xmath25 and the framings of @xmath10 and @xmath11 are given by the integers @xmath32 and @xmath6 , respectively , then the integer for @xmath31 will be @xmath33 where @xmath34 denotes the linking number and the sign is @xmath35 precisely when the orientations of @xmath10 and the copy of @xmath11 fit together to orient @xmath31 .
( this requires us to orient @xmath10 and @xmath11 , but the answer is easily seen to be independent of the choice of orientations . ) the terms _
handle addition _ and _ subtraction _
are used to distinguish the two cases of sign ( @xmath35 and @xmath36 , respectively ) in the formula .
any statement about obtaining @xmath23-manifolds by surgery on a link will have to take account of this move , which we continue to call a handle - slide , in deference to its role in @xmath22-dimensional handle theory .
suppose @xmath37 is obtained from components @xmath10 and @xmath11 by the handle - slide of @xmath10 over @xmath11 as described above .
let @xmath38 and @xmath39 be the dual knots to @xmath10 and @xmath11 .
it will be useful to note this counterintuitive but elementary lemma : [ lemma : dual ] the link in @xmath28 that is dual to @xmath40 is @xmath41 , where @xmath42 is obtained by a handle - slide of @xmath43 over @xmath44 .
it suffices to check this for the simple case in which the @xmath23-manifold is a genus @xmath0 handlebody , namely a regular neighborhood of @xmath10 , @xmath11 , and the arc between them along which the band - sum is done .
a sketch of this is shown in figure [ fig : dual2 ] .
the dual knots @xmath45 , @xmath43 and @xmath46 are displayed as boundaries of meridian disks for regular neighborhoods of @xmath10 , @xmath11 and @xmath47 respectively . at 390 543 at 100 543 at 171 512 at 465 500 @xmath48 at 90 435 @xmath49 at 380 435 at 245 545 at 225 435 alternatively
, a @xmath0-dimensional schematic of the @xmath22-dimensional process is shown in figure [ fig : dual ] .
the handle corresponding to @xmath10 is shown half - way slid across the handle corresponding to @xmath11 .
each disk in the figure is given the same label as its boundary knot in @xmath24 or @xmath28 as appropriate . at 37 183 at 137 86 at 346 168 at 238 122 @xmath50 at 162 238 at 310 50 at 396 101
let @xmath51 denote the connected sum of @xmath6 copies of @xmath4 .
the generalized property r conjecture ( see ( * ? ? ?
* problem 1.82 ) ) says this : [ conj : genr ] suppose @xmath7 is an integrally framed link of @xmath52 components in @xmath53 , and surgery on @xmath7 via the specified framing yields @xmath51 .
then there is a sequence of handle slides on @xmath7 that converts @xmath7 into a @xmath2-framed unlink . in the case
@xmath54 no slides are possible , so conjecture [ conj : genr ] does indeed directly generalize theorem [ thm : propr ] . on the other hand , for @xmath55
it is certainly necessary to include the possibility of handle slides .
figure [ fig : squareknot ] shows an example of a more complicated link on which @xmath2-framed surgery creates @xmath1 . to see this , note that the kirby move shown , band - summing the square knot component to a copy of the unknotted component , changes the original link to the unlink of two components , for which we know surgery yields @xmath1 .
even more complicated links with this property can be obtained , simply by using kirby moves that complicate the link rather than simplify it .
see figure [ fig : squareknot2b ] ; the free ends of the band shown can be connected in an arbitrarily linked or knotted way .
@xmath2 at 72 89 @xmath2 at 137 120 @xmath2 at 281 18 @xmath2 at 338 120 @xmath2 at 432 18 @xmath2 at 432 120 at 108 -11 the conjecture can be clarified somewhat by observing that the only framed links that are relevant are those in which all framings and linking numbers are trivial .
there is a straightforward @xmath22-dimensional proof , using the intersection pairing on the trace of the surgery .
here is an equally elementary @xmath23-dimensional proof : [ prop : frame ] suppose @xmath7 is a framed link of @xmath52 components in @xmath53 , and surgery on @xmath7 via the specified framing yields @xmath51 .
then the components of @xmath7 are algebraically unlinked and the framing on each component is the @xmath2-framing .
it follows immediately from alexander duality that @xmath56 in particular , filling in the solid tori via whatever framing we are given yields an epimorphism , hence an isomorphism @xmath57 . for each torus component @xmath58 of @xmath59 ,
the filling will kill some generator of @xmath60 , so the homomorphism @xmath61 is not injective .
it follows that the homomorphism @xmath62 can not be injective and , moreover , @xmath63 must contain the framing curve .
but @xmath63 must be contained in the subgroup generated by the standard longitude , since this is the only subgroup that is trivial when the other components of @xmath64 are just left in place and not removed from @xmath53 .
it follows that the framing at each component is that of the standard longitude , namely the @xmath2-framing .
since the longitude of each @xmath58 is null - homologous in @xmath65 it follows that all linking numbers are trivial .
there is also an immediate topological restriction on the link itself , which carries over to a restriction on the knots that can appear as individual components of such a link ( see ( * ? ? ?
* theorem 2 ) ) : [ prop : slice ] suppose @xmath7 is a framed link of @xmath52 components in @xmath53 , and surgery on @xmath7 via the specified framing yields @xmath51
. then @xmath7 bounds a collection of @xmath6 smooth @xmath0-disks in a @xmath22-dimensional homotopy ball bounded by @xmath53 .
it follows from freedman s proof of the @xmath22-dimensional topological poincar conjecture @xcite that @xmath7 ( and so each component of @xmath7 ) is topologically slice in @xmath66 .
see also the earlier @xcite for the case @xmath54 .
consider the @xmath22-manifold trace @xmath67 of the surgery on @xmath7 .
@xmath68 has one end diffeomorphic to @xmath53 and the other end , call it @xmath69 , diffeomorphic to @xmath51 .
@xmath67 has the homotopy type of a once - punctured @xmath70 .
attach @xmath71 to @xmath69 via the natural identification @xmath72 .
the result is a homotopy @xmath22-ball , and the cores of the original @xmath6 @xmath0-handles that are attached to @xmath7 are the required @xmath6 @xmath0-disks . somewhat playfully , we can turn the generalized property r conjecture , which is a conjecture about links , into a conjecture about knots , and also stratify it by the number of components , via the following definition and conjecture .
a knot @xmath3 has * property nr * if it does not appear among the components of any @xmath6-component counterexamples to the generalized property r conjecture .
that is , whenever @xmath5 is a component of an n - component link @xmath73 and some integral surgery on @xmath7 produces @xmath51 , then there is a sequence of handle slides on @xmath7 that converts @xmath7 into a @xmath2-framed unlink .
all knots have property nr .
thus the generalized property r conjecture is that the property nr conjecture is true for all @xmath52 .
following proposition [ prop : slice ] any non - slice knot has property nr for all @xmath6 .
the first thing that we will show ( theorem [ thm : main ] ) is that if there are any counterexamples to property 2r , a least genus such example can not be fibered .
we already know that both of the genus one fibered knots ( the figure 8 knot and the trefoil ) can not be counterexamples , since they are not slice , and we show below ( proposition [ prop : unknot ] ) that the unknot can not be a counterexample .
so these simplest of fibered knots do have property 2r .
these considerations suggest that , if one is searching for a potential counterexample to property nr and hope to exploit a knot s geometric properties , the simplest candidate could be the square knot , viewed as a potential counterexample to property 2r .
after all , the square knot is symmetric , slice , fibered , and both its genus and its crossing number are low .
in fact it turns out that there is a relatively simple range of possibilities for what a @xmath0-component link containing the square knot and surgering to @xmath1 can look like .
the background for this comes from the theory of heegaard splittings and is discussed in sections [ sect : heeg ] and [ sect : squareknot ] . in section [
sect : nonstandard ] we present strong evidence that property 2r does indeed fail for the square knot .
since the square knot is fibered , it would then follow from theorem [ thm : main ] that there is a counterexample to property 2r among genus one knots .
on the other hand , the square knot may well satisfy a somewhat weaker property , which is still useful in the connection of property r to such outstanding @xmath22-manifold questions as the smooth poincar conjecture and the smooth schoenflies conjecture .
for this weaker property , copies of canceling hopf pairs may be added to the link before the handle slides and then removed after the handle slides . beyond our specific example
, we discuss the mechanics of how addition and later removal of a hopf pair can be used to reduce the complexity of a surgery description .
almost nothing is known about generalized property r , beyond the elementary facts noted in propositions [ prop : frame ] and [ prop : slice ] that the framing and linking of the components of the link are all trivial and the links themselves are slice in a possibly exotic ball . a bit more is known about property 2r .
the first proposition was shown to us by alan reid : suppose @xmath73 is a @xmath0-component link with tunnel number @xmath74 .
if surgery on @xmath7 gives @xmath1 then @xmath7 is the unlink of two components .
the assumption that @xmath7 has tunnel number @xmath74 means that there is a properly embedded arc @xmath75 so that @xmath76 is a genus @xmath0 handlebody @xmath77 .
let @xmath78 .
there is an obvious epimorphism @xmath79 ( fill in a meridian disk of @xmath80 ) and an obvious epimorphism @xmath81 ( fill in solid tori via the given framing ) .
but any epimorphism @xmath82 is an isomorphism , since free groups are hopfian , so in fact @xmath83 .
it is then a classical result that @xmath7 must be the unlink of two components .
this first step towards the property 2r conjecture is a bit disappointing , however , since handle - slides ( the new and necessary ingredient for generalized property r ) do not arise .
in contrast , figure [ fig : squareknot ] shows that handle slides are needed in the proof of the following : [ prop : unknot ] the unknot has property 2r .
suppose @xmath7 is the union of two components , the unknot @xmath10 and another knot @xmath11 , and suppose some surgery on @xmath7 gives @xmath1 .
following proposition [ prop : frame ] the surgery is via the @xmath2-framing on each and , since @xmath10 is the unknot , @xmath2-framed surgery on @xmath10 alone creates @xmath4 .
moreover , the curve @xmath84 that is dual to @xmath10 is simply @xmath85 for some point @xmath86 .
a first possibility is that @xmath11 is a satellite knot in @xmath4 , so @xmath11 lies in a solid torus @xmath5 in such a way that the torus @xmath87 is essential in @xmath88 .
since there is no essential torus in @xmath1 , @xmath87 compresses after the surgery on @xmath11 . since @xmath1 contains no summand with finite non - trivial homology , it follows from the main theorem of @xcite that @xmath11 is a braid in @xmath5 and that surgery on @xmath11 has the same effect on @xmath4 as some surgery on @xmath5 .
proposition [ prop : frame ] shows that the surgery on @xmath5 must be along a longitude of @xmath5 , but that would imply that @xmath11 has winding number @xmath74 in @xmath5 .
the only braid in a solid torus with winding number @xmath74 is the core of the solid torus , so in fact @xmath11 is merely a core of @xmath5 and no satellite .
so we conclude that @xmath89 can not be a satellite knot .
consider the manifold @xmath88 .
if it is irreducible , then it is a taut sutured manifold ( see , for example , @xcite ) and two different fillings ( trivial vs. @xmath2-framed ) along @xmath90 yield reducible , hence non - taut sutured manifolds .
this contradicts @xcite .
we conclude that @xmath88 is reducible .
it follows that @xmath11 is isotopic to a knot @xmath91 lying in a @xmath23-ball in @xmath92 and that surgery on @xmath93 creates a summand of the form @xmath4 . by property r , we know that @xmath91 is the unknot in @xmath94 .
hence @xmath95 is the unlink of two components .
the proof , though , is not yet complete , because the isotopy of @xmath11 to @xmath91 in @xmath4 may pass through @xmath44 .
but passing @xmath11 through @xmath84 can be viewed as band - summing @xmath11 to the boundary of a meridian disk of @xmath44 in @xmath4 .
so the effect back in @xmath53 is to replace @xmath11 with the band sum of @xmath11 with a longitude of @xmath10 . in other words ,
the knot @xmath91 , when viewed in @xmath53 , is obtained by from @xmath11 by a series of handle slides over @xmath10 , a move that is permitted under generalized property r. in a similar spirit , the first goal of the present paper is to prove a modest generalization of proposition [ prop : unknot ] .
a pleasant feature is that , since the square knot is fibered , figure [ fig : squareknot2b ] shows that the proof will require handle slides of _ both _ components of the link .
[ thm : main ] no smallest genus counterexample to property 2r is fibered .
echoing the notation of proposition [ prop : unknot ] , suppose there is a @xmath0-component counterexample to generalized property r consisting of a fibered knot @xmath10 and another knot @xmath11 .
let @xmath24 be the @xmath23-manifold obtained by @xmath2-framed surgery on @xmath10 alone ; @xmath24 will play the role that @xmath4 played in the proof of proposition [ prop : unknot ] . since @xmath10 is a fibered knot , @xmath24 fibers over the circle with fiber @xmath96 , a closed orientable surface of the same genus as @xmath10 . the dual to @xmath10 in @xmath24 is a knot @xmath44 that passes through each fiber exactly once .
the hypothesis is that @xmath2-framed surgery on @xmath97 creates @xmath1 . following ( * ? ? ?
* corollary 4.2 ) , either the knot @xmath11 lies in a ball , or @xmath11 is cabled with the surgery slope that of the cabling annulus , or @xmath11 can be isotoped in @xmath24 to lie in a fiber , with surgery slope that of the fiber .
if @xmath11 were cabled , then the surgery on @xmath5 would create a lens space summand , which is clearly impossible in @xmath51 .
if @xmath11 can be isotoped into a ball or into a fiber , then , as argued in the proof of proposition [ prop : unknot ] , the isotopy in @xmath24 is realized in @xmath53 by handle - slides of @xmath11 over @xmath10 , so we may as well regard @xmath11 as lying either in a ball that is disjoint from @xmath44 or in a fiber @xmath98 . the former case , @xmath11 in a ball disjoint from @xmath44 would , as in proposition [ prop : unknot ] , imply that the link @xmath99 is the unlink .
so we can assume that @xmath100 .
the surgery on @xmath11 that changes @xmath24 to @xmath1 has this local effect near @xmath96 : @xmath24 is cut open along @xmath96 , creating two copies @xmath101 of @xmath96 , a @xmath0-handle is attached to the copy of @xmath11 in each @xmath102 , compressing the copies of the fiber to surfaces @xmath103 .
the surfaces @xmath104 are then glued back together by the obvious identification to give a surface @xmath105 .
( see the surgery principle lemma [ lemma : surgprin ] below for more detail . )
this surface has two important features : each component of @xmath106 ( there are two components if and only if @xmath11 is separating in @xmath96 ) has lower genus than @xmath96 ; and @xmath106 intersects @xmath44 in a single point .
let @xmath107 be the dual knot to @xmath11 and let @xmath108 be the component of @xmath106 that intersects @xmath44 .
@xmath43 intersects @xmath106 in some collection of points ( in fact , two points , but that is not important for the argument ) .
each point in @xmath109 can be removed by a handle - slide of @xmath43 over @xmath44 along an arc in @xmath108 .
let @xmath110 be the final result of these handle - slides .
then @xmath108 is an orientable surface that has lower genus than @xmath96 , is disjoint from @xmath110 and intersects @xmath44 in a single point .
following lemma [ lemma : dual ] the handle - slides of @xmath43 over @xmath44 in @xmath1 correspond in @xmath53 to handle - slides of @xmath10 over @xmath11 .
call the knot in @xmath53 that results from all these handle - slides @xmath111 . since @xmath108 is disjoint from @xmath110 , and intersects @xmath44 in a single meridian , @xmath112 is a surface in @xmath113 whose boundary is a longitude of @xmath31 . in other words ,
the knot @xmath31 , still a counterexample to property 2r , has @xmath114 as required .
we have just seen that a fibered counterexample to property 2r would not be a least genus counterexample .
we now explore other properties of potential fibered counterexamples . in this section
we consider what can be said about the monodromy of a fibered knot in @xmath53 , and the placement of a second component with respect to the fibering , so that surgery on the @xmath0-component link yields @xmath1 . perhaps surprisingly , the theory of heegaard splittings is useful in answering these questions .
much of this section in fact considers the more general question of when @xmath1 can be created by surgery on a knot in a @xmath23-manifold @xmath24 that fibers over a circle .
the application to property 2r comes from the special case in which the manifold @xmath24 is obtained from @xmath2-framed surgery on a fibered knot in @xmath53 .
suppose @xmath96 is a @xmath0-sided surface in a @xmath23-manifold @xmath24 and @xmath115 is an essential simple closed curve in @xmath96 .
a tubular neighborhood @xmath116 intersects @xmath96 in an annulus ; the boundary of the annulus in @xmath117 defines a slope on @xmath118 .
let @xmath119 denote the manifold obtained from @xmath24 by surgery on @xmath120 with this slope and let @xmath106 be the surface obtained from @xmath96 by compressing @xmath96 along @xmath120 .
[ lemma : surgprin ] @xmath119 can be obtained from @xmath24 by the following @xmath23-step process : 1 .
cut @xmath24 open along @xmath96 , creating two new surfaces @xmath101 in the boundary , each homeomorphic to @xmath96 .
2 . attach a @xmath0-handle to each of @xmath101 along the copy of @xmath120 it contains .
this changes each of the new boundary surfaces @xmath101 to a copy of @xmath106 .
denote these two surfaces @xmath103 .
glue @xmath121 to @xmath122 via the natural identification .
the surgery itself is a @xmath0-step process : remove a neighborhood of @xmath118 , then glue back a copy of @xmath123 so that @xmath124 is the given slope .
the first step is equivalent to cutting @xmath96 along an annulus neighborhood @xmath125 of @xmath120 in @xmath96 , creating a torus boundary component as the boundary union of the two copies @xmath126 of @xmath125 .
thinking of @xmath127 as the boundary union of two intervals , the second step can itself be viewed as a two - step process : attach a copy of @xmath128 to each annulus @xmath129 along @xmath130 ( call the attached copies @xmath131 ) , then identify the boundary disks @xmath132 with @xmath133 in the natural way .
this creates a three - stage process which is exactly that defined in the lemma , except that in the lemma @xmath134 is first cut apart and then reglued by the identity .
the case in which @xmath24 fibers over a circle with fiber @xmath96 is particularly relevant .
we will stay in that case throughout the remainder of this section ( as always , restricting to the case that @xmath24 and @xmath96 are orientable ) and use the following notation : 1 .
@xmath135 is the monodromy homeomorphism of @xmath24 .
@xmath120 is an essential simple closed curve in @xmath96 .
@xmath106 is the surface obtained by compressing @xmath96 along @xmath120 4 .
@xmath119 is the manifold obtained by doing surgery on @xmath24 along @xmath136 using the framing given by @xmath96 .
note that @xmath106 may be disconnected , even if @xmath96 is connected .
[ prop : isotopic1 ] suppose @xmath137 is isotopic to @xmath120 in @xmath96 * if @xmath120 is non - separating in @xmath96 , or if @xmath120 is separating and the isotopy from @xmath137 to @xmath120 reverses orientation of @xmath120 , then @xmath138 , where @xmath139 fibers over the circle with fiber @xmath106 . *
if @xmath120 separates @xmath96 so @xmath140 , and the isotopy from @xmath137 to @xmath120 preserves orientation of @xmath120 , then @xmath141 , where @xmath142 ( resp .
@xmath143 ) fibers over the circle with fiber @xmath144 ( resp @xmath145 ) .
we may as well assume that @xmath146 and consider first the case where @xmath147 is orientation preserving . in this case
, the mapping torus of @xmath120 in @xmath24 is a torus @xmath58 containing @xmath120 .
the @xmath23-stage process of lemma [ lemma : surgprin ] then becomes : 1 .
@xmath24 is cut along @xmath58 to give a manifold @xmath148 with two torus boundary components @xmath149 .
@xmath148 fibers over the circle with fiber a twice - punctured @xmath106 .
( @xmath106 is connected if and only if @xmath120 is non - separating . ) 2 .
a @xmath0-handle is attached to each torus boundary component @xmath150 , turning the boundary of @xmath148 into two @xmath0-spheres .
the two @xmath0-spheres are identified .
the second and third stage together are equivalent to filling in a solid torus along each @xmath150 , giving an @xmath106-fibered manifold @xmath28 , then removing a @xmath23-ball from each solid torus and identifying the resulting @xmath0-spheres . depending on whether @xmath106 is connected or not , this is equivalent to either adding @xmath4 to @xmath28 or adding the two components of @xmath28 together .
the case in which @xmath147 is orientation reversing is only slightly more complicated .
since @xmath24 is orientable , the mapping torus of @xmath147 is a @xmath74-sided klein bottle @xmath5 , so @xmath151 is a single torus @xmath58 .
the argument of lemma [ lemma : surgprin ] still mostly applies , since @xmath120 has an annulus neighborhood in @xmath5 , and shows that the surgery can be viewed as attaching two @xmath0-handles to @xmath58 along parallel curves , converting the boundary into two @xmath0-spheres , then identifying the @xmath0-spheres .
this is again equivalent to filling in a solid torus at @xmath58 ( which double - covers @xmath5 ) and then adding @xmath4 . but
filling in a solid torus at @xmath152 changes the fiber from @xmath96 to @xmath106 .
( note that if @xmath120 separates @xmath96 , so @xmath140 , then since @xmath153 is orientation preserving on @xmath96 but orientation reversing on @xmath120 , @xmath153 must exchange @xmath144 and @xmath145 .
so @xmath139 also fibers over the circle with fiber @xmath154 . )
[ cor : isotopic2 ] if @xmath155 and @xmath137 is isotopic to @xmath120 in @xmath96 , then @xmath96 is a torus . according to proposition [ prop : isotopic1 ] ,
the hypotheses imply that @xmath4 fibers over the circle with fiber ( a component of ) @xmath106 . but this forces @xmath156 and so @xmath157 .
surgery on fibered manifolds also provides a natural connection between the surgery principle and heegaard theory : suppose @xmath158 , @xmath159 are two copies of a compression body @xmath77 and @xmath160 is a given homeomorphism .
then the union of @xmath158 , @xmath159 along their boundaries , via the identity on @xmath161 and via @xmath162 , is called the _ heegaard double _ of @xmath77 ( via @xmath153 ) .
note that the closed complement @xmath163 of @xmath164 in a heegaard double is a manifold in which @xmath158 is attached to @xmath159 by only identifying @xmath165 and @xmath166 ( via @xmath153 ) .
but such an identification constitutes a heegaard splitting @xmath167 of @xmath163 .
lemma [ lemma : surgprin ] gives this important example : [ example : double ] for @xmath168 as above , let @xmath77 be the compression body obtained from @xmath169 by attaching a @xmath0-handle to @xmath170 along @xmath120
. then @xmath119 is the heegaard double of @xmath77 via @xmath153 .
here is the argument : delete from @xmath24 regular neighborhoods of both the fiber @xmath96 that contains @xmath120 and another fiber @xmath171 .
the result is two copies @xmath172 of @xmath169 each containing a copy @xmath173 of @xmath96 and a copy @xmath174 of @xmath171 .
one can reconstruct @xmath24 by identifying @xmath175 and @xmath176 by the natural identification and identifying @xmath177 with @xmath178 via @xmath153 .
now apply the surgery principle to the surface @xmath179 in @xmath24 .
each @xmath180 becomes a compression body @xmath181 in which @xmath174 is @xmath182 and the copy @xmath104 of @xmath173 compressed along @xmath120 is @xmath161 .
next is a sample application , using heegaard theory .
first a preliminary lemma : [ lemma : singlehandle ] suppose @xmath77 is a compression body built with a single @xmath0-handle , whose core @xmath183 is bounded by a simple closed curve @xmath184 .
suppose @xmath185 is any properly embedded disk whose boundary is essential in @xmath186 . 1 .
if @xmath120 is separating in @xmath186 then @xmath187 is isotopic to @xmath120 in @xmath186 .
if @xmath120 is non - separating in @xmath186 then either @xmath187 is isotopic to @xmath120 in @xmath186 or @xmath188 cuts off a solid torus from @xmath77 whose meridian disk is bounded by @xmath120 .
dually , @xmath77 can be viewed as constructed by adding a single @xmath74-handle to , say , @xmath189 .
the co - core of the @xmath74-handle is the disk @xmath183 , with boundary @xmath120 . a fairly standard innermost disk
, outermost arc argument shows that @xmath188 may be isotoped to be disjoint from @xmath183 : the central observation is that an outermost subdisk @xmath190 cut off from @xmath188 by @xmath183 will have boundary that is inessential in @xmath191 , since @xmath191 is incompressible in @xmath192 , and the disk that @xmath193 bounds in @xmath191 ca nt contain an entire copy of @xmath183 , since @xmath187 intersects both copies of @xmath194 in the same number of points .
( these copies of @xmath183 can be viewed as the two disks on which the original @xmath74-handle was attached . )
once @xmath188 is disjoint from @xmath183 , it lies entirely in @xmath195 with @xmath196 .
since @xmath195 has incompressible boundary , @xmath187 is inessential in @xmath197 ; since @xmath187 was essential in @xmath186 we conclude that @xmath187 bounds a disk in @xmath197 containing either one or both copies of @xmath183 that lie in @xmath197 . in the first case ( which always occurs
if @xmath120 is separating , so the copies of @xmath183 lie on different components of @xmath198 ) , @xmath188 is parallel to @xmath199 in @xmath200 . in the second case ,
when @xmath187 bounds a disk in @xmath198 containing both copies of @xmath183 , @xmath188 is separating and cuts off a solid torus from @xmath77 that contains @xmath183 as a meridian disk .
[ prop : monodromy ] for @xmath168 as above , suppose some surgery on @xmath120 gives a reducible manifold .
then the surgery slope is that of @xmath96 and either 1 .
@xmath137 can be isotoped in @xmath96 so that it is disjoint from @xmath120 or 2 .
@xmath115 is non - separating and @xmath201 , where * @xmath139 fibers over the circle with fiber @xmath106 and * @xmath7 is either @xmath53 or a lens space .
note in particular that possibility ( 2 ) is not consistent with @xmath155 .
choose distinct fibers @xmath202 in @xmath24 , with @xmath115 . via (
* corollary 4.2 ) and the proof of theorem [ thm : main ] we know that the surgery on @xmath120 must use the framing given by the fiber @xmath96 , so the result of surgery is @xmath119 .
example [ example : double ] shows that @xmath119 is a heegaard double via @xmath153 , so the complement @xmath203 of a regular neighborhood of @xmath204 has a heegaard splitting @xmath205 .
that is , @xmath206 . if @xmath156 , so @xmath157 , then the compression bodies @xmath158 and @xmath159 in the heegaard
splitting of @xmath148 are each punctured solid tori and @xmath119 is obtained from @xmath148 by identifying the @xmath0-spheres at the punctures . hence @xmath119 is the connected sum of @xmath4 with a closed @xmath23-manifold of heegaard genus one , either @xmath53 , a lens space , or @xmath4 .
but the last happens only if the same curve in @xmath171 compresses in both @xmath158 and @xmath159 ; in our context , that implies @xmath120 and @xmath137 are isotopic in @xmath96 , and so can be isotoped to be disjoint . if @xmath207 , choose a reducing sphere for @xmath119 with a minimal number of intersection curves with @xmath106 .
if the reducing sphere is disjoint from @xmath106 , then @xmath148 is reducible . if , on the other hand , the reducing sphere intersects @xmath106 , then at least one copy of @xmath106 in @xmath208 must be compressible in @xmath148 .
we conclude that in either case the heegaard splitting @xmath205 of @xmath148 is weakly reducible ( and possibly reducible ) , see @xcite .
that is , there are essential disjoint simple closed curves @xmath209 in @xmath210 which compress respectively in @xmath158 and @xmath159 . *
case 1 : * the curve @xmath120 is separating . in this case , since the compression bodies @xmath181 each have only the @xmath0-handle with attaching circle @xmath115 attached , it follows from lemma [ lemma : singlehandle ] that any curve in @xmath211 that compresses in @xmath181 is isotopic to @xmath212 . in particular , fixing the identification @xmath213 , @xmath214 must represent @xmath120 in @xmath171 and @xmath215 represents @xmath137 .
hence @xmath120 and @xmath137 are disjoint .
* case 2 : * the curve @xmath120 is non - separating , and so is at least one of the curves @xmath209 .
if both curves @xmath216 are non - separating then it follows from lemma [ lemma : singlehandle ] that both @xmath215 and @xmath214 , when viewed in the handlebodies @xmath217 , must each be isotopic to @xmath212 and the case concludes as case 1 did .
if @xmath214 is non - separating , and @xmath215 is separating , then it follows from lemma [ lemma : singlehandle ] that @xmath214 is isotopic to @xmath218 whereas @xmath215 bounds a punctured torus @xmath219 on which @xmath137 lies .
if @xmath214 is disjoint from @xmath58 , then @xmath120 and @xmath137 are disjoint , as required .
if @xmath214 lies in @xmath58 then @xmath220 also bounds a disk in @xmath159 .
that is , each @xmath181 can be viewed as the boundary sum , along a disk whose boundary is @xmath220 of a copy of @xmath221 and a solid torus .
when the @xmath174 are identified to get @xmath148 , the union of the disks in @xmath158 and @xmath159 bounded by @xmath220 constitute a sphere that decomposes @xmath148 into @xmath222 , where @xmath67 is heegaard split by @xmath58 into two solid tori , with meridian disks bounded by @xmath120 and @xmath137 respectively .
@xmath119 is then obtained by identifying the two copies of @xmath106 in @xmath208 .
hence @xmath223 , where @xmath139 fibers over @xmath127 with fiber @xmath106 .
if @xmath224 then @xmath67 is a lens space . if @xmath225 then @xmath226 . if @xmath227 then @xmath137 is disjoint from @xmath120 . * case 3 : * the curve @xmath120 is non - separating , but both @xmath209 are separating . in this case
, much as in case 2 , each @xmath216 cuts off a torus @xmath150 from @xmath228 , with @xmath229 and @xmath230 . since the @xmath216 are disjoint , the two tori either also are disjoint ( and the proof is complete ) or the two tori coincide .
if the two tori coincide , the argument concludes as in case 2 .
in applying proposition [ prop : monodromy ] to the manifold @xmath24 obtained from @xmath2-framed surgery on a fibered knot @xmath3 , note that the isotopy in the proposition takes place in a fiber @xmath96 of @xmath24 , the closed manifold obtained by @xmath2-framed surgery on @xmath5 , not in the fiber @xmath231 of the knot @xmath5 itself .
the importance of the distinction is illustrated by the following proposition which , without the distinction , would ( following propositions [ prop : unknot ] , [ prop : monodromy ] and corollary [ cor : genustwo](1 ) ) seem to guarantee that all genus @xmath0 fibered knots have property 2r .
[ prop : strictmono ] suppose @xmath232 is a fibered knot , with fiber the punctured surface @xmath233 and monodromy @xmath234 .
suppose a knot @xmath235 has the property that @xmath2-framed surgery on the link @xmath236 gives @xmath1 and @xmath237 can be isotoped to be disjoint from @xmath11 in @xmath238 .
then either @xmath11 is the unknot or @xmath239 .
_ case 1 : _ @xmath11 bounds a disk in @xmath238 or is parallel in @xmath238 to @xmath240 . in this case ,
@xmath2-framed surgery on @xmath236 would be @xmath241 , where @xmath139 is the result of @xmath2-framed surgery on @xmath10 .
our hypothesis is that @xmath242 which , by classical property r @xcite , implies that @xmath10 is the unknot . hence @xmath243 .
_ case 2 : _
@xmath11 is essential in @xmath238 .
if @xmath238 is a punctured torus , then the fact that @xmath11 is essential and @xmath237 can be isotoped off of @xmath11 imply that @xmath237 is isotopic to @xmath11 , and we may as well assume that @xmath244 .
the mapping torus of @xmath245 is then a non - separating torus in @xmath53 , which is absurd .
suppose @xmath238 is a punctured genus @xmath0 surface , and let @xmath96 denote the closed surface obtained by capping off the puncture .
we may as well assume that @xmath246 , and , following corollary [ cor : isotopic2 ] , @xmath247 is not isotopic to @xmath11 in @xmath96 .
in particular , @xmath11 must be non - separating .
since @xmath11 and @xmath247 are non - separating and disjoint in @xmath238 , but not isotopic in @xmath96 , if @xmath238 is compressed along both @xmath11 and @xmath247 simultaneously , @xmath238 becomes a disk .
apply the surgery principle lemma [ lemma : surgprin ] to @xmath11 and conclude that @xmath232 bounds a disk after @xmath2-framed surgery on @xmath11 .
in particular , if @xmath139 is the @xmath23-manifold obtained by @xmath2-framed surgery on @xmath11 alone , then surgery on @xmath236 would give @xmath241 .
for this to be @xmath1 would require @xmath242 hence , again by classical property r , @xmath248 would be the unknot .
return to the general case of fibered manifolds and surgery on a curve @xmath120 in the fiber , and consider the case in which the fiber has genus two . according to corollary [ cor : isotopic2 ] , if the result of surgery on @xmath120 is @xmath1 , then @xmath137 is not isotopic to @xmath120 in @xmath96 .
the following proposition is a sort of weak converse .
[ prop : nonisotopic ] for @xmath168 as above , suppose @xmath96 has genus @xmath0 and @xmath137 can be isotoped off of @xmath120 in @xmath96 .
if @xmath137 is not isotopic to @xmath120 in @xmath96 then @xmath249 , where @xmath7 is @xmath53 , @xmath4 , or a lens space .
we may as well assume that @xmath137 is disjoint from @xmath120 but not isotopic to @xmath120 in @xmath96 .
since @xmath96 is genus two , this immediately implies that @xmath120 is non - separating .
take the heegaard viewpoint of example [ example : double ] .
the complement @xmath148 of a regular neighborhood of @xmath106 in @xmath119 has a heegaard splitting @xmath205 , with the splitting surface @xmath171 a fiber not containing @xmath120 .
since @xmath137 can be isotoped off of @xmath120 in @xmath171 , the heegaard splitting is a weakly reducible splitting , with @xmath250 bounding a disk in @xmath159 and @xmath251 bounding a disk in @xmath158 .
now do a weak reduction of this splitting .
that is , consider the @xmath0-handles @xmath252 with boundary @xmath250 and @xmath253 with boundary @xmath137 .
since @xmath120 and @xmath137 are disjoint , @xmath148 can also be regarded as the union of compression bodies @xmath254 and @xmath255 .
each @xmath256 can be regarded as obtained from @xmath221 by attaching a single @xmath0-handle ; e. g. @xmath257 is obtained from @xmath221 by attaching the @xmath0-handle @xmath258 at one end . since @xmath96 was of genus two ,
@xmath106 is a torus and so each @xmath256 is a once - punctured solid torus .
hence @xmath119 is a heegaard double of a once - punctured solid torus , and this is easily seen to be the connected sum of @xmath4 with a manifold @xmath7 that has a genus one heegaard splitting .
[ cor : nonisotopic ] for @xmath168 as above , suppose @xmath96 has genus @xmath0 and @xmath119 is reducible .
if @xmath137 is not isotopic to @xmath120 in @xmath96 then @xmath259 , where @xmath7 is @xmath53 , @xmath4 , or a lens space and @xmath28 is either @xmath4 or a torus bundle over the circle . via ( * ? ? ?
* corollary 4.2 ) and the proof of theorem [ thm : main ] we know that the surgery on @xmath120 must use the framing given by the fiber in which it lies .
apply proposition [ prop : monodromy ] .
if the first conclusion holds , and @xmath137 can be isotoped off of @xmath120 in @xmath96 , then proposition [ prop : nonisotopic ] can be applied and that suffices .
if the second conclusion holds then @xmath120 is non - separating , so @xmath106 is a torus , as required .
[ cor : genustwo ] suppose @xmath232 is a genus two fibered knot and @xmath248 is a disjoint knot .
then @xmath2-framed surgery on @xmath236 gives @xmath1 if and only if after possible handle - slides of @xmath11 over @xmath10 , 1 .
@xmath11 lies in a fiber of @xmath10 ; 2 . in the closed fiber @xmath96 of the manifold @xmath24 obtained by @xmath2-framed surgery on @xmath10 ,
@xmath247 can be isotoped to be disjoint from @xmath11 ; 3 .
@xmath247 is not isotopic to @xmath11 in @xmath96 ; and 4 .
the framing of @xmath11 given by @xmath96 is the @xmath2-framing of @xmath11 in @xmath53 .
suppose first that @xmath2-framed surgery on @xmath236 gives @xmath1
. apply ( * ? ? ?
* corollary 4.2 ) as in the proof of theorem [ thm : main ] to handle - slide @xmath11 over @xmath10 until it lies in the fiber of @xmath10 in a way that the @xmath2-framing on @xmath11 is the framing given by the fiber in which it lies .
proposition [ prop : monodromy ] shows that @xmath247 satisfies the second condition and corollary [ cor : isotopic2 ] gives the third : @xmath247 is not isotopic in @xmath96 to @xmath11 .
for the other direction , suppose @xmath11 lies in a fiber of @xmath10 and the four conditions are satisfied .
the last condition says that the surgery on @xmath11 is via the slope of the fiber . by proposition [ prop : nonisotopic ]
, the surgery gives @xmath260 , for @xmath7 either @xmath53 , a lens space , or @xmath4 .
but @xmath10 and @xmath11 are algebraically unlinked in @xmath53 ( push @xmath11 off of @xmath96 ) , so @xmath2-framed surgery on @xmath236 must give a homology @xmath1 .
this forces @xmath7 to be a homology @xmath4 , hence @xmath4 precisely .
the square knot @xmath261 is the connected sum of two trefoil knots , @xmath262 and @xmath263 , respectively the right - hand trefoil knot and the left - hand trefoil knot .
there are many @xmath0-component links containing @xmath261 so that surgery on the link gives @xmath1 .
figure [ fig : squareknot ] shows ( by sliding @xmath261 over the unknot ) that the other component could be the unknot ; figure [ fig : squareknot2b ] shows ( by instead sliding the unknot over @xmath261 ) that the second component could be quite complicated . in this section
we show that , up to handle - slides of @xmath11 over @xmath261 , there is a straight - forward description of all two component links @xmath264 , so that surgery on @xmath264 gives @xmath1 .
the critical ingredient in the characterization of @xmath11 is the collection of properties listed in corollary [ cor : genustwo ] , which apply because @xmath261 is a genus two fibered knot .
let @xmath24 be the @xmath23-manifold obtained by @xmath2-framed surgery on the square knot @xmath261 , so @xmath24 fibers over the circle with fiber the closed genus @xmath0 surface @xmath96 .
there is a simple picture of the monodromy @xmath135 of the bundle @xmath24 , obtained from a similar picture of the monodromy on the fiber of a trefoil knot , essentially by doubling it ( * ? ? ?
* section 10.i ) : @xmath96 at 100 260 @xmath163 at 300 230 at 156 167 at 90 330 @xmath265 at 250 220 at 190 172 regard @xmath96 as obtained from two spheres by attaching @xmath23 tubes between them .
see figure [ fig : rhosigma ] .
there is an obvious period @xmath23 symmetry @xmath266 gotten by rotating @xmath267 around an axis intersecting each sphere in two points , and a period @xmath0 symmetry ( the hyperelliptic involution ) @xmath268 obtained by rotating around a circular axis that intersects each tube in two points .
then @xmath269 is an automorphism of @xmath96 of order @xmath270 .
the quotient of @xmath96 under the action of @xmath271 is a sphere with @xmath22 branch points , each of branching index @xmath23 .
let @xmath163 be the @xmath22-punctured sphere obtained by removing the branch points .
a simple closed curve in @xmath163 is _ essential _ if it does nt bound a disk and is not @xmath272-parallel . put another way , a simple closed curve in @xmath163 is essential if and only if it divides @xmath163 into two twice - punctured disks .
it is easy to see that there is a separating simple closed curve @xmath273 that is invariant under @xmath274 and @xmath271 , and hence under @xmath153 , that separates @xmath96 into two punctured tori @xmath275 and @xmath276 ; the restriction of @xmath153 to @xmath275 or @xmath276 is the monodromy of the trefoil knot .
the quotient of @xmath277 under @xmath271 is shown as the brown curve in figure [ fig : rhosigma ] .
( digressing a bit , it is instructive to compare in this example the monodromy of the closed surface @xmath96 with that of the seifert surface @xmath234 appearing in proposition [ prop : strictmono ] .
to recover @xmath278 from the given description of @xmath135 , consider @xmath238 as the complement of a point @xmath279 in @xmath96 .
the image of the point @xmath280 under @xmath153 is another point in @xmath277 ; to see the monodromy on @xmath281 , compose @xmath153 with an isotopy of @xmath96 that slides @xmath282 back to @xmath280 along a subarc of @xmath277 .
typically , even a simple closed curve in @xmath96 that is carried by @xmath153 to a disjoint simple closed curve will intersect its image once @xmath282 is slid back to @xmath280 . )
an important property of the hyperelliptic involution @xmath268 is that it fixes the isotopy class of each curve .
that is , for any simple closed curve @xmath115 , @xmath283 is isotopic to @xmath120
. the isotopy preserves orientation on @xmath120 if and only if @xmath120 is separating .
this immediately gives for any simple closed curve @xmath115 , @xmath137 is isotopic in @xmath96 to a curve disjoint from @xmath120 if and only if @xmath284 is isotopic in @xmath96 to a curve disjoint from @xmath120 . since @xmath285 is isotopic to @xmath120 in @xmath96 , @xmath137 is isotopic to a curve disjoint from @xmath120 if and only if @xmath137 is isotopic to a curve disjoint from @xmath286 .
applying @xmath287 shows that @xmath137 is isotopic to a curve disjoint from @xmath286 if and only if @xmath288 is isotopic to a curve disjoint from @xmath120 .
[ lemma : overline]for the @xmath23-fold branched covering @xmath289 described above : 1 .
an essential simple closed curve @xmath115 has the property that @xmath284 can be isotoped to be disjoint from @xmath120 if and only if @xmath120 is isotopic to a lift of an essential simple closed curve @xmath290 in @xmath163 .
2 . a lift @xmath120 of an essential simple closed curve @xmath290 in @xmath163 is separating in @xmath96 if and only if @xmath290 separates the pair of branch points coming from @xmath276 from the branch points coming from @xmath275 .
a single lift @xmath120 of an essential simple closed curve @xmath290 in @xmath163 is non - separating in @xmath96 if and only if @xmath120 projects homeomorphically to @xmath290 .
since @xmath290 is embedded , any lift @xmath120 of @xmath290 will either be carried by @xmath271 to a disjoint curve , or to @xmath120 itself . in either case , @xmath284 can be isotoped off of @xmath120 . on the other hand ,
suppose @xmath284 can be isotoped off of @xmath120 .
give @xmath96 the standard hyperbolic metric , which is invariant under @xmath271 , and isotope @xmath120 to a geodesic in @xmath96 .
then @xmath284 is also a geodesic and , since they can be isotoped apart , either @xmath291 or @xmath292 .
apply @xmath293 to deduce that @xmath294 either all coincide or are all disjoint . in either case , @xmath120 projects to an embedded curve in @xmath163 .
this proves the first part of the lemma . to prove the second and third parts , we first establish : * claim : * an essential
simple closed curve @xmath295 lifts to a curve @xmath120 that is invariant under @xmath271 if and only if @xmath120 is separating in @xmath96 .
_ proof of claim .
_ since @xmath296 are not eigenvalues of the monodromy matrix for @xmath297 , no non - separating curve is left invariant by @xmath153 or @xmath271 . on the other hand ,
if the essential curve @xmath120 ( and hence also @xmath284 ) is separating and @xmath284 is disjoint from @xmath120 then , since @xmath96 only has genus @xmath0 , @xmath120 and @xmath284 must be parallel in @xmath96 .
this proves the claim .
the claim establishes the third part of the lemma . for the second part ,
we restate the claim : whether @xmath290 lifts to a non - separating or to a separating curve is completely determined by whether @xmath290 is covered in @xmath96 by three distinct curves or is thrice covered by a single curve . that , in turn , is determined by whether @xmath290 represents an element of @xmath298 that is mapped trivially or non - trivially to @xmath299 under the homomorphism @xmath300 that defines
the branched covering @xmath301 .
one way to view such a curve @xmath290 in @xmath163 is as the boundary of a regular neighborhood of an arc @xmath302 between two branch points of @xmath303 , namely a pair of branch points ( either pair will do ) that lie on the same side of @xmath290 in @xmath304 . from that point of view
, @xmath290 represents a trivial element of @xmath299 if and only if the normal orientations of the branch points at the ends of the arc @xmath302 disagree ; that is , one is consistent with a fixed orientation of the rotation axis and one is inconsistent .
put another way , @xmath290 represents a trivial element if and only if one endpoint of @xmath302 is a north pole in figure [ fig : rhosigma ] and the other is a south pole .
the brown curve in the figure , which lifts to the curve separating @xmath275 from @xmath276 , separates both north poles from both south poles .
this establishes the second part of the lemma .
[ lemma : framing ] if @xmath305 projects to a simple closed curve @xmath290 in @xmath163 then the framing of @xmath120 given by the seifert surface @xmath233 of @xmath261 is the @xmath2-framing .
here is an equivalent conclusion : if @xmath306 is a parallel copy of @xmath120 in @xmath238 , then @xmath307 in @xmath53 .
this self - linking number @xmath308 is independent of the orientation of @xmath120 ( since reversing orientation reverses also the orientation of its push - off @xmath306 ) and does nt change when @xmath120 is isotoped in @xmath238 .
one way of viewing @xmath308 is this : take an annular neighborhood of @xmath120 in @xmath238 and ask whether in @xmath53 the annulus is right - hand twisted ( @xmath309 ) , left - hand twisted ( @xmath310 ) or not twisted at all ( @xmath311 ) . here
are some useful properties : 1 .
let @xmath312 be a simple closed curve in @xmath238 that is parallel to @xmath313
. then @xmath314 .
+ use as a representative annular neighborhood of @xmath312 the annulus between @xmath312 and @xmath313 .
then the complement of that annulus in @xmath238 provides a null - cobordism of one end of the annulus that is disjoint from the other end of the annulus , hence the two ends of the annulus have trivial linking number .
2 . suppose @xmath120 is a simple closed curve in @xmath238 and @xmath306 is a curve obtained by band - summing @xmath120 to the curve @xmath312 just described , via a band that lies in @xmath238 .
then @xmath315 + first note that a slight push - off of @xmath238 from itself gives a null - cobordism of @xmath312 that is disjoint from @xmath120 , so @xmath316 .
now a direct argument in a spirit similar to that of the framing addition formula for handle - slides ( see @xcite ) shows @xmath317 3 .
[ enum : isotopy ] suppose @xmath318 are simple closed curves in @xmath238 that are isotopic _ in @xmath96 _ ( not necesarily in @xmath238 ) . then @xmath319 thus @xmath320 can be extended unambiguously to isotopy classes of curves in @xmath96 .
+ an isotopy in @xmath96 can be broken up into a series of isotopies in @xmath238 and ( when the puncture in @xmath96 corresponding to @xmath313 is crossed ) band sums to the curve @xmath312 described above .
we have just seen that both moves leave @xmath320 unchanged .
[ enum : sigma ] for any simple closed curve @xmath120 in @xmath96 and @xmath268 the hyperelliptic involution , @xmath283 is isotopic to @xmath120 in @xmath96 , so @xmath321 5 .
[ enum : rho ] for any simple closed curve @xmath120 in @xmath96 and @xmath266 the period @xmath23 symmetry described above , @xmath322 + the monodromy @xmath323 does nt change @xmath320 , since the fiber structure gives an isotopy in @xmath53 of any annulus in @xmath238 to its image under the monodromy . according to ( [ enum : sigma ] ) @xmath274 also does nt change @xmath320 , so then neither does @xmath324 . now consider the following automorphism @xmath325 of @xmath53 that preserves the seifert surface @xmath238 of @xmath261 , as viewed in the left side of figure [ fig : tau ] : reflect through the plane of the knot projection ( this reflection makes the right - hand trefoil a left - hand trefoil , and vice versa ) , then do a @xmath326 rotation around an axis in @xmath53 that passes through the center of the arc @xmath327 that separated @xmath262 from @xmath263 .
the automorphism @xmath325 is orientation preserving on the plane of reflection , since the reflection is the identity on this plane and the rotation is orientation preserving .
hence @xmath325 is orientation preserving also on the seifert surface , since , for example , we can suppose the seifert surface has a point of tangency with the plane of reflection throughout the construction .
on the other hand , @xmath325 is orientation reversing on @xmath53 . because of the orientation reversal of @xmath53 we deduce that for any simple closed curve @xmath120 in @xmath238 , @xmath328 or ,
more concretely , @xmath325 changes a right - hand twist in the annulus about @xmath120 in @xmath96 into a left - hand twist and vice versa .
a visual example of this phenomenon is how @xmath325 carries the right - hand trefoil summand , with right - hand twists in its part of the seifert surface , to the left - hand trefoil summand with left - hand twists in its part of the seifert surface .
the automorphism @xmath329 descends to an automorphism @xmath330 of @xmath303 given by @xmath326-rotation about the blue axis shown in figure [ fig : tau ] .
this rotation exchanges @xmath275 and @xmath276 but , up to isotopy and orientation , leaves invariant the simple closed curves in @xmath163 . in particular ,
if @xmath115 projects to a simple closed curve @xmath295 , then @xmath331 projects to @xmath332 which is isotopic to @xmath290 in @xmath163 . the isotopy lifts to @xmath96 , so it follows that @xmath331 is isotopic in @xmath96 to some @xmath333 .
hence by ( [ enum : isotopy ] ) and ( [ enum : rho ] ) above @xmath334 this implies that @xmath311 . at 285 125 at 190 90
[ cor : enumerate ] suppose @xmath335 is the square knot with fiber @xmath233 and @xmath248 is a disjoint knot .
then @xmath2-framed surgery on @xmath264 gives @xmath1 if and only if , after perhaps some handle - slides of @xmath11 over @xmath261 , @xmath11 lies in @xmath238 and the quotient map @xmath336 projects @xmath11 homeomorphically to an essential simple closed curve in @xmath163 .
note that , according to lemma [ lemma : overline ] , essential simple closed curves @xmath290 in @xmath163 that are such homeomorphic projections are precisely those for which one branch point of @xmath276 ( or , equivalently , one branch point from @xmath275 ) lies on each side of @xmath290 .
so another way of saying that @xmath337 projects @xmath11 homeomorphically to an essential simple closed curve in @xmath163 is to say that @xmath11 is the lift of an essential simple closed curve in @xmath163 that separates one branch point of @xmath276 ( or , equivalently @xmath275 ) from the other .
this is an application of proposition [ cor : genustwo ] .
suppose that @xmath2-framed surgery on @xmath264 gives @xmath1 .
then , after possible handle - slides over @xmath261 , @xmath11 lies in @xmath238 . according to proposition [ cor : genustwo ] , in the closed surface @xmath96
, @xmath247 can be isotoped to be disjoint from @xmath11 but is not isotopic to @xmath11 .
the result then follows from lemma [ lemma : overline ] .
conversely , suppose that after some handle - slides of @xmath11 over @xmath261 , @xmath11 lies in @xmath238 and @xmath337 projects @xmath11 homeomorphically to an essential simple closed curve @xmath290 in @xmath338 .
@xmath339 then ca nt be isotopic to @xmath11 in @xmath96 ( or else in fact @xmath340 and so @xmath341 would not be a homeomorphism ) but @xmath339 is disjoint from @xmath11 , since @xmath290 is embedded .
thus the second and third conditions of proposition [ cor : genustwo ] are satisfied .
lemma [ lemma : framing ] establishes the fourth condition .
to understand why the square knot @xmath261 is likely to fail property 2r , it is useful to ascend to 4 dimensions .
we use an unusual family of handlebodies from @xcite to construct a family of 2-component links in @xmath53 , each containing @xmath261 as one component .
each of these links will fail to satisfy generalized property r if a certain associated presentation of the trivial group fails ( as seems likely ) to satisfy the andrews - curtis conjecture .
these links will also be the centerpiece of our discussion of the slice - ribbon conjecture in section [ slice ] . in @xcite , the first author provided unexpected examples of handle structures on homotopy @xmath22-spheres which do not obviously simplify to give the trivial handle structure on @xmath342 .
at least one family is highly relevant to the discussion above .
this is example ( * ? ? ?
* figure 1 ) , reproduced here as the left side of figure [ fig : gompffig1b ] .
( setting @xmath343 gives rise to the square knot . )
the two circles with dots on them represent @xmath74-handles , indicating that the @xmath22-manifold with which we begin is @xmath344 with boundary @xmath1 given by @xmath2-surgery .
( think of the two dotted unknotted circles as bounding disjoint unknotted disks in the @xmath22-ball ; the dots indicate one should scoop these disks out of @xmath345 rather than attach @xmath0-handles to @xmath345 . )
the circles without dots represent @xmath0-handles attached to @xmath1 with the indicated framing .
a sequence of kirby operations in @xcite shows that the resulting @xmath22-manifold has boundary @xmath53 . at 156 758 at 451 758 @xmath6 at 154 678 @xmath6 at 448 678 @xmath2 at 177 718 @xmath346 at 117 718 @xmath347 at 85 718 @xmath348 at 232 718 @xmath347 at 377 718 @xmath348 at 527 718 @xmath349
$ ] at 470 718 @xmath349 $ ] at 336 772 @xmath349 $ ] at 336 666 @xmath350 $ ] at 422 705 @xmath2 at 422 733 @xmath2 at 508 635 we will be interested in the @xmath22-manifold that is the trace of the @xmath0-handle surgeries , the manifold that lies between @xmath1 and @xmath53 . our viewpoint , though , is dual to that in @xcite ; the @xmath22-manifold is thought of as starting with @xmath53 ( actually , with @xmath351 ) to which two @xmath0-handles are attached along a link @xmath20 to get @xmath1 .
this dual viewpoint puts the construction solidly in the context of this paper .
a recipe for making the switch in perspective is given in ( * ? ? ?
* example 5.5.5 ) .
the result is shown in the right half of figure [ fig : gompffig1b ] .
the circles representing @xmath74-handles are changed to @xmath2-framed @xmath0-handles and then all @xmath0-handles have their framing integers bracketed .
we call these circles bracketed circles ; they form a complicated surgery description of @xmath53 .
new @xmath2-framed @xmath0-handles are added , each linking one of these original @xmath0-handles .
these small linking @xmath0-handles will eventually be the link @xmath20 , after we eliminate the other components by following the recipe for turning the original diagram into a trivial diagram of the @xmath23-sphere .
this process allows handle slides of bracketed or unbracketed circles over bracketed ones , but does not allow handle slides of bracketed circles over unbracketed circles .
the reduction is done in figure [ fig : gompffig2b ] and roughly mimics ( * ? ? ?
* figures 7 and 8) : the top ( red ) circle is handle - slid over the bottom ( black ) circle @xmath352 via a band that follows the tall thin @xmath349$]-framed circle . in order to band - sum to a @xmath2-framed push - off of @xmath352 , the @xmath6 twists that are added in the lower twist box are undone by adding @xmath353 twists via a new twist box to the left .
the hopf link at the bottom of the left diagram is removed in a @xmath0-stage process : the unbracketed component is handleslid over @xmath352 , so it is no longer linked with the @xmath349$]-framed component of the hopf link . then @xmath352 is canceled with the @xmath349$]-framed circle of the hopf link .
the result is the picture on the right .
the apparent change is both that the hopf link has been removed and the framing of @xmath352 has been changed from @xmath349 $ ] to @xmath2 . at 150 520 at 450 520 @xmath6 at 455 440 @xmath6 at 150 440
@xmath353 at 25 425 @xmath353 at 337 425 @xmath2 at 176 360 @xmath349 $ ] at 25 476 @xmath352 at 275 476 @xmath2 at 340 476 @xmath349 $ ] at 140 360 @xmath349 $ ] at 40 530 @xmath349 $ ] at 352 530 @xmath350 $ ] at 125 500 @xmath350 $ ] at 433 500 @xmath2 at 180 495 @xmath2 at 480 500 @xmath347 at 78 476 @xmath348 at 227 476 @xmath347 at 380 476 @xmath348 at 527 476 next ( figure [ fig : gompffig3b ] ) the large red @xmath349$]-framed circle is handle - slid over the central red @xmath350 $ ] -framed circle .
that gives the figure on the left .
note that the twist boxes guarantee we have pushed off with framing @xmath346 as required .
since there is now only one strand going through the top twist - box , the box can be ignored .
the handle - slide changes the @xmath349$]-framing on the slid circle to a @xmath354$]-framing , since the two curves , if oriented in the same direction through the twist boxes , have linking number @xmath355 and , with that orientation , the handle - slide corresponds to a handle subtraction . to get the figure on the right , the lower black circle is moved into position and a flype moves the lower twist - box to a new upper twist - box .
( note the sphere enclosing the left half of the diagram , intersecting the @xmath6-twist box and two other strands . )
if we ignore the black circles , the red curves form an unlink . blowing these down
would yield @xmath53 with the black curves realizing the desired link @xmath20 .
@xmath6 at 155 673 @xmath6 at 435 730 @xmath353 at 35 656 @xmath353 at 310 650 @xmath2 at 40 705 @xmath2 at 320 705 @xmath356 $ ] at 50 755 @xmath356 $ ] at 325 755 @xmath350 $ ] at 158 740 @xmath350 $ ] at 435 690 @xmath347 at 85 705 @xmath348 at 230 705 @xmath347 at 360 705 @xmath348 at 509 705 aim now for the square knot , by setting @xmath343 and changing the full twists to half twists by a flype .
figure [ fig : gompffig4 ] then shows a series of isotopies ( clockwise around the figure beginning at the upper left ) that simplify the red @xmath357$]-framed circles , while dragging along the @xmath2-framed black circles . @xmath6 at 150 365 @xmath6 at 435 370 @xmath6 at 445 180 @xmath6 at 200 175 @xmath353 at 30 410 @xmath353 at 320 415 @xmath353 at 40 230 @xmath353 at 320 235 @xmath2 at 35 460 @xmath2 at 325 460 @xmath2 at 325 300 @xmath2 at 40 280 @xmath356 $ ] at 45 510 @xmath356 $ ] at 340 510 @xmath356 $ ] at 355 270 @xmath356 $ ] at 70 260 @xmath350 $ ] at 130 470 @xmath350 $ ] at 455 410 @xmath350 $ ] at 400 240 @xmath350 $ ] at 240 220 @xmath2 at 160 450 @xmath2 at 430 400 @xmath2 at 470 325 @xmath2 at 220 270 finally , figure [ fig : gompffig5 ] shows that the middle @xmath2-framed component , when the two red @xmath357$]-framed components are blown down , becomes the square knot @xmath335 .
( the other component becomes an interleaved connected sum of two torus knots , @xmath358 , as shown in figure [ fig : ln1 ] . )
this leaves two natural questions .
@xmath356 $ ] at 75 530 @xmath350 $ ] at 215 460 _ question one : _ as described , @xmath359 does not obviously lie on a seifert surface for @xmath261 . according to corollary [ cor : enumerate ] ,
some handle slides of @xmath359 over @xmath261 should alter @xmath359 so that it is one of the easily enumerated curves that do lie on the seifert surface , in particular it would be among those that are lifts of ( half of ) the essential simple closed curves in the @xmath22-punctured sphere @xmath163 . which curves in @xmath163 represent @xmath359 for some @xmath6 ?
_ question two : _ is each @xmath360 , @xmath361 , a counter - example to generalized property r ?
this second question is motivated by figure [ fig : gompffig1b ] . as described in @xcite , the first diagram of that figure exhibits a simply connected 2-complex , presenting the trivial group as @xmath362 where @xmath363 is some word in @xmath364 depending on @xmath347 and equal to @xmath365 when @xmath366 .
if the 2-component link @xmath20 of figure [ fig : gompffig3b ] ( obtained by blowing down the two bracketed circles ) can be changed to the unlink by handle slides , then the dual slides in figure [ fig : gompffig1b ] will trivialize that picture , showing that the above presentation is andrews - curtis trivial . for @xmath366 , for example , this is regarded as very unlikely when @xmath361 .
( for @xmath367 , see section [ slice ] . ) since surgery on @xmath20 is @xmath368 by construction , this suggests an affirmative answer to question two , which ( for any one @xmath6 ) would imply : [ conj : not2r ] the square knot does not have property 2r .
the invariant we have implicitly invoked here can be described in a purely 3-dimensional way .
suppose @xmath7 is an @xmath6-component framed link that satisfies the hypothesis of generalized property r. then surgery on @xmath7 yields @xmath369 , whose fundamental group @xmath370 is free on @xmath6 generators .
its basis @xmath371 is unique up to nielsen moves .
if we pick a meridian of each component of @xmath7 , attached somehow to the base point , we obtain @xmath6 elements @xmath372 that normally generate @xmath370 ( since they normally generate the group of the link complement ) .
thus , we have a presentation @xmath373 of the trivial group .
changing our choices in the construction changes the presentation by andrews - curtis moves .
if we change @xmath7 by sliding one component @xmath374 over another @xmath375 , then @xmath376 ceases to be a meridian , but it again becomes one via the dual slide over a meridian to @xmath374 ( cf .
figure [ fig : dual2 ] ) .
this is again an andrews - curtis move , multiplying @xmath376 by a conjugate of @xmath377 .
thus , the link @xmath7 up to handle slides determines a balanced presentation of the trivial group up to andrews - curtis moves .
unfortunately , there is presently no way to distinguish andrews - curtis equivalence classes from each other .
when such technology emerges , it should be able to distinguish handle - slide equivalence classes of links satisfying the hypothesis of generalized property r , such as the links @xmath20 . for a related perspective , see also @xcite .
a link in @xmath378 is called _ ( smoothly ) slice _ if it bounds a collection of smooth , disjoint disks in @xmath66 .
it is called _ ribbon _ if the disks can be chosen so that the radial function on @xmath66 restricts to a morse function without local maxima on the disks .
both conditions are preserved by ( 0-framed ) handle slides , the net effect being to band - sum one disk with a parallel copy of another .
band - summing link components to create a knot similarly preserves both conditions .
slice knots are somewhat rare , and their constructions typically generate slice disks that are easily seen to be ribbon .
in fact , the slice - ribbon conjecture , asserting that every slice knot should be ribbon , is still open after more than three decades .
the corresponding question for links is also still open .
potential counterexamples to the generalized property r conjecture are a natural source of potential counterexamples to the slice - ribbon conjecture , since the conclusion of the former conjecture immediately implies a link and its band - sums are ribbon , whereas its hypothesis only guarantees that the link and its band - sums are slice in some possibly exotic smoothing of @xmath66 ( proposition [ prop : slice ] ) . in our case ,
the smoothing is standard , so we can construct slice knots with no obvious approach to proving that they are ribbon .
this method appears to be the only currently known source of potential counterexamples to the slice - ribbon conjecture ( for knots or links ) . in light of akbulut s recent work @xcite ,
more complicated potential counterexamples can also be generated by this same method ; cf .
@xcite . at 170 215 at 165 105 at 165 90 at 95 25 @xmath379 at 75 125 @xmath380 at 115 70
recall that the link @xmath20 is obtained from the right side of figure [ fig : gompffig3b ] by blowing down the red 2-component unlink .
this was done explicitly for @xmath366 to obtain figure [ fig : ln1 ] . for general @xmath347 , we at least simplify the red unlink to obtain figure [ fig : lnk ] . to interpret this diagram ,
first consider the black unlink in the figure .
the large component is the boundary of a disk in the plane that is wrapped in a spiral with @xmath347 turns .
thus , the circle runs around the spiral with @xmath381 strands , occurring in oppositely oriented pairs . to obtain @xmath20 , thread the red curves through the twist boxes as shown , then twist as indicated on the red unlink so that one component becomes the sum of torus knots @xmath382 .
( this can be done explicitly , but results in a messier picture . )
the link @xmath20 was first shown in a complicated diagram of @xmath53 , figure [ fig : gompffig1b ] , as the two meridians of the 2-handles .
these obviously bound disjoint disks in the pictured 4-manifold , that are the cores of the 2-handles in the dual picture , explicitly exhibiting the proof of proposition [ prop : slice ] in our case .
since the pictured 4-manifold is actually diffeomorphic to a 4-ball @xcite , it follows that @xmath20 is smoothly slice .
the authors do not know whether @xmath20 is ribbon except in the special cases @xmath383 or @xmath384 or @xmath385 .
( we continue to assume @xmath386 without loss of generality , since the symmetry of figure [ fig : gompffig1b ] , @xmath326-rotation about a horizontal axis in the plane of the paper , interchanges @xmath6 and @xmath387
. however , it may be useful to consider @xmath388 . ) in each of the special cases , the corresponding presentation is andrews - curtis trivial .
( this is clear except for the last case , which is due to gersten @xcite and reproduced for convenience at the end of this section . ) in each case , the algebra can be followed by a trivialization of the handlebody figure [ fig : gompffig1b ] that introduces no new handles , and the dual computation shows that the corresponding @xmath20 satisfies the conclusion of generalized property r. thus , the special cases , and any knots created from them by band - sum , are ribbon . in principle , the ribbon disks can be explicitly determined from the construction , although the details may be laborious .
in contrast , the remaining cases are mysterious . in general ,
each component of @xmath20 is ribbon : a band move turns the small component in figure [ fig : gompffig3b ] into a meridian of each of the bracketed circles , becoming an unlink when these are blown down .
the corresponding move in figure [ fig : lnk ] is to cut through the spiral disk .
the other component is ribbon since it is the sum of a knot with its mirror .
unfortunately , the individual ribbon disks seem to interfere with each other .
one could try to trace the given slice disks through the construction , but it seems likely that the 2 - 3 handle pair needed for canceling the handlebody creates troublesome local maxima . on the other hand , the given slice disks have the unusual property that their complement is a regular neighborhood of a wedge of two circles ; perhaps there is a pair of ribbon disks whose complement has more complicated fundamental group . in summary
, we have the following : _ questions : _ for @xmath389 , @xmath390 and @xmath391 , is @xmath20 a ribbon link ?
are the slice knots made by band - summing its components always ribbon ?
are they ever ribbon ?
a simple example with @xmath392 is given in figure [ fig : sliceknot ] .
gersten trivializes the presentation @xmath393 corresponding to @xmath394 by multiplying the first relator by @xmath395,\ ] ] obtaining ( after conjugation ) @xmath396 . replacing the generator @xmath397 by @xmath398 changes the new relator to @xmath399 , killing @xmath400 and hence @xmath401 .
the evidence of section [ sect : nonstandard ] suggests that property 2r , and hence the generalized property r conjecture , may fail for a knot as simple as the square knot .
there are other reasons to be unhappy with this set of conjectures : for one thing , there is no clear relation between property nr and property ( n+1)r : if @xmath5 has property nr , there is no reason to conclude that no @xmath402-component counterexample to generalized property r contains @xmath5 .
after all , gabai has shown that _ every knot has property 1r_. conversely , if @xmath5 does not have property nr , so there is an @xmath6-component counterexample @xmath7 to generalized property r , and @xmath403 , it may well be that even adding a distant @xmath2-framed unknot to @xmath7 creates a link that satisfies generalized property r. so @xmath5 might still satisfy property ( n+1)r .
a four - dimensional perspective on this last possibility is instructive .
( see also ( * ? ? ?
* section 3 ) . )
suppose @xmath7 is an @xmath6-component link on which surgery gives @xmath51 .
suppose further that , after adding a distant @xmath2-framed @xmath404-component unlink to @xmath7 , the resulting link @xmath405 can be reduced to an @xmath406-component unlink by handle - slides .
consider the closed @xmath22-manifold @xmath67 obtained by attaching @xmath0-handles to @xmath345 via the framed link @xmath7 , then attaching @xmath71 to the resulting manifold along their common boundary @xmath51 . via @xcite
we know there is essentially only one way to do this .
the result is a simply - connected ( since no @xmath74-handles are attached ) homology @xmath22-sphere , hence a homotopy @xmath22-sphere .
if the @xmath0-handles attached along @xmath7 can be be slid so that the attaching link is the unlink , this would show that @xmath407 , since it implies that the @xmath0-handles are exactly canceled by the @xmath23-handles .
but the same is true if handle - slides convert @xmath405 to the unlink : the @xmath22-manifold @xmath67 is the same , but first attaching @xmath404 more trivial @xmath0-handles and then canceling with @xmath404 more @xmath23-handles has no effect on the topology of the associated @xmath22-manifold .
so from the point of view of @xmath22-manifolds , this weaker form of generalized property r ( in which a distant @xmath404-component unlink may be added to the original link ) would suffice . from the @xmath22-manifold point of view
there is a dual operation which also makes no difference to the topology of the underlying @xmath22-manifold : adding also pairs of canceling @xmath74- and @xmath0- handles . from the point of view of kirby calculus , each such pair is conventionally noted in two possible ways ( see section 5.4 of @xcite and figure [ fig : canceling ] below ) : * a dumb - bell shaped object .
the ends of the dumb - bell represent @xmath23-balls on which a @xmath74-handle is attached ; the connecting rod represents the attaching circle for a canceling @xmath0-handle . * a hopf link .
one component of the link is labeled with a dot and the other is given framing @xmath2 .
the dotted component represents the @xmath74-handle and the @xmath2-framed component represents the attaching circle for a canceling @xmath0-handle .
call such an object a canceling hopf pair and call the union of @xmath408 such pairs , each lying in a disjoint @xmath23-ball , a _ set of @xmath408 canceling hopf pairs_. ( the rules for handle - slides of these dotted components in canceling hopf pairs are fairly restricted ; see @xcite .
for example , they can not slide over non - dotted components .
note that while we require the 2-handle of a canceling hopf pair to have framing 0 , we can change it to any even framing by sliding the 2-handle over the dotted circle to exploit the nonzero linking number . ) @xmath2 at 160 650 @xmath2 at 510 650 the @xmath22-manifold perspective suggests the following weaker and more awkward version of generalized property r : [ conj : weakgenr ] suppose @xmath7 is a framed link of @xmath52 components in @xmath53 , and surgery on @xmath7 yields @xmath51 .
then , perhaps after adding a distant @xmath404-component @xmath2-framed unlink and a set of @xmath408 canceling hopf pairs to @xmath7 , there is a sequence of handle - slides that creates the distant union of an @xmath409 component @xmath2-framed unlink with a set of @xmath408 canceling hopf pairs .
the move to weak generalized property r destroys the andrews - curtis invariant of the previous section , since adding a canceling hopf pair leaves the surgered manifold , and hence @xmath370 , unchanged , but introduces a new relator that is obviously trivial .
in fact , we will see in the next section that the links @xmath20 are not counterexamples to this weaker conjecture ( even if we set set @xmath410 , @xmath411 ) . on the other hand , stabilization by adding a distant unknot to @xmath7 changes the corresponding presentation by adding a generator @xmath412 and relator @xmath413 , and this preserves the andrews - curtis class .
thus , the hopf pairs seem crucial to the conjecture , whereas it is unclear whether the distant unlink is significant .
( of course , if we allow 1-handles in our original diagram as in figure [ fig : gompffig1b ] , then the andrews - curtis problem seems to make the distant unlink necessary .
) a knot @xmath3 has * weak property nr * if it does not appear among the components of any @xmath6-component counterexample to the weak generalized property r conjecture .
we have seen that property nr and weak property nr are probably quite different , due to the hopf pairs .
the latter property also exhibits nicer formal behavior due to the distant unlink , in that if @xmath5 has weak property ( n+1)r then it has weak property nr : suppose @xmath5 appears among the components of an @xmath6-component framed link @xmath7 and surgery on @xmath7 via the specified framing yields @xmath51 .
then the corresponding statement is true for the union @xmath405 of @xmath7 with a distant @xmath2-framed unknot .
if @xmath5 has weak property ( n+1)r then the distant union @xmath414 of @xmath405 with the @xmath2-framed @xmath404-component unlink and a set of @xmath408 canceling hopf pairs has a sequence of handle - slides that creates the @xmath2-framed unlink of @xmath415 components and a set of @xmath408 canceling hopf pairs .
but @xmath414 can also be viewed as the union of @xmath7 with the @xmath2-framed @xmath416-component unlink and a set of @xmath408 canceling hopf pairs .
thus @xmath5 satisfies weak property nr . the weak generalized property r conjecture is closely related to the smooth ( or pl ) 4-dimensional poincar conjecture , that every homotopy 4-sphere is actually diffeomorphic to @xmath342 .
for a precise statement , we restrict attention to homotopy spheres that admit handle decompositions without 1-handles .
the weak generalized property r conjecture is equivalent to the smooth 4-dimensional poincar conjecture for homotopy spheres that admit handle decompositions without 1-handles . while there are various known ways of constructing potential counterexamples to the smooth 4-dimensional poincar conjecture , each method is known to produce standard 4-spheres in many special cases .
( the most recent developments are @xcite , @xcite . )
akbulut s recent work @xcite has eliminated the only promising potential counterexamples currently known to admit handle decompositions without 1-handles . for 3-dimensional renderings of the full smooth 4-dimensional poincar conjecture and other related conjectures from 4-manifold theory , see @xcite .
suppose @xmath417 is a homotopy sphere with no 1-handles and @xmath6 @xmath0-handles .
since there are no @xmath74-handles , the @xmath0-handles are attached to some framed @xmath6-component link @xmath73 in the boundary of the unique @xmath2-handle @xmath345 in @xmath417 . since @xmath417 has euler characteristic 2 , there are @xmath6 3-handles attached to the resulting boundary , showing that surgery on @xmath7 is @xmath369
. if the weak generalized property r conjecture is true , there are some @xmath418 so that when a distant @xmath2-framed @xmath404-component unlink and a distant set of @xmath408 canceling hopf pairs is added to @xmath7 ( call the result @xmath405 ) then after a series of handle slides , @xmath405 becomes the distant union of an @xmath409 component unlink and a set of @xmath408 canceling hopf pairs . to the given handle decomposition of @xmath417 ,
add @xmath404 copies of canceling @xmath0- and @xmath23- handle pairs and @xmath408 copies of canceling @xmath74- and @xmath0-handle pairs .
after this change of handle decomposition , the manifold @xmath419 that is the union of @xmath2- , @xmath74- and @xmath0-handles can be viewed as obtained by the surgery diagram @xmath420 .
after a sequence of handle - slides , which preserve the diffeomorphism type of @xmath419 , @xmath405 is simplified as above ; it can be further simplified by canceling the @xmath74- and @xmath0-handle pairs given by the set of @xmath408 canceling hopf pairs .
what remains is a handle description of @xmath419 given by @xmath2-framed surgery on an unlink of @xmath409 components .
since @xmath417 has euler characteristic 2 , it is obtained by attaching exactly @xmath406 @xmath23-handles and then a single @xmath22-handle to @xmath419 .
if we view @xmath419 as obtained by attaching @xmath0-handles via the @xmath2-framed @xmath406-component unlink , there is an obvious way to attach some set of @xmath406 @xmath23-handles so that they exactly cancel the @xmath0-handles , creating @xmath342 .
it is a theorem of laudenbach and poenaru @xcite that , up to handle - slides , there is really only one way to attach @xmath406 @xmath23-handles to @xmath419 .
hence @xmath417 is diffeomorphic to @xmath342 as required .
conversely , if @xmath2-framed surgery on an @xmath6-component link @xmath73 gives @xmath51 then a smooth homotopy @xmath22-sphere @xmath417 can be constructed by attaching to the trace of the surgery @xmath6 @xmath23-handles and a @xmath22-handle
. if @xmath417 is @xmath342 , then standard cerf theory says one can introduce some canceling @xmath74- and @xmath0- handle pairs , plus some canceling @xmath0- and @xmath23- handle pairs to the given handle description of @xmath417 , then slide handles until all handles cancel . but
introducing these canceling handle pairs , when expressed in the language of framed links , constitutes the extra moves that are allowed under weak generalized property r. ( to see that the framings in the canceling hopf pairs can be taken to be 0 , first arrange them to be even by choosing the diagram to respect the unique spin structure on @xmath417 , then slide as necessary to reduce to the 0-framed case . )
with the weak generalized property r conjecture in mind , return now to the square knot example of section [ sect : nonstandard ] .
in @xcite it is shown that the introduction of a canceling @xmath421-handle pair does make figure [ fig : gompffig1b ] equivalent by handle - slides to the corresponding canceling diagram . in our dualized context , that means that the introduction of a canceling @xmath422 handle pair ( in our terminology , a canceling hopf pair ) should allow @xmath423 ( or more generally any @xmath20 ) to be handle - slid until the result is the union of the unlink and a canceling hopf pair .
in particular , the square knot examples provide no evidence against the weak generalized property r conjecture .
it is a bit mysterious how adding a hopf pair can help to simplify links .
since the dotted circle is not allowed to slide over the other link components , it may seem that any slides over its meridian would only need to be undone to remove extraneous linking with the dotted circle . in this section
, we give a general procedure that exploits a nontrivial self - isotopy of the meridian to create families of potential counterexamples to generalized property r. the examples @xmath20 of section [ sect : nonstandard ] arise as a special case for each fixed @xmath347 .
this suggests that generalized property r probably fails frequently for links constructed by this method , whereas weak generalized property r does not .
begin with an oriented 3-manifold @xmath24 containing a framed link @xmath424 .
we will construct an infinite family of framed links in @xmath24 that may not be handle - slide equivalent , but become so after a hopf pair is added .
( in our main examples , @xmath24 will be @xmath53 with @xmath7 a 0-framed 2-component link . )
let @xmath28 be the manifold made from @xmath24 by surgery on @xmath405 .
let @xmath425 be a generic immersion that is an embedding on each fiber @xmath426 , @xmath427 .
denote the image circle @xmath428 by @xmath429 .
since @xmath430 is generic , it is transverse both to @xmath414 and to the cores of the surgery solid tori in @xmath28 . in particular ,
* only finitely many circles @xmath431 intersect @xmath414 * each such circle @xmath431 intersects @xmath414 in exactly one point * the curves @xmath431 are pairwise disjoint and * each @xmath431 lies in the link complement @xmath432 . the immersion @xmath430 determines an infinite family of framed links @xmath433 in @xmath24 , @xmath434 : for each @xmath435 , there is an annulus @xmath436 in @xmath437 centered on @xmath431 , transverse to @xmath430 near @xmath438 and intersecting @xmath414 in an arc , with @xmath436 oriented so that its positive normal direction corresponds to increasing @xmath439 .
see figure [ fig : ai ] , where the left and right edges of each diagram are glued together .
the annuli @xmath436 are unique up to isotopy , and we can assume they are pairwise disjoint . let @xmath440 be the framed link obtained from @xmath414 by diverting each arc @xmath441 so that it spirals @xmath442 times around in @xmath436 ( relative to its original position ) as in figure [ fig : ai ] , starting with a right turn if @xmath386 .
( thus @xmath443 . )
we specify the framings by taking them to be tangent to @xmath436 everywhere inside these annuli ( before and after ) and unchanged outside .
note that @xmath444 depends on our initial choices of how to isotope the curves @xmath431 off of the surgery solid tori , but changing the choices only modifies @xmath444 by handle slides of @xmath440 over @xmath405 . in our main examples below ,
@xmath430 is an embedding .
whenever this occurs , there is a self - diffeomorphism of @xmath28 sending @xmath414 onto @xmath440 , namely the @xmath6-fold dehn twist along @xmath447 parallel to the circles @xmath429 .
this shows directly in this case that the manifolds obtained by surgery on @xmath444 are all diffeomorphic , regardless of @xmath6 . in general
, we can obtain more specific information : [ prop : ln ] for @xmath448 and @xmath430 a generic immersion as above , suppose that @xmath449 is disjoint from all other circles @xmath429 and bounds an embedded disk in @xmath24 whose algebraic intersection with nearby curves @xmath429 is @xmath296 .
then all of the framed links @xmath444 become handle - slide equivalent after the addition of a single canceling hopf pair .
since @xmath449 is nullhomologous in @xmath24 , it has a canonical longitude , determined by any seifert surface in @xmath24 .
thus , the last hypothesis can be restated by saying that @xmath449 is an unknot in @xmath24 for which the immersion @xmath430 determines the @xmath296 framing . to see that the hypotheses are not unreasonable , note that they are satisfied for @xmath450 , @xmath405 empty and @xmath449 a @xmath451-curve in the standard torus @xmath58 . in this case , the links @xmath452 are obviously all equivalent for any fixed @xmath414 .
( since @xmath58 is embedded , the links @xmath452 are all related as above by dehn twisting along @xmath58 in @xmath453 .
these dehn twists are isotopic to the identity by an isotopy fixing one complementary solid torus of @xmath58 and twisting the other by both of its circle actions . )
the more interesting examples @xmath20 that arise in @xcite will be exhibited below .
these all arise from the same general procedure : start with a 0-framed unlink in @xmath53 .
slide handles to create a more complicated link @xmath7 satisfying generalized property r. locate a suitable immersion @xmath430 , then apply proposition [ prop : ln ] to create an infinite family @xmath454 of links that still satisfy the hypothesis of generalized property r , but not necessarily the conclusion for most @xmath6 .
one can generalize still further , for example replacing the domain of @xmath430 by a klein bottle , although the authors presently have no explicit examples of this .
since @xmath430 is an immersion and @xmath449 is disjoint from all other @xmath429 , there is a neighborhood of @xmath449 in @xmath24 intersecting @xmath447 in an annulus .
let @xmath455 be a circle in this neighborhood , parallel to @xmath449 and disjoint from @xmath447 .
thus , it is pushed off of @xmath449 by the @xmath296 framing in @xmath24 .
( all signs in the proof agree with the one in the statement of the theorem . )
let @xmath456 be the framed link obtained from @xmath444 by adding a @xmath457-framed component along @xmath449 and a dotted circle along @xmath455 .
we will show that @xmath458 is obtained from @xmath444 by adding a canceling hopf pair and sliding handles , and that the links @xmath458 are all handle - slide equivalent .
( as usual , the dotted circle is not allowed to slide over the other link components . )
this will complete the proof .
to reduce @xmath459 to @xmath444 , slide @xmath449 over the dotted circle @xmath455 using the simplest band connecting the parallel circles .
this changes @xmath449 to a 0-framed meridian of @xmath455 .
( to compute the framing , recall that the dotted circle is automatically 0-framed , and that if @xmath449 and @xmath455 have parallel orientations , we are subtracting handles with linking number @xmath296 . )
the dotted circle @xmath455 bounds a disk @xmath460 and now has a 0-framed meridian , but this is not yet a canceling hopf pair since the link @xmath7 may still intersect @xmath188 .
however , we can easily remove each intersection by sliding the offending strand of @xmath7 over the meridian .
this does not change @xmath7 but liberates the hopf pair as required . to see that the links @xmath458 are handle - slide equivalent
, we wish to isotope their common component @xmath352 at @xmath449 around the continuous family @xmath429 , returning to the original position at @xmath461 . by construction
, @xmath352 will never meet the dotted circle at @xmath455 , but it will meet components of @xmath444 . since the isotopy lies in @xmath28 , each encounter with @xmath405 consists of a push across a surgery solid torus in @xmath28 , which is precisely a handle slide in @xmath24 of @xmath352 over a component of @xmath405 . on the other hand , @xmath452 transversely intersects the path of @xmath352 , once at each annulus @xmath436 , so we must somehow reverse the crossing of @xmath352 with @xmath440 there . since the framing of @xmath352 has @xmath296 twist relative to @xmath430 , and hence to @xmath436 , we can reverse the crossing by sliding @xmath440 over @xmath352 .
this is shown by figure [ fig : cross ] ( or its mirror image if the sign is @xmath346 ) , where @xmath436 is drawn as in figure [ fig : ai ] , but after a diffeomorphism so that @xmath440 appears as a vertical line .
( to see that the framing transforms as drawn , note that the right twist in the framing of the curve parallel to @xmath352 cancels the left twist introduced by the self - crossing added to @xmath440 by the band - sum . )
when @xmath352 returns to its original position at @xmath461 , the link has been transformed to @xmath462 by handle slides , completing the proof .
the examples @xmath20 from section [ sect : nonstandard ] form a family as in proposition [ prop : ln ] for each fixed @xmath347 , generated by the handle - slide trivial case @xmath464 .
figure [ fig : gompf6 ] shows the starting point .
it is basically the right side of figure [ fig : gompffig3b ] with both a twist - box and one of the 0-framed links moved to a more symmetric position .
let @xmath24 be the manifold diffeomorphic to @xmath53 obtained by surgery on the bracketed ( red ) circles .
( recall that these form an unlink in @xmath53 . )
let @xmath405 and @xmath452 be the small and large 0-framed circles , respectively .
there is a vertical plane , perpendicular to the page , that bisects the picture .
this plane bisects both @xmath465 twist boxes and contains @xmath405 .
we interpret this plane as a 2-sphere in the background @xmath53 .
surgery on @xmath405 splits the sphere into a disjoint pair of spheres .
neither of these lies in @xmath28 since each is punctured by both bracketed circles .
however , we can remove the punctures by tubing the spheres together along both bracketed curves , obtaining an embedded torus in @xmath28 .
the fibers @xmath429 of this torus include a meridian of each bracketed circle , and the torus induces the 0-framing on these relative to the background @xmath53 .
to see these in @xmath24 , ignore the link @xmath444 and unwind the bracketed link .
blowing down reveals the 0-framed meridian of the @xmath466$]-framed curve to be a @xmath296-framed unknot in @xmath24 ( see ( * ? ? ?
* figure 5.18 ) ) .
thus , we can choose either meridian to be @xmath449 , with framing @xmath457 in @xmath24 , i.e. , @xmath296 in @xmath53 .
( in either case , the framing in @xmath53 is preserved as we travel around the torus , so it becomes 0 in @xmath24 when we reach the other meridian , showing that we can not get off the ride in the middle . ) since @xmath440 intersects the torus at two points , at the @xmath465-twist boxes , it is easy to verify that the procedure turns @xmath458 into @xmath467 , with the sign depending both on the choice of @xmath449 and the direction of motion around the torus
. it is instructive to view the procedure explicitly as a sequence of four handle slides in figure [ fig : gompf6 ] .
it is not completely obvious how to trivialize the link by handle slides when @xmath464 . in the @xmath366 case , figures [ fig : gompffig4 ] and [ fig : gompffig5 ] isotope the link to figure [ fig : squareknot ] , solving the problem with a single handle slide . for general @xmath347 ,
it is helpful to begin from a somewhat simpler picture of @xmath20 , figure [ fig : example ] , which is obtained from figure [ fig : gompf6 ] by an isotopy that rotates both @xmath468 twist boxes 180 degrees outward about vertical axes .
now set @xmath469 and consider the two handle slides in figure [ fig : example2 ] ( following the arrows , framed by the plane of the paper ) .
the slides first complicate @xmath405 , but after an isotopy shrinking the strands shown in green , @xmath405 simplifies , so that a slide over the @xmath350$]-framed circle changes it to a meridian of the @xmath354$]-framed circle .
the @xmath468-twist boxes then cancel , showing that the link is trivial .
this construction was derived from the dual construction in @xcite , showing that the relevant homotopy balls are standard by introducing a canceling 2-handle/3-handle pair . while it was not easy to explicitly dualize the construction , in the end the two constructions appear rather similar .
the construction in @xcite also involves introducing a new link component and moving it around a torus . in that case , the new component was a 0-framed unknot , the attaching circle of the new 2-handle .
akbulut s recent proof @xcite that infinitely many cappell - shaneson homotopy 4-spheres are standard again relies on such a trick . in each case
, the key seems to be an embedded torus with trivial normal bundle , to which a 2-handle is attached with framing @xmath296 , forming a _
fishtail neighborhood_. this neighborhood has self - diffeomorphisms that can undo certain cut - and - paste operations on the 4-manifold , as we implicitly saw in the case of proposition [ prop : ln ] with @xmath430 an embedding .
while this procedure has been known in a different context for at least three decades , underlying the proof that simply connected elliptic surfaces with fixed @xmath470 are determined by their multiplicities , it still appears to be underused .
in fact , one can obtain a simpler and more general proof that many cappell - shaneson 4-spheres are standard by going back to their original definition and directly locating fishtail neighborhoods @xcite . | if there are any @xmath0-component counterexamples to the generalized property r conjecture , a least genus component of all such counterexamples can not be a fibered knot .
furthermore , the monodromy of a fibered component of any such counterexample has unexpected restrictions .
the simplest plausible counterexample to the generalized property r conjecture could be a @xmath0-component link containing the square knot .
we characterize all two - component links that contain the square knot and which surger to @xmath1 .
we exhibit a family of such links that are probably counterexamples to generalized property r. these links can be used to generate slice knots that are not known to be ribbon . |
the formation and evolution of massive early - type galaxies remains poorly understood , despite much recent progress .
constraints from the local universe indicate that their stellar ages are very old ( @xmath510 gyr ) , indicating that they formed the bulk of their stellar masses at z@xmath52 @xcite , subsequently quenching star - formation , and remaining quenched , until the present .
constraints on stellar abundance ratios ( high @xmath6/fe ) indicate their stars formed in rapid bursts of star - formation @xcite .
additionally , it has been observed that galaxy morphology and star - formation properties are highly correlated , such that this quenched nature in massive galaxies appears coincident with morphological transformation to ellipsoidal , velocity dispersion dominated stellar structure ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
despite efforts to study this transition from star - forming galaxy to quenched ellipsoid , we have gained very little insight into both the transformational quenching process , as well as the mechanisms preventing further star - formation for the majority of the universe s history . of particular importance to this effort
are constraints from observing the progenitors of these massive early - type galaxies at z@xmath51 , shortly after their transformation from star - forming galaxy .
recently quenched galaxies ( qgs ) begin to appear in large numbers at z@xmath72 @xcite , and have been observed even out to z@xmath73 - 4 @xcite .
the properties of these quenched high - redshift galaxies provide significant insight into both the formation process of the galaxies during the star - formation phase , as well as the quenching mechanisms causing their transformation .
the most striking feature of these recently quenched galaxies at high - redshift is their stellar structure ; while already having built up a similar amount of stellar mass as their z@xmath70 counterparts , they are remarkably compact in stellar density @xcite .
the overwhelming majority of qgs ( @xmath580% ) at [email protected] have stellar densities higher than the lower 1@xmath8 of passive ( early - type ) galaxies at z@xmath70 at the same stellar mass @xcite .
additionally , they are much smaller than the majority of massive star - forming galaxies at the same epoch @xcite , and in fact , one of the strongest predictors of quiescence among high - redshift galaxies is centrally concentrated light ( i.e. a measure of compactness ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it appears , therefore , that compactness and quenched nature at high - redshift are inextricably linked .
however , the physical reason for this correlation is also poorly understood .
does the existence of the compact qgs at high - redshift imply something very fundamental about quenching ?
in other words , does some physical process associated with stellar compactness predispose galaxies to quench ? alternatively ,
are the earliest galaxies to form in the universe and complete their evolution simply the densest because the universe was denser at early times , or , because of some highly dissipative gaseous process that could take place predominantly at high redshift ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
there are some physically motivated reasons to believe that the former may be true .
first , high stellar density implies a previous epoch of high surface density of star - formation , which would mean a higher energy input into the interstellar medium ( ism ) of compact galaxies , than might be present in larger , extended galaxies .
@xcite have made this argument , based on the observation that there appears to be a maximum stellar density of _ any _ structure in the universe ( @xmath9 10@xmath10 m@xmath11 kpc@xmath12 ) .
this limit in stellar density exists despite covering 8 orders of magnitude in stellar mass , from star clusters within galaxies , to the z@xmath52 compact qgs .
this empirical limit argues for some stellar feedback process , such as massive stellar winds , that truncate the star - formation and prevent further growth beyond this density limit .
studies of objects with high surface density of star - formation , where such extreme stellar feedback might be expected , have in fact found evidence of feedback in the form of very fast ( @xmath71000 km / s ) galactic - scale outflows from extremely compact star - forming regions that approach the eddington limit @xcite .
the second , alternative scenario exists , that this connection is a very simple consequence of structure formation in the @xmath13cdm paradigm .
the most massive galaxies in the early universe have formed the earliest in cosmic time , and therefore evolve to the end of their star - forming lifetime earliest . in this scenario ,
the density of the parent halo of the quenched population at any redshift reflects the density of the universe at its formation epoch ( e.g. * ? ? ?
* ) , and therefore will progressively increase in size ( and therefore stellar density ) over time .
such a scenario , known as progenitor bias , would also contribute to the increasing size evolution of qgs over cosmic time @xcite ( see also * ? ? ?
in this scenario , the significance of compactness is irrelevant for the quenching , and rather , galaxies quench when they ve reached sufficient mass that they no longer support star - formation .
the quenching mechanism then may be related to halo mass , or some other mass related mechanism to cut of gas supply for future star - formation ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . distinguishing between these scenarios is highly important to understanding the evolution of early - type galaxies .
each of these two scenarios have empirical predictions for the properties of qgs .
in the stellar - density regulated star - formation scenario , galaxies with high enough surface density of star - formation will quench , and produce remnants with high stellar density .
thus at any given epoch , the most recently quenched objects should also be the densest ( e.g. * ? ? ?
there is no explicit prediction for a trend of stellar density with stellar age or mass ( e.g. * ? ? ?
however , progenitor bias explicitly predicts that stellar age , and stellar density be correlated , with the densest objects also being the oldest at any given epoch ( in the absence of size growth via merging ; * ? ? ?
* ) . in this paper , we seek to distinguish between these two scenarios , and in the process , gain insight into why galaxies quench their star - formation early in cosmic time . in section 2
we present the data used in this study . in section 3
we present our results , and in section 4 we discuss these results in the context of quenching and the formation of the qgs . throughout this paper
we assume a cosmology with @xmath14=0.7 , @xmath15=0.3 , and h@xmath16 = 70 km / s / mpc . [ ! t ]
we select qgs from the cosmic assembly near - infrared deep extragalactic legacy survey ( candels ) data @xcite in goods - south according to the selection outlined in @xcite , which identified 179 qgs at z@xmath51 with m*@xmath510@xmath17m@xmath11 , specific sfr @xmath18 10@xmath12 gyr@xmath19 and centrally concentrated , spheroidal morphologies . for this study , we have measured the properties of this sample using the candels multi - wavelength photometry @xcite following the sed fitting procedure outlined in lee et al .
( in preparation ) using the speedymc bayesian sed - fitting software @xcite where star - formation history is left as a free parameter .
we use the morphologies measured from the hst / wfc3 f160w _ h_-band imaging from candels using galfit @xcite presented in @xcite .
in particular , we define the size to be the circularized half - light radius in kpc , r@xmath20 , where @xmath21 is the length of the semi - major axis in arc seconds converted to kpc using the spectroscopic redshift , and @xmath22 is the axis ratio . from this parent sample , we identify 61 qgs with complementary publicly available spectra ( see next section ) .
we define a compactness cut on the qg sample according to the local z@xmath70 mass - size relation for local galaxies from the sloan digital sky survey ( sdss ) .
we define qgs as compact if their size ( at a given mass ) is smaller than the lower 1-@xmath8 of the local early - type galaxy size - mass relation @xcite .
this roughly corresponds to a stellar mass surface density of @xmath93x10@xmath23m@xmath11kpc@xmath12 .
we refer to any qg with a stellar density higher than this threshold as compact .
qgs that are more extended , and thus similar to the majority of local sdss early - type galaxies in mass and size , we refer to as normal. the position of the qgs in the mass - size diagram , illustrating the compactness cut from sdss , are presented in figure [ masssize ] .
there are 28 compact galaxies and 33 normal - sized galaxies .
average redshift for the samples are @[email protected] and 1.13 for compact and normal , respectively .
our spectroscopic data for this sample was obtained at the european southern observatory s very large telescope ( vlt ) as part of the great observatories origins deep survey ( goods ; * ? ? ?
* ) spectroscopic program .
in particular , we use spectra from programs with fors2 @xcite , and vimos @xcite .
the fors2 spectra have an instrument resolution of r@xmath7660 , which is 13 at 8600 observed ( average 5.9 at 3900 restframe ; all spectra between redshifts 1 and 1.4 ) .
the spectra we use from vimos in the @xcite release are all using the medium resolution grism , which has comparable resolution . to produce average composite spectra ( stacks ) we perform the following procedure .
we first individually transform each spectrum into the rest frame using the published redshifts @xcite .
we then flux normalize each spectrum using the median flux value measured between restframe 4000@xmath254050 .
finally , we stack the normalized spectra using the _ scombine _ package in iraf , and rebin to a common dispersion ( @xmath261.4pixel@xmath19 , rest frame ) , and performed a 3@xmath8 clipping during the stack .
our stacked composite spectra of the two qg samples are presented in figure [ stack ] .
there are 6 out of the 28 compact qgs that do not have spectral coverage at red ward of rest frame 4000 , thus we have excluded them from the stack and its analysis presented in section 3.2 .
however their spectra are suitable for the analysis presented in sections 3.3 and 3.4 , which include the full sample of 28 galaxies . [ !
t ] we have estimated the error on the stacks as the sample standard deviations of each spectroscopic sample . to measure the standard deviations , we repeat the stacks of each sample using jackknife resampling , each time removing one spectrum and stacking the rest of the spectra following the identical procedure described above .
the final sample error is the standard deviation of the jackknifed stacks at each spectral point .
the stacks presented in figure [ stack ] exhibit many features typical of old stellar populations , namely strong balmer absorption typical of post - starburst galaxies , strong g - band absorption , and a prominent 4000 break .
additionally , very weak [ ] emission is occasionally present ( as we discuss later ) , indicating little ( if any ) star - formation in some cases .
we discuss the [ ] properties of the galaxies further in section 3.3 . in the following section 3.2 ,
we show the properties of the stellar populations , including average stellar population ages , in these two samples . we have measured the average age of the stellar populations of the two qg samples from their stacks , using two age diagnostics : the 4000 break ( d@xmath274000 ) , and h@xmath28 absorption . for the d@xmath274000 diagnostic ,
we have used the age calibration presented in @xcite , which takes the ratio of the mean flux between 4000 and 4100 , to the mean flux between 3850 and 3950 in the stack . we measure the error on the d@xmath274000 diagnostic by generating 1000 gaussian deviates of each spectral point in the stack using the observed sample error described in section 2.3 , and repeating the d@xmath274000 measurement each time .
the error is then the standard deviation of the sample of gaussian deviated d@xmath274000 measurements .
we find that the compact sample have a d@xmath274000 @xmath29 1.45 @xmath30 0.03 , a larger measure ( i.e. older age ) than for the normal sample , from which we have measured d@xmath274000 @xmath29 [email protected] .
we have compared these d@xmath274000 measurements to those made with single - age stellar population models using the same procedure ( figure [ kauff03 ] ; * ? ? ?
adapted from @xcite , this figure ( top left panel ) presents the evolution of this diagnostic as a function of stellar age at solar metallicity for a single instantaneous starburst of star - formation according to the stelib library ( * ? ? ?
* solid line ) , the @xcite library ( dotted line ) , and ( * ? ? ? * dashed line ) . in the top right panel ,
the age evolution of the diagnostic is presented from the stelib library for bursts of solar metallicity ( solid line ) , 20% solar ( dotted ) and 2.5 times solar ( dashed line ) .
for all models , the d@xmath274000 of the compact sample implies older stellar age than that of the normal sample . [ !
t ] similarly , we have made the comparison using the age - sensitive h@xmath28 absorption feature .
unlike the d@xmath274000 diagnostic , which grows monotonically with age , the h@xmath28 ( as well as other balmer ) absorption feature peaks for stellar populations of age @xmath71 gyr ( a - type stars ) . as the stellar population ages , the balmer absorption starts to decrease in strength .
we measure h@xmath31 using the lick absorption line index @xcite . to estimate the error on our measurement of h@xmath31 we follow the same procedure as for d@xmath274000 , by re - measuring the index from 1000 gaussian deviates of each spectral point , and taking the standard deviation of these h@xmath31 measurements .
we find that the compact sample has h@xmath31 @xmath29 0.33 @xmath30 0.31 , a smaller value ( i.e. older age ) than for the normal sample , from which we have measured h@xmath31 @xmath29 2.74 @xmath30 0.06 .
the bottom panels of figure [ kauff03 ] show how these measured values compare to those from the evolutionary models explored in @xcite .
although the trends between the age diagnostics and the stellar population age are obvious for each of the models shown in figure [ kauff03 ] , it is also obvious that the diagnostics depend on other features of the models such as metallicity . therefore , we do not attempt to use these measurements as average age measurements of the galaxy populations . rather , we seek to gain insight into _ relative _ age differences between the two samples using the few models presented in figure [ kauff03 ] . at face value
, the average age implied by the 6 models from d@xmath274000 suggests an age difference between the two galaxy samples of roughly @xmath70.3 gyr . the h@xmath31 index suggests a larger age difference , with the compact sample are @xmath72.5 gyr older than the normal qg sample ( we discuss in section 3.2.2
why the difference implied by h@xmath31 is likely overestimated ) .
however , it is clear that both age indicators imply that the compact sample , on average , have older stellar ages than the normal sample . in the literature , it has been extensively discussed that other spectral features adjacent to the h@xmath28 absorption line can affect the measurement of the h@xmath31 lick index ( e.g. * ? ? ?
* ; * ? ? ?
. in particular , stellar continuum absorption by molecular ch and cn lines to the blue and red of h@xmath28 can depress the pseudo continuum regions used to measure the index , resulting in an underestimated value for the index .
the net consequence is an overestimated age based on the line .
although the nature of this continuum absorption is not well understood , it is likely originating in old stellar populations , from metal enriched cool stars @xcite . as can be seen in figure [ stack ] , we observe the compact qg stack to exhibit a prominent peak immediately blue ward of the h@xmath28 line , followed by a band of continuum absorption between 4100 - 4200 .
the blue peak is an indication of the true continuum level , unaffected by the cn absorption .
such prominent features are not seen in the normal qg stack .
this feature is undoubtedly affecting the h@xmath31 measurement in the compact sample , such that the implied age is overestimated ( and , therefore , the relative age difference as well ) .
however , this evidence of old , metal enriched stellar populations in the compact sample is in general agreement with the evidence for an older average stellar age than the normal qg sample .
we discuss further constraints on this age differential between the two samples in the following section . [ !
t ] an alternative estimate for the age differential between the two populations can be obtained by fitting stellar population templates to the stacks . in this section ,
we use the penalized pixel - fitting ( ppxf ) software @xcite to explore the average ages of the two galaxy samples , again with the goal of gaining insight into _ relative _ age differences . as templates for the fitting
, we use the @xcite ssp models based on the medium resolution int library of empirical spectra ( miles ; sanchez - blazquez et al .
due to the poorly understood nature of the cn and ch stellar continuum absorption features , we choose to use these templates because they are composed of empirical stellar spectra where these features are observed ( e.g. * ? ? ?
we allow ppxf to choose the optimal combination of templates , along with multiplicative and additive polynomials , to fit the continuum shape , to fit each stacked spectrum , where the initial mass function ( imf ) is fixed with a slope of 1.3 @xcite , and we constrain the maximum age of any template to be the age of the universe at z=1.2 .
we allow ppxf to choose from a range of templates of varying ages , which are weighted and summed together to construct the best fitting model .
the signal to noise of our stacks starts to deteriorate at wavelengths longer than rest - frame 4800 from a combination of decreased number of spectra with wavelength coverage in that region , and possibly poor subtraction of telluric features , and so we constrain the spectral region provided to ppxf to 3500 - 4800 , masking out 40 surrounding the [ ] @xmath323727 emission . to choose the appropriate metallicity range for the fitting , we use constraints from the mass - stellar metallicity relation out to [email protected] @xcite at our average stellar masses ( @xmath3310.7 for both samples ) . only the solar metallicity models in (
* using the padova+00 isochrones of @xcite ) fall within the confidence intervals of the @xcite mass - metallicity relation at this average mass .
therefore , we limit our fits to these models .
ppxf measures the luminosity weighted age based on the sum of 2 templates , which assumes galaxies may be composed of several stellar populations .
those individual populations may not necessarily follow the mass - stellar metallicity relation .
however , we note that our results do not strongly depend on either the wavelength range or metallicity assumed by the fit .
we estimate errors on the age of the stellar populations given by the fit as the variance of a series of bootstrap resampled templates added to the fit residuals .
we find that the compact stack has an average age of [email protected] gyr ( reduced @xmath34=15.3 ) , while the normal sample are [email protected] gyr ( reduced @xmath34=30.2 ) , an age differential that is in general agreement with the findings using the lick indices in section 3.2.1 in that the age measured from the compact sample is older .
these results are shown in figure [ ppxf ] . to conclude this section , our results from both the spectral indices and the full spectral fitting _ consistently _
indicate that there is evidence that the average age of the compact qg sample may be older than that of the normal qg sample .
t ] [ !
t ] in the stacked spectra presented in figure [ stack ] , we find that each qg sample exhibits weak [ ] @xmath323727 emission .
we have visually inspected each of the individual spectra in each sample , and found that only a fraction of the individual spectra exhibit detected [ ] emission , and in general it always appears weak .
the right panel of figure [ oiifluxdist ] shows estimated sfr from the [ ] flux , assuming that the emission is produced entirely by star - formation @xcite .
the [ ] flux was measured from the individual spectra using the iraf routine _
splot_. the [ ] is always weak ( corresponding to @xmath352 m@xmath11 yr@xmath19 ; for the majority of the sample this is less than predicted from the photometry from the best - fitting sed ) .
it is therefore unlikely that this [ ] emission is contributed to significantly by an agn ( see section 3.4 for additional constraints ) .
there is no obvious difference in the line luminosities or sfrs between the two samples ( figure [ oiifluxdist ] ) .
however , the occurrence of [ ] emission is much more frequent among the normal qg sample than the compact sample ( top panel of figure [ masssize ] , and left panel of figure [ oiifluxdist ] ) .
this is the primary reason the stronger average emission in the normal qg stack . out of 33 normal qgs , roughly half ( 15 )
galaxies exhibit detected [ ] emission . among the 28 compact qgs only 6
have detected [ ] ( 21% ) .
we conduct the following analyses to assess the significance of the [ ] detection rate among qgs of different stellar density .
a priori we do not have reason to believe [ ] emission should depend on compactness , and so we test the hypothesis that the chance of [ ] emission is random ; i.e. there is an equal chance ( 50% ) that any galaxy emits [ ] as there is it does nt . to test this hypothesis
, we conduct a monte carlo simulation where we randomly draw samples equal to the number in each qg sample from a binomial distribution .
we create 10,000 realizations , and compare the success rate ( i.e. an [ ] detection ) to the observed [ ] detection rate in each qg sample .
the results of this test are presented in figure [ oii ] .
the histograms in the top panel show the frequency of [ ] detection from a sample of size 33 ( normal qgs ; blue ) and that of a sample of size 28 ( all compact qgs ; red ) .
the dotted lines show the standard deviations of the realizations .
the solid lines show the observed frequency ; it is clear from the figure that the normal qgs show an [ ] frequency that is consistent with this hypothesis that [ ] detection is random ( 50% detection rate ) . in contrast , the compact qgs show an [ ] frequency significantly lower , inconsistent at the 3-@xmath8 level .
whatever energy source produces the [ ] emission in qgs , it appears to be significantly less frequent in compact galaxies than in the normal sample . [ !
th ] the x - ray properties of the two qg samples provides further insight into residual energy sources in these galaxies , which may or may not be related to the quenching process that shut down the star - formation , or , the origin of the [ ] emission when present .
the majority of our qgs are undetected ( 27 out of 28 compact qgs ; 25 out of 33 normal qgs ) in the _ chandra _ 4 ms data in goods - south @xcite . for the x - ray counterpart identification
, we have used the catalog by @xcite , which is able to include significantly fainter x - ray sources than blind detections ( e.g. * ? ? ?
* ) , due to a novel technique based on prior information of the positions of optical - nir sources .
the x - ray detected galaxies are identified by boxes in figure [ masssize ] , where we have also identified those with significantly detected [ ] emission .
we note that among the compact sample , _ none _ of the [ ] detections come from an x - ray detected galaxy .
the one x - ray detection exhibits no [ ] emission . in the normal sample ,
a large fraction ( 6 out of 15 ) of the [ ] detections come from an x - ray detected galaxy .
the majority of x - ray detections ( 6 out of 8 x - ray detections ) are [ ] emitters .
we present the x - ray fluxes , hardness ratios , and [ ] luminosities for all x - ray detected sources in table [ xdet ] . in general , it does not appear that [ ] detected galaxies exhibit any obvious difference in their x - ray properties from those without [ ] emission .
similarly , there does not appear to be an obvious distinction in x - ray properties between the one detected compact qg and the normal qg detections ( see also figure [ xray ] )
. however , there is some indication that [ ] luminosity in the normal sample may increase with increasing x - ray luminosity ( see right panel of figure [ xray ] ) .
this may indicate a relationship between the physical process driving the [ ] emission and the higher x - ray luminosity in those galaxies where both are present , however , we note that the dynamic range in [ ] luminosity in our sample is relatively small .
we also plot upper limits to the [ ] line luminosity for x - ray detected galaxies where no [ ] emission was detected . to estimate the upper limits to the [ ] line flux , we first measure the rms in the continuum in the vicinity of the line , and multiply by the square root of the number of pixels per resolution element ( 4 and 5 pixels for fors2 and vimos , respectively ) .
we measure hardness ratios from hard and soft band fluxes for the detections as ( hard - soft)/(hard + soft ) , where the fluxes are in counts / sec . in general , agn , depending on type , can span a range of hardness ratios ( e.g. * ? ? ?
* ) , as do our detected qgs , although the luminosities of our qgs are relatively weak for agn @xcite .
the left panel of figure [ xray ] shows that the hardness ratio of our sample increases with hard x - ray luminosity , indicating that in general an increase in hardness ratio in these samples is driven by the increase in hard - band luminosity . for comparison we have included all x - ray detections in the catalog of @xcite , where detections have been gray - scaled according to their photometric redshift ( or spectroscopic redshift , if available ) . by number ,
the x - ray detections from @xcite are dominated by agn , but at the low - luminosity end , star - forming galaxies are abundant . for the non - detected sources , we stack the x - ray images at the position of the optical - nir sources to gain insight into the average x - ray properties of the two samples using the _ chandra _ stacking analysis tool cstack v4.3 .
detected galaxies were excluded from the stack .
we have removed one compact qg from the stack due to its proximity to a very off - axis , bright x - ray source , where we suspect that leaking flux after source removal using the psf has affected the counts in our stack .
the results of the stacking analysis is presented in table [ xstack ] .
we find significant stacked flux in the soft band for both samples , but both samples are essentially undetected in the hard band .
the stacked soft fluxes for both samples do not differ significantly from each other .
the luminosities , assuming the average redshifts in each sample , are very low level , inconsistent with the presence of powerful agn .
the stacked soft - band flux ( and lack of hard - band detection ) , which represent the average of the majority of our samples , seem to indicate that the x - ray emission is soft and may be likely due to stellar or low ionization nebular emission regions ( liner ; * ? ? ?
* ) emission sources .
llllllll galaxy sample & soft flux ] ] & soft luminosity ] & hard flux & hard luminosity & hardness ratio & [ ] luminosity ] ] & + compact & 1.90e-17@xmath30 5.94e-18 & 8.00e+40 & 4.39e-17@xmath30 3.01e-17 & 1.84e+41 & -0.38 & -99 + normal & 8.90e-18@xmath30 1.37e-17 & 2.22e+40 & 1.04e-15@xmath30 2.30e-16 & 2.59e+42 & 0.92 & -99 + & 5.39e-17@xmath30 1.67e-17 & 7.34e+40 & 2.17e-16@xmath30 9.47e-17 & 2.95e+41 & -0.12 & 1.63e+07 + & 4.35e-17@xmath30 1.71e-17 & 6.22e+40 & 2.41e-16@xmath30 1.17e-16 & 3.45e+41 & 0.04 & 1.80e+07 + & 2.35e-17@xmath30 7.79e-18 & 3.26e+40 & 1.53e-16@xmath30 5.17e-17 & 2.13e+41 & 0.12 & -99 + & 5.47e-18@xmath30 9.18e-18 & 9.46e+39 & 4.78e-16@xmath30 1.24e-16 & 8.26e+41 & 0.89 & 2.85e+07 + & 4.06e-17@xmath30 1.20e-17 & 7.96e+40 & 1.14e-16@xmath30 6.23e-17 & 2.23e+41 & -0.29 & 1.48e+07 + & 1.55e-17@xmath30 5.95e-18 & 2.14e+40 & 7.43e-18@xmath30 1.89e-17 & 1.03e+40 & -0.83 & 2.06e+07 + & 5.04e-17@xmath30 1.33e-17 & 8.24e+40 & 7.28e-17@xmath30 7.06e-17 & 1.19e+41 & -0.56 & 9.08e+06 + [ xdet ]
there are three main results in this study : 1 ) in the redshift range that we have considered , @xmath36 , massive compact qgs have stellar populations that are , on average , older than normally sized ones ; 2 ) the frequency of [ ] detection and x - ray detection are significantly lower among the compact galaxies ; and 3 ) the x - ray properties generally disfavor the presence of strong agn in both samples of recently qgs ( low luminosities @xmath37 erg / sec from both the few x - ray detections , and average stacked emission ) . while x ray detected normal qgs often also have [ ] emission , not a single compact galaxy with [ ] emission is individually detected in chandra images .
this strongly argues against agn as the power source of the [ ] emission in compact galaxies , and favors instead either stellar remnants and/or residual star formation , or liner emission , possibly powered by stellar sources ( e.g. * ? ? ?
* ; * ? ? ?
these mechanisms may also be active in the normal galaxies .
although it is unclear what is the energizing power source , it is clear that they are less active among the compact sample . taken all together , these lines of evidence paint a picture in which compact galaxies formed and evolved earlier than normal ones and , consequently , quenched star - formation earlier .
llllll & & & + flux [ ergs / s/@xmath38 & galaxy sample & soft band & hard band & + & & & & + & compact & [email protected]@xmath39&[email protected]@xmath40 + & normal & [email protected]@xmath39&[email protected]@xmath40 + luminosity [ ergs / s ] & galaxy sample & soft band & hard band & + & & & + & compact & [email protected]@xmath41&[email protected]@xmath42 + & normal & [email protected]@xmath41&[email protected]@xmath42 + [ xstack ] the evidence of the age difference between normal and compact qgs we have presented here comes from two independent age diagnostics , the @xmath43 and the h@xmath31 , as well as stellar population synthesis modeling .
the @xmath43 is larger in compact galaxies than normal ones , corresponding to an age difference of @xmath70.3 gyr , using the calibrations by @xcite .
the difference in h@xmath31 implies a larger age differential ( @xmath72.5 gyr ) but as discussed in section 3 , may be somewhat overestimated due to continuum absorption in old , metal - rich stars .
although these age conversions are model dependent , the evidence for an age _ difference _ between samples is independent of the adopted stellar libraries and assumed metallicity for the range of values found here .
stellar population synthesis modeling with ppxf provide consistent results with the lick indices .
all age diagnostics considered here imply age differentials in the same direction , i.e. more compact passive galaxies are older .
the age differential between normal and compact qgs that we discuss here is at redshift @xmath44 .
however , the result is in qualitative and quantitative agreement with the observations by @xcite at @xmath45 that at a given redshift , the largest galaxies ( in radius , although related to stellar density ) are among the youngest , suggesting that the property is a general feature of passive , massive galaxies at high redshift .
additionally , @xcite have arrived at the same result via the opposite analysis from us ; by separating the oldest qgs during the epoch 1@xmath18z@xmath182 from the youngest ( @[email protected] - 2gyr ) they find the youngest to reside on the local z@xmath70 early - type galaxy mass - size relation ( like our normal qg sample ) , whereas the old qgs are denser , with radius a factor of 2.5 - 3 smaller than local early - type galaxies .
although , they note that their qg samples differ in mass with the younger sample being less massive ( see also * ? ? ?
* ; * ? ? ?
the fact that we observe the same trend with essentially an identically mass - matched sample indicates that mass is not the primary factor related to the age differential in the population , and rather the stellar density or size may be the primary factor ( * ? ? ?
* find a similar result ) . in a complementary study , @xcite
find that at lower redshifts than our sample ( 0.2@[email protected] ) this trend of age and compactness in qgs persists in the stellar mass range explored here . at high redshift ,
compact galaxies dominate the population of qgs at the high mass end @xcite .
this has prompted investigations of scenarios where quenching is more efficient in galaxies with high stellar density because of increased stellar feedback ( e.g. * ? ? ?
however , a causal relationship between high stellar density ( compactness ) and likelihood of quenching does not directly predict an age density correlation ; rather , at any given epoch , the densest galaxies should be the most likely to quench , but not necessarily the oldest ( see e.g. * ? ? ?
* ; * ? ? ?
the presence of a relationship between age and stellar density where by denser galaxies are older is precisely the prediction of the progenitor bias scenario @xcite .
the density of a galaxy reflects the density of the universe when the galaxy formed ( assuming very little , or average , structural disruption ) and therefore older galaxies should be denser .
thus , at face value , our results support that galaxies which form earlier , and completed their evolution earlier , were simply denser than larger galaxies that form and evolve later , without the density necessarily having anything directly to do with the cessation of their star formation activity .
star - formation may be then affected by some other quenching agent ( e.g. halo or mass quenching ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this idea is extensively discussed in @xcite who show by means of a simple toy model that this scenario will naturally explain the correlations between galaxy structure and star - formation properties , without the need of a stellar density - related quenching mechanism .
an independent , but complementary , piece of information comes from the x ray and [ ] properties of our two samples . in themselves
the [ ] emission line and the x ray data do not provide any firm indication as to the causes of quenching .
the average x ray luminosity for the majority of our sample does not show evidence of any powerful agn , but one could have been present prior to reaching the current , very low level of star formation activity .
there is evidence that the sources of ionizing radiation that are still present in the two samples are different at the time of observation ; namely the compact galaxies show a significantly lower detection rate of [ ] and of x - ray emission than the normal qgs .
this difference is fully consistent with a scenario where the quenching occurred earlier in the compact sample and have therefore had a longer time to fade .
larger qgs from the normal sample are more likely to exhibit emission from energizing sources simply because on average , quenching in larger galaxies was initiated more recently . in general
, our data do not provide any conclusive constraints on the quenching mechanisms that truncated the star - formation in these galaxies , nor if the quenching mechanisms differ with stellar density .
it remains an important goal of galaxy evolution to understand the quenching mechanisms in massive galaxies . over the last several years
, efforts have been made to identify star - forming progenitors of soon - to - be qgs at z@xmath52 .
these efforts have relied on the fortuitous observation that the first galaxies to quench are compact , making it relatively easy to identify their immediate star - forming progenitors among compact star - forming galaxies @xcite .
it is interesting to note that various studies have come to very different conclusions about the nature of feedback in compact star - forming galaxies .
@xcite have claimed that the agn fraction in compact sfgs may be up to 50% , suggesting that the presence of agn may truncate the star - formation on short timescales .
however , @xcite have discovered that a subset of this sample , despite being on the star - forming main sequence , have much lower molecular gas masses than normal main - sequence galaxies , suggesting that star - formation will be quenched on short timescales due to simple gas exhaustion if the influx of gas has been suppressed .
additionally , @xcite found no evidence for agn among compact star - forming galaxies , but instead detected both faster outflow velocities in the ism , and extreme , redshifted lyman-@xmath6 emission among compact star - forming progenitors .
( similar ly@xmath6 signatures were identified among quenching galaxies by * ? ? ?
they interpreted these observational signatures as related to the compact galaxies having enhanced feedback in the ism , due to higher surface density of star - formation than their more extended counterparts ( see also * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , plausibly leading to the truncation of future star - formation . from our data
it is clear , however , that in the time since quenching was initiated in these qg samples , major energizing sources have already dissipated .
future detailed studies of galaxies closer to their quenching epoch may identify the physical processes that shut down star - formation at high - redshift .
an important feature of both samples is that some of the qgs have x - ray emission detected in the deep chandra 4ms data of the cdfs .
only a minority of the galaxies are individually detected ( see table 1 ) ; stacking the images at the position of the the non - detected galaxies , however , yields measurable flux in the soft band , but not in the hard one .
figure 7 shows a plot of the hardness ratio versus the x - ray luminosity ( left ) and versus the luminosity ( right ) .
the former correlation seems rather striking given the relatively small sample of galaxies with x - ray detections , and barring some unforeseen selection effects , it may be used as a diagnostic of the physical nature of the x - ray emission .
the latter correlation is much less obvious due to the limited dynamic range of the luminosity , and while a physical correlation could still be present , we note that the second most luminous galaxy in x - ray has the weakest emission while the most luminous one has no measurable line flux .
agn have long been suspected to be the key agent behind the quenching of star formation , as well as prevention of future star formation in galaxies ( e.g. * ? ? ?
thus an obvious question is whether or not the x - ray properties of our qgs are consistent with this idea . while the range of hardness ratio spanned by our galaxies is the same as narrow line agn and normal galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ) , the correlation between hardness ratio and x - ray luminosity shown by our galaxies qualitatively looks very different from the distribution of the parent population of x - ray detections ( see gray points in figure [ xray ] ; * ? ? ?
* ) . in number ,
these x - ray sources are dominated by agn with luminosities @xmath5 10@xmath46 erg / s ( see also * ? ? ? * for the distribution among more powerful obscured agn ) .
the distribution exhibits large scatter in hardness ratio and x - ray luminosity , most likely a reflection of the diversity of obscuration , orientation , spectral slope , power and selection effects of the agn population . while our data points seem to follow the general trend of other faint x - ray sources in their redshift bins , they seem to have much less scatter , especially given the small size of the sample ( only 9 points ) .
this , taken together with the fact that we observe stacked x - ray emission in the soft band but not the hard one ( i.e. on average our galaxies are soft x - ray sources ) suggests that the x - ray emission is unlikely powered by agn .
a possibility is that the x ray luminosity of our qgs is primarily powered by emission from hot gas , e.g. coronal gas that formed as a combination of outflows , stellar winds , and also low mass x
ray binaries ( lmxb ) .
at this time we do not have sufficient information to distinguish which one is more likely to be dominant .
coronal emission and lmxb produce x - ray radiation that is generally softer than agn , and the value of hardness ratio that we measure from our sample is generally consistent with both , although the hardness ratio of agn covers a very broad range @xcite . in the case of emission from hot gas
a correlation between x - ray luminosity and hardness ratio would simply reflect a dependence of the spectral index and the gas temperature .
this is observed , for example , in the emission from coronal gas around individual stars @xcite . in our galaxies ,
the gas would come collectively from the stellar remnants produced during the phase of star formation , or hot halo gas , both of which could be still present in the galaxies .
an in depth study of the x ray emission of our sample and the comparison with local counterparts is beyond the scope of this paper , which is devoted to the galaxies stellar age .
we simply note that the possibility that we are observing coronal emission from individual quenched galaxies at @xmath44 is very important to understanding the evolution of the interstellar medium and circum - galactic medium in early - type galaxies . in a forthcoming paper
, we plan to accurately quantify selection effects and further investigate the x - ray properties of qgs .
if confirmed , it will provide a powerful diagnostic of the feedback that took place during the star formation phase and that is preventing the galaxies from forming stars again .
we also note that even if the observed x ray emission does indeed turn out to be from hot gas , this still does not rule out agn as the cause of quenching , since the agn activity could have been quenched together with the star - formation , for example , as a result of the cessation of gas accretion into the galaxies .
the question of whether or not agn is responsible for star - formation quenching in galaxies thus remains an open one .
we have presented analysis based on optical spectroscopy of two samples of [email protected] qgs ; those which are compact relative to local early - type galaxies , and larger , more extended
normal qgs .
we find evidence that compact qgs on average have older stellar ages than normal - sized qgs .
this observed density - age correlation is an empirical prediction of the progenitor bias scenario , whereby the first galaxies to evolve and quench are simply the most compact due to their early formation time , and extended qgs form later and quench later .
we also study the [ ] emission and x - ray properties of the two samples , finding evidence for lower incidence rate of residual energy sources in the compact sample compared to normal qgs .
this is consistent with the idea that the compact sample quenched earlier , having had longer time to fade .
we do not find explicit evidence for compactness - driven quenching , and suggest that future studies of galaxies at their quenching epoch may help illuminate whether the feedback or regulation of star - formation differs with stellar density .
ccw acknowledges support from the jwst / nircam contract to the university of arizona , nas5 - 02015 .
this research made use of astropy , a community - developed core python package for astronomy @xcite , and cstack ( http://lambic.astrosen.unam.mx/cstack/ ) developed by takamitsu miyaji . | we report the detection of morphology dependent stellar age in massive quenched galaxies at redshift @xmath0 .
the sense of the dependence is that compact quenched galaxies are 0.52 gyr older than normal - sized ones .
the evidence comes from three different age indicators , d@xmath1 , h@xmath2 and fits to spectral synthesis models , applied to the stacked optical spectra of massive , quenched galaxies observed during the candels project .
all age indicators consistently show that the dominant stellar populations of compact passive galaxies are older than those of their normally
sized counterparts , with the d@xmath1 and model fitting in excellent quantitative agreement .
we detect weak [ ] emission in a fraction of passive galaxies , and the strength of the line , when present , is similar between the two samples ; however , compact galaxies have a significantly lower frequency of [ ] emission than normal ones .
a fraction of both compact and normal galaxies are also individually detected in the 4 ms chandra x
ray images , with luminosity in the range @xmath3@xmath4 erg / sec ; while both [ ] emitters and non - emitters are found to be x ray sources among normal galaxies , no compact galaxy with [ ] emission is an x ray source , arguing against an agn powering the line in compact galaxies .
we interpret the [ ] properties as further evidence that compact galaxies are older and are further along into the process of quenching star formation and suppressing gas accretion .
finally , we argue that the older stellar age of the compact passive galaxies is further evidence of progenitor bias : compact passive galaxies are older because they reflect the smaller sizes of galaxies at their earlier quenching epoch , with stellar density most likely having nothing directly to do with the cessation of their star formation . |
the most fascinating explanation of recent observations of the polarization of the cmb radiation @xcite is the hypothesis of primordial gravitational waves .
however , there are other sources of polarization - dependent deformations of the cmb spectrum that have been pointed out ( galactic dust @xcite and topological defects @xcite ) . in this paper
we explore a similar effect : the polarization - dependent distortion of the cmb spectrum in the presence of magnetic fields .
the coupling of the magnetic field to the cmb is due to the field - dependent qed vacuum polarization . in order to determine precisely this effect we quantize electromagnetic waves propagating on the magnetic background .
then we consider the properties of the thermal state of the quantized radiation . in our analysis
we shall use the results obtained previously in the classical theory @xcite .
our starting point is a relativistic lagrangian @xmath0 where @xmath1 and @xmath2 are the scalar and the pseudoscalar invariants of the electromagnetic field , @xmath3 we will employ the heisenberg - euler lagrangian @xmath4 derived in qed @xcite under the assumption that the electric and magnetic fields _ do not change _ in space and time , @xmath5,\end{aligned}\ ] ] where @xmath6 is the fine structure constant and @xmath7v / m is the characteristic value of the electric field built from the electron mass and the universal constants , called the critical field by heisenberg and euler .
the corresponding magnetic critical field is @xmath8 t .
the invariants @xmath1 and @xmath2 enter the lagrangian ( [ he ] ) through the expressions : @xmath9 the heisenberg - euler lagrangian can also be used for fields whose variation in space and time is slow on the characteristic quantum scale : @xmath10 m and @xmath11s .
we apply the theory based on this lagrangian to the description of the cosmic microwave background ( cmb ) radiation in the presence of a magnetic field . following the procedure introduced in ref .
@xcite , we separate the electromagnetic field into a strong , constant magnetic field @xmath12 and a weak wave @xmath13 , @xmath14 next , we expand the lagrangian keeping only the quadratic terms in @xmath13 ( linear terms do not contribute to the equations of motion ) , @xmath15,\end{aligned}\ ] ] where @xmath16 these derivatives of the heisenberg - euler lagrangian are to be evaluated at @xmath17 and @xmath18 .
aiming at the quantization of the @xmath13 fields , we find it more convenient to use not the lagrangian but the hamiltonian formulation .
the classical hamiltonian expressed as a function of divergenceless canonical fields @xmath19 and @xmath20 has the form : @xmath21,\end{aligned}\ ] ] where @xmath22 is the projector on the direction of the magnetic field and @xmath23 the canonical evolution equations for @xmath24 and @xmath20 , obtained from the lagrangian @xmath25 , are : @xmath26 these equations have a complete set of solutions in the form of normalized monochromatic complex mode vectors .
each mode has the @xmath27 part and the @xmath28 part , @xmath29 where @xmath30 following adler @xcite , we called these modes parallel ( the @xmath20 field lies in the @xmath31-plane ) and perpendicular ( the @xmath20 field perpendicular to this plane ) .
the mode frequencies are given by the following formulas : @xmath32 where the two refractive indices are : @xmath33 and @xmath34 is the angle between the direction of propagation @xmath35 and the direction of the magnetic field .
the whole information about the nature of nonlinearity is contained in two functions of the magnitude of the magnetic field , @xmath36 and @xmath37 .
for the heisenberg - euler lagrangian these functions do not have a closed form ; they were evaluated numerically using _ mathematica _ @xcite .
note that the magnetic field has no effect on waves propagating along the field ( @xmath34=0 ) . in fig .
[ fig1 ] we show the changes @xmath38 and @xmath39 in the refractive indices as functions of the magnetic field @xmath40 when the effect is the largest ; the direction of the field is perpendicular to the direction of observation . in fig .
[ fig2 ] we show the variation of the refractive indices with the angle of observation for the critical magnetic field @xmath41 .
having found a complete set of modes , we can construct field operators as the following linear combinations of the annihilation and creation operators , @xmath42 where the index @xmath43 takes on the values @xmath44 and @xmath45 .
the standard commutation relations between the annihilation and creation operators , @xmath46=\delta_{\lambda\lambda'}\delta({\bm k}-{\bm k}'),\end{aligned}\ ] ] guarantee the correct canonical commutation relations between the field operators , @xmath47=i\hbar\epsilon_{ijk}\partial_k\delta({\bm r}-{\bm r}').\end{aligned}\ ] ] these relations confirm that the fields @xmath24 and @xmath20 ( and not @xmath48 and @xmath20 ) form the canonically conjugate pair of fields @xcite .
their commutation relations are universal since they do not depend on the properties of the medium ( in our case the background magnetic field ) .
the substitution of the field operators into the formula ( [ ham ] ) gives the quantum energy operator , @xmath49
in the thermal state at the temperature @xmath50 the density matrix @xmath51 is an exponential function of the hamiltonian ( [ qham ] ) , @xmath52 in the presence of the magnetic field , the hamiltonian ( [ qham ] ) contains the modified frequencies ( [ freq ] ) .
the average density of the cmb photons per unit volume @xmath53 in the thermal state with the wave vector @xmath54 and polarization @xmath43 is : @xmath55 the influence of the magnetic field on the spectrum shows through the modification of the frequencies . since both refractive indices @xmath56 are greater than one , the spectrum of the cmb radiation looks like coming from regions of higher effective temperature @xmath57 .
an important characteristic feature of the spectrum is its dependence on the polarization ; the effective temperature for the perpendicular polarization is larger than the effective temperature for the parallel polarization .
the change in the temperature due to the change of the refractive index is quite large for strong magnetic field .
already for the critical field the increase in temperature is 2.4mk for the perpendicular mode and 0.9mk for the parallel mode .
these values exceed by far the standard deviations in the measurements of the cmb temperature
. the characteristic dependence of the observable temperature of the cmb radiation on polarization would be helpful in identifying regions where there exist very strong magnetic fields .
the strongest magnetic fields in the universe exist in the vicinity of magnetars @xcite .
unfortunately , the predicted distortion of the cmb spectrum by those fields can not be detected with the present observational means .
the problem lies in the smallness of the region where the magnetic field is strong , coupled with the large distance to the nearest magnetar .
one may hope that in the future the qed effects of strong magnetic fields will be accessible to observations due to improvements in angular resolution and in the precision of the measurements of temperature fluctuations in cmb . at present
one may hope to use the calculated effects of magnetic fields on the polarization only when the extension of regions with strong magnetic fields is much larger than that in the vicinity of magnetars .
we thank the anonymous referee for pointing out the difficulties in observing the effects predicted in this paper .
this research was financed by the polish national science center grant no .
2012/07/b / st1/03347 . 00
p. a. r. ade _ et .
lett . * 112 * , 241101 ( 2014 ) .
m. m. mortonson and u. seljak , j. cosmol .
. phys . * 10 * , 035 ( 2014 ) .
j. lizarraga _ et .
_ , phys.rev .
* 112 * , 171301 ( 2014 ) .
a. moss and l. pogosian , phys.rev .
112 * , 171302 ( 2014 ) .
z. bialynicka - birula and i. bialynicki - birula , phys .
d * 2 * , 2341 ( 1970 ) .
s. adler , ann .
( n.y . ) * 67 * , 599 ( 1971 ) .
w. heisenberg and h. euler , z. phys .
* 98 * , 714 ( 1936 ) .
wolfram research , inc .
, _ mathematica _ , version 9 , champaign , illinois , ( 2012 ) .
i. bialynicki - birula and z. bialynicka - birula , _ quantum electrodynamics _ ( pergamon , oxford , 1975 ) , chap . 9 .
s. a olausen and v. m. kaspi , astrophys . j. suppl . ser .
* 212 * , 6 ( 2014 ) . | the changes in the cosmic microwave background ( cmb ) spectrum seen as an increase of temperature due to a strong magnetic field are determined and their influence on the polarization of the radiation is exhibited .
the effect is due to the coupling of the cmb photons to the magnetic field in the qed vacuum via the interaction with virtual pairs . in spite of the fact that the distortion of the cmb spectrum for magnetic fields that exist in the vicinity of magnetars is quite large ,
this effect is very difficult to detect at present because the required angular resolutions is not yet available . |
the development of jetset @xcite , containing several of the components that later were merged with pythia @xcite , was begun in 1978 .
thus the current pythia 6 generator @xcite is the product of almost thirty years of development , and some of the code has not been touched in a very long time .
new options have been added , but old ones seldom removed .
the basic structure has been expanded in different directions , well beyond what it was once intended for , making it rather cumbersome by now . from the onset
, all code has been written in fortran 77 .
for the lhc era , the experimental community has made the decision to move heavy computing completely to c++ .
fortran support may be poor to non - existing , and young experimenters will not be conversant in fortran any longer .
therefore it is logical also to migrate pythia to c++ , and in the process clean up and modernise various aspects .
a first attempt in this direction was the pythia 7 project @xcite .
however , priority came to be given to the development of a generic administrative structure , renamed thepeg @xcite and adopted by the herwig++ @xcite group , while the physics parts of pythia 7 remained underdeveloped .
pythia 8 is a clean new start , to provide a successor to pythia 6 .
it is a completely standalone generator , thus not relying on thepeg or any other external library .
some optional hooks for links to other programs are provided , however .
the version presented here is the first operational one in the pythia 8 series . as such
it is not yet tested and tuned enough to have reached the same level of maturity as pythia 6 , so we expect the two to coexist for a while .
it is only by an increasing use of the new version that it will mature , however , so we encourage a critical try - out , and look forward to feedback .
the intention is to release a version in time for comparisons with first lhc data .
thus some areas , like @xmath0 and @xmath1 physics , are not yet addressed .
further , some intended processes remain to be implemented . actually , with the rise of automatic matrix - element code generation and phase - space sampling , input of process - level events via the les houches accord ( lha ) @xcite and with les houches event files ( lhef ) @xcite reduces the need to have an extensive process library inside pythia itself .
thus emphasis is more on providing a good description of subsequent steps of the story , involving elements such as initial- and final - state parton showers , multiple parton
parton interactions , string fragmentation , and decays .
the current article provides an introduction to pythia 8 .
the programming aspects are covered in more detail in a set of interlinked html ( or alternatively php ) pages that comes in the same package as the program files , see below .
much of the physics aspects are unchanged relative to the pythia 6.4 manual @xcite , and so we refer to it and to other physics articles for that .
instead what we here give is an overview for potential users who already have some experience with event generators and want to understand how to get going with pythia 8 .
section [ sec : physics ] contains an ultra - brief summary of the physics of pythia 8 , with particular emphasis on aspects that are different relative to pythia 6 .
the program structure ( including flow , files , documentation , and a few important warnings ) is described in section [ sec : structure ] ; summaries of the main user methods , including the event record and particle classes , in section [ sec : main ] .
section [ sec : databases ] is concerned with the databases of flags , modes , parameters , processes , and particle data which exist in pythia 8 .
those who wish to link to external programs , e.g. to gain access to external parton distributions , standard input / output formats , and much more , will find the relevant information in section [ sec : external ] .
a brief how - to on getting going is then included in section [ sec : how - to ] .
section [ sec : outlook ] rounds off with an outlook .
this article is not intended to provide a complete description of the physics content . for this
we primarily refer to the pythia 6 manual @xcite and associated literature .
we would like to draw attention to some key points of difference , however .
further details are available on the html / php pages in the program distribution .
some new physics aspects will eventually be covered in separate articles .
the physics components are controlled by many parameters .
these have been assigned sensible default values , based on previous experience with pythia 6 and some first studies with the new code .
we look forward to more extensive tunes by the experimental community , however .
currently the program only works with @xmath2 , @xmath3 , @xmath4 and @xmath5 incoming beams .
in particular , there is no provision for @xmath6 collisions or for incoming photon beams , neither on their own nor as flux around an electron .
the list of processes currently implemented is summarised further down ; it corresponds to most of the ones in pythia 6 , with the exception of the supersymmetry and technicolor sectors , which are yet to come .
the cross - section expressions should be identical , but default scale choices have been changed , so that cross sections may be somewhat different for that reason . the default parton distribution remains cteq 5l , but ones found in the lhapdf library @xcite can easily be linked .
it is now possible to use separate pdf sets for the hard interaction , on one hand , and for the subsequent showers and multiple interactions , on the other .
the initial- and final - state algorithms are based on the new @xmath7-ordered evolution introduced in pythia 6.3 @xcite , while the older mass - ordered ones have not been implemented .
it is now additionally possible to have a branching of a photon to a fermion pair as part of the final - state evolution . already in pythia 6.3 the initial - state evolution and
the multiple interactions were interleaved into one common decreasing @xmath7 sequence .
now also the final - state evolution is interleaved with the other two . in this context
, some of that final - state radiation gets to be associated with dipoles stretched between a final - state parton and the `` hole '' left by an initial - state one , which therefore now can take a recoil .
the initial - state - radiation algorithm remains unchanged in this respect , with recoils always taken by the hard scattering subsystem as a whole .
the multiple - interactions machinery as such contains the full functionality introduced in pythia 6.3 @xcite .
rescaled parton densities are defined after the first interaction , that take into account the nature of the previous partons extracted .
currently there is only one scenario for colour - reconnection in the final state , in which there is a certain probability for the partons of two subscatterings to have their colours interarranged in a way that reduces the total string length .
( this is intermediate in character between the original strategy @xcite and the more recent ones @xcite . )
the description of beam remnants is based on the new framework .
in addition to the standard qcd @xmath8 processes , the possibility of multiple interactions producing prompt photons , charmonia and bottomonia , low - mass drell - yan pairs , and @xmath9-channel @xmath10 exchange is now also included . for dedicated studies of two low - rate processes in coincidence
, two hard interactions can now be set in the same event .
there are no sudakov factors included for these two interactions , similarly to normal events with one hard interaction .
hadronisation is based solely on the lund string fragmentation framework @xcite ; older alternative descriptions have been left out .
particle data have been updated in agreement with the 2006 pdg tables @xcite .
this also includes a changed content of the scalar meson multiplet .
some updated charm and bottom decay tables have been obtained from the delphi and lhcb collaborations . the be@xmath11 model for bose
einstein effects @xcite has been implemented , but is not on by default .
standardised procedures have been introduced to link the program to various external programs for specific tasks , see section [ sec : external ] .
finally , some of the old jet finders and other analysis routines are made available .
also included is a utility to generate , display and save simple one - dimensional histograms .
the physics topics that have to come together in a complete event generator can crudely be subdivided into three stages : 1 . the generation of a `` process '' that decides the nature of the event . often it would be a `` hard process '' , such as @xmath12 , that is calculated in perturbation theory , but a priori we impose no requirement that a hard scale must be involved .
only a very small set of partons / particles is defined at this level , so only the main aspects of the event structure are covered .
the generation of all subsequent activity on the partonic level , involving initial- and final - state radiation , multiple parton parton interactions and the structure of beam remnants .
much of the phenomena are under an ( approximate ) perturbative control , but nonperturbative physics aspects are also important . at the end of this step ,
a realistic partonic structure has been obtained , e.g. with broadened jets and an underlying - event activity .
3 . the hadronisation of this parton configuration , by string fragmentation , followed by the decays of unstable particles .
this part is almost completely nonperturbative , and so requires extensive modelling and tuning or , especially for decays , parametrisations of existing data .
it is only at the end of this step that realistic events are available , as they could be observed by a detector .
this division of tasks is not watertight parton distributions span and connect the two first steps , to give one example but it still helps to focus the discussion .
( 430,370)(-215,10 ) ( -215,350)(215,380)0.9 ( 0,365)[]the user ( @xmath13 main program ) ( -215,300)(215,330)0.9 ( 0,315)[]`pythia ` ( -215,250)(-170,280)0.9 ( -192.5,265)[]`info ` ( -130,250)(-20,280)0.9 ( -75,265)[]`event process ` ( 20,250)(215,280)0.9 ( 105,265)[]`event event ` ( -215,110)(-85,230)0.9(-215,200)(-85,200 ) ( -150,215)[]`processlevel ` ( -150,185)[]`processcontainer ` ( -150,165)[]`phasespace ` ( -150,145)[]`lhainit , lhaevnt ` ( -150,125)[]`resonancedecays ` ( -65,110)(65,230)0.9(-65,200)(65,200 ) ( 0,215)[]`partonlevel ` ( 0,185)[]`timeshower ` ( 0,165)[]`spaceshower ` ( 0,145)[]`multipleinteractions ` ( 0,125)[]`beamremnants ` ( 85,110)(215,230)0.9(85,200)(215,200 ) ( 150,215)[]`hadronlevel ` ( 150,185)[]`stringfragmentation ` ( 150,165)[]`ministringfrag ` ( 150,145)[]`particledecays ` ( 150,125)[]`boseeinstein ` ( -130,60)(-20,90)0.9 ( -75,75)[]`beamparticle ` ( 20,60)(200,90)0.9 ( 110,75)[]`sigmaprocess , sigmatotal ` ( -215,10)(215,40)0.9 ( 0,25)[]`vec4 , rndm , hist , settings , particledatatable , resonancewidths , ` ( 0,350)(0,332 ) ( -150,300)(-150,232 ) ( 0,300)(0,232 ) ( 150,300)(150,280 ) ( 150,280)(150,250)4 ( 150,250)(150,232 ) ( -192.5,230)(-192.5,248 ) ( -192.5,280)(-192.5,300 ) ( -192.5,300)(-192.5,330)4 ( -192.5,330)(-192.5,348 ) ( -107.5,230)(-107.5,248 ) ( -42.5,250)(-42.5,232 ) ( -42.5,280)(-42.5,300 ) ( -42.5,300)(-42.5,330)4 ( -42.5,330)(-42.5,348 ) ( 42.5,230)(42.5,248 ) ( 107.5,250)(107.5,232 ) ( 182.5,230)(182.5,248 ) ( 182.5,280)(182.5,300 ) ( 182.5,300)(182.5,330)4 ( 182.5,330)(182.5,348 ) ( -107.5,100)(-107.5,108 ) ( -107.5,100)(-107.5,92 ) ( -42.5,100)(-42.5,108 ) ( -42.5,100)(-42.5,92 ) ( 42.5,100)(42.5,108 ) ( 42.5,100)(42.5,92 ) ( -160,50)(-160,108 ) ( -160,50)(0,50 ) ( 0,50)(19,59 ) the structure of the pythia 8 generator , as illustrated in fig .
[ fig : generatorstructure ] , is based on this subdivision .
the main class for all user interaction is called ` pythia ` .
it calls on the three classes ` processlevel ` , ` partonlevel ` and ` hadronlevel ` , corresponding to points 1 , 2 and 3 above .
each of these , in their turn , call on further classes that perform the separate kinds of physics tasks .
information is flowing between the different program elements in various ways , the most important being the event record , represented by the ` event ` class . actually , there are two objects of this class , one called ` process ` , that only covers the few partons of the `` hard '' process of point 1 above ( i.e. , containing information corresponding to what might be termed the `` matrix element '' level ) , and another called ` event ` , that covers the full story from the incoming beams to the final hadrons .
a small ` info ` class keeps track of useful one - of - a - kind information , such as kinematical variables of the hard process .
there are also two incoming ` beamparticle`s , that keep track of the partonic content left in the beams after a number of interactions and initial - state radiations , and rescales parton distributions accordingly .
the process library , as well as parametrisations of total , elastic and diffractive cross sections , are used both by the hard - process selection machinery and the multiple - interactions one .
the ` settings ` database keeps track of all integer , double , boolean and string variables that can be changed by the user to steer the performance of pythia , except that ` particledatatable ` is its own separate database .
finally , a number of utilities can be used just about anywhere , for lorentz four - vectors , random numbers , jet finding and simple histograms , and for a number of other `` minor '' tasks .
orthogonally to the subdivision above , there is another , more technical classification , whereby the user interaction with the generator occurs in three phases : * initialisation , where the tasks to be performed are specified . *
generation of individual events ( the `` event loop '' ) . *
finishing , where final statistics is made available .
again the subdivision ( and orthogonality ) is not strict , with many utilities and tasks stretching across the borders , and with no finishing step required for many aspects .
nevertheless , as a rule , these three phases are represented by different methods inside the class of a specific physics task .
the code is subdivided into a set of files , mainly by physics task .
each file typically contains one main class , but often with a few related helper classes that are not used elsewhere in the program .
normally the files come in pairs . 1 . a header file , `
.h ` in the ` include ` subdirectory , where the public interface of the class is declared , and inline methods are defined .
2 . a source code file , ` .cc ` in the ` src ` subdirectory , where the lengthier methods are implemented . during compilation , related dependency files ,
` .d ` , and compiled code , ` .o ` are created in the ` tmp ` subdirectory . in part
the ` .xml ` documentation files in the ` xmldoc ` subdirectory have matching names , but the match is broken by the desire to group topics more by user interaction than internal operation .
these files contain information on all settings and particle data , but not in a convenient - to - read format .
instead they are translated into a corresponding set of ` .html ` files in the ` htmldoc ` subdirectory and a set of ` .php ` files in ` phpdoc ` .
the former set can easily be read if you open the ` htmldoc / welcome.html ` file in your favourite web browser , but offers no interactivity .
the latter set must be installed under a webserver ( like a homepage ) to function properly , and then provides a simple graphical user interface if you open the ` phpdoc / welcome.php ` file in a web browser . for output to the hepmc event record format @xcite , an interface is provided in the `
hepmcinterface ` subdirectory .
there are also interfaces to allow parton distribution functions to be used from the lhapdf library @xcite and hard processes from external programs .
the installation procedure is described in a ` readme ` file ; it involves running a ` configure ` script , whereafter an ordinary ` makefile ` is used .
the former should be invoked with command - line arguments ( or be edited ) to provide the path to the hepmc library if this is going to be used .
compiled libraries are put in the ` lib ` subdirectory .
default is to build archive libraries , but optionally also shared - object ones can be built .
the standard setup is intended for linux systems , but a simplified alternative is provided for windows users . finally , some examples of main programs , along with input files , or `` cards '' , for them , are found in the ` examples ` subdirectory .
this directory contains its own ` configure ` script and ` makefile ` which will allow you to build executables , see the ` examples / readme ` file . as above ,
command - line arguments or brute - force editing allows you to set the lhapdf and pythia 6.4 paths , if so required .
the executables are placed in the ` bin ` directory , but with links from ` examples ` . playing with the files in the ` examples ` subdirectory
is encouraged , to familiarise oneself with the program . modifying the ` configure ` files may be required during installation . for the rest
, files should not be modified , at least not without careful consideration of consequences . in particular , the ` .xml ` files are set read - only , and should not be tampered with .
interspersed in them , there are lines beginning with ` < flag ` , ` < mode ` , ` < parm ` , ` < word ` , ` < particle ` , ` < channel ` , or ` < a ` .
they contain instructions from which ` settings ` and ` particledatatable ` build up their respective databases of user - accessible variables , see further below .
any stupid changes here will cause difficult - to - track errors !
further , sometimes you will see two question marks , `` ? ? '' , in the text or code .
this is for internal usage , to indicate loose ends or preliminary thoughts
. please disregard .
the ` pythia ` class is the main means of communication between the user and the event - generation process .
we here present the key methods for the user to call , ordered by context .
firstly , at the top of the main program , the proper header file must be included : + ` # include pythia.h ` + to simplify typing , it also makes sense to declare + ` using namespace pythia8 ; ` + given this , the first step in the main program is to create a generator object , e.g. with + ` pythia pythia ; ` + in the following we will assume that the ` pythia ` object has been created with this name , but of course you are free to pick another one .
when this object is declared , the ` pythia ` constructor initialises all the default values for the ` settings ` and the ` particledatatable ` data bases .
these data are now present in memory and can be modified in a number of ways before the generator is initialised ( see below ) .
most conveniently , pythia s settings and parameters can be changed by the two methods + ` pythia.readstring(string ) ; ` + for changing a single variable , and + ` pythia.readfile(filename ) ; ` + for changing a set of variables , one per line in the input file . the allowed form for a string / line
will be explained as we consider the databases in the next section .
further , methods will be introduced to list all or only the changed settings and particle data . at this stage
you can also optionally hook up with some external facilities , see section [ sec : external ] .
after this , in the initialisation call all remaining details of the generation are to be specified .
the ` pythia.init ( ... ) ` method allows a few different input formats , so you can pick the one convenient for you : + ` pythia.init(ida , idb , ea , eb ) ; ` + lets you specify the identities and energies of the two incoming beam particles , with a ( b ) assumed moving in the @xmath14 ( @xmath15 ) direction ; + ` pythia.init(ida , idb , ecm ) ; ` + is similar , but you specify the cm energy , and you are assumed in the rest frame ; + ` pythia.init(lhainit * , lhaevnt * ) ; ` + assumes lha initialisation information is available in an ` lhainit ` class object , and that lha event information will be provided by the ` lhaevnt ` class object , see below ; + `
pythia.init(filename ) ; ` + assumes that the file obeys the lhef standard format and that information can be extracted from it accordingly ; and finally + ` pythia.init ( ) ; ` + will take its values from the beam specification stored in the ` settings ` database .
it is when the ` init ( ... ) ` call is executed that all the settings values are propagated to the various program elements , and used to precalculate quantities that will be used at later stages of the generation .
further settings changed after the ` init ( ... ) ` call will be ignored ( unless methods are used to force a partial or complete re - initialisation ) . by contrast , the particle properties database is queried all the time , and so a later change would take effect immediately , for better or worse .
the bulk of the code is concerned with the event generation proper . however , all the information on how this should be done has already been specified .
therefore only a command + ` pythia.next ( ) ; ` + is required to generate the next event .
this method would be located inside an event loop , where a required number of events are to be generated .
the key output of the ` pythia.next ( ) ` command is the event record found in ` pythia.event ` , see below .
a process - level summary of the event is stored in ` pythia.process ` .
when problems are encountered , in ` init ( ... ) ` or ` next ( ) ` , they can be assigned one of three degrees of severity .
abort is the highest . in that case
the call could not complete its tasks , and returns the value ` false ` .
if this happens in ` init ( ... ) ` it is then not possible to generate any events at all . if it happens in ` next ( ) ` only the current event must be skipped . in a few cases
the abort may be predictable and desirable , e.g. when a file of lha events comes to an end .
errors are less severe , and the program can usually work around them , e.g. by backing up one step and trying again .
should that not succeed , an abort may result .
warnings are of informative character only , and do not require any corrective actions ( except , in the longer term , to find more reliable algorithms ) . at the end of the generation process
, you can call + ` pythia.statistics ( ) ; ` + to get some run statistics , both on cross sections for the subprocesses generated and on the number of aborts , errors and warnings issued .
the ` event ` class for event records is not much more than a wrapper for a vector of ` particle`s .
this vector can expand to fit the event size .
the index operator is overloaded , so that ` event[i ] ` corresponds to the ` i`th particle of an ` event ` object called ` event ` . for instance , given that the pdg identity code @xcite of a particle is provided by the ` i d ( ) ` method , ` event[i].id ( ) ` returns the identity of the ` i`th particle .
line 0 is used to represent the event as a whole , with its total four - momentum and invariant mass , but does not form part of the event history , and only contains redundant information .
when you translate to another event - record format where the first particle is assigned index 1 , such as hepmc , this line should therefore be dropped so as to keep the rest of the indices synchronised .
it is only with lines 1 and 2 , which contain the two incoming beams , that the history tracing begins .
that way unassigned mother and daughter indices can be put 0 without ambiguity . in this section , first the ` particle ` methods are surveyed , and then the further aspects of the event record .
a ` particle ` corresponds to one entry / slot / line in the event record .
its properties therefore mix ones belonging to a particle - as - such , like its identity code or four - momentum , and ones related to the event - as - a - whole , like which mother it has .
the following properties are stored for each particle , listed by the member functions you can use to extract the information : * ` i d ( ) ` : the identity of a particle , according to the pdg particle codes . * `
status ( ) ` : status code .
the full set of codes provides information on where and why a given particle was produced .
the key feature is that a particle is assigned a positive status code when it is created , which then is negated if later it branches into other particles .
the mechanism of this branching can be inferred from the status code of the daughters .
thus , at any given stage of the event - generation process , the current final state consists of the particles with positive status code . * ` mother1 ( ) , mother2 ( ) ` : the indices in the event record where the first and last mothers are stored , if any .
a few different cases are possible , to allow for one or many mothers .
the ` motherlist(i ) ` method ( see below ) can return a vector with all the mother indices , based on this info . * ` daughter1 ( ) , daughter2 ( ) ` : the indices in the event record where the first and last daughters are stored , if any .
a few different cases are possible , to allow for one or many daughters .
the ` daughterlist(i ) ` method ( see below ) can return a vector with all the daughter indices , based on this info .
* ` col ( ) , acol ( ) ` : the colour and anticolour tags , lha style . * ` px ( ) , py ( ) , pz ( ) , e ( ) ` : the particle four - momentum components ( in gev , with @xmath16 ) , alternatively extracted as a ` vec4 p ( ) ` . *
` m ( ) ` : the particle mass ( in gev ) . * ` scale ( ) ` : the scale at which a parton was produced ( in gev ) ; model - specific but relevant in the processing of an event . * ` xprod ( ) , yprod ( ) , zprod ( ) , tprod ( ) ` : the production vertex coordinates ( in mm or mm/@xmath17 ) , alternatively extracted as a ` vec4 vprod ( ) ` .
* ` tau ( ) ` : the proper lifetime ( in mm/@xmath17 ) . the same method names , with a value inserted between the brackets , set these quantities .
in addition , a number of derived quantities can easily be obtained , but can not be set , such as : * ` isfinal ( ) ` : ` true ` for a remaining particle , i.e. one with positive status code , else ` false ` .
* ` pt ( ) , pt2 ( ) ` : ( squared ) transverse momentum .
* ` mt ( ) , mt2 ( ) ` : ( squared ) transverse mass . * ` pabs ( ) , pabs2 ( ) ` : ( squared ) three - momentum magnitude . * ` theta ( ) , phi ( ) ` : polar and azimuthal angle ( in radians ) . * ` y ( ) , eta ( ) ` : rapidity and pseudorapidity . * ` xdec ( ) , ydec ( ) , zdec ( ) , tdec ( ) ` : the decay vertex coordinates , assuming free - streaming propagation , alternatively extracted as a ` vec4 vdec ( ) ` . each ` particle ` contains a pointer to the respective ` particledataentry ` object in the particle data tables .
this pointer gives access to properties of the particle species as such .
it is there mainly for convenience , and should be thrown if an event is written to disk , to avoid any problems of object persistency .
this pointer is used by member functions such as : * ` name ( ) ` : the name of the particle , as a string . * ` spintype ( ) ` : @xmath18 , or 0 where undefined spin . * ` charge ( ) , chargetype ( ) ` : charge , and three times it to make an integer .
* ` ischarged ( ) , isneutral ( ) ` : ` bool`s for charged or not .
* ` coltype ( ) ` : 0 for colour singlets , 1 for triplets , @xmath19 for antitriplets and 2 for octets . * ` m0 ( ) ` : the nominal mass of the particle species . while the ` particle ` vector is the key component of an ` event ` , a few further methods are available .
the event size can be found with ` size ( ) ` , i.e. valid particles are stored in the range @xmath20`i`@xmath21 ` event.size ( ) ` .
a listing of the whole event is obtained with ` list ( ) ` .
the basic identity , status , mother , daughter , colour , four - momentum and mass data are always given , but optional arguments can be set to provide further information , on the complete lists of mothers and daughters , and on production vertices .
the user would normally be concerned with the ` event ` object that is a public member ` event ` of the ` pythia ` class .
thus ` pythia.event[i].id ( ) ` would be used to return the identity of the ` i`th particle , and ` pythia.event.size ( ) ` to give the size of the event record .
a ` pythia ` object contains a second event record for the hard process alone , similar to the lha process specification , called ` process ` .
this record is used as input for the generation of the complete event .
thus one may e.g. call either ` pythia.process.list ( ) ` or ` pythia.event.list ( ) ` . to distinguish those two rapidly at visual inspection ,
the `` pythia event listing '' header is printed out differently , adding either `` ( hard process ) '' or `` ( complete event ) '' .
there are also a few methods with an individual particle index ` i ` as input , but requiring some search operations in the event record , and therefore not possible to define as methods of the ` particle ` class .
the most important ones are ` motherlist(i ) ` , ` daughterlist(i ) ` and ` sisterlist(i ) ` . these return a ` vector < int > ` containing a list of all the mothers , daughters or sisters of a particle
. this list may be empty or arbitrarily large , and is given in ascending order .
one data member in an event object is used to keep track of the largest ` col ( ) ` or ` acol ( ) ` tag set so far , so that new ones do not clash .
the event record also contains two further sets of vectors .
these are intended for the expert user only , so only a few words on each .
the first is a vector of junctions , i.e. vertices where three string pieces meet .
this list is often empty or else contains only a very few per event .
the second is a storage area for parton indices , classified by subsystem .
such information is needed to interleave multiple interactions , initial - state showers , final - state showers and beam remnants .
it can also be used in the hadronisation .
a set of one - of - a - kind pieces of event information is stored in the ` info ` object ( an instance of the class ` info ` ) in the ` pythia ` class .
this is mainly intended for processes generated internally , but some of the information is also available for external processes .
you can use ` pythia.info.method ( ) ` to extract e.g. the following information : * ` list ( ) ` : list some information on the current event . * ` ecm ( ) , s ( ) ` : the cm energy and its square . * ` name ( ) , code ( ) ` : the name and code of the subprocess . * ` id1 ( ) , id2 ( ) ` : the identities of the two partons coming in to the hard subprocess . * ` x1 ( ) , x2 ( ) ` : @xmath22 fractions of the two partons coming in to the hard subprocess . * ` pdf1 ( ) , pdf2 ( ) , qfac ( ) , q2fac ( ) ` : parton densities @xmath23 evaluated for the two incoming partons , and the associated factorisation scale @xmath24 and its square . * ` mhat ( ) , shat ( ) , that ( ) , uhat ( ) ` : the invariant mass of the hard subprocess and the mandelstam variables for @xmath8 processes . * ` pthat ( ) , thetahat ( ) ` : transverse momentum and polar scattering angle of the hard subprocess for @xmath8 processes . * ` alphas ( ) , alphaem ( ) , qren ( ) , q2ren ( ) ` : @xmath25 and @xmath26 values for the hard process , and the associated renormalisation scale @xmath24 and its square . * ` ntried ( ) , naccepted ( ) , sigmagen ( ) , sigmaerr ( ) ` : the number of trial and accepted events , and the resulting estimated cross section and estimated statistical error , in units of mb , summed over the included processes . in other classes
there are also methods that can be called to do a sphericity or thrust analysis or search for jets with a clustering or simple cone jet finder .
these take the event record as input .
inevitably one wants to be able to modify the default behaviour of a generator .
currently there are two pythia 8 databases with modifiable values .
one deals with general settings , the other specifically with particle data .
the key method to set a new value is + ` pythia.readstring(string ) ; ` + the typical form of a string is + ` variable = value ` + where the equal sign is optional and the variable begins with a letter for settings and a digit for particle data .
a string not beginning with either is considered as a comment and ignored . therefore inserting an initial ! , # , $ , % , or another such character , is a good way to comment out a command . for non - commented strings ,
the match of the name to the database is case - insensitive .
strings that do begin with a letter or digit and still are not recognised cause a warning to be issued , unless a second argument ` false ` is used in the call .
any further text after the value is ignored , so the rest of the string can be used for any comments . for variables with an allowed range , values below the minimum or above the maximum are set at the respective border . for ` bool ` values , the following notation may be used interchangeably : ` true ` = ` on ` = ` yes ` = ` ok ` = 1 .
everything else gives ` false ` ( including but not limited to ` false ` , ` off ` , ` no ` and ` 0 ` ) .
the ` readstring ( ... ) ` method is convenient for changing one or two settings , but becomes cumbersome for more extensive modifications .
in addition , a recompilation and relinking of the main program is necessary for any change of values .
alternatively , the changes can therefore be collected in a file , for historical reasons often called a `` card file '' , where each line is a character string defined in the same manner as above ( without quotation marks ) .
the whole file can then be read and processed with a command + ` pythia.readfile(filename ) ; ` + as above , comments can be freely interspersed .
we distinguish four kinds of user - modifiable variables , by the way they have to be stored : 1 . a ` flag ` is an on / off switch , and is stored as a ` bool ` .
2 . a ` mode ` corresponds to an enumeration of separate options , and is stored as an ` int ` .
3 . a ` parm ` short for parameter takes a continuum of values , and is stored as a ` double ` .
a ` word ` is a text string ( with no embedded blanks ) and is stored as as a ` string ` . collectively the four above kinds of variables are called settings . not surprisingly , the class that stores them is called ` settings ` . each variable stored in
` settings ` is associated with a few pieces of information .
these are : * the variable name , of the form ` class : name ` ( or ` file : name ` , or ` task : name ` , usually these agree ) , e.g. ` timeshower : ptmin ` . * the default value , set in the original declaration , and intended to represent a reasonable choice .
this value is not user modifiable . * the current value . during construction of the ` settings ` object
, this value is set equal to the default value .
it can subsequently be modified , e.g. by the ` pythia.readstring ( ) ` or ` pythia.readfile ( ) ` methods discussed above . during the ` pythia.init ( ) ` initialisation this value will be stored as a local copy in the class(es ) where it is used , and thereby also control the subsequent generation . * an allowed range of values , represented by meaningful minimum and maximum values .
this has no sense for a flag or a word , is usually rather well - defined for a mode , but less so for a parameter .
either of the minimum and maximum may be left free , giving an open - ended range .
often the allowed range exaggerates the uncertainty in our current knowledge , so as not to restrict too much what the user can do .
all the same , this information should not be modified by the user . technically , the ` settings ` class is implemented with the help of four separate maps , one for each kind of variable , with the name used as key .
the default values are taken from the ` .xml ` files in the ` xmldoc ` subdirectory .
the ` settings ` class is purely static , i.e. exists only as one global copy , that you can interact with directly by ` settings::command(argument ) ` .
however , a ` settings ` object is a public member of the ` pythia ` class , so an alternative notation would be ` pythia.settings.command(argument ) ` .
as already mentioned , for input the ` pythia.readstring ( ... ) ` method is to be preferred , since it also can handle particle data .
a typical example would be + ` pythia.readstring(timeshower:ptmin = 1.0 ) ; ` you may obtain a listing of all variables in the database by calling + ` pythia.settings.listall ( ) ; ` + the listing is strictly alphabetical , which at least means that names in the same area are kept together , but otherwise may not be so well - structured : important and unimportant ones will appear mixed .
a useful alternative is + ` pythia.settings.listchanged ( ) ; ` + which will only print out a list of those variables that differ from their defaults .
all internal processes available in pythia 8 can be switched on and off via the ordinary settings machinery just discussed , using flags of the generic type ` processgroup : processname ` .
a complete list of processes currently implemented is given in table [ processes ] .
by default all processes are off .
a whole group can be turned on by a ` processgroup : all = on ` command , then overriding the individual flags .
note that processes in the ` softqcd ` group are of a kind that can not be input via the lha , while essentially all other kinds could .
each process is assigned an integer code .
this code is not used in the internal administration of events ; it is only intended to allow a simpler user separation of different processes .
also the process name is available , as a string . for many processes it makes sense to apply phase space cuts .
the ones currently available ( in the ` settings ` database ) in particular include * ` phasespace : mhatmin , phasespace : mhatmax ` : the range of invariant masses of the scattering process . * ` phasespace : pthatmin , phasespace : pthatmax ` : the range of transverse momenta in the rest frame of the process for @xmath8 and @xmath27 processes ( for each of the products ) .
in addition , for any resonance with a breit - wigner mass distribution , the allowed mass range of that particle species is taken into account , both for @xmath28 , @xmath8 and @xmath27 processes , thereby providing a further cut possibility .
note that the ` softqcd ` processes do not use any cuts but generate their respective cross sections in full .
the following particle properties are stored in the ` particledatatable ` class for a given pdg particle identity code ` i d ` , here presented by the method used to access this property : * ` name(id ) ` : particle and antiparticle names are stored separately , the sign of ` i d ` determines which of the two is returned , with `` void '' used to indicate the absence of an antiparticle . * ` hasanti(id ) ` : ` bool ` whether a distinct antiparticle exists or not . * ` spintype(id ) ` : @xmath18 for particles with defined spin , else 0 . * ` chargetype(id ) ` : three times the charge ( to make it an integer ) ; can also be read as a ` double charge(id ) = chargetype(id)/3 ` . * ` coltype(id ) ` : the colour type , with 0 uncoloured , 1 triplet , @xmath19 antitriplet and 2 octet . * ` m0(id ) ` : the nominal mass @xmath29 ( in gev ) .
* ` mwidth(id ) ` : the width @xmath30 of the breit - wigner mass distribution ( in gev ) . * ` mmin(id ) , mmax(id ) ` : the allowed mass range generated by the breit - wigner , @xmath31 ( in gev ) . * ` tau0(id ) ` : the nominal proper lifetime @xmath32 ( in mm/@xmath17 ) . * ` constituentmass(id ) ` : the constituent mass for a quark , hardcoded as @xmath33 , @xmath34 , @xmath35 and @xmath36 gev , for a diquark the sum of quark constituent masses , and for everything else the same as the ordinary mass . *
` mrun(id , massscale ) ` : the running mass for quarks , else the same as the nominal mass . * ` maydecay(id ) ` : a flag telling whether a particle species may decay or not , offering the main user switch ( whether a given particle of this kind then actually will decay also depends on other flags in the ` particledecays ` class ) .
similar methods can also be used to set most of these properties .
each particle kind in the ` particledatatable ` also has a a vector of ` decaychannel`s associated with it .
the following properties are stored for each decay channel : * ` onmode ( ) ` : whether a channel is on ( 1 ) or off ( 0 ) , or on only for particles ( 2 ) or antiparticles ( 3 ) . * ` bratio ( ) ` : the branching ratio . * ` memode ( ) ` : the mode of processing this channel , possibly with matrix - element information ; 0 gives isotropic phase space .
* ` multiplicity ( ) ` : the number of decay products in a channel , at most 8 . * ` product(i ) ` : a list of the decay products , 8 products @xmath20`i`@xmath37 , with trailing unused ones set to 0 .
the original particle data and decay table is read in from the ` particledata.xml ` file .
the ` particledatatable ` class is purely static , i.e. exists as one global copy , that you can interact directly with by ` particledatatable::command(argument ) ` .
however , a ` particledata ` object of the ` particledatatable ` class is a public member of the ` pythia ` class , which offers an alternatively notation .
as already mentioned , for input the ` pythia.readstring(string ) ` method is to be preferred , since it also can handle settings .
it is only the form of the ` string ` that needs to be specified slightly differently than for settings , as + ` id : property = value ` .
+ the ` i d ` part is the standard pdg particle code , i.e. a number , and ` property ` is one of the ones already described above , with a few minor differences : ` name ` , ` antiname ` , ` spintype ` , ` chargetype ` , ` coltype ` , ` m0 ` , ` mwidth ` , ` mmin ` , ` mmax ` , ` tau0 ` , ` maydecay ` , ` isresonance ` , ` isvisible ` , ` doexternaldecay ` , and ` doforcewidth ` .
as before , several commands can be stored as separate lines in a file , say + ` 111:name = pizero !
normal notation pi0 ` + ` 3122:maydecay = false !
lambda0 stable ` + ` 431:tau0 = 0.15 ! d_s proper lifetime ` + and then be read with ` pythia.readfile(filename ) ` . for major changes of the properties of a particle , the above one - at - a - time changes can become rather cumbersome .
therefore a few extended input formats are available , where a whole set of properties can be given after the equal sign , separated by blanks and/or by commas .
one line like + ` id : all = name antiname spintype chargetype coltype m0 mwidth mmin mmax tau0 ` + replaces all the current information on the particle itself , but keeps its decay channels , if any , while using ` new ` instead of ` all ` also removes any previous decay channels .
( the flags ` maydecay ` , ` isresonance ` , ` isvisible ` , ` doexternaldecay ` , and ` doforcewidth ` are in either case reset to their defaults and would have to be changed separately . ) in order to change the decay data , the decay channel number needs to be given right after the particle number , i.e. the command form becomes + ` id : channel : property = value ` .
+ recognised properties are ` onmode ` , ` bratio ` , ` memode ` and ` products ` , where the latter expects a list of all the decay products , separated by blanks , up until the end of the line , or until a non - number is encountered .
the property ` all ` will replace all the information on the channel , i.e. + ` id : channel : all = onmode bratio memode products ` + to add a new channel at the end , use + ` id : addchannel = onmode bratio memode products ` + to remove all existing channels and force decays into one new channel , use + ` id : onechannel = onmode bratio memode products ` + a first ` onechannel ` command could be followed by several subsequent ` addchannel ` ones , to build up a completely new decay table for an existing particle .
+ it is currently not possible to remove a channel selectively , but setting its branching ratio vanishing is as effective . often one may want to allow only a specific subset of decay channels for a particle .
this can be achieved e.g. by a repeated use of ` id : channel : onmode ` commands , but there also is a set of commands that initiates a loop over all decay channels and allows a matching to be carried out .
the ` id : onmode ` command can switch ` on ` or ` off ` all channels .
the ` id : onifany ` and ` id : offifany ` will switch on / off all channels that contain any of the enumerated particles . for instance + ` 23:onmode = off ` + ` 23:onifany = 1 2 3 4 5 ` + first switches off all @xmath38 decay modes and then switches back on any that contains one of the five lighter quarks .
other methods are ` id : onifall ` and ` id : offifall ` , and ` id : onifmatch ` and ` id : offifmatch ` , where all the enumerated products must be present for a decay channel to be switched on / off .
the difference is that the former two allow further non - matched particles in a decay channel while the latter two do not .
there are also further methods to switch on channels selectively either for the particle or for the antiparticle .
when a particle is to be decayed , the branching ratios of the allowed channels are always rescaled to unit sum .
there are also methods for by - hand rescaling of branching ratios .
you may obtain a listing of all the particle data by calling + ` pythia.particledata.listall ( ) ` .
+ the listing is by increasing ` i d ` number . to list only those particles that have been changed , instead use + ` pythia.particledata.listchanged ( ) ` .
+ to list only one specific particle ` i d ` , use ` list(id ) ` .
it is also possible to ` list ` a ` vector < int > ` of ` i d ` s .
while pythia 8 itself is self - contained and can be run without reference to any external library , often one does want to make use of other programs that are specialised on some aspect of the generation process .
the html / php documentation accompanying the code contains full information on how the different links should be set up . here
the purpose is mainly to point out the possibilities that exist . the les houches accord for user processes ( lha ) @xcite is the standard way to input parton - level information from a matrix - elements - based generator into pythia .
the conventions for which information should be stored has been defined in a fortran context , as two commonblocks . here a c++ equivalent is defined , as two separate classes .
the ` lhainit ` and ` lhaevnt ` classes are base classes , containing reading and printout methods , plus a pure virtual method ` set ( ) ` each . derived classes have to provide these two virtual methods to do the actual work .
currently the only examples are for reading information at runtime from the respective fortran commonblock or for reading it from a les houches event file ( lhef ) @xcite .
the ` lhainit ` class stores information equivalent to the ` /heprup/ ` commonblock , as required to initialise the event - generation chain .
the ` lhaevnt ` class stores information equivalent to the ` /hepeup/ ` commonblock , as required to hand in the next parton - level configuration for complete event generation .
the ` lhainitfortran ` and ` lhaevntfortran ` are two derived classes , containing ` set ( ) ` members that read the respective lha fortran commonblock for initialisation and event information .
this can be used for a runtime link to a fortran library . as an example
, an interface is provided to the pythia 6.4 process library .
the ` lhainitlhef ` and ` lhaevntlhef ` are two other derived classes , that can read a file with initialisation and event information , assuming that the file has been written in the lhef format .
you do not need to declare these classes yourself , since a shortcut is provided by the ` pythia.init(filename ) ` command . if you create ` lhainit ` and ` lhaevnt ` objects yourself , pointers to those should be handed in with the ` init ( ... ) ` call , then of the form ` pythia.init(lhainit * , lhaevnt * ) ` .
when you implement new processes via the les houches accord you do all flavour , colour and phase - space selection externally , before your process - level events are input for further processing by pythia .
however , it is also possible to implement a new process in exactly the same way as the internal pythia ones , thus making use of the internal phase - space selection machinery to sample an externally provided cross - section expression .
the matrix - element information has to be put in a new class that derives from one of the existing classes , ` sigma1process ` for @xmath28 processes , ` sigma2process ` for @xmath8 ones , and ` sigma3process ` for @xmath27 ones , which in their turn derive from the ` sigmaprocess ` base class .
note that ` pythia ` is rather good at handling the phase space of @xmath28 and @xmath8 processes , is more primitive for @xmath27 ones and does not at all address higher multiplicities .
this limits the set of processes that you can implement in this framework .
the produced particles may be resonances , however , so it is possible to end up with bigger `` final '' multiplicities through sequential decays , and to include further matrix - element weighting in those decays . in your new class
you have to implement a number of methods .
chief among them is one to return the matrix - element weight for an already specified kinematics configuration and another one to set up the final - state flavours and colour flow of the process .
further methods exist , some of more informative character , such as providing the name of the process .
should you actually go ahead , it is strongly recommended to shop around for a similar process that has already been implemented , and to use that existing code as a template .
once a class has been written , a pointer of type ` sigmaprocess * ` to a ` new ` instance of your class needs to be created in the main program , and handed in with the ` pythia.setsigmaptr ( ... ) ` method .
from there on the process will be handled on equal footing with internally implemented processes . if your new process introduces a new particle you have to add it and its decay channels to the particle database , as already explained .
this only allows for a fixed width and fixed branching ratios , however , with only some minor generalisations . to obtain a dynamical calculation , where the width and the branching ratios can vary as a function of the currently chosen mass
, you must also create a new class for it that derives from the ` resonancewidths ` class . in it
you have to implement a method that returns the partial width for each of the possible decay channels .
the structure is simpler than for the ` sigmaprocess ` case , but again it may be convenient to use a similar existing resonance as a template .
you then hand in a pointer to an instance of this new class with the ` pythia.setresonanceptr ( ... ) ` method .
the ` pdf ` class is the base class for all parton distribution function parametrisations , from which specific ` pdf ` classes are derived .
currently the selection of sets that comes with the program is very limited ; for protons only cteq 5l ( default ) and grv 94l are available .
however , a built - in interface to the lhapdf library @xcite allows a much broader selection , if only lhapdf is linked together with pythia . should this not be enough
, it is possible to write your own class derived from the ` pdf ` base class , wherein you implement the ` xfupdate ( ... ) ` member to do the actual updating of pdfs .
once you have created two distinct ` pdf ` objects , ` pdfa ` and ` pdfb ` , you should supply pointers to these as arguments in a ` pythia.setpdfptr(pdfa * , pdfb * ) ` call . a word of warning : to switch to a new pdf set implies that a complete retuning of the generator may be required , since the underlying - event activity from multiple interactions and parton showers is changed .
there is an option that allows a replacement of the pdf for the hard process only , so that this is not required .
inconsistent but convenient .
while ` pythia ` is set up to handle any particle decays , decay products are often ( but not always ) distributed isotropically in phase space , i.e. polarisation effects and nontrivial matrix elements usually are neglected in pythia . especially for the @xmath39 lepton and for some @xmath40 mesons
it is therefore common practice to rely on dedicated decay packages @xcite . to this end , ` decayhandler ` is a base class for the external handling of decays .
the user - written derived class is called if a pointer to it has been given with the ` pythia.setdecayptr(decayhandler * , vector < int > ) ` method . the second argument to this method should contain the ` i d ` codes of all the particles that should be decayed by the external program .
it is up to the author of the derived class to send different of these particles on to separate packages , if so desired . the ` decay ( ... ) `
method in the user - written ` decayhandler ` class should do the decay , or return ` false ` if it fails . in the latter case
` pythia ` will try to do the decay itself .
thus one may implement some decay channels externally and leave the rest for ` pythia ` , assuming the ` pythia ` decay tables are adjusted accordingly .
sometimes it may be convenient to step in during the generation process : to modify the built - in cross sections , to veto undesirable events or simply to collect statistics at various stages of the evolution .
there is a base class ` userhooks ` that gives you this access at a few selected places .
this class in itself does nothing ; the idea is that you should write your own derived class for your task .
a few very simple derived classes come with the program , mainly as illustration .
there are four distinct sets of routines
. ordered by increasing complexity , rather than by their appearance in the event - generation sequence , they are : * ones that gives you access to the event record in between the process - level and parton - level steps , or in between the parton - level and hadron - level ones
. you can study the event record and decide whether to veto this event .
* ones that allow you to set a scale at which the combined multiple - interactions , initial - state and final - state parton - shower downwards evolution in @xmath7 is temporarily interrupted , so the event can be studied and either vetoed or allowed to continue the evolution .
* similar ones that instead gives you access after the first few parton - shower branchings of the hardest subprocess . * ones that gives you access to the properties of the trial hard process , so that you can modify the internal pythia cross section by your own correction factors . `
rndmengine ` is a base class for the external handling of random - number generation .
the user - written derived class is called if a pointer to it has been handed in with the ` pythia.setrndmengineptr(rndmengine * ) ` method .
since the default marsaglia - zaman algorithm is quite good , there is absolutely no physics reason to replace it , but this may still be required for consistency with other program elements in big experimental frameworks . the hepmc event format @xcite is a standard format for the storage of events in several major experiments .
the translation from the pythia 8 ` event ` format should be done after ` pythia.next ( ) ` has generated an event .
therefore there is no need for a tight linkage , but only to call the + ` hepmc::i_pythia8::fill_next_event ( pythia.event , hepmcevt ) ` + conversion routine from the main program written by the user .
version 1 of hepmc makes use of the clhep library @xcite for four - vectors , while version 2 is standalone ; this requires some adjustments in the interface code based on which version is used .
pythia 8 does not contain a machinery for calculating masses and couplings of supersymmetric particles from some small set of input parameters . instead the susy les houches accord ( slha ) @xcite is used to provide this information , as calculated by some external program .
you need to supply the name of the file where the slha information is stored , in an appropriate setting , and then the rest is taken care of automatically .
( or at least will be , once susy processes are implemented . ) it is possible to replace the existing timelike and/or spacelike showers in the program by your own .
this is truly for experts , since it requires a rather strict adherence to a wide set of rules .
these are described in detail in the html / php documentation accompanying the code .
the vincia program @xcite offers a first example of a plug - in of an external ( timelike ) shower .
after you download the ` pythia8100.tgz ` ( or later ) package from the pythia webpage , + ` http://www.thep.lu.se/`@xmath41`torbjorn/pythia.html ` + you can unpack it with ` tar xvfz pythia8100.tgz ` , into a new subdirectory ` pythia8100 ` .
the rest of the installation procedure is described in the ` readme ` file in that directory .
it is assumed you are on a linux system ; so far there is hardly any multiplatform support .
after this , the main program is up to the user to write .
a worksheet ( found on the webpage ) takes you through as step - by - step procedure , and sample main programs are provided in the ` examples ` subdirectory .
these programs are included to serve as inspiration when starting to write your own program , by illustrating the principles involved .
the information available if you open ` htmldoc / welcome.html ` in your web browser will help you explore the program possibilities further . if you install the ` phpdoc ` subdirectory under a web server you will also get extra help to build a file of commands to the ` settings ` and ` particledatatable ` machineries , to steer the execution of your main program .
such `` cards files '' are separate from the main programs proper , so that minor changes can be made without any recompilation .
it is then convenient to collect in the same place some run parameters , such as the number of events to generate , that could be used inside the main program .
therefore some such have been predefined , e.g.`main : numberofevents ` .
whether they actually are used is up to the author of a main program to decide .
as already explained in the introduction , pythia 8.1 is not yet a complete replacement of pythia 6.4 , but it is getting there , and already contains some new features not found elsewhere . in many cases
the quality of the physics should be comparable between the two versions , but obviously the objective is that soon pythia 8 should offer the overall better alternative .
this will occur by further improvements of the existing framework and by the gradual addition of new features .
the support and kind hospitality of the sft group at cern is gratefully acknowledged by ts .
mikhail kirsanov has developed the configure files , the makefiles and the interface to hepmc , and made several valuable suggestions .
ben lloyd has written the php webpage framework .
bertrand bellenot has provided a simple makefile for win32/nmake .
marc montull has helped write the extended higgs sector .
sm and ps are supported by fermi research alliance , llc , under contract no . de - ac02 - 07ch11359 with the united states department of energy .
this work was supported in part by the european union marie curie research training network mcnet under contract mrtn - ct-2006 - 035606 .
bengtsson , computer physics commun .
* 31 * ( 1984 ) 323 ; + h .- u .
bengtsson and g. ingelman , computer physics commun .
* 34 * ( 1985 ) 251 ; + h .- u .
bengtsson and t. sjstrand , computer physics commun .
* 46 * ( 1987 ) 43 ; + t. sjstrand , computer physics commun .
* 82 * ( 1994 ) 74 | the pythia program is a standard tool for the generation of high - energy collisions , comprising a coherent set of physics models for the evolution from a few - body hard process to a complex multihadronic final state .
it contains a library of hard processes and models for initial- and final - state parton showers , multiple parton - parton interactions , beam remnants , string fragmentation and particle decays .
it also has a set of utilities and interfaces to external programs .
while previous versions were written in fortran , pythia 8 represents a complete rewrite in c++ .
the current release is the first main one after this transition , and does not yet in every respect replace the old code .
it does contain some new physics aspects , on the other hand , that should make it an attractive option especially for lhc physics studies .
pacs : 13.66.-a , 13.85.-t , 12.38.-t , 12.15.-y , 12.60.-i event generators , multiparticle production , parton showers , multiple interactions , hadronisation cern - lcgapp-2007 - 04 + lu tp 07 - 28 + fermilab - pub-07 - 512-cd - t + october 2007 + , , * new version program summary * _ manuscript title : a brief introduction to pythia 8.1 _ + _ authors : torbjrn sjstrand , stephen mrenna , peter skands _
+ _ program title : pythia 8.1 _ + _ journal reference : _
+ _ catalogue identifier : _
+ _ licensing provisions : gpl version 2 _ + _ programming language : c++ _ + _ computer : commodity pcs _ + _ operating systems : linux ; should also work on other systems _ + _ ram : 8 _ megabytes + _ keywords : event generators , multiparticle production , parton showers , multiple interactions , hadronisation _ + _ pacs : 13.66.-a , 13.85.-t , 12.38.-t , 12.15.-y , 12.60.-i _ + _ classification : 11.2 phase space and event simulation _
+ _ catalogue identifier of previous version : adnn_v1_0 _ + _ journal reference of previous version : t. sjstrand , p. edn , c. friberg , l. lnnblad , g. miu , s. mrenna and e. norrbin , computer physics commun .
* 135 * ( 2001 ) 238 _ + _ does the new version supersede the previous version ? : yes , partly _
+ + _ nature of problem : high - energy collisions between elementary particles normally give rise to complex final states , with large multiplicities of hadrons , leptons , photons and neutrinos .
the relation between these final states and the underlying physics description is not a simple one , for two main reasons .
firstly , we do not even in principle have a complete understanding of the physics .
secondly , any analytical approach is made intractable by the large multiplicities .
_ + + _ solution method : complete events are generated by monte carlo methods . the complexity is mastered by a subdivision of the full problem into a set of simpler separate tasks .
all main aspects of the events are simulated , such as hard - process selection , initial- and final - state radiation , beam remnants , fragmentation , decays , and so on .
therefore events should be directly comparable with experimentally observable ones .
the programs can be used to extract physics from comparisons with existing data , or to study physics at future experiments . _ + + _ reasons for the new version : improved and expanded physics models , transition from fortran to c++ _ + + _ summary of revisions : new user interface , transverse - momentum - ordered showers , interleaving with multiple interactions , and much more _
+ + _ restrictions : depends on the problem studied _
+ + _ unusual features : none _ + + _ running time : 101000 events per second , depending on process studied _ + |
we assume that we are given a set g consisting of @xmath0 points , along with their coordinates , in n - dimensional x - space .
our task is to partition all the @xmath0 points by using planes such that no two points are left un - separated by some plane .
our method is such that each point in g , is looked at mostly once and once only , then it is put in a another set s which consists of other points and planes , s will contain all the coordinate information of the points in it as well as the equations of the planes which are in it .
we start with a set s which has a few number of points @xmath4 which were drawn from g , but positioned in s such that they have the same coordinates in s as they had in g , and also , s , has initially an adequate number of planes @xmath5 , so chosen that they separate all the points @xmath4 presently in s. after which points are chosen randomly from g ; a point thus chosen can be immediately put in s ( but at a position which has the same coordinate as the point had in g ) if it will be separated from all other points in s by the planes presently in s , if this is not so then the point will be put temporarily in a set t. points from g are thus picked and put in ( s or t ) , one by one , at some time it may be necessary to add a new plane such that , after the inclusion of this new plane ( making @xmath6 ) some @xmath1 points currently in t are separated from one another by one or other of the @xmath7 planes currently present , and thus can be transfered to s. all this is only possible because we have discovered a quick method which reveals whether a particular point p in s is separated from other points such as r in s or not .
we are able to do this because we associate an orientation vector @xmath8 , with every point p in s , this orientation vector is a hamming vector of dimension @xmath9 the current number of planes in s. the orientation vector has the property that two points will not have a plane between them if and only if @xmath10 .
that is @xmath11 automatically implies that p and r are separated by at least one of the @xmath9 planes in s. since orientation vectors are hamming vectors of dimension @xmath9 , the condition @xmath11 can be quickly verified , to ensure that points p and r are separable in s. the above process of including points , one by one , chosen randomly from g is done in such a way that all the points in s always remain separated from one another by the current number of planes @xmath9 , in s. this condition is ensured by following a procedure which we will now describe .
however , as related earlier , we require the services of another set t , and we will use this set t to temporarily hold certain `` candidate '' points and line segments till the `` candidate '' points are qualified to be transfered to s. we now describe the procedure : let us say that s contains n points and q planes , because of the process followed so far all the points @xmath3 are separable by the @xmath9 planes in s. we now choose randomly a new point , say p , from set g , this point p is then considered as a candidate for being the @xmath12 point for inclusion in s , however for p to be actually included in s we must calculate the orientation vector @xmath8 and ensure that @xmath13 points @xmath14 .
since there are @xmath3 points such as r , the preceding conditions on @xmath8 involve @xmath3 hamming vector comparisions . if p satisfies all these conditions then this implies that p is separable from all points currently existing in s and can thus be safely be added to s so p is transferd to s and then @xmath15 and a new candidate point
point p is chosen from g.
however if the test for @xmath8 fails this means there is some point @xmath16 s.t .
@xmath17 , this indicates that p and a are not separated by @xmath9 and thus p can not be put in s. at this point we rename the point p as b for convenience and put b in t and we also store the `` pair '' ( a , b ) in t. we will treat ( a , b ) as a line segment connecting two points a and b where @xmath16 and @xmath18 , we then keep a counter which keeps a count of the number of pairs in t. after this we randomly choose another point p from g as the next possible `` candidate '' for s and repeat the procedure .
it is important that at each stage of the algorithmic process , the orientation vector of points in s are calculated and stored in s along with the candidate points and their orientation vectors in another set t. the algorithm keeps adding points and calculating planes to be put in s step by step all the time ensuring that s contains only points that are separated from one another by the @xmath9 planes presently contained in s. each point when thus chosen either is separate from other points or if not , as we have seen matters are so arranged , that it is identified `` as a special point '' which is not yet separated , such points such as b , are placed in t along with its `` pair '' and with the segment ( a , b ) .
matters can always be so arranged that each point in t has one and only one point in s which is its neighbor .
( two points are neighbors if they are un - separated ) .
it is obvious that a point such as b can not have two points say a and c as neighbors , with both a and c belonging to s , because this means @xmath19 and @xmath20 which implies @xmath21 which is impossible because a and c , if they both belong to s , can not have their orientation vectors equal to each other since s contains points which are separated from one another by @xmath9 planes .
the collection proceeds ( at random ) as new points are added to s or t depending on whether the new point is either separate or labelled as `` special '' .
as soon as n `` special '' points are collected in t , the process of collecting points is temporarily halted .
now the set t is examined , since the counter has stopped at @xmath22 , there will be in t , precisely @xmath1 pairs of points @xmath23 and @xmath24 .
+ _ we now have arrived at a very crucial stage in the algorithm _
, till now we have been merely adding points , now it is time to add a new plane to s. we do so now , and we do it in such a manner that the new plane will pass through all the midpoints @xmath25 of all the @xmath1 segments @xmath26 that we have collected in t. if we do this , we would have in a single stroke , not only separated all the @xmath1 points @xmath27 from their respective pairs the @xmath28 , by this new plane , but also from one another .
it can be easily proved that all the @xmath29 will be separated from one another if we do not permit two @xmath30 to be the neighbor of the same a. ( this situation is rare in large n dimension space but is discussed later and shown that it can be handled ) .
note this requirement automatically means the two @xmath30 are separated because if @xmath31 is a neighbor of @xmath32 and if @xmath27 is not a neighbor of @xmath32 this means @xmath33 and @xmath34 implying @xmath35 .
thus if we do not permit two @xmath30 to be neighbors of the same @xmath36 , then after the @xmath7 plane bisects all the @xmath1 segments then we can include the @xmath7 plane in s , thus enabling all the @xmath1 points @xmath37 to be included in s too .
the separation of @xmath1 line segments by a single extra plane , can be done only because we are in n - dimension space .
the calculation of the coefficients of this @xmath38 plane can be done by ( say ) gaussian elimination and involves @xmath39 multiplications .
then the new plane with its coefficients are added to s. all the @xmath1 recently separated points @xmath27 are added to s so @xmath40 after this all the @xmath30 are removed from t along with the pairs ( because they are all separate - a reconfirmation is done ) . and importantly , after adding this @xmath7 new plane to s the orientation vectors of all points in s have to be modified wrt this new plane all the orientation vectors are now of dimension @xmath7
-it only means adding a bit to each vector .
+ this noniterative algorithm is possible because we use the concept of an orientation vector [ 3 ] which is a hamming vector and has the following properties : ( i ) every point is associated with its own orientation vector , ( ii ) the orientation vector for a point p in s keeps track of how p is positioned with respect to all the planes in s and ( iii ) the orientation vector is unique for each point in s and ( iv ) the orientation vectors of two points say p and q are unequal if and only if they are separated by at least one plane in s ; equality implies that p and q are not separated .
the method is a kind of pigeon hole principle , except that each point is a pigeon and the pigeon holes ( region between half spaces ) are added in installments ( much more than one at a time ) only when they are needed , to accommodate new pigeons . at each stage of the algorithm
there is only one pigeon in its hole , and as the algorithm proceeds the size of some pigeon holes may reduce as they are partitioned , and many pigeon holes are created and new pigeons are added to occupy some of the empty holes and at the end of the algorithm all the pigeons are accommodated .
we give the necessary definitions and briefly prove the fundamental algorithm by which planes which separate a given set of points could be found .
our task is to determine the equations of the planes that can separate the points in such a way that every point is separated from another by at least one plane .
assumption : it is assumed that we have specified a ( defined ) normal direction , a point which lies on the positive side of the normal is said to lie on the positive side of the plane , on the other hand if the point lies on the other side it is said to lie on the negative side.this direction is easily found if the equation of the plane is known for example if @xmath41 is the equation to some @xmath1- dimensional plane then a point p whose coordinates are @xmath42 will be said to be on the positive side if @xmath43 and in the negative side if @xmath44 each component of the orientation vector of a point p in x space , is on the positive side or negative side of each plane.(the`orientation vector ' , is defined in the next section , below . ) * orientation vector * : suppose we have a point p in n - dimension space * ( also called x - space ) * and suppose we are given 3-planes .
we define a orientation vector as a hamming vector whose components precisely specify on which side the point p lies with respect to the 3 planes .
example : if the point p lies on the positive side of plane 1 , negative side of plane 2 and positive side of plane 3 , then we define the orientation vector associated with p as @xmath8 and define @xmath45 . similarly a point q which lies on the negative side of plane 1 , negative side of plane 2 and positive side of plane 3 will have an orientation vector @xmath46 .
points , whose orientation vector differ from another by at least one component can be said to be separated by at least one plane .
* q space or hamming space : * we will also call the space spanned by the orientation vector as hamming space or q space .
since each point p in x - space has an orientation ( hamming ) vector associated with it , we can imagine that all the points in x space are mapped to a point in hamming space .
of course this mapping is many to one , but the fact to notice is the following : let p be some point in x - space , then all points , r , in x space which are not separated from p by planes will all have the same orientation vector as p ie .
@xmath10 and we describe this by saying : `` points p and r belong to the same ` quadrant ' '' this statement is certainly true for the `` images '' of p and r in q - space , but we will loosely use this terminology for the points in x - space and say p and r are in the same `` quadrant '' , what we really mean is that p and r are points in x - space and that they are not separated by planes ; sometimes we will use the term * neighbors * and say p and r are neighbors .
the point to remember is that p and r will be neighbors if and only if @xmath10 . therefore if one wishes to find out if two points a and b are separated in x - space which contains planes , all one needs to do is to compare their orientation vectors : @xmath47 with @xmath48 . * a saturated plane : * we consider a plane in @xmath1 dimension space to be `` saturated '' if it has already been constrained to pass through @xmath1 points and hence can not be adjusted to pass through a new point , the coefficients of such a plane are completely determinable .
a plane in 3 dimension gets saturated if it is made to pass through three points .
we will suppose we are given all the data containing @xmath0 points , and that all all their coordinates in @xmath1 dimension space are known to us .
our task is then to find the planes numbering @xmath49 , that can separate all these points from one another and also to find all the coefficients of the equations which define each one of these planes . *
input to the algorithm * we are given a set g containing a total of @xmath0 points in @xmath1 dimensional space .
that is we know the coordinates of all these points .
these points may also belong to different classes , so they may be given a label or ` color ' .
however , since we are isolating all points from one another regardless of their class , the labels do not matter here , but will be useful in certain applications .
* without , too much of a loss of generality we will assume that the coordinates of all the n points in g are rational numbers , i.e they are either integers or fractions .
* we will see that this assumptions makes the coefficients of all the @xmath9 planes that separate the @xmath3 in g , also rational numbers . * output items of the algorithm : * the equations of the planes which divide each point in such a way that every point is separated from another . and an integer @xmath50 which is equal to the total number of planes required .
the equations of the @xmath51 plane will be of the form : @xmath52 where all the @xmath53 are coefficients of the plane all these coefficients are determined by the algorithm and @xmath54 is a known number nearly equal to unity . * storage requirements for running the algorithm : * a set set s which will contain a list of @xmath3 points and @xmath9 planes .
s will contain the orientation vectors of all the points in s and the identification numbers ( labels ) for the planes , along with the coefficients which define each plane .
s contains an array v for storing orientation vectors ; and ` counter ' : an integer number . in all stages of the algorithm s
, will contain only those points , ( n in number ) which are completely separated from one another by the q planes contained in s. this condition of separability of the n points in s is never violated even though n and and q change as the algorithm progresses . when the algorithm ends s will contain the necessary output .
another set t which contains points and their orientation vectors and an array m : `` list of midpoints '' which are the coordinates of the mid points of certain line segments . at each stage of the algorithm this set t , will contain pairs of points one not in s and the other in s. we are sure that each such point has only one such neighbor ( in s ) , because the points in s are always separate from one another and hence two points in the same quadrant can not both belong to s. each such point stored in t will also store the coordinate of the point which is the midpoint of the line joining it to its neighbor and the line segment with its neighbor .
the points in t are the candidates waiting to be put in s but they first have to be separated from all the points in s. we shall see that as soon as t has a collection of n pairs of such points , then a new plane is drawn which separates all these points so that they become eligible to be incorporated in s. ( later on , we will permit a given point in s to have two or even three neighbors in t , these too will be candidates to be put in s , but for a first reading of this algorithm , it is simpler to assume that a point in s can have at most one neighbor in t. ) we also need another set d , this is the ` dust - bin ' set , it removes all ` accumulation ' points when detected from set g , ( as explained later this is done so that the algorithm does not go into an infinite loop ; which may happen if g contains a sequence of points converging to an ` accumulation ' point or ` limit ' point - we assume that such points do not exist in g ) .
step 1 : initially collect a small number of initial points numbering @xmath55 and choose a set of @xmath56 planes that separate each one of these @xmath55 points and put these planes in s. the coefficients defining the equations of these @xmath56 planes should be determined ( or randomly chosen but ensuring that they separate the @xmath55 ) , call @xmath4 and @xmath5 . store
all orientation vectors of points in s in array v. in addition we need a set t which will contain points which can not become immediate members of s but are prospective members and will become members of s eventually . set
t will contain points which are neighbours of a point which already is a member of s. to make things simple we will assume that at the start , all the @xmath56 planes are saturated .
( we do not want to adjust any of these planes ) and @xmath55 s.t . @xmath57 and @xmath58
. put counter = 0 .
step 2 : if no more points in g go to step 7 , else : randomly choose a new point from g , which could be a candidate point to be put in s , from the remaining points not in s , go to step 3 .
step 3 : check if this new point is in a new ` quadrant ' , this involves finding its orientation vector and comparing with those in s. if the point is in a new quadrant , put this point in s and store its orientation vector in v , put @xmath59 go to step 2 , if not , it means it has a neighbor in its quadrant .
go to step 4 .
step 4 : ( you will come here only if the current point has a neighbor in s. notation and procedure for this step : we keep count of the number of members of s which have first neighbors .
such points are called @xmath60 , its first neighbor will be called @xmath61 , if @xmath60 has a second neighbour this will be called @xmath62 and the third will be called @xmath63 . ) first : find the ` distance ' of this new point from its neighbor , call this@xmath64 * if * @xmath65 then remove the new point from g itself and put it in set d , and go to step 2 ; * else * put this new point in set t. check if this new point is in a quadrant which has already a pair in t , if not , put @xmath66 define @xmath67 , call the point with which this new point is a neighbor as @xmath60 .
however , if this new point already is a neighbor of some point @xmath68 already in s and if @xmath69 exist call this new point @xmath70 , if only @xmath71 exists call the new point @xmath72 go to step 2 . however , if @xmath73 , @xmath74 as well as @xmath75 exists this means the new point will become a 4th neighbor of @xmath68 , a situation which we do not permit , so put back the new point in g and go to step 2 . if @xmath62 and if @xmath63 do not exist , it means this point @xmath60 has only the present point as neighbor .
calculate the mid point of the segment@xmath76 , if not already calculated , and call it @xmath77 , add the coordinates @xmath77 , in a `` list of midpoints '' .
( note @xmath78 only keeps track of the first neighbors of the the @xmath79 . ) if @xmath22 go to step 5 , if less than n , go to step 2 .
step 5 : since @xmath22 , this means you have collected n points in t , each of which have to be separated from their neighbors .
however we , first re - check if all the @xmath1 points collected in t are in different quadrants ( this step is only cautionary , and not necessary if step 4 has been done properly),if say @xmath61 and @xmath80 in t are in the same quadrant , then mark one of them say @xmath80 as @xmath62 , if the latter does nt exist , else as @xmath63 ( if both @xmath62and @xmath63both exist there is no place for @xmath80 put it back in g , restore the count of points in g as well as restore in either case@xmath81 , check similarly for other @xmath82 and then go to step 2 . ) , if there are @xmath9 planes there are @xmath83 quadrants , hence the chances of two points randomly chosen to be in the same quadrant is @xmath84 .
we permit max .
no of neighbors,3 , just to ensure that even in the freakiest conditions the algorithm comes to a halt successfully .
of course the the other possibility is the presence of points of ` accumulation ' , however this does not happen when points represent integers . ] now in this step we will separate the n pairs collected in t by introducing a new plane which passes through all the @xmath1 mid points , whose coordinates are available in `` list of midpoints '' in the array m : `` list of midpoints '' and determine the coefficients of the new plane , by solving the@xmath1 constraint eqs . to pass it through the @xmath1 midpoints ; call this plane as plane number @xmath7 and put @xmath6 , and include this new plane in s along with all the @xmath1 points labeled @xmath85 and their coordinates , put @xmath86 , step 6 : now check the neighbors of all the @xmath60 , @xmath87 , in t , if only @xmath62 exists then it gets promoted to first neighbour .
but there is a slight complication after @xmath60 and @xmath88 separated @xmath62 can be a neighbor either of @xmath60 or @xmath61 so if it is the former call @xmath89 as @xmath90 else call @xmath90 as @xmath91 and @xmath92 as @xmath90 the same kind of investigation needs to be done for @xmath93 because it can now belong to the old @xmath94 quadrant or belong to the new @xmath95 quadrant , in either case the @xmath96 gets promoted as second neighbour and it will be called @xmath62
. calculate the new mid point @xmath77 for the just created segment @xmath76 and store . after finishing this task , the number , @xmath97 of second neighbours that were promoted as first neighbours will be known , after drawing the plane @xmath7 , then call @xmath98 .
update v , the orientation vectors of points now stored in s wrt to this new plane ( their dimensions become @xmath7 ) ; clear the data in m : `` list of midpoints '' of all points t you have transfered to s and remove data , of these @xmath99 points that have been transfered to s , from set t ( most of the time @xmath100 ) and go to step 2 .
step 7 : ( you will come here only if no more points are left in g @xmath101 and counter is not yet @xmath102 ) if counter = 0 , stop else introduce a new plane passing through the midpoints of segments collected so far , ( t will contain as many points as the value of counter ) , since counter @xmath103 , some of the coefficients of this plane can be randomly chosen ; ( take action like step 6 to check all these new points are in their own quadrants ) , update v , the orientation vectors of points now stored in s wrt to this new plane ( their dimensions become @xmath7 ) ; call this plane @xmath6 , counter = 0 , clear the data in m : `` list of midpoints '' then if now @xmath101 i.e. g is empty , stop ; else go back to step 2 .
* at the end of section 6.2 , we mention that the case when @xmath104 and g is empty which is one of the possibilities in step 7 ; the situation can be dealt with in an easier manner . *
end of algorithm the algorithm works swiftly most of the time , since in our case all the point are discrete and there are no limit points or ther is no `` accumulation points '' in g the algorithm will always come to a halt successfully .
the general case when the input data is say prepared by another program , then certain conditions need to be met to ensure the correct halting of the algorithm .
these are idscussed in the cted reference .
the following crucial concepts will immediately clarify the steps of the whole algorithm : : in n dimension space , if we are given n line segements ; @xmath105 , then a single plane passing through the mid points of each segment will separate the n points @xmath106 from the n points @xmath107 that is the @xmath79 and the @xmath82 will lie on opposite sides of the plane . * but this does not ensure that the @xmath79 are separate from each other and the @xmath82 are separate from each other . * in order to ensure this we have imposed the condition that each of the @xmath79 belong to set s and are * already separate from each other * and have all have different orientation vectors ( in the space of q planes which are in s ) , * similarly by imposing the condition that each @xmath108 , ( each of which belong to t ) , has only one neighbor in s * , ( for the moment ignore the second and third neighbours@xmath109 and @xmath110 ) we are making sure that each @xmath73 is separate from other @xmath82 though , each @xmath73 is not separate from its neighbor in s namely @xmath111 .
thus each pair of points @xmath112 have the same orientation vector . now
if the @xmath113)st plane is drawn then it will ensure that all the @xmath114 points i.e. the set of all the @xmath79 and the set of all the @xmath82 are not only separate but are separated from each other .
we can then include plane @xmath7 and the @xmath1 points viz .
the @xmath82 in s. this intuitive result is mathematically easily demonstrable : * * * proof : * since we have included the @xmath113)st plane , all the orientation vectors have gained one more dimension and have @xmath7 dimensions , the last , @xmath113)st component of the orientation vector of point @xmath111 will differ from the last component of the orientation vector of its ex - neighbor @xmath115 because they are now on either sides of plane @xmath113 ) . thus proving that all the @xmath114 points are now separate**. qed * * the rare cases of three points belonging to the same quadrant ( viz @xmath116 ) is permitted but not four , we do this only to avoid unecessary return of points to g , once randomly chosen from g. however , one can simplify the algorithm if one just returns the second point to g and then choose a new point at random . at the start
s only contains points which are separable from each other , because that is the way they have been selected . since @xmath5
have been selected initially each of the orientation vectors of the points in s , are @xmath9 dimension vectors .
they are all different .
now till the time we come to step 5 we are just collecting points which are separable and putting them s or in case thay have a neighbor in s then we are putting the points in t. this can go on till there are n points in t and we arrive at step 5 .
at this point there are n points in t and n points in s. firstly , there can never be two points in s which are neighbors to a single point in t. this is because every point in s has a different orientation ( hamming ) vectors so they belong to different quadrants in hamming space and one point in t can not belong to two quadrants .
there are now only two possibilities ( i ) each point in t has one distinct neighbor in s or ( ii ) there are two or three points in t which have the _ same _ neighbor in s. we will consider the first possibility ( i ) : since by choice , every point in t must have one neighbor in s , they all belong to different quadrants , since there are n points which can be separated by the @xmath7 st plane after implementing step 5
. then in the new @xmath7 space all the points ( the old points in s and the n new points in t ) are all separable form one another and hence all the n points in t which awere first neighbours the @xmath61 s can now be included in s , we must add the @xmath7 st component to each orientation vector in s to make them into orientation vector wrt to the @xmath7 st plane . now coming to the second possibility ,
if there are two or three points in t which have the same neighbor in s then after introducing the new @xmath117 plane only one of the points in t namely the one which is the first neighbor of @xmath60 say @xmath61can be transfered to s and the other points @xmath118 @xmath63 need to be kept in t with one of the points @xmath60 or @xmath61 as its neighbor and the newly calculated midpoint.(because the @xmath7 plane has separated the @xmath60and @xmath61 they are on opposite sides of plane @xmath7 , so we have to now determine on which sides @xmath118 @xmath63 are ; the seemingly complicated arguments are only to ensure that we make the right decision about @xmath118 @xmath63 ) .
however , the situation ( ii ) is rare in high dimension n space , since we are choosing points in random , and can best be avoided by not choosing two points in t which are neighbors to the same point in s ( this means ejecting the second point from t and putting it back among @xmath0 in g , or by keeping it in t and use it , later , only after you have introduced a new plane or even after a few new planes . ) in brief , the situation ( ii ) can be avoided though it just causes difficulties in programming .
so we see after step 5 and step 6 we will have an enhanced set s with more points and one more plane and all of which are separated by these @xmath7 planes .
so we can call @xmath117 as @xmath9 and then re - do steps 2 to 6 till all points @xmath0 in g are exhausted .
then s will contain all the points and all the planes which separate them .
step 7 will happen in the end when all the points are exhausted but there are less than _ n _ points in t and these must be separated .
* * the logic of step 7 is similar to step 6**. * * * it may be mentioned here that step 7 can be completely avoided . *
if we find that there are say less than n points in t. let us say r is a number such @xmath119 .
all we have to do is choose r random points ( these should not have been in g ) each of which has as a neighbor in its quadrant , one of the points which was drawn from g but now in s. then put each of these r points along with its neighbor as a pair in t. since you have already @xmath78 number of pairs in t , with the addition of these r pairs , we have @xmath1 pairs . now since @xmath22 , we can choose the last plane separating all the @xmath1 pairs .
so we see at the end of the algorithm s will contain all the points along with the equations of the @xmath9 planes that separate them , * * which are the needed outputs . * * qed * * in the above algorithm it is assumed that every new plane does not ` split ' some other point ( pass through ) . however if this happens ( it is a very rare event ) , to some point , we will have to shift the ` mid points ' of the segments slightly to one side ( because to cut a segment into two parts it is not necessary that the plane passes exactly though its mid point ) so that the plane passes to one side of the offending point,[multiblock footnote omitted ] and then proceed with the algorithm . before completing the proof of the algorithm , it better to briefly discuss the conditions when the algorithm will stop and when it will not and to make sure that it will always stop . in order to better understand what is happening :
we will pretend that the first part of the * if ... then * statement in step 4 is removed .
ie the ` distance ' is not calculated and all new points which arrive at this step are never removed and put in d even if it is close to its neighbor .
( this analysis will then reveal the need of set d ) .
it is then clear from the algorithm that as we randomly pick points from g each point picked finds a new quadrant or finds itself in t. if it finds itself in a new quadrant there is no problem : it is put in s and the algorithm goes to step 2 , and a new point is picked from g. but if it lands itself in t , then it is paired with another point which is already in s. let us say that such a point is q and it has been paired with p which is in s and we pair(p , q ) . now the algorithm will not attempt to transfer q to s from t until it has collected @xmath1 such pairs
. then the algorithm draws a plane through the mid points of @xmath1 pairs .
every thing is fine if all the pairs are in different quadrants the new plane separates all the @xmath114 points and the @xmath1 points which were collected in t ( and now separated)are transfered to s. the algorithm ensures that in the set of @xmath1 pairs if there are at most 3 ( @xmath120 ) points which are neighbors of a point in s ( in the same quadrant ) , for example @xmath121 can be neighbor of @xmath122 which is in s , then things would be fine .
things will not be fine if every new point that is subsequently picked will all belong to the same quadrant as q !
this strange phenomena will happen when : 1 .
all the q s which are chosen all belong to a sequence which has an `` accumulation '' point .
that is the remaining points in g which are left after all the others are separated are only those points which tend to be close to some unknown accumulation point ( or limit point ) close to p. this will happen since we are treating x - space as real space and all the coordinates @xmath123 are either rational numbers ( or even real numbers which are uncountable as opposed to integers which are countable ) .
and then we are trying to separate an infinity of points or uncountable points in a sequence by a countable ( or finite ) set of planes an impossibility !
2 . when all the points in g are not distinct and there are repetitions in data .
they way to get rid of both the situations is to take a bit of care in preparing the input data . in real life situations the input set g may be points generated by another computer program .
the following are obvious suggestions : 1 .
we have characterized all the components of the @xmath1 space as real numbers , that is all the @xmath124 in the coordinate array of point p : @xmath123 , are considered real . it is better that one of the components is an unique integer number .
for example , in the case of the medical data which we describe in section 3.2 ( below ) it is better to label each point with an unique integer number ie every patient p is given a unique integer label @xmath125 , and we put this value as ( say ) the @xmath126 component of the coordinate of p. that is we put @xmath127 .
if this happens no two points will ever have the same coordinate , ( because they differ in their @xmath126 component ) and the number of points will always be an integer number , and be kept finite hence countable .
they will always be separable because the minimum euclidean ( or manhattan)distance would be 1 .
2 . for every prospective candidate for t
, we check its ` distance ' @xmath64 from it neighbor and if @xmath128 , we remove the point from g and put it into the ` dust - bin ' set d. we use a small threshold value for @xmath129 , we need only calculate the ` manhattan distance ' for @xmath64 .
this process removes all ` accumulation ' points from the data and thus explains the need of the first part of * if ... then * statement in step 4 .
ensure that there are no repetitions in input data i.e. the points in g are all distinct .
once it is ensured that the input data , set g , is prepared as above , the algorithm will run smoothly .
as we have seen , as the algorithm progresses and points are being transferred from g to s and new pairs are being created , one will have to draw a new plane as soon as @xmath1 pairs are collected .
let these @xmath1 segments be called : @xmath130 , and let the mid points of each segment such as @xmath131 be denoted as @xmath132 thus we have the list of midpoints for all the @xmath1 segments : @xmath133 , where the coordinate of the @xmath134 mid point is denoted as @xmath135 now we require that the @xmath136 plane : @xmath137 should pass through all the above @xmath1 mid points , we thus have the @xmath1 constraint equations , from @xmath138 : @xmath139 the above eqs.(4)represent @xmath1 linear equations which can be solved by suitable numerical techniques , eg .
gaussian elimination , to obtain the @xmath1 unknown coefficients @xmath140 , for @xmath141 .
it may be noted that since all the @xmath132 are midpoints of the segments @xmath131 and the coordinates of @xmath142 and @xmath143 are rational numbers the coordinates of @xmath132 are also rational numbers , this makes all the coefficients @xmath140 rational numbers .
this observation has profound implications as we have described in the footnote 9 at end of page 12 . * in appendix a we have worked out a simple example on the working of the algorithm in great detail .
it may be consulted by anyone who needs to quickly get an hands on experience of the algorithm and program it . * the computational complexity of the algorithm is easily determined . *
given : * @xmath0 : total number of points ; @xmath50 : total number of planes used ; @xmath1 ; dimension of x - space .
\1 . to determine the coefficients of each plane by gaussian elimination we require @xmath144 multiplications , and @xmath144 additions therefore for @xmath50 planes we have @xmath145 multiplications and @xmath145 additions . 2 .
to compute the orientation vector of each point with respect to @xmath50 planes require per point per plane : @xmath1 linear evaluations , each consisting of @xmath1 multiplications and @xmath1 additions .
therefore total : @xmath146 multiplications and @xmath146 additions .
when every time a new point is put into s , it s orientation vector is compared with the orientation vector of the points in s. therefore there are @xmath147 hamming vector comparisons .
but just to see if two hamming vectors of dimension @xmath50 are not equal it does not require us to check all the @xmath50 components , the moment the rth component differs the hamming vectors are declared unequal .
so we may assume that 99 percent of the time only about 8 components are checked on the average .
so the number of bit comparisons made are : @xmath148 bit comparisons .
we must calculate the manhattan distance of every candidate member which arrives at t from its neighbor ( step 4 ) .
for every plane there will be @xmath1 such points arriving at t. each manhattan distance calculation requires @xmath1 subtractions and @xmath1 additions , therefore for @xmath9 planes we have @xmath149 additions .
we ignore the additions required to detect accumulation points which we had put in the dust - bin set d , assuming they are small in number .
now we need to guess as to how many planes are require to separate @xmath0 points , we use the estimate of boland and urrutia ( 1995 ) , ref [ 2 ] , also see [ 1],and assume that @xmath150 , hence we come to the conclusion that the order of computations involved by the algorithm are : @xmath151 multiplications ; @xmath152 additions and @xmath153 bit comparisons .
we will now briefly describe how this algorithm can be applied to data retrieval .
suppose one has @xmath3 points in a @xmath1 dimension x - space and the algorithm was used to separate all n points and it was found after running the algorithm that @xmath9 planes were finally necessary .
we will now show that this information can be used as a data retrieval device .
that is all the data can now be stored in such a manner that retrieval can be done with great efficiency .
now for purposes of this illustration we will assume that each point n represents @xmath1 dimension image ( say we had used a passport size photograph which represents a face using precisely @xmath1 pixels ) now we wish to store these images in such a manner that retrieval becomes easy .
the idea is simple : after we have solved the problem , we will have the exact information of how each of the data points @xmath3 reside in x - space with respect to each other and the separation planes , because we have the orientation vector for each point stored in s. we can use this information to ( i ) store the images in such a manner that ( ii ) it is possible to retrieve it easily .
what we mean is that after the storage is done and if we are given a fresh ( approximate ) image of a person whose face is stored in the storage receptacle in the repository , it possible to use this new ( approximate ) image to retrieve the stored image . we show that if the new photo which is a point p in x - space , is close to the original photo , say a point l stored in the receptacle then all we need to do is calculate @xmath8 and compare with @xmath154
, we will show presently by using the neural engine in the picture below the task requires only @xmath155 multiplication @xmath155 additions to retrieve l. fig ref : fig - fig - a in the figure @xmath156 represent the equation to the first plane stored in s ; we assume that the equation to this plane is gven by @xmath41 we then define @xmath157 if we use the unit step function defined as : @xmath158 defined s.t .
@xmath159 , if @xmath160 and @xmath161 if @xmath162 and define @xmath163 , then @xmath164 is 1 if the point p is on the + ve side of plane 1 and is zero if it is on the -ve side of plane 1 .
the same goes for the other planes @xmath165 .
so the output array @xmath166 is a binarized version of the orientation vector @xmath8 .
therefore our `` storage plan '' for each point l in is to use a binary representation of the orientation vector @xmath154 of a point @xmath167 as its label as shown in the receptacle in the figure and then store information about l in the space next to this code .
`` retrieval plan '' : present an approximate image of l say p to the network .
it is then possible to immediately retieve the information about l when a nearby point p has the same ov as l ie it is in the same quadrant as l in x - space .
thus the coordinates @xmath168 of p when presented to the neural engine retrieves l. the retrieval just takes @xmath155 multiplication @xmath155 additions .
a similar application can be thought of in medicine . in this case
a person p may be represented as a point in n - dimensional feature space and the @xmath1 numbers in the array @xmath168 may represent , values of platelet count , wbc count , rbc count , mcv , mch , etc .
a total of @xmath1 variables .
so a doctor may be looking at patient p , and wondering if there is any other person in the data base whose medical condition is similar to that of the present patient p. all the doctor has to do is to input @xmath168 to the retrieval system shown in fig 1 , and a l will be extracted s.t @xmath169 , hence it is very likely that p and l share similar ailments since they share the same quadrant .
the doctor can then examine the entire case sheet of person l which is retrieved by the the repository shown to the extreme right .
thanks to the advancement of computer technology , algorithms such as this one could be very attractive . in order to be aware that such possibilities could soon become realities ,
let us obtain an approximate estimate of the numbers involved : taking the case of medical records ; assuming @xmath170 for such a problem since @xmath1 is large it is reasonable to expect that approx .
@xmath171 , we see that the number of calculations for finding these planes and separating all the n points is approx @xmath172 multiplications , since computers have achieved around 1 tflops this translates to around 20 seconds . to retrieve the data :
for example the history of patient l , which is similar to patient p , from a data base of 10 billion , would take @xmath155 multiplications i.e. 2000 multiplications , which is practically instantaneous .
in these two subsections we speak about one of the properties that the algorithm has and some interesting aspects of the algorithm .
claude shannon , held the view that a bit is an information and one must use just the right amount of bits necessary to solve a problem .
( a mathematical theory of communication bell sys . tech .
379 - 423 , and 623 - 656 , 1948 ) .
we show below that the algorithm gathers and uses the information optimally : we will define a completed ` stage ' of the algorithmic process as that state when a new plane say the @xmath51 has been inducted into s along with the other points which it has just separated and which are now in s along with all the orientation vectors of the points now in s. to get to the next ` stage ' the algorithm transfers enough number of points and garners just sufficient information for it to draw the @xmath38 plane
. we will now prove that the new information collected in bits is exactly equal to that which will be stored after the next stage is completed that is after the @xmath38 plane is drawn .
no extra bits are collected , so the algorithm always makes optimum use of information it collects ` stage ' by ` stage ' right up to the end when the problem is completely solved .
let us imagine : we are at a ` stage ' when we had just added the @xmath51 plane which then separated all @xmath3 points in s , since we have calculated all the @xmath3 , orientation vectors w.rt .
each of the planes we have used @xmath173 bits of information to arrive at this ` stage ' .
now let us say as we proceed with the algorithmic process : we added @xmath99 new points which happened to fall in empty quadrants and @xmath1 new points which are paired with the old .
when the @xmath38 plane is drawn all these @xmath174 new points are added to @xmath3 , to make the number of points in s equal to @xmath175 .
but since we need to determine the orientation vectors of these new points wrt the @xmath117 planes now in s , we would be acquiring @xmath176 bits of information , in addition , since we have to update the orientation vectors of the old @xmath3 points wrt this new @xmath117plane , we have to add one bit for each such old vector ( because the old orientation vectors were of dimension @xmath9 and now it has to be increased by one to accommodate the new plane ) so we have acquired @xmath3 bits .
so the new bits of information acquired is : @xmath177 .
so if we add this to the old @xmath173 bits already acquired we have a total of @xmath178 bits which is exactly equal to @xmath179 bits now stored as orientation vectors by the @xmath180 points now in s. so we see , that in the process of acquisition and usage and then storage of bits of information there has been no loss nor any redundancy . and even though the points in g are picked at random , shanon s principle has been followed , strictly , in an optimal manner in going from @xmath9 planes to @xmath7 planes and then onwards till the final @xmath50 plane is determined and drawn .
the algorithm uses the geometrical properties of @xmath181dimension space , to judiciously choose a plane passing through @xmath1 mid - points simultaneously , at an appropriate time , to make all this possible
. we will now speak about a few interesting aspects of the algorithm : \1 . once you have solved the problem ,
there is no need to start from the very beginning , if at some later date you need to separate a new set of @xmath182 points .
all you have to do is include this new set into set g ( which is presently empty ) and start running the algorithm from step 2
. then these new points will be inducted into s one by one and new planes added when ever necessary , till all the points are separated and g is once again empty.also since for large dimension @xmath1 we have the relationship @xmath183 , when you have solved the problem of @xmath3 points with @xmath9 planes then if at some later date the number of points become @xmath184 then @xmath185 only and when @xmath186 then @xmath187 .
this is a huge advantage .
suppose at some future time you need to increase the dimension of data say from @xmath1 to @xmath188 ?
once again there is no need to worry , assume that all the existing points have the same coordinates for the first @xmath1 components but for the @xmath189 st coordinate we define its value as zero .
do the same thing for the planes , the @xmath189 coefficient of each plane is defined to be zero .
that is for any existing planes we just define the coefficient @xmath190 eg .
see eq(1 ) .
now all the data has become @xmath189 dimensional and there is no need to change anything else and as and when new data points which are now dimension @xmath188 comes in , we put them into g and the algorithm can proceed from step 2 , just as before .
both the points 1 .
and 2 , above raises interesting possibilities .
suppose each data point of @xmath1 dimension , represented an image of dimension @xmath1 , you can incorporate new images at any point of time .
in fact you can incorporate images of larger size .
let us extend 3 .
a bit further : now if you want to incorporate a different kind of data which are @xmath97 dimensional , let us say they are different because they are data pertaining to spoken words , again there is no problem ! just increase the dimension of data to dimension @xmath191 , so each point is an entity in @xmath191 dimension space .
the old data will live in the first @xmath1 dimensions with their components from @xmath188 to @xmath191 defined as zero ; whereas the new data pertaining to words will again be defined in this @xmath191 dimension space but will have their first @xmath1 components as zero and have their components from @xmath188 to @xmath191 as mostly non zero .
so we see new data can always be added not only of the same type but of different types and the old data is never discarded only the dimension of space increases and this can go on from generation to generation .
the set s will be the repository of all knowledge in perpetuity ! as they say knowledge never dies !
in this paper we have reported the discovery of a new algorithm for separating a given set of @xmath0 points by planes in n - dimensional space .
the algorithm is noniterative and will always halt successfully .
it has the property of restart , if new points are needed to be separated the algorithm can continue from where it left off . and at some later stage if the dimension of the data is increased ( @xmath1 to @xmath191 ) is increased the algorithm can still continue from where it left off , ( after some adjustments ) and tackle the new data points which are of a higher dimension .
the algorithm is insensitive to the type of data ; the points may represent , images , words or any other type .
it can handle even the types are mixed as discussed in section 4 .
a rigorous proof has been provided in the body of the paper and a worked example is given in appendix a. the computational complexity is of @xmath192 , where @xmath3 is the given number of points and @xmath1 is the dimension of space . *
an informal essay on the new algorithm and its future applications is contained in appendix b *
the author thanks the management of sreenidhi institute of science and technology for their sustained support .
he proclaims his grateful thanks to his wife suhasini , for her willingness to be a sounding board on innumerable occasions and listen to his monologues without which he does not think this work would have ever been done .
\1 . william b. johnson and joram lindenstrauss :
extensions of lipschitz mappings on to a hilbert space , contemporary mathematics , 26 , pp 189 - 206 ( 1984 ) \2 .
ralph p. boland and jorge urrutia : separating collection of points in euclidean spaces , information processing letters , vol 53 , no.4 , pp , 177 - 183 ( 1995 ) \3 .
k.eswaran:a system and method of classification etc .
patents filed ipo no.(a ) 1256/che july 2006 and ( b ) 2669/che june 2015 \4 .
fischer : `` the statistical utilization of multiple measurements '' , annals of eugenics , 8 , 376 - 386 ( 1938 ) ; also annals eugenics , 7 , 179 - 188 ( 1936 ) \5 .
mahalanobis : `` on the generalized distance in statistics '' , proc .
inst . of sc .
india , 12 , 49 - 55 ( 1936 ) \6 .
mcculloch , w. and pitts , w. a : `` logical calculus of the ideas immanent in nervous activity '' , bulletin of mathematical biophysics , 5:115133 .
( 1943 ) \7 .
a. n. kolmogorov : on the representation of continuous functions of many variables by superpositions of continuous functions of one variable and addition .
doklay akademii nauk ussr , 14(5):953 - 956 , ( 1957 ) .
translated in : amer .
28 , 55 - 59 ( 1963 ) .
paul werbos : `` beyond regression : new tools for prediction and analysis in the behavioral sciences '' , phd thesis , harvard university , 1974 \9 .
rumelhart , d. e. , hinton , g. e. , and r. j. williams : learning representations by back - propagating errors .
nature , 323 , 533536 ( 1986 ) .
j. schmidhuber : deep learning in neural networks : an overview .
75 pages , www.arxiv.org/abs/1404.7828 ( 2014 ) .
yoshua bengio : learning deep architectures for ai .
foundations and trends in machine learning : vol .
1 , pp 1 - 127 ( 2009 ) .
k. eswaran : `` on the storage and retrieval of primes using n - dimensional geometry '' , sent for publication .
hawkins with s. blakeslee : `` on intelligence '' , publ . henry holt and co. ny .
( 2004 ) \14 .
d. george and j.c .
hawkins : trainable hierarchical memory system and method , january 24 2012 .
url https:/ www.google.com patents us8103603 .
us patent 8,103,603
* we will now show the working of our algorithm in complete detail .
the algorithm was used to separate the points shown in fig * [ fig : fig - b ] it was discovered that the above points which are 29 in number in 2d space , can be separated by 8 planes , by using the algorithm .
we will now show how this was done .
however , instead of using unlabeled points we will use fig [ fig : fig - c ] which has a label for each point but in the same position .
( fig [ fig : fig - b ] and fig [ fig : fig - c ] are identical except for labels ) .
the 29 points belong to three classes , ( class a , b , c ) .
we have numbered each point by a number , these numbers is the sequence with which each point was chosen to be in set s ( or t ) when we solved the problem .
we now retain the labels while we describe the process that was followed so that the method can be better understood .
we describe the process in stages .
* stage1 : * the first stage is when we started with an initial set of points which are separated by planes . in our commentary below we use the present tense in describing what was actually already done , this change of tense from past to present has been adopted so that the commentary reads better .
this initial set of planes are two in number and numbered 1 and 2 as shown in fig [ fig : fig - i ] below . and the initial set of points , three in number are @xmath193 . we store the coefficients of the two equations for the planes 1 and 2 in s. the equations for the two planes 1 and 2 will be of the form : @xmath194 @xmath195 we compute the orientation vectors of these three points which are shown in fig [ fig : fig - i ] . and store in s and put @xmath196 and @xmath197 .
the first component of the orientation vector of @xmath198 is @xmath199 because it lies on the positive side of the normal of plane 1 , indicated by black arrow , similarly the second component is also @xmath199 because @xmath198 is also on the positive side of plane 2 , therefore we write @xmath200 , the other orientation vectors of @xmath201 and @xmath202 are depicted in fig [ fig : fig - i ] .
the first component of @xmath203 is @xmath204 , because it lies on the negative side of plane 2 but its second component is @xmath199 because it lies on the positive side of plane 2 , hence @xmath205 , the ov of point @xmath201 is also shown in the fig [ fig : fig - i ] . notice all the three point have different orientation vectors and therefore they all are in separate quadrants , of course this is as it should be , otherwise they would not have qualified to be members of s. now applying step 2 of our algorithm we choose point @xmath206 , and then go to step 3 , and check whether this point is in the same quadrant as the other three points , in order to do this we have to find the orientation vector of this point : this is done by substituting the @xmath207 of the point @xmath206 in eq .
( 2 ) , for plane 1 , the lhs will evaluate to some -ve value which will signify that @xmath206 is on the -ve side of plane 1 , then we substitute @xmath207 of the point @xmath206 in eq . ( 3 ) which is the equation for plane 2 , the lhs will again evaluate to some -ve value which will signify that @xmath206 is on the -ve side of plane 2 . in this manner
we obtain the orientation vector of point @xmath206 as @xmath208 . in n dimension space
this is the only way we can determine orientation vectors . and
determining an orientation vector of a point in n - space where q planes are present would involve evaluating q linear equations of type eq.(1 ) , and therefore would need a total of @xmath155 multiplications and @xmath155 additions .
however when n = 2 ( or 3 ) and we have a diagram such as the one below present , we can write down the orientation vectors by sight , we can see that @xmath206 is on the -ve side of the plane 2 because it is on the opposite side of the normal direction of plane 1 , which is indicated by the black arrow on the top of the figure , similarly @xmath206 is on the -ve side of the plane 2 because it is on the opposite side of the normal direction of plane 2 , which is indicated by the black arrow in the middle right corner of the figure , hence we can conclude , correctly , that @xmath208 . * from now on we will write down the orientation vectors by sight .
* we now complete the process given in step 3 with regard to this new point @xmath206 , we see that by comparing the orientation vectors of this point @xmath206 with the orientation vectors of the points which are currently in s , namely @xmath193 , and finding that it is different we are sure that @xmath206 is in its own quadrant space and thus point @xmath206 is added to s , and its orientation vector is stored in v and we put @xmath209 and go to step 2 .
we now see that there are no more free regions where a point randomly chosen from g that is points from fig [ fig : fig - c ] can be put in s. and hence when we go to step 2 and choose a new point , we will be directed to go to step 3 , as we shall soon see in stage 2 .
* stage 2 * we are now in step 2 , we randomly select a point say @xmath210 notice it has a neighbor @xmath211 .
this is discovered by sight , but in actuality if we are in @xmath1 dimension space , we calculate the orientation vector of this new point , @xmath212 and compare with all other orientation vectors of points already in s , namely @xmath213 and then discover that @xmath214 , this implies that in q - space @xmath211and @xmath210 are mapped to the same point , hence in x - space they are not separated by any planes thus we discover that point @xmath211 is a neighbor of the randomly chosen point @xmath210 .
the above is the standard procedure of discovering whether a new randomly chosen point has a neighbor in s , this procedure is done in step 3 for all randomly chosen points .
_ this standard procedure of discovering neighbors in s , is assumed to be always adopted , _ * though for the purpose of this example we will from hence forth , do the `` neigbour detection '' in s , by sight . *
now that @xmath210 has a neighbor we go to step 4 , calculate the mid point coordinate @xmath215 and store it in the list of midpoints and put @xmath210 in t. put @xmath216 and then go to step 2 .
we now choose a new point at random say , @xmath217 it has a neighbor point @xmath218 which is already in s , we find the midpoint @xmath219 put @xmath217 in t and increase @xmath220 , but this time we can not go back to step 2 .
since we are in two dimension space ( n=2 ) , we need to immediately separate the points in t from their respective neighbors in s. as can be seen that the algorithm initiates procedures to start including a new plane , in s , as soon as @xmath22 ( where @xmath1 is the dimension of space ) . which are stored in the `` list of midpoints '' .
the calculation of the coefficients of this new plane is done in step 5 and the order of computations to find the newplane ( say ) by using gaussian elimination is @xmath39 , so if we have to intoduce @xmath50 planes in the problem the multiplications are in the order of @xmath221 . ] to proceed with our algorithm we need to go to step 5 .
step 5 says that we must now find the equation to a new plane which passes through the mid points @xmath215 and @xmath219 , this is easliy done by assuming that the equation to this new plane is : @xmath222 ; with unknown coefficients @xmath223 then these two unknown coefficients can be found by using the condition that the line must pass through @xmath215and @xmath219 , whose coordinates are @xmath224 and @xmath225 to obtain the eqs : @xmath226 @xmath227 solving the above we determine @xmath223 which completely defines the equation for plane 3 , which is now include s and we drawn this line as shown in fig [ fig : fig - d ] . since the coefficients @xmath223 are known the direction of the normal is also known , the normal is also drawn ( black arrow ) in the figure .
now @xmath228 .
going to step 6 , we check to see if all the points are in their own quadrants ( ie are separated ) by actually updating the finding the orientation vectors of the old points wrt to the 3 planes .
the orientation vectors are now of dimension 3 and given by : @xmath229 @xmath230 @xmath231 .
we see that all the orientation vectors of @xmath232 are all different so that they are all in their own quadrants in q - space and hence separated from one another in x - space .
this ofcourse , is plainly visible by sight in fig [ fig : fig - d ] .
now we complete the tasks indicated in step 6 .
we clear the `` list of mid points '' , include the new points @xmath233 in s , include plane 3 in s put q=3 , clear the set t because its contents @xmath233 have been transferd to s. then we go to step 2 .
the situation arising is shown in fig [ fig : fig - d ] below : * stage 3 : * we are now in step 2 .
since no more points can be added , the situation is exactly the same as that in the beginning of stage 2 , so we repeat what was done in stage 2 except that we randomly choose two new points @xmath234 and @xmath235 which have neighbors @xmath236 and @xmath217 which are already in s. so we follow the steps similar to that described in stage 2 and end up with a new plane , viz .
plane number 4 , put q = q+1=4 ; and s now contains 8 points all their orientation vectors in the new 4 dimensional q space are all different and therefore all the 8 points are separated by these 4 planes .
see fig [ fig : fig - e ] .
* stage 4 ; * from now on we will hurry through the steps .
we are in step 2 .
two new points : @xmath237 which is neighbor of @xmath206 and @xmath238 which is a neighbor of @xmath211 are introduced and a new plane plane no .
5 is introduced and transfered to s. after this we see that the next point @xmath239 occurs in an empty quadrant hence it is immediately added to s , the next two points @xmath240 have neighbors @xmath238 and @xmath241 resp . and therefore need a new plane , viz plane 6 , and are all transfered to s along with plane 6 .
the next point @xmath242 happens to be in its own quadrant so is included in s. the next two points @xmath243 have neighbors @xmath244 and @xmath245 in s and determine a new plane 7 and are transferred to s along with 7 .
see fig [ fig : fig - f ] * stage 5 : * we are in step 2 : 7 planes have been just added and s contains 16 points .
all the orientation vectors are now dimension 7 .
they are all different for the 16 points , as may be checked tediously or by sight .
the new randomly chosen points @xmath246 all are in their own quadrants and can be directly included in s with out troubling ourselves to add a new plane see fig [ fig : fig - g ] .
( this sort of situation when we can include a whole sequence of points before we introduce a new plane occurs when the number of planes q increase , because the number of quadrants in q space increase . for q planes we have @xmath83 quadrants in q space , therefore there is space for more points in q space . )
dimension space is much easier , this is because whenever you introduce a plane say from @xmath9 to @xmath7 the number of quadrants in x - space double from @xmath247 to @xmath248 , this doubling in x - space stops after @xmath249 after which you will get some confined regions . now an interesting question arises : if have just added a plane ( say ) the @xmath51 and you already have n points in s , how many points can you add to s before you need the next plane ?
the answer for large @xmath1 is that you can on the average add @xmath250 points . ] after this we arrive at @xmath251 and @xmath252 which have neighbors already in s namely @xmath218 and @xmath217 .
these determine the last plane plane 8 .
see fig [ fig : fig - g ] * stage 6 : * we are in step 2 and have just introduced plane 8 .
there are 24 points in s and 8 planes .
we see that we now randomly choose the sequence of points @xmath253 and @xmath254 all of them are in their own quadrants and are already separated from other points in s and from each other and hence can be included in s without adding any more planes .
and now there are no more points in g and so we have @xmath255and @xmath256 .
the final solution is given in fig [ fig : fig - h ] in this subsection we will , for the sake of completeness , tackle some of the situations that we had discussed before , in section 4,but with reference to this * worked example .
* how shall we tackle new points ?
suppose we have once solved a problem that is , we have separated @xmath3 points given in an original set g , in n dimension space and have found that @xmath9 planes have done the job .
after this we make a retrieval and storage engine as depicted in fig [ fig : fig - a ] and discussed in section 3 .
we begin to use the retrieval engine for some time , and afterwards we encounter new data .
* case 1 : new data has the same dimension as the old data * we have already spoken about this situation : we had said that there is no need to start from the very beginning , the algorithm can start from where it left off.we just add these new points to set g , which until now was empty , and start the algorithm from step 2 keeping the set s containing the latest number of points and planes which we had ( i.e. @xmath257 and @xmath258 , with all the @xmath259 ) , just before we encounter this new data .
we already depicted the situation , in our example , figure [ fig : fig - h ] , we have separated 29 points with 8 planes .
now we encounter 2 more points ( say ) @xmath260 and @xmath261 .
we just continue as before ( assuming just for convenience that both these have neighbors ) : treat these two as temporary candidates in t with neighbors as shown in [ fig : fig - j ] and draw a new plane number 9 .
this separates all the 31 points and we have an additional plane , and these two new points all of which now can be included in s. notice new quadrants are created , these new quadrants could capture some new points , which are are still in g. the algorithm continues from step 2 till g is empty and then stops . if g becomes empty at @xmath262 then the situation in s is as shown in fig [ fig : fig - j ] . *
case 2 : new data does not have the same dimension as the old data , but has one more dimension @xmath263 * suppose we are as before in a situation depicted by fig .
[ fig : fig - h ] i.e. s contains 29 points and 8 planes .
now we encounter three new points of dimension @xmath263 , since all our original data is 2d we can assume that all the points are in the x - y plane .
we assume that these three new points belong to the new data set and are of dimension @xmath188 .
we tackle the problem by first converting all the 2-d data in s to 3 d data , this is simply done by embedding the 2d data in 3d space . with reference to fig
[ fig : fig - c ] step c1 : convert all the 2-d coordinates by defining for each point whose original coordinates are @xmath264 a new coordinate @xmath265 such that @xmath266 step c2 : convert all the 2d equations of the @xmath9 planes to corresponding equations valid as 3d equations eg if the equation of the @xmath267nd plane is ( say ) : @xmath195 we should convert it to : @xmath268 and define @xmath269 this process should be done for all the @xmath9 planes . by this process
a 2d plane will become a 3d plane , however its normal will be in the x - y plane .
if we assume that x - y plane to be ` horizontal ' then all the planes will become vertical walls containing the original 2d lines .
these planes are now shown as blue lines .
all the original quadrants at this stage will become 3-d regions defined by vertical walls shown as blue lines .
we start the algorithmic process : we go to step 2 of the algorithm after putting @xmath270 and @xmath271 and putting a , b , c in g. and s will contain @xmath9 3d planes and n points along with their ` orientation vectors ' . for simplicity we will assume that none of the three points a , b , c fall in a new ` quadrant ' so they will all end up in t and each with their respective neighbors for example the neighbor of a and b are shown to be : @xmath244 and @xmath272 resp .
we will have three mid points which we call @xmath273 .
now since @xmath274 , we go to step 5 and find the equation to the plane which passes through these three midpoints .
the new ` genuine ' 3d planest plane is the first plane among all the planes in s , whose normal is not lying in the x - y plane .
of course as we proceed with the algorithm and be adding points in 3d space there will be many such planes in s. ] along with the old 8 planes will now separate all the n=29 points from a , b , c and from each other .
so we can now add a , b , c to s and include this new plane so @xmath275 and @xmath276 and calculate all the orientation vectors of the 32 points ( i.e. update v ) and then go to step 2 of the algorithm after clearing t and the ` list of midpoints ' .
so we see now s has 32 points all separated by 9 , 3d planes if there are more points in g , the algorithm proceeds or else stops .
fig [ fig : fig - k ] we have thus seen how the algorithm can restart from the last step made earlier , even when the dimension of the new points are increased by one . * end of worked example *
* what the algorithm does * : imagine we are given a set @xmath277 of n points in n - dimensional space .
basically the algorithm finds planes , in n - dimension space such that they can separate all the n points , in such a manner that every point is separated from every other point by at least one plane .
the output of the algorithm is a set s containing the points and also the equations of the planes that separate them .
( in general for large dimension space the number of planes @xmath278required is approx.@xmath279 ) . *
how its done : * let us imagine the set @xmath277 as an n - dimensional @xmath280space , containing @xmath3 points .
we create another n - dimensional @xmath281-space called @xmath282 .
we then transfer points randomly from @xmath277 to @xmath282 one by one , so that they occupy the same coordinate position in @xmath282 as they had occupied in @xmath277 ( their coordinates do not change ) and also planes are drawn in s. the algorithm makes sure that after a new plane is drawn , all the points in s at this stage are separated .
the algorithm proceeds stage by stage transferring , new points from @xmath277 and drawing new planes in s till eventually s contains all the n points as well as the q planes needed to separate all the n from one another . *
beauty : * there is a certain beauty in the algorithm , there are no redundancies and duplications .
if it required say 97 bits of information to draw a plane , then it will acquire exactly all these 97 bits of information and draw the plane , then store the 97 bits , before it seeks more information to draw another plane ) , sec 4.1 , p. 15
, contains the proof .
it adopts shannons principle that each bit is an information , so use it as far as possible , before you seek another bit of information , ( but these are technicalities that i will pass on for some other time ) .
just to put the paper in its proper perspective , we list its contents : 1 .
it has a very strict mathematical proof , of how @xmath3 points in n - dimensional space can be separated by q planes . 2 .
the complexity is of @xmath283 .
this itself is somewhat unique because very few algorithms have this efficiency to name a few:(i ) the fft , complexity : @xmath284 , ( ii ) euclidean algorithm of gcd complexity : @xmath285 , @xmath3 being the larger of the two numbers , ( iii ) quick sort algorithm complexity on the average @xmath284 worst case @xmath286 , ( iv ) gauss elimination for solving @xmath3 linear equations @xmath287 , ( vi ) primality testing algorithm of agarwal et al , complexity @xmath288 , ( vi ) rsa algorithm @xmath287 , here @xmath3 is the number of bits .
it contains the conditions under which the algorithm can be made to work along with a proof .
we have simple worked example in the appendix a and outlined applications in sec 3 and its interesting properties in sec 4 .
many researchers , from statisticians and neural network scientists have long been trying to separate clusters by planes or discriminant functions .
this has been ever since the time of fisher [ 4 ] and mahalanobis [ 5 ] and mcculloch and pitts[6 ] , kolmogorov [ 7 ] , werbos[8 ] , and rumelhart , hinton and williams [ 9 ] and others [ 10 ] , also the deep learning people[11 ] , famous names in the fields .
but all have had great difficulties in large n - dimensional space and they found that it is not easy .
so the phrase like the curse of dimensionality and np hard complexity , has become a part of folk lore . however , from the very beginning it was never very easy to separate clusters by planes .
this is mostly because , a cluster is not well defined , every cluster has its own shape and in @xmath1-dimensions you could have long thin filaments and all kinds of snake like dragon like shapes which constitute a cluster .
though statisticians try to approximate the shape of each cluster as ellipsoids or even simple spheres , clusters in general would require more parameters to define their shapes than that required to define the planes which are supposed to separate them !
so all along it was , perhaps , very naive of all of us to have tried to separate clusters when such entities are not mathematically well defined .
i felt that it is far better to separate the individual points and to use the enormous space and degrees of freedom that is available in @xmath1-dimension space to separate each point , rather than try to separate clusters which will never be well defined .
this was the genesis of the idea that gave an impetus to do the kind of research work reported in this paper .
as an illustration , in the last section we have considered a cluster of 29 points in 2-d , see fig 3 , and used the algorithm to separate all the 29 points using 8 planes see fig 9 , ( it can be proved that the theoretical minimum for 29 points in 2-d , is 7 planes ) .
another fact , that only adds to the prospect of success in this new direction is that : for large n dimension space where @xmath289 ( @xmath3 being the number of points ) , you will find the number of planes @xmath9 , needed for separating @xmath3 points is far less than the number of points itself , in fact @xmath290 imagine : even a highly reduced small passport size image of @xmath291 pixels is a point in a @xmath292 dimension space . and
in such a space there are @xmath293 quadrants i.e. approximately @xmath294 quadrants . and
even if you put one image in one quadrant ( thus automatically separating them from others ) you will never be able to fill up this space : there are only about @xmath295 atoms in the universe so how will you make so many photographs ? the answer is : we wanted all the readers to concentrate on the algorithm and its proof and not rile them with matters not germane to the task on hand ; for after all a mathematical paper makes its greatest impact by cold logic and rigid proofs rather than tall talk and philosophy . .. in sec 3.2 where we describe a possible application to medical records of 10 billion people- the number of planes , @xmath9 , you require would be @xmath296 or @xmath297 planes .
most of the time you will require very few planes to separate all@xmath3 points .
this is counter intuitive but will always happen , for large @xmath1 the number of planes @xmath9 will be @xmath298 . the following explains the phenomena : consider an empty @xmath1-dimensional space , void of planes , then if you put one plane , it divides the space to 2 quadrants , the 2nd plane will divide the space to 4 quadrants , the 3rd to 8 and the qth plane to@xmath83 quadrants .
( the doubling stops only when @xmath249 , afterwards some closed regions are formed ) , so you see you have sufficient number of planes @xmath9 to handle @xmath3 points even if you put one point in a single quadrant all you need is @xmath279planes . in all these applications one must some how employ mappings to reduce multi - dimensional data , temporal data or any other kind of data to image points in n - dimensional space ,
it is only then that the methods of the algorithm described in this paper can be usefully employed for classification or decision making .
an example illustrating this method is given in ref.[12 ] , by which any @xmath181digit prime number can be depicted as a point in @xmath181 dimension space ; and therefore a prime number repository for storage and easy retrieval of primes can be created .
the figure 12 , below shows how all 2-digit prime numbers can be considered as points in 2-d space and separated by just 10 planes .
chess games:**we can treat each chess game as a point in @xmath114-dimension .
its coordinates being @xmath299 the @xmath1 moves made by white and black ( it is not hard to convert the moves @xmath300 etc . into numbers
so the above array which represents a single chess game , is represented as a single point in 2n - dimension space ) .
it is possible to store numerous chess games , say @xmath3 , for easy retrieval using the same technique as described for medical data .
the number of planes involved would be only @xmath301 . in the above cases the data is static and not dynamic .
in many life situations it is necessary to tackle dynamic data viz .
sequences , like a sequence of events , tracking of time signal , or even such matters as health monitoring of machines , ( eg . see jeff hawking [ 13]-[14 ] , who has underlined the importance of storing and recalling a sequence of events in order to further research in artificial intelligence ) .
we then need a method using which we can compare two sequences , say , one a shorter sequence , @xmath302 , of length , @xmath303 , with a longer sequence @xmath304 of length @xmath97 , which is stored in the memory .
the @xmath72 may be the condition of a machine in the year , @xmath305 , of its working life . so that the possible future values of the shorter sequence@xmath72
can be predicted by comparing it with a longer sequence@xmath306 , which is the record of some machine which has already lived its life and whose records are now stored in a memory along with the historical records of many such machines .
we must imagine that the longer second sequence has been extracted from the data base , by the retrieval engine ( see sec 3.2 ) , because it happens to have its first @xmath303 values somewhat close to the @xmath303 values of the shorter sequence .
we describe a method of mapping a sequence to points in a @xmath1-dimensional space .
this mapping is necessary to tackle dynamic situations and tracking / memorizing a sequence of events .
the data contained in the sequence are then mere points in @xmath1-dimensional space which are then separated by using the algorithm using @xmath9 planes , the orientation vectors are used to store the data in a repository sec 3.2 ) . but in order to perform all these tasks we need a scheme to convert a sequence to points in @xmath1-dimensional space ; we now demonstrate how this can be done .
we will suppose@xmath1 is the max number of years that each record has .
it is possible to store the sequence as points in an @xmath1 dimensional space using the following scheme .
we define the coordinates in @xmath1-dimension space for each member of the sequence as follows : note : each of the above is a point in @xmath1-dimension space .
hence the sequence of points given above is like a world - line in @xmath1-space . ( even though , we considered the @xmath72 as a single number ; there is no difficulty if this is ( say ) @xmath316 dimensional then the actual value of the bigger space will @xmath317 dimensional i.e. be @xmath318 . and a similar scheme for the sequence@xmath319 : similarly , we can think of the sequence of points above as a world - line in n - space .
the problem of comparing two sequences @xmath326 and @xmath319 has been reduced to the comparison of two world - lines .
since we have separated every point such as the @xmath327 by using planes , we can easily retrieve any sequence such as @xmath319 by using the retrieval engine . now , suppose we are presented a point @xmath328 ( as described in sec 3.2 ) , whose coordinates are @xmath329 ; the retrieval engine will retrieve a point @xmath167 whose coordinates @xmath330 are closest to point @xmath328 .
this implies you have detected another world - line @xmath319 , which had had similar experiences in its first@xmath303 episodes of its ` life ' as that of world - line @xmath326 ; and hence it is quite possible that the fate of @xmath326 would be similar to what befell @xmath331 thus making it possible to retrieve the entire record @xmath332 , thus enabling us to predict the possible values of @xmath333 when @xmath334 by looking at the other values @xmath335 which are now been made available . in the above example , you could also think of world line @xmath326 as the medical record of some person who is presently living and is @xmath303 years old . and the world line @xmath319 as the medical record of some person who had probably lived and died , but whose first @xmath303 years of life she / he had a similar medical history as that of @xmath326 .
this is how we can predict the ` future life ' of a presently occurring sequence of events by using the repository containing historical data of sequences that have occurred in the past .
* 3 . other possibilities :
* we could similarly convert decision trees and/or logical trees to sequences which can then be mapped as points in a large @xmath1 dimension space , so we see the algorithm can help in decision making and imitative learning etc .
of large complex data . | we demonstrate an algorithm that can separate any number of points in n - dimensional space from one another by planes . given a set g of @xmath0 points , along with their coordinates , in n - dimensional x - space , the algorithm partitions all the @xmath0 points by using planes such that no two points are left un - separated by some plane .
the algorithm proceeds by transferring points from g to another set s in n - dimensional x - space , to the same coordinate position in s as it was in g. however initially s has very few points and very few planes , but the few planes are so chosen that all the points in s are already separated from each other by the few planes contained in it .
each point is then chosen at random from g and then transferred to s one by one , matters are so arranged that if necessary new planes are drawn in s so that the points in s will always be separated . in order to do this each point from g
is labeled as a member of s if it is separated from others in s , otherwise it is labeled as a `` special '' point which is not yet separated . the collection proceeds ( at random ) as new points are added to s from g and each new point is either separate and becomes a member of s or labeled as `` special '' .
as soon as @xmath1 `` special '' points are collected they are all separated by a single extra plane , which is then added to the collection .
the algorithm is made possible because a new concept called orientation vector is used .
this vector is a hamming vector and is associated with each point and has all the information necessary to ascertain if two points are separate or not .
the algorithm proceeds till s contains all the planes which separate all the points .
the method is non iterative , it will always halt successfully and the algorithm strictly follows shannon s principle of making optimal use of information as it advances stage by stage .
it has the property of restart , if new points are needed to be separated the algorithm can continue from where it left off . at some later stage
if the dimension of the data ( n to n+r ) is increased the algorithm can still continue from where it left off , and tackle the new data points which are of a higher dimension , after some minor modifications . a proof is provided with a worked example . the computational complexity is of @xmath2 , where @xmath3 is the given number of points and $ n$ is the dimension of space . in summary
this paper describes a non - iterative algorithm of separating a given set of points in n - dimension by planes .
its application to data retrieval problems in very large medical data bases is also given .
possible future applications are also identified .
dept . of computer science ,
sreenidhi institute of science and technology , yamnampet , ghatkesar , hyderabad , 500010 india |
the quasi - one - dimensional spin chain and ladder ( la , y , sr , ca)@xmath0cu@xmath1o@xmath2 compounds have attracted considerable interest since the discovery of a quantum critical phase transition in their phase diagram.@xcite these compounds are the first superconducting copper oxide materials with a non - square lattice .
they are composed of alternately stacked chain and ladder planes .
the chains are made up of cuo@xmath3 linear edge - sharing cuo@xmath4 squares and the ladder planes consist of two zig - zag strings of corner - sharing cu@xmath3o@xmath5 squares ( see fig .
[ fig1 ] ) .
the parent compound sr@xmath0cu@xmath1o@xmath2 is naturally doped with six holes per formula unit ( @xmath6 ) .
( color online ) 3d view along the * c * crystal axis of four 316-atom af unit cells of sr@xmath0cu@xmath1o@xmath2 .
the structure is composed of alternately stacked @xmath7 layers , constituted of cuo@xmath3 linear edge - sharing cuo@xmath8 squares , and @xmath9 planes , made up of two zig - zag strings of cu@xmath3o@xmath5 corner - sharing squares .
they are parallel to the ( * a*,*c*)-plane and separated by sr atoms .
cu / o / sr atoms are colored in red / blue / yellow.,width=283 ] determining exactly how the holes are distributed in the system is hindered by the complexity of the crystal structure of sr@xmath0cu@xmath1o@xmath2 and electron correlation effects.@xcite room temperature optical conductivity,@xcite near - edge x - ray absorption fine structure ( nexafs),@xcite x - ray emission spectroscopy,@xcite and hall coefficient measurements@xcite on sr@xmath0cu@xmath1o@xmath2 estimated that at least @xmath10 holes per @xmath6 reside in chains and at most @xmath11 in ladders .
high density of holes in chains is compatible with the low temperature t@xmath12 @xmath13 200 k@xcite charge ordering with the 5-fold @xmath7 periodicity accompanying the antiferromagnetic ( af ) spin - dimerization , observed by inelastic neutron scattering.@xcite in ladders , a gapped spin - liquid and the charge density wave ( cdw ) appear in the ground state , in the spin and charge sectors , respectively .
whereas the spin - liquid and its collective spin excitations ( triplons ) are well understood theoretically and experimentally,@xcite the low hole density in the charge sector is at variance with the 5-fold periodicity in @xmath9 cell units ( wigner hole crystal ) observed by low temperature resonant x - ray diffraction measurements at the o k edge if a 4k@xmath14 cdw picture is assumed.@xcite on the other hand , a revisited interpretation of nexafs spectra@xcite claimed a distribution of 2.8 holes in ladders and 3.2 holes in chains , which satisfactorly explains the cdw in ladders , while it fails to explain af dimer order in chains . the apparently contradictory results are mainly related to the lack of a suitable theoretical model .
all published nexafs analyses are based on semi - empirical models that do not take explicitly into account the complex interplay between spins and holes in ladders and chains .
thus the basic question of where the holes reside in sr@xmath0cu@xmath1o@xmath2 has remained unresolved up to now .
the hole distribution between chains and ladders is the key to stabilization of interdependent electronic phases in these two subsystems .
it is a prerequisite to understanding the low - energy physics of the system and such open issues as the occurrence of zhang - rice singlets,@xcite the evolution of spin order and the origin of the superconductivity in the ca - doped compound.@xcite we bridge the gap between theory and experiment by performing low temperature electron yield polarization dependent o k pre - edge nexafs measurements and state - of - the - art electronic structure calculations . in the theoretical
modeling we consider the full 316-atom af unit cell , include the core - hole effects , core level shift and correlations in a dft+u framework .
this method has proven to be fairly successful in reproducing x absorption spectra in correlated metals with well localized orbitals @xcite .
we show that , contrary to previous claims , correlation effects and af order stabilize holes on chains , induce hole - depletion in ladders , where rung sites become less populated than leg sites .
high - quality single crystals of sr@xmath0cu@xmath1o@xmath2 were grown by the traveling - solvent floating - zone method and characterized by x - ray diffraction measurements .
a well oriented sample was cleaved @xmath15-@xmath16 along the ( * a*,*c * ) plane under a pressure of 10@xmath17 mbar .
the o k edge nexafs measurements were performed using the helical undulator dragon beam line id08 at the european synchrotron radiation facility ( esrf ) in the total electron yield mode with 200 mev resolution .
the incident light was normal to the sample surface and its polarization was parallel to the * a*/*c * sample axis .
nexafs calculations are performed by the xspectra@xcite code based on density functional theory ( dft)@xcite and all - electron wave function reconstruction.@xcite correlation effects are simulated in a mean field dft+u framework @xcite using a hubbard on - site energy of u@xmath18 ev for copper and u@xmath19 ev for oxygen .
the chain and ladder atomic positions were taken from gotoh et al.@xcite without imposing any superstructure modulations , either along the chains or along the ladders .
we performed a specific nexafs calculation for @xmath20 of the five inequivalent oxygen sites existing in the structure , two in the ladders , and three in the chains .
we find a negligible core - level - shift .
( color online ) polarization dependent o k nexafs pre - edge intensity for * * e**@xmath21*a * and * * e**@xmath21*c * measured at 150 k. ) .
subscript @xmath22 is for experiment.,width=340 ] ( color online ) temperature dependence of the o k nexafs pre - edge intensity for * * e**@xmath21*c * showing that for this polarization spectra change below and above the charge ordering temperature ( @xmath13 200 k ) .
the inset shows that the * * e**@xmath21*a * spectra temperature change is negligible.,width=340 ]
the low temperature ( 150 k , see fig . [ fig2 ] ) o k nexafs pre - edge spectra has two distinct structures , a@xmath23 , centered at 527.9 ev and b@xmath23 , at 529.5 ev ( @xmath22 is for `` experiment '' ) , in agreement with previously published room - temperature spectra.@xcite the shape of the two structures is dependent on the polarization of the incident photon : for * * e**@xmath21*a * , a@xmath23 is stronger than b@xmath23 , while for * * e**@xmath21*c * ( direction along the chains and the ladders ) it is the opposite . for * * e**@xmath21*c * structure a@xmath23 has a second , well resolved peak at 528.5 ev , a@xmath23 .
the intensity of this structure is only weakly affected by the charge ordering transition ( see fig . [ fig3 ] ) .
the * * e**@xmath21*a * spectra are even less affected at t@xmath12 ( see fig .
[ fig3 ] inset ) .
the attribution of the features in these spectra has so far been very controversial.@xcite theoretical modeling using the af unit cell with 5-fold periodicity in chains is shown in fig .
the two , a@xmath23 and b@xmath23 , structures are well reproduced in the calculations ( labeled a@xmath24 and b@xmath24 where @xmath25 is for the calculated spectra ) , although their separation is only 0.7 ev , compared to 1.6 ev for the experiment .
this discrepancy is attributed to the well - known underestimation of the hubbard gap in dft+u simulations .
the agreement between theory and experiment is particularly good for the * * e**@xmath21*c * polarization as in the * c*-direction the charges are less localized . for * * e**@xmath21*c * ( i ) the b@xmath24 cross section is stronger than the a@xmath24 cross section , and ( ii ) the high - energy shoulder a@xmath24 is clearly visible . for * * e**@xmath21*a * polarization , the b@xmath24 structure has smaller cross section than the a@xmath24 peak .
the two insets in fig .
[ fig4 ] show the calculated cross section contribution of chain ( c ) , ladder - leg ( l ) and ladder - rung ( r ) oxygens to the nexafs spectra .
the contribution from ladder - rungs is small or negligible for all features except for the b@xmath24 peak in the * * e**@xmath21*a * geometry .
this result is in line with the analysis by ncker et al.@xcite and is in strong disagreement with the results of rusydi et al.@xcite where the contribution of ladder rungs was over - estimated .
our results demonstrate that _ ladder - rungs are hole - depleted in _ sr@xmath0cu@xmath1o@xmath2 . for both polarizations the low energy a@xmath24 peak
is equally composed of chain and ladder - leg states , with a minor contribution from ladder - rungs in the * * e**@xmath21*a * geometry .
its high energy a@xmath24 shoulder is composed of both ladder - leg and @xmath7 contributions .
this result is in stark disagreement with all previous claims based on semi - empirical nexafs analysis @xcite that attribute the a@xmath24 structure solely to ladders .
it points to the fact that the o k edge resonant diffraction results , performed at the photon energy corresponding to ac , could be related to the modulation reflections due to the chain - ladder lattice mismatch@xcite .
the wigner - hole crystallization analysis of the o k edge resonant soft x - ray scattering data in ref .
now becomes questionable .
( color online ) calculated o k nexafs pre - edge cross section on the af unit cell with 5 fold periodicity in chains , for * * e**@xmath21*a * and * * e**@xmath21*c*. insets : contribution of chain ( c ) , ladder - leg ( l ) and ladder - rung ( r ) oxygens for * * e**@xmath21*a * ( upper inset ) and * * e**@xmath21*c * ( lower inset ) .
zero energy corresponds to a photon energy of 527.4 ev .
subscript @xmath25 stands for calculated.,width=340 ] ( color online ) calculated o k nexafs pre - edge cross section on the af unit cell with 4 fold periodicity in chains , for * * e**@xmath21*a * and * * e**@xmath21*c*. insets : contribution of chain ( c ) , ladder - leg ( l ) and ladder - rung ( r ) oxygens for * * e**@xmath21*a * ( upper inset ) and * * e**@xmath21*c * ( lower inset ) .
zero energy corresponds to a photon energy of 527.4 ev.,width=340 ] calculations using the af unit cell with 4-fold periodicity in chains@xcite is shown in fig .
the two insets show the calculated cross section contribution of chain ( c ) , ladder - leg ( l ) and ladder - rung ( r ) oxygens , equivalent to these in fig .
the experimental results are less well reproduced compared to the 5-fold chain - periodicity unit cell .
first , the a@xmath24 - b@xmath24 energy separation is smaller .
further , for the * * e**@xmath21*c * geometry a@xmath24 and b@xmath24 are almost equal in intensity , contrary to the experiment .
finally , a@xmath24 appears as a low energy shoulder to structure b@xmath24 and not as a high energy shoulder to a@xmath24 . for these reasons ,
the 4-fold chain - periodicity unit cell can be eliminated as a model for chain spin ordering .
( color online ) ( upper panel ) total dos of sr@xmath0cu@xmath1o@xmath2 without a core - hole , calculated on the af unit cell with 5 fold chain periodicity .
the fermi level ( e@xmath14 ) corresponds to 0 ev .
( lower panels ) local chain- and ladder - dos projected to cu ( red line ) and o ( blue lines ) .
the second dotted line indicates where e@xmath14 would be if the system had 6 additional electrons per @xmath6 , i.e. no holes ( e@xmath14 = e@xmath14+@xmath26 , where @xmath26 = 0.31 ev corresponds to additional 6 x 4 = 24 electrons).,width=302 ] ( color online ) chain site dependent hole ( spin ) density : ( a ) chains have natural 5-fold hole periodicity in the paramagnetic cell without correlations ; ( b ) when the 4-fold chain spin periodicity is imposed , the 5-fold hole periodicity is destroyed ; ( c ) in the case of the 5-fold spin periodicity , the natural hole periodicity is preserved , on oxygen and as well on copper sites.,width=321 ] to understand the role of core - hole effects we also performed density of states ( dos ) calculations for the system without a core - hole ( see fig .
[ fig4 ] ) .
contrary to previous theoretical works,@xcite performed on a minimal cell without af ordering , we find the occurrence of a 0.25 ev gap appearing in ladder states .
this finding is in agreement with the insulating behavior of the ladders @xcite and points out the need of including the full af crystal structure in the simulation .
the occurrence of a 0.25 ev hubbard gap in ladder states results in a strong hole depletion of this sub - system .
indeed , we find that only about 0.4 holes reside in ladders , while the chains accommodate about 5.3 holes per @xmath6 . a more detailed analysis of hole distribution is presented in table [ proba ] .
the mean value of the probability of accommodating a hole on one atom is schematically presented in fig .
[ fig8 ] by the thickness of the contour surrounding atoms of each group ( chain , ladder - leg , ladder - rung ) .
.[proba ] calculated fraction ( @xmath27 ) of 6 holes per @xmath6 residing in chain- , ladder- and in sr - inter - layers of sr@xmath0cu@xmath1o@xmath2 ; number of atoms in each group ( @xmath28 ) ; distribution of holes between cu- and o - sites in chains and ladders ( @xmath29 ) .
[ cols="^,^,^,^,>,^,^,^ " , ] the role of correlations is even more evident when analyzing hole and spin order along the chains in the paramagnetic dft calculations and in the dft including u and af ordering .
when correlation effects are switched - on the number of holes in the ladder is decreased by a factor of five .
_ hole - depletion of ladders is thus a pure correlation effect_. the distribution of more than 5 holes in chains and less than 1 in ladders is compatible not only with the af unit cell with 5-fold chain periodicity , but also with the af model with 4-fold chain periodicity.@xcite but the latter model turns out to be irrelevant since in this case the nexafs calculations could not reproduce the experimental data ( see sect.[sec : nexcalc ] ) . moreover , the 4-fold chain - spin periodicity destroys the natural 5-fold hole ordering , while the 5-fold spin periodicity preserves it , as shown in fig.[fig7 ] .
we would like to stress once more that taking into account the proper long - range af order is crucial to achieve a correct description of o k - edge nexafs spectra .
it is worthwhile noting that our dos is in good agreement with experiments as we find the occurrence of a strong -2 ev band with mixed chain / ladder contributions and large cu - ladder weight , in accordance with angle resolved photoemission spectroscopy ( arpes ) measurements.@xcite furthermore the two ladder - dos structures at -0.25 ev and -0.1 ev have also been observed in a recent arpes @xcite experiment and were identified as the quasi-1d underlying fermi surface .
finally , the tendency of hole - depleting ladder - rungs found in our work , provides a satisfactory explation for the 2d to 1d cross - over when the number of holes in the system is reduced , like in under - doped compounds@xcite or at low temperature where the less localized ladder sub - system suffers from the migration of holes to chains ( back - transfer ) as reported in the nmr.@xcite on the fully doped side we suggest that the ca doping has a twofold effect : it destroys long - range af order in chains @xcite and reduces hole - depletion in ladders .
increased population of the rung sites reinforces the 2d character of the ladders . for sufficient ca - induced ladder - doping the cdw ground state
is then suppressed in favor of the 2d superconductivity under pressure.@xcite
in conclusion , our experimental nexafs spectra and first - principles theoretical modeling including the complete 316 atom af unit cell , core - hole and correlation effects demonstrate that holes mainly reside on chains and thus ladders are hole - depleted .
this finding resolves long standing controversial interpetations by several authors based on semi - empirical models .
the analysis by ncker et al.@xcite is in agreement with our findings based on an @xmath30-@xmath31 approach .
it is clear , however , that their arguments were based on a too limited description since correlations and long range af behavior consisting of a 5-fold chain periodicity must be included to obtain the corrrect interpretation of the experimental spectra . in ladders ,
majority of holes reside in leg sites , while only a tiny minority populates rung sites .
further experiments and calculations are under consideration to verify our suggestion that the small affinity of rung sites in the parent compound explains the 1d - 2d cross - over going from the under - doped to the fully doped side in the phase diagram of ( la , y , sr , ca)@xmath0cu@xmath1o@xmath2 .
our work unambiguously answers the out - standing question of where the holes reside in sr@xmath0cu@xmath1o@xmath2 .
this is crucial to understanding all phenomena occurring in the phase diagram of this family of compounds , ranging from charge - ordered antiferromagnetism to superconductivity and is a prerequisite to all investigations based on correlated models of spin ladder and chain compounds.@xcite
m. c. and c. g. acknowledge discussions with f. mauri .
calculations were performed at the idris superconducting center ( project number 96053 ) .
s. t. acknowledges support from the croatian ministry of science , education and sports under grant 035 - 0000000 - 2836 .
n. ncker , m. merz , c. a. kuntscher , s. gerhold , s. schuppler , r. neudert , m. s. golden , j. fink , d. schild , s. stadler , v. chakarian , j. freeland , y. u. idzerda , k. conder , m. uehara , t. nagata , j. goto , j. akimitsu , n. motoyama , h. eisaki , s. uchida , u. ammerahl , a. revcolevschi , phys .
b , 14384 ( 2000 ) s. notbohm , p. ribeiro , b. lake , d. a. tennant , k. p. schmidt , g. s. uhrig , c. hess , r. klingeler , g. behr , b. bchner , m. reehuis , r. i. bewley , c. d. frost , p. manuel , and r. s. eccleston , phys . rev .
lett . * 98 * , 027403 ( 2007 ) j. schlappa , t. schmitt , f. vernay , v. n. strocov , v. ilakovac , b. thielemann , h. m. ronnow , s. vanishri , a. piazzalunga , x. wang , l. braicovich , g. ghiringhelli , c. marin , j. mesot , b. delley , and l. patthey , phys .
lett . * 103 * 047401 ( 2009 ) | the hole distribution in sr@xmath0cu@xmath1o@xmath2 is studied by low temperature polarization dependent o k near - edge x - ray absorption fine structure measurements and state of the art electronic structure calculations that include core - hole and correlation effects in a mean - field approach .
contrary to all previous analysis , based on semi - empirical models , we show that correlations and antiferromagnetic ordering favor the strong chain hole - attraction . for the remaining small number of holes accommodated on ladders , leg - sites are preferred to rung - sites .
the small hole affinity of rung - sites explains naturally the 1d - 2d cross - over in the phase diagram of ( la , y , sr , ca)@xmath0cu@xmath1o@xmath2 . |
discrete - time quantum walks ( dtqws ) are unitary quantum automata and can be viewed as formal generalizations of classical random walks .
they were first considered in a systematic way by meyer @xcite in the context of quantum computing , following the seminal work of feynman @xcite , who was looking for path discretizations to build a fermionic path integral , and aharonov @xcite , who pointed out the parallel with classical random walks .
dtqws have been realized experimentally with a wide range of physical objects and setups @xcite , and are studied in a large variety of contexts , ranging from fundamental quantum physics @xcite to quantum algorithmics @xcite , solid - state physics @xcite and biophysics @xcite .
it has been shown recently that the continuous limit of several dtqws defined on regular ( 1 + 1)d spacetime lattices coincides with the dynamics of dirac fermions coupled , not only to electric fields @xcite , but also to arbitrary non - abelian yang - mills gauge fields @xcite and to relativistic gravitational fields @xcite .
gravitational waves ( gws ) @xcite are of great interest , both theoretically and experimentally , and their effects on dtqws are thus worth investigating .
linear gws ressemble electromagnetic waves .
in particular , gws can be expanded as a superposition of plane waves and each plane wave as a superposition of two polarization states , both polarizations being perpendicular to the direction of propagation .
the effect of these plane gws on matter is thus typically studied in the polarization plane and it makes little sense to envisage the action of gws on @xmath1d dtqws . performing a valid study of how gws influence dtqws thus requires building dtqws coupled to ( 1 + 2)d gravitational fields .
we start by presenting a new family of @xmath0d dtqws whose continuous limit coincides with the dynamics of a dirac fermion coupled to arbitrary ( 1 + 2)d gravitational fields .
the construction of this family is inspired by the @xmath1d construction presented in @xcite .
the dtqws in the @xmath0d spacetime depend on two parameters , which code for the mass of the walker and for the finite spacing of the lattice , and on four time- and space - dependent angles .
we then show how to choose these four angles to describe linear gws .
we finally focus on pure shear gws , which couple the dynamics in the two directions of the spatial lattice , and examine in detail their action on the dtqws .
our main results are the following . on large spatial scales , pure shear gws do not modify the polarizations of the fermion eigen - modes , but rescale all energies by a common factor . on smaller scales typically comparable to two or three lattice steps , both polarizations and energies
are modified in a non - trivial way .
this effect is particularly salient on interference patterns between the fermion eigen - modes .
a final section discusses the construction of the dtqws and mentions several avenues open to further study .
consider a quantum walker moving on a two - dimensional lattice ( discrete space ) with nodes labelled by @xmath2 .
let @xmath3 label discrete time and @xmath4 be a certain time- and position - independent basis of the two - dimensional coin hilbert space of the walker , which will be used implicitly in the rest of the paper .
the state of the walker at time @xmath5 and point @xmath6 is described by a two - component wave function @xmath7 .
the collection @xmath8 will be denoted by @xmath9 .
the time evolution of @xmath9 is fixed by a time - dependent unitary operator @xmath10 : @xmath11 this unitary operator involves two real positive parameters @xmath12 and @xmath13 and four time- and space - dependent angles @xmath14 , @xmath15 , @xmath16 and @xmath17 . the choice and construction of this operator will be discussed at length in the final section of this paper .
the operator @xmath10 consists essentially in a rather complicated combination of rotations in spin space and of translations in physical space along the two directions of the lattice .
its precise definition is as follows : @xmath18 \boldsymbol{\pi } \\ & \ \ \ \ \ \times \big [ \mathbf{w}_2(\theta^{21}_{j})\mathbf{w}_1(\theta^{11}_{j})\big ] \mathbf{q } \left ( \nu , m - t ( \boldsymbol{\theta } ) /4 \right ) , \nonumber \end{aligned}\ ] ] where @xmath19 for @xmath20 , and @xmath21 .
the operator @xmath22 is @xmath23.\ ] ] the operators @xmath24 , @xmath25 , are defined by @xmath26 \
\mathbf{r}(\theta_j ) \ , , \end{aligned}\ ] ] where @xmath27 is the spin - dependent translation operator in the @xmath28 spatial direction , while @xmath29 and @xmath30 are rotations in the coin hilbert space : @xmath31 \nonumber \\ & \left(\mathbf{s}_{2 } \psi_j \right)_{p_1 , p_2 } = \left [ \begin{array}{c } \psi^{-}_{j , p_1,p_2 + 1 } \\
\psi^{+}_{j , p_1,p_2 - 1 } \end{array } \right ] , \end{aligned}\ ] ] @xmath32 with @xmath33 \nonumber \\ \mathbf{r}(\theta ) & = \left [ \begin{array}{cc } i \cos(\theta/2 ) & i \sin(\theta/2 ) \\ - \sin(\theta/2 ) & \cos(\theta/2 ) \end{array } \right].\end{aligned}\ ] ] finally , @xmath34,\ ] ] and @xmath35 with @xmath36 for any quantity @xmath37 defined on the spacetime lattice , and @xmath38= \left[\cos \theta^{kl } \right ] .
\label{eq : defe}\ ] ] note that @xmath39 ( i ) is non local in the time @xmath5 and ( ii ) vanishes if @xmath40 $ ] is either diagonal or independent of @xmath5 .
it is convenient to introduce two three - dimensional ob- jects , @xmath41 $ ] and @xmath42 $ ] , @xmath43 , whose spatial parts coincide respectively with @xmath44 and @xmath45 . more precisely , we define @xmath46 , @xmath47 and @xmath48 , @xmath49 for @xmath50 .
these two objects allow to rewrite @xmath51 in the more compact form @xmath52 where @xmath53 = \mbox{diag}(1 , -1 , -1)$ ] , @xmath54 is the totally antisymmetric symbol , with @xmath55 , and the finite difference operator @xmath56 has arbitrary spatial components .
the einstein summation convention has also been adopted in ( [ eq : tcomp ] ) . in the continuous limit
, the parameter @xmath13 will be interpreted as the mass of the quantum walker and the four time- and space - dependent angles will code for the components of a curved metric in @xmath0-dimensional spacetime .
the @xmath57 s will then code for a @xmath58-bein or triad and the @xmath59 s will code will code for its dual .
finally , the parameter @xmath12 will play the role of an infinitesimal which goes to zero with the spacing of the lattice .
to investigate the continuous limit of our walk , ( [ eq : walk ] ) , we proceed as in @xcite .
we introduce a dimensionless spacetime step @xmath60 , and interpret @xmath61 and @xmath62 , @xmath63 , @xmath64 , @xmath65 as the values taken by a wave function @xmath66 and by four functions @xmath14 , @xmath15 , @xmath16 , @xmath17 at spacetime point @xmath67 . the limit @xmath68 is investigated by taylor expanding ( [ eq : walk ] ) at first order in @xmath60 . for the continuous limit to exist
, the zeroth - order terms must cancel each other .
this is possible only if @xmath12 goes to zero with @xmath60 .
considering now the scaling @xmath69 , the first - order terms deliver a schrdinger - like equation for @xmath66 : @xmath70 where @xmath71 + q \
, , \ ] ] with @xmath72 and @xmath73 \nonumber \\
\gamma^1 & = \left [ \begin{array}{cc } 0 & 1 \\ -1 & 0 \end{array } \right ] \\
\gamma^2 & = \left [ \begin{array}{cc } i & 0 \\ 0 & -i \end{array } \right ] .
\nonumber\end{aligned}\ ] ] equation ( [ eq : hamiltonian_dirac ] ) is the hamiltonian form @xcite of the dirac equation for a spin-@xmath74 fermion of mass @xmath13 in @xmath0d spacetime with metric @xmath75 where the @xmath57 s are the components of the @xmath58-bein or triad @xmath76 on the coordinate basis @xmath77 in the tangent space , and the @xmath78 s are the components of the dual triad @xmath79 on the dual coordinate basis @xmath80 . all components of the dual triad with one and only one of the indices equal to the time index @xmath81 vanish , and @xmath82 , so that : @xmath83 = \left [ \begin{array}{c c c } 1 & 0 & 0 \\ 0 & g_{11 } & g_{12 } \\ 0 & g_{21 } & g_{22 } \end{array } \right ] , \label{eq : defg}\ ] ] where @xmath84 also of interest is the explicit expression of the matrices @xmath85 and @xmath86 in terms of the angles defining the dtqw : @xmath87 \nonumber \\ b^2 & = \left [ \begin{array}{cc } -\cos \theta^{21 } & -i \cos \theta^{22 } \\
i \cos \theta^{22 } & \cos \theta^{21 } \end{array } \right ] .
\end{aligned}\ ] ]
gws represent weak gravitational fields propagating in spacetime .
technically , they are particular solutions of the einstein equations linearized around the vacuum . after the proper choice of coordinate system ( gauge ) , these einstein equations linearized around the vacuum essentially simplify into poisson equation for two independent metric components .
gws can thus be expanded in fourier modes .
hence , of particular interest are the plane gws . in the so - called traceless gauge with
coordinates @xmath88 , the metric of a monochromatic plane gw of pulsation @xmath89 propagating along the @xmath90 direction takes the standard form @xmath91 with @xmath92.\ ] ] here , the @xmath93 s are the components of the minkowski metric in inertial coordinates , @xmath94 and @xmath95 are two arbitrary functions which represent two polarization states and @xmath96 is a small parameter which traces the perturbative nature of the waves .
the gw affects the metric only in planes of constant @xmath90 parallel to the @xmath97 plane and it is therefore customary to study its effect on matter located in these planes .
the reduced ( 1 + 2)d metric in a plane of constant @xmath90 , say @xmath98 , takes the form : @xmath99 where @xmath100 and @xmath101 .
let @xmath102 and @xmath103 be two arbitrary real constants and perform the following change of coordinates : @xmath104 since gws are perturbative solutions of einstein equations valid only at first order in @xmath96 , the components of the metric in the new coordinate system @xmath105 need only be computed at the same order in @xmath96 and one finds : @xmath106 we can choose the triad @xmath107 such that its only non - vanishing components on the coordinate basis @xmath108 are @xmath109 , @xmath110 , and @xmath111 .
identify now @xmath112 introduced in the preceding section with @xmath105 .
a ( 1 + 2)d dtqw presented in sec .
[ sec : first ] will simulate a dirac fermion propagating in the metric ( [ eq : metrictxy ] ) if @xmath113 since @xmath96 is an infinitesimal and we are working at first order in @xmath96 , the third equation can be solved immediately by choosing @xmath114 .
now , the perturbative computation of the gw from einstein equations only makes sense if @xmath115 is bounded .
one can then always find @xmath102 and @xmath103 which make both @xmath116 and @xmath117 positive .
the first and second equations are then solved by @xmath118 and @xmath119 . we will now focus on pure shear gws for which @xmath120 . we can then choose @xmath121 .
these are particularly interesting because they the couple the dynamics in the @xmath122 and @xmath123 directions .
the dtqws are entirely defined by the single angle @xmath114 , the mass parameter @xmath13 and the value of @xmath124 . to make the discussion definite , we set @xmath125 and @xmath126 . the continuous limit
is then recovered by considering wave functions which vary only on scales much greater than unity .
since the advancement operator @xmath10 depends only on the time t and not on @xmath122 and @xmath123 , the dtqws are best analyzed in fourier space .
let @xmath127 be an arbitrary function defined on the infinite lattice .
one can write ( see @xcite ) : @xmath128 with @xmath129 each fourier mode @xmath130 , @xmath131 , is dynamically independent of the other ones .
its evolution between time @xmath132 and time @xmath133 is fixed by the operator @xmath10 in fourier representation , that depends on @xmath134 and @xmath135 through exponentials which describe in fourier space the translations in physical space . since the dtqws perform pairs of jumps in each direction ( see ( [ eq:1d_block ] ) ) , the time - evolution operator in fourier space in best conceived as a function of @xmath136 and @xmath137 .
we will denote this operator by @xmath138 , making also explicit the dependence with respect to @xmath96 and @xmath139 .
a straightforward computation at first order in @xmath96 delivers @xmath140 , with @xmath141\ ] ] and @xmath142,\ ] ] with @xmath143 and where the superscript @xmath144 denotes complex conjugation .
this operator controls the entire dtqw dynamics .
in particular , at time @xmath145 , the energies corresponding to any given @xmath146 are the eigenvalues of @xmath138 and the associated eigenvectors define the eigen - polarizations of the spin-@xmath74 fermion .
thus , a generic shear gw changes in a time - dependent way the polarization and the energy of the quantum walker .
the only modes which are not affected by pure shear gws are those for which both @xmath147 and @xmath148 vanish .
a direct computation shows that , in the brillouin zone @xmath149 , for @xmath134 and @xmath135 ] both functions vanish simultaneously only for @xmath151 , @xmath152 .
for all these wave vectors , the unperturbed operator @xmath153 coincides with unity .
thus , the only modes on which a shear gw has no influence are those which are unaffected by the dynamics of the unperturbed ( _ i.e. _ ` free ' ) dtqw .
they are all ` small - scale ' ( _ i.e. _ of the same order of magnitude as the lattice spacing ) except the @xmath154 mode , which corresponds to a uniform wave function .
this mode can not be populated if space is infinite .
the operator @xmath155 can be further characterized by its eigen - values .
they are complex conjugate to each other and have as common modulus @xmath156 . the function @xmath157 is plotted in figure 1 .
it admits four absolute maxima of identical amplitude approximately equal to @xmath158 .
these maxima are approximately located at @xmath159 .
generally speaking , the higher @xmath160 , the greater the influence of gws on dtqws .
thus , pure shear gws typically maximally influence dtqws on small scales comparable to two or three lattice steps ( one lattice - step corresponds to @xmath161 ) . of the eigen - values of @xmath155 . ]
we will now present in detail two examples .
the first one deals with large - scale modes _
i.e. _ modes which vary on scales much larger than the lattice spacing .
the other example is not restricted to large - scale modes and shows how the interference pattern between two fermion eigen - polarizations is changed by a gw .
large scales are those relevant to the continuous limit and correspond to values of @xmath162 and @xmath163 close to zero .
one finds , at first order in @xmath162 and @xmath163 : @xmath164 . \nonumber\end{aligned}\ ] ] the eigenvalues @xmath165 and associated eigenvectors of this operator read , at first order in @xmath96 , @xmath162 , and @xmath163 : @xmath166 and @xmath167,\ ] ] which do not actually depend on @xmath96 and @xmath168 . since the eigen - energies
@xmath169 are by definition related to the eigenvalues @xmath165 by @xmath170 , one finds that , at first order in @xmath96 : @xmath171 thus , at large scales , the lowest - order effect of the pure shear gw is simply to dilate space isotropically by a factor @xmath172 .
this changes the eigen - energies accordingly , but not the polarization of the associated eigen - spinors .
as mentionned above , a shear gw generically changes both the frequency and the polarization of a fermion eigen - mode .
one way to visualize this effect in a physically meaningful manner is to show how gws modify the interference patterns between two fermion eigen - modes .
consider for example the two wave vectors @xmath173 and @xmath174 and the two initial polarizations @xmath175 \nonumber \\ \psi^{2 } & = & \left [ \begin{array}{c } -\frac{i}{\sqrt{2}}\\ \frac{1}{\sqrt{2 } } \end{array } \right].\end{aligned}\ ] ] the first is an eigen - polarization of @xmath176 , associated to the eigenvalue @xmath177 and the second is an eigen - polarization of @xmath178 , associated to the eigenvalue @xmath179 .
in particular , both eigenvalues are the same if @xmath180 and the free dtqw then does not modify the interference pattern of the two fermion eigen - modes .
this is not so anymore in the presence of a gw , which changes the interference pattern at each time step @xmath5 by a contribution proportional to @xmath181 .
let @xmath182 be the initial density of the two interfering fermion eigen - modes with @xmath180 and @xmath183 the density after one time step of the dtqw .
the interesting quantity is the relative density difference per unit @xmath184 _ i.e. _ @xmath185 a direct computation shows that @xmath186 and @xmath187 figure 2 shows contours of @xmath182 and @xmath188 in the @xmath189 plane for @xmath190 , which corresponds to the maximum effect of the gw ( see below ) .
both @xmath182 and @xmath188 depend on @xmath122 and @xmath123 only through @xmath191 .
figure 3 shows how the profile of @xmath192 as a function of @xmath193 changes with @xmath194 .
the net effect of the gw can be measured by @xmath195 .
one finds that : @xmath196 this function is even in @xmath194 and symmetric with respect to @xmath190 .
figure 4 displays its profile on the interval @xmath197 .
the absolute maximum of @xmath198 is reached for @xmath190 , which corresponds to an original wave vector @xmath199 _ i.e. _ to twice the scale of the lattice spacing .
the factor @xmath200 with respect to the lattice spacing comes from the fact that the dtqw makes two jumps in each direction at each time step .
the function @xmath201 admits two local maxima , situated approximately at @xmath202 and @xmath203 , which are much smaller than the absolute maximum at @xmath190 .
the function @xmath201 tends to zero with @xmath194 _ i.e. _ at large spatial scales .
this is so because the interfering polarizations have identical energy in the absence of gws and we have seen in the preceding section that the only effect gws at large scales is to multiply all energies by a common factor , and this does not change the interference pattern of the two waves under consideration .
finally , @xmath201 vanishes for @xmath204 , @xmath205 and for @xmath161 _ i.e. _ for the original wave number @xmath206 , which corresponds to the lattice spacing . and of the the relative density variation @xmath207 in the @xmath189 plane .
the mesh represents the lattice on which the dtqw is defined . ] and of the the relative density variation @xmath207 in the @xmath189 plane .
the mesh represents the lattice on which the dtqw is defined . ] as a function of @xmath193 for various values of @xmath194 . ]
we have introduced new two - dimensional dtqws whose continuous limit coincides with the dynamics of massive spin-1/2 dirac fermions propagating in a curved spacetime .
we have then shown how to use these dtqws to simulate the influence of gws on dirac fermions .
we have finally focused on pure shear gws and investigated in detail how these influence dtqws on both large and small scales . on large spatial scales ,
pure shear gws do not modify the polarizations of fermion eigen - modes , but rescale all energies by a common factor . on smaller scales typically comparable to two or three lattice steps , both polarizations and energies
are modified in a non - trivial way .
this effect is particularly salient on interference patterns between fermion eigen - modes .
let us now discuss various aspects of our work .
the new ( 1 + 2)d dtqws introduced in this paper are strongly inspired by the ( 1 + 1)d dtqws introduced in @xcite . as shown this earlier work , the coupling of a dtqw with a gravitational field in ( 1 + 1)d can only be retained if the walker performs at each time - step not one , but two jumps in the spatial direction .
thus , one would expect that coupling dtqws with gravitational fields in ( 1 + 2)d could be obtained by letting the walker perform two jumps in each spatial direction at each time - step .
these jumps are represented by the operators @xmath208 and @xmath209 which we will discuss in more detail below .
these @xmath210 jumps however do not suffice and do not take into account the shear generated by generic gravitational fields .
this shear couples directions and manifest itself by non - diagonal metric coefficients which do not vanish identically . to take the shear into account
, the quantum walker has to perform two extra jumps in each direction at each time - step .
these two extra - jumps are represented by the operators @xmath211 and @xmath212 .
now , choosing any given representation of the ( 1 + 2)d clifford algebra implies breaking the symmetry between the two space directions @xmath213 and @xmath214 and the operator @xmath22 is included to trace this symmetry breaking .
the operator @xmath215 endows the dtqw with a mass @xmath13 and provides the ` mass - like ' extra - term @xmath216 which is present in the ( 1 + 2)d dirac equation .
the operators @xmath217 themselves deserve some more comments .
each operator involves two jumps ( represented by the operators @xmath27 s ) and two mixing operators ( represented by the @xmath30 s ) . as shown in @xcite for the @xmath1d , involving only these operations would deliver the correct dirac dynamics in the continuous limit , but not in a standard ` fixed ' spin basis .
the operators @xmath218 and @xmath29 compensate for this basis difference .
dtqws simulating the interaction of a dirac field with a gravitational one can also be constructed in a different way .
the other construction was presented in @xcite for the ( 1 + 1)d case and an extension to higher space - time dimensions has just been completed .
this extension also allows fermions to have internal degrees of freedom .
both constructions use a regular lattice , but they differ in nearly all other aspects .
the construction of dtqws used in this paper is tailor - made to deliver dtqws whose continuous limit coincides with the dirac equation .
the other construction considers very general unitary quantum automata in large - enough hilbert spaces .
these admit a whole class of pdes as continuous limits , including the dirac equation . as a consequence
, each construction has its pros and cons , which we shall now discuss .
as apparent from the above material , our construction works only in synchronous coordinate systems , for which all mixed time - space metric components identically vanish and the time - time component is identical to unity .
this does not restrict the gravitational fields one can take into account , since all gravitational fields admit local synchronous coordinates .
but it forces the lattice to be regular in precisely these synchronous coordinates , whereas the other construction allows the lattice to be regular in arbitrary coordinates . on the other hand , the time - advancement operator @xmath10 used in our construction only involves gauge - invariant _ i.e. _ intrinsic aspects of the gravitational fields , while the time - advancement operator used in the other construction mixes both gauge - invariant and gauge - dependent aspects of the gravitational field . thus ,
if one adopts the construction used in this paper , the operator @xmath10 itself can be viewed as a discrete gauge - invariant gravitational field .
the gauge dependence of standard continuous physics then manifests itself only in the choice of coordinates when one performs the continuous limit _
i.e. _ in the correspondence between the continuous coordinates @xmath219 and the lattice indexes . on the other hand , the other construction encodes both the gauge - invariant aspects of the gravitational fields and a choice of coordinates into a single object _
i.e. _ in the time - advancement operator .
and one is still free to choose continuous coordinates as one wishes .
also , the other construction requires doubling the number of components of the wave function , the reduction to the usual number of components appearing only in the continuous limit .
in particular , our 2-component dtqw wave functions can be considered as discrete generalizations of the usual ( 1 + 2)d spinors , whereas the wave functions used in the other construction , having 4 instead of 2 components , are more like couples of spinors .
it thus seems that both constructions have their own strong and weak points .
the construction used in this paper is perhaps more physically grounded and the other one may be of more interest to computer scientists .
let us finally mention of few extensions of this work .
the first one is rather straightforward and consists in allowing the operators @xmath220 and @xmath221 to be arbitrary members of @xmath222 . the extra angles which will appear code for arbitrary discrete electromagnetic fields ( see @xcite ) .
the other extension consists in incorporating in the model extra internal dimensions .
this should in particular deliver dtqws where the time - advancement operator unifies arbitrary discrete yang - mills fields with discrete gravitational fields ( see @xcite for ( 1 + 1)d dtqws simulating dirac fermions in arbitrary yang - mills fields ) .
g. s. engel , t. r. calhoun , r. l. read , t .- k .
ahn , t. manal , y .- c .
cheng , r. e. blankenship , and g. r. fleming .
evidence for wavelike energy transfer through quantum coherence in photosynthetic systems . , 446:782786 , 2007 . | a new family of discrete - time quantum walks ( dtqws ) propagating on a regular @xmath0d spacetime lattice is introduced .
the continuous limit of these dtqws is shown to coincide with the dynamics of a dirac fermion interacting with an arbitrary relativistic gravitational field .
this family is used to model the influence of arbitrary linear gravitational waves ( gws ) on dtqws .
pure shear gws are studied in detail . on large spatial scales ,
pure shear gws do not modify the polarizations of the fermion eigen - modes , but rescale all energies by a common factor . on smaller scales typically comparable to two or three lattice steps , both polarizations and energies
are modified in a non - trivial way .
this effect is particularly salient on interference patterns between the fermion eigen - modes . |
quantum error correction ( qec ) @xcite and fault - tolerant quantum computation @xcite demonstrate that quantum information can in principle be stored and manipulated coherently for arbitrarily long times despite the presence noise .
the general framework of qec is the following .
redundancy is introduced by encoding the information of system @xmath1 into a larger system @xmath2 .
the image of @xmath1 in @xmath2 characterizes a code , while a particular embedding of @xmath1 into @xmath2 is called an encoding .
the system @xmath2 is subjected to some noise .
partial measurements whose outcomes are known as theerror syndrome " are performed on @xmath2 . conditioned on this error syndrome , a recovery operation is applied to @xmath2 in order to restore its original information .
this last step , called decoding " , is the subject of the present study . in the absence of structure in the code , we know from a classical result @xcite
that finding the optimal recovery is np - hard . for practical purposes ,
one must either use codes with lots of structure which typically offer poorer performances or settle for suboptimal recovery .
residual errors after decoding are therefore of two varieties : those due to the information - theoretic limitations of the code and those arising from suboptimal decoding procedures . in the past decades
, considerable progress has been made towards understanding this tradeoff in the classical setting ( see e.g. @xcite and references therein ) .
central to these advancements is the use of the message passing decoding algorithm pioneered by gallager @xcite which often leads to near - optimal decoding .
this technique was recently introduced in the quantum realm by ollivier et al .
@xcite for the decoding of low density parity check ( ldpc ) codes ( see also @xcite for related work ) .
concatenation of block codes is widely used in quantum information science and is a key component of almost all fault tolerant schemes ( a noticeable exception is topological quantum computing @xcite ) . as the name suggests , the system @xmath2 that redundantly encodes the information of system @xmath1 can itself be encoded in a yet larger system @xmath3 , adding an extra layer of redundancy .
provided the initial error rate is below a threshold value , every extra level of concatenation should reduce the probability of error after decoding , so concatenation can in principle be repeated until the error is below any desired value . in this article , we demonstrate an efficient message passing algorithm that achieves optimal ( maximum likelihood ) decoding for _ concatenated block codes _ with uncorrelated noise .
we numerically investigate the message passing algorithm using the 5 qubit code @xcite and steane s 7 qubit code @xcite and compare their performances to the commonly used blockwise minimal - distance decoder ( based on a local rather than global optimization ) .
the advantages of the message passing algorithm are substantial . on the one hand , for the 5 qubit code used on a depolarizing channel
, the message passing algorithm can correctly decode the information for a noise level up to at least 0.1885 ( the exact threshold is probably the hashing bound @xmath4 ) compared to the values 0.1376 previously established using blockwise decoding @xcite .
for steane s code , this enhancement is even greater going from 0.0969 @xcite to at least 0.188 . on the other hand , away from these noise thresholds , the probability of error decreases at a significantly higher rate using optimal decoding .
for instance , for a 0.1 depolarizing channel and using 4 levels of concatenation of the 5 qubit code , the probabilities that the blockwise decoding and the optimal decoding fail to correctly identify the error differ by more that 3 orders of magnitude . as a consequence , a decoding error probability @xmath5 for any @xmath6
can be achieved with a substantially reduced error correction overhead .
our presentation of the stabilizer formalism follows @xcite , see @xcite for the general theory .
denote by @xmath7 , @xmath8 and @xmath9 the three pauli matrices and by @xmath10 the @xmath11 identity matrix .
the group @xmath12 is the multiplicative group generated by the pauli matrices and the imaginary unit @xmath13 .
the @xmath14-qubit pauli group @xmath15 is the @xmath14-fold tensor product of @xmath12 .
we denote @xmath16 the pauli matrix @xmath7 acting on the @xmath17th qubit for @xmath18 and similarly for @xmath8 and @xmath9 .
note that the @xmath16 s and the @xmath19 s are a generating set of @xmath15 , i.e. @xmath20 . the clifford group on @xmath14 qubits @xmath21 is the largest subgroup of the unitary group @xmath22 that maps @xmath15 to itself under the adjoint action .
the encoding of @xmath23 qubits into @xmath14 qubits can be specified by a matrix @xmath24 .
@xmath25 is a unitary matrix acting on @xmath14 qubits that are distributed in 3 different sets .
the first @xmath23 logical " qubits contain the information to be encoded in the @xmath14 qubits ; the next @xmath26 stabilizer " qubits are set to the state @xmath27 ; and finally the remaining @xmath28 gauge " qubits are in arbitrary states . the image of the pauli operators acting on the first @xmath23 qubits are known as logical pauli operators @xmath29 and @xmath30 .
the image of the @xmath9 pauli operators acting on qubits @xmath31 are called stabilizer generators @xmath32 whereas the image of the @xmath7 operators acting on those qubits are called pure errors @xmath33 . finally , the image of the pauli operators acting on the remaining @xmath34 qubits are called gauge operators @xmath35 and @xmath36 .
the stabilizer generators @xmath37 mutually commute , so can be simultaneously measured .
the outcome of that measurement is called the error syndrome @xmath38 .
since the @xmath26 stabilizer qubits are all in state @xmath39 prior to encoding , we conclude that in the absence of noise the encoded state should be a @xmath40 eigenstate of all stabilizer generators , thus the error syndrome should be all ones .
a non - trivial syndrome therefore indicates that an error has corrupted the register , and the task of decoding consists in finding the optimal recovery procedure given an error syndrome .
to address the decoding problem , note that @xmath41 .
in other words , any element @xmath42 can be written , up to an irrelevant phase , as @xmath43 where @xmath44 is a product of logical pauli operators , @xmath45 is a product of pure errors , and @xmath46 is a product of gauge operators and stabilizer elements .
moreover , this decomposition can be found by running the circuit @xmath25 backward , which is efficient since @xmath24 @xcite .
@xmath45 is completely determined by the syndrome : @xmath47 appears in @xmath45 if and only if the @xmath17th syndrome bit is @xmath48 .
the value of @xmath46 is irrelevant because the information encoded in the @xmath14 qubits is invariant under the action of any @xmath46 .
this reflects the fact that the stabilizer qubits are initially set to @xmath39 and that the gauge qubits are in random states .
thus , to undo the effect of an error @xmath49 , one needs to identify the most likely value of @xmath50 given @xmath51 , or equivalently given @xmath52 . for simplicity , we will focus on pauli channels where errors @xmath49 are elements of @xmath15 distributed according to @xmath53 . given this probability @xmath53 over
@xmath15 one can compute the conditional probability @xmath54 using @xmath55\delta[\cl(e ) = l ] p(e ) \\ & = & \sum_g p(e = ltg ) , \label{eq : prob_block}\end{aligned}\ ] ] where @xmath56 denotes the indicator function and @xmath57 takes all possible combinations of stabilizer generators and gauge operators .
given a finite block size @xmath14 , these probabilities can be computed and the optimal decoding @xmath58 can be evaluated .
decoding a block code thus consists of looking in a table containing the values of @xmath59 for each @xmath60 .
typically and in particular for a non - degenerate code over the depolarization channel @xmath59 corresponds to the minimal distance decoder @xmath61 where @xmath62 is the error acting on the fewest number of qubits and that is compatible with the observed syndrome .
concatenation is realized by encoding the @xmath14 qubits of the code in an other code .
there is no need for this other code to be identical to the original one .
however to simplify the presentation , we will assume that the same code is used at every concatenation layer and that it encodes a single qubit in @xmath14 qubits ; generalizations are straightforward .
this procedure can be repeated @xmath63 times at the expense of an exponentially growing number of physical qubits @xmath64 .
the number of stabilizer generators grows roughly as @xmath65 ( it is a geometric sum ) ; thus the syndrome can take @xmath66 different values .
thus , even for moderate values of @xmath63 , it is not feasible to construct a lookup table giving the optimal decoding procedure for each syndrome value . what is generally done to circumvent this double exponential blowup
is to apply the optimal recovery _
independently for each concatenation layer _ ( see e.g. ( * ? ? ?
* chap . 6 ) and references therein ) .
one first measures the syndrome from each of the @xmath67 blocks of @xmath14 qubits of the last layer of concatenation , and optimally decodes them using the lookup table .
one then moves one layer up and applies the same procedure to the @xmath68 blocks of the second - to - last layer , etc .
when the initial error rate is below a certain threshold value , the probability @xmath69 that this procedure fails to correctly identify @xmath44 decreases doubly - exponentially with @xmath63 .
hence , this decoding scheme based on hard decisions for each concatenation layer is efficient , leads to a good error suppression , but is nonetheless suboptimal .
let @xmath70 be the syndrome of the @xmath17th block of the @xmath71th concatenation layer .
denote @xmath72 the collection of syndromes whose stabilizers act non - trivially on the physical qubits associated to the @xmath17th block of the @xmath71th concatenation layer : these sets can be defined recursively by @xmath73 with the initialization @xmath74 .
finally , denote @xmath75 all the syndromes from the layers @xmath71 to @xmath63 .
( see the above figure for a pictorial representation of @xmath76 , @xmath77 , and @xmath78 . ) then , @xmath79 is the set of all syndromes and maximum likelihood decoding consists in finding @xmath80 .
this probability can be factorized by conditioning on the logical errors of the second layer @xmath81 : @xmath82 \frac{p(\bl_2,\bs_1)}{p(\bs_1 ) } \nonumber \\ & = & \sum_{\bl_2 } \delta[l_1 = \cl(\bl_2 ) ] \frac{p(s_1|\bl_2,\bs_2)p(\bl_2,\bs_2)}{p(s_1,\bs_2 ) } \nonumber \\ & = & \sum_{\bl_2 } \delta[l_1 = \cl(\bl_2 ) ] \delta[s_1 = \cs(\bl_2 ) ] \frac{p(\bl_2|\bs_2 ) p(\bs_2)}{p(s_1,\bs_2 ) } \nonumber\\ & = & \sum_{\bl_2 } \frac{\delta[l_1 = \cl(\bl_2 ) ] \delta[s_1 = \cs(\bl_2)]}{p(s_1|\bs_2 ) } \prod_{j=1}^n p(l^{(j)}_2|\bs^{(j)}_2 ) .
\nonumber\end{aligned}\ ] ] above , @xmath83 denotes the syndrome associated to the error pattern @xmath84 .
this series of manipulations repeatedly uses bayes rule and the fact that the syndrome and logical error of level @xmath71 are completely determined given the logical errors of layer @xmath85 .
the last step relies on the important assumption that the channel is memoryless , or more specifically , that the noise model does not correlate qubits across distinct blocks ( errors on qubits in the same block could be correlated ) .
equation ( [ eq : factor ] ) shows that by conditioning on the logical errors of each concatenation layer , the _ factor graph _ associated to the function @xmath86 is a tree , as depicted in the above graph .
we have thus reduced optimal decoding to a sum product problem ( known as tensor network contraction in quantum information science @xcite ) on a tree graph which can be solved exactly and efficiently in the number of variables using a message passing algorithm ( also known as belief propagation ) ; see @xcite , and references therein .
let us describe this algorithm in a general setting .
the factor graph is a bipartite graph , and vertices from the two partitions are decorated with circles and boxes .
circle vertices are labeled @xmath87 and each one carries a variable @xmath88 with value in a discrete set .
box vertices are labeled @xmath89 , and each one contains a function @xmath90 that depends on the variables @xmath88 from the adjacent circles @xmath91 , collectively denoted @xmath92 .
the goal is to compute marginals @xmath93 where @xmath94 indicates that @xmath88 is omitted from the set and @xmath9 is a normalization factor . to this end , messages @xmath95
are passed from the circles to the boxes and messages @xmath96 are passed from the boxes to the circles following the rules @xmath97,\end{aligned}\ ] ] where @xmath98 means all neighbors of @xmath99 excluding @xmath100 , and similarly for @xmath101 .
note that these messages are functions of the discrete variables @xmath88 ( i.e. they are vectors ) .
the @xmath95 messages are initialized to the constant function 1 . for a tree graph
, the desired marginal is obtained from these messages after a number of steps equal to the depth of the variable @xmath88 and is given by @xmath102 , where @xmath23 is a normalization factor . in the case of interest , circles carry logical operators and a box labeled @xmath103 carries the function @xmath104\delta[s_{m+1}^{(j ) } = \cs(l_m^{(nj - n+1)},\ldots , l_m^{(nj)})]$ ] , where the syndrome is fixed by the measurements . to complete the picture , extra box vertices carrying the function eq .
( [ eq : prob_block ] ) need to be attached to the bottom leaves of the graph .
the factor @xmath105 can be evaluated by normalizing the obtained distribution .
thus , we can efficiently evaluate @xmath106 , for all @xmath103 .
computing only @xmath106 leads to simplification of the message passing rules ; in particular , messages need only to flow from the bottom to the top of the graph . ] and the optimal recovery is the @xmath107 maximizing this function .
the advantage of the message passing algorithm over the minimal distance decoder comes from the fact that it does not throw away useful information @xcite . instead of computing the most likely recovery and passing it on to the next level of coding
, the entire list of probability of possible recoveries , conditioned on the observed syndrome , is passed on .
in other words , the original channel is composed with the syndrome measurement , and projected onto the logical algebra to yield a conditionally renormalized " channel .
following the tradition for benchmarking qec techniques , we investigate the performance of the message passing decoding algorithm using a depolarization channel , where each qubit is independently subjected to the channel @xmath108 we use the 5 qubit code @xcite concatenated with itself up to @xmath109 times , for an overhead of 9,765,625 physical qubits per logical qubit .
pauli errors @xmath110 are generated by picking each @xmath14 single - qubit operator independently according to the probability @xmath111 , @xmath112 .
the associated logical error @xmath44 and syndromes @xmath113 are computed exactly .
these syndromes are used by a blockwise decoding routine yielding an estimate @xmath114 and by a message passing routine yielding the optimal decoding @xmath115 .
a decoding is declared incorrect when its estimate differs from @xmath44 .
this is repeated a large number of times ( @xmath116 ) to evaluate the probability @xmath69 that the decoding gives an incorrect estimate . as a function of the level of concatenation @xmath63 for different depolarization rate @xmath117 and @xmath118 .
the diamonds are from the message passing algorithm , and the circles are from the blockwise decoding .
all data are from samples of @xmath119 encoded qubits.,height=201 ] figure [ fig : flow ] shows the probability of incorrect decoding @xmath69 for both the blockwise and the optimal decoding as a function of the level of concatenation @xmath63 and for different channel parameters @xmath120 ranging from 0.13 to 0.19 . for the blockwise decoding , @xmath69 ceases to decrease with @xmath63 for values of @xmath121 .
this reflects the fact that the threshold of this decoding technique for this particular code is about 0.1376 @xcite , so all curves except the 0.13 one are above the threshold . on the other hand ,
optimal decoding succeeds in decreasing the error probability for values of @xmath120 up to at least 0.1885 , but appears to fail at @xmath122 .
we conjecture that the exact value of this threshold is the hashing bound @xmath123 , where the single - qubit coherent information vanishes and is the highest threshold any non degenerate code can achieve @xcite .
results obtained from steane s code @xcite show a quite similar behavior , with at least 94@xmath0 increase of threshold going from 0.0969 @xcite to at least 0.188 and appears to fail at 0.1885 .
an interesting feature of the @xmath124 curves obtained from optimal decoding is their non - monotonicity .
blockwise decoding , on the other hand , always yields monotonic curves for this type of channel ; thus its global behavior under concatenation can be predicted from a single level of coding .
this is because decoding is performed independently on each concatenation layer . with the optimal decoder , information about the syndromes
is propagated from one layer of concatenation to the next through the conditionally - renormalized channel that ceases to be depolarizing and varies from one qubit to the other .
thus , non - monotonicity of the @xmath124 curves is a signature of the global optimization performed by the message passing algorithm . for @xmath125 and 0.05 .
diamonds obtained from samples of @xmath126 encoded qubits .
circle were produced using an exact numerical technique similar to that of ref .
@xcite.,height=201 ] figure [ fig : exponential ] shows the behavior of @xmath69 as a function of @xmath63 away from the threshold values , i.e. in the natural operating regime of the code .
again , the advantages of the message passing algorithm are considerable . after 4 rounds of concatenation for @xmath125 ,
message passing fails with a probability of roughly @xmath127 , whereas this probability is well above @xmath128 for blockwise hard decoding .
it takes @xmath129 layers of concatenation for the blockwise decoding to reach comparable performances .
again , results obtained from steane s code show an even larger gap .
finally , we once again stress that the message passing outputs the probability of an error @xmath130 rather than a particular value of @xmath130 . a hard decision can then be made based on this probability .
we observe that when decoding succeeds , @xmath131 is typically very close to one ( e.g. 0.999 for @xmath132 ) whereas when it fails it is relatively low ( typically 0.7 ) ; the algorithm knows that it is failing . this flagging " of errors offers a great advantage when post - selection is an option .
the possibility of operating the algorithm with soft inputs , i.e. noisy syndrome measurements , is also of interest in several circumstances .
we have demonstrated an efficient message passing algorithm for the optimal decoding of concatenated quantum block codes on a memoryless channels .
numerical results show substantial benefits of our approach over the widely used blockwise hard decoding , including an increase of error thresholds and a greater error suppression rate .
message passing algorithms have been used on graphs with loops ( describing e.g. ldpc codes , turbo - codes , or channels with memory ) and often yield near - optimal decoding .
the quantum generalization of these schemes , including quantum ldpc codes @xcite and quantum turbo - codes @xcite , are promising avenues for the realization of a quantum information technologies .
techniques reminiscent of message passing have been used to beat the hashing bound but were not efficiently implementable @xcite : efficient decoding may now be within reach using our techniques .
a hard " message passing scheme was also used in @xcite to obtain high fault - tolerant error thresholds : a full - fledge message passing scheme although not optimal for correlated errors that are typically present in fault - tolerant schemes should further improve this threshold and may significantly reduce the resource overhead .
i thank harold ollivier for several useful conversations on message passing algorithms and quantum error correction , and graeme smith and jon yard for comments on this paper .
this work was supported in part by the gordon and betty moore foundation through caltech s center for the physics of information , by the national science foundation under grant no .
phy-0456720 , and by canada s nserc . | we consider the problem of optimally decoding a quantum error correction code that is to find the optimal recovery procedure given the outcomes of partial check " measurements on the system . in general , this problem is np - hard .
however , we demonstrate that for concatenated block codes , the optimal decoding can be efficiently computed using a message passing algorithm .
we compare the performance of the message passing algorithm to that of the widespread blockwise hard decoding technique .
our monte carlo results using the 5 qubit and steane s code on a depolarizing channel demonstrate significant advantages of the message passing algorithms in two respects .
1 ) optimal decoding increases by as much as 94@xmath0 the error threshold below which the error correction procedure can be used to reliably send information over a noisy channel .
2 ) for noise levels below these thresholds , the probability of error after optimal decoding is suppressed at a significantly higher rate , leading to a substantial reduction of the error correction overhead . |
recent observations by the fermi satellite have confirmed that both long and short gamma - ray bursts are capable of emission at energies @xmath1 gev .
the rest - frame energies inferred for gamma rays from long - duration gamma - ray burst ( grbs ) 080916c and 090902b @xcite and short grb 090510 @xcite are all in excess of @xmath260 gev .
these findings raise the question of whether emission might occur at still higher energies , and these gamma rays might be observable by fermi , or by ground - based instruments such as magic , veritas , and hess .
electron - positron pair - production from interactions between gamma rays and uv - ir photons is an important effect at these energy scales , modifying the spectra of distant extragalactic sources and shrouding the highest - energy emission .
this effect could thus be a powerful technique for understanding cosmological background radiation @xcite the photon - photon optical depth per unit path length for a gamma ray of energy @xmath3 is @xmath4 where @xmath5 is the local density of target photons at energy @xmath6 , and @xmath7 is the angle of interaction . here
@xmath8 is the cross section for interaction , which peaks at twice the minimum energy for pair creation @xmath9 at energies below 1 tev , it is the optical , uv , and near ir photons that are responsible for attenuation .
higher energy gamma rays have sufficient energy to produce electron positron pairs in interactions with longer - wavelength ir photons . in general , attenuation at a cosmological scale is expected to dominate the pair - production opacity , as the strength of the effect scales with path length through the intervening photon field .
analyses of the interstellar radiation fields within the milky way find significant attenuation of high energy photons only at several tens of tev , with negligible absorption at lower energies @xcite .
there is strong evidence , however , that long - duration grbs are associated with deaths of massive stars , and therefore occur in star - forming regions of galaxies , and that the hosts of these objects often exhibit considerable star - formation rates @xcite , and therefore significant number densities of uv and optical photons .
the prevailing theory for these bursts is that they are associated with some subset of type 1c supernovae resulting from the deaths of short - lived stars @xcite . in this paper
, we examine the optical depth for gamma - rays from a grb produced by interactions with `` local '' photons within a host galaxy . loosely speaking
, the opacity from pair - production is proportional to the product of path length and density of target photons at the relevant energy scale .
if the line - of - sight to the grb passes through a region of intense flux from nearby stars , the resulting optical depth could conceivably be equal to or greater than that from intergalactic background photons .
attenuation effects from photon fields on several different length scales could be pertinent .
in particular , we will discuss here absorption at sub - parsec scales from the photon field of a binary companion to the grb progenitor , parsec - scale absorption within the star - forming region of the progenitor site , and finally opacity due to the kiloparsec - scale galactic disk of the host galaxy . in [ approximations ] , we present some simple estimates of the opacity arising on each of these scales . in [ numcalcs ] , we describe a numerical calculation performed to calculate the attenuation for star - forming galactic disks , followed by our conclusions .
as a simple preliminary approximation , an estimate of @xmath10 can be found for a given region by the product @xmath11 where @xmath12 is the typical size of the region of interest and therefore the approximate pathlength of the gamma ray through the photon field , @xmath13 is the number density of photons near energies corresponding to the peak cross section ( maximum in @xmath8 ) , and @xmath14 is this approximated cross section .
here @xmath15 is the thomson cross section .
here we consider the attenuation of gamma rays by a high - mass binary companion of the grb progenitor .
one class of grb creation models invokes a companion star at a few stellar radii to form a rapidly spinning helium core @xcite , a distance of @xmath16 to @xmath17 cm for a high - mass stars .
the effect of radiation from this companion on gamma - ray flux is limited by the radius @xmath18 at which the high energy flux is emitted from the outflow of the progenitor .
most models place this emission at a distance @xmath19 cm @xcite , typically greater than the binary separation .
we have then for the opacity @xmath20 where @xmath21 is the companion luminosity , and @xmath7 is the angle of photon - photon interactions from equation ( [ eq : opdep ] ) .
if we consider the opening angle of the jet from relativistic beaming @xmath22 , then the maximum @xmath23 interaction angle possible is @xmath24 the radius @xmath18 , for models in which the burst is triggered by the action of the companion , is expected to be larger than the progenitor
companion distance @xmath25
. meanwhile the high bulk lorentz factor , @xmath26 , in most grb models @xcite leads to a tightly beamed radiation cone .
this angular factor therefore greatly reduces the likelihood of interaction for gev - scale grb emission . in figure
[ fig : odbin ] , we show a summary of our results for gamma - ray opacity based on binary separation and emission radius , assuming that the direction of emission is perpendicular to the orbital plane . lines in this plot indicate the minimum gamma - ray production distance at which photon - photon opacity is less than unity , for a given separation distance between the progenitor and its companion .
we assume here a low - metallicity main - sequence companion at two different stellar masses , 10 and 100 m@xmath27 , which are modeled as blackbody sources . from @xcite , we assign these two cases luminosities of @xmath28 and @xmath29 l@xmath27 , and effective blackbody temperatures of @xmath30 and @xmath31 k. these results suggest that while significant attenuation of gamma - rays can occur very close to the progenitor , it is unlikely to be dominant at the the radius at which gamma - ray are usually thought to be produced . for the case of attenuation by the photon field within a large molecular cloud
, we can look to surveys of these complexes within the milky way ( mw ) and nearby galaxies to estimate the extent and magnitude of a typical photon field .
clouds in the mw typically have sizes @xmath32 of @xmath33 to @xmath34 pc , and an h@xmath35 gas mass as high as @xmath36 to @xmath37 m@xmath27 @xcite , and other authors confirm similar results for m31 and the large magellanic cloud @xcite . with the assumption that 5 to 10% of this gas forms stars within a time comparable to the lifetime of high - mass stars , 20 - 30 myrs @xcite , we can estimate the highest value likely for our estimate of @xmath10 in equation ( [ eq : tauest ] ) . for a salpeter imf @xcite @xmath38
we have @xmath39 high - mass ( @xmath40 m@xmath27 ) stars per solar mass of gas converted into stars .
if we assume that these stars have a typical luminosity of @xmath41 l@xmath27 which is emitted mostly in the uv and is responsible for most of this radiation at these wavelengths , then we estimate that the most massive molecular clouds could have @xmath42 massive stars producing @xmath43 l@xmath27 in optical - uv photons .
the energy density in the cloud can be approximated as @xmath44 about 6.7 x @xmath45 erg @xmath46 under our assumptions .
if we assume that this energy exists in the form of uv - optical ( i.e. 3 ev ) photons , in a cloud of size @xmath47 pc , then we have an photon number density of 1.4 x @xmath48 @xmath46 .
applying equation ( [ eq : tauest ] ) to find the opacity over a path through the cloud , we estimate @xmath49 from uv photons in the most massive stellar birth clouds .
we have not accounted for dust extinction in these systems , which can be substantial and would reduce the uv flux and @xmath10 .
while it is generally accepted that long - duration grbs are associated with core - collapse supernovae , the relation between grb host galaxies and the population of star - forming galaxies appears to be complex , and recent studies seem to dismiss the idea that grb rates follow star formation in an unbiased manner . surveys of host galaxies can provide a guide for us to choose general parameters in this project .
a survey of grb host galaxies at a variety of redshifts by @xcite found that hosts can typically be described by an exponential profile with a median scale radius of 1.7 kpc ; none followed the de vaucouleurs profile associated locally with elliptical galaxies .
@xcite analyzed 30 grb hosts , and found typical star - formation rates of 0.01 to 10 m@xmath50yr from an assumed unobscured uv , though the inclusion of dust could raise these values considerably .
this paper also mentions that some galaxies in their sample with more poorly constrained seds could have star formation rates as high as a few hundred m@xmath50yr .
a study of 46 hosts by @xcite found star - formation rates of 0.01 - 36 m@xmath27/yr , with a mean of 2.5 , after correction for dust extinction .
stellar masses in these samples ranged from @xmath37 to @xmath51 m@xmath27 in the former , and @xmath52 to @xmath53 m@xmath27 in the latter .
however , @xcite has found that even galaxy hosts for obscured `` dark '' grbs typically have normal optical colors overall , suggesting that the dust is unevenly distributed or local to the grb . in their survey ,
@xcite estimate that @xmath225% of grb hosts have dust extinction a@xmath54 .
@xcite found an average dust extinction of @xmath55@xmath56@xmath57 in a subset of 10 galaxies in their sample ; 2 of these galaxies had @xmath58 . following our method of the last two sections
, we can perform an order - of - magnitude estimate of @xmath10 for a rapidly star - forming galaxy . for a given star - formation rate
, we can estimate the uv flux using the approximation @xcite @xmath59 for a star - formation rate of 100 m@xmath50yr , and a galaxy size of 1 kpc , we estimate a total uv output of 2 x @xmath60 erg / s , and consequently a photon energy density similar to that of the molecular cloud in the previous section . from equation ( [ eq : tauest ] )
, we estimate a gamma - ray opacity @xmath61 if this energy is emitted as 3 ev photons .
these finding suggest that the emission from a star - forming galaxy , acting over a path length of a kpc or more , could be a barrier to gev gamma - rays from a grb within a galaxy .
however , this is only an order - of - magnitude estimate of @xmath10 , with a high star - formation rate and no account of the absorption and re - emission of starlight by dust .
higher star - formation rates than we assume here have been seen in grb - selected submillimeter galaxies @xcite , and in extreme cases galaxies of this type can have rates exceeding 1000 m@xmath50yr @xcite , albeit with heavy dust obscuration . to better understand the circumstances under which the photon flux from stars and dust in a galaxy might lead to significant attenuation of gamma rays on a kiloparsec scale , we have performed a set of numerical calculations which are described in the next section .
in order to accurately estimate the contribution to opacity arising from a compact , rapidly star - forming galaxy , we have performed a simple simulation of a galactic disk , and calculated numerically the resulting uv - ir photon field and from this the gamma - ray optical depth along a given line - of - sight .
the stellar spectra are calculated using the population synthesis code of .
we assume that the stellar population in our galaxy is formed at a constant rate over a characteristic timescale @xmath62 .
stars are then distributed in a galactic disk of exponential scale length @xmath63 and maximum radius @xmath64 , and with thickness @xmath65 .
note that the half - light radius for a exponential profile is @xmath66 .
our calculation includes a simple model for dust absorption and re - emission .
starlight is absorbed using the model of , in which starlight is absorbed by 2 components : an interstellar medium ( ism ) component affecting all stars , and a molecular cloud component affecting only stars younger than @xmath37 yrs .
we assume the standard ratio for these opacities , in which the ism is responsible for 30% of the opacity for young populations , and the rest arises from absorption within molecular clouds . the ism dust component is assumed to have a homogeneous distribution in the same disk as the stellar population .
all energy absorbed by dust is re - emitted in the ir .
we calculate the emission at wavelengths above 4 @xmath67 m using the templates of @xcite , which are based on _ spitzer _ observations of local galaxies .
we consider face - on dust extinction factors of 0.1 , 1.0 , and 3.0 in the v - band attenuation @xmath68 , broadly corresponding to very low , moderate , and large amounts of reddening . from this normalization ,
the extinction at other wavelengths is determined by multiplying by a power law in wavelength with index -0.7 . in galaxies with @xmath69 , greater than 90% of starlight from stellar populations younger than 500 myrs
is absorbed by dust and re - emitted in the ir , so this corresponds to a ( u)lirg - like mode with highly - obscured star - formation .
all of our galaxy spectra are created using a metallicity of [ fe / h ] = -1.65 .
this is somewhat less than the typical metallicity of 1/6 solar that was seen in the sample of @xcite ; we do not expect this to notably impact our final results .
figure [ fig : galspect ] shows the sample spectra for galaxies of two different star - formation timescales ( @xmath62 ) and three values for face - on dust extinction ( @xmath70 , 1.0 , and 3.0 ) , normalized to the same mass . for our calculation , we have considered timescales over the range 10 myr @xmath71 10 gyr . in figure
[ fig : galspect_comp ] , we show a comparison with the sed for a starbursting grb host , grb000418 , as presented in @xcite . this source is representative of the type of grb hosts that have the highest inferred star - formation rate and
are therefore expected to have the largest opacity to gamma rays .
we do not include a luminosity contribution from agn in our calculation . however , our templates for dust re - emission are based on observations of local galaxies , and therefore some degree of agn contamination in the ir unavoidably exists .
some degree of evolution in the typical ir properties of galaxies likely occurs between local redshifts and those near the peak of the star - formation era ( z@xmath22 ) ; @xcite estimates that these are not larger than a factor of two . in our model
we assume a scaling law for the galactic disk radius with stellar mass of @xmath72 with normalization of 500 pc at @xmath73 m@xmath27 ; this slope is based upon a rough fit to the half - light radius results from a survey of star - forming galaxies in @xcite , and is consistent with the more compact galaxies seen in the sample presented in @xcite .
this particular scaling produces galaxies that are more compact than 18 of 20 grb hosts with @xmath74 ( @xmath75 ) fits in the study of @xcite .
since the disk height is also proportional to @xmath63 in our model , it should be pointed out that for a given set of stellar and dust properties the photon flux density scales as @xmath76 , and therefore gamma - ray opacity is proportional to @xmath77 , after accounting for the increase in path length .
the uv , optical , and near - ir spectral properties of a unit mass of stars are defined in our model by the assumed dust attenuation factors and star - formation timescale , therefore for gamma - ray energies affected solely by these wavelengths ( @xmath78 gev ) the gamma - ray opacity scales as @xmath79 when these factors are held fixed .
the primary ingredients of our model are the star - formation timescale , dust attenuation factor , and galaxy size and stellar mass which are conjoined in equation ( [ eq : galmasssize ] ) , as discussed in the previous section .
the other important considerations are the location of the grb within the disk and the geometry of the line - of - sight ( los ) path from the grb to the observer .
we fix the location of the grb at a radius @xmath80 ( 25 % of the simulated radius @xmath81 ) and centered within the disk height .
we consider two possibilities here for the los : a `` skimming '' mode in which the path is toward the galactic center at an angle of incidence of 10 degrees to the disk , and a geometry in which the path is face - on relative to the disk , with angle of incidence 90 degrees .
the former gives a nearly maximal amount of attenuation for a given galaxy , while the second gives a closer to average amount of absorption .
we have calculated opacities for gamma rays from energies of 1 gev to 100 tev . in figure
[ fig : opdep ] , we show the opacity @xmath10 for gamma rays at a particular energy for several sets of galaxy properties . changing from the skimming los to one that is perpendicular to the galactic disk
is found to reduce @xmath10 by a factor of 2 3 .
the results in figure [ fig : galspect ] for a galaxy of mass @xmath82 m@xmath27 indicate that attenuation at energies @xmath55 tev is significantly less than @xmath83 even for the optimal `` skimming '' geometry and a fast stellar buildup time of 30 myrs .
as mentioned in the previous section , the attenuation is expected to be approximately proportional to @xmath84 , if we assume that galaxy volume scales with mass .
the only significant attenuation in these cases is due to the population of mid- and far - ir photons created by thermal and pah emission ; these being much more numerous than the uv , optical , and near - ir photons of direct starlight . in figures [ fig : opthick1 ] and [ fig : opthick2 ] we examine the relationship between galaxy mass , star - formation timescale ( or star - formation rate ) and the minimum gamma - ray energy for which @xmath85 ; these two plots are respectively for cases of low and moderate dust extinction ( @xmath86 0.1 and 1.0 ) .
results for the largest dust extinction , @xmath87 , are not substantially different from the 1.0 case and are not shown .
displaying our results in this way can tell us the approximate energy scale at which photon - photon opacity becomes an important effect on the high - energy spectrum of radiation escaping from the vicinity of the galaxy .
in general , the opacity only becomes significant at multi - tev energies , where gamma - rays have sufficient energy to interact with the re - emitted ir photons .
as seen in figure [ fig : opdep ] , the opacity curve is approximately flat between 100 gev and 1 tev .
opacity at these energies is generally less than 1 , and we only find optically - thick galaxies in cases of minimal dust extinction , high mass , and short star - formation timescale , with star - formation rate @xmath88 m@xmath50yr .
our results suggest that it is unlikely the radiation field within a single star - forming region could impact the high - energy ( gev and tev ) gamma - ray emission from a grb , although our understanding of these regions is limited to nearby galaxies .
significant attenuation due to the emission of a massive binary companion is possible only if the radius of high - energy photon emission is less than predicted by most radiation models .
large , rapidly star - forming galaxies can form an optically - thick barrier only at multi - tev energies , where attenuation with mid - ir photons can occur , regardless of line - of - sight geometry or dust extinction .
however , photons at these energies would already be strongly attenuated by extragalactic background light for all but the closest grbs ( z ) .
creating enough photons in the uv , optical and near ir to significantly impact transmission of gev photons is only possible with star - formation rates well in excess of 1000 m@xmath27/yr , even with a minimal amount of dust attenuation .
this requirement could be lowered if galaxies are much more compact than we have assumed , though the galaxy sizes used here are already more compact than most of those in the sample of @xcite , and these authors mention that the densest objects at high redshift tend to be elliptical galaxies with much lower photon densities .
clumpy or irregular star formation is one possibility that we have not considered in our numerical calculations , which assume a smooth exponential disk .
the total amount of attenuation through the host galaxy is proportional to the photon density over the line - of - sight path .
therefore the total @xmath10 for the galaxy would only be increased significantly from our predictions if a large fraction of the total stars in the galaxy were placed in a much more compact distribution than we have assumed with our exponential profiles .
this region would have to contain a much higher density of young stars than the star - forming regions seen in the mw and lmc , based on the order - of - magnitude analysis of [ mc ] . merely rearranging star formation into clumpy regions throughout the galaxy
would not , in general , affect the total photon field or @xmath10 , unless those regions were were positioned near the line - of - sight path in a stochastic irregularity . in conclusion
, both long and short grb observations can be safely used as powerful tools for understanding the extragalactic background light .
we thank j. granot , x. prochaska , and j. primack for useful discussions , m. michalowski and d. watson for providing us with their spectral fits to grb host galaxies , and the anonymous referee for comments which helped improve this paper . r.g .
acknowledges support from a fermi guest investigator grant during part of this research .
e.r - r acknowledges nsf grant ast-0847563 , and thanks the packard foundation for support . | high - energy photons emitted from gamma - ray bursts ( grbs ) are subject to pair - production interactions with lower energy photons , leading to an effective optical depth . in this paper
, we estimate the opacity resulting from photon fields located at various distances from long grb sites : that of the binary companion to the massive stellar progenitor , that of the star - forming molecular cloud containing the grb , and the total photon field of the host galaxy
. the first two photon fields are found to be transparent for most reasonable sets of assumptions about these systems . in the case of galactic radiation fields
, we have performed several numerical simulations to calculate the expected opacities for 2 different line - of - sight geometries through the host galaxy , and include a full accounting of the infrared radiation produced by the absorption and re - radiation of starlight by dust .
the optical depth for gev gamma - rays , due to direct starlight is less than unity for all host galaxies . at higher energies , @xmath010 tev
, a spectral cutoff can occur due to the rapidly increasing number of mid- to far - ir intra - galactic photons reradiated by dust .
photons in the extragalactic background light therefore remain the only relevant source of photon - photon opacity for ongoing grb observations with fermi lat , and potential future detections with ground - based gamma - ray telescopes . |
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 and fis2011 - 28853-c02 - 02 , and the generalitat valenciana in the program prometeo , 2009/090 .
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 .
this work is also partly supported by the national natural science foundation of china under grant 11105126
. 999 m.ablikim _ et al . _
[ bes collaboration ] , phys .
lett . b * 685 * , 27 ( 2010 ) .
j. beringer et al .
( particle data group ) , phys . rev . *
d86 * , 010001 ( 2012 ) .
b. aubert _ et al . _ [ babar collaboration ] , phys .
d * 74 * , 091103 ( 2006 ) .
a. martinez torres , k. p. khemchandani , l. s. geng , m. napsuciale and e. oset , phys .
d * 78 * , 074031 ( 2008 ) .
l. alvarez - ruso , j. a. oller and j. m. alarcon , phys .
d * 80 * , 054011 ( 2009 ) .
s. coito , g. rupp and e. van beveren , phys . rev .
d * 80 * , 094011 ( 2009 ) .
geng , and e. oset , phys .
d * 79 * , 074009 ( 2009 ) .
r. molina , d. nicmorus and e. oset , phys .
d * 78 * , 114018 ( 2008 ) . c. garcia - recio , l. s. geng , j. nieves and l. l. salcedo , phys .
d * 83 * , 016007 ( 2011 ) .
m. albaladejo and j. a. oller , phys .
* 101 * , 252002 ( 2008 ) .
a. v. anisovich , c. a. baker , c. j. batty , d. v. bugg , l. montanet , v. a. nikonov , a. v. sarantsev and v. v. sarantsev _ et al .
_ , phys .
b * 542 * , 19 ( 2002 ) .
j. a. oller and u. g. meissner , phys .
b * 500 * , 263 ( 2001 ) .
j. c. nacher , e. oset , h. toki and a. ramos , phys .
b * 455 * , 55 ( 1999 ) . | the bes data on the @xmath0 reaction show a clear enhancement in the @xmath1 mass distribution close to the threshold of this channel .
such an enhancement is usually a signature of a @xmath2 resonance around threshold , which in this case would correspond to an @xmath3 state with quantum numbers @xmath4 .
a state around @xmath5 results from the interaction of the @xmath6 using the local hidden gauge approach .
we show that the peak observed in @xmath0 naturally comes from the creation of this @xmath3 state with mass and width around @xmath7 and @xmath8 , respectively .
a second analysis , model independent , corroborates the first result , confirming the relationship of the enhancement in the invariant mass spectrum with the @xmath3 resonance .
the decay @xmath0 was measured for the first time by the bes collaboration @xcite with the aim of searching for the @xmath9 resonance through the decay @xmath10 .
however , no clear enhancement in the @xmath1 mass distribution was found near @xmath11 gev .
although unknown at the time of the experiment , it is easier to understand at present that the @xmath9 , now catalogued as @xmath12 in the pdg @xcite , with quantum numbers @xmath13 , does not couple to @xmath1 .
the resonance couples strongly to the @xmath14 @xcite , and it was found in refs .
@xcite , using fadeev calculations and related methods , that it was made of a cluster of @xmath15 , with the @xmath16 highly correlated into an @xmath17 . on the other hand , since the @xmath18 and the @xmath19 mesons have quantum numbers @xmath20 and @xmath21 , respectively , the decay @xmath0 constitutes the ideal reaction to look for an @xmath3 state , with quantum numbers @xmath22 , coupling to an @xmath23wave @xmath24 pair .
both on the theoretical and experimental sides , the study of @xmath3 states above @xmath25 gev is very scarce @xcite . in ref .
@xcite , the work of ref .
@xcite on the @xmath26 interaction was extended to @xmath27 using the local hidden gauge formalism for vector
vector interaction and a unitary approach in coupled channels .
this interaction generates resonances in different strangeness - isospin - spin channels . in the @xmath22 sector , a resonance was found with mass and width around @xmath28 and @xmath29 mev , respectively . this resonance is dynamically generated from the interaction of @xmath30 and @xmath31 . as shown in ref .
@xcite , given its quantum numbers , it can not couple to other vector vector or pseudoscalar pseudoscalar channels , which makes its observation difficult .
this state can not be clearly identified with any of the @xmath3 states listed in the pdg , and it is not found in ref .
@xcite ( or , rather , it is pushed down to lower energy ) .
reference @xcite uses different dynamics , based on spin - flavour @xmath32 symmetry , and a regularization of the loops which is different from the one in ref .
@xcite .
both approaches share qualitatively many features , but differ somewhat in the numerical results . a different dynamical approach to the vector - vector interactions
is provided in @xcite , using a massive yang - mills field theoretical framework . on the experimental side , in an analysis of @xmath33 annihilation channels done in ref .
@xcite , an @xmath3 state with mass and width @xmath34 and @xmath35 mev is found , very far from the one predicted in ref .
@xcite . on the other hand , a careful inspection of the bes results in ref .
@xcite , shows an enhancement in the invariant mass spectrum of the @xmath1 around @xmath36 mev ( see more details from fig . 4 of ref .
@xcite ) , which might hint to a sizeable contribution from an @xmath3 state strongly coupling to the @xmath1 channel . however , as explained above , the search of the @xmath12 resonance was the main aim of ref .
@xcite , and no attention was paid to this enhancement .
decay .
[ feydgm],scaledwidth=47.0% ] in the present work , following the formalism of ref .
@xcite , we shall make a first study of the role of the @xmath3 state ( which is dynamically generated by the @xmath30 and @xmath37 interaction ) in the @xmath0 decay .
the decay mechanism is shown in fig .
[ feydgm ] .
let us denote by @xmath38 the bare production vertex for @xmath0 .
we assume that this bare vertex is of a short range nature , _
i.e. _ , just a coupling constant in the field theory language . to take into account the final state interaction of the @xmath6 pair ( which , in the end , will generate the @xmath3 resonance )
, one has to make the resummation of the diagrams shown in fig .
[ feydgm ] .
let us also denote by @xmath39 and @xmath40 ( where @xmath41 is the @xmath6 invariant mass ) the full amplitude and the bare interaction vertex , respectively , of an @xmath42 @xmath6 pair .
these are related by : @xmath43 where @xmath44 is the loop function for the @xmath6 pair .
this function is divergent , and it can be regularized both with a cutoff prescription or with dimensional regularization in terms of a subtraction constant @xcite .
here we shall make use of the dimensional regularization scheme , which introduces a subtraction constant , @xmath45 , where @xmath46 is a regularization scale , taken here as @xmath47 ( notice that there is only one free parameter ) .
moreover , since the @xmath30 and @xmath37 have large total decay widths , they should be taken into account .
for that purpose , the @xmath48 function is the loop function for two stable particles of masses @xmath49 and @xmath50 , @xmath51 , but convoluted in the masses @xmath49 and @xmath50 with the mass distribution of the two vectors , as done in refs . @xcite .
the distribution is considered in the range @xmath52 , where @xmath53 and @xmath54 are the nominal mass and width of the @xmath30 meson , respectively .
explicit expressions for @xmath55 and @xmath48 can be found in ref .
@xcite . with all these definitions ,
the full amplitude @xmath56 for the process @xmath57 can be written , according to the diagrams in fig .
[ feydgm ] , as : @xmath58 where the last equality follows from eq . .
then we can easily get the @xmath59 invariant mass spectrum for the @xmath60 as @xcite : @xmath61 with @xmath62 a global constant factor ( that absorbs the unknown vertex @xmath38 ) which will be fitted to the data , and @xmath63 is the @xmath19 momentum in the @xmath18 rest frame , @xmath64 where @xmath65 is the khlen or triangle function . in eq .
, @xmath66 is the momentum of the @xmath67 in the @xmath59 center of mass system , but , as for the @xmath48 function , we must take into account the large width of the @xmath30 meson .
that is , @xmath66 is the convolution with the mass distribution of the two @xmath67 ( in the range @xmath68 ) , of the momentum @xmath69 , given by : @xmath70 the @xmath1 invariant mass spectrum , eq . , depends on the amplitude @xmath71 , and the most important ingredient in this amplitude is the @xmath72 transition potential @xmath55 .
we will discuss two approaches to this potential in this work . in principle , this potential is fixed from the hidden gauge unitary approach of ref . @xcite .
there , it was found that : @xmath73 where @xmath74 , being @xmath75 and @xmath76 the mass of the @xmath77 meson and the pion weak decay constant , respectively .
the term @xmath78 comes from the four vector contact term , whereas the term proportional to @xmath79 comes from the exchange of vector mesons .
the constant @xmath79 is determined by the masses of the vector mesons ( @xmath77 , @xmath80 , @xmath81 and @xmath30 ) and its value turns out to be @xmath82 . if one uses eq .
there are two free parameters for calculating the @xmath1 invariant mass spectrum of eq . .
one is the global constant @xmath62 , which does not influence the amplitude @xmath71 .
the other one is the subtraction constant @xmath45 , that completely determines now the amplitude @xmath71 .
this constant , in the absence of data and by similarity with other channels , was fixed in ref .
@xcite to @xmath83 , but this value does not allow to obtain a good reproduction of the bes data for the @xmath1 invariant mass spectrum .
it produces a peak that is narrower and lower in mass than the bump seen in the data . instead
, we find a good agreement with the data using the rather natural range @xmath84 .
the second possibility is to consider a constant potential ( in contrast to the one in eq . , which depends on @xmath85 ) , fitted to reproduce the data .
a constant potential is acceptable here given the small range of @xmath41 considered in this work ( @xmath86 ) , and would render our analysis more model independent . by considering a constant potential ,
the subtraction constant @xmath45 is no longer a free parameter .
this is so because it is an additive term in the @xmath48 function , and then any shift in this subtraction constant can be exactly absorbed in the constant potential , as can be seen in eq . .
for this reason , we fix @xmath87 in this case ( the central value of the range used above ) .
then , when using the constant potential , two free parameters need to be determined : the global constant @xmath62 , and the potential @xmath55 itself . in both approaches ,
the free parameters are fitted by means of a @xmath88 function to reproduce the @xmath1 invariant mass spectrum data of the bes collaboration , shown in fig .
[ fig : dgdm ] .
actually , the points in the latter represent the number of events once the sideband events are subtracted from the experimental points for the @xmath89 mass distribution given in ref .
@xcite . , but the data for the mass distribution shown in fig . 4 of ref .
@xcite are given without this subtraction .
] as a further constraint , we consider in the @xmath88 function the total number of events in the invariant mass range considered .
invariant mass spectrum of @xmath0 decay .
the data points are taken from ref .
@xcite .
the different lines represent the output of eq .
for the different approaches considered in this work .
the short - dashed line and the associated error band ( light blue ) represent the results of the constant potential .
the ( red ) solid , long - dashed and dot - dashed lines represent the results for the potential in eq . with @xmath90 , @xmath91 and @xmath92 , respectively .
finally , the ( green ) dotted line , and the associated error band ( dark green ) is the prediction for phase space alone.[fig : dgdm],scaledwidth=49.0% ] .values of some of the parameters used or determined in this work.[tab : params ] [ cols="^,^,^,^,^",options="header " , ] we start with the discussion of the second approach ( constant potential ) .
the best fit values of the potential @xmath55 and the global constant @xmath62 are shown in table [ tab : params ] . the blue short - dashed line in fig .
[ fig : dgdm ] represents the spectrum obtained with these parameters .
the errors of the parameters as well as the error band of the curve are given by the condition @xmath93 , where @xmath94 is the minimum @xmath88 .
it is worth noting , by inspection of fig .
[ fig : dgdm ] , that the reproduction of the data is good .
the @xmath95 is rather small , given the large errors in the experimental data .
we stress that this approach is rather model independent , since we do not assume any underlying model for the transition potential @xmath55 , but take it to be a constant .
we now consider the approach in which the potential is taken from the local hidden gauge theory , eq .
, for three different values of the subtraction constant , @xmath90 , @xmath91 and @xmath92 .
the results are shown in fig .
[ fig : dgdm ] with solid , long - dashed , and dot - dashed lines , respectively .
the fitted global constant for each case is shown in table [ tab : params ] .
the @xmath95 is similar to that obtained with the constant potential and , inspecting fig .
[ fig : dgdm ] , it is also clear that the reproduction of the data is also good .
both approaches ( constant and local hidden gauge potentials ) give results in good agreement with the data , and give hints for the presence of a resonance .
yet , we would like to see if the experimental @xmath1 invariant mass spectrum in fig .
[ fig : dgdm ] can be accounted for by solely phase space .
this is easily achieved by putting @xmath96 in eq . , and adjusting @xmath62 .
this turns out to be @xmath97 , and the fit has a larger @xmath98 .
the dotted curve and the associated error band ( dark green band ) in fig . [ fig : dgdm ] represent this result .
it is easy to see that the inclusion of an @xmath3 state , generated dynamically from the @xmath24 interaction , is crucial to achieve a fairly good description of the bes data .
the phase space alone clearly does not follow the trend of the data , even the corresponding @xmath98 could be statistically acceptable , which is due to the large errors of the experimental data .
but one observes that below @xmath99 almost every datum is above the phase space band , while they are mostly below for energies above that one .
, for @xmath72 , for the different approaches considered in this work . the notation of the lines as in fig .
[ fig : dgdm].[fig : ts],scaledwidth=49.0% ] now , we use the parameters obtained in the fits to see if the amplitude @xmath71 obtained by means of eq . with both approaches
has or has not a resonance .
for this , we plot the modulus squared of the scattering amplitude of @xmath72 in fig . [
fig : ts ] . for the case in which the potential is taken as a constant ,
the peak is located roughly at @xmath100 , whereas its width is around @xmath101 . in the local hidden gauge potential case
, we find instead @xmath102 and @xmath103 . in both cases ,
the @xmath71 matrix has a resonant shape , and the differences of about @xmath104 in the position of the peak can be accepted as systematic uncertainties of the resonance mass in our analysis .
definitely , an experimental study with more statistics would help reduce these uncertainties .
note that , because of the large error bars in the data , the statistical error band ( shown in fig .
[ fig : ts ] for the constant potential ) is quite broad .
however , this barely affects the position of the maximum , which moves by less than @xmath105 . in summary
, we have studied the @xmath106 decay with the aim of determining the existence of an @xmath3 state through @xmath107 decay . in particular , by using a constant potential and a local hidden gauge potential for the @xmath108 transition
, we calculate the distributions of the differential decay width , with which we obtain the line shape of the @xmath1 invariant mass spectrum . by fitting our results to the bes experimental data , we find that these can be fairly described by including the contributions from an @xmath3 state , while the phase space alone can not describe the data , especially the bump structure near threshold .
we also show the corresponding results for the modulus squared of the scattering amplitude of @xmath72 , from where we can get the mass and width of this @xmath3 state around @xmath109 and @xmath110 , respectively .
the analysis done here shows clearly enough that the bes data , with the enhancement of the @xmath1 invariant mass spectrum close to the threshold , call for an @xmath3 resonance with the properties given by our fit .
yet , the data could be considerably improved , and in view of the results of the present paper , it would be most advisable to do so , in order to improve on the present statistics , which would also revert into smaller systematic errors in the determination of the resonance properties . |
we consider a system interacting with its bath via a qnd interaction .
the hamiltonian is @xmath84 the second term on the rhs of the above equation is the free hamiltonian of the environment , while the third term is the @xmath7-@xmath8 interaction hamiltonian .
the last term on the rhs of eq .
( [ 2a ] ) is a renormalization inducing ` counter term ' . since @xmath85=0 $ ] , eq .
( [ 2a ] ) is of qnd type . following ref .
@xcite , taking into account the effect of the environment modelled as a squeezed thermal bath ( eqs .
( [ eq : reservoir ] ) , ( [ 2e ] ) and ( [ 2f ] ) ) , the reduced dynamics of the system can be shown to be where the explicit forms of @xmath87 and @xmath88 can be obtained from ref .
@xcite . here
@xmath89 is the eigenstate of the system hamiltonian defined in the system eigenbasis .
it can be seen from the above equation that whereas the off - diagonal elements of the reduced density matrix decay with time , the diagonal elements remain unaffected .
clearly , this feature makes a qnd @xmath7-@xmath8 interaction unsuitable for quantum deletion .
we note that eq .
( [ 2 g ] ) does not depend on the specific form of the system hamiltonian . for the case of a two - level system , eq .
( [ 2 g ] ) can be recast in the form of bloch vectors as follows @xcite : @xmath90 where @xmath91 ( @xmath92 if and only if the environmental interaction is absent ) is calculated for the @xmath32 and high @xmath46 case in ref .
it suffices for our present purpose to note that the action described by eqs .
( [ eq:2g0 ] ) is that of contracting the bloch sphere along the @xmath14-axis .
an initial pure state is driven to a mixed state that is diagonal in the computational basis , _ even at @xmath32_. thus , this is not useful for our purpose .
100 m. nielsen and i. chuang , _ quantum computation and quantum information _ , cambridge ( 2000 ) .
p. w. shor , siam j. sci .
* 26 * , 1484 ( 1997 ) ; eprint quant - ph/9508027 ; l. k. grover , phys .
* 79 * , 325 ( 1997 ) .
w. k. wooters and w. h. zurek , nature * 299 * , 802 ( 1982 ) .
a. k. pati and s. braunstein , nature * 404 * , 164 ( 2000 ) .
r. ct , nature physics * 2 * , 583 ( 2006 ) ; a. andr , d. demille , j. m. doyle , _ et al .
_ , _ ibid . , _ 636 ( 2006 )
p. zanardi and m. rasetti , phys .
lett . * 79 * , 3306 ( 1997 ) .
p. w. shor , phys .
a * 52 * , r2493 ( 1995 ) ; a. r. calderbank and p. w. shor , phys .
a * 54 * , 1098 ( 1996 ) ; a. steane , proc .
soc . , london , ser .
a * 452 * , 2551 ( 1996 ) .
l. vaidman , l. goldenberg , and s. wiesner phys .
a * 54 * , r1745 ( 1996 ) .
l. viola and s. lloyd , phys .
a * 58 * , 2733 ( 1998 ) ; d. vitali and p. tombesi , phys . rev .
a * 65 * , 012305 ( 2001 ) .
g. s. agarwal , phys . rev . a * 61 * , 013809 ( 1999 ) ; g. s. agarwal , m. o. scully and h. walther , phys . rev
. lett . * 86 * , 4271 ( 2001 ) .
d. gottesman , phys .
a * 57 * , 127 ( 1998 ) .
d. a. lidar , i. l. chuang and k. b. whaley , phys .
* 81 * , 2594 ( 1998 ) .
cai , chin .
* 21 * , 1189 ( 2004 ) ; eprint quant - ph/0401005 .
r. romano and d. dalessandro , phys .
rev . lett . * 97 * , 080402 ( 2006 ) .
r. landauer , ibm j res .
dev . * 5 * , 183 ( 1961 ) .
b. piechocinska , phys .
a * 61 * , 062314 ( 2000 ) . c. h. bennett , int .
* 21 * , 905 ( 1982 ) . c. m. caves and b. l. schumaker , phys .
a * 31 * , 3068 ( 1985 ) ; b. l. schumaker and c. m. caves , phys .
a * 31 * , 3093 ( 1985 ) .
m. s. kim and v. buek , phys .
a * 47 * , 610 ( 1993 ) .
breuer and f. petruccione , _ the theory of open quantum systems _
( oxford university press , 2002 ) .
s. banerjee and r. srikanth , eprint quant - ph/0611161 .
m. ziman , p. telmachovi , v. buek , m. hillery , v. scarani and n. gisin , phys .
a * 65 * , 042105 ( 2002 ) .
a. v. taichenachev , a. m. tumaikin , and v. i. yudin and l. hollberg , phys .
a * 63 * , 033402 ( 2001 ) .
m. atatre , j. dreiser , a. badolato , _ et al .
_ , science * 312 * , 551 ( 2006 ) . c. j. myatt , b. e. king , q. a. turchette , _ et al .
_ , nature * 403 * , 269 ( 2000 ) . | environment - induced decoherence presents a great challenge to realizing a quantum computer .
we point out the somewhat surprising fact that decoherence can be useful , indeed necessary , for practical quantum computation , in particular , for the effective erasure of quantum memory in order to initialize the state of the quantum computer .
the essential point behind the deleter is that the environment , by means of a dissipative interaction , furnishes a contractive map towards a pure state .
we present a specific example of an amplitude damping channel provided by a two - level system s interaction with its environment in the weak born - markov approximation .
this is contrasted with a purely dephasing , non - dissipative channel provided by a two - level system s interaction with its environment by means of a quantum nondemolition interaction .
we point out that currently used state preparation techniques , for example using optical pumping , essentially perform as quantum deleters .
quantum computation is well known to solve certain types of problems more efficiently than classical computation @xcite .
a basic challenge facing the realization of a quantum computer is that of fighting decoherence .
although quantum mechanical linearity endows a quantum computer with greater - than - classical power @xcite , it also imposes certain restrictions , such as the prohibition on perfect cloning @xcite and on deleting the copy of an arbitrary quantum state perfectly @xcite .
a quantum computational task can be broadly divided into three stages : ( 1 ) initializing the quantum computer , by preparing all qubits in a standard ` blank state ' ; ( 2 ) executing the unitary operation that performs the actual computation ; ( 3 ) performing measurements to read off results . the well known difficulty in realizing a quantum computer is that of shielding it from the environment during step ( 2 ) @xcite .
this is the problem of fighting decoherence , the loss of coherence in a system ( here a qubit ) , due to interaction with its environment , which has been the subject of intense research .
a variety of techniques , including quantum error avoidance @xcite , quantum error correction @xcite , quantum error prevention @xcite , dynamic decoupling @xcite , frequency modulation of the heat bath @xcite , fault tolerant quantum computation @xcite , decoherence - free subspaces @xcite , among others , exist to combat decoherence .
our main aim here , however , is to point out that in one of the stages of quatum computation , decoherence is useful , even necessary . in step ( 1 ) , we must be able to erase quantum memory of the state inherited from a previous computational task , in order to prepare the state of a quantum computer for a subsequent task .
what is required is a quantum mechanism that with high probability allows us to prepare standard ` blank states ' , usually designated by the pure state @xmath0 .
it is clear that no unitary process can achieve this , since true deletion would be irreversible , and hence non - unitary . given two distinct states @xmath1 and @xmath2 , a purported deleting operation @xmath3 should effect @xmath4 , which can not be satisfied unless @xmath5 .
further , the no - deleting theorem implies that no qubit state can be erased against a copy @xcite . a direct method for initializing the quantum computer would be to measure all qubits in the computational basis .
this results in a statistical mixture of @xmath0 s and @xmath6 s , but there is no unitary way in a closed system to flip the @xmath6 s while retaining the @xmath0 s . however , open quantum systems can effect non - unitary evolution on a sub - system of interest .
we are thus led to conclude that decoherence is in fact _ necessary _ for step 1 .
in particular , decoherence must furnish a contractive map that drives any initial state of the system towards a fixed pure state to serve as the blank state . in this work
, we use this insight to argue that decoherence can be useful to quantum computation .
in particular , we show that a dissipative interaction with an environment can serve as an effective deleter of quantum information .
our result is comparable to that presented in ref .
@xcite , where the evolution of a qubit - probe system is modified by interaction with an environment , which enables the qubit to be driven to a target state , though in function , their work aims for quantum control , whereas the present proposal is aimed at state preparation .
we briefly introduce the interaction of a two - level system ( a qubit ) with an environment ( bath ) of harmonic oscillators via a dissipative as well as non - dissipative interaction .
our model assumes that the qubits of the computer are mutually non - interacting and decohere independently .
however , it may be noted that in many situations this is a reasonable assumption .
in fact , conventional quantum error - correction codes @xcite were developed primarily to deal with independent , incoherent errors . in the dissipative case
the system - environment ( @xmath7-@xmath8 ) interaction is treated in a standard born - markov approximation . in the non - dissipative case ,
the @xmath7-@xmath8 interaction is of a quantum nondemolition ( qnd ) type .
these environmental interactions are representative of open system effects , and further , correspond to two of the most important noisy channels in quantum information theory . in particular , the dissipative type of system - environment interaction yields the ( generalized ) amplitude damping channel , while the non - dissipative interaction yields the phase damping channel @xcite .
we point out that whereas the former is useful to engineer a quantum deleter , the latter leads to a mixed state , which is unsuitable for state preparation .
this observation is compatible with the fact that erasing information is an irreversible process that dissipates energy @xcite , affirming the connection between thermodynamics and information theory @xcite .
the total hamiltonian is @xmath9 , where @xmath10 , @xmath11 and @xmath12 stand for the hamiltonians of the system , reservoir and @xmath7-@xmath8 interaction , respectively . here
the system hamiltonian is given by @xmath13 , with @xmath14 being the usual pauli matrix .
for the reservoir hamiltonian we use the standard form of a bath of harmonic oscillators , i.e. , @xmath15 .
we assume separable initial conditions , i.e. , @xmath16 , and the reservoir is assumed to be initially in a squeezed thermal state , i.e. , a squeezed thermal bath , with an initial density matrix @xmath17 given by @xmath18 where @xmath19 \exp\left(-\beta \hbar\omega_k b^{\dag}_kb_k\right ) % \sum\limits_n e^{-n\beta \hbar \omega_k } |n\rangle \langle n| \label{2e}\ ] ] is the density matrix of the thermal bath , and @xmath20 \label{2f}\ ] ] is the squeezing operator with @xmath21 , @xmath22 being the squeezing parameters @xcite .
squeezing of the bath has been shown to be useful in the suppression of decay of quantum coherence @xcite , and to modify the evolution of the geometric phase of two - level quantum systems @xcite .
hence it is of relevance to study its possible influence on the behavior of the quantum deleter .
the system - environment interaction is taken to be dissipative and of the weak born - markov type @xcite leading to a standard lindblad equation , which in the interaction picture has the following form @xcite @xmath23 where @xmath24 , @xmath25 and @xmath26 , is the planck distribution giving the number of thermal photons at the frequency @xmath27 . here
@xmath28 , and the quantities @xmath29 and @xmath30 are the environmental squeezing parameters and @xmath31 . if @xmath32 , so that @xmath33 , then @xmath34 vanishes , and a single lindblad operator suffices . from this , the bloch vectors can be obtained as @xcite @xmath35 \beta(t ) \langle \sigma_1 ( 0 ) \rangle \nonumber\\ & - & \sin(\phi ) \sinh({\gamma_0 a t \over 2 } ) e^{-{\gamma_0 \over 2}(2n + 1)t } \langle \sigma_2 ( 0 ) \rangle , \nonumber \\
\langle \sigma_2 ( t ) \rangle & = & \left[1 + { 1 \over 2 } \left(e^{\gamma_0 a t } - 1\right ) ( 1 - \cos(\phi))\right ] \beta(t ) \langle \sigma_2 ( 0 ) \rangle \nonumber\\ & - & \sin(\phi ) \sinh({\gamma_0 a t \over 2 } ) e^{-{\gamma_0 \over 2}(2n + 1)t } \langle \sigma_1 ( 0 ) \rangle , \nonumber \\
\langle \sigma_3 ( t ) \rangle & = & e^{-\gamma_0 ( 2n + 1)t } \langle \sigma_3 ( 0 ) \rangle - \frac{\left(1 - e^{-\gamma_0 ( 2n + 1)t } \right)}{2n+1 } , \end{aligned}\ ] ] where @xmath36 , @xmath37 , @xmath38 , and @xmath39 is a constant typically denoting the system - environment coupling strength , while @xmath27 is the system frequency . in eqs .
( [ eq:2b0 ] ) , the quantities @xmath40 @xmath41 are the expectation values of the respective pauli operators with respect to the intial state @xmath42 it is seen from eqs .
( [ eq:2b0 ] ) that for fixed squeezing , the asymptotic equilibrium state @xmath43 , given by @xmath44 where @xmath45 $ ] , is approached faster for stronger coupling , as might be expected .
the entire bloch sphere shrinks towards @xmath43 . in the absence of squeezing and at zero temperature ( @xmath46 )
, this action corresponds exactly to an amplitude damping channel , for which @xmath47 and thus @xmath48 .
this corresponds to the point representing the state @xmath0 ( the bloch sphere south pole in our notation ) .
for the case of finite @xmath46 but zero squeezing , this corresponds to a generalized amplitude damping channel , for which @xmath49 , giving @xmath50 , i.e. , a mixed state @xcite . in the limit of infinite temperature , @xmath51 and @xmath52 , so that @xmath43 tends to the maximally mixed state , represented by the center of the bloch sphere .
thus , the interaction with the environment provides a contractive map , such that the asymptotic state is pure ( @xmath48 ) or mixed ( @xmath50 ) , depending on environmental conditions .
it is worth stressing that the quantum deleter requires a dissipative interaction with its environment .
this may be contrasted with a nondissipative interaction , in order to shed light on why the former is necessary for our purpose . in the case of a non - dissipative , quantum nondemolition ( qnd )
interaction , the environment acts as a purely quantum dephasing channel that leaves the energy of the system unchanged @xcite .
this is illustrated briefly in appendix [ sec : p ] .
of the bloch vector ( eqs .
( [ eq:2b0 ] ) ) for various initial states , with @xmath53 , @xmath54 , temperature @xmath32 and the squeezing parameters set to zero .
the small - dashed , dot - dashed and large - dashed curves represent initial pure states with @xmath55 , @xmath56 and @xmath57 , respectively ( in this case , there is no dependence of @xmath58 on azimuthal angle @xmath59 ) .
the solid curve represents the intially maximally mixed state . in the case of the pure states ,
note that after at first becoming mixed on account of entanglement with the environment , the qubit returns to purity ( in the state @xmath0).,width=264 ] fig .
[ fig : rbma ] depicts the length of the bloch vector from eqs .
( [ eq:2b0 ] ) for various initial states ( both pure as well as maximally mixed ) as a function of time .
it can be shown from eqs .
( [ eq:2b0 ] ) that for all pure states except @xmath0 , the quantity @xmath60 , where @xmath61 is the bloch vector , is negative at @xmath62 , implying that the tip of the bloch vector plunges initially into the bloch sphere on its way towards the point representing the stationary state @xmath0 .
this is reflected in the initial reduction of the length of the bloch vector , during which time , the qubit becomes increasingly entangled with the environment , before gradually factoring out .
the proposed quantum deleter works as follows . to clear the memory of the quantum computer , the control processes brought into play to shield it from environmental decoherence ,
are turned off for a short time @xmath63 , during which each qubit in the quantum computer is assumed to interact independently with and equilibriate with its environment , assumed to be an unsqueezed vacuum bath .
we may characterize the performance of the quantum deleter in terms of fidelity , a measure of closeness of two quantum states @xcite . from eqs .
( [ eq:2b0 ] ) , we find that with probability exponentially approaching unity , the fidelity @xmath64 of a qubit approaches the designated standard ` blank state ' @xmath0 according to @xmath65^{1/2}\end{aligned}\ ] ] where @xmath66 and @xmath67 is the expectation value of @xmath14 at time @xmath62 .
observe that at any time @xmath63 , fidelity is smallest when @xmath68 is largest , i.e. , 1 .
thus , no matter what the initial state , we can lower bound fidelity as @xmath69^{1/2}.\ ] ] observe that if and only if temperature and squeezing are zero , implying @xmath47 , the asymptotic value of @xmath70 is 1 , i.e. , @xmath71 .
if any of these two conditions is not met , then @xmath49 and hence @xmath72 . therefore ,
given zero temperature and zero squeezing , one may select the ` off - shield ' duration @xmath63 that guarantees fidelity equal to or higher than any pre - selected value ( less than 1 ) . )
falls as a function of temperature ( @xmath46 , in units where @xmath73 ) until it reaches the value @xmath74 corresponding to a maximally mixed state . the case shown here corresponds to @xmath55 , @xmath53 , @xmath54 and time @xmath75 .
here we set the squeezing parameters @xmath29 and @xmath30 to zero.,width=264 ] , the latter representing an initial pure state with @xmath76 , the former an initially maximally mixed state .
the large - dashed ( maximally mixed initially ) and small - dashed ( initially pure state with @xmath76 ) lines represent the evolution of the bloch vector for the same initial states , but with @xmath77 . in all cases , @xmath32 , @xmath54 and @xmath78.,width=264 ] as seen from eq .
( [ eq : asymp ] ) , the effect of increasing temperature is to make the asymptotic state @xmath43 ever more mixed .
as @xmath79 , @xmath43 tends to the maximally mixed state .
fig .
[ fig : fidtemp ] depicts the effect of increasing temperature on fidelity .
fig .
[ fig : rbma0 ] depict the effect of squeezing on the evolution of the bloch vector length ( from eqs .
( [ eq:2b0 ] ) ) .
it can be seen that due to the presense of squeezing , the system does not return to purity in the state @xmath0 .
therefore , for optimal performance of the quantum deleter , i.e. , for high purity of the output state of the system qubit , both the squeezing parameter and temperature must be set as close to zero as possible ( i.e. , @xmath32 and @xmath80 , yielding @xmath81 and hence @xmath48 ) .
it is interesting to note that a contractive map finds another practical , but quite different ( cryptographic ) , use . in the universal quantum homogenizer ( uqh ) ,
proposed by ziman _
et al . _
@xcite , a sequence of interactions of a system qubit with a reservoir of qubits , initially prepared in an identical state @xmath82 , drives an arbitrary state towards @xmath82 . as a result ,
the uqh acts as a _ quantum safe _ with a classical key , consisting of the interaction sequence .
one might also consider the uqh being turned into a quantum deleter by setting @xmath83 .
however this would require a sequence of highly controlled operations , which would render it relatively difficult to implement , in comparison to allowing a system to cool via interaction with its environment , as required in our case .
we believe that the most interesting aspect of our work is the idea that the unavoidable problem posed by decoherence is shown to be not only useful , but indeed necessary for quantum computation , in particular for initial state preparation . by allowing a role to be played by a realistic environment , in particular , one in which squeezing is absent , it improves the chance that a quantum computer can be practically realized . in practice , what is needed is to turn off in step ( 1 ) control processes used to protect the quantum information in step ( 2 ) against decoherence due to a vacuum environment .
for example , in a quantum information processing system using ultracold atoms in a magneto - optic trap , one may allow spontaneous decay of excitations .
however , one should not switch off the trap , since this would effectively bring the system in contact with a finite temperature environment .
it would at first appear that the requirement of a vacuum bath ( @xmath32 ) places a technological hurdle . however , we note that state preparation techniques in atom - optical experiments , quantum dots , etc
. essentially involve a quantum deletion process , usually furnished by a spontaneous emission .
some examples are , preparation of a zeeman - state in an optical lattice by two - dimensional sideband raman cooling @xcite , and by laser cooling of the spin of an electron trapped in a semiconductor quantum dot @xcite . in this context , the notable experiment by myatt _ et al . _ @xcite , in which they engineered a ( nearly ) zero - temperature reservoir , is worth pointing out .
we are thankful to prof .
hema ramachandran for useful discussions . |
magnetite , fe@xmath1o@xmath2 , is the oldest magnetic material recognized 2500 years ago and is one of the most fascinating materials in present - day solid state physics .
it is because fe@xmath1o@xmath2 exhibits many interesting physical properties @xcite such as mixed valence and a metal - insulator ( mi ) transition known as the verwey transition @xcite .
furthermore , fe@xmath1o@xmath2 has potential industrial applications in magnetic multilayer devices @xcite since it has full spin polarization with a high magnetic ordering temperature and its electronic structure at ambient conditions is predicted by band calculations to be half - metallic@xcite . above the mi transition temperature , @xmath3 , fe@xmath1o@xmath2 has an inverse spinel structure with cubic symmetry where eight fe@xmath4 ions occupy the a sites with local tetragonal symmetry , while the b sites , which locally exhibit octahedral symmetry , are occupied by eight fe@xmath4 and fe@xmath5 ions in a simple ionic chemical formula .
the results of resonant x - ray scattering @xcite show that the fe ions on the b sites are electronically equivalent for time scales lower than 10@xmath6 s. the mi transition occurs at @xmath3 @xmath7 120 k accompanied by a structural change where the conductivity decreases sharply by two orders of magnitude @xcite .
early explanations of the mi transition invoked charge ordering on the fe b sites below @xmath3 . however , nmr relaxation results indicate that the states of fe ions on the b sites are mixed strongly even below @xmath3 @xcite .
moreover , recent resonant x - ray and inelastic neutron scattering studies show a fractional charge ordering on the fe b sites below @xmath3 @xcite , which is predicted by band calculations @xcite .
thus , some studies suggest that only the fe b sites are involved in the change in valence across the verwey transition @xcite while other investigations claim that the fe a sites play also an important role @xcite . therefore , despite continuing experimental and theoretical efforts in the last 70 years , many questions regarding the mi transition still remain open .
magnetic compton scattering using synchrotron radiation is an established technique for probing the spin - uncompensated momentum density distribution in a material with spontaneous magnetization @xcite .
the total electron momentum density , @xmath8 , is @xmath9 where @xmath10 and @xmath11 give the momentum densities of electrons in the majority and minority spin bands , respectively , and @xmath12 is the electron momentum .
owing to the presence of spin - dependent terms in the scattering cross - section , the spin - uncompensated momentum density distribution , @xmath13 , can be measured using circularly polarized x - rays .
it is obtained in the form of a projection of @xmath14 on the x - ray scattering vector , the so - called magnetic compton profile , defined as @xmath15 here the @xmath16-axis is taken to be parallel to the x - ray scattering vector and @xmath17 is the projection of @xmath12 on the @xmath16-axis .
the integral of @xmath18 over @xmath17 yields the spin - magnetic moment of the material . for electronic structure studies in a material with spontaneous magnetization
, a measurement of @xmath14 is very useful because @xmath14 reflects the spin polarization of occupied bands in momentum space .
it is possible to image @xmath14 from a finite number of @xmath18s using a reconstruction technique . in this work
we report magnetic compton scattering studies of a single crystal of magnetite .
the amplitude of the two - dimensional spin - uncompensated momentum density distribution is shown to develop negative excursions above the verwey temperature .
we conclude on this basis that a substantial amount of @xmath19 minority electrons become itinerant on fe b sites above @xmath3 .
a high - quality single crystal of fe@xmath1o@xmath2 was synthesized using the floating zone melting method .
the crystal was annealed at around 1200 @xmath20c in the controlled atmosphere of a mixture of co and co@xmath21 in order to homogenize the composition of the specimen and to release internal strain .
we used a disk - like single crystal sample of approximately 10 mm diameter and 3 mm thickness with its axis parallel to the cubic [ 101 ] axis .
magnetic compton scattering experiments were carried out at 12 k and 300 k on beamline bl08w at spring-8 , japan .
elliptically polarized x - rays emitted from an elliptical multipole wiggler were monochromatized to 174 kev by a bent si(620 ) crystal .
energy spectra of compton scattered x - rays from the sample with a scattering angle of 178.0@xmath20 were measured using a 10-segmented ge solid - state detector with external magnetic field of 20 koe .
the estimated momentum resolution is 0.50 @xmath22 0.01 a.u .
full width at half - maximum .
the directional @xmath18 is extracted from the difference between two spectra taken under the same experimental conditions with alternating directions of magnetization of the sample , aligned by an external magnetic field .
the observed spectra were corrected for the energy - dependent scattering cross section , efficiency of the detector and absorption of the sample .
since severe twinning occurs at @xmath3 in the single crystalline sample due to the structural change , de - twinning of the sample below @xmath3 was accomplished by simultaneous application of external magnetic field cooling and squeezing to extract directional magnetic profiles @xmath18 at 12 k. an external magnetic field of 20 koe was applied along the cubic [ 001 ] axis to establish the @xmath23-axis below @xmath3 .
the use of a special cu sample holder enabled us to squeeze the sample along the cubic [ 11@xmath24 direction as a result of the larger cu thermal contraction .
changes in the extracted @xmath18 at 12 k were observed below @xmath25 @xmath7 2 a.u .
resulting from this mechanical de - twinning .
thus the mechanical de - twinning is effective although we have not estimated the twin ratio in the sample at 12 k with and without mechanical de - twinning .
figure 1 shows the extracted @xmath18 along the three cubic principal axes at 12 and 300 k. these profiles show a substantial directional dependence of the magnetic spin momentum density in good accord with the corresponding theoretical predictions discussed below .
figure 2 shows the extracted @xmath18 at 12 k and 300 k between the cubic [ 001 ] and [ 101 ] axes .
note that the dip in @xmath18 below @xmath17 @xmath7 1 a.u . observed from 9@xmath20 to 27@xmath20 at 300
k disappears at 12 k. this result reveals that @xmath26 changes across @xmath3 even though the saturation magnetic moment changes by less than 0.1 % at @xmath3 @xcite .
we have reconstructed the two - dimensional spin - uncompensated momentum density distribution , @xmath0 . for this purpose
, we adopted the so - called direct fourier - transform method to reconstruct @xmath0 from several directional @xmath18s @xcite . in the reconstruction procedure , we assumed the cubic [ 010 ] axis to possess a four - fold rotational symmetry , even though fe@xmath1o@xmath2 at 12 k has a lower than cubic symmetry @xcite .
this assumption was made as we measured @xmath18s only from the cubic [ 001 ] to the [ 101 ] axis due to the mechanical de - twinning using the special cu sample holder .
the reconstructed @xmath0s are shown in figs .
3(a ) and ( b ) , which correspond to projections of @xmath26 on the cubic ( 010 ) plane . the reconstructed @xmath27 looks like a volcano at 12 and 300 k , where @xmath28 represents the projection of @xmath12 on the cubic ( 010 ) plane . at 300 k , the reconstructed @xmath27 in the crater region ( below @xmath29}|$ ] and @xmath30}|$ ] @xmath7 0.4 a.u . )
is negative , while highly positive values are found in the region of @xmath31 @xmath7 1.7 a.u . with peaks at @xmath29}|$ ] = @xmath30}|$ ] @xmath7 1.2 a.u . , where @xmath32}$ ] and @xmath33}$ ] represent the components of @xmath12 along the cubic [ 001 ] and [ 100 ] axes , respectively .
the reconstructed @xmath27 below @xmath29}|$ ] and @xmath30}|$ ] @xmath7 0.4 a.u . becomes positive at 12 k as seen in fig .
highly positive values in the reconstructed @xmath27 at 12 k appear in the region of @xmath31 @xmath7 1.6 a.u . with peaks at @xmath29}|$ ]
= @xmath30}|$ ] @xmath7 1.1 a.u . ,
similar to the situation at 300 k.
in a 3@xmath34 transition metal compound with spontaneous magnetization , the spin - magnetic moment is the difference in occupancy between the majority and minority spin bands , which mostly have 3@xmath34 character .
electrons in fully occupied majority and minority spin bands do not contribute to the spin - magnetic moment .
however , at any point in momentum space spin components of these band electrons are not completely compensated because hybridization effects depend on the spin state of the band electrons .
these uncompensated spin components contribute to @xmath26 @xcite .
accordingly , @xmath26 is the spin - uncompensated component of all occupied band electrons and the area under @xmath18 is the spin - magnetic moment of the compound .
fe@xmath1o@xmath2 is ferrimagnetic below 858 k where the magnetic moments of the fe a and b sites in the inverse spinel structure are antiparallel , resulting in a saturation magnetic moment of about 4 @xmath35/formula unit , which does not change drastically at @xmath3 @xcite .
the magnetic electrons are mostly 3d electrons from the fe sites .
consequently , the extracted difference in the reconstructed @xmath27 results from a change in the band electrons with mainly 3@xmath34 character .
since the symmetry of fe@xmath1o@xmath2 above @xmath3 is cubic , the negative values below @xmath29}|$ ] and @xmath30}|$ ] @xmath7 0.4 a.u . in the reconstructed @xmath27 at 300 k stem from electrons in the minority spin bands below @xmath36 @xmath7 0.7 a.u . , a region in the first brillouin zone of the inverse spinel structure .
accordingly , these electrons contribute to the electrical conductivity of fe@xmath1o@xmath2 because they have itinerant character .
the reconstructed @xmath27 below @xmath29}|$ ] and @xmath30}|$ ] @xmath7 0.4 a.u .
is positive at 12 k and the positive values in the reconstructed @xmath27 in the region of @xmath31 @xmath7 1.6 a.u at 12 k are smaller than those at 300 k. the population of electrons with itinerant character decreases discontinuously at @xmath3 reflecting the remarkable decrease in the conductivity @xcite .
therefore , the observed difference in the reconstructed @xmath27 provides clear experimental evidence that electrons with @xmath36 @xmath37 0.7 a.u . in the minority spin bands
are localized below @xmath3 and that the electronic structure of fe@xmath1o@xmath2 with the inverse spinel structure is half - metallic in character .
the mi transition is characterized by the localization of the electrons in the minority spin bands which contribute to the spin magnetic moment of fe@xmath1o@xmath2 .
this mi transition with the structural change does not strongly affect the electrons with high momenta , which have a localized character even at 300 k. these results are consistent with a small change in the saturation magnetic moments at @xmath3 @xcite .
the behavior of the reconstructed @xmath27 at 12 k can not be understood using the simple ionic model with the inverse or normal spinel structure .
we have evaluated the number of electrons , @xmath38 , whose wave functions localize from the positive region in the difference of reconstructed @xmath27 between 12 k and 300 k. the value of @xmath38 is found to be 0.590(2 ) per unit cell ( containing two fe b sites ) and should be compared to the charge ordering on the iron b sites of about 0.4 electrons per unit cell @xcite .
the higher value of @xmath38 indicates that the iron - oxygen hybridization plays also a role in the mi transition .
we have carried out kkr band structure and momentum density calculations on fe@xmath1o@xmath2 in the inverse spinel structure within the framework of the local density approximation ( lda ) in order to gain insight into the nature of @xmath0 @xcite . a reasonable level of agreement is seen in fig . 1 between the computed @xmath18 and the measured spectra at 300 k along the three cubic axes , especially above @xmath17 @xmath7 2 a.u . figure 3(c )
shows the theoretical @xmath0 after it has been broadened to account for experimental resolution .
the features of the calculated @xmath27 look similar to those of the reconstructed @xmath27 at 300 k. however , in contrast with the experimental spectrum , the positive peaks at @xmath29}|$ ] = @xmath30}|$ ] @xmath7 1.2 a.u .
in @xmath0 are sharper in the calculations and the value of the calculated @xmath0 at @xmath31 @xmath7 0 is almost zero .
these discrepancies can not be understood within the ionic picture , but could arise from exaggerated hybridization in the lda and/or correlation effects missing in our lda - based calculations . as shown in fig .
4 , our computed spectrum before it is convoluted with the experimental resolution does display a negative value of the spin momentum density in the first brillouin zone due to the presence of minority @xmath19 itinerant electrons appearing above @xmath3 , but this negative excursion is smeared out by the effect of resolution broadening . the observed deeper crater - like feature in the reconstructed @xmath27
could be explained by the partial occupation of @xmath39 orbitals reflecting correlation effects beyond the lda , since the majority spin @xmath39 orbitals give a positive contribution to @xmath27 at @xmath31 @xmath7 0 .
correlation effects are known to produce partial occupation of the natural orbitals and smearing of the electron momentum density @xcite .
we have carried out magnetic compton scattering experiments to investigate the spin - uncompensated electron momentum density @xmath26 of fe@xmath1o@xmath2 at 12 k and 300 k. mechanical de - twinning was used to overcome severe twinning in the single crystal sample below @xmath3 .
the reconstructed @xmath27 below @xmath29}|$ ] and @xmath30}|$ ] @xmath7 0.4 a.u .
changes from being negative at 300 k to positive at 12 k. this result reveals that electrons in the minority spin bands with @xmath36 @xmath37 0.7 a.u . are localized below @xmath3 , while electrons with high momenta , which have a localized character at 300 k , are not affected strongly by the mi transition .
interestingly , the observed behavior of the reconstructed @xmath27 at 12 k can not be understood within the simple ionic model or the standard lda - based band theory picture , and indicates the presence of correlation effects beyond the lda leading to particularly strong smearing of the spin - uncompensated momentum density distribution .
the experiments were performed at spring-8 with the approval of the japan synchrotron radiation research institute ( jasri ) under proposal numbers 2002b0381-nd3-np and 2003a0308-nd3-np .
the work at northeastern university was supported by the us department of energy , office of science , basic energy sciences contract de - fg02 - 07er46352 and benefited from the allocation of time at the nersc supercomputing center .
it was also sponsored by the stichting nationale computer faciliteiten ( ncf ) for the use of supercomputer facilities , with financial support from the nederlandse organisatie voor wetenschappelijk onderzoek ( netherlands organization for scientific research ) . h. k. thanks dr .
y. tanaka for providing the reconstruction programs .
f. walz , j. phys . condens .
matter * 14 * , r285 ( 2002 ) and references therein . j. garcia and g subias , j. phys . condens .
matter * 16 * , r145 ( 2004 ) and references therein .
verwey , nature ( london ) * 144 * , 327 ( 1939 ) .
s. kale , s.m .
bhagat , s.e .
lofland , t. scabarozi , s.b .
ogale , a. orozco , s.r .
shinde , t. venkatesan , b. hannoyer , b. mercey , and w. prellier , phys .
b * 64 * , 205413 ( 2001 ) .
a. yanase and n. hamada , j. phys .
jpn * 68 * , 1607 ( 1999 ) .
j. garcia , g. subias , m.g .
proietti , h. renevier , y. joly , j.l .
hodeau , j. blasco , m.c .
sanchez , and j.f .
berar , phys .
* 85 * , 578 ( 2000 ) .
p. novak , h. stepankova , j. englich , j. kohout , and v.a.m .
brabers , phys .
b*61 * , 1256 ( 2000 ) .
huang , h .- j .
lin , j. okamoto , k.s .
chao , h .-
jeng , g.y .
m , hsu , c .- m .
huang , d.c .
ling , w.b .
wu , , c.s .
yang , and c.t .
chen , phys .
lett . * 96 * , 096401 ( 2006 ) e. nazarenko , j.e .
lorenzo , y. joly , j.l .
hodeau , d. mannix , and c. marin , phys .
* 97 * , 056403 ( 2006 ) .
j. schlapp , c. schssler - langeheine , c.f .
chang , h. ott , a. tanaka , z. hu , m.w .
haverkort , e. schierler e. weschke , g. kaindl , and l.h .
tjeng , phys .
lett . * 100 * , 026406 ( 2008 ) .
mcqueeney , m. yethiraj , w. montfrooij , j.s .
gardner , p. metcalf , and j.m .
honig , phys .
b*73 * , 174409 ( 2006 ) .
mcqueeney , m. yethiraj , s. chang , w. montfrooij , t.g .
perring , m. honig , and p. metcalf , phys .
lett . * 99 * , 246401 ( 2007 ) .
i. leonov , a.n .
yaresko , v.n .
antonov , m.a .
korotin , and v.i .
anisimov , phys .
lett . * 93 * , 146404 ( 2004 ) .
jeng , g.y .
guo , and d.j .
huang , phys .
lett . * 93 * , 156403 ( 2004 ) .
z. lodziana , phys .
lett . * 99 * , 206402 ( 2007 ) .
j. schlappa , c. schssler - langeheine , c. f. chang , h. ott , a. tanaka , z. hu , m.w .
haverkort , e. schierle , e. weschke , g. kaindl , and l. h. tjeng , phys .
lett . * 100 * , 026406 ( 2008 ) .
rozenberg , m. p. pasternak , w. m. xu , y. amiel , m. hanfland , m. amboage , r. d. taylor , and r. jeanloz , phys .
lett . * 96 * , 045705 ( 2006 ) .
rozenberg , y. amiel , w. m. xu , m. p. pasternak , r. jeanloz , m. hanfland , and r. d. taylor , phys .
b*75 * , 020102 ( 2007 ) .
n. sakai , j. appl .
* 29 * , 81 ( 1996 ) .
m. cooper , j. radiat .
phys . chem . *
50 * , 63 ( 1997 ) .
z. kakol , j. solid state chem . * 88 * , 104 ( 1990 ) .
r. suzuki , m. osawa , s. tanigawa , m. matsumoto and n. shiotani , j. phys .
jpn * 58 * , 3251 ( 1989 ) .
y. tanaka , n. sakai , y. kubo , and h. kawata , phys .
lett . * 70 * , 1537 ( 1993 ) .
m. iizumi , t.f .
koetzle , g. shirane , s. chikazumi , m. matsui , and s. todo , acta cryst .
b*38 * , 2121 ( 1982 ) .
wright , j.p .
attfield , and p.g .
radaelli , phys .
lett . * 87 * , 266401 ( 2001 ) t. nagao , y. kubo , a. koizumi , h. kobayashi , m. itou , and n. sakai , j. phys . condens .
matter * 20 * , 055201 ( 2008 ) .
y. li , p.a .
montano , b. barbiellini , p.e .
mijnarends , s. kaprzyk , and a. bansil , j. phys .
solids * 68 * , 1556 ( 2007 ) .
b. barbiellini , a. bansil , j. phys .
solids * 62 * , 2181 ( 2001 ) .
@xmath18 of fe@xmath1o@xmath2 along the three cubic principal axes as a function of @xmath17 .
open and filled circles with error bars represent the measured @xmath18s at 300 and 12 k , respectively .
solid lines give the calculated @xmath18s in the cubic spinel structure which have been broadened to reflect experimental resolution @xcite .
] ( color online ) ( a ) and ( b ) reconstructed @xmath27s of fe@xmath1o@xmath2 from measured @xmath18s at 300 k and at 12 k , respectively . ( c ) calculated @xmath27 of fe@xmath1o@xmath2 in the inverse spinel structure with cubic symmetry which has been broadened to reflect experimental resolution . ] reconstructed and calculated @xmath27s of fe@xmath1o@xmath2 along the cubic [ 101 ] axis at 300 k and in the cubic spinel structure with and without convolution of experimental resolution , respectively .
filled circles represent the reconstructed @xmath27 at 300 k. solid and broken lines give the calculated @xmath27s in the cubic spinel structure without and with convolution of experimental resolution , respectively . ] | two - dimensional spin - uncompensated momentum density distributions , @xmath0s , were reconstructed in magnetite at 12k and 300k from several measured directional magnetic compton profiles .
mechanical de - twinning was used to overcome severe twinning in the single crystal sample below the verwey transition .
the reconstructed @xmath0 in the first brillouin zone changes from being negative at 300 k to positive at 12 k. this result provides the first clear evidence that electrons with low momenta in the minority spin bands in magnetite are localized below the verwey transition temperature . |
for a given weight function @xmath2 supported on a set @xmath3 , a cubature formula of degree @xmath4 is a finite sum , @xmath5 , that provides an approximation to the integral and preserves polynomials of degree up to @xmath4 ; that is , @xmath6 where @xmath7 denotes the space of polynomials of total degree at most @xmath8 in @xmath9 variables .
the points @xmath10 are called _ nodes _ and the numbers @xmath11 are called _ weights _ of the cubature .
our primary interests are gaussian type cubature , which has minimal or nearer minimal number of nodes . for @xmath12
, it is well known that gaussian quadrature of degree @xmath4 needs merely @xmath13 nodes and these nodes are precisely the zeros of the orthogonal polynomial of degree @xmath8 with respect to @xmath2 .
the situation for @xmath14 , however , is much more complicated and not well understood in general . as in the case of @xmath12
, it is known that a cubature of degree @xmath4 needs at least @xmath15 number of nodes , but few formulas are known to attain this lower bound ( see , for example , @xcite ) .
in fact , for the centrally symmetric weight function ( symmetric with respect to the origin ) , it is known that the number of nodes , @xmath16 , of a cubature of degree @xmath4 in two dimension satisfies the lower bound @xmath17 known as mller s lower bound @xcite .
it is also known that the nodes of a cubature that attains the lower bound , if it exists , are necessarily the common zeros of @xmath18 orthogonal polynomials of degree @xmath8 with respect to @xmath2 .
similar statements on the nodes hold for cubature formulas that have number of nodes slightly above mller s lower bound , which we shall call cubature of _ gaussian type_. these definitions also hold in @xmath9-dimension , where the lower bound for the number of nodes for the centrally symmetric weight function is given in @xcite .
there are , however , only a few examples of such formulas that are explicitly constructed and fewer still can be useful for practical computation .
the best known example is @xmath19^d$ ] with the weight function @xmath20 and only when @xmath21 . in this case , several families of gaussian type cubature are explicitly known , they were constructed ( @xcite ) by studying the common zeros of corresponding orthogonal polynomials , which are product chebyshev polynomials of the first kind and the second kind , respectively .
furthermore , interpolation polynomials bases on the nodes of these cubature formulas turn out to possess several desirable features ( @xcite , and also @xcite ) . on the other hand , studying common zeros of orthogonal polynomials of several variables is in general notoriously difficult . in the case of
, the product chebyshev polynomials have the simplest structure among all orthogonal polynomials , which permits us to study their common zeros and construct cubature formulas in the case @xmath22 , but not yet for the case @xmath23 or higher .
the purpose of the present paper is to provide a completely different method for constructing cubature formulas with respect to @xmath24 and @xmath25 .
it uses the discrete fourier analysis associated with lattice tiling , developed recently in @xcite .
this method has been used in @xcite to establish cubature for _ trigonometric functions _ on the regular hexagon and triangle in @xmath26 , a topic that has been studied in @xcite , and on the rhombic dodecahedron and tetrahedron of @xmath27 in @xcite .
the cubature on the hexagon can be transformed , by symmetry , to a cubature on the equilateral triangle that generates the hexagon by reflection , which can in turn be further transformed , by a nontrivial change of variables , to gaussian cubature formula for algebraic polynomials on the domain bounded by steiner s hypercycloid .
the theory developed in @xcite uses two lattices , one determines the domain of integral and the points that defined the discrete inner product , the other determines the space of exponentials or trigonometric functions that are integrated exactly by the cubature . in @xcite
the two lattices are taken to be the same . in this paper
we shall choose one as @xmath28 itself , so that the integral domain is fixed as the cube , while we choose the other one differently . in @xmath21 ,
we choose the second lattice so that its spectral set is a rhombus , which allows us to establish cubature formulas for trigonometric functions that are equivalent to gaussian type cubature formulas for @xmath24 and @xmath25 . in the case of @xmath29
, we choose the rhombic dodecahedron as a tiling set and obtain a cubature of degree @xmath4 that uses @xmath30 nodes , worse than the expected lower bound of @xmath31 but far better than the product gaussian cubature of @xmath32 nodes .
this cubature with @xmath30 nodes has appeared recently and tested numerically in @xcite .
we will further study the lagrange interpolation based on its nodes , for which the first task is to identify the subspace that the interpolation polynomials belongs .
we will not only identify the interpolation space , but also give the compact formulas for the fundamental interpolation polynomials .
one immediate question arising from this study is if there exist cubature formulas of degree @xmath4 with @xmath33 nodes on the cube .
although examples of cubature formulas of degree @xmath4 with @xmath34 nodes are known to exist for special non - centrally symmetric regions ( @xcite ) , we are not aware of any examples for symmetric domains that use @xmath35 nodes . from our approach of tiling and discrete fourier analysis
, it appears that the rhombic dodecahedron gives the smallest number of nodes among all other fundamental domains that tile @xmath27 by translation .
giving the fact that this approach yields the cubature formulas with optimal order for the number of nodes , it is tempting to make the conjecture that a cubature formula of degree @xmath4 on @xmath0 ^ 3 $ ] needs at least @xmath30 nodes .
the paper is organized as follows . in the following section we recall the result on discrete fourier analysis and lattice tiling in @xcite .
cubature and interpolation for @xmath21 are developed in section 3 and those for @xmath23 are discussed in section 4 , both the latter two sections are divided into several subsections .
we recall basic results in @xcite on the discrete fourier analysis associated with a lattice .
a lattice of @xmath36 is a discrete subgroup that can be written as @xmath37 , where @xmath38 is a @xmath39 invertible matrix , called the generator of the lattice .
a bounded set @xmath40 is said to tile @xmath36 with the lattice @xmath41 if @xmath42 where @xmath43 denotes the characteristic function of the set @xmath44 .
the simplest lattice is @xmath28 itself , for which the set that tiles @xmath36 is @xmath45 we reserve the notation @xmath46 as above throughout the rest of this paper .
the set @xmath46 is chosen as half open so that its translations by @xmath28 tile @xmath36 without overlapping .
it is well known that the exponential functions @xmath47 form an orthonormal basis for @xmath48 .
these functions are periodic with respect to @xmath28 ; that is , they satisfy @xmath49 let @xmath50 be a @xmath39 matrix such that all entries of @xmath50 are integers
. denote @xmath51 it is known that @xmath52 , where @xmath53 denotes the cardinality of the set @xmath44 . we need the following theorem ( * ? ? ?
* theorem 2.5 ) .
[ thm:2.1 ] let @xmath50 be a @xmath54 matrix with integer entries . define the discrete inner product @xmath55 for @xmath56 , the space of continuous functions on @xmath46 .
then @xmath57 for all @xmath58 in the finite dimensional subspace @xmath59 the dimension of @xmath60 is @xmath61 .
this result is a special case of a general result in @xcite , in which @xmath46 is replaced by @xmath62 for an invertible matrix @xmath38 , and the set @xmath63 is replaced by @xmath64 with @xmath65 and @xmath16 is assumed to have integer entries . since we are interested only at the cube @xmath66^d$ ] in this paper , we have chosen @xmath38 as the identity matrix
. we can also use the discrete fourier analysis to study interpolation based on the points in @xmath63 .
we say two points @xmath67 congruent with respect to the lattice @xmath68 , if @xmath69 , and we write @xmath70 .
we then have the following result : [ thm : interpolation ] for a generic function @xmath71 defined in @xmath72 , the unique interpolation function @xmath73 in @xmath74 that satisfies @xmath75 is given by @xmath76 where @xmath77 the proof of this result is based on the second one of the following two relations that are of independent interests : @xmath78 and @xmath79 for proofs and further results we refer to @xcite . throughout this paper
we will write , for @xmath80 , @xmath81 and @xmath82 .
in this section we consider the case @xmath21 . in the first subsection ,
the general results in the previous section is specialized to a special case and cubature formulas are derived for a class of trigonometric functions .
these results are converted to results for algebraic polynomials in the second subsection .
results on polynomial interpolation are derived in the third subsection .
we choose the matrix @xmath50 as @xmath83 \qquad \hbox{and } \qquad b^{-1 } = \frac{1}{2n } \left [ \begin{matrix } 1 & - 1\\ 1 & 1 \end{matrix } \right].\ ] ] since @xmath50 is a rotation , by 45 degree , of a constant multiple of the diagonal matrix , it is easy to see that the domain @xmath84 is defined by @xmath85 which is depicted in figure 1 below .
, scaledwidth=40.0% ] from the expression of @xmath86 , it follows readily that @xmath87 , where @xmath88 the cardinality of @xmath89 is @xmath90 .
we further denote the space @xmath74 by @xmath91 , which is given by @xmath92 [ inner1 ] define the set @xmath93 then for all @xmath94 , @xmath95 ^ 2 } f(x)\overline{g(x ) } dx.\ ] ] changing variables from @xmath96 to @xmath97 , or @xmath98 and @xmath99 , then , as @xmath100 and @xmath101 need to be integers and @xmath102 , @xmath103 , we see that @xmath104 hence , as @xmath105 , we conclude that @xmath106 and this theorem follows as a special case of theorem [ thm:2.1 ]
. the set @xmath89 lacks symmetry as the inequalities in its definition are half open and half closed .
we denote its symmetric counterpart by @xmath107 , which is defined by @xmath108 we also denote the counterpart of @xmath91 by @xmath109 , which is defined by @xmath110 along the same line , we also define the counterpart of @xmath111 as @xmath112 it is easy to see that @xmath113 , whereas @xmath114 .
we further partition the set @xmath115 into three parts , @xmath116 where @xmath117 is the set of interior points of @xmath115 , @xmath118 consists of those points in @xmath115 that are on the edges of @xmath119 ^ 2 $ ] but not on the 4 vertices or corners , while @xmath120 consists of those points of @xmath115 at the vertices of @xmath119 ^ 2 $ ] .
[ inner2 ] define the inner product @xmath121 then for all @xmath94 , @xmath122 ^ 2 } f(x)\overline{g(x ) } dx = \langle f , g \rangle_n = \langle f , g \rangle_n^*.\ ] ] evidently we only need to show that @xmath123 . since @xmath124 for @xmath125 , the partial sums over interior points of the two sums agree .
the set @xmath126 of boundary points can be divided into two parts , @xmath127 , where @xmath128 consists of points in @xmath111 that are on the edges of @xmath129 , but not equal to @xmath130 , and @xmath131 is the complementary of @xmath128 in @xmath126 . evidently , if @xmath132 , then either @xmath133 or @xmath134 belongs to @xmath131 .
hence , if @xmath71 is a periodic function , @xmath135 for @xmath136 , then @xmath137 furthermore , for @xmath138 , @xmath115 contains all four vertices @xmath139 . since a periodic function takes the same value on all four points , @xmath140 .
consequently , we have proved that @xmath141 if @xmath142 are periodic functions . as a consequence of the above two theorems
, we deduce the following two cubature formulas : [ thm : cuba1 ] for @xmath143 , the cubature formulas @xmath144 ^ 2 } f(x ) dx = \frac{1}{2n^2 } \sum_{k \in x_n^ * } c_k^{(n ) } f(\tfrac{k}{2n } ) \quad\hbox{and } \quad \int_{[-\frac12,\frac12]^2 } f(x ) dx = \frac{1}{2n^2 } \sum_{k \in x_n } f(\tfrac{k}{2n})\ ] ] are exact for @xmath145 .
it suffices to proof that both cubature formulas in are exact for every @xmath146 with @xmath147 . for this purpose
, we first claim that for any @xmath148 , there exist @xmath149 and @xmath150 such that @xmath151 .
indeed , the translations of @xmath84 by @xmath152 tile @xmath26 , thus we have @xmath153 for certain @xmath154 and @xmath150 . since all entries of the matrix @xmath50 are integers , we further deduce that @xmath155 .
next assume @xmath147 .
clearly the integral of @xmath146 over @xmath46 is @xmath156 . on the other hand ,
let us suppose @xmath157 with @xmath149 and @xmath150 .
then it is easy to see that @xmath158 for each @xmath159 .
consequently , we obtain from theorem [ inner2 ] that @xmath160 since @xmath161 implies @xmath162 which gives @xmath163 , we further obtain that @xmath164 .
this completes the proof of .
we note that the second cubature in is a so - called chebyshev cubature ; that is , all its weights are equal .
the set @xmath107 is symmetric with respect to the mappings @xmath165 and @xmath166 .
it follows that both the spaces @xmath167 are subspaces of @xmath109 .
recall that chebyshev polynomials of the first kind , @xmath168 , and the second kind , @xmath169 , are defined , respectively , by @xmath170 they are orthogonal with respect to @xmath171 and @xmath172 over @xmath0 $ ] , respectively .
both are algebraic polynomials of degree @xmath8 in @xmath173 .
recall the definition of @xmath24 and @xmath25 in . under the changing of variables
@xmath174 ^ 2,\ ] ] the subspace @xmath175 becomes the space @xmath176 of polynomials of degree @xmath8 in the variables @xmath177 , @xmath178 and the orthogonality of @xmath179 over @xmath46 implies that @xmath180 are orthogonal polynomials of two variables , @xmath181 ^ 2 } t_j^k(t ) t_{j'}^{k'}(t ) w_0(t ) dt = \begin{cases } 1 , & k = k'=j = j'=0,\\ \tfrac12 , & ( k , j ) = ( k',j ' ) \hbox { and } ( k - j)j=0,\\ \tfrac14 , & k = k'>j = j'>0,\\ 0 , & ( k , j ) \ne ( k',j ' ) . \end{cases}\ ] ] we note also that the subspace @xmath182 becomes the space @xmath183 in the variables @xmath184 , and the orthogonality of @xmath179 also implies that @xmath185 are orthogonal polynomials of two variables , @xmath181 ^ 2 } u_j^k(t ) u_{j'}^{k'}(t ) w_1(t ) dt = \frac{1}{4 } \delta_{j , j'}\delta_{k , k'}.\ ] ] the symmetry allows us to translate the results in the previous subsection to algebraic polynomials .
since @xmath186 are even in both variables , we only need to consider their values over @xmath187 . hence , we define @xmath188 and , under the change of variables , @xmath189 furthermore , we denote by @xmath190 the subset of interior points of @xmath191 , by @xmath192 the set of points in @xmath191 that are on the boundary of @xmath0 ^ 2 $ ] but not on the four corners , and by @xmath193 the set of points in @xmath191 that are on the corners of @xmath0 ^ 2 $ ] .
the sets @xmath194 , @xmath195 and @xmath196 are defined accordingly .
a simple counting shows that @xmath197 [ thm : cubat ] the cubature formula @xmath198 ^ 2 } f(t ) w_0(t ) dt = \frac{1}{2 n^2}\sum_{k \in \xi_n } \l_k^{(n ) } f(z_{k_1},z_{k_2 } ) , \quad \l_k^{(n ) } : = \begin{cases } 4 , & k \in \xi_n^\circ , \\ 2 , & k \in \xi_n^e , \\ 1 , & k \in \xi_n^v , \end{cases}\ ] ] is exact for @xmath199 .
we note that @xmath115 is symmetric in the sense that @xmath200 implies that @xmath201 and @xmath202 .
let @xmath203 .
then @xmath204 is even in each of its variables and @xmath205 .
notice that @xmath206 implies @xmath207 . applying the first cubature formula to @xmath208
, we see that follows from the following identity , @xmath209 to prove this identity , let @xmath210 denote the set of distinct elements in @xmath211 ; then @xmath212 takes the same value on all points in @xmath210 . if @xmath200 , @xmath213 and @xmath214 , then @xmath210 contains 4 points ; @xmath215 if @xmath125 , @xmath216 if @xmath217 , and @xmath218 if @xmath219 .
if @xmath220 and @xmath214 or @xmath221 and @xmath222 , then @xmath223 contains 2 points ; @xmath224 if @xmath125 and @xmath225 if @xmath217 . finally , if @xmath226 then @xmath210 contains 1 point and @xmath227 has coefficient 1 .
putting these together proves the identity . by , the number of nodes of the cubature formula is just one more than the lower bound .
we can also write into a form that is more explicit . indeed ,
if @xmath228 , then can be written as @xmath229 ^ 2 } f(t ) w_0(t ) dt \\ & \qquad\qquad = \frac{2}{n^2 } \sideset { } { '' } \sum_{i=0}^{m } \sideset{}{''}\sum_{j=0}^m f(z_{2i},z_{2j } ) + \frac{2}{n^2 } \sum_{i=0}^{m-1 } \sum_{j=0}^{m-1 } f(z_{2i+1},z_{2j+1 } ) , \notag\end{aligned}\ ] ] where @xmath230 means that the first and the last terms in the summation are halved .
if @xmath231 , then can be written as @xmath232 ^ 2 } f(t ) w_0(t ) dt \\ & \qquad\qquad = \frac{2}{n^2 } \sideset { } { ' } \sum_{i=0}^{m } \sideset{}{'}\sum_{j=0}^m f(z_{2i},z_{2j } ) + \frac{2}{n^2 } \sideset { } { ' } \sum_{i=0}^{m } \sideset{}{'}\sum_{j=0}^{m } f(z_{n - 2i},z_{n-2j } ) , \notag\end{aligned}\ ] ] where @xmath233 means that the first term in the sum is divided by 2 .
the formula appeared in @xcite , where it was constructed by considering the common zeros of orthogonal polynomials of two variables . from the cubature formula
, we can also derive cubature formulas for the chebyshev weight @xmath25 of the second kind .
[ thm : cubau ] the cubature formula @xmath234 ^ 2 } f(t ) w_1(t ) dt = \frac{2}{n^2 } \sum_{k \in \xi_n^\circ } \sin^2 \tfrac{k_1 \pi}{n}\sin^2 \tfrac{k_2 \pi}{n } f(z_{k_1},z_{k_2})\ ] ] is exact for @xmath235 .
we apply the first cubature formula in on the functions @xmath236 for @xmath237 , where @xmath238 and @xmath239 as in .
clearly these functions are even in both @xmath240 and @xmath241 and they are functions in @xmath242 .
furthermore , they are zero when @xmath243 or @xmath244 , or when @xmath245 are on the boundary of @xmath115 .
hence , the change of variables shows that the first cubature in becomes for @xmath246 .
a simple counting shows that @xmath247 .
the number of nodes of the cubature formula is also one more than the lower bound . in this case , this formula appeared already in @xcite . as shown in @xcite
, there is a close relation between interpolation and discrete fourier transform .
we start with a simple result on interpolation by trigonometric functions in @xmath91 .
[ prop:3.4 ] for @xmath248 define @xmath249 then @xmath250 for all @xmath251 .
for @xmath252 define @xmath253 . from the relation ,
@xmath252 is equivalent to @xmath251 with @xmath254 . as a result
, we can write @xmath255 as @xmath256 and the interpolation means @xmath257 for @xmath252 . for @xmath258 , @xmath259 by . for our main result , we need a lemma on the symmetric set @xmath115 and @xmath107 . recall that @xmath260 is defined for @xmath200 .
since the relation clearly extends to @xmath261 we define @xmath262 whenever @xmath263 and @xmath96 are so related . comparing to , we then define @xmath264 we also introduce the following notation : for @xmath217 , we denote by @xmath265 the point on the opposite edge of @xmath115 ; that is , @xmath266 and @xmath267 or @xmath268 .
furthermore , we denote by @xmath269 the index corresponding to @xmath265 under .
[ prop:3.5 ] the function @xmath270 satisfies @xmath271 as in the proof of the previous theorem , we can write @xmath272 as @xmath273 by using . let @xmath274 . for all @xmath275 , @xmath276 since @xmath277 for any @xmath278 , we derive by using a similar argument as in theorem [ inner2 ] that @xmath279 by , @xmath280 if @xmath281 . if @xmath282 then @xmath283 , so that if @xmath284 then @xmath285 . the same holds for the case of @xmath286 and @xmath287 .
if both @xmath288 , then @xmath289 .
using the relation , we have shown that @xmath290 when @xmath291 and 0 otherwise , from which the stated result follows .
it turns out that the function @xmath292 satisfies a compact formula .
let us define an operator @xmath293 by @xmath294.\ ] ] for @xmath295 , it follows immediately that @xmath296 [ lem : d_n ] for @xmath297 , @xmath298 - \frac{1}{4 } ( \cos 2\pi n x_1 + \cos 2\pi n x_2 ) , \ ] ] where @xmath299 using the values of @xmath300 and the definition of @xmath301 , it is easy to see that @xmath302 - \sum_{\nu \in \lambda^v } { \mathrm{e}}_v(x).\ ] ] since @xmath303 contains four terms , @xmath304 and @xmath305 , the sum over @xmath303 becomes the second term in . on the other hand , using the symmetry of @xmath107 and , @xmath306 where @xmath233 means that the terms in the sum are halved whenever either @xmath307 or @xmath308 , from which the second equal sign in follows from ( * ? ? ?
* ( 4.2.1 ) and ( 4.2.7 ) ) .
our main result in this section is interpolation over points in @xmath309 with @xmath310 defined in .
for @xmath297 define @xmath311(x)\ ] ] with @xmath312 given in .
then @xmath313 is even in both variables and it satisfies @xmath314 as shown in the proof of proposition [ prop:3.5 ] , @xmath315 satisfies @xmath316 when @xmath317 and 0 otherwise .
hence , if @xmath318 then @xmath319^{-1 } \d_{k , j}$ ] .
if @xmath320 then the number of terms in the sum of @xmath321 depends on whether @xmath322 is zero ; if @xmath323 then @xmath324 = \frac12 \d_{k , j}= [ \l_k^{(n)}]^{-1 } \d_{k , j}$ ] , whereas if @xmath325 then @xmath326^{-1 } \d_{k , j}$ ] . for @xmath327 or @xmath328 in @xmath196 , we have @xmath329 = \d_{k , j}$ ] ; for @xmath330 we have @xmath331 = \d_{k , j}$ ] ; finally for @xmath332 , it is evident that @xmath333 .
putting these together , we have verified that @xmath334 for all @xmath335 , which verifies the interpolation of @xmath336 . as in the case of cubature
, we can translate the above theorem to interpolation by algebraic polynomials by applying the change of variables .
recall @xmath191 defined in .
for @xmath297 , let @xmath337 then @xmath338 and it satisfies @xmath339 for all @xmath340 .
furthermore , under the change of variables , the fundamental polynomial @xmath341 satisfies @xmath342(x ) - \frac{1}{4 } \left[(-1)^{k_1 } t_{k_1}(t_1)+(-1)^{k_2 } t_{k_2}(t_2 ) \right].\ ] ] that @xmath336 interpolates at @xmath340 is an immediate consequence of the change of variables , which also shows that @xmath338 .
moreover , @xmath343 , which verifies the formula of @xmath341 .
the polynomial @xmath336 belongs , in fact , to a subspace @xmath344 of dimension @xmath345 , and it is the unique interpolation polynomial in @xmath346 . in the case of @xmath8 is odd , this interpolation polynomial was defined and studied in @xcite , where a slightly different scheme with one point less was studied in the case of even @xmath8 .
recently the interpolation polynomials in @xcite have been tested and studied numerically in @xcite ; the results show that these polynomials can be evaluated efficiently and provide valuable tools for numerical computation .
for @xmath21 , the choice of our spectral set @xmath84 and lattice in the previous section ensures that we end up with a space close to the polynomial subspace @xmath176 ; indeed , monomials in @xmath176 are indexed by @xmath347 , a quarter of @xmath107 . for @xmath348
, the same consideration indicates that we should choose the spectral set as the octahedron @xmath349 .
the octahedron , however , does not tile @xmath27 by lattice translation ( see , for example , @xcite ) . as an alternative
, we choose the spectral set as _ rhombic dodecahedron _ , which tiles @xmath27 by lattice translation with _
face centered cubic ( fcc ) lattice_. in @xcite , a discrete fourier analysis on the rhombic dodecahedron is developed and used to study cubature and interpolation on the rhombic dodecahedron , which also leads to results on tetrahedron .
in contrast , our results will be established on the cube @xmath350 ^ 3 $ ] , but our set @xmath84 is chosen to be a rhombic dodecahedron .
the strict inequality in the definition of @xmath84 reflects our requirement that the tiling of the spectral set has no overlapping . from the expression of @xmath86
, it follows that @xmath353 is given by @xmath354 it is known that @xmath355 .
furthermore , @xmath356 is given by @xmath357 we denote the space @xmath74 by @xmath91 , which is given by @xmath358 then @xmath359 .
changing variables from @xmath96 to @xmath97 , or @xmath362 , then , as @xmath363 are integers and @xmath364 , @xmath365 , @xmath366 , we see that @xmath367 from which we conclude that @xmath368 .
consequently , this theorem is a special case of theorem [ thm:2.1 ] .
just like the case of @xmath21 , we denote the symmetric counterpart of @xmath111 by @xmath115 which is defined by @xmath369 a simple counting shows that @xmath370 .
the set @xmath115 is further partitioned into four parts , @xmath371 where @xmath117 is the set of interior points , @xmath372 contains the points in @xmath115 that are on the faces of @xmath119 ^ 3 $ ] but not on the edges or vertices , @xmath118 contains the points in @xmath115 that are on the edges of @xmath119 ^ 3 $ ] but not on the corners or vertices , while @xmath120 denotes the points of @xmath115 at the vertices of @xmath119 ^ 3 $ ] .
the proof follows along the same line as the proof of theorem [ inner2 ] .
we only need to show @xmath374 if @xmath375 is periodic .
the interior points of @xmath111 and @xmath115 are the same , so that @xmath376 for @xmath125 .
let @xmath377 , @xmath378 , and @xmath379 .
each point @xmath263 in @xmath372 has exactly one opposite point @xmath380 in @xmath372 under translation by @xmath381 and only one of them is in @xmath111 , so that @xmath382 $ ] if @xmath71 is periodic , which is why we define @xmath383 for @xmath384 .
evidently , only three edges of @xmath115 are in @xmath385 .
each point in @xmath126 corresponds to exactly four points in @xmath126 under integer translations @xmath381 and only one among the four is in @xmath111 , so we define @xmath386 for @xmath217 .
finally , all eight corner points can be derived from translations @xmath387 points , used repeatedly , and exactly one , @xmath388 , is in @xmath389 , so that we define @xmath390 for @xmath219 .
[ thm : cuba1 - 3d ] for @xmath143 , the cubature formulas @xmath395 ^ 3 } f(x ) dx = \frac{1}{2n^3 } \sum_{k \in x_n^ * } c_k^{(n ) } f(\tfrac{k}{2n } ) \quad\hbox{and } \quad \int_{[-\frac12,\frac12]^3 } f(x ) dx = \frac{1}{2n^3 } \sum_{k \in x_n } f(\tfrac{k}{2n})\ ] ] are exact for @xmath145 .
assume now @xmath399 .
clearly the integral of @xmath146 over @xmath46 is @xmath156 . on the other hand ,
let us suppose @xmath151 with @xmath149 and @xmath398 . then it is easy to see that @xmath158 for each @xmath159 .
consequently , we get from theorem [ inner2 - 3d ] that @xmath400 since @xmath161 implies @xmath163 , we further obtain that @xmath401 .
this states that the cubature formulas are exact for each @xmath146 with @xmath399 , which completes the proof
. we can also translate the cubature in theorem [ thm : cuba1 - 3d ] into one for algebraic polynomials .
for this we use the change of variables @xmath402 ^ 3.\ ] ] under , the functions @xmath403 become algebraic polynomials @xmath404 , which are even in each of its variables .
the subspace of @xmath109 that consists of functions that are even in each of its variables corresponds to the polynomial subspace @xmath405 notice that @xmath115 is symmetric in the sense that if @xmath406 then @xmath407 for all @xmath408 , where @xmath409 . in order to evaluate functions that are even in each of its variables on @xmath115
we only need to consider @xmath410 .
hence , we define , @xmath411 and , under the change of variables , define @xmath412 moreover , we denote by @xmath413 , @xmath414 , @xmath192 and @xmath193 the subsets of @xmath191 that contains interior points , points on the faces but not on the edges , points on the edges but not on the vertices , and points on the vertices , of @xmath0 ^ 3 $ ] , respectively , and we define @xmath194 , @xmath415 , @xmath195 and @xmath196 accordingly .
a simple counting shows that @xmath416 [ thm : cubat3d ] write @xmath417 . the cubature formula @xmath418 ^ 3 } f(t )
w_0(t ) dt = \frac{1}{2 n^2}\sum_{k \in \xi_n } \l_k^{(n ) } f(z_k ) , \quad \l_k^{(n ) } : = \begin{cases } 8 , & k \in \xi_n^\circ , \\ 4 , & k \in \xi_n^f , \\ 2 , & k \in \xi_n^e , \\ 1 , & k \in \xi_n^v , \end{cases}\ ] ] is exact for @xmath419 . in particular , it is exact for @xmath420
. let @xmath421 .
then @xmath204 is even in each of its variables and @xmath422 . applying the first cubature formula in to @xmath208
, we see that follows from the following identity , @xmath423 this identity is proved in the same way that the corresponding identity in theorem [ thm : cubat ] is proved .
let @xmath210 denote the set of distinct elements in @xmath424 ; then @xmath212 takes the same value on all points in @xmath210 .
if @xmath200 , @xmath425 for @xmath426 , then @xmath210 contains 8 points ; if exactly one @xmath427 is zero then @xmath210 contains 4 points ; if exactly two @xmath427 are zero then @xmath210 contains one point ; and , finally , if @xmath428 then @xmath210 contains one point . in the case of @xmath429 for @xmath430 , @xmath431 if @xmath125 , @xmath215 if @xmath384 , and @xmath432 if @xmath217 .
the other cases are treated similarly .
thus , holds for @xmath419 .
we note that @xmath419 contains @xmath420 as a subspace , but it does not contain @xmath434 since @xmath435 is in @xmath434 but not in @xmath420 . hence , the cubature is of degree @xmath4 .
a trivial cubature formula of degree @xmath4 for @xmath24 can be derived by taking the product of gaussian quadrature of degree @xmath4 in one variable , which has exactly @xmath32 nodes .
in contrast , according to , the number of nodes of our cubature is in the order of @xmath30 , about a quarter of the product formula . as far as we know
, this is the best that is available at the present time . on the other hand ,
the lower bound for the number of nodes states that a cubature formula of degree @xmath4 needs at least @xmath436 nodes .
it is , however , an open question if there exist formulas with number of nodes attaining this theoretic lower bound .
recall the cubature is derived by choosing the spectral set as a rhombic dodecahedron .
one natural question is how to choose a spectral set that tiles @xmath27 by translation so that the resulted cubature formula is of degree @xmath4 and has the smallest number of nodes possible . among the regular lattice tiling
, the rhombic dodecahedron appears to lead to the smallest number of nodes .
[ thm : cubau3d ] the cubature formula @xmath438 ^ 3 } f(t ) w_1(t ) dt = \frac{4}{n^3 } \sum_{k \in \xi_n^\circ } \sin^2 \tfrac{k_1 \pi}{n}\sin^2 \tfrac{k_2 \pi}{n } \sin^2 \tfrac{k_3 \pi}{n}f(z_k)\ ] ] is exact for @xmath439 ; in particular , it is exact for @xmath440 . in order to obtain the compact formula for the interpolation function , we follow @xcite and use homogeneous coordinates and embed the rhombic dodecahedron into the plane @xmath441 of @xmath442 . throughout the rest of this paper , we adopt the convention of using bold letters , such as @xmath443 , to denote the points in the space @xmath444 in other words , the bold letters such as @xmath443 and @xmath445 will always mean homogeneous coordinates .
the transformation between @xmath446 and @xmath447 is defined by @xmath448 in this homogenous coordinates , the spectral set @xmath84 becomes @xmath449 we now use homogeneous coordinates to describe @xmath392 defined in .
let @xmath450 and @xmath451 in order to keep the elements as integers , we make the change of variables @xmath452 for @xmath453 .
it then follows that @xmath392 in homogeneous coordinates becomes @xmath454 we could have changed variables without the factor 2 , setting @xmath455 etc .
we choose the current change of variables so that we can use some of the computations in @xcite .
in fact , the set @xmath456 is used in @xcite .
the main result of this subsection is a compact formula for the partial sum @xmath457 where @xmath458 and @xmath443 are related by and the middle equality follows from the fact that @xmath459 under this change of variables .
in fact , by and , we have @xmath460 = \frac{1}{4 } \jb \cdot \tb\end{aligned}\ ] ] where in the last step we have used the fact that @xmath447 . the compact formula of @xmath461
is an essential part of the compact formula for the interpolation function . by definition
, @xmath466 is a subset of @xmath467 that contains elements with all indices being even integers .
for technical reasons , it turns out to be easier to work with @xmath468 .
in fact , the sum over @xmath467 has already been worked out in @xcite , which is @xmath469 thus , we need to find only the sum over odd indices , that is , the sum @xmath470 just as in @xcite , the index set @xmath471 can be partitioned into four congruent parts , each within a parallelepiped , defined by @xmath472 for @xmath473 .
furthermore , for each index set @xmath474 , @xmath475 , define @xmath476 then we have @xmath477 using the inclusion - exclusion relation of subsets , we have @xmath478 fix @xmath479 , using the fact that @xmath480 , we have @xmath481 since @xmath482 implies , in particular , @xmath483 , we obtain @xmath484 where @xmath485 .
the last equation needs a few words of explanation : if @xmath486 , then using the fact that @xmath487 , @xmath488 for @xmath489 and @xmath490 , we see that @xmath491 , which is odd by the definition of @xmath492 ; on the other hand , assume that @xmath493 is odd , then we define @xmath494 for all @xmath495 and define @xmath496 , so that all components of @xmath263 are odd and @xmath489 . the condition that @xmath497 is
an odd integer means that the last term is not a simple product of sums .
setting @xmath498 we see that , up to a permutation , only products @xmath499 and @xmath500 are possible for triple products ( @xmath501 ) , only @xmath502 is possible for double products ( @xmath503 ) , only @xmath504 is possible ( @xmath505 ) , and there is a constant term .
thus , using the fact that @xmath506 , we conclude that @xmath507 \\ & + d_n^o(t_2 ) \left [ d_n^e(t_1)d_n^e(t_3)d_n^e(t_4 ) - ( d_n^e(t_1)-1)(d_n^e(t_3)-1)(d_n^e(t_4)-1 ) \right ] \\ & + d_n^o(t_3 ) \left [ d_n^e(t_1)d_n^e(t_2)d_n^e(t_4 ) - ( d_n^e(t_1)-1)(d_n^e(t_2)-1)(d_n^e(t_4)-1 ) \right ] \\ & + d_n^o(t_4 ) \left [ d_n^e(t_1)d_n^e(t_2)d_n^e(t_3 ) - ( d_n^e(t_1)-1)(d_n^e(t_2)-1)(d_n^e(t_3)-1 ) \right],\end{aligned}\ ] ] where the first sum is over all distinct triple integers in @xmath508 .
assume that @xmath8 is an even integer .
a quick computation shows that @xmath509 furthermore , we see that @xmath510 = - \prod_{j=2}^4 \frac{\sin \pi n t_i } { \sin 2 \pi t_i } \frac{\sin \pi ( n-2 ) t_1}{\sin 2 \pi t_1}.\end{aligned}\ ] ] adding the two terms together and then summing over the permutation of the sum , we end up the formula for @xmath511 when @xmath8 is even .
the case of @xmath8 odd can be handled similarly .
let us write down explicitly the function @xmath512 defined in in @xmath458-variables . using the elementary trigonometric identity and
, we see that @xmath513 so that we end up with the compact formula @xmath514 where @xmath515 @xmath516 and @xmath517 in which @xmath518 is given in terms of @xmath519 in . as a result of this explicit expression
, we see that @xmath512 is an even function in each @xmath520 . in order to develop the interpolation on the set @xmath115 , we will need to understand the structure of the points on the boundary of @xmath521 . as @xmath84 is a rhombic dodecahedron , we need to understand the boundary of this 12-face polyhedron , which has been studied in detail in @xcite . in this subsection , we state the necessary definitions and notations on the boundary of @xmath84 ,
so that the exposition is self - contained .
we refer to further details and proofs to @xcite .
again we use homogeneous coordinates . for @xmath522 and @xmath523 ,
the ( closed ) faces of @xmath84 are @xmath524 there are a total @xmath525 distinct @xmath526 , each represents one face of the rhombic dodecahedron . for nonempty subsets @xmath527 of @xmath508 ,
define @xmath528 it is shown in @xcite that @xmath529 if and only if @xmath530 , and @xmath531 if @xmath532 and @xmath533 .
these sets describe the intersections of faces , which can then be used to describe the edges , which are intersections of faces , and vertices , which are intersections of edges .
let @xmath534 we now define , for each @xmath535 , the boundary element @xmath536 of the dodecahedron , @xmath537 it is called a face if @xmath538 , an edge if @xmath539 , and a vertex if @xmath540 if @xmath541 and @xmath542 .
in particular , it follows that @xmath543 and , for example , @xmath544 for distinct integers @xmath545 .
let @xmath546 denote the permutation group of four elements and let @xmath547 denote the element in @xmath548 that interchanges @xmath549 and @xmath96 ; then @xmath550 .
for a nonempty set @xmath551 , define @xmath552 , where we take @xmath553 and take @xmath554 as the identity element .
it follows that @xmath555 forms a subgroup of @xmath556 of order @xmath557 . for @xmath558 , we then define @xmath559 } [ \b_{i , j } ] : = \bigcup_{\sigma\in \mathcal{g}_{i\cup j } } \b_{i , j}\sigma.\end{aligned}\ ] ] it turns out that @xmath560 $ ] consists of exactly those boundary elements that can be obtained from @xmath536 by congruent modulus @xmath50 , and @xmath561\cap [ \b_{i_1,j_1 } ] = \emptyset$ ] if @xmath562 for @xmath563 and @xmath564 .
more importantly , we define , for @xmath565 , @xmath566 } \begin{split } \b^{i , j}:= \bigcup_{(i , j ) \in \ck_0^{i , j } } [ \b_{i , j } ] \quad \hbox{with}\quad \ck^{i , j}_0 : = & \ , \left\{(i , j ) \in
\ck_0:\ |i| = i,\ |j| = j \right\}. \end{split}\end{aligned}\ ] ] then the boundary of @xmath567 can be decomposed as @xmath568 the main complication is the case of @xmath569 , for which we have , for example , @xmath570 } [ \b_{\{1\},\{2,3\ } } ] = \b_{\{1\},\{2,3\ } } \cup \b_{\{2\},\{1,3\ } } \cup \b_{\{3\},\{1,2\}}.\end{aligned}\ ] ] the other cases can be written down similarly . furthermore , we have @xmath571 with @xmath572 if @xmath538 then @xmath573 is a face and @xmath574 \cup [ \b_{\{1\},\{3\}}]\cup [ \b_{\{1\},\{4\}}]\cup [ \b_{\{2\},\{3\}}]\cup[\b_{\{2\},\{4\}}]\cup [ \b_{\{3\},\{4\}}]\end{aligned}\ ] ] if @xmath575 then @xmath536 is an edge and we have @xmath576 \cup [ \b_{\{1\},\{2,4\}}]\cup [ \b_{\{1\},\{3,4\}}]\cup [ \b_{\{2\},\{3,4\}}],\\ & \b^{2,1 } = [ \b_{\{1,2\},\{3\}}]\cup [ \b_{\{1,2\},\{4\}}]\cup [ \b_{\{1,3\},\{4\ } } ] \cup [ \b_{\{2,3\},\{4\ } } ] . \end{split}\end{aligned}\ ] ] if @xmath577 , then @xmath578 , \quad \b^{2,2 } = \left[\{(\tfrac{1}{2 } , \tfrac{1}{2 } , -\tfrac{1}{2 } , -\tfrac{1}{2}\})\right]\\ & \b^{3,1 } = \left[\{(\tfrac{3}{4 } , -\tfrac{1}{4 } , -\tfrac{1}{4 } , -\tfrac{1}{4})\}\right ] . \end{split}\end{aligned}\ ] ] recall that @xmath466 is @xmath579 in homogeneous coordinates .
we now consider the decomposition of the boundary of @xmath466 according to the boundary elements of the rhombic dodecahedron .
first we denote by @xmath580 the points inside @xmath466 , @xmath581 we further define , for @xmath582 , @xmath583 the set @xmath584 describes those points @xmath585 in @xmath466 such that @xmath586 are in @xmath587 of @xmath588 .
it is easy to see that @xmath589 if @xmath590 and @xmath591 next we consider interpolation on the symmetric set of points @xmath115 . for this
we need to modify the kernel function @xmath600 .
recall that , under the change of variables , @xmath601 becomes @xmath466 in homogeneous coordinates .
we define @xmath602 where @xmath458 and @xmath443 are related by , @xmath603 is defined by @xmath604 under the change of indices , and @xmath605 if @xmath606 , @xmath607 if @xmath608 ; more explicitly @xmath609 for each @xmath263 on the boundary of @xmath115 , that is , @xmath610 on the boundary of @xmath350 ^ 3 $ ] , let @xmath611 which contains the points on the boundary of @xmath115 that are congruent to @xmath263 under integer translations .
[ thm:3dinterp ] for @xmath248 define @xmath612 then for each @xmath613 , @xmath614 in homogeneous coordinates , the function @xmath615 is a real function and it satisfies @xmath616,\end{aligned}\ ] ] from which the formula for @xmath617 follows from and . by , we need to verify the interpolation at the points @xmath618 for @xmath619 . by definition ,
we can write @xmath620 it is easy to see that @xmath621 if @xmath622 is related to @xmath585 by . hence , as in the proof of theorem 3.15 in @xcite , we conclude that @xmath623 now , for @xmath624 , there exist @xmath625 and @xmath626 such that @xmath627 .
consequently , it follows from that @xmath628 by , we have verified that @xmath629 which proves the interpolation part of the theorem . in order to prove the compact formula , we start with the following formula that can be established exactly as in the proof of theorem 3.15 in @xcite : @xmath630.\notag\end{aligned}\ ] ] let us define @xmath631 for @xmath632 and also define @xmath633 : = \{\kb \in \gg_n : \tfrac{\kb}{4n } \in [ \b_{i , j}]\}$ ] .
it follows from , and that @xmath634 \qquad\hbox{and}\qquad \left[\gg_n^{i , j } \right ] = \bigcup_{\sigma\in \mathcal{g}_{i\cup j } } \gg_n^{i , j}\sigma.\ ] ] in order to compute the sums in , we need to use the detail description of the boundary elements of @xmath84 in the previous subsection .
the computation is parallel to the proof of theorem 3.15 in @xcite , in which the similar computation with @xmath466 replaced by @xmath635 is carried out .
thus , we shall be brief . using @xmath441 and the explicit description of @xmath636
, we get @xmath637 } \phi_\kb(\tb ) = \sum_{\kb \in \gg_n^{\{1\},\{2,3\ } } } { \mathrm{e}}^{\frac{\pi i}{2}\kb\cdot \tb } + \sum_{\kb \in \gg_n^{\{2\},\{1,3\ } } } { \mathrm{e}}^{\frac{\pi i}{2}\kb\cdot \tb } + \sum_{\kb \in \gg_n^{\{3\},\{1,2\ } } } { \mathrm{e}}^{\frac{\pi i}{2}\kb\cdot \tb } \\ &
\qquad = \sum_{j=1 , j \mathrm{even}}^{n-1 } e^{-2
\pi i j t_4 } \left ( e^{2 n \pi i ( t_1+t_4 ) } + e^{2 n \pi i ( t_2+t_4 ) } + e^{2 n \pi i ( t_3+t_4 ) } \right ) \\ & \qquad = \frac{\sin 2 \pi \lfloor \frac{n-1}{2 } \rfloor t_4}{\sin 2 \pi t_4 } { \mathrm{e}}^{- 2\pi i \lfloor \frac{n+1}{2 } \rfloor t_4 } \left(e^{2 \pi i n ( t_1+t_4 ) } + e^{2 \pi i n ( t_2+t_4)}+ e^{2 \pi i n ( t_3+t_4 ) } \right),\end{aligned}\ ] ] similarly , we also have @xmath638 } \phi_\kb(\tb ) = \sum_{\kb \in \gg_n^{\{1\},\{2,3\ } } } { \mathrm{e}}^{\frac{\pi i}{2}\kb\cdot \tb } + \sum_{\kb \in \gg_n^{\{2\},\{1,3\ } } } { \mathrm{e}}^{\frac{\pi i}{2}\kb\cdot \tb } + \sum_{\kb \in \gg_n^{\{3\},\{1,2\ } } } { \mathrm{e}}^{\frac{\pi i}{2}\kb\cdot \tb } \\ & \qquad = \frac{\sin 2 \pi \lfloor \frac{n-1}{2 } \rfloor t_4}{\sin 2 \pi t_4 } { \mathrm{e}}^ { 2\pi i \lfloor \frac{n+1}{2 } \rfloor t_4 } \left(e^{-2 \pi i n ( t_1+t_4 ) } + e^{-2 \pi i
n ( t_2+t_4)}+ e^{-2 \pi i n ( t_3+t_4 ) } \right).\end{aligned}\ ] ] from these and their permutations , we can compute the sum over @xmath639 and @xmath640 . putting them together ,
we obtain @xmath641 using , we see that , @xmath642 and , if @xmath8 is even then @xmath643 and @xmath644 , whereas if @xmath8 is odd , then @xmath645 . as a result , it follows that @xmath646 where we have used the fact that @xmath441 , and @xmath647 if @xmath8 is even , whereas it is equal to @xmath648 if @xmath8 is odd . following the standard procedure , we see that @xmath653 ^ 3 } \sum_{k \in x_n^ * } \left|\phi_n^*(x -\tfrac{k}{4n})\right|.\ ] ] using the formula of @xmath292 in , it is easy to see that it suffices to prove that @xmath654 ^ 3 } \sum_{k \in x_n^ * } \left|d_n^*(x -\tfrac{k}{2n})\right| \le c ( \log n)^3 , \qquad n \ge 0.\ ] ] furthermore , using the explicit formula of @xmath511 and ( 3.19 ) in @xcite , we see that our main task is to estimate the sums in the form of @xmath655 and three other similar estimates @xmath656 , @xmath657 and @xmath658 , respectively , as well as similar sums in which the denominator becomes product of @xmath659 and @xmath8 in the numerator is replace by @xmath660 or @xmath661 . here
@xmath662 is the image of @xmath0 ^ 3 $ ] under the mapping ; that is , @xmath663 changing the summation indices and enlarging the set @xmath115 , we see that @xmath664 } } \left ( \frac{1}{2n } \sum_{k = 0}^{2 n } \left | \frac{\sin n \pi ( t -\tfrac{k}{2 n } ) } { \sin \pi ( t -\tfrac{k}{2 n } ) } \right| \right)^3 \le c ( \log n)^3,\end{aligned}\ ] ] where the last step follows from the standard estimate of one variable ( cf .
ii , p. 19 ) ) .
the main outcome of theorem [ thm : interp3d ] in the previous section is that we can derive a genuine interpolation by trigonometric polynomials based on the set of points in @xmath665 defined at .
the development below is similar to the case of @xmath21 .
we define @xmath666 as shown in , @xmath315 satisfies @xmath316 when @xmath667 and 0 otherwise .
hence , if @xmath318 then @xmath668^{-1 } \d_{k , j}$ ] .
if @xmath669 , then we need to consider several cases , depending on how many components of @xmath585 are zero , which determines how many distinct terms are in the sum @xmath321 and how many distinct @xmath263 can be obtained from @xmath96 by congruent in @xmath670 .
for example , if @xmath671 and none of the components of @xmath96 are zero , then there are 2 elements in @xmath672 , @xmath96 and the one in the opposite face , and the sum @xmath673 contains 8 terms , so that @xmath674^{-1 } \delta_{k , j}$ ] . the other cases can be verified similarly , just as in the case of @xmath22 .
we omit the details .
the theorem states that the interpolation space for the point set @xmath191 is exactly @xmath346 , which consists of monomials that have indices in the positive quadrant of the rhombic dodecahedron , as depicted in figure 3 below .
the set @xmath191 consists of roughly @xmath680 points .
the interpolation polynomial @xmath677 is about a total degree of @xmath681 .
the compact formula of the fundamental interpolation polynomial provides a convenient way of evaluating the interpolation polynomial .
furthermore , the lebesgue constant of this interpolation process remains at the order of @xmath682 , as the consequence of theorem [ lebesgueh ] and the change of variables . | several cubature formulas on the cubic domains are derived using the discrete fourier analysis associated with lattice tiling , as developed in @xcite .
the main results consist of a new derivation of the gaussian type cubature for the product chebyshev weight functions and associated interpolation polynomials on @xmath0 ^ 2 $ ] , as well as new results on @xmath0 ^ 3 $ ] . in particular , compact formulas for the fundamental interpolation polynomials are derived , based on @xmath1 nodes of a cubature formula on @xmath0 ^ 3 $ ] . |
let @xmath0 be a finite state irreducible and aperiodic markov chain ( mc ) with transition probability matrix @xmath1 whose elements are the transition probabilities @xmath2 this mc admits a stationary distribution @xmath3 , which is the unique solution of @xmath4 the entropy rate @xmath5 is a measure of the average entropy growth of a sequence generated by the process @xmath0 and is defined as @xmath6 for the particular case of an irreducible and aperiodic mc , the limit above is equal to @xcite @xmath7 where @xmath8 denotes the @xmath9 row of @xmath1 and where @xmath10 is the _ local entropy _ of state @xmath11 .
note that , throughout this paper , we use @xmath12 as a shorthand for the markov chain whose transition probability matrix is @xmath1 .
we follow the setting of @xcite closely .
we define a _ random trajectory _ @xmath13 of a mc as a path with initial state @xmath14 , final state @xmath15 , and no intermediate state @xmath15 , i.e. , the trajectory is terminated as soon as it reaches state @xmath15 . using the markov property
, we express the probability of a particular trajectory @xmath16 given that @xmath17 as @xmath18 let @xmath19 be the set of all trajectories that start at state @xmath14 and end as soon as they reach state @xmath15 . as
the mc defined by the matrix @xmath1 is finite and irreducible , we have @xmath20 so the discrete random variable @xmath13 has as support the set @xmath19 , with the probability mass function @xmath21 .
subsequently , we use @xmath21 as a shorthand for @xmath22 . we can now express the entropy of the random trajectory @xmath23 as @xmath24 we define the matrix of trajectory entropies @xmath25 where @xmath26 .
ekroot and cover @xcite provide a general closed - form expression for the matrix @xmath25 of an irreducible , aperiodic and finite state mc .
the entropy @xmath27 of a trajectory from @xmath14 to @xmath15 given that it goes through @xmath28 is defined by @xmath29 where @xmath30 is the set of all trajectories in @xmath19 with an intermediate state @xmath28 @xmath31 the major challenge is to compute efficiently the entropy @xmath27 .
even the costly approach of computing all the terms of the sum ( @xmath32 ) is not always possible because the set @xmath33 has an infinite number of members in the case where , after removing state @xmath15 , the transition graph of the mc is not a dag .
it is important to emphasize that the entropy @xmath27 is not the entropy of the random variable @xmath13 given another random variablea quantity which is easy to computebut the entropy of @xmath13 conditional on the realization of a dependent random variable . in figure
[ fig : mcexample ] , we show an example of a finite - state irreducible and aperiodic mc .
note that the presence of cycles implies that the set of trajectories between some pair of states might have infinite cardinality ( @xmath34 for example ) .
therefore , in addition to being complex , the naive approach of enumerating all trajectories is not always possible . using the results of @xcite
, we obtain the matrix of trajectory entropies @xmath35 the zero elements of the matrix @xmath25 correspond to deterministic trajectories such as @xmath36 , which is equal to the path @xmath37 with probability @xmath38 since no other path allows a walk to go from @xmath39 to @xmath40 .
the entropy of the random trajectory @xmath41 is 1.56 bits .
now imagine that we have an additional piece of information stating that the trajectory @xmath41 goes through state @xmath42 .
intuitively , we would be tempted to argue that the entropy @xmath43 of the trajectory @xmath41 conditional on going through state @xmath42 is equal to @xmath44 , but this additivity property does not hold . indeed , the conditional entropy @xmath43 is zero because the trajectory @xmath41 , conditional on the intermediate state 4 , can only be equal to the path @xmath45 , whereas @xmath46 bits , hence @xmath47 bits . in the next section
, we study the entropy of markov trajectories conditional on _ multiple _ intermediate states and derive a general expression for this entropy .
let @xmath48 denote the probability that the random trajectory @xmath13 goes through the state @xmath28 at least once : @xmath49 this is also equal to the probability that a walk reaches the state @xmath28 before the state @xmath15 , given that it started at @xmath14 . in order to compute @xmath48 , the technique from @xcite
is to make the states @xmath28 and @xmath15 absorbing ( a state @xmath11 is absorbing if and only if @xmath50 ) and compute the probability to be absorbed by state @xmath28 given that the trajectory has started at state @xmath14 .
our first step towards computing @xmath27 is to express it as a function of quantities that are much simpler to compute .
the idea is to relate the entropy of a trajectory conditional on a given state to its entropy conditional on _ not _ going through that state .
therefore , we define the entropy @xmath51 of a trajectory from @xmath14 to @xmath15 given that it does _ not _ go through @xmath28 to be @xmath52 using the chain rule for entropy , we can derive the following equation which relates @xmath27 to @xmath53 and @xmath48 : @xmath54 for all @xmath28 , where @xmath55 is the entropy of a bernoulli random variable with success probability @xmath48 .
first , we define the indicator variable @xmath56 by @xmath57 using the chain rule for entropy , we express the joint entropy @xmath58 in two different ways , @xmath59 because @xmath56 is a deterministic function of @xmath13 .
so the entropy of the random trajectory @xmath13 can be expressed as @xmath60 since @xmath61 , we obtain @xmath62 as we know from @xcite how to compute @xmath63 and @xmath48 , if we are able to compute @xmath51 , we can use ( [ eq : chainruleentropy ] ) to find @xmath27 . however , generalizing ( [ eq : chainruleentropy ] ) to trajectories conditional on passing through _ multiple _ intermediate states turns out to be difficult , hence we propose an approach that circumvents this problem .
as we will see , the difficulty of our approach also boils down to computing the entropy of a trajectory conditional on _ not _ going through a given state .
first , we define @xmath64 , the set of all trajectories in @xmath19 that exhibit the sequence of intermediate states @xmath65 , i.e. @xmath66 for an arbitrary sequence of states @xmath65 , satisfying @xmath67 , we prove the following lemma .
[ thm : cond_entropy_as_sum ] @xmath68 where @xmath69 . first , given @xmath70 , the random trajectory @xmath13 can be expressed as a sequence of random sub - trajectories @xmath71 .
therefore , the conditional entropy @xmath72 , which we denote by @xmath73 , can be written as a joint sub - trajectory entropy @xmath74 applying the chain rule for entropy , we obtain successively @xmath75 the markovian nature of the process generating the trajectory @xmath13 implies that each of the sub - trajectories @xmath76 is independent of the preceding ones , given its starting point @xmath77 .
since the sequence @xmath78 defines the starting point of each sub - trajectory , we can therefore write that @xmath79 using , the expression for the conditional entropy becomes @xmath80 note that for each trajectory @xmath76 , the only restriction imposed by the event @xmath81 is that the final state @xmath15 can not be an intermediate state of any of the first @xmath82 trajectories @xmath83 . as a result , @xmath84 where @xmath85 now , if we are able to compute @xmath86 , we can use ( [ eq : condientropy ] ) to derive @xmath72 .
the following lemma shows how the conditional entropy @xmath86 can be obtained by a simple modification of the mc .
+ we consider a mc whose transition probability matrix is @xmath1 , and @xmath14 , @xmath28 and @xmath15 three distinct states such that @xmath87 .
let @xmath88 be the transition matrix of the same mc but where both states @xmath28 and @xmath15 are made absorbing , and whose entries are thus @xmath89 next , we define a second matrix @xmath90 , obtained by a transformation of the matrix @xmath88 @xmath91 [ thm : trajcondentropy ] @xmath92 the matrix @xmath90 is stochastic and @xmath93 if @xmath94 is a random trajectory defined on the mc whose transition probability matrix is @xmath90 then @xmath95 @xmath92 the matrix @xmath88 is the transition probability matrix of a mc where the states @xmath28 and @xmath15 are absorbing .
we can therefore introduce the vectors of absorption probability @xmath96 and @xmath97 where @xmath98 and @xmath99 are , respectively , the probability of being absorbed by @xmath28 and @xmath15 , given that the trajectory starts at @xmath11 .
these vectors are eigenvectors of @xmath88 associated with the unit eigenvalue @xcite @xmath100 moreover as @xmath101 has only two absorbing states @xmath28 and @xmath15 , for all @xmath11 , @xmath102 .
recall that for all @xmath11 , @xmath103 hence @xmath104 can be written as @xmath105 note that all transitions leading to state @xmath28 in @xmath106 will have zero probability in @xmath107 .
in fact , consider a state @xmath11 such that @xmath108 and @xmath109 . in the new matrix @xmath90 , the probability of transition from @xmath11 to @xmath28
will be @xmath110 , which is zero because @xmath111 .
proving that @xmath90 is stochastic is now straightforward : first , the entries of @xmath90 are positive ; second , they are properly normalized and sum up to one . indeed ,
if we consider a state @xmath11 such that @xmath112 , we have that @xmath113 whereas if @xmath114 , we have that @xmath115 because of ( @xmath116 ) .
let @xmath117 and @xmath118 be the probability measures defined , respectively , for @xmath119 and @xmath107 on the same sample space @xmath19 .
any trajectory from the set @xmath19 has the form @xmath16 . if @xmath120 , @xmath121 since we have constructed @xmath107 such that all transitions leading to state @xmath28 have zero probability .
if @xmath122 , we have @xmath123 but @xmath124 as the probability to be absorbed by state @xmath15 , given that we have started at this same state , is @xmath38 . moreover , we know from that @xmath125 , for all @xmath126
as we have supposed that the trajectory @xmath127 does not admit either @xmath28 or @xmath15 as intermediate states , @xmath128 . rewriting ( [ eq : trajcondprob ] ) yields @xmath129 combining and , we have therefore proven , for all @xmath130 , that @xmath131 consequently , if the random variable @xmath94 describes the trajectory between @xmath14 and @xmath15 in @xmath107 , implies that @xmath132 for the particular case where @xmath133 , we still can use lemma [ thm : trajcondentropy ] to express the conditional entropy @xmath134 : we modify the mc by removing the incoming transitions of @xmath14 and creating a new state @xmath135 that will inherit them .
the conditional entropy @xmath134 in the original mc is equal to @xmath136 in the modified one and , since @xmath137 , we can use lemma [ thm : trajcondentropy ] to express it .
building on lemma [ thm : cond_entropy_as_sum ] and lemma [ thm : trajcondentropy ] , we can now state the main result of this paper : a general expression for the entropy of markov trajectories conditional on multiple intermediate states .
[ thm : main_result ] let @xmath1 be the transition probability matrix of a finite markov chain and @xmath138 a sequence of states such that @xmath67 .
then , we have the following equality @xmath139 where @xmath69 , and @xmath140 is a random trajectory defined on the markov chain whose transition probability matrix @xmath141 is defined as follows @xmath142 the matrix @xmath141 is obtained from @xmath1 using , which is equivalent to applying successively and where the starting , intermediate and ending states are , respectively , @xmath77 , @xmath15 and @xmath143 .
therefore , using lemma [ thm : trajcondentropy ] , we have @xmath144 for all @xmath145 . consequently ,
we can write that @xmath146 where @xmath69 . using lemma [ thm : cond_entropy_as_sum ] , we finally obtain @xmath147 now that we have derived a general expression for the entropy of markov trajectories conditional on multiple states , we introduce , in the next section , a method that allows us to compute this expression .
the closed - form expression for the entropy of markov trajectories proposed by ekroot and cover @xcite is valid only if the markov chain studied is irreducible .
however , the markov chain @xmath107 obtained from @xmath119 after the transformations and is not necessarily irreducible : all transitions leading to state @xmath28 have zero probability , which implies that possibly many states do not admit any path leading to @xmath15 .
therefore , we need an expression for the entropy of markov trajectories that is valid under milder conditions . in order to identify these conditions , we study the properties of @xmath107 .
let @xmath148 be the set of all states in @xmath107 and let @xmath149 and @xmath150 be two subsets that partition @xmath148 in the following manner @xmath151 the set @xmath149 is closed as no one - step transition is possible from any state in @xmath149 to any state in @xmath150 .
in fact , if @xmath152 and @xmath153 , @xmath104 yields that @xmath154 clearly , all trajectories leading to state @xmath15 are composed of states belonging to @xmath149 .
now , we propose a closed - form expression for the entropy of markov trajectories that is valid under the weaker condition that the destination state @xmath15 can be reached from any other state of the mc .
moreover , we prove that the trajectory entropy can be expressed as a weighted sum of local entropies .
we also provide an intuitive interpretation of the weights .
[ thm : trajentropy2 ] let @xmath1 be the transition probability matrix of a finite state mc such that there exists a path with positive probability from any state to a given state @xmath15 .
let @xmath155 be a sub - matrix of @xmath1 obtained by removing the @xmath156 row and column of @xmath1 .
@xmath157 & & \vdots \\ \hline p_{d1 } & \cdots & p_{dd } \\ \end{array}. \right)\ ] ] for any state @xmath158 , the trajectory entropy @xmath63 can be expressed as @xmath159 where @xmath160 is the local entropy of state @xmath161 .
first , observe that the matrix @xmath155 is a sub - matrix of @xmath1 corresponding to all states except state @xmath15 and that we use @xmath155 to derive the entropy of all trajectories ending at @xmath15 . applying the chain rule for entropy
, we express the entropy of a trajectory as the entropy of the first step plus the conditional entropy of the rest of the trajectory given this first step @xmath162 we expand this equality further by recursively expanding the entropy @xmath163 as follows @xmath164 with @xmath165 .
observe that the matrix @xmath155 describes the markov chain as long as it does not reach state @xmath15 .
moreover , the matrix @xmath155 has a finite number of states and there is a path with positive probability from each state to state @xmath15 . as a consequence
, the markov process will enter state @xmath15 with probability @xmath38 , i.e. , @xmath166 ( zero matrix ) .
in addition , since @xmath167 we can easily verify that @xmath168 replacing ( [ eq : locentropy2 ] ) in ( [ eq : locentropy1 ] ) , we have @xmath169 we have shown that the entropy of a family of trajectories can be expressed as a weighted sum of the states local entropies .
the weights are given by the matrix @xmath170 .
in the markovian literature , the matrix @xmath170 is referred to as the fundamental matrix @xcite . in fact , the @xmath171 element of the fundamental matrix ( defined with respect to the destination state @xmath15 ) can be seen as the expected number of visits to the state @xmath161 before hitting the state @xmath15 , given that we started at state @xmath14 . as a result , the entropy of the random trajectory @xmath23 is the sum over the chain states of the expected number of visits to each state multiplied by its local entropy .
this is a remarkable observation since it links a global quantity , which is the trajectory entropy , to the local entropy at each state .
recall that in the example shown in figure [ fig : mcexample ] , we found that the entropy of the trajectory @xmath41 is equal to @xmath172 bits .
we can retrieve this result by computing the fundamental matrix with respect to state @xmath40 .
the @xmath173 element of this matrix is equal to the expected number of visits to state @xmath174 before hitting state @xmath40 , given that we started at state @xmath11 .
multiplying the first row of the fundamental matrix @xmath175 by the column vector of local entropies @xmath176 yields @xmath177 bits .
the following algorithm defines the set of steps to compute the entropy of markov trajectories conditional on a set of intermediate states : [ alg : cond_entr ] matrix of transition probability @xmath1 , source state @xmath14 , destination state @xmath15 , sequence of intermediate states @xmath178 @xmath73 @xmath179 compute @xmath180 from @xmath1 using compute @xmath181 from @xmath180 using lemma [ thm : trajentropy2 ] @xmath182 compute @xmath183 from @xmath1 using lemma [ thm : trajentropy2 ] @xmath184 @xmath73 the worst - case running time for the algorithm is @xmath185 where @xmath186 is the number of states of @xmath119 , and @xmath82 the length of the sequence of intermediate states @xmath187 .
this complexity is dominated by the cost of computing the inverse of the matrix @xmath188 , which is needed to compute the entropy @xmath63 in .
however , since we need only the @xmath189 row of the matrix @xmath188 to compute the trajectory entropy @xmath63 , we can solve a system ofpotentially sparselinear equations .
moreover , many iterative methods @xcite take advantage of the structure of the matrix representing the system of linear equations in order to solve them efficiently . coming back to the example shown in figure [ fig : mcexample ]
, we use the algorithm above to compute the conditional entropy @xmath190 bit .
we leave no ambiguity about the trajectory @xmath41 when we condition on both states @xmath191 and @xmath39 and find that @xmath192 bits .
[ [ conditioning - on - a - set - of - states ] ] conditioning on a set of states + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in this paper , we focused on computing the entropy of markov trajectories conditional on a _ sequence _ of states . a natural extension is the computation of this entropy conditional on a _ non ordered _ set of states . finding a general expression for this conditional entropy appears very hard and there is no simple relation linking it to the entropy conditional on a sequence .
we provide an example , shown in figure [ fig : mcexample_2 ] , that illustrates an interesting and counter - intuitive result about conditioning on a set of states .
intuitively , we would expect that the entropy of a random trajectory conditional on a sequence of states is always less than the entropy of the same trajectory conditional on the set formed by these states .
however , this is not true . we take the mc shown in figure [ fig : mcexample_2 ] as an example and we compute , using theorem [ thm : main_result ] , the entropy of the random trajectory @xmath41 conditional on going through the sequence of intermediate states @xmath193 @xmath194 where @xmath195 is the entropy of a bernoulli random variable with success probability @xmath196 . to compute the entropy of the random trajectory @xmath41 conditional on going through the set of states @xmath197 ,
we apply the chain rule for entropy and express the entropy of a trajectory as the entropy of the first two steps plus the conditional entropy of the rest of the trajectory given these first two steps @xmath198 since @xmath199 , we have that @xmath200 using and , we can write @xmath201 this difference can therefore be lower bounded by @xmath202 as a consequence , if @xmath203 , the entropy of the random trajectory @xmath41 conditional on going through the sequence @xmath193 is strictly greater than the entropy of the same trajectory conditional on going through the set of states @xmath197 .
the reason is that conditioning on the sequence @xmath193 implies that the random trajectory @xmath41 is composed of a random sub - trajectory @xmath204 whose entropy can be made arbitrary large by increasing the parameter @xmath205 .
more generally , this example illustrates the absence of a simple relation between the entropy of random trajectories conditional on a sequence of states and the entropy of the same trajectory conditional on the set formed by these same states .
in this paper , we address the problem of computing the entropy of conditional markov trajectories .
we propose a method based on a transformation of the original markov chain into a markov chain that yields the desired conditional entropy .
we also derive an expression that allows us to compute the entropy of markov trajectories , under conditions weaker than those assumed in @xcite .
furthermore , this expression links the entropy of markov trajectoriesa global quantityto the local entropy of states .
these results have applications in various fields including mobility privacy of the users of online services .
in fact , using our framework , we are able to quantify the predictability of a user s mobility and its evolution with locations updates : we represent a location as a state of a markov chain .
a sequence of visited locations is therefore a markovian trajectory , and location - updates amount to conditioning this trajectory on a set of intermediate states . in this setting
, we can quantify the evolution of the user s mobility predictability as she / he discloses some of the locations she / he visited by computing the entropy of conditional markov trajectories .
consequently , users are empowered with an objective technique to protect their privacy : they are able to anticipate the evolution of their mobility predictability as they reveal a subset of the locations they visited .
the authors would like to thank olivier lvque and emre telatar for their feedback about this paper . l. yen , m. saerens , a. mantrach , and m. shimbo , `` a family of dissimilarity measures between nodes generalizing both the shortest - path and the commute - time distances , '' in _ proceedings of the 14th sigkdd international conference on knowledge discovery and data mining _ , 2008 . mohamed kafsi(s12 ) is a ph.d .
student in the school of computer and communication sciences at epfl .
his research interests lie at the intersection of graph theory , information theory and data mining .
he received the m.s . and
degrees in communication systems from epfl . during his bachelor
s studies , he spent one year at the electrical and computer engineering department of carnegie mellon university ( cmu ) .
he is a member of the winning team of the 2013 nokia mobile data challenge .
matthias grossglauser(m95 ) is an associate professor in the school of computer and communication sciences at epfl .
he received his diplme dingnieur en systmes de communication degree from epfl in 1994 , the m.sc .
degree from the georgia institute of technology in 1994 , and the ph.d . from the university pierre
et marie curie ( paris 6 ) in 1998 .
his research interests are in social and information networks , privacy , mobile and wireless networking , and network traffic measurement and modeling .
he received the 1998 cor baayen award from the european research consortium for informatics and mathematics ( ercim ) , the ieee infocom 2001 best paper award , and the 2006 conext / sigcomm rising star award .
he served on the editorial board of ieee / acm transactions on networking , and on numerous technical program committees . from 2007 - 2010 , he was with the nokia research center ( nrc ) in helsinki , finland , holding the positions of laboratory director , then of head of a tech - transfer program focused on data mining , analytics , and machine learning .
in addition , he served on nokia s ceo technology council , a technology advisory group reporting to the ceo . from 2003 - 2007 , he was an assistant professor at epfl . from 1998 to 2002 , he was a senior , then principal member of research staff in the networking and distributed systems laboratory at at&t research in new jersey . from 1995 to 1998 , he was a ph.d .
student at inria sophia antipolis , france .
patrick thiran(s89 m96
sm12 ) received the electrical engineering degree from the universit catholique de louvain , louvain - la - neuve , belgium , in 1989 , the m.s .
degree in electrical engineering from the university of california at berkeley , usa , in 1990 , and the ph.d .
degree from epfl , in 1996 .
he is a full professor at epfl .
he became an adjunct professor in 1998 , an assistant professor in 2002 , an associate professor in 2006 and a full professor in 2011 .
from 2000 to 2001 , he was with sprint advanced technology labs , burlingame , ca .
his research interests include communication networks , performance analysis , dynamical systems , and stochastic models .
he is currently active in the analysis and design of wireless and plc networks , in network monitoring , and in data - driven network science .
thiran served as an associate editor for the ieee transactions on circuits and systems in 1997 - 99 , and as an associate editor for the ieee / acm transactions on networking in 2006 - 10 .
he was the recipient of the 1996 epfl ph.d .
award and of the 2008 crdit suisse teaching award . | to quantify the randomness of markov trajectories with fixed initial and final states , ekroot and cover proposed a closed - form expression for the entropy of trajectories of an irreducible finite state markov chain .
numerous applications , including the study of random walks on graphs , require the computation of the entropy of markov trajectories conditional on a set of intermediate states .
however , the expression of ekroot and cover does not allow for computing this quantity . in this paper
, we propose a method to compute the entropy of conditional markov trajectories through a transformation of the original markov chain into a markov chain that exhibits the desired conditional distribution of trajectories .
moreover , we express the entropy of markov trajectoriesa global quantityas a linear combination of _ local entropies _ associated with the markov chain states . the entropy of conditional markov trajectories entropy , markov chains , markov trajectories .
the randomness of markov trajectories has applications in graph theory @xcite and in statistical physics @xcite , as well as in the study of random walks on graphs @xcite .
the need to quantify the randomness of markov trajectories first arose when lloyd and pagels @xcite defined a measure of complexity for the macroscopic states of physical systems .
they examine some intuitive properties that a measure of complexity should have and propose a universal measure called _
depth_. they suggest that the depth of a state should depend on the complexity of the process by which that state arose , and prove that it must be proportional to the shannon entropy of the set of trajectories leading to that state .
subsequently , ekroot and cover @xcite studied the computational aspect of the depth measure . in order to quantify the number of bits of randomness in a markov trajectory , they propose a closed - form expression for the entropy of trajectories of an irreducible finite state markov chain .
their expression does not allow , however , for computing the entropy of markov trajectories conditional on the realisation of a set of intermediate states .
computing the conditional entropy of markov trajectories turns out to be very challenging yet useful in numerous domains , including the study of mobility predictability and its dependence on location side information .
consider a scenario where we are interested in quantifying the predictability of route - choice behaviour .
we can model the mobility of a traveller as a weighted random walk on a graph whose vertices represent locations and edges represent possible transitions @xcite
. we can therefore model a route as a sample path or trajectory in a markov chain . if we suppose that we know where the traveller starts and ends her / his route , the randomness of the route she / he would follow is represented by the distribution of trajectories between the source and destination vertices . consequently , the predictability of her / his route is captured by the entropy of markov trajectories between these two states .
now , if we obtain side information stating that the traveller went ( or has to go ) through a set of intermediate vertices , quantifying the evolution of her / his route predictability requires the computation of the trajectory entropy conditional on the set of known intermediate states .
the conditional entropy is also a way to quantify the informational value of the intermediate states revealed .
for example , if the entropy conditional on the set of known intermediate states is zero , then this set reveals the whole trajectory of the traveller . in our work ,
we propose a method to compute the entropy of markov trajectories conditional on a set of intermediate states .
the method is based on a transformation of the original markov chain so that the transformed markov chain exhibits the desired conditional distribution of trajectories .
we also derive an expression that enables us to compute the entropy of markov trajectories , under conditions weaker than those assumed in @xcite .
moreover , this expression links the entropy of markov trajectories to the local entropies at the markov chain states . |
the advent of the _ chandra x - ray observatory _ enables us to study the intracluster medium ( icm ) in clusters of galaxies with superb spatial resolution .
_ chandra _ has revealed the dynamics of substructures in the central region of clusters . in some clusters , sharp contact surfaces between the cold gas in merging subcluster cores and the hot icm have been observed @xcite .
these contact surfaces are often called ` cold fronts ' .
they are interpreted as cool cores from subclusters which are moving through the main body of the cluster without losing their identity @xcite .
the existence of those cold fronts also provides information about fluid dynamics , transport properties , and magnetic fields in the icm .
@xcite argued that magnetic fields provide a surface tension that stabilizes the cold fronts against the development of kelvin - helmholtz ( kh ) instabilities .
_ chandra _ has also provided more detailed information on cluster merger shocks than the previous observations with _
rosat _ and _ asca _ ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . for abell 665
, @xcite showed that radio emission comes from high temperature shock regions ; this suggests that relativistic electrons are accelerated in merger shocks .
another important result from _ chandra _ is the discovery of complex x - ray structures in the central regions of cooling flow clusters with radio sources .
recent _ chandra _ images and earlier results from _ rosat _ show that central radio sources in cooling flow clusters can inflate bubbles of nonthermal plasma which displace the x - ray emitting gas @xcite .
the x - ray images of the central regions of the hydra a and perseus clusters revealed minima in the x - ray emission coincident with the radio lobes @xcite .
subsequently , similar x - ray minima , corresponding to radio bubbles , were found in abell 2052 and mkw3s @xcite ; a particularly strong interaction between the icm and radio plasma was found in abell 2052 @xcite .
for hydra a , @xcite suggested that the agn activity may prevent the gas from cooling to low temperatures ( see also * ? ? ? * ) .
this suggests that these central regions are not simple , spherically - symmetric , hydrostatic systems .
a wealth of previous x - ray observations has shown that gas in the cores of many clusters is cooling through at least part of the x - ray emitting temperature range from @xmath2@xmath4 to @xmath2@xmath5 k ( see * ? ? ? * for a review ) . since the cooling time of gas in many cluster cores is well below the age of the universe , cooling flows are the natural consequence of radiative cooling if there are no balancing heat sources .
more recent spectral observations suggest that the gas does not continue to cool down to low temperatures at the rates suggested previously @xcite .
observations had also suggested this ( see * ? ? ? * for a review ) .
abell 133 is an x - ray luminous cluster at @xmath6 @xcite with a central cooling flow @xcite .
the central cd galaxy is a radio source @xcite .
there also is a diffuse , very steep spectrum radio source @xmath230 kpc north of the center of the cd galaxy , which has no optical counterpart and is classified as a radio relic @xcite .
the radio relic source has an unusual filamentary structure .
a comparison of previous _ rosat _ x - ray and _ vla _ radio observations revealed an excess of x - ray emission near to the cluster center which was spatially coincident with the radio relic @xcite .
@xcite argued that the enhanced emission was the result of the interaction between a jet from the cd nucleus with hot icm .
@xcite indicated that the excess x - ray emission in the @xmath7 kev band is too high to be attributed to the inverse compton ( ic ) emission from the high energy particles in the relic alone .
on the basis of _ rosat _ data alone , one can not draw a conclusion about the interaction between the radio plasma and the hot icm because of _ rosat _ s limited spatial resolution and bandwidth . in this paper , we present _ chandra _ observations of the central region of abell 133 , which we use to study the x - ray morphology , temperature distribution , and the contribution of nonthermal emission . we assume @xmath8 , @xmath9 , and @xmath10 unless otherwise mentioned . at a redshift of 0.0562 , 1corresponds to 1.07 kpc .
abell 133 was observed with _
chandra _ on 2000 october 1314 for a useful exposure time of 35.5 ks .
the observation was made with the center of the cluster located between node boundaries near the aim point on the back - illuminated chip s3 .
the detector temperature during the observation was @xmath11 c. hot pixels , bad columns , and events with grades 1 , 5 , and 7 were excluded from the analysis .
a few periods of time containing background flares were removed using the lc_clean software provided by maxim markevitch . ] .
only data from the s3 chip are discussed here .
figure [ fig : csm]a shows the raw _ chandra _ image of the inner 2@xmath122 region of the cluster in the @xmath13 kev energy band , uncorrected for exposure or background .
figure [ fig : csm]b shows an adaptively smoothed image of the same region .
chandra _ interactive analysis of observation ( ciao 2.0 ) routine csmooth was used .
the image has a minimum signal - to - noise ratio of three per smoothing beam and was corrected for exposure , vignetting , and background .
the background was taken from blank sky observations compiled by maxim markevitch .
complex x - ray structures are seen within @xmath14 of the center of the cluster .
the most prominent is a `` tongue''-like feature which extends from the cluster center to the northwest , and ends in two knots of emission .
this tongue can be identified with the excess x - ray emission feature observed by _ rosat _
1b in @xcite and fig .
12b in @xcite ) .
to either side of the tongue , the x - ray surface brightness is rather low . at the southeast edge of the faint region to the east of the tongue
, there is a diffuse arc of brighter x - ray emission which curves toward the northeast .
the core region appears to have a relatively sharp , curved edge to the southeast . within the core region ,
there are several bright knots of x - ray emission , as well as the two knots at the end of the tongue .
several of these knots appear to be small enough to be due to unresolved point sources .
we used the wavelet source detection method to detect possible point sources in the image .
the ciao wavdetectroutine was used with the significance threshold set at @xmath15 . within the central region of the cluster
, three sources were detected with sizes consistent with point sources ; they are indicated in figure [ fig : point ] .
source c is the western knot at the tip of the tongue of x - ray emission . in figure
[ fig : radio ] , the 1.4 ghz radio image from @xcite is shown in green , superposed on the adaptively smoothed @xmath16 x - ray image ( figure [ fig : csm]b ) in red .
contours from the radio image are also shown in figure [ fig : point ] .
a fairly compact radio source is associated with the central cd galaxy .
this appears in yellow in figure [ fig : radio ] as this is also a peak in the x - ray image .
this figure also shows the extended filamentary radio relic source to the nnw .
the bright knots at the tip of the x - ray tongue are projected on the radio relic source , to the west of the brightest portion , and the yellow - orange color indicates that this region is bright in both radio and x - rays .
however , most of the radio relic lies to the east and west of the tongue , projected on regions of relatively low x - ray surface brightness .
an interesting aspect of the radio image are the long filaments extending to the east and south - east of the main portion of the relic .
these filaments appear to wrap around the `` wing '' of brighter x - ray emission in this direction .
also , there is some evidence that a shorter radio filament wraps around the northern edge of the tip of the x - ray tongue . in figure
[ fig : xopt ] , the x - ray contour plot of the central region of abell 133 is overlaid on the optical digital sky survey ( dss ) image . as can be seen
, there is no optical counterpart of the x - ray tongue , neither for the entire tongue , nor for the point - like knots ( including source c ) at the end of the tongue .
x - ray source b corresponds to a galaxy ( 2masxi j0102379@xmath17215305 ) ; the optical position agrees with the x - ray position to within 1 and the galaxy is a member of abell 133 @xcite .
this galaxy is also a compact radio source @xcite .
the x - ray source a is close to the optical center of the cd galaxy and the core of the central radio source in the cd galaxy , but is separated by about @xmath18 to the south .
it is not clear whether source a is really a distinct point source , perhaps due to the radio active galactic nucleus ( agn ) of the cd galaxy , or whether it is just the peak in the cluster surface brightness profile .
the position of the cd galaxy , central radio source , and central x - ray peak are all shifted from the larger scale center ( see [ sec : spatial_profiles ] ) of the x - ray emission toward the south . to help visualize the faint x - ray structures around the cd galaxy , we compare the observed cluster to a smooth elliptical isophotal model .
elliptical isophotes are fitted to the image using the iraf / stsdas task ellipse . before the fitting , we smoothed the image with a gaussian kernel with a width of 1 .
since the x - ray structure around the cd is very complicated , the ellipticity , position angle , centroid , and intensity of the elliptical isophotes are all allowed to vary , except that the centroid of the innermost elliptical isophote is fixed to the position of the cd galaxy .
the fits to the elliptical isophotes are used to create a smooth model of the image , which is subtracted from the original x - ray image of the cluster .
the counts of the model image are multiplied by 0.5 before the subtraction to avoid oversubtraction .
figure [ fig : residu ] shows the subtracted image of the cluster .
the cluster is nearly axisymmetric about the tongue ; faint wings extend from the center to the ne and sw .
this `` bird - like '' structure and the shift of the cd galaxy toward the south or southwest may indicate that the cd galaxy and its environment are moving toward the south or southeast ( see [ sec : disc_tongue_kh ] ) .
note that if we do not multiply by 0.5 before the subtraction , the sw structure becomes unclear because the sw structure is a `` local '' excess , while on the average , the region to the north of the cd is brighter than the south in the innermost region around the cd . unlike the other merging galaxy clusters abell 2142 @xcite , abell 3667 @xcite , rx j1720.1 + 2638 @xcite , abell 2256 @xcite , ms 1455.0 + 2232 @xcite , and 1e0657@xmath1756 @xcite , abell 133 does not have clear surface brightness discontinuities except around the nw tongue and the complex structure within @xmath19 from the center ( figures [ fig : csm ] and [ fig : xopt ] ) . in order to study the x - ray structure quantitatively
, we extracted surface brightness profiles for the @xmath13 kev band for three sectors shown in figure [ fig : point ] .
the resulting surface brightness profiles are shown in figure [ fig : surf ] .
they were derived by accumulating counts in circular annular wedges .
the range of radii covered by each annulus is shown by the horizontal bars attached to each point .
there are vertical error bars , but they are too small to be easily seen in the figure . because of the complexity of the structure in the central regions , it is difficult to select a center for the sectors .
the center that we selected ( @xmath20 ; @xmath21 , j2000 ) is the centroid of the bright core of the x - ray emission , which we took to be outlined by the contour having a surface brightness of 4 counts / pixel ( 0.310 kev ) .
as is obvious from figure [ fig : point ] , the sector center does not coincide with the position of the cd galaxy ( @xmath22 ; @xmath23 , j2000 ) , nor does it lie on any of the brighter knots of x - ray emission within the core . however , it does lie along the extension of the x - ray tongue into the center of the core , which makes it easier to isolate the surface brightness profile of the tongue ( nw2 sector ) . with this center
, the nw1 profile passes through much of the radio - emitting area outside of the region of the tongue .
also , the drop in the surface brightness associated with the edge of the bright core region ( figures [ fig : csm ] , [ fig : xopt ] , and [ fig : residu ] ) can be studied with this center .
the se profile shows that there is a small jump corresponding to a possible bow - shaped feature about 30 from the central cd ( figures [ fig : csm ] , [ fig : xopt ] , [ fig : residu ] , and [ fig : surf ] ) . while the se profile is relatively smooth
, the nw1 profile shows a depression at @xmath24 relative to a smooth interpolation of the declining surface brightness from smaller to larger radii .
the position of this depression corresponds to that of the radio relic ( figure [ fig : point ] ) .
the nw1 sector also features a high surface brightness region associated with one of the x - ray knots near the center of the cd ( at 8) .
the nw2 profile shows a hump or plateau at about the same radius of the depression , followed by a very rapid drop in surface brightness at about 40 ( figure [ fig : surf ] ) .
the hump represents the tip of the x - ray tongue . for the se sector
, we also determined the electron density profile by deprojecting the surface brightness profiles in figure [ fig : surf ] assuming that the density distributions are spherically symmetric and the density is constant in spherical shells ( figure [ fig : density ] ) . in the se sector , the density declines smoothly except for a small jump at a radius of about 30 , which corresponds to the small jump in figure [ fig : surf ] .
for the nw sectors , we show the density profiles outside the tongue and the relic because the assumption of spherical symmetry is not correct in the central regions .
although we analyze the spectra in detail for a number of regions , a sense of the overall temperature distribution of the cluster can be found from an x - ray color image . for three energy bands ( @xmath25 , @xmath26 , and @xmath27 kev ) , we make adaptively smoothed images as in figure [ fig : csm]b .
we allocate the colors of red , green , and blue to the @xmath25 , @xmath26 , and @xmath27 kev band images , respectively , and combine the three images .
figure [ fig : color ] shows the result ; the cluster center and the tongue are red , implying that these contain cooler gas . in particular
, the tip of the tongue appears to be the coolest .
there are no strong , sharp features which are blue ( hard spectra due to hot gas ) and which might be attributed to shocks .
we extracted spectra in the @xmath28 kev band in pi channels from selected regions of the _ chandra _ image using the ciaosoftware package .
we used the gain file acisd2000-01-29gainn0001.fits and the fef file acisd2000-01-29fef-pin0001.fits .
both the response matrix files and the ancillary response files were calculated using the calcrmf / calcarf 1.07 package written by alexey vikhlinin and jonathan mcdowell .
the package weighted the response files by the x - ray brightness over the corresponding image region .
the spectra were grouped to have a minimum of 20 counts per bin , and fitted with one or two thermal models ( mekal , * ? ? ?
* ; * ? ? ?
* ) using xspec 11.00 .
in general , we allowed the absorbing column to vary from the galactic value . in some regions , we fitted the spectra with a power - law model .
errors on fitted spectral parameters are given at the 90% confidence level .
first , we studied the average spectrum extracted from the innermost circular region of 25 radius .
we found that this spectrum can not be reproduced by a single thermal component ( table [ tab : tot ] ) , because this cluster has a temperature gradient ( [ sec : spectra_rad ] ) .
the temperature at the cluster center is much lower than that in the outer region , so we fitted the spectrum with two thermal components with the same absorption .
we assumed that the metallicities of the two thermal components were the same , because we could not constrain the metal abundance when they were fitted independently .
we show the result in table [ tab : tot ] . for comparison , we present the previous results on abell 133 from the _ einstein _ and _ exosat _ observatories @xcite . the higher of the two temperatures obtained with _
chandra _ is consistent with the temperatures obtained from _
einstein _ and _
exosat_. because of their relatively large fields , _ einstein _ and _ exosat _ appear to be affected by the temperature of the outer region .
the _ chandra _
flux is smaller than that measured previously because our spectrum applies only to the central portion of the cluster .
hydrogen column density is higher than the galactic value ( @xmath29 ; * ? ? ?
figure [ fig : temp ] shows the radial profiles of temperature for four sectors .
the position angles are @xmath30 ( ne ) , @xmath31 ( se ) , @xmath32 ( sw ) , and @xmath33 ( nw ) , where the angles are measured from north through the east . the center of the sectors is the same as that of the sectors in figure [ fig : point ] . for the nw sector
, we derive the temperature of the x - ray tongue separately from that of the surrounding emission .
the spectra were fitted by a mekal model with variable absorption . for most regions ,
the single temperature model can reproduce the spectra ( @xmath34/dof @xmath35 ) , where dof is the degrees of freedom , except for the region of the tongue ( @xmath34/dof @xmath36 ) . in figure
[ fig : temp ] , the temperature increases gradually outward except for the nw sector . for the nw sector
, the temperature shows a sudden rise with radius at @xmath37 for the profile excluding the tongue .
a radius of @xmath37 corresponds to the southern edge of the deficit of x - ray emission associated with the radio relic ( figure [ fig : surf ] ) .
the fitted absorbing column density in the central regions is @xmath38 .
the profile is not inconsistent with a uniform absorbing column .
the temperature profiles in figure [ fig : temp ] do not change significantly even if we fix the absorbing column density at the galactic value ( @xmath29 ; * ? ? ?
we investigated more closely the x - ray spectra of the regions containing the radio relic and the tongue .
the regions are shown in figure [ fig : relic ] . in the radio relic ,
our main purpose was to detect or limit the contribution of nonthermal inverse compton ( ic ) emission to the spectrum .
we first fitted each region s spectrum using a single temperature model with a variable absorbing column density @xmath39 .
this model is referred to as `` 1 t '' .
the free parameters in this model are the temperature , @xmath40 , the metallicity of the gas , @xmath41 , measured relative to the solar photosphere values of @xcite , the absorbing column density , @xmath39 , and the normalization , @xmath42 . in the `` 1tpl '' model
, we added a power - law component to the 1 t model .
thus , the normalization and the power - law photon index ( @xmath43 ) were added to the free parameters . on the other hand
, we fixed the metallicity in the 1tpl model at the value of the 1 t model , because it could not be constrained .
given the radial temperature gradient ( figure [ fig : temp ] ) , spectral complexity may also result from several different emission temperatures along the line of sight .
thus , we also tried a two - temperature model `` 2 t '' with a variable absorption .
we assume that metallicities of the two thermal components are the same . for the radio relic region , the model with a power - law nonthermal component ( 1tpl )
is slightly favored ( table [ tab : region ] and figure [ fig : sp]a ) .
however , the difference in @xmath34 is not very significant . given the likely calibration uncertainties , we do not regard the detection of a power - law component as established .
instead , we take the upper limit on the flux of this component as an upper limit on any power - law nonthermal spectral component
. the upper limit on the @xmath13 kev flux of the power - law component is @xmath44 .
we discuss the implications of this in [ sec : disc_relic_b ] .
the spectrum of the x - ray tongue is not fitted very well by a single temperature model , but is fitted adequately by either the 2 t or 1tpl model .
the 2 t model is a somewhat better fit . in the 1tpl model
, the fraction of the flux in the power - law component is relatively small ( 24% of the 0.310 kev flux ) , so most of the emission is thermal and relatively cool in any case .
we note that the emission from the tongue can be explained by the superposition in projection of cold thermal emission from the small volume of the tongue , and the ambient hotter thermal emission in the cluster in front of and behind the tongue .
this would account naturally for the need for two temperature components .
to test this idea , we tried a model in which we fixed the absorbing column , the temperature , and the metallicity of the hotter gas spectral component to those derived from the spectrum of the region surrounding the tongue .
we also fix the normalization of the hotter component to the expected value given the foreground and background volume of the tongue region .
this model is listed as `` 2t@xmath45 '' in table [ tab : region ] .
the f - test indicates that this constrained model is as good a fit as the unconstrained two temperature model .
we find that the intrinsic temperature of the cold gas in the tongue is 1.3 kev in the 2t@xmath45 model . in this model
, the metallicity of the cooler component in the tongue is allowed to differ from that of the surrounding gas ; the best - fit value is slightly lower but agrees within the errors .
nonthermal emission from the tongue is not required ( table [ tab : region ] and figure [ fig : sp]b ) . in this projection model , we assume that the tip of the tongue is a sphere with a radius of 8 kpc @xmath46 .
the required electron density in the tongue is approximately @xmath47 , which implies that the tongue is nearly in pressure equilibrium with the ambient medium at the same radius , whose pressure is @xmath48 kev @xmath49 @xmath50 dyn @xmath51 . while there is a peak in the x - ray surface brightness at a position that nearly coincides with the center of the cd galaxy which is the host of a radio source , it is uncertain whether this is a peak in the density of the diffuse x - ray gas or x - ray emission from an agn .
we determined the x - ray spectrum from a circular region with a radius of 5 centered on the x - ray brightness peak .
the emission appears to be purely thermal , and does not show any evidence for a nonthermal , power - law component .
the spectrum was consistent with an absorbing column that is identical to that for the surrounding region .
these both suggest that the peak is emission from diffuse hot gas , and not primarily due to an agn . in the central regions of the abell 133 cluster , the gas densities are high and the radiative cooling times are short .
thus , we fitted the spectra of the inner regions using a cooling flow model .
for ease of comparison with figure [ fig : temp ] , the spectra were accumulated in a number of annuli , centered on the center of the sectors in figure [ fig : point ] .
the radii of the three annuli were selected to cover the central bright region except the tongue , the region containing the tongue , and the largest outside region fitting completely on the s3 chip . as a spectral model
, we use the mkcflow model based on the mekal plasma emission code .
we include any intrinsic absorption , @xmath52 , that may be acting on the cooling flow in the core of the cluster using the zwabs model .
moreover , we add a mekal model representing the emission of the icm outside the cooling flow , subject to galactic absorption using the wabs model .
we fix the galactic absorption at @xmath29 @xcite .
the combination of the models is similar to that of @xcite : @xmath53 \:.\end{aligned}\ ] ] the free parameters are given in parentheses .
the intrinsic absorption @xmath54 acts only on the cooling flow emission .
the normalization of the mkcflow model , @xmath55 , is the so - called cooling rate or mass deposition rate of the flow .
we fix the metallicity @xmath41 and initial gas temperature @xmath56 in the mkcflow component to the values of metallicity and temperature of the mekal component , respectively .
the results are shown in table [ tab : cool ] .
the total mass deposition rate , @xmath57 , increases outward .
we note that the mass deposition rate of the whole region ( @xmath58 ; @xmath59 kpc ) is somewhat larger than the deposition rate of @xmath60 ( corrected to @xmath61 ) from @xcite , although the errors overlap . for the entire region ,
the luminosity of the cooling flow component is @xmath62 .
we note that if we do not use the data below 0.9 kev , @xmath54 is significantly reduced ( parenthesis in table [ tab : cool ] ) , and at the same time , @xmath55 also decreases .
both @xmath54 and @xmath55 strongly depend on the data below 0.9 kev and they correlate with one another .
this effect suggests that there may still be some calibration uncertainties for _ chandra _ below 0.9 kev .
thus , we regard the evidence for excess absorption as preliminary .
the most prominent feature of the _ chandra _ image of abell 133 is the tongue of colder x - ray emitting gas that extends from the center .
cold filaments similar to the tongue are seen in a1795 and mkw3s with _
@xcite and in virgo with _ xmm - newton _ ( @xcite ) and _ rosat _ ( @xcite ) .
we discuss several possible explanations for the tongue .
the four panels in figure [ fig : ori ] illustrate these possibilities schematically .
one possible origin of the tongue is a cooling wake ( figure [ fig : ori]a ) .
that is , the tongue might be gas which is cooling from the hot icm , attracted into a wake along the path of the moving cd galaxy @xcite . however , the temperature discontinuity between the tongue and the ambient hot gas is very sharp .
we extracted the x - ray spectra of @xmath63 wide regions along but just outside both sides of the tongue and find the temperature is @xmath64 kev , which is significantly higher than that of the tongue ( @xmath65 kev ) .
this shows that the temperature of the gas changes on a scale of @xmath66 kpc . if the cooling wake scenario is correct , we expect a smooth temperature profile around the tongue .
thus , our result does not seem to support the idea that the tongue is the cooling gas accreted from the ambient hot gas .
recently , numerical simulations of cluster mergers by @xcite showed that ram pressure induced by a cluster merger can displace the core gas from the cluster s potential center . as a result ,
the gas becomes convectively unstable , and in some cases , a convective plume of low entropy gas is formed . in the region of this plume , the variations in entropy of the gas are much greater than the variations in the pressure ; thus , the low entropy plume has a high density , a low temperature , and a very high x - ray emissivity .
the plume and wings with low entropy gas shown at @xmath67 and 6.00 gyr in figure 14 of @xcite are remarkably similar to the x - ray distribution of abell 133 ( figures [ fig : csm ] and [ fig : residu ] ) .
thus , based on the x - ray image alone , merger - induced convection would be a strong candidate to explain the origin of the x - ray tongue and overall x - ray morphology of the core of abell 133
. however , other observations are not fully consistent with this scenario , at least if the simulations in @xcite are typical . in this mechanism
, the plume of dense gas must be falling towards the center of the cluster potential , and figure 14 in @xcite shows that the tip of the plume should be at the center of the cluster potential .
since the stars in the cd galaxy will not be displaced by ram pressure , the optical center of the cd galaxy should be near to the center of the cluster potential .
however , the _ chandra _ image shows that the cd galaxy is not at the tip of the x - ray tongue ( figure [ fig : point ] ) .
it is possible that the cd galaxy has orbital motion which results in a separation from the center of the cluster potential . in support of this idea
is the observation that the cd is displaced from the center of the cluster determined on larger scales from the x - ray image ( [ sec : spatial_profiles ] ) . the time when the convective plume is seen corresponds to the time between the @xmath68 , @xmath692 gyr and @xmath693.5 gyr panels in figures 6 and 7 of @xcite . at this time
, a smaller merging subcluster should be located @xmath70 mpc to the northwest of the main cluster center .
merger shocks would be located to the southeast ( @xmath71 mpc ) and to the northwest near the smaller subcluster .
there is no clear optical evidence for a concentration of galaxies or a cd galaxy to the northwest where the subcluster should be located .
moreover , an x - ray image obtained with _ rosat _ , which has a larger field of view than _ chandra _ , showed that there is no significant substructure at larger radii @xcite .
if this merger scenario can be applied to abell 133 , the radio relic may be induced by a shock associated with the same merger .
this would probably require that the merger be at a slightly later stage than suggested by the @xcite simulations , and that the smaller subcluster ( located to the northwest ) is now falling back into the main cluster ( just after the @xmath72 $ ] , @xmath693.5 gyr panels in figures 6 and 7 of @xcite ) .
a weak bow shock would be formed ahead of this subcluster ; this shock would now be located between the subcluster and the cd galaxy of abell 133 , and might be accelerating particles to form the observed radio relic .
it would be a coincidence that the tip of the x - ray tongue and the radio relic overlapped .
moreover , we do not find a temperature jump corresponding to a shock at the position of the radio relic .
below , we argue more generally that the radio relic is not produced by a merger shock ( [ sec : disc_relic_origin ] ) . the complicated x - ray structures in the inner region of the cluster ( @xmath73 ) and in the tongue suggest that some instability is developing at the cluster center . in this subsection , we show that the kelvin - helmholtz ( kh ) instability is a possible candidate .
in fact , the shift of the cd galaxy from the x - ray geometrical center to the se and the bird - like feature in figure [ fig : residu ] may suggest that the galaxy is moving toward the se .
in contrast with abell 133 , distinct smooth discontinuities of surface brightness or cold fronts have been found for abell 2142 , abell 3667 , rx j1720.1 + 2638 , a2256 , and ms 1455.0 + 2232 @xcite . for these clusters
, the cold front is interpreted as the boundary between cold gas around the central cd galaxies and the ambient hot gas . from the shape of the cold front
, the authors concluded that the cd and any associated subcluster are moving in the direction of the cold front .
@xcite indicated that the kh instabilities on small scales must be suppressed over an extended region ( angular width @xmath230@xmath74 ) near the stagnation point at the cold front for abell 3667 , and that magnetic fields are likely to be responsible for this suppression . on larger scales ,
the relatively smooth distributions of the cold gas at the forward edge of these cold fronts suggest that the kh instability on the scale of the entire cold core is also suppressed or has not developed to the level that it significantly changes the core shape for these clusters
. when a gravitational field exists , perturbations at a sharp interface with wavenumber less than @xmath75 are stable if @xmath76 where @xmath77 is the gravitational acceleration at the interface between a high density region with density @xmath78 which is moving with a relative velocity @xmath79 through a surrounding medium with a lower density @xmath80 @xcite , and the density contrast is @xmath81 .
if the high density gas is approximately isothermal and hydrostatic in the gravitational field of the cd galaxy and/or an associated subcluster , and the gravitational field is approximated by an isothermal sphere , then the gravitational acceleration is roughly @xmath82 where @xmath83 is the boltzmann constant , @xmath84 is the temperature of the high density region , @xmath85 is the mean molecular weight ( 0.61 ) , @xmath86 is the hydrogen mass , and @xmath87 is the radius of the high density region .
we note that equation ( [ eq : grav ] ) may somewhat overestimate the gravitational acceleration .
this can occur if the outer part of a subcluster is stripped by tidal forces , and only the core with a flat density distribution remains . on the other hand ,
if the center of the subcluster has a density cusp rather than a core , equation ( [ eq : grav ] ) will give a more accurate estimate ( e.g. fig . 6 in * ? ? ?
we are interested in the largest scale kh instabilities with @xmath88 which would lead to a significant disruption of the entire cold front .
then , equations ( [ eq : kh ] ) and ( [ eq : grav ] ) can be rewritten to show that the gas will be stable if @xmath89 on the other hand , if gravity is unimportant , the growth rate in the linear regime at a flat interface between two incompressible fluids with different densities and relative motion is given by @xmath90 @xcite . here , we set @xmath91 .
@xcite showed that conditions ( [ eq : khu ] ) and ( [ eq : khw ] ) are consistent with results of numerical simulations . in table
[ tab : kh ] , we compare the values of @xmath92 and @xmath93 for the cold fronts in abell 133 , abell 2142 , abell 3667 , rx j1720.1 + 2638 , ms 1455.0 + 2232 and 1e0657@xmath1756 .
the observed parameters @xmath84 , @xmath87 , and @xmath94 used to derive the kh instability parameters were taken from @xcite , @xcite , and @xcite and @xcite .
because @xcite do not give the density contrast @xmath94 for abell 2256 due to the complicated geometry of the substructure , we do not discuss its cold front in the following .
we assume that the cd galaxy of abell 133 is moving toward the se .
however , there is no clear large density jump in that direction ( figure [ fig : density ] ) , and we assume that the density jump was disrupted by the kh instability . as an upper limit , we assume the density jump prior to disruption is less than the change in density from @xmath95 to 30 , which is the region over which the temperature rises rapidly . however , there is a small jump in the density @xmath96 at a radius of @xmath97 ( figure [ fig : density ] ) .
if we instead adopt this value of @xmath94 , then the maximum velocity to avoid disruption by the kh instability is reduced to @xmath98 . for abell 2142 ,
abell 3667 , and rx j1720.1 + 2638 , we also present the velocities of the cold cores , @xmath99 , obtained from the stagnation condition at the cold front @xcite .
we used @xmath99 to derive @xmath100 .
for abell 133 , we derived @xmath100 by assuming that the core velocity is larger than the upper limit on @xmath92 . for ms 1455.0 + 2232 , @xcite do not present @xmath99 because of large temperature uncertainties .
thus , we derived @xmath100 by assuming @xmath101 .
the small value of @xmath92 for abell 133 shows that the cold core is more vulnerable to the kh instability than in abell 2142 , abell 3667 , rx j1720.1 + 2638 , and ms 1455.0 + 2232 .
this may show that there is more variety in the depths of the potentials of substructures around cd galaxies than might be expected from the similarity in their optical properties .
note that table [ tab : kh ] suggests that gravity can not completely suppress the kh instability on large scales even for abell 2142 and abell 3667 ( @xmath102 ) .
magnetic fields may have aided in the suppression of large scale as well as small scale instabilities in these cases ( see * ? ? ?
alternatively , if the cores started to move very recently , the kh instability may have just started to develop @xcite , and the cores might not have been disrupted yet . in table
[ tab : kh ] , we also present the values for 1e0657@xmath1756 , which was recently observed by @xcite .
for this cluster , @xmath100 is extremely low because of the high velocity , @xmath99 .
in fact , most of the cold gas around the substructure appears to be removed @xcite .
abell 133 seems to be intermediate between 1e0657@xmath1756 and the other four clusters in the level of kh instability and in the size of the cold front . to assess the degree of kh instability more quantitatively
, we calculate the ratio @xmath103 , where @xmath104 is the crossing time of the cool core in the cluster @xcite .
the larger @xmath105 is , the more vulnerable the core is to the kh instability .
we estimate the crossing time @xmath104 from the core velocity and the maximum cluster radius that a core can reach , @xmath106 .
let the current radius of the core be @xmath107 , and let @xmath108 be the cluster radius within which the average mass density is 200 times the critical density of the universe ( @xmath109 ) : @xmath110 where @xmath111 is the mass of the cluster within @xmath108 .
for this estimate of @xmath104 , we will assume that the core has a radial orbit .
then , the maximum radius @xmath106 of the core is given by @xmath112 where @xmath113 is the cluster gravitational potential at @xmath107 .
if a cluster has an isothermal density distribution ( density @xmath114 ) , the potential difference is given by @xmath115 from equations ( [ eq : kin ] ) and ( [ eq : pot ] ) , we obtain @xmath116 the cluster mass @xmath111 can be derived from cluster temperature @xmath117 using an empirical relation .
@xcite found the relation of @xmath118 from the x - ray data of 88 clusters .
the observed cool cores appear to be near the cluster centers .
of course , we do not know where in the orbit the core started , or how many times it has passed near the cluster center .
thus , we crudely estimate the crossing times as @xmath119 . in table
[ tab : kh ] , we present the values of @xmath117 , @xmath106 , @xmath104 , and @xmath120 .
we assume @xmath121 for all the clusters . for abell 133 , we assume that @xmath122 is larger than the upper limit on @xmath92 , while for ms 1455.0 + 2232 we assume that @xmath123 .
table [ tab : kh ] shows that the value of @xmath120 for abell 133 is intermediate between that for 1e0657@xmath1756 and those for the other four clusters .
this supports our argument about the relative level of kh instability in these clusters .
ram pressure from the ambient gas is also expected to affect the cold gas around a cd galaxy .
@xcite indicated that ram pressure would affect the cold gas distribution at a radius which satisfies @xmath124 where @xmath125 is the static pressure of the gas around the cd galaxy . since the cd galaxy is not at the geometrical center of abell 133 ( figure [ fig : point ] ) , we expect that the relation is satisfied even at @xmath126 .
figures [ fig : density ] and [ fig : temp ] show that the pressure in the central region of the cluster is @xmath127 .
since there is no clear cold front for this cluster , we take the gas density around the cluster center , @xmath128 , as the density of the gas streaming relative to the cd galaxy ( figure [ fig : density ] ) . in this case
, the relative velocity of the cd galaxy to the surrounding gas must be larger than about @xmath129 to satisfy the relation ( [ eq : ram ] ) at @xmath126 .
this velocity is very similar to @xmath92 .
we note that the criteria for those two mechanisms , the kh instability and ram pressure , should be similar because both involve the competition between the momentum in the flow and gravitational acceleration .
since the x - ray substructure around the cd galaxy of abell 133 is fragile as is shown by the large @xmath120 , it is unlikely that it is a merged subcluster that has survived a violent merger .
as proposed for rx j1720 + 2638 and ms 1455.0 + 2232 , an alternative scenario which may explain the presence of the moving substructure is that it may be the result of the collapse of two different perturbations in the primordial density field on two different linear scales at nearly the same location in space @xcite .
as the density perturbations grew , the smaller scale perturbation collapsed first and formed the substructure around the cd galaxy .
the larger scale perturbation collapsed more recently , and forms the main body of abell 133 .
the substructure around the cd galaxy was initially located slightly offset from the center of the cluster and it is now oscillating around the minimum of the cluster potential well . since the initial position of the substructure is near to the cluster center , it is likely that the velocity is subsonic .
recently , @xcite showed quantitatively that substructures tend to be located near the cluster centers .
@xcite estimated the formation redshift of the central subclump in rx j1720 + 2638 using a relation between the formation redshift and the temperature of a cluster .
however , @xcite showed that this relation has a large dispersion , because for a given spatial scale the amplitude of initial perturbations can vary significantly .
even with this caveat , the low temperature of the gas around the cd galaxy of abell 133 ( @xmath130 kev ) suggests a large formation redshift for the substructure ( @xmath131 ) if the virial radius was smaller than 0.5 mpc ( figure 5b in * ? ? ? * ) .
since the cd galaxy is located to the se of the radio relic and it seems to be moving to the se , it might be responsible for the relic . @xcite
estimated that the age of the relic is @xmath132 yr .
if the cd galaxy were located at the position of the relic when the relic was formed , the velocity of the galaxy must be @xmath133 in the plane of the sky .
this velocity is larger than @xmath92 ( table [ tab : kh ] ) . on the other hand , the relative velocity between the cd galaxy and the whole cluster is @xmath134 along the line of sight @xcite
this would require that the cd velocity be at an angle of @xmath135 with respect to the plane of the sky .
the value of @xmath136 may be somewhat larger than @xmath137 yr , the transverse velocity may be somewhat smaller than 700 km s@xmath3 , and/or the radio source may have been formed at some distance from the center of the cd galaxy , as is generally true of radio lobes . on the other hand ,
if the radio relic itself is moving to the nw with the velocity of @xmath138 ( mach number @xmath139 ) as a result of buoyancy ( [ sec : disc_tongue_bubble ] ) , the kh model may be more easily reconciled with the radio relic age .
recent _ chandra _ x - ray observations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and earlier _ rosat _ images ( e.g. , * ? ? ?
* ) show that radio sources at the centers of several clusters are inflating bubbles in the icm , which presumably consist of relativistic particles , magnetic fields , and possibly extremely hot thermal gas of low density .
the radio bubbles have low x - ray surface brightnesses , presumably since the radio plasma has displaced the icm .
these hot bubbles are buoyant and should be moving outward in the cluster gravitational potential .
recent numerical simulations show that these hot bubbles may uplift cooler gas from the cluster center @xcite .
the nonthermal plasma in the bubble can produce radio emission , while the uplifted dense , cool thermal gas will emit strongly in the x - ray band .
the predicted configuration of a radio emitting bubble and uplifted x - ray gas is very similar to the one observed for abell 133 ( figure [ fig : point ] ) and also in the virgo cluster @xcite .
thus , the x - ray and radio observations may suggest that a bubble including the radio relic is uplifting the cold tongue from the cluster center .
the deficit in x - ray surface brightness at the radio relic can also be explained by this picture ( figure [ fig : surf ] ) .
moreover , the fact that the tongue seems to be in pressure equilibrium with the ambient medium is also favorable to this model @xcite .
however , assuming the tongue results from the buoyant motion of the bubble , there should be another umbrella - like x - ray excess emission at the head of the bubble due to compression of cool gas ahead of the bubble , if the buoyant motion is in the plane of the sky ( see the upper right figure of figure 8 in * ? ? ?
* ) . since the emission is not observed , the tongue and the bubble may not be in the plane of the sky as is shown in the bottom right figure of figure 8 in @xcite . for example
, if the buoyant motion were at an angle @xmath140 with respect to the plane of the sky and the bubble s depth along the line of sight is the same as its width ( @xmath141 kpc ) , the observed decrease in surface brightness ( @xmath142 % ; figure [ fig : surf ] ) would be consistent with the idea that the bubble is deficient in x - ray - emitting gas .
if we assume the age of the relic is @xmath143 yr @xcite , and the bubble including the relic was formed around the cd galaxy , the relative upward velocity must be @xmath144 .
more realistically , if the velocity vector of the relic is at @xmath140 to the plane of the sky and the relic was formed by jet activity away from the cd galaxy , say at a distance of @xmath145 kpc from the cd , the upward velocity is still @xmath146 .
the sound velocity of the surrounding gas ( @xmath147 kev ) is @xmath148 .
thus , the mach number of the rising bubble is similar to the values of @xmath149 predicted by @xcite . assuming that the density of the tip of the tongue is @xmath150 and the temperature is 1.3 kev ( [ sec : spectra_reg ] ) , the radiative cooling time is @xmath151 yr .
this is about an order of magnitude longer than the estimated age of the radio relic .
thus , if the tongue was uplifted by the buoyant motion of the radio relic , the low temperature of the tongue can not be the result of radiative cooling at its current position .
the gas in the tongue presumably cooled radiatively when the gas was located closer to the center of the cd before it was uplifted .
in addition , adiabatic cooling may have lowered the temperature further as the gas was lifted to regions with a lower ambient gas pressure . given the high density of gas in the tongue and its large volume , one concern is whether the radio bubble could lift such a large mass and remain buoyant .
for the purpose of this estimate , let us assume that the radio relic is filled with nonthermal and possibly thermal plasma having a very low density so that its x - ray emissivity is low .
we further assume that the extent of the radio bubble along the line of sight is similar to its largest diameter in the plane of the sky ; we treat the radio bubble as an oblate spheroid .
this is consistent with the flattened , `` mushroom cloud '' shape expected for a buoyant bubble @xcite .
then , the total volume of the radio bubble is @xmath152 kpc@xmath153 . on the other hand ,
we assume that the extent of the tongue along the line of sight is similar to its narrower dimension ; we treat it as a prolate spheroid , which is consistent with the uplifted trail below a convective bubble @xcite .
then , the volume of the tongue is @xmath154 kpc@xmath153 . for an average tongue density of @xmath155 ,
this implies a total mass of @xmath156 , thus , the average density of the material in the radio bubble and tongue combined is @xmath20.001 @xmath49 .
this is still much lower than the typical gas densities at similar distances on either side of the cd galaxy in abell 133 ( figure [ fig : density ] ) .
thus , it is at least plausible that the radio bubble could have lifted the tongue via buoyancy .
it is interesting to estimate the impact of the radio bubble on the cooling flow of abell 133 .
if bubbles transport a large amount of cold gas from the cluster centers to the periphery of cooling flows , they may explain the lack of cooled gas in the cooling flows @xcite . as noted above
, the total mass of the tongue is @xmath157 .
this material has been lifted over the lifetime of the radio relic , which implies an average outflow rate of @xmath158 yr@xmath3 .
this is somewhat smaller than the cooling inflow rate derived from the x - ray spectrum on the same radial scale ( table [ tab : cool ] ) .
however , the total cooling radius for abell 133 is much larger than the size of the x - ray tongue @xcite .
thus , the tongue seems to reside well within the cooling radius of the cluster , and would therefore not actually transport gas outside of the cooling radius .
moreover , as is shown in table [ tab : cool ] , the mass deposition rate derived by spectral analysis is increasing outward at the position of the tongue .
so , buoyant mass transport may affect the local mass cooling inflow rate near the center of abell 133 , but it probably does not affect the overall cooling rate . the energy associated with the expansion and motion of the radio bubble might affect the energetics of the x - ray gas , and might partially counteract radiative cooling @xcite . assuming that the radio bubble is adiabatic , the work done by the radio bubble on the surrounding gas is @xmath159 where @xmath160 and @xmath161 are the pressure and the volume of the radio bubble , respectively , and @xmath162 is the adiabatic constant .
the subscripts 0 and 1 stand for the values when the radio bubble is formed and those at present , respectively .
if we assume that the radio bubble was formed near the cluster center and has been nearly in pressure equilibrium with the surrounding gas , @xmath163 and @xmath164 ( figures [ fig : density ] and [ fig : temp ] ) . moreover , if we assume that the volume of the observed radio bubble is @xmath165 and @xmath166 , the initial volume is @xmath167 and the work is @xmath168 .
moreover , the gravitational energy released by matter displacement is @xmath169 where @xmath170 is the density of the gas surrounding the tongue , and @xmath171 is the length of the tongue .
the gravitational acceleration around the tongue is given by @xmath172 thus , the gravitational energy is @xmath173 , and the total energy released by the bubble motion is @xmath174 .
if the work has been done during the past @xmath137 yr , the rate of work is @xmath175 . on the other hand ,
the luminosity of the cooling flow component is @xmath176 ( see [ sec : spectra_cool ] ) , which is smaller than @xmath177 .
thus , the energy dissipated by the expansion of the radio bubble may compensate the energy radiated by cooling gas at least for this cluster at present , although it is uncertain what fraction of the dissipated energy is converted to thermal energy ( but see * ? ? ?
although the buoyant radio bubble model may explain the observed tongue , at present it is not clear what conditions are required to produce uplifting of dense cool gas . for example , the numerical simulations of @xcite do not show the formation of a cold x - ray tongue . assuming that the unusual radio source to the north of the cluster center in abell 133 is related to the cluster , there have been at least two suggestions as to its origin .
first , the source might be a cluster radio relic ( e.g. , * ? ? ?
radio relics are diffuse radio sources with steep spectra which are not associated with a radio galaxy in the cluster and which are generally located far from the center of the cluster ( e.g. , * ? ? ?
generally speaking , radio relic sources tend to be moderately polarized .
this and their outer location distinguish radio relics from radio halos .
radio relics ( and radio halos ) are generally associated with clusters which appear to be undergoing a merger .
a common theoretical interpretation is that radio relics are produced by acceleration , re - acceleration , or adiabatic compression of relativistic electrons and magnetic fields by merger shocks .
some of the radio properties of the source in abell 133 seem at variance with those of other cluster radio relics .
first , the radio source is projected only @xmath178 40 kpc from the cd galaxy of abell 133 .
of course , it is possible that it is actually at a much larger radial distance , and only appears near the center in projection .
there are several other radio relics seen in projection near cluster centers , including the sources in abell 13 and abell 4038 @xcite .
second , the integrated polarization of the radio relic source in abell 133 is low ( @xmath2 2.3% * ? ? ?
* ) , unlike most radio relics .
however , the low integrated polarization is mainly the result of variations in the direction of the polarization across the source .
figure [ fig : radio_polar ] shows the radio polarization at 1.465 ghz as a function of position in the radio relic source ; the data comes from the same observations discussed in @xcite .
the data in figure [ fig : radio_polar ] are not corrected for faraday rotation . in a number of comparatively large areas within the main body of the relic the polarized fraction varies between 3 and 20 percent , the larger fractions suggesting that either the faraday rotation in these areas is not large or that it is fairly uniform .
note that even rotation measures as small as @xmath179 rad m@xmath180 would produce faraday rotations of @xmath181 at 1.4 ghz .
in general , radio sources located near the centers of cluster cooling flows show very high faraday rotations ( @xmath182 rad m@xmath180 ; e.g. , * ? ? ?
* ; * ? ? ?
the fact that strong polarization is observed in the abell 133 relic without correction for faraday rotation may indicate the relic is located in front of , rather than within or behind , the cluster cool core .
if the radio relic were produced by a merger shock and the magnetic field were strongly affected by compression in the shock , one would expect the magnetic field direction ( perpendicular to the electric vector ) to be mainly parallel to the shock front .
indeed , we see in figure [ fig : radio_polar ] that the electric vectors closest to the concave northern edge of the relic have the required orientation to suggest that this may be a shock front with a tangential magnetic field direction
. of course , any other mechanism which produced a compression of the northern edge of the radio source might have the same effect ; if the radio source has risen buoyantly to the north , one might also expect some compression of this edge .
however , this must be an uncertain interpretation until the electric vectors can be corrected for faraday rotation .
several other features in figure [ fig : radio_polar ] are of interest .
the polarization near the bright knots at the tip of the x - ray tongue ( @xmath183 ; @xmath184 , b1950 ) and further down the tongue to the se is quite strong and particularly variable in direction .
perhaps this indicates that the relic lies behind the tongue , and that the emission from those regions of the relic suffers strong and variable faraday rotation by the comparatively cool gas in the tongue . in the two filaments of radio emission extending to the east and south - east , the polarizations are mainly perpendicular to the filaments , which implies that the magnetic field is parallel to them .
the origin of these filaments may be connected to the magnetic field lines ( e.g. , * ? ? ?
* ; * ? ? ?
the radio relic in abell 133 is rather filamentary .
although there are several possible origins for this structure , @xcite and @xcite suggested that the structure results from the compression of old radio plasma by merger shocks . on the other hand , our _ chandra _ observations are largely inconsistent with a merger shock origin for the radio relic .
first of all , no evidence for a shock is seen in the image ( figure [ fig : point ] ) , x - ray colors ( figure [ fig : color ] ) , or spectra ( figure [ fig : sp ] ) of the radio relic region . if the radio relic coincided with a shock , this region should be unusually bright in x - rays , due mainly to compression of the cluster gas in the shock . actually , most of the radio relic region is unusually faint in x - rays ( figure [ fig : surf ] ) .
the exception is the region where the x - ray tongue overlaps the radio relic .
however , the tongue is due to cool thermal gas , and is thus not a result of a merger shock .
an alternative explanation of the radio relic is that it is an old radio lobe from the central radio source associated with the cd galaxy in abell 133 @xcite .
an early radio map of the source suggested a bridge of emission connecting the cd radio source with the radio relic @xcite , but this does not appear in a more sensitive , higher resolution image @xcite .
many of the radio properties of the relic are consistent with its interpretation as an old radio lobe .
similar radio lobes associated with cd galaxies in cooling flow clusters generally have rather steep radio spectra ( e.g. , * ? ? ?
* ; * ? ? ?
. the displacement of the radio relic from the central cd might be due to buoyancy ( [ sec : disc_tongue_bubble ] ) , or to the motion of the central cd , perhaps as a result of a subcluster motion ( [ sec : disc_tongue_kh ] ) .
one difficulty is that one usually finds two radio lobes , presumably produced by a double - sided jet .
however , if the central cd is moving to the southeast , the two bubbles would be displaced in the same direction and might either overlap in projection or have merged together .
our _ chandra _ observations are probably more consistent with the interpretation of this radio source as a relic radio lobe formerly energized by the central cd , rather than a merger - shock generated cluster radio relic .
radio lobes associated with cd galaxies at the centers of cooling flow clusters produce holes in the x - ray emission ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , suggesting that the radio plasma has displaced the thermal , x - ray emitting gas . in abell 133 ,
the radio relic region is also somewhat fainter in x - rays than the surrounding region ( figure [ fig : surf ] ) . as noted in [ sec : disc_tongue_bubble ] , many properties of the x - ray tongue are consistent with gas which has been uplifted by a buoyant radio bubble .
in fact , the overall radio and x - ray properties of the relic source in abell 133 are similar to the eastern radio bubble ( the `` ear '' ) of m87 in the virgo cluster @xcite .
this mushroom - shaped radio bubble has a shape similar to the radio source in abell 133 , and also has a very steep radio spectrum .
the m87 ear is filamentary , although mainly around its edge . in m87
, there is a linear extension of x - ray emission due to cool , thermal gas which connects the center of the cluster with the radio bubble @xcite , similar to the tongue in abell 133 .
the radio and x - ray structure of m87 have been interpreted as the result of a buoyant radio bubble @xcite .
one difference between m87 and abell 133 is that there is radio emission connecting the ear in m87 with the nucleus of the galaxy , while the radio relic in abell 133 does not show any strong radio bridge with the central cd .
this might be a result of a lower level of current agn activity in the cd galaxy in abell 133 compared to m87 . in summary ,
our x - ray observations are more consistent with the interpretation of the radio relic in abell 133 as a displaced radio lobe from the central cd , rather than as a merger shock - induced relic .
in [ sec : spectra_reg ] , we found an ( at best ) marginal detection of nonthermal x - ray emission from a radio relic .
we now use these spectral results to constrain the magnetic field in the radio relic , under the assumption that this nonthermal emission is due to inverse compton ( ic ) scattering . in the model with a power - law component ( model 1tpl in [ sec : spectra_reg ] ) , the x - ray flux of the power - law component is @xmath185 for @xmath186 kev and the photon index is @xmath187 .
the radio synchrotron flux density of the relic is 137 mjy and flux density spectral index is @xmath188 ( flux density proportional to @xmath189 ) , both at @xmath190 ghz @xcite . for emission from a single power - law electron energy distribution ,
the ic and synchrotron emission have the same spectral index .
this would require @xmath191 , which is not consistent with our measured values . however , the ic x - ray emission in the _ chandra _ band would be due to relatively low energy relativistic electrons , which would produce very low frequency radio emission . both observations and models suggest that the radio spectrum is considerably flatter at lower frequencies @xcite , and thus the x - ray and radio spectral indices might be consistent
. we will adopt the x - ray spectral index and compare the x - ray flux with the radio flux density of @xmath192 jy at 80 mhz , the lowest frequency at which the flux density is determined accurately .
this radio flux density includes the cd galaxy and another nearby radio source , but they are much weaker than the radio relic . using the standard expressions for the ic and synchrotron emission assuming a power - law electron energy distribution ( e.g. , * ? ? ?
* ) , we find that the required magnetic field in the relic is @xmath193 .
this is about one order of magnitude smaller than the magnetic field @xcite derived from models for the radio spectral index ( @xmath194 ) or minimum energy arguments applied to the radio flux ( @xmath195 ) .
this might indicate that the magnetic field is filamentary and highly variable , which may also be suggested by the radio image of the the relic .
however , given the marginal detection of the nonthermal x - ray emission and the disparity between the magnetic field values required by the radio and x - ray observations , we believe it is best to treat the x - ray spectral result as an upper limit of the nonthermal flux .
this implies a lower limit on the radio relic magnetic field of @xmath196 , which is consistent with the radio results but not constraining .
we have presented the results of a _
chandra _ observation of the cluster of galaxies abell 133 .
the x - ray image shows a tongue of emission extending from the cd galaxy to the nw ; the tongue extends in the direction of and partially overlaps the radio relic .
there is no optical counterpart of the tongue .
the x - ray surface brightness at the position of the radio relic is smaller than the surrounding region except for the tongue , suggesting that the radio plasma has displaced the thermal gas in this region .
the x - ray spectrum indicates that the emission from the tongue is thermal , and the temperature is distinctly lower than that of the ambient hot gas .
we discussed several possible origins for the tongue ; two of these seem reasonably consistent with the x - ray and radio data .
first , the tongue may be the result of a kelvin - helmholtz instability acting on the cool core surrounding the cd galaxy as the cd moves through surrounding , lower density gas .
we show that the cd galaxy and the cold gas around it must be moving with a velocity of @xmath197 km s@xmath3 for a large scale kh instability to develop between the cold core and the ambient hot intracluster medium .
the critical velocity for the kh instability is several times smaller for abell 133 than for other clusters which show smooth , well - defined cold fronts .
the required motion of the cd galaxy is consistent with its offset position in the cluster relative to the x - ray centroid determined at large radii .
this offset can be explained if the velocity of the cd galaxy is @xmath198 .
thus the onset of a kh instability is at least plausible .
the second possibility is that the radio relic is a buoyant radio bubble , and that the tongue was uplifted by the motion of this bubble .
recent numerical simulations indicate that buoyant bubbles of nonthermal plasma produced by agn activity can uplift cold gas from the cluster center @xcite . for abell 133
, we suggest that the bubble is the radio relic , and that it uplifts the x - ray tongue .
the velocity of the radio bubble expected from the age of the radio relic is consistent with that predicted numerically .
the energy dissipated by the moving bubble may affect the cooling flow in abell 133 .
the origin of the filamentary radio relic source in abell 133 remains uncertain .
it may be a radio source that is generated by a shock wave from merging clusters , similar to the relics which are often seen in the outer regions of clusters .
in this case , the radio emitting electrons would have been accelerated , re - accelerated , and/or compressed by the cluster merger shock .
alternatively , the radio relic might be a displaced radio lobe ( or lobes ) from the radio source associated with the central cd galaxy in the cluster .
the lack of evidence for a merger shock near the relic , the anti - correlation between the radio and x - ray emission across most of the relic , and the association of the relic with the tongue of cool , dense x - ray gas all are more consistent with the relic being a displaced radio lobe or bubble , rather than being a cluster merger - shock induced radio relic .
we are grateful to e. blanton , s. randall , i. tanaka , and s. oya for useful comments .
support for this work was provided by the national aeronautics and space administration through @xmath16 award numbers go1 - 2122x and go1 - 2123x , issued by the @xmath16 x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of nasa under contract nas8 - 39073 .
l. r. acknowledges support from nsf grant ast-0071167 to the university of minnesota .
a. thanks conacyt for financial support under grant 27602-e .
cccccc chandra & @xmath199 & @xmath200 & @xmath201&0.75 & 452.8/315 + chandra & @xmath202 & @xmath203 & @xmath204&0.75 & 347.1/313 + einstein & @xmath205 & 0.5 & 1.58 & 1.31 & + exosat & @xmath206 & @xmath207 & @xmath208 & 1.43 & 24.60/17 + cccccccc relic & 1 t & @xmath209 & & @xmath210 & & @xmath211 & 156.0/128 + & 1tpl & @xmath212 & & 0.62 & @xmath213 & @xmath214 & 147.2/127 + & 2 t & @xmath215&@xmath216 & @xmath217 & & @xmath218 & 151.1/126 + tongue & 1 t & @xmath219 & & @xmath220 & & @xmath221 & 75.8/57 + & 1tpl & @xmath222&&0.28 & @xmath223 & @xmath224 & 59.1/56 + & 2 t & @xmath225&@xmath226 & @xmath227 & & @xmath228 & 52.6/55 + & 2t@xmath45&@xmath229 & 3.2 & @xmath230 ( 0.72 ) & & 1.3 & 54.6/58 + cd center&1 t & @xmath231 & & @xmath232 & & @xmath233 & 62.4/58 + cccccc @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 & 195.7/154 + & @xmath239 & @xmath240 & @xmath241 & @xmath242 & ( 164.1/140 ) + @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & 178.2/136 + & @xmath248 & @xmath249 & @xmath250 & @xmath251 & ( 154.5/127 ) + @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & 227.5/224 + & @xmath257 & @xmath258 & @xmath259 & @xmath260 & ( 215.3/210 ) ccccccccccc abell 133 & 1.4 & 25 & @xmath261 & @xmath262 & & @xmath263 & 4.0 & @xmath264 & @xmath265 & @xmath266 + abell 2142 & 7 & 250 & 2 & 720 & 900 & 5.8 & 9.4 & 0.57 & 6.2 & 1.1 + abell 3667 & 4 & 400 & 3.9 & 850 & 1400 & 6.9 & 6.5 & 0.66 & 4.6 & 0.67 + rx j1720.1 + 2638 & 6 & 200 & 2 & 670 & 500 & 8.3 & 5.6 & 0.40 & 7.8 & 0.94 + ms
1455.0 + 2232 & 5 & 200 & 2.7 & 760 & & @xmath267 & 4.8 & @xmath268 & @xmath269 & @xmath270 + 1e0657@xmath1756 & 7 & 25 & 3.8 & 1100 & 3500 & 0.2 & 14.8 & 2.1 & 5.9 & 30 + | we present the analysis of a _
chandra _ observation of the galaxy cluster abell 133 , which has a cooling flow core , a central radio source , and a diffuse , filamentary radio source which has been classified as a radio relic .
the x - ray image shows that the core has a complex structure .
the most prominent feature is a tongue " of emission which extends from the central cd galaxy to the northwest and partly overlaps the radio relic .
spectral analysis shows that the emission from the tongue is thermal emission from relatively cool gas at a temperature of @xmath0 kev .
one possibility is that this tongue is produced by kelvin - helmholtz ( kh ) instabilities through the interaction between the cold gas around the cd galaxy and hot intracluster medium .
we estimate the critical velocity and time scale for the kh instability to be effective for the cold core around the cd galaxy .
we find that the kh instability can disrupt the cold core if the relative velocity is @xmath1 .
we compare the results with those of clusters in which sharp , undisrupted cold fronts have been observed ; in these clusters , the low temperature gas in their central regions has a more regular distribution .
in contrast to abell 133 , these cluster cores have longer timescales for the disruption of the core by the kh instability when they are normalized to the timescale of the cd galaxy motion .
thus , the other cores are less vulnerable to kh instability .
another possible origin of the tongue is that it is gas which has been uplifted by a buoyant bubble of nonthermal plasma that we identify with the observed radio relic . from the position of the bubble and the radio estimate of the age of the relic source
, we estimate a velocity of @xmath2700 km s@xmath3 for the bubble .
the structure of the bubble and this velocity are consistent with numerical models for such buoyant bubbles .
the energy dissipated by the moving bubble may affect the cooling flow in abell 133 .
the combination of the radio and x - ray observations of the radio relic suggest that it is a relic radio lobe formerly energized by the central cd , rather than a merger - shock generated cluster radio relic .
the lobe may have been displaced from the central cd galaxy by the motion of the cd galaxy or by the buoyancy of the lobe . |
interest in quantum information processing has brought significant attention to the problem of measurement of tunneling currents via a microscopic system that can be modeled by a two - level hamiltonian ( such as a quantum dot , or a molecule or an atom with a localized spin @xcite ) . in the case of a single spin ,
measurements of the tunneling current in such a system provide information on spin orientation and its dynamics , and they constitute an example of indirect - continuous quantum measurement @xcite .
a fundamental question that arises is what signatures of the spin dynamics are encoded in the tunneling current and how this current affects the spin dynamics .
scanning tunneling microscopy ( stm ) experiments @xcite on a single molecule with a spin , in the presence of a magnetic field @xmath0 , have reported a peak in the current noise power spectrum ( i.e. , current - current correlation function ) @xmath15 at the larmor frequency @xmath16 , where @xmath17 is the gyromagnetic ratio .
experiments were done at room temperature and the authors found that the signal - to - noise ratio @xmath11 ( ratio of the power at the peak frequency to the shot noise power ) exceeded unity and was almost independent of the orientation of the applied magnetic field @xmath0 . in the non - relativistic approach the tunneling electrons couple to the spin by exchange interaction . in this case , electrons with a spin polarization along @xmath0 do not couple with the oscillatory components of the spin ( which are perpendicular to @xmath0 ) .
the experimental results @xcite are difficult to explain in the framework of a single - spin non - relativistic model since electrons in the leads were polarized by the same magnetic field which acted on the spin .
possible relevance of spin - orbit interaction to explain these data was discussed by shachal and manassen and later on by balatsky and martin @xcite .
recently , levitov and rashba @xcite noticed that in systems with low space symmetry ( as in dots or a molecule near a surface ) the nonvanishing orbital moment of electrons near the spin ( in the dot or molecule ) provides a strong coupling of the tunneling electrons to the spin , and they speculated that this mechanism may lead to a significant effect of the spin oscillatory component on the tunneling current .
in fact , in order to fit the experimental data @xcite one needs a model of electron tunneling via a single spin which is able to explain not only the existence of a peak in the current power spectrum , but also a significant signal - to - noise ratio @xmath18 almost independent of the orientation of @xmath0 .
in the following we consider a model with exchange ( non - relativistic ) coupling of a single spin @xmath19 and the tunneling electrons , including spin - orbit coupling . in the framework of this model
we analyze the dependence of @xmath11 and linewidth @xmath20 on the applied voltage @xmath1 between leads , the applied magnetic field @xmath0 , the temperature @xmath2 , and on the degree and orientation @xmath21 of electron polarization in the right ( @xmath4r ) and left ( @xmath4l ) leads in the steady state .
this state establishes in a transient time after the voltage or tunneling matrix elements are switched on .
the problem of tunneling via a single localized spin @xmath19 is similar to the problem of tunneling via a double quantum dot ( two - level system ) extensively studied by korotkov and averin @xcite and by ruskov and korotkov @xcite .
indeed , they were interested in the question of how quantum oscillations in the two - level system , representing a qubit , may be detected by a tunneling current .
such a system can be described by a model of electron tunneling via a spin when leads electrons are fully polarized . using the bloch equations describing the ensemble - averaged evolution of the density matrix for the coupled qubit - detector system
, they obtained oscillations at the larmor frequency in the noise power spectrum @xmath22 , with @xmath23 in the case of weak interaction between the qubit and the detector .
a similar quasi - classical approach for tunneling of electrons via a single spin was used in ref .
in such an approximation , the action of the electrons on the spin is replaced by the action of an effective _ classical _ magnetic field with a shot noise spectrum depending on the tunneling current between the leads .
this approach has sense at high currents ( voltages ) , while at low currents it does not account for the quantum nature of the tunneling electrons and hence a more elaborate solution of the quantum transport equation for both spin and electrons is desired .
here the situation is formally similar to the case of the photo - electric effect where in some circumstances a full quantum mechanical treatment of the problem provides the same information as a semi - classical approach , while in other cases , such as the measurement of the statistics of the photo - current fluctuations , the predictions of the two approaches differ qualitatively @xcite .
an important step toward a full quantum mechanical treatment of electron tunneling via a spin was recently given by parcollet and hooley @xcite .
assuming that the spin interacts only with the tunneling electrons , and electrons in the leads are in a voltage driven thermal equilibrium , they computed the spin magnetization of a quantum dot in the two - leads kondo model as a function of the temperature of the electrons , magnetic field , and voltage between the leads ( which exceeds the kondo temperature @xmath24 ) in the steady state , i.e. , in the long time limit .
they considered the situation where electrons in the leads are polarized in the same direction as the magnetic field acting on the spin , @xmath25 , @xmath26 .
their calculations confirmed previously obtained quasi - classical result @xcite that , in the steady state even at zero order in the spin - leads coupling , the magnetization is not given by the thermal equilibrium expression @xmath27 when the spin is more strongly coupled to the leads than to any other thermal bath . at long times , the state of the spin is completely determined by the characteristics of the tunneling electrons , and since the system is out of equilibrium , the steady state thus achieved is described by a steady state distribution function which differs from the gibbs distribution .
it turns out that this distribution function depends upon the voltage between the leads .
they also showed that the correct way to calculate the perturbative corrections to the spin distribution function and the spin decoherence rate @xmath20 in the steady state is by solving a quantum transport equation self - consistently .
we will extend the treatment of parcollet and hooley to arbitrary orientations of @xmath28 , and @xmath29 with respect to @xmath0 , and account for a direct tunneling of electrons between leads .
the main goals in this article are ( i ) to examine under what tunneling conditions the current - current correlation function @xmath22 exhibits a peak at the larmor frequency in the steady state , and ( ii ) to calculate @xmath30 and @xmath31 in the regime where @xmath32 , with @xmath33 denoting the charge of the electron . as in ref .
@xcite we use the keldysh formalism @xcite and the majorana - fermion representation @xcite for the spin to find the spin distribution function and the current - current correlation function .
we note , from a more general perspective , that the problem of electron tunneling via a spin represents an example of indirect quantum measurement .
the spin is probed by the tunneling electrons whose correlation function is measured in a continuous fashion with a classical apparatus ( which is not included ) . the goal of quantum measurement is to determine the frequency of precession of the isolated spin ( energy separation in the isolated two - level system ) and to obtain information on the initial state of the system .
however , when the spin is decoupled from the environment , tunneling via the spin changes its state as the measurement goes on . after switching on the voltage , or the tunneling matrix elements , there is an initial period of time where the tunneling current depends upon the initial state of the spin .
after some transient time the steady state , which is in general independent of the initial conditions , establishes . in the steady state
the frequency of precession is renormalized by the tunneling electrons , and the width of the peak ( at the precession frequency ) and corresponding signal - to - noise ratio depend upon the precession frequency and characteristics of the tunneling electrons . to extract information about the precession of the isolated spin one needs to have complete information on the tunneling electrons .
we show how to relate the precession frequency of the isolated spin to results of current measurements in the steady state ( peak frequency , width of the peak and signal - to - noise ratio ) .
to obtain information on the initial state of the spin , measurements of the current during the initial transient period are needed . here
we do not consider the spin dynamics in this short time interval .
the plan of the article is the following . in the next section we present the model for tunneling via a single spin 1/2 , with and without spin - orbit coupling , and a model describing tunneling via a two - level quantum dot .
then we introduce majorana fermions and the keldysh technique . in the subsequent section
we will compute the spin distribution function , broadening of the spin precession due to the tunneling of electrons , and the spin - spin correlation function .
then we will calculate the dependence of the tunneling current on the spin and determine the current - current correlation function @xmath22 .
finally , we discuss the stm experimental results and compare our results to those obtained in the quasiclassical approach @xcite and other theoretical works @xcite .
for the system consisting of a spin coupled to the leads by an exchange mechanism ( in the non - relativistic approximation , i.e. , neglecting the spin - orbit interaction ) , we use the hamiltonian of the two - leads kondo model @xcite with a direct tunneling term included ( which we call tunneling - via - spin ( tvs ) model ) @xmath34 c^{\dagger}_{\alpha n \sigma}c^{\;}_{\alpha n \sigma ' } , \nonumber \\ { \cal h}_s&=&-g\mu_b { \bf b}_0\cdot{\bf s } , \nonumber \\ { \cal h}_{\sf ref}&=&\!\!\!\ ! \sum_{\alpha , n , n ' \sigma , \sigma'}\!\!\!\!\ !
c^{\dagger}_{\alpha n \sigma}(\hat{t}_{\alpha\alpha})_{\sigma\sigma ' } c^{\;}_{\alpha n ' \sigma ' } , \ \ \
\hat{t}_{\alpha\alpha}=t_{\alpha\alpha}^{({\sf ex})}{\bf s}\cdot\vec{\bm{\sigma}}_{\sigma\sigma ' } , \nonumber \\ { \cal h}_{{\sf tr}}&=&\!\!\!\!\sum_{n , n ' , \sigma,\sigma ' } \!\!\!\ ! c^{\dagger}_{{\rm r } n \sigma } ( \hat{t}_{{\rm rl}})_{\sigma\sigma ' } c^{\;}_{{\rm l } n ' \sigma ' } \ ! + { \rm h.
c. } , \nonumber \\ & & ( \hat{t}_{{\rm rl}})_{\sigma\sigma'}= t_0\delta _ { \sigma\sigma'}+ t_{{\rm rl}}^{({\sf ex})}{\bf s } \cdot \vec{\bm{\sigma}}_{\sigma\sigma ' } , \nonumber\end{aligned}\ ] ] where @xmath35 ( @xmath36 ) creates ( annihilates ) an electron in the left or right lead ( depending on @xmath37 ) in the eigenstate @xmath38 , and with spin @xmath39 .
further , @xmath40 , where @xmath41 is the energy in the state @xmath38 and @xmath42 is the chemical potential in the lead @xmath43 , while @xmath44 represents the three pauli matrices .
@xmath45 , @xmath46 and @xmath47 are tunneling matrix elements due to the exchange interaction for the electron tunneling from the leads to the molecule with the spin 1/2 , while @xmath48 is the direct tunneling matrix element .
we take them as real numbers .
the spin localized in the molecule is described by the operator @xmath49 .
figure [ fig1 ] sketches the physical setup we want to study and which basically represents the model hamiltonian @xmath50 . in the following we use energy units for the bias voltage @xmath1 , i.e.
, we denote @xmath51 by @xmath1 .
also , by @xmath52 we mean @xmath53 , write @xmath2 instead of @xmath54 , and @xmath55 represents @xmath56 .
we describe the leads by a free electron gas with a density of states @xmath57 of bandwidth @xmath58 , so that @xmath41 is the bare energy of the electron in the eigenstate @xmath38 , the same in both leads .
we assume weak tunneling , @xmath59 , where @xmath60 is the density of states per spin ( dos ) of the leads at the fermi level ( when the leads are different @xmath61 where @xmath62 is the dos in the lead @xmath43 ) .
the applied voltage is represented by the difference in chemical potentials , @xmath63 .
we also assume @xmath64 , and @xmath65 ( @xmath66 ) .
we assume in the following that @xmath67 and @xmath68 ; in other words , the lead - electron bandwidth @xmath69 is the largest energy scale .
the leads are supposed to be in a voltage driven thermal equilibrium at temperature @xmath2 .
@xmath70 describes spin - flip scattering of an electron from one lead back into the same lead , and @xmath71 represents both direct and spin - assisted tunneling between leads . in the case of fully polarized leads
, the summation in the hamiltonian is taken only over one type of electron spin . in the model with spin - orbit interaction proposed by levitov and rashba
@xcite the spin precession induces electron density oscillations inside the dot , and thus an electric field inside and around the dot , @xmath72 , where @xmath73 is the orbital momentum , while @xmath74 is a polar vector allowed by the symmetry of the system .
this electric field modulates the tunneling barrier between the leads and the dot .
hence , the tunneling matrix element acquires , in addition to the non - relativistic exchange term , a dependence on the localized spin @xmath75 which is similar for both spin components @xmath39 of the tunneling electrons . to describe such relativistic corrections we need to add to the transfer matrix elements in the hamiltonian , eq .
( [ ham ] ) , the spin - orbit terms .
the compound matrix elements are finally given as @xmath76 here the unit pseudo - vector @xmath77 indicates what projections of the localized spin are modulating the tunneling matrix elements depending upon the geometry of the system .
we can estimate @xmath78 , where @xmath79 is an energy of the order of an atomic energy , while @xmath80 is a small dimensionless parameter which describes the weak overlap of the lead - electron wave - function to that of the molecule with the spin . for spin - orbit tunneling
we have an additional small relativistic multiplicative factor @xmath81 , i.e. , @xmath82 , where @xmath83 is a typical electron velocity .
= 6.0 cm finally , we presume that the spin also couples to the environment due to the spin - phonon coupling , the coupling with nuclear spins , etc . to account for this spin relaxation mechanism
we introduce the relaxation rate @xmath84 . in order to put our model in perspective
let us compare it to other models used in the literature .
for the quantum dot studied in refs .
@xcite the hamiltonian for spinless fermion tunneling is given by @xmath85 \
c_{{\rm l}m}^\dagger c^{\;}_{{\rm r } n } + { \rm h.c .
} , \nonumber\end{aligned}\ ] ] where @xmath86 ( @xmath87 ) creates ( annihilates ) an electron in the state 1 or 2 of the dot . here
@xmath88 is the energy splitting between states 1 and 2 , @xmath89 is the tunneling matrix element connecting them , and tunneling of electrons between the leads depends on the population of the states 1 and 2 .
the difference between this and the tvs model ( with @xmath90=0 ) is that the couplings to the leads only depend upon the difference of fillings of states 1 and 2 or , making correspondence with the spin operators , the hamiltonian @xmath91 is @xmath92 while the coupling , @xmath93 depends only on one spin component , @xmath94 of the probed system , instead of all of them as in the exchange term @xmath95
. moreover , the leads electrons are fully polarized along the @xmath96-axis .
it is assumed that for sufficiently long times the composite ( spin - electrons ) system reaches a dc ( non - equilibrium ) steady state ( which does not depend upon the initial conditions ) .
we use the majorana fermion representation for the spin and the keldysh diagrammatic technique @xcite to describe it .
this technique is nowadays a standard method used in non - equilibrium phenomena and its basic description can be found in standard textbooks such as ref .
its application to kondo - like problems is described in ref .
@xcite . since spin operators are not amenable to the application of the wick s theorem , and hence are not appropriate for diagrammatic techniques , we represent spin-@xmath19 operators @xmath97 in terms of three majorana fermions @xmath98 , where @xmath99 , satisfying the following mapping and relations @xmath100 the hilbert space for majorana fermions is eight - dimensional ( 8@xmath101 ) and the transformation from the spin 1/2 to the majorana fermion representation may be performed in the following way . in general , the spin expectation values that we need to calculate are of the form of a trace , @xmath102 $ ] , over the two - dimensional ( 2@xmath101 ) spin hilbert space .
here @xmath103 is some functional of the spin operators .
we replace the spin operators by majorana fermion operators using eq .
( [ majorana ] ) and we represent the majorana fermion operators as @xmath104 where @xmath105 and @xmath106 are the creation and annihilation operators of a type @xmath107 fermion satisfying the canonical anti - commutation relation @xmath108 .
this representation guarantees that eq .
( [ relation ] ) for the majorana fermion operators are satisfied . a possible orthonormal basis for the 8@xmath101 hilbert space
@xmath109 is @xmath110 , where @xmath111 is the vacuum , @xmath112 , @xmath113 , @xmath114 , @xmath115 , @xmath116 , @xmath117 , and @xmath118 .
@xmath109 may be written as a direct sum of four orthogonal 2@xmath101 subspaces , @xmath119 , which are invariant under the action of the spin-1/2 operators @xmath97 . here ,
basis sets for @xmath120 , @xmath121 , @xmath122 , and @xmath123 are : @xmath124 , @xmath125 , @xmath126 , and @xmath127 , respectively . in other words , the majorana fermion representation of the spin-@xmath19 operator
@xmath128 is an irreducible one in any of the four mentioned subspaces .
furthermore , these representations are equivalent to each other .
therefore , we can write down @xmath129=\frac{1}{4}{\rm tr}[\tilde{f}\{\eta_\mu\}],\ ] ] where the second trace is computed in the extended hilbert space @xmath109 .
( the trace is invariant under a change of basis in @xmath109 ) .
now one can prove wick s theorem in the same way as it was proved for the free fermi and bose fields .
the time - ordered averages in the keldysh technique are taken with the help of the evolution operator @xmath130 along a closed time contour ( its part running from @xmath131 to @xmath132 denoted by @xmath133 and the part from @xmath132 to @xmath131 by @xmath134 ) .
one defines the four ( non - independent ) real - time green s functions @xmath135 which can be compactly written as a matrix @xmath136 . here
@xmath2 and @xmath137 are the time- and anti - time - ordering operators on the @xmath133 and @xmath134 parts of the keldysh contour and @xmath138 represents an arbitrary fermionic field .
they are related through : @xmath139 , and the brackets @xmath140 have the meaning of either a pure state or a distribution of the available phase space of the interacting system .
the four green s functions ( the `` @xmath141 basis '' ) can be expressed in terms of the retarded , advanced and keldysh green s functions ( `` larkin - ovchinnikov ( lo ) basis '' ) @xmath142 \rangle , \label{kel}\end{aligned}\ ] ] by use of the transformation @xmath143 in general , the retarded and advanced functions are hermitian conjugates of each other , i.e. , @xmath144 .
furthermore , in thermal equilibrium the fluctuation - dissipation theorem connects them through @xmath145 , \ ] ] with the thermal distribution function of fermions @xmath146 . as in ref .
@xcite we assume that for sufficiently long times the spin reaches a dc ( non - equilibrium ) steady state which does not depend on the initial conditions .
in such dc steady state the correlation functions depend upon the time difference and the new distribution function @xmath147 differs from @xmath148 . it should be determined as a stationary solution of the quantum kinetic equation .
firstly , we introduce the bare green functions for electrons in the presence of an applied magnetic field ( or internal exchange field in the case of magnetically ordered leads ) @xmath149 which induces a spin polarization of electrons in the directions @xmath150 .
since the leads are assumed to be coupled to a thermal bath , the bare lead - electron green s functions are in a voltage driven thermal equilibrium .
they have the matrix form in the space of electron spin ( @xmath151_{\sigma\sigma ' } = -i{\langle\!\langle{c^{\;}_{\alpha\sigma}(t ) c^{\dagger}_{\alpha\sigma'}(t')}\rangle\!\rangle}$ ] , where @xmath152 ) @xmath153 , \nonumber\\ g_{0 \alpha\kappa}^r(\omega)&=&\int d \epsilon \ \frac{\rho_{\alpha\kappa}(\epsilon)}{\omega - \epsilon + i0^+ } , \nonumber\\ g_{0 \alpha\kappa}^k ( \omega ) & = & 2 \pi i \
h_{0 \alpha}(\omega ) \
\rho_{\alpha\kappa}(\omega ) , \nonumber\\ h_{0\alpha } ( \omega ) & = & h_{\sf eq}(\omega- \mu _ { \alpha})=-\tanh[\frac{\omega-\mu_{\alpha}}{2 t } ] , \nonumber \\ \rho_{\alpha\kappa}(\epsilon)&=&\rho(\epsilon-\kappa b_\alpha-\mu_{\alpha } ) , \nonumber\end{aligned}\ ] ] where @xmath154 and @xmath155 is the bare electron density of states ( per spin ) which we assume to be similar in both leads .
we consider here only local green s functions , so the eigenstate index @xmath38 can be dropped .
note , that for fully polarized electrons ( @xmath156 ) we obtain @xmath157 , while for weakly spin polarized electrons ( @xmath158 ) @xmath159 .
the propagators for the majorana fields can be compactly written as @xmath160 .
the bare propagators satisfy @xmath161 , @xmath162 , @xmath163 , due to the symmetry under rotations about the @xmath96-axis .
a useful property holds for the keldysh components of the spin green s functions : since majorana fermion operators are hermitian , from the properties of the commutator , it turns out that @xmath164 .
in the frequency @xmath55 representation , for time - translation invariant solutions , this means that @xmath165 , and therefore @xmath166 is an odd function of @xmath55 . to write down dyson s equation , it is convenient to use a basis in which the bare propagator is diagonal . defining the canonical fermion operators @xmath167 and @xmath168 , with the associated green s functions @xmath169 and @xmath170 , one has the following expressions for the bare propagators in the @xmath171 basis in the steady state in the lowest order of perturbation theory @xmath172 where @xmath173 is a small regulator ( a width due to an infinitesimally small coupling to a thermal bath ) , and @xmath52 is the total effective magnetic field acting on the spin ( see next section ) . in the thermal equilibrium state @xmath174 .
the perturbative calculations of various physical quantities are very conveniently performed by using the matrix form of the green s functions in the keldysh space .
graphical representations can then be done by feynman diagrams with lines representing matrices of green s functions , where at each vertex of an internal point is assigned an additional @xmath133 or @xmath134 factor ( due to the opposite direction of time integration for the points on the @xmath134 part of the schwinger - keldysh contour ) .
the interaction vertices of majorana fermions with electrons follow from the form of the tunneling hamiltonian ( [ ham ] ) @xmath175 accounting for the relation @xmath176 .
we represent the electron and spin green s functions as well as the vertices in fig .
[ fig2 ] .
the interacting green s function for the spin , @xmath177 , satisfies the dyson s equation @xmath178 for the @xmath179 matrices ( in the tensor product of the @xmath180 space and the keldysh space ) , where the free propagator is given by eqs .
( [ free ] ) .
to lowest order , the imaginary part of the self - energy @xmath181 is of order @xmath182 ( its @xmath55-dependent real part induces a trivial small energy shift that we neglect in the following ) .
calculating @xmath183 and neglecting terms of order @xmath184 we can take the @xmath185 blocks @xmath186 , @xmath187 , @xmath188 , @xmath189 , @xmath190 and @xmath191 as zeroes .
then , @xmath192 has nonvanishing off - diagonal blocks @xmath193 and
@xmath194 , @xmath195 and the diagonal block @xmath196 .
inverting such block matrix @xmath197 , we obtain the matrix @xmath183 with a similar block structure , where the block @xmath193 is given as @xmath198 while for @xmath199 we replace @xmath200 and @xmath201 .
for @xmath202 we replace @xmath196 for @xmath203 and put @xmath204 .
parcollet and hooley @xcite have found that if one starts from the equilibrium distribution functions for majorana fermions , the perturbation theory breaks down when one takes the limit of zero coupling to the leads prior to taking the limit of zero coupling to the thermal bath .
the reason for this non - commutativity of the limits is that in a non - equilibrium steady state the distribution function of the system can deviate from the equilibrium one significantly in the long - time limit even for very weak tunneling .
hence , at long times it can not be calculated perturbatively .
one can , nevertheless , use the perturbation theory built upon the appropriate bare green s functions with correct zeroth - order distribution functions @xmath205 , and @xmath206 which are stable with respect to weak perturbations .
these distributions should be obtained self - consistently , i.e. , @xmath205 and @xmath206 computed using second order perturbation theory have to be the same as the zero order distribution function . hence , assuming that after a long time the system is in a steady state , we can define a distribution function @xmath207 obeying the self - consistency equations @xmath208 , \nonumber\\ h_f(\omega)&= & \frac{\sigma ^k_{f^{\dagger}f}(\omega ) } { \sigma^a _ { f^{\dagger}f}(\omega)-\sigma^r_{f^{\dagger}f}(\omega ) } , \label{hb}\end{aligned}\ ] ] where @xmath209 is calculated in the second order perturbation theory with respect to the electron - spin interaction . @xmath210 and @xmath205 have to be obtained in a similar way .
therefore , one needs to take the zeroth - order green s functions for the spin in the form of eq .
( [ free ] ) , calculate the second order self - energies @xmath211 and @xmath212 , and solve eq .
( [ hb ] ) .
this procedure , as well as taking the limits @xmath213 and @xmath214 , was carefully outlined by parcollet and hooley @xcite .
a useful relation between the distribution functions @xmath207 and @xmath215 can be derived . using the definitions in eq .
( [ kel ] ) and the commutator and anticommutator properties one can show that @xmath216 and using the definition of eq .
( [ hb ] ) one obtains that @xmath217 so that @xmath218 .
the diagrams shown in fig .
[ fig3 ] lead to a renormalization of the effective dc magnetic field acting on the spin when the electrons are polarized .
to first order in the electron - spin interaction ( the first diagram in fig .
[ fig3 ] ) the additional dc magnetic field caused by the tunneling electrons is given by @xmath219 neglecting spin - orbit corrections . here
@xmath220 is the degree of the electron spin polarization in the lead @xmath43 , i.e. , the ratio of the electron magnetization to the optimal magnetization .
this additional magnetic field originates due to the exchange interaction between the spin and the electrons in the leads , and it is voltage independent .
we may estimate this contribution using the expression for the tunneling resistance @xmath221 . for @xmath222 = 100 mohm
we obtain @xmath223 and taking @xmath224 we estimate at @xmath225 ev@xmath226 the additional magnetic field @xmath227 t. when electrons are polarized by the magnetic ( or exchange ) field @xmath228 such that @xmath229 , we estimate @xmath230 . the second diagram in fig .
[ fig3 ] is smaller by a factor @xmath231 and , thus , may be neglected . in the following
we choose the @xmath96-axis along the total field @xmath232 . by @xmath3
we denote the direction of electron spin polarization in this new coordinate system . in the case that @xmath233 , and @xmath234 has the components @xmath235 in the coordinate system with the @xmath96-axis aligned along the bare field @xmath0
, we get that @xmath236 has coordinates @xmath237 in the coordinate system where the @xmath96-axis is aligned along @xmath6 with @xmath238 the diagrams shown in figs .
[ fig4 ] determine the imaginary part of the majorana fermion self - energies which depend upon @xmath2 , @xmath1 , @xmath7 and @xmath6 .
we start by considering only the interaction of the spin with the tunneling electrons , assuming @xmath239 . in this case , the spin is precessing freely between consecutive passings of tunneling electrons which change the phase of precession randomly , preserving time - translation invariance on the average in the steady state . in the quasi - classical approach
@xcite these passings are replaced by an effective classical magnetic field representing white noise . on the other hand , in the quantum description the electrons are treated as tunneling particles with initial energy @xmath240 and final energy @xmath241 .
the tunneling process now has similarities with the scattering of an individual particle by the spin .
the difference is that both the initial and final energy of the tunneling electrons are not fixed , in our model being restricted only by the bands of the leads - electrons .
the diagrams shown in fig .
[ fig4 ] lead to the self - energy @xmath242 g_{f^{\dagger } f}^{+-}(\epsilon)+ \nonumber\\ & & 2{\rm tr}[\sigma^+{\bf g}_{\alpha}^{+-}(u ) \sigma^-{\bf g}_{\beta}^{-+}(v ) ] g_{zz}^{+-}(\epsilon)\ } .\end{aligned}\ ] ] the expression for @xmath243 can be similarly obtained .
note , that @xmath244 depends on @xmath245 via @xmath246 , while @xmath247 depends on @xmath248 .
then , eq . ( [ hb ] ) is a closed equation for @xmath207 .
the other diagrams in fig .
[ fig4 ] give @xmath249 g_{f^{\dagger}f}^{+-}(\epsilon)+ \nonumber\\ & & { \rm tr}[\sigma^+{\bf g}_{\alpha}^{+-}(u ) \sigma^-{\bf g}_{\beta}^{-+}(v ) ] g_{ff^{\dagger}}^{+-}(\epsilon)\}.\end{aligned}\ ] ] the traces for the spin matrices in the expressions for the self - energy can be performed with the help of the relations @xmath250&= & \nonumber \\ & & \hspace*{-3.5cm}-{\bf m}_{\alpha}\cdot{\bf m}_{\beta } \
\delta_{\mu\nu}+m_{\alpha \mu}m_{\beta \nu}+m_{\alpha \nu } m_{\beta \mu } , \nonumber\\ \sum_\nu m_{\alpha \nu}{\rm tr}[\sigma_\mu\sigma_\nu]=2m_{\alpha \mu } , \ & & \!\ ! { \rm tr}[\sigma_x\sigma_y\sigma_z]=2i.\end{aligned}\ ] ] the transformation from the `` @xmath141 basis '' to the `` lo basis '' for the self - energy is the same as for the inverse green s function @xmath251 which implies that @xmath252 and , therefore , @xmath253 and @xmath254 . in this way , for the fully spin - polarized case , @xmath255 , we obtain @xmath256 \nonumber \\ & & \hspace*{-2.3cm}-(1- m_{\alpha z } m_{\beta z } \pm m_{\alpha z } \mp m_{\beta z } ) ( \mu _ { \beta } -\mu _ { \alpha } + \omega ) \}\end{aligned}\ ] ] and @xmath257_{ff^{\dagger},f^{\dagger}f}(\omega)\!\!\!&=&\!\!\
! i\frac{\pi}{2}\rho_0 ^ 2\sum _ { \alpha , \beta}t_{\alpha \beta}^2 \times \\ & & \hspace*{-3.9 cm } \{(1 + 2m_{\alpha z } m_{\beta z } -{\bf m}_{\alpha } \cdot { \bf m}_{\beta } ) \nonumber\\ & & \hspace*{-3.9 cm } [ t\phi(\frac{\mu_{\beta}-\mu _ { \alpha}+\omega \pm b}{t } ) \pm h_{f}(b)(\mu _ { \beta } -\mu _ { \alpha } + \omega \pm b ) ] \nonumber\\ & & \hspace*{-3.9 cm } + ( 1- m_{\alpha z } m_{\beta z } \pm m_{\alpha z } \mp m_{\beta z } ) t\phi ( \frac{\mu _ { \beta } -\mu _ { \alpha } + \omega}{t } ) \ } , \nonumber\end{aligned}\ ] ] with the property @xmath258 assuming that @xmath259 , we calculated the integrals over electron energies , @xmath260 in the limit of large electron bandwidth , @xmath261 , and using the relation @xmath262 ( where @xmath263^{-1}$ ] is the fermi distribution function ) the resulting function @xmath264 is @xmath265 in the weak polarization limit , @xmath266 , there are corrections to these expressions due to terms consisting of integrals of the form @xmath267 to estimate them we assume that near the fermi energy the electron density of states may be approximated as @xmath268
. then , these integrals are proportional to the small parameter @xmath269 and they vanish in the limit @xmath261 . we obtain the expressions for self - energies in the weak polarization limit from the fully polarized results by putting @xmath270 and multiplying the resulting expressions by a factor 4 that accounts for both spin projections of the electron . in the next section
we show that the imaginary part of the self - energy @xmath271 , @xmath272 , describes dephasing of the spin precession caused by the tunneling electrons .
it determines the current noise power as shown later .
therefore , at frequency @xmath273 , eq . ( [ hb ] )
gives us a self - consistent equation for @xmath274 in the case of full spin polarization .
this equation has a unique solution for the distribution function ( @xmath275 , @xmath276 ) @xmath277 provided that we do not have one of the cases @xmath278 . in the latter two cases , the self - consistency equation is an identity and the spin steady state can be any . when @xmath279 we reproduce the results of ref .
@xcite for unpolarized electrons ( see fig . [ fig5 ] ) .
if we account for environment - driven relaxation , @xmath280 , the distribution function will be closer to that in thermal equilibrium .
when such relaxation dominates , the distribution function @xmath207 coincides with the equilibrium one .
we compute first the average spin magnetization @xmath281 ( time - independent in the dc steady state ) , where @xmath282 is the antisymmetric unit tensor .
to lowest order , the values @xmath283 and @xmath284 are proportional to the small parameters @xmath182 and can be neglected . for @xmath285 ,
we get ( to zeroth order ) @xmath286 when @xmath287 ( except @xmath288 ) , we obtain the same result as for unpolarized electrons found by parcollet and hooley @xmath289 here @xmath290 . at @xmath291 , and @xmath292 we get @xmath293 , i.e. , the spin is in the ground state because electrons do not have enough energy to flip the spin .
when @xmath10 the electrons can flip the spin , thus reducing the spin magnetization to @xmath294 , where @xmath295 .
it drops as @xmath296 for large @xmath1 .
as noticed by shnirman and makhlin @xcite the spin - spin correlation function @xmath297 may be expressed via the @xmath298-fermion green s functions .
we rewrite the spin operators in the following fashion @xmath299 the operator @xmath300 commutes with the spin operators @xmath97 and hence with the hamiltonian , while @xmath301 .
( notice that the operator @xmath302 maps the orthogonal subspaces @xmath303 of section [ 3 ] among themselves . ) as a result , @xmath304[f^{\dagger}(0)+f(0 ) ] \rangle \nonumber \\ & = & \frac{i}{4}[g_{ff^{\dagger}}^{-+}(t)+g_{f^{\dagger}f}^{-+}(t)+ g_{f^{\dagger}f^{\dagger}}^{-+}(t)+g_{ff}^{-+}(t)].\nonumber\end{aligned}\ ] ] similarly , @xmath305[f^{\dagger}(0)-f(0)]\rangle \nonumber \\ & = & \frac{i}{4}[g_{ff^{\dagger}}^{-+}(t)+g_{f^{\dagger}f}^{-+}(t)- g_{f^{\dagger}f^{\dagger}}^{-+}(t)-g_{ff}^{-+}(t ) ] .
\nonumber \\ { \cal s}_{zz}(t)&= & \frac{i}{2 } \ g_{zz}^{-+}(t ) .
\label{szz}\end{aligned}\ ] ] and , in general , @xmath306 . to zeroth order in perturbation theory we obtain @xmath307 \label{bare } \\ & = & \frac{\pi}{4 } \{[1-h_f(b)]\delta(\omega -b)+[1+h_f(b)]\delta(\omega + b)\ } , \nonumber\end{aligned}\ ] ] while @xmath308 . to describe the contribution which displays a peak at the larmor frequency in the correlation function @xmath309 we need to account for the imaginary part of the self - energy .
for @xmath310 we have @xmath311 at the weak tunneling condition , @xmath312 , when the relaxation due to the environment is negligible , we can also neglect the contributions from @xmath313 and @xmath314 to @xmath315 with respect to those from @xmath316 and @xmath317 .
we obtain @xmath318 , \label{gamma}\ ] ] where , at @xmath291 , for fully polarized electrons along the @xmath319-axis we get @xmath320 when @xmath292 this gives the width of the precession resonance due to quantum fluctuations of the current ( the current operator does not commute with the hamiltonian , and the current fluctuates between the leads and the spin ) , @xmath321 while for @xmath10 the decoherence rate has an additional contribution due to the voltage induced transport current @xmath322 we estimate @xmath323 , in a junction with @xmath324 and tunneling resistance @xmath325 mohm . to determine @xmath326 in the unpolarized case we need to put @xmath270 and multiply by 4 in eq .
( [ gammasxx ] ) .
thus , we see that the spin - spin correlation functions @xmath327 and @xmath328 exhibit oscillations at a renormalized larmor frequency @xmath5 , with a peak width depending on @xmath1 , @xmath52 and @xmath2 .
when the relaxation due to the environment dominates , @xmath329 , we need to replace @xmath326 by @xmath84 in eq .
( [ gamma ] ) .
the important point is that @xmath326 provides a lower bound for the decoherence rate @xmath20 , i.e. , for the width of the peak at @xmath5 in the spin - spin correlation function .
let us now turn to the evaluation of the average current and current power spectrum .
the current operator is given as @xmath330 .\ ] ] the average current ( involving the complete spin - electrons hamiltonian @xmath331 ) is determined from @xmath332 in the following , we will assume that @xmath333 is turned on adiabatically and we are interested only in the average current up to second order in @xmath333 .
( remember that we are only interested in the dc steady state solution . )
the following derivation is similar to that explained in ref .
@xcite for the quantum dot tunneling . denoting @xmath334
we obtain ( assuming that the leads are not superconducting ) @xmath335\rangle . \nonumber\end{aligned}\ ] ] assuming the dependence @xmath336 near the fermi energy we obtain for fully polarized electrons , @xmath156 , @xmath337 for weakly polarized electrons , @xmath338 , approximating @xmath268 near the fermi energy , we get @xmath339 , while @xmath340 . when @xmath341 the term of order @xmath342 vanishes because @xmath343 is parallel to @xmath344 .
the next spin - dependent term in the average current , @xmath345 , is of order @xmath346 and , in the fully polarized case along the @xmath319-axis , is given by @xmath347\}. \nonumber\end{aligned}\ ] ] here @xmath348 is the causal green s function . using fourier transforms and eq .
( [ sxxo ] ) the result , at @xmath291 , is @xmath349 \ \theta(v - b).\ ] ] hence , when @xmath292 the average current is not affected by the spin because electrons have not enough energy to flip it .
the spin remains in the ground state and thus can not affect tunneling electrons and be probed . in the case that there is no reflection , i.e. , @xmath350 , the spin - dependent contribution is related to the spin - flip rate @xmath351 , @xmath352 because the spin - dependent contribution to the current is determined by the matrix element of the operator @xmath353 between states with opposite spin projections .
hence , each tunneling electron needs to flip the localized spin in order to go through . in the large @xmath1 limit
we obtain @xmath354 in agreement with the quasiclassical results @xcite . at @xmath291
, the contribution @xmath345 leads to a discontinuous derivative of @xmath8 at @xmath355 due to the opening of a new channel corresponding to the tunneling of electrons via the spin .
the change is more pronounced in the second order derivative of @xmath8 with respect to @xmath1 .
accounting for broadening of the spin energy , @xmath326 , we derive @xmath356 we see that this second order derivative of the average tunneling current shows a peak ( see eq .
( [ sxxo ] ) ) at @xmath355 .
hence , measurements of the @xmath357-@xmath1 tunneling characteristics provide information on the spectral density of the spin - spin correlation function . similar behavior for the average current with respect to bias voltage
was used previously for inelastic tunneling spectroscopy of phonons @xcite .
it is based on an equation similar to eq .
( [ peak ] ) relating @xmath358 to the phonon density of states .
note , that in the case of unpolarized electrons @xmath359 is small , but the second derivative of @xmath8 also shows a similar peak @xcite .
however , now its amplitude has an additional small factor @xmath360 .
next we calculate the current - current correlation function . the particular fluctuation operator one needs to evaluate depends upon the nature of the measurement itself @xcite , see also @xcite .
in the general case the result of measurements depends on the correlation functions describing emission , ( + ) , and absorption , ( - ) , by the tunneling contact @xmath361 we obtain @xmath362 , \nonumber\end{aligned}\ ] ] where the term @xmath363 describes the shot noise to second order in the matrix elements @xmath48 and @xmath364 , while @xmath365 represents the contribution which displays a peak at the renormalized larmor frequency of order @xmath366 ( see next subsection ) .
we calculate @xmath363 in the framework of standard time - dependent perturbation theory ( as we did to calculate the average current ) . in the equation above
, @xmath367 means average over electron degrees of freedom , after the perturbation due to the spin - electron interaction is accounted for , while @xmath368 means average over the localized spin degrees of freedom .
averages over spin operators are performed using the known spin correlation functions determined in previous sections .
contributions proportional to @xmath369 and @xmath370 give the standard expression @xcite modified by the presence of the localized spin , in the same way as the average current eq .
( [ avcur ] ) . for fully polarized electrons along the @xmath319-axis
we obtain , at @xmath291 , @xmath371 for unpolarized electrons we need to multiply this result by a factor 4 .
we obtain for the symmetrized shot noise power ( @xmath372 ) @xmath373/2=e(i_0+i_s^{(1)}).\ ] ] the terms proportional to @xmath374 describe the effect of the spin dynamics on the shot noise , @xmath375 . a diagram contributing to @xmath375 is shown in fig . [ fig6 ] . for electrons fully polarized along the @xmath319-direction
we derive for the emission part at @xmath291 and @xmath376 @xmath377 , \nonumber\end{aligned}\ ] ] where @xmath378 emission is possible at frequencies @xmath379 , when the spin is in the ground state , and frequencies @xmath380 , when the spin is in an excited state .
the second order derivative of @xmath381 with respect to @xmath55 , when @xmath10 , has peaks with amplitudes @xmath382 and widths @xmath326 at frequencies @xmath383 .
this provides a way to detect the presence of the localized spin .
the absorption part at @xmath291 and @xmath376 is given as @xmath384 \nonumber \\
+ 2\theta(b - v ) & & \!\!\!\!\!\!\!\!\ ! [ f(\omega+v - b)+f(\omega - v - b ) ] \ \ } , \label{spinnn}\end{aligned}\ ] ] with the result @xmath385/2=e i^{(2)}_{s \perp}\ ] ] for @xmath386 .
figure [ fig7 ] displays the emission and absorption contributions .
= 7.5 cm for unpolarized electrons we derive @xmath387 where @xmath388 is given by eq .
( [ spinn ] ) with the same @xmath389 as for fully polarized electrons .
this noise power also exhibits anomalies at frequencies @xmath383 .
we see that @xmath390 has no peak at the larmor frequency neither for polarized nor for unpolarized electrons contrary to the results of ref .
@xcite .
the fourth order terms , @xmath391 , in @xmath22 correspond to 10 skeleton diagrams some of which are shown in figs .
[ fig8 ] and [ fig9 ] . in these 4-vertex diagrams the external frequency @xmath392 ( via the current vertex ) may enter in any of the two vertices .
diagrams in fig .
[ fig8 ] ( all six ) , as well as the diagram shown in fig .
[ fig6 ] , contain integration over the energy @xmath393 entering in all the electron and spin lines of the diagram .
hence , these contributions change at least on the frequency scale of order @xmath394 as the term @xmath390 .
they are small in comparison to @xmath390 , and may be neglected .
= 8.5 cm to derive this peak we consider the 4 diagrams of the type shown in fig .
[ fig9 ] , which give contributions of order @xmath395 . in these diagrams integration over energy @xmath393 in the spin lines is independent of the integrations in the electron lines and they provide a contribution to the current - current correlation function @xmath396 which has the same peaks at @xmath397 as the spin - spin correlation function . to calculate this contribution we use the keldysh technique as done by shnirman _
we will assume that to study the peak at the renormalized larmor frequency experimentally one can measure the instantaneous values of the current over a long period of time and then obtain the symmetrized current power spectrum defined as @xmath398 .
\label{si}\ ] ] to calculate the first term in eq .
( [ ii ] ) , we put @xmath399 and @xmath400 . for electrons fully polarized along the @xmath319-axis the first term
is given as @xmath401 -t_{\rm rl}^2{\hat a}(t)s_x(t){\hat b}(0)s_x(0)\rangle , \label{3 t } \\ & & { \hat b}(t)=\sum_n c_{{\rm r}nx}^+(t)c_{{\rm l}nx}(t ) , \nonumber \\ & & { \hat a}(t)=\sum_n c_{{\rm r}nx}^+(t ) \sigma _ x c_{{\rm l}nx}(t ) . \nonumber\end{aligned}\ ] ] now , in the framework of the keldysh technique , we expand @xmath402 up to @xmath403 order in the first term , up to @xmath404 in the second and third terms and up to @xmath369 in the last term . let us consider the first term which corresponds to the second diagram in fig .
this diagram leads to the contribution @xmath405 here the product @xmath406 is the sign factor @xmath407 if @xmath408 and @xmath409 otherwise , while @xmath410 \ , \\ & & \lambda^{\mu\nu}(\omega)= [ \lambda_{\rm rl}^{\mu\nu}(\omega)-\lambda_{\rm lr}^{\mu\nu}(\omega ) ] , \end{aligned}\ ] ] where the trace is taken over spin variables .
further , we denote by @xmath411 the combination of majorana fermion green s functions which are represented by broken lines in the diagrams in fig .
[ fig9 ] , @xmath412 with the property @xmath413 adding the term with interchanged r and l and symmetrizing with respect to time we obtain for the first term @xmath414,\end{aligned}\ ] ] similarly we calculate the other three terms @xmath415 , \\ & & { \cal p}_p^{(3)}(\omega)=-\frac{1}{2}(t_0t_{{\rm rl}})^2 \sum_{\mu,\nu}\mu\nu[\lambda^{-\mu}(\omega ) { \cal s}^{\mu-}(\omega)\lambda^{\nu+}(0 ) \nonumber\\ & & + \lambda^{-\mu}(-\omega){\cal s}^{\mu-}(-\omega)\lambda^{\nu+}(0 ) ] , \\ & & { \cal p}_p^{(4)}(\omega)=-\frac{1}{2}(t_0t_{{\rm rl}})^2 \sum_{\mu,\nu}\mu\nu \lambda^{-\mu}(0)[{\cal s}^{-+}(\omega)+ \nonumber \\ & & s^{-+}(-\omega)]\lambda^{\nu+}(0).\end{aligned}\ ] ] let us calculate @xmath411 using the expressions for @xmath416 and @xmath417 in terms of the zeroth - order functions @xmath418 , @xmath419 and @xmath420 @xmath421 + \nonumber \\ & & g_{yy}^k(\epsilon+\omega)[g_{zz}^a(\epsilon)-g_{zz}^r(\epsilon)]\}. \nonumber\end{aligned}\ ] ] since @xmath422 , we have @xmath423 .
therefore , the contribution to @xmath424 from the terms containing @xmath425 is negligible , but the last term does contribute .
we get @xmath426\delta(\omega -b ) + [ 1+h_f ( - \omega ) ] \delta ( \omega + b ) \ } , \nonumber\\ & & { \cal s}^{++}(\omega)=\frac{1}{2}ih_f(\omega ) { \rm p}\frac{\omega}{\omega^2-b^2}+\frac{\pi}{4}\delta^+(\omega ) , \\ & & { \cal s}^{--}(\omega)=-\frac{1}{2}ih_f(\omega){\rm p } \frac{\omega}{\omega^2-b^2}+\frac{\pi}{4}\delta^+(\omega),\end{aligned}\ ] ] where @xmath427 .
next , we take into account the broadening of the resonances at @xmath397 by replacing @xmath428 such procedure follows from the fact that accounting for higher order terms of perturbation theory in @xmath364 we need to replace the product @xmath429 in eq .
( [ rep ] ) by @xmath430 .
this function has the same frequency dependence as @xmath431 because from one hand @xmath432 while from the other this spin - spin correlation function is determined by eq .
( [ tr1 ] ) . at @xmath291
we obtain , neglecting terms of order @xmath433 , @xmath434 , @xmath435 summing up all terms for @xmath396 we obtain zero for the amplitude of the peak at @xmath292 .
this same result was obtained by shnirman _
it means that quantum fluctuations of the spin in the ground state which are present in the spin - spin correlation function can not be probed by tunneling electrons . when @xmath10 we obtain , taking into account the broadening of the resonance , @xmath436 .
\nonumber\end{aligned}\ ] ] more generally , the coefficient 4 is substituted by @xmath437 ^2 $ ] for fully polarized electrons in the leads , while for weakly polarized electrons we get the factor @xmath438 .
accounting also for spin - orbit interaction we obtain an additional term in @xmath439 which is proportional to the small factor @xmath440 $ ] .
we see that @xmath396 exhibits a peak at the renormalized larmor frequency with a width @xmath326 only when the electrons are polarized in a direction which is not parallel to @xmath0 .
this peak is caused by precession of the spin excited by tunneling electrons .
tunneling of electrons excites the spin , resulting in the modulation of the tunneling current , but they also cause decoherence of the spin precession with a rate @xmath326 . therefore , the spin precession can not be seen in the current - current correlation function without spin - orbit interaction when the leads - electron polarization is parallel to @xmath0 , or the electrons are unpolarized , despite the fact that the spin - spin correlation functions for the @xmath353 and @xmath441 spin components exhibit oscillations at @xmath5 .
the reason for this is that to observe oscillations in @xmath22 one needs to probe with the current a component of the spin perpendicular to @xmath6 . for
that the electrons need to have a component of polarization @xmath442 or @xmath443 that couples to @xmath353 or @xmath441 .
spin - orbit interaction leads to a signal when the electrons are unpolarized or @xmath444 if @xmath0 is not parallel to @xmath77 .
however , the signal in that case is much weaker than in the case of fully polarized electrons .
for the symmetrized signal the signal - to - noise ratio reaches its maximum at low temperatures when @xmath445 , while @xmath446 , and both are perpendicular to @xmath6 .
indeed , electrons with components of the spin polarization parallel to @xmath6 do not contribute to the signal but they do enhance @xmath326 . when @xmath447 the relaxation rate @xmath448 ( neglecting relaxation due to the environment ) is determined by eq .
( [ gammasxx ] ) .
the signal - to - noise ratio at the peak position , for the case of perpendicular spin polarization at @xmath291 and @xmath10 , is given by the expression @xmath449 here @xmath450 is the height of the peak at the renormalized larmor frequency in the symmetrized current - current correlation function , while @xmath451 is the current power spectrum for the shot noise neglecting the smaller contribution @xmath452 .
@xmath11 reaches its maximum value @xmath453 for @xmath454 . in the case of symmetrical leads ,
@xmath455 we get @xmath456 , and the function @xmath457 for @xmath458 is shown in fig .
[ fig11 ] .
when the spin relaxation due to the environment becomes dominant , @xmath459 , the signal - to - noise ratio for the case of fully polarized electrons is smaller by the factor @xmath460 .
if the perpendicular components of @xmath3 were absent , we see that @xmath11 becomes as small as @xmath461 , while for weak polarization in the direction perpendicular to @xmath6 it is as small as @xmath269 .
we start summarizing our main results for the model of a localized spin interacting with the tunneling electrons only via the exchange interaction , i.e. , when the tunneling matrix element contains the term @xmath462 in addition to the spin - independent term @xmath48 . *
1*. the larmor frequency is renormalized from the one determined by the external magnetic field @xmath463 to @xmath464 ( see eq .
( [ bren ] ) ) , when tunneling electrons are polarized .
an estimate for @xmath465 is given after eq .
( [ bren ] ) .
* 2*. in the steady state the @xmath357-@xmath1 characteristics , namely @xmath466 , exhibits a peak at @xmath355 when the electrons are polarized in a direction different from @xmath6 .
this peak is due to the opening of a new channel for electron tunneling via the localized spin ( spin - flip tunneling ) .
the width of this peak is determined by the spin decoherence rate @xmath326 , see eq.([gammasxx ] ) .
* 3*. when the electrons are polarized and @xmath10 , on the background of frequency - independent shot noise caused by the discrete nature of the electrons , there is a peak in the current noise power @xmath467/2 $ ] at the renormalized larmor frequency , caused by the precession of the spin excited by the tunneling electrons .
this peak is absent when @xmath292 because the spin is in the ground state since electrons do not have enough energy to flip it .
* 4*. the signal - to - noise ratio for this peak , @xmath468 , strongly depends upon the degree of spin polarization of the leads - electrons and the orientation of this polarization with respect to the external magnetic field @xmath0 acting on the localized spin .
the maximum value of @xmath11 is reached when the electrons are polarized in a direction perpendicular to the effective magnetic field acting on the spin because in this case they probe in an optimal way the precessing spin components .
when the relaxation of the localized spin with the environment is weak , @xmath11 reaches the values of order unity * only * if the leads are almost fully polarized in a direction close to perpendicular to @xmath6 .
* 5*. in the case of full polarization and @xmath469 at low temperatures @xmath470 , the dependence of @xmath11 on @xmath295 is given by eq .
( [ sis ] ) , where @xmath471 comes from the leads - electron reflection via the localized spin ( see eqs . ( [ hbgenp ] ) ) . the maximum value , @xmath472 , is reached when @xmath454 , see figs .
[ fig11 ] and [ fig12 ] .
the signal - to - noise ratio is reduced by the presence of the electron reflection couplings ( @xmath473 ) as expected , since this process does not contribute to the current but does affect the width of the peak .
* 6*. when the spin relaxation due to the environment is negligible , the peak width in the noise power spectrum ( at the renormalized larmor frequency ) depends weakly on the orientation of the polarization .
it increases linearly with @xmath1 , when @xmath474 and the tunneling electrons cause spin flips , see fig .
[ fig10 ] . at high temperatures
@xmath475 , the width of the peak is determined by @xmath2 ( @xmath476 ) .
* 7*. in points 2 through 4 we assumed that the spin relaxation with the environment , @xmath84 , was negligible in comparison to the relaxation due to the leads - electrons , @xmath326 .
this assumption is crucial to obtain a maximum @xmath11 of order unity because in this case @xmath477 , and @xmath478 , so that @xmath11 does not depend on the small parameter @xmath479 , see eq .
( [ sis ] ) .
if the spin relaxation rate due to the environment dominates we get @xmath480 .
* 8*. for weak polarization of the electrons , @xmath11 becomes proportional to the degree of polarization , @xmath481 , where @xmath228 is the magnetic or exchange field acting on the leads - electrons and @xmath58 represents their bandwidth .
this is because only polarized tunneling electrons contribute to the signal at the larmor frequency , while all tunneling electrons contribute to the broadening of this signal .
we estimate @xmath482 for @xmath483 g and @xmath484 ev and this estimate leads to @xmath485 . *
9*. spin - orbit ( i.e. , relativistic @xmath486 ) corrections of any type @xcite can only provide a very small signal - to - noise ratio , @xmath487 , in the situation where the exchange coupling leads to a vanishing signal ( i.e. , the electrons are unpolarized or they are polarized parallel to @xmath0 ) . the signal in this case
is determined by the relativistic coupling between the spin and the tunneling electrons , while the broadening is determined by the dominant non - relativistic exchange coupling .
hence , for a tunneling process via a single localized spin in the steady state , the signal - to - noise ratio becomes of order unity * only * when + a ) the spin is decoupled from the environment , and + b ) the electrons are almost fully polarized in a direction perpendicular to the effective field acting on the localized spin .
* 10*. we derived the correction to the standard shot noise power caused by the presence of the localized spin when the leads - electrons are unpolarized .
this correction is small when @xmath488 , but it changes behavior at @xmath355 , see eqs .
( [ spinn ] ) , and ( [ spinun ] ) . on the other hand ,
the standard shot noise power increases linearly with v. this new result can be observed , in principle , experimentally .
we are now in position to compare our results to the experimental data obtained by manassen et al .
@xcite and durkan and welland @xcite , and decide whether the model of tunneling via a single localized spin can explain these data . in these experiments
the polarization of the electrodes was weak , @xmath11 was of order unity or larger almost independently of the orientation of the applied magnetic field , and the position of the peak in the noise power spectrum @xmath15 was found at the larmor frequency corresponding to the applied magnetic field .
our results show that a combination of weak polarization and a signal - to - noise ratio of order unity are incompatible for the dc steady state described by our model , which accounts for both exchange and spin - orbit coupling of the tunneling electrons and the localized spin . from the point of view of the theory of quantum measurement
we see that after a long time , any measurement usually leads to a steady state where information on the initial state of the quantum system is probably lost .
hence , an important question is what is the transient time for a given measurement .
we note that @xmath358 and the contribution @xmath396 in the current - current correlation function are proportional to the spin - spin correlation function @xmath489 and @xmath490 , respectively ( if the electrons are spin polarized along the @xmath319-axis ) .
suppose that one wants to probe a single spin , say a , which is part of an ensemble of other spins interacting with each other .
measurement of the @xmath357-@xmath1 characteristics , namely @xmath358 , and current noise power , @xmath22 , at low voltages will carry information on the dynamics of that particular ( coupled ) spin a , whenever tunneling occurs via that single spin . effectively , because of the coupling , that measurement provides information on the dynamics of the whole system and eqs .
( [ peak ] ) and ( [ gr ] ) form the background for such tunneling spectroscopy of quantum systems .
note , however , that the subject of study , the spin - spin correlation function , is affected by tunneling measurements ( see the way the bare spin - spin correlation function , eq .
( [ bare ] ) , was modified in the case of a single - spin system , eq .
( [ gamma ] ) ) . we have shown how this modification may be accounted for in the case of steady - state measurements of a single - spin system . in a similar way it may be accounted for in more complicated systems , and corrections to obtain the bare correlation function of the system studied can , in principle , be made .
we note that we presented here a theoretical description of the tunneling spectroscopy of a single localized spin-1/2 system .
however , such a description is also valid for any two - level system ( quantum dot ) , and eq .
( [ transl ] ) describes the mapping between the two - level system and the spin-1/2 system . in conclusion
, we described the characteristics of the tunneling current via an isolated single spin 1/2 ( two - level system ) in the steady state , i.e. , in the long time limit after switching on the voltage or the tunneling matrix elements .
we found optimal conditions when the tunneling current carries maximum information about dynamic quantum fluctuations of the localized spin .
we showed what type of information about the isolated spin may be extracted from measurements of the tunneling current in the steady state .
we express our thanks to d.j .
scalapino , m.n .
kiselev , d. averin , a. andreev , a. shnirman , i. martin , d. mozyrsky , b. spivak , a.m. tsvelik , and a.v .
balatsky for useful discussions .
we acknowledge a. shnirman and yu .
makhlin for sharing their results on the spin - spin correlation function , eq .
( [ tr1 ] ) , prior to publication .
one of us ( m.h . ) is grateful for the hospitality of the los alamos national laboratory .
this work has been supported by the u.s .
bulaevskii and g. ortiz , phys .
lett . , * 90 * , 040401 ( 2003 ) .
o. parcollet and c. hooley , phys .
b , * 66 * , 085315 ( 2002 ) .
a.n . korotkov and d.v .
averin , phys .
b * 64 * , 165310 ( 2001 ) .
r. ruskov and a.n .
korotkov , phys .
b , * 66 * , 041401 ( 2002 ) .
y. manassen , r.j .
hamers , j.e . demuth and a.j .
castellano , jr .
lett . * 62 * , 2531 ( 1989 ) . c. durkan and m.e .
welland , appl .
. lett . * 80 * , 458 ( 2002 ) .
d. shachal and y. manassen , phys .
b * 44 * , 11528 ( 1991 ) ; y. manassen , e. ter - ovanesyan , d. shachal , and s. richter , phys . rev .
b * 48 * , 4887 ( 1993 ) ; a.v .
balatsky and i. martin , cond - mat/0112407 .
levitov and e.i .
rashba , cond - mat/0209507 a.c . funk and m.
beck , am .
* 65 * , 492 ( 1997 ) .
l. glazman and a. kaminsky , private communication .
j. schwinger , j. math .
* 2 * , 407 ( 1961 ) .
keldysh , sov .
jetp * 20*,1018 ( 1965 ) [ zh .
47*,1515 ( 1964 ) ] .
a.m. tsvelik , phys .
lett . * 69 * , 2142 ( 1992 ) .
balatsky , y. manassen , and r. salem , phyl . mag .
b , * 82 * , 1291 ( 2002 ) ; phys . rev .
b , * 66 * , 195416 ( 2002 ) . g.d .
mahan , _ many - particle physics _
( kluwer , new york , 2000 ) ; a.m. zagoskin , _ quantum theory of many - body systems _ ( springer , new york , 1998 ) .
a. shnirman and yu .
makhlin , cond - mat/0305064 .
yanson , sov .
jetp * 39 * , 506 ( 1974 ) ; i.o .
kulik , a.n .
omelyannchuk , and r.i .
shekhter , sov .
j. low temp .
* 3 * , 1543 ( 1977 ) ; b.c .
stipe , m.a .
rezael , and w. ho , science * 280 * , 1732 ( 1998 ) ; j.r .
hahn and w. ho , phys .
lett . * 87 * , 166102 ( 2001 ) ; d. mozyrsky and i. martin , cond - mat/0203521 .
balatsky , ar .
abanov , j .- x .
zhu , cond - mat/0211448 .
g. b. lesovik and r. loosen , jetp lett . *
65 * , 295 ( 1997 ) .
kogan , _ electronic noise and fluctuations in solids _ ( cambridge univ .
press , cambridge , 1996 ) .
d. rogovin and d.j .
scalapino , annals of phys . *
86 * , 1 ( 1974 ) .
a. shnirman , i. martin and d. mozyrsky , cond - mat/0211618 . | we consider the problem of tunneling between two leads via a localized spin 1/2 or any other microscopic system ( e.g. , a quantum dot ) which can be modeled by a two - level hamiltonian .
we assume that a constant magnetic field @xmath0 acts on the spin , that electrons in the leads are in a voltage driven thermal equilibrium and that the tunneling electrons are coupled to the spin through exchange and spin - orbit interactions . using the non - equilibrium keldysh formalism
we find the dependence of the spin - spin and current - current correlation functions on the applied voltage between leads @xmath1 , temperature @xmath2 , @xmath0 , and on the degree and orientation @xmath3 of spin polarization of the electrons in the right ( @xmath4r ) and left ( @xmath4l ) leads .
we show that a ) the spin - spin correlation function exhibits a peak at the larmor frequency , @xmath5 , corresponding to the effective magnetic field @xmath6 acting upon the spin as determined by @xmath0 and the exchange field induced by tunneling of spin polarized electrons , b ) if the @xmath7 s are not parallel to @xmath6 the second order derivative of the average tunneling current @xmath8 with respect to @xmath1 is proportional to the spectral density of the spin - spin correlation function , i.e. , exhibits a peak at the voltage @xmath9 , c ) in the same situation when @xmath10 the current - current correlation function exhibits a peak at the same frequency , d ) the signal - to - noise ( shot noise ) ratio @xmath11 for this peak reaches a maximum value of order unity , @xmath12 , at large @xmath1 when the spin is decoupled from the environment and the electrons in both leads are fully polarized in the direction perpendicular to @xmath6 , and e ) @xmath13 if the electrons are weakly polarized , or if they are polarized in a direction close to @xmath0 , or if the spin interacts with the environment stronger than with the tunneling electrons .
our results of a full quantum - mechanical treatment of the tunneling - via - spin model when @xmath14 are in agreement with those previously obtained in the quasi - classical approach .
we discuss also the experimental results observed using stm dynamic probes of the localized spin . |
the sloan digital sky survey ( sdss ) is a digital photometric and spectroscopic survey which will cover 10,000 deg@xmath4 of the celestial sphere in the north galactic cap and produce a smaller ( @xmath5 225 deg@xmath4 ) but much deeper survey in the southern galactic hemisphere ( york _ et al .
_ 2000 and references therein ) .
the survey sky coverage will result in photometric measurements for about 50 million stars and a similar number of galaxies .
the flux densities of detected objects are measured almost simultaneously in five bands ( @xmath6 , @xmath7 , @xmath8 , @xmath9 , and @xmath10 ; @xcite ) with effective wavelengths of 3551 , 4686 , 6166 , 7480 , and 8932 , 95% complete0.1 arcsec . ] for point sources to limiting magnitudes of 22.0 , 22.2 , 22.2 , 21.3 , and 20.5 in the north galactic cap and @xmath11 because the absolute calibration of the sdss photometric system ( dependent on a network of standard stars ) is still uncertain at the @xmath12 level .
the sdss filters themselves are referred to as @xmath13 and @xmath10 .
all magnitudes are given on the ab@xmath14 system ( oke & gunn 1983 , for additional discussion regarding the sdss photometric system see @xcite , and stoughton _ et al .
_ 2002 ) . ] .
astrometric positions are accurate to about 0.1 arcsec per coordinate ( rms ) for sources brighter than 20.5@xmath15 ( pier _ et al .
_ 2001 ) , and the morphological information from the images allows robust star - galaxy separation to @xmath5 21.5@xmath15 ( lupton _ et al .
_ 2002 ) .
sdss , although primarily designed for observations of extragalactic objects , will significantly contribute to studies of the solar system objects , because asteroids in the imaging survey must be explicitly detected to avoid contamination of the samples of extragalactic objects selected for spectroscopy .
( 2001 , hereafter i01 ) analyzed sdss commissioning data and showed that sdss will increase the number of asteroids with accurate five - color photometry by more than two orders of magnitude ( to about 100,000 ) , and to a limit about seven magnitudes fainter than previous multi - color surveys ( e.g. the eight color asteroid survey by zellner , tholen & tedesco 1985 ) .
the main results derived from these early sdss observations are 1 . a measurement of the main - belt asteroid size distribution to a significantly smaller size limit ( 1 km ) than possible before .
the size distribution resembles a broken power - law , independent of the heliocentric distance : @xmath16 for 0.4 km @xmath17 5 km , and @xmath18 for 5 km @xmath17 40 km .
2 . a smaller number of asteroids compared to previous work . in particular ,
the number of asteroids with diameters larger than 1 km is about @xmath19 , or up to three times less then suggested by earlier studies .
the distribution of main - belt asteroids in 4-dimensional sdss color space is strongly bimodal , and the two groups can be associated with s ( rocky ) and c ( carbonaceous ) type asteroids .
a strong bimodality is also seen in the heliocentric distribution of asteroids : the inner belt is dominated by s type asteroids centered at @xmath20 2.8 au , while c type asteroids , centered at @xmath20 3.2 au , dominate the outer belt .
i01 demonstrated that the sdss photometric pipeline ( _ photo _ , lupton _ et al .
_ 2002 ) is a robust and highly efficient automated tool for finding moving objects . i01
present a detailed discussion of the completeness and reliability of the sdss asteroid catalog based on the known sdss photometric and astrometric precision , but this discussion is incomplete because it does not attempt a direct comparison with catalog of known asteroids on an object - by - object basis .
the importance of such an analysis is evident from the realization that there are 90,000 _ cataloged _ asteroids with @xmath21 , while i01 expect to find only 58,000 in the same absolute magnitude range .
the implied low completeness of 64% ( compare to 98% estimated by i01 ) can only be verified by direct matching of known and sdss asteroids .
another motivation for cross - correlating sdss asteroids and known asteroids is a large potential increase in the number of asteroids with both accurate multi - color photometry and known orbital parameters ( the sdss data themselves are insufficient for accurate orbit determination , but will provide serendipitous photometric measurements for a substantial fraction of known asteroids ) .
additionally , the sdss color information may be utilized to study the chemical segregation in the full orbital parameter space , rather than only as a function of heliocentric distance as done by i01 .
this paper presents the first results on cross - correlating sdss asteroids and cataloged asteroids .
section 2 describes the analyzed data , a software pipeline used for generating the positions of known asteroids at the time of sdss observations , and their positional matching to objects automatically detected by the sdss photometric pipeline .
section 3 discusses the statistics of matched objects , and compares the sdss photometric measurements with the apparent magnitudes predicted from cataloged absolute magnitudes . in section 4 we discuss the chemical segregation of asteroids in orbital element space , and in section 5 we summarize the main results .
we use asteroid data ( c.f .
i01 ) from the sdss early data release , described in detail by stoughton _
_ 2002 ( hereafter sdssedr ) .
these data include equatorial observing runs 94 , 125 , 752 and 756 ( see i01 and sdssedr ) ; the boundaries are given by @xmath22 , and @xmath23 = 351@xmath24 56@xmath24 ( runs 94 and 125 ) , or @xmath23
= 145@xmath24 250@xmath24 ( runs 752 and 756 ) . for a footprint in ecliptic coordinates , see figure 1 in i01 .
this region has an area of 432 deg@xmath4 and includes 12,668 moving objects selected as in i01 ( the velocity @xmath25 deg / day and @xmath26 ) .
we use this sample for all quantitative estimates of the catalog completeness , and for photometric comparison . in order to increase the number of objects
when studying the color segregation in the asteroid belt ( section [ colseg ] ) , we also add additional objects from seven currently unreleased equatorial observing runs that roughly double the matched sample size .
the sdss photometric data include a list of objects flagged as moving by the photometric pipeline . for each object ,
the position and time of observation , and the magnitudes and associated errors in five sdss photometric bands ( @xmath27 ) are recorded .
the sdss imaging data is obtained in the time - delay - and - integrate ( tdi , or drift - scanning " ) mode , and thus each observed position corresponds to a different observing time ( as opposed to the starring mode where all objects from a given image are observed at the same time ) . in a general case , the correlation between a position and time is easiest to compute in the great circle coordinate system . since all the scans discussed here were obtained along the celestial equator , this dependence becomes particularly simple , and is given by @xmath28 where @xmath29 is the time corresponding to position @xmath23 ( e.g. julian day ) , and the zeropoints @xmath30 and @xmath31 are constants for a given run ( all reported positions correspond to the @xmath8 band ) .
for an arbitrary run , this expression is easily generalized by using an appropriate coordinate system .
the cataloged asteroid magnitudes are reported in the johnson @xmath32 band .
preliminary transformations from the sdss photometric system to the johnson bands are given by fukugita _
( 1996 ) and krisciunas , margon & szkody ( 1998 ) .
we use a recent updated version of these transformations ( m. fukugita , priv .
comm . ; the new transformations produce v magnitudes that agree to better than 0.1 mag with the version from fukugita _ et al .
_ 1996 ) to synthesize the johnson b and v band magnitudes @xmath33 for typical values of for asteroids ( 0.40.8 ) , @xmath34 is in the range 0.16 to 0.33 .
the overall accuracy of these transformations and sdss photometry is better than 0.05 mag , as determined by direct comparison with non - sdss observations obtained in the johnson system ( m. fukugita , e. grebel , j. holtzman , unpublished ) . for positional matching of the moving objects observed by sdss with cataloged asteroids we select the asteroid orbital elements database ( astorb ) .
astorb is a catalog of high - precision osculating orbital elements and other information on all numbered and a large number of unnumbered asteroids ( but does not contain information on known numbered comets ) .
the astorb catalog is distributed by the lowell observatory in the form of an ascii file , containing single line records for each asteroid ( bowell 2001 ) .
the catalog is updated daily for addition of newly discovered objects , deletion of duplicate objects , and improvement of parameters of known objects . in this work
we use the six osculating orbital elements , absolute magnitude and the phase - correction slope parameter which are necessary to predict the asteroid position and apparent magnitude at the time of sdss scans .
we also utilize the arc ( in days ) spanned by the observations used to compute the orbit to estimate the accuracy of orbital elements .
orbital elements given in astorb are heliocentric and have been computed by a variable - timestep differential orbit correction algorithm , based on astrometric observations obtained from the minor planet center .
the perturbations from all the major planets , the moon and the three largest asteroids ( ceres , pallas and vesta ) have been taken into account .
absolute magnitude corresponding to the johnson @xmath32 band ( @xmath35 , see i01 ) is listed as numbers rounded to two , one , or no decimal places , with the number of decimal places reflecting assumed reliability .
however , for unnumbered asteroids @xmath35 is given to two decimal places regardless of its real accuracy .
the slope parameter @xmath36 ( see section [ apmags ] below ) is given for asteroids for which it is known , and for others , which are the overwhelming majority of the sample , assumed to be 0.15 ( dimensionless ) .
we note that besides astorb there are several other asteroid orbital elements databases ( e.g. minor planet center orbit database , asteroid dynamics site ) .
however , because of its widespread use and significant additional information supplied for each asteroid , we have chosen to use astorb as a referent catalog for this study .
the astorb version used here is from september 18 , 2001 , and contains 141,110 objects ( 29,074 of which are numbered asteroids ) .
the asteroid identification pipeline consists of three parts : 1 .
propagation of the asteroid osculating orbital elements from astorb to the epoch of the sdss observation 2 .
computation of the asteroid positions and apparent magnitudes at the time of sdss scan 3 .
positional matching of sdss moving objects with the known asteroids the astorb catalog contains osculating orbital elements computed for epoch near the current epoch , where `` current '' corresponds to the publishing date of the astorb catalog file .
the orbital elements are propagated to the epoch of observation using the proele routine of orbfit v1.8 _ propag _ library ( @xcite ) .
we use the default dynamical model supplied with the orbfit package .
it includes gravitational perturbations of the sun , moon and the major planets and relativistic corrections , when necessary .
the positions and masses for major planets are derived from jpl ephemeris de405 ( standish 1998 ) .
although the orbfit package offers multiple choices for the numerical integration schemes , we use its automated selection procedures , which has been proven to be sufficiently accurate in practice ( @xcite ) .
the end result of the orbit propagation is a catalog of osculating orbital elements at the time of sdss observation , which are then used for quick two - body ( sun and asteroid ) computations of the ephemeris .
the calculations of positions are performed using a modified preobs routine from the same orbfit package ( the modification consists of enabling a two - body approximation ) .
when calculating the positions of cataloged asteroids , they must satisfy the condition given by eq .
[ tdrift ] .
we solve the problem iteratively , until the difference between the positions calculated in two successive iterations is smaller than 0.001 arcsec . in practice
the computations converge rapidly and usually reach the required accuracy with two to three iterations .
the apparent magnitudes are computed by orbfit s appmag subroutine from the cataloged values of the absolute magnitude @xmath35 and the phase - correction slope parameter @xmath36 .
the phase - corrected apparent visual `` catalog '' magnitude , @xmath37 , is obtained from ( bowell 1999 ) @xmath38 where @xmath20 is the heliocentric , and @xmath39 is the geocentric distance ( expressed in au ) , @xmath40 is the `` solar '' angle ( the angle between the sun and the earth as viewed from the asteroid , see i01 ) , and @xmath41 and @xmath42 are the `` phase - correction '' functions .
the latter are obtained from standard approximations @xmath43 ; \,\ , i = 1 , 2 \\ a_1 = 3.33\ \,\ a_2 = 1.87 \non b_1 = 0.63\ \,\ b_2 = 1.22 \non\end{aligned}\ ] ] note that these approximations are formally different from the one used by i01 ( eq .
however , they are similar numerically ; the difference is less than 0.05 mag for @xmath44 and @xmath45 , with the maximum value of 0.15 mag at @xmath46 .
the matching algorithm compares the list of predicted positions for the known asteroids with the list of sdss objects flagged as moving by the photometric pipeline . after finding the nearest - neighbor pairs from the two lists
, it requires that the distance between the two positions is less than 30 arcsec .
the final sample is very insensitive to the precise value of this cut because for the majority ( 86% ) of matched objects the distance between the two positions is less than 3 arcsec ( centered on zero in both coordinates ) .
the probability that a randomly chosen position would fall within 3 arcsec from a moving sdss object is of the order 10@xmath47 in the regions with the highest asteroid number density ( 64 deg@xmath48 , i01 ) , implying that only about one of the matches within 3 arcsec is a random association .
the associated objects are taken to be positive positional identifications of an sdss moving object with a cataloged asteroid .
the sdss photometric pipeline identified 12,668 moving objects selected as described by i01 .
there are 2801 cataloged asteroids whose computed positions are within the boundaries of the sdss scans considered here , and 1633 have within 30 arcsec an object flagged by sdss photometric pipeline as moving . of these ,
only 1325 are unique objects because some asteroids were observed by sdss more than once ( i01 ) .
detailed statistics for each run are listed in table 1 .
this comparison , taken at face value , implies that the sdss sample is 58% complete .
this is a much lower completeness than claimed by i01 ( 98% ) .
we have visually inspected sdss images around the positions of all 1168 astorb objects not found in the sdss catalog ( moving objects are easily recognized on 1x1 arcmin g - r - i color composites thanks to their peculiar appearances , see i01 ) .
there are 219 objects which are clearly moving , but were not recognized as such by the photometric pipeline .
the most common reason for missing them are complex environments ( bright stars with diffraction spikes , compact galaxy clusters , meteor and airplane trails , etc . ) which are hard to `` deblend '' into individual sources .
the revised , true completeness of the sdss sample is thus 88% , somewhat lower than 98% determined by i01 .
we note that i01 determined completeness using only data from run 756 . for this run
the completeness determined here is 91% .
furthemore , for all runs except for run 94 the completeness is higher than 90% . the overall completeness is below 90% only for run 94 ( 81% ) , which is one of the earliest sdss commissioning runs , obtained while the telescope still did not have proper optics . for the remaining 949 ( 34% of the total ) astorb objects there is no visible sdss source within 30 arcsec from the predicted position down to the sensitivity limit of 22.5 .
the most plausible explanation for these `` missing '' sources is inaccurate orbital elements ( that is , their true position during sdss observations is further than 30 arcsec from the predicted position ) .
indeed , the inspection of their entries in the astorb catalog indicates that the majority are either faint ( @xmath49 ) , have a small observational arc used to determine the orbital elements , or both .
sdss observations demonstrate that a large fraction of known asteroids listed in the astorb catalog do not have sufficiently accurate orbital elements for reliable identification .
the observational arc used to determine the orbital elements can be utilized as a rough estimate of their quality the longer the arc spanned by the observations used to compute the elements , the more accurate are the positions .
the astorb catalog also provides a more quantitative estimate for the accuracy of the orbital elements the _ current ephemeris uncertainty ( @xmath50 ) _ parameter ( 1-@xmath51 absolute positional uncertainty ) , for an epoch near the catalog publishing date .
however , the @xmath50 can not be propagated to the time of observation without the knowledge of covariance matrices for the orbital elements solution ( cf .
@xcite ) , and we use arc to make a quality cut .
the relationship between the @xmath50 and the arc is shown in figure [ fig_arc ] .
the top panel shows each object from the astorb catalog as a dot , and the bottom panel shows the arc histograms for all objects by a full line , and separately for objects with @xmath52 30 arcsec as a shaded region . as evident , the requirement that arc @xmath53 300 days successfully eliminates most of the asteroids with large ceus , and selects 1634 objects ( out of 2801 ) .
note that the resulting sample is insensitive to the precise value of this cutoff because of the non - continuous distribution of data .
a further constraint is based on the predicted apparent magnitude .
we only select asteroids that are reliably detectable by sdss ; they satisfy @xmath54 ( i01 ) .
this cut has only a minor importance and selects 1612 objects ( relaxing the limit to 22.5 adds 3 objects ) .
the absolute magnitudes for selected ( hereafter `` reliable '' ) and removed asteroids ( hereafter `` unreliable '' ) are compared in figure [ fig_hlog ] .
the solid line shows the absolute magnitude distribution for all 2801 asteroids , and the dashed and dotted lines show the distributions for the 1612 ( 58% ) reliable and 1189 ( 42% ) unreliable asteroids , respectively .
it is evident that the asteroids with unreliable orbital parameters are mostly faint , and dominate the sample for @xmath55 .
the subsample of reliable asteroids appears complete for @xmath56 .
the removal of the sources with unreliable orbital elements is efficient , but not perfect .
the matching statistics for the subsample of astorb asteroids with reliable orbits is listed in table 2 . out of 1612 objects ,
1335 are matched within 30 arcsec to objects automatically recognized by the photometric pipeline as moving . of the remaining 277 objects ,
173 are visually identified as moving , and the remaining 104 objects do not have an sdss moving source within 30 arcsec . for each of these 104 objects ,
the predicted position has been independently verified using the on - line astdys orbital calculator . in summary ,
the arc @xmath53 300 days cut decreases the fraction of astorb catalog asteroids with unreliable orbital elements from 34% to 6% .
the sdss completeness based on the reduced sample is 89% ( = 1335/1508 ) , consistent with the estimate based on the full sample ( 88% ) .
since the removed objects are predominantly faint , this agreement indicates that the sdss completeness is not strongly dependent on apparent magnitude .
the sample of 1335 matched objects that were detected by the photometric pipeline , together with the sample of 173 visually confirmed objects from the astorb catalog , can be used to determine whether the completeness of the sdss automatically detected sample depends on magnitude .
the top panel in figure [ fig_hlog_det ] compares the absolute magnitude distribution , as listed in the astorb catalog , for the 1335 matched objects ( dashed line ) to the distribution of all 1612 objects ( full line ) .
the bottom panel shows the fraction of matched sources in each magnitude bin , shown by full line with error bars .
the horizontal line is added to guide the eye and represents the overall completeness of 89% . as evident
, there is no significant correlation between the fraction of matched sources and the magnitude .
we have also tested for correlations between the fraction of matched sources and the phase angle , the object s color and apparent magnitude and did not find any significant dependences .
the top panel in figure [ fig_vlog ] compares the apparent magnitude distribution for asteroids from the astorb catalog that are within the boundaries of sdss scans discussed here ( full line ) , with the apparent magnitude distributions for moving objects detected by sdss .
the distribution for all sdss objects is shown by the dashed line and the distribution for matched objects by the dotted line .
we use the calculated @xmath37 magnitudes for the astorb asteroids and synthesized @xmath57 magnitudes for sdss asteroids ( including the matched sample ; this is why the dotted line is above the solid line for @xmath58 ) . at first sight
it appears that the sdss sample is significantly incomplete .
the bottom panel shows the number ratio of astorb and sdss asteroids in each magnitude bin , and indicates that the sdss completeness is as low as 60% for @xmath59 .
however , as demonstrated in the previous section ( see figure [ fig_hlog_det ] ) , the sdss catalog _ includes _ 89% of astorb asteroids , and the mismatch in figure [ fig_vlog ] is simply due to an offset in apparent magnitude scales .
the existence of such an offset is further corroborated by the direct comparison of the calculated ( @xmath37 , based on the astorb catalog ) and synthesized ( @xmath57 , based on sdss observations ) magnitudes for the 1335 matched objects .
the top panel in figure [ fig_mdiff_hist ] shows the histogram of the difference @xmath60 for all 1335 objects by the full line .
the median offset is 0.41 mag ( the sdss values are fainter ) , with the root - mean - scatter of 0.35 mag .
the dotted and dashed lines show the @xmath60 histograms for the 400 blue asteroids and the 935 red asteroids selected by their color as described in i01 .
it is evident that the apparent magnitude offset depends on the asteroid color ; the median offsets are 0.34 and 0.44 for the blue and red subsample , respectively .
this dependence is further illustrated in the bottom panel , where the @xmath60 difference is plotted as a function of the asteroid color .
the @xmath60 histograms shown in figure [ fig_mdiff_hist ] are not symmetric and give a hint of two components : one centered at 0.1 is independent of color , and the other one centered at 0.4 - 0.5 that appears to be color - dependent .
this bimodality is more pronounced for @xmath61 , as discernible in figure [ fig_mdiff ] .
it is of obvious interest to find out whether the offset between predicted and observed magnitudes is particular to the astorb catalog , or perhaps also present in other available databases . for this test
we chose the minor planet center orbit database from september 5 , 2001 .
it contains similar data to the astorb catalog for 115,797 objects ; 115,583 of these objects are common to both catalogs .
the top panel in figure [ m.mdiff_cats ] compares the listed absolute magnitudes for the objects in common . as evident ,
the catalogs contain two `` populations '' of object .
for about half of the sample the difference between the two values is less than 0.1 mag , while for the remaining 57,719 objects it is larger than 0.1 .
the median value for the latter is 0.4 .
we did not find any correlation between the difference in absolute magnitude and other cataloged quantities .
in particular , there is no correlation between this difference and the mean absolute magnitude .
given only two catalogs , it is not possible to tell which one is responsible for the offset in magnitude scale .
fortunately , the catalogs can be compared to sdss observations of the matched objects for which the photometric errors are not larger than 0.05 mag .
the bottom panel in figure [ m.mdiff_cats ] compares the @xmath60 histograms obtained with the data from the astorb catalog ( full line ) and from the mpc catalog ( dashed line ) . as evident , the median discrepancy between the cataloged magnitudes and the magnitudes measured by sdss is about twice as large for the astorb catalog as for the mpc catalog ( 0.4 mag vs. 0.2 mag ) . based on this finding alone , it would seem that the mpc catalog should be used for the matching purposes .
however , the mpc catalog contains fewer objects than the astorb catalog ; out of the 1633 matched objects only 1387 are found in the mpc catalog ( implying an upper limit on the completeness of the mpc catalog of 85% ) .
the discrepancy in the absolute magnitude scale significantly affects the number of objects brighter than a given limit . assuming that the number counts follow a @xmath62 relation ( i01 ) ,
the mpc catalog will imply 1.32 more objects than sdss measurements , and the astorb catalog will imply 1.74 more objects than sdss measurements .
although the uncertainty in the values of @xmath57 synthesized from the sdss photometry is probably smaller than 0.05 mag , it is prudent to verify the offset in magnitude scales using non - sdss observations .
krisciunas , margon and szkody ( 1998 ) obtained photometric measurements in sdss - like bands for 15 bright asteroids .
table 3 lists their measurements , and the synthesized , @xmath57 , and predicted , @xmath37 , magnitudes .
the @xmath60 difference is shown as a function of @xmath57 in figure [ fig_krisc ] .
asteroids with @xmath35 in the astorb catalog given to two decimal places are marked as circles , and those with one decimal place are marked by triangles .
the former are predominantly at the bright end where they scatter around zero offset , indicating that the photometric transformations from sdss to johnson v band are not biased .
the latter are found at the faint end and indicate a similar offset in the magnitude scales as discussed above .
the correlation between the magnitude offset and the asteroid color indicates that the discrepancy is probably due to differences in the standard johnson v magnitude and the `` visual '' magnitude listed in the astorb catalog .
a most likely cause is the fact that the majority of current asteroid surveys ( which produced most of the entries in the catalogs ) are obtained with open , unfiltered ccds since they are mostly concentrated on astrometric measurements .
thus , the measured magnitudes are not equivalent to the standard johnson v band , and furthermore , the discrepancy depends on the asteroid color .
indeed , as discussed by jedicke , larsen & spahr ( 2002 ) , the reporting of observed asteroid magnitudes is notoriously inconsistent .
for example , the minor planet center regularly weights magnitudes reported by different groups according to their historical photometric accuracy and transforms them to the @xmath32 band , while the astorb database uses the unweighted photometry in observed bands in determining the absolute magnitude . correcting the predicted magnitudes for a median offset of 0.41 mag shifts the @xmath37 histogram for astorb asteroids shown in figure [ fig_vlog ] by the right amount to make the matched fraction consistent with the estimate shown in figure [ fig_hlog_det ] .
we show the shifted histogram in figure [ fig_vlog_mod ] .
as discussed in section [ sdssdata ] , we joined the 1335 matched objects from the sdss early data release with 1306 matched objects from unreleased runs , to form the final sample of 2641 unique matched objects .
this is the sample used in our analysis of the color segregation in the asteroid belt .
asteroid dynamical families are groupings of asteroids in proper orbital elements space , widely thought to be the results of collisional disruptions of parent bodies ( larger asteroids ) .
pioneered by hirayama ( 1918 ) , the number and understanding of asteroid families has increased dramatically thanks to a number of factors a large increase in the number of discovered asteroids , more advanced knowledge of celestial mechanics and more sophisticated cluster - discovery techniques . for a detailed review see chapman _ et al .
_ 1989 . in a recent study , zappala _ et al .
( 1995 ) identified 32 families and 31 clumps " using a hierarchical clustering method in a sample of 12,487 asteroids .
however , even a rough taxonomic classification exists only for a small number of these objects .
the knowledge of the chemical composition of each family could provide further clues about the physical properties of the parent body .
sdss photometric data can remedy this problem because the asteroids segregate in sdss multi - dimensional color space . in this section
we analyze the correlation between sdss colors and the distribution in orbital parameter space for 2641 matched asteroids . in order to study the correlation between sdss colors and the various taxonomic systems ,
we have cross - referenced the 1335 matched asteroids from the sdss edr sample with a catalog of asteroids with known taxonomy obtained at the planetary data system small bodies node .
the catalog contains information on 1199 objects , with tholen ( 978 objects ) , barucci ( 438 objects ) , tedesco ( 357 objects ) , howell ( 112 objects ) and xu ( 221 objects ) taxonomies .
there are only seven asteroids with taxonomical information present in the sdss sample , and are listed in table 4 .
this small number is not surprising given that the sdss saturation limit ( @xmath63 ) is usually below the faint limit for most taxonomy surveys .
the seven asteroids have sdss colors in agreement with their taxonomic types ( i01 ) .
the 2641 known asteroids observed by sdss have been cross - referenced with the catalog of proper orbital elements produced by milani _
( 1999 ) and distributed through astdys .
the resulting sample includes 1768 asteroids with available proper elements .
hereafter we separate this sample into 531 blue asteroids ( @xmath64 , see i01 ) , and 1237 red asteroids ( @xmath65 ) . as discussed by i01 , there is a possibility that red asteroids with @xmath66 form a separate group in color - color space , and we treat separately the 131 asteroids with such colors . in the remainder of this section , we analyze the differences in the distributions of these subsamples in the three - dimensional space spanned by the semi - major axis , @xmath67 , inclination , @xmath68 , and eccentricity , @xmath69 .
figure [ fig_a_i ] displays the distribution of asteroids in the @xmath67 vs. @xmath70 plane .
the top panel shows 67,917 asteroids from the astorb database which have known proper orbital elements ( from milani _ et al .
_ 1999 ) , marked as dots .
the major dynamical families ( also known as the hirayama families ) eos , koronis and themis , are clearly visible at @xmath71 au ; their approximate ( @xmath72 ) positions are ( 3.0 , 0.18 ) , ( 2.9 , 0.03 ) and ( 3.15,0.02 ) , respectively .
the 4:1 , 3:1 , 5:2 and 2:1 mean motion resonances with jupiter at 2.065 , 2.501 , 2.825 and 3.278 au
( the latter three correspond to the kirkwood gaps ) are also evident .
the @xmath73 resonance is visible as a strong cuttof of asteroid density at high inclinations in the @xmath74 region . due to using a sample which is about five times larger , the various families are more easily discerned than in zappala _
( 1995 ) ; however , the distribution shown here agrees with their results .
the remaining two panels show the same distribution as isodensity contours , with the matched asteroids shows by open circles .
the middle panel shows blue asteroids ( @xmath75 0 ) , and the bottom panel shows red asteroids ( @xmath53 0 ) .
a subset of red asteroids with @xmath76 is shown by crosses ; most are found around @xmath67 2.22.5 and @xmath70 0.12 .
figures [ fig_a_e ] and [ fig_i_e ] are analogous to figure [ fig_a_i ] , and show the asteroid distributions in the @xmath69 vs. @xmath67 and @xmath69 vs. @xmath70 planes .
figure [ fig_i_e ] shows the distributions in three belt regions selected by the positions of resonances . for studying the chemical composition of various families
, we adopt the family definitions by zappala _
the distribution of blue and red asteroids in the @xmath67-@xmath70-@xmath69 space confirms the earlier results for the three largest families .
koronis and eos family seem to be mostly composed of red asteroids : 40 out of 45 ( 89% ) asteroids associated with the koronis family are red , and 77 of 98 ( 79% ) are red for the eos family .
on the other hand , the themis family is predominantly blue : 41 out of 43 ( 95% ) members have @xmath77 .
all three major families are located in the outer belt . from other families in the outer belt
we find two members of the brasilia family , both are blue .
the middle portion of the belt , between the 3:1 and 5:2 resonances ( 2.5 @xmath78 2.8 ) contains a number of smaller families .
the sdss sample discussed here provides good color information for five of them .
the maria family is found to be red with 16 out of 21 asteroids with @xmath79 .
the more numerous eunomia family also displays red characteristics ( 97 red and 7 blue members ) . on the other hand ,
the adeona family and the dora family are blue ( 12 out of 15 , and all 11 members are blue , respectively ) . the ceres family seems to be a mix of both asteroid types , with 11 blue and 24 red asteroids .
other smaller families have fewer identified members ; we find marginal evidence that the hoffmeister , rafita and hestia families are predominantly red , while the misa and taiyuan families are predominantly blue .
the inner part of the belt is dominated by the vesta / flora complex and at lower inclinations by the nysa / polana and massalia families .
the nysa / polana group is found to include a mix of red and blue asteroids , with red asteroids dominating the sample in approximately 3:1 ratio ( 51 red and 19 blue ) .
this is the strongest mixing of blue and red asteroids found for the families studied here , and supports the claim by cellino _
( 2001 ) that the nysa / polana group includes two independent families , one composed of s ( red ) asteroids and the other including low - albedo f - like ( presumably blue ) objects .
we note that for many families we find a substantial fraction ( @xmath80 10% ) of the minor component ( e.g. 20% of the red eos family members are blue ) .
it is not easy to determine whether this mixing is due to the background contamination ( i.e. by asteroids that are not family members ) , or due to non - homogeneous structure of the parent bodies .
a robust analysis necessarily involves precise definitions of families , and a careful study of the multi - dimensional color distribution for each family .
such an analysis is beyond the scope of this work and will be presented in a future publication .
i01 analyzed the distribution of 316 asteroids in sdss color space by producing synthetic sdss colors from the spectral measurements obtained by the smass project ( xu _ et al .
in addition to the two major color types , there was an indication that red asteroids with extremely blue colors may form a separate class ( see figure 10 in i01 ) .
this notion was also supported by the taxonomic classification of asteroids with such colors ; they were all classified as the j and v asteroids associated with the vesta family ( binzel & xu 1995 ) .
the proper orbital elements for 131 objects with @xmath76 discussed here can be used to test whether these objects cluster in the @xmath67-@xmath70-@xmath69 space .
the vesta family occupies a very small volume in the orbital parameter space centered at @xmath81 , @xmath82 , and @xmath83 .
if the asteroids with extreme colors can be interpreted as j and v taxonomic types , then they should be concentrated in that region , rather than scattered throughout the belt as blue and red asteroids are .
the 131 red asteroids with @xmath76 are marked in figures [ fig_a_i ] , [ fig_a_e ] and [ fig_i_e ] as crosses ; it is evident that they are clustered around the vesta family .
this is better seen in figure [ m.vestoids ] where we show only the small region of the orbital parameter space that is relevant for the vesta family ( 2.1 @xmath84 2.5 and 0.09 @xmath850.15 ) .
the red asteroids are shown as circles and the subset of 99 objects with @xmath76 as crosses .
the large dot marks the position of vesta .
the box outlined by the solid line is the core region ( 2.28 @xmath84 2.4 and 0.1 @xmath850.13 ) , and the box outlined by the dashed line is the tail region ( 2.4 @xmath84 2.49 ) and 0.1 @xmath850.13 ) , as proposed by zappala _
( 1995 ) . since a very high fraction of objects with @xmath76 ( 99 out of 131 in the displayed region , and 50 and 29 , respectively , in the outlined core and tail regions )
is found in the region of the orbital parameter space that is associated with the vesta family , we conclude that the color provides a reasonably reliable method for selecting objects from the vesta family .
this work demonstrates the feasibility of significantly increasing the number of asteroids with both accurate multi - color photometry and known orbital elements by matching sdss - detected asteroids with the catalogs of known asteroids .
the sample discussed here includes 2600 asteroids ; when sdss is completed the final catalog could be more than ten times larger .
such a large sample can be used for a compositional analysis of the asteroid families at a level of detail which was not previously possible .
we show here that sdss color information is sufficient to reveal strong color segregation among the asteroid families : most are dominated by either blue or red objects .
we showed that the sdss photometric pipeline automatically flags 90% of observed moving objects , with a contamination level of only a few percent .
this is a better performance than for any other similar code .
the directly determined completeness is lower than the 98% claimed by i01 , though without any significant consequences for their results .
it is somewhat surprising that only 66% of the known asteroids listed in the astorb catalog have sufficiently accurate orbital elements to predict their positions to better than 30 arcsec .
however , most objects with unreliable orbital elements can be efficiently removed .
it is much more perplexing that there is a large offset in magnitude scales , not only with respect to sdss , but also among the standard asteroid databases .
this offset results in an overestimate of the number of asteroids brighter than a given magnitude limit by factor 1.7 , and resolves the discrepancy between the sdss asteroid count normalization and the number of asteroids listed in astorb catalog .
we thank robert jedicke for fruitful discussions regarding the discrepancy in asteroid counts between sdss and astorb database , and dejan vinkovi for helping with the visual inspection of sdss images .
we thank princeton university for generous financial support of this research .
the sloan digital sky survey ( sdss ) is a joint project of the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the max - planck - institute for astronomy , the max - planck - institute for astrophysics , new mexico state university , princeton university , the united states naval observatory , and the university of washington .
apache point observatory , site of the sdss telescopes , is operated by the astrophysical research consortium ( arc ) .
funding for the project has been provided by the alfred p. sloan foundation , the sdss member institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , monbusho , and the max planck society .
the sdss web site is http://www.sdss.org/. jedicke , r. , larsen , j. & spahr , t. 2002 , _ observational selection effects in asteroid surveys and estimates of asteroid population sizes _ , proceedings of the asteroids 2001 `` from piazzi to the 3@xmath86 millennium '' conference , june 1116 , 2001 , palermo , italy , university of arizona press , in press .
crrrrr [ tblmatching ] 94 & 663 & 2757 & 395 & 94 & 174 + 125 & 648 & 2844 & 426 & 48 & 174 + 752 & 712 & 3422 & 373 & 33 & 306 + 756 & 778 & 3645 & 439 & 44 & 295 + all & 2801 & 12668 & 1633 & 219 & 949 + unique & 2170 & & 1325 & & + crrrrrr [ tblmatchingconstrained ] 94 & 391 & 2757 & 290 & 82 & 19 + 125 & 402 & 2844 & 350 & 32 & 20 + 752 & 374 & 3422 & 318 & 25 & 31 + 756 & 445 & 3645 & 377 & 34 & 34 + all & 1612 & 12668 & 1335 & 173 & 104 + unique & 1239 & & 960 & & + rlcccccccr [ tbl_krisc ] 446 & aeternitas & a & 30 sep & 11:09 & 14.33 & 0.61 & 14.58 & 13.43 & 1.14 702 & alauda & c & 30 sep & 11:47 & 12.75 & 0.45 & 12.93 & 12.90 & 0.03 82 & alkmene & s & 28 sep & 10:57 & 12.43 & 0.64 & 12.69 & 12.72 & -0.02 774 & armor & s & 28 sep & 03:52 & 12.97 & 0.71 & 13.26 & 12.94 & 0.32 371 & bohemia & as & 4 oct & 05:05 & 12.64 & 0.65 & 12.90 & 12.83 & 0.07 349 & dembowska & r & 29 sep & 11:54 & 10.28 & 0.70 & 10.57 & 10.57 & -0.00 433 & eros@xmath87 & s & 30 sep & 12:26 & 13.30 & 0.67 & 13.57 & 13.09 & 0.48 480 & hansa & s & 29 sep & 09:59 & 12.25 & 0.63 & 12.51 & 12.61 & -0.10 10 & hygeia & c & 28 sep & 12:05 & 11.43 & 0.46 & 11.61 & 11.54 & 0.06 683 & lanzia & c & 1 oct & 03:35 & 14.23 & 0.48 & 14.42 & 13.78 & 0.64 68 & leto & s & 27 sep & 10:37 & 10.11 & 0.66 & 10.38 & 10.53 & -0.15 149 & medusa & s & 29 sep & 06:12 & 12.57 & 0.68 & 12.85 & 12.96 & -0.11 196 & philomela & s & 28 sep & 08:51 & 11.41 & 0.64 & 11.67 & 11.56 & 0.10 314 & rosalia & c & 5 oct & 03:15 & 13.95 & 0.46 & 14.13 & 13.81 & 0.32 138 & tolosa & s & 5 oct & 04:11 & 11.57 & 0.69 & 11.85 & 11.92 & -0.07 rrrr [ tbl_taxonomy ] ( 220 ) stephania & xc ( tholen ) & -0.062 & -0.012 + ( 1628 ) strobel & x ( smass ) & 0.024 & 0.079 + ( 1679 ) nevanlinna & x ( smass ) & -0.048 & 0.036 + ( 1711 ) sandrine & s ( tholen ) & 0.074 & 0.042 + ( 1772 ) gagarin & s ( smass ) & 0.135 & 0.003 + ( 2215 ) sichuan & s ( smass ) & 0.097 & 0.011 + ( 5118 ) elnapoul & s ( smass ) & 0.151 & -0.022 + | we positionally correlate asteroids from existing catalogs with a sample of @xmath018,000 asteroids detected by the sloan digital sky survey ( sdss , ivezi _ et al .
_ 2001 ) .
we find 2641 unique matches , which represent the largest sample of asteroids with both accurate multi - color photometry and known orbital parameters .
the matched objects are predominantly bright , and demonstrate that the sdss photometric pipeline recovers 90% of the known asteroids in the observed region . for the recovered asteroids
we find a large offset ( 0.4 mag ) between johnson v magnitudes derived from sdss photometry and the predicted catalog - based visual magnitudes .
this offset varies with the asteroid color from 0.34 mag for blue asteroids to 0.44 mag for red asteroids , and is probably caused by the use of unfiltered ccd observations in the majority of recent asteroid surveys .
this systematic photometric error leads to an overestimate of the number of asteroids brighter than a given absolute magnitude limit by a factor of 1.7 .
the distribution of the matched asteroids in orbital parameter space indicates strong color segregation .
we confirm that some families are dominated by a single asteroid type ( e.g. the koronis family by red asteroids and the themis family by blue asteroids ) , while others appear to be a mixture of blue and red objects ( e.g. the nysa / polana family ) .
asteroids with the bluest colors , which can be associated with the vesta family , show particularly striking localization in orbital parameter space .
# 1@xmath1 # 1@xmath2 # 1 @xmath3 # 1 # 1_#1 # 1/#2 u i # 1 # 1 |
it was recently shown that a non - equilibrated energy transfer between kinetic and configuration degrees of freedom lies at the origin of a breakdown of the classical stokes - einstein relation in suspensions of particles at arbitrary concentrations @xcite .
this breakdown is quantified in terms of an effective temperature , @xmath0 , that depends on the local confining forces that a tracer particle experiences during its motion through the fluid @xcite . when a constant external force , @xmath1 , is applied on the particles ,
the coupling between configuration and kinetic degrees of freedom is stronger and , consequently , the corresponding effective temperature contains the effect of this external force , @xmath2 , @xcite .
a similar effect has been recently reported in simulations on active non - linear microrheology of supercooled liquids in refs .
@xcite in which the correspondence between simulation results and theory is remarkable , see , for instance , ref .
it is convenient to notice that , although the methodology adopted in ref.@xcite , where the authors use the potential energy landscape concept , differs notably from that of refs .
@xcite , in which correlation functions of dynamic variables are analyzed , the same effect associated with thermal fluctuations and the effective temperature appears .
we will back to this point below .
if @xmath3 is the probability density in @xmath4space , with @xmath5 and @xmath6 the position and velocity of a driven particle , then the heart of the statistical description of this problem is that for systems in or near an arrested state , like concentrated particle suspensions and supercooled liquids , the assumption of adiabatic decoupling @xmath7 _ is no longer valid _ , and therefore we have to use the exact definition @xmath8 which means that the different time scales of the dynamics are not well separated .
this simple fact has huge implications on the scaling of the effective temperature and mobility of glasses and glass like systems . in this work ,
we discuss these implications . in doing this , we provide a theoretical and conceptual framework that serves to understand , on the basis of a non - equilibrium description based on fokker - planck and smoluchowski equations , the origin of the already mentioned effective temperature and mobility , as well as to formulate a generalized version of the fluctuation - dissipation theorem in terms of the stokes - einstein relation , valid for far from equilibrium systems under the action of an external force .
we illustrate our results by considering an ensemble of tracer particles in non - linear microrheology in atomic glasses . the article is organized as follows . in section
ii we explain the physics behind the non - adiabatic decoupling of the fokker - planck description .
then , insection iii we show the origin of the force dependent temperature whereas section iv is devoted to show the origin of the effective mobility within the mean field description .
finally , section v discusses the generalization of the stokes - einstein relation and the summary and conclusions are presented in section vi .
in order to proceed , it is convenient to introduce the well - known kramers - fokker - planck equation @xmath9 which describes the dynamics of an ensemble of non - interacting brownian degrees of freedom , that is , the velocity @xmath6 and the position @xmath5 of the driven particles in contact with a heat bath . in eq .
( [ fokker - planck ] ) @xmath10 stands for the specific energy landscape , @xmath11 the mass of the driven particles , @xmath12 is the temperature of the heat bath and @xmath13 the boltzmann s constant .
it is convenient to stress that @xmath14 is known as the damping coefficient ( friction per unit of mass ) of a single particle which may differ , in the general case , from the effective friction coefficient @xmath15 of the ensemble , with @xmath16 the mobility .
our aim is to calculate the explicit expression of @xmath17 .
usually , this is done starting from the constrained description in terms of the reduced density @xmath18.@xcite note here that , the normalization of the probability is expressed by @xmath19 the configurational probability density evolves according to@xmath20 . % \ ] ] equation ( [ configurational probability ] ) has been obtained by partial integration of eq .
( [ fokker - planck ] ) over velocity and thus , implicitly defines the current @xmath21 . for times
@xmath22 , the constitutive relation for @xmath23 can be obtained by taking the time derivative of @xmath24 , using eq .
( [ fokker - planck ] ) and integrating by parts assuming that the currents vanish at the boundaries .
thus , after neglecting the time derivative @xmath25 , the resulting expression is @xmath26 - \beta^{-1 } \rho \nabla v.\ ] ] large external forces cause local inhomogeneities in the system in such a way that there is no clear time separation between the relaxation of the fast and the slow ( configuration ) variables , a situation which corresponds to the underdamped motion of the driven particles . in these conditions , @xmath27 is not given by a maxwellian and thus , according to the equipartition theorem , the second moment of @xmath27 yields a nonequilibrium temperature @xmath28 @xmath29 since it is not possible to know a _
priori _ the explicit dependencies of @xmath27 , it is convenient to make an overall estimate of this effective temperature . defining @xmath30 and using eq .
( [ fokker - planck ] ) , we obtain the modified rayleigh equation @xmath31 + \frac{\partial}{\partial u}\left [ \phi \langle\nabla v\rangle\right ] , \ ] ] where we have introduced the volume averaged specific force : @xmath32 . the rayleigh equation ( [ rayleigh ] ) is valid up to zero order in the expansion @xmath33 where @xmath34 is a characteristic length of the system .
( [ rayleigh ] ) describes the relaxation of the spatially - averaged distribution @xmath35 and , as a consequence of this , it includes , over average , the effect of the external forces on the dynamics of the fast variable through the _ time dependent quantity _ @xmath36 .
the solution of eq . ( [ rayleigh ] ) can be found by using the method of characteristics .
thus , in the absence of diffusion , the velocity @xmath6 is determined by the initial velocity @xmath37 through the relation @xmath38 a change of variables from @xmath6 to @xmath39 transforms eq .
( [ rayleigh ] ) into @xmath40 a second change , @xmath41 @xmath42 leads to@xmath43 whose solution is a gaussian@xcite that , written in the original variables , reads @xmath44 @xmath45 ^{2}}{\frac{2k_{b}t_{0}}{m}\left ( 1-e^{-2\beta t}\right ) } \right\}\ ] ] at long time the dependence of @xmath35 on @xmath37 disappears , and if @xmath36 is a slow varying function of is argument such that@xmath46 then the long time _ quasi - stationary _ limit of eq .
( [ gaussian ] ) becomes@xmath47 ^{2}}{\frac{2k_{b}t_{0}% } { m}}\right\}.\ ] ] the second moment of eq .
( [ gaussian3 ] ) is then given by @xmath48^{2}\equiv \frac{k_{b}t_{eff}(t)}{m}.\ ] ] this equation is the spatial average of the nonequilibrium temperature @xmath28 introduced in eq .
( [ nonequilibrium temperature ] ) : @xmath49 .
it is worth emphasizing that eq .
( [ second moment ] ) contains the damping coefficient @xmath14 in a form that weighs the corrective effect that external forces induce on the thermal energy that particles have at their disposal in order to perform brownian motion .
this extra energy comes from the non - equilibrated energy distribution associated to velocities.@xcite thus , eq .
( [ second moment ] ) can be interpreted as an equipartition theorem defining the thermal energy of the system that is proportional to the non - equilibrium mean - field effective temperature @xmath0 .
when the specific potential energy is of the form @xmath50 , due to the application of an external force @xmath1 , eq .
( [ second moment ] ) reduces to @xmath51 in the literature reporting simulations on nonlinear - active microrheology , different strategies were used to introduce the effective temperature similar to eq .
( [ meanfield temperature ] ) .
for instance in refs .
@xcite , the authors inferred its existence by analyzing the behavior of the mean square displacement and the diffusion coefficient in terms of the externally applied force .
more recently , in ref .
@xcite the potential energy landscape approach was used in order to calculate the population distribution of different metabasin energies in the presence of the external force .
it was shown that the energy distribution of the system maintains a gaussian form but becomes shifted towards higher energies , fact that yield the authors to characterize the effect in terms of an effective temperature whose dependence on the force coincides with eq .
( [ meanfield temperature ] ) .
adopting now a mean - field description by using @xmath2 in the first term at the right hand side of eq .
( [ constituttive ] ) yields @xmath52 , % \ ] ] which , after substituted into eq .
( [ configurational probability ] ) , yields the _ mean - field smoluchowski equation _ @xmath53\right\ } , \ ] ] where we have introduced the mean - field ( self ) diffusion coefficient @xmath54 where we have introduced the bare mobility coefficient @xmath55 that depends on the damping coefficient @xmath14 .
the calculation of the effective mobility coefficient can now be done as follows .
first , we define the local velocity @xmath56 as @xmath57 such that @xmath58 and according to eq .
( [ diffusion current ] ) , see refs .
@xcite , @xmath59}dx , % \ ] ] where the characteristic length of a cage @xmath34 , coincides with the periodicity of the potential energy @xmath60 . in first approximation
, we may substitute the stationary solution @xmath61 of eq .
( [ smoluchowski ] ) in order to obtain the following expression for the effective mobility @xmath62 : @xmath63 the factor @xmath64 in eq .
( [ effective mobility ] ) is defined by @xmath65 and has dimensions of length . in the limit @xmath66
, it has been shown that @xmath67 .
equation ( [ effective mobility ] ) shows that the effective mobility @xmath16 is a non - linear function of the external force @xmath1 , no matter if the mean field temperature @xmath0 is or not a function of @xmath1 . assuming that the periodic potential energy is of the form @xmath68+\frac{8}{10}\sin^2\left[\frac{2\pi x}{l}\right]\right)%\ ] ] and approximating the arguments of the exponentials in eq .
( [ a ] ) by their first order series expansion in terms of @xmath69 , it is possible to compare our results with the simulation data for the dependence of the mobility coefficient on the externally applied force @xcite , see fig .
other potentials may be used to make the interpolation of simulation data ( not shown ) .
the set of parameters change but the fit is equally good .
we assumed the potential ( [ periodic potential ] ) , shown in fig .
1a , because it has a combination of large and small energy barriers .
it is convenient to mention that the potential energy @xmath70 is an effective potential , a mean force potential , whose parameters may depend upon the temperature @xcite . for the interpolation of the simulation data in in fig .
1a we have assumed @xmath71 and @xmath72 in all cases .
for the blue circles ( data ) and the blue solid line we have used @xmath73 , @xmath74 and @xmath75 ; for the orange squares and the orange dashed line we have used @xmath76 , @xmath77 and @xmath78 ; for the green diamonds and the green dotted - dashed line we have used @xmath79 , @xmath80 and @xmath81 ; for the red upward triangles and the red solid line we have used @xmath82 , @xmath83 and @xmath84 , for the purple downward triangles and the purple dashed line we have used @xmath85 , @xmath86 and @xmath87 and , finally , for the purple downward triangles and the purple dashed line we have used @xmath88 , @xmath89 and @xmath87 .
units are arbitrary .
an interesting question regarding these results is related with the fluctuation - dissipation theorem . in ref .
@xcite , it was shown for colloidal suspensions at arbitrary volume fractions that the stokes - einstein relation has to be modified by considering the natural scaling of all the quantities involved , the effective diffusion and viscosity coefficients and the effective temperature , in order to cope with experimental data .
following this idea , we propose that for molecular glasses a generalized stokes - einstein relation similar to eq .
( 3 ) of ref .
@xcite may hold , that is : @xmath90 where we have introduced the long - time collective diffusion coefficient @xmath91 .
fig 1b shows the scaling of this quantity in terms of the effective temperature @xmath0 .
this result means that the effective temperature is the natural scaling variable of the system , in similar form as the effective volume fraction is the natural scaling variable of concentrated colloidal suspensions @xcite .
other calculations showing correction of the diffusion coefficient when brownian particles move along periodic and tilted potentials are reported in ref .
as a summary , in this work we have shown the distinction between the damping coefficient and the effective non - linear mobility of driven particles in active micro - rheology of supercooled liquids .
in addition , we have shown that the origins of the effective temperature and of the effective mobility also differ .
the effective temperature is a consequence of the dynamic coupling between kinetic and configurational degrees of freedom .
the effective mobility arises as a collective effect that gives insight into the energy landscape of the system . from the interplay between these quantities it was possible to formulate a generalized stokes - einstein relation that may be used to determine the collective diffusion coefficient in non - linear microrheology of atomic glasses .
comparison of our theoretical results with simulation data is remarkably good .
we believe that our findings may be of utility in understanding the thermo - dynamics of glassy systems .
this work was supported by the dgicyt of spanish government under grant no .
fis2015 - 67837-p and by unam dgapa grant no . in113415 .
we also thank the academic mobility program between the university of barcelona and the national autonomous university of mexico . | the distinction between the damping coefficient and the effective non - linear mobility of driven particles in active micro - rheology of supercooled liquids is explained in terms of individual and collective dynamics .
the effective mobility arises as a collective effect which gives insight into the energy landscape of the system . on the other hand ,
the damping coefficient is a constant that modulates the effect of external forces over the thermal energy which particles have at their disposition to perform brownian motion . for long times , these thermal fluctuations become characterized in terms of an effective temperature that is a consequence of the dynamic coupling between kinetic and configurational degrees of freedom induced by the presence of the strong external force .
the interplay between collective mobility and effective temperature allows to formulate a generalized stokes - einstein relation that may be used to determine the collective diffusion coefficient .
the explicit relations we deduce reproduce simulation data remarkably well . |
half - metallic ferromagnets ( hmf ) @xcite attract now a growing attention of researchers , first of all , because of their importance for `` spintronics '' , or spin - dependent electronics @xcite .
hmf have metallic electronic structure for one spin projection ( majority- or minority - spin states ) , but for the opposite spin direction the fermi level lies in the energy gap @xcite .
therefore the corresponding contributions to electronic transport properties have different orders of magnitude , which can result in a huge magnetoresistance for heterostructures containing hmf @xcite .
a discovery of the half - metallic ferromagnetism in colossal magnetoresistance ( cmr ) materials like la@xmath6sr@xmath7mno@xmath8 @xcite has increased considerably the interest in this topic .
transport properties of hmf are the subject of numerous experimental investigations ( see , e.g. , recent works for cro@xmath9 @xcite and nimnsb @xcite , and the reviews @xcite ) . at the same time
, the theoretical interpretation of these results is still a problem . as for electronic scattering mechanisms , the most important difference between hmf and `` standard '' itinerant electron ferromagnets like iron or nickel
is the absence of one - magnon scattering processes in the former case @xcite .
two - magnon scattering processes have been considered many years ago for both the broad - band case ( weak @xmath10 exchange interaction ) @xcite and narrow - band case ( `` double exchange model '' ) @xcite .
the obtained temperature dependence of the resistivity have the form @xmath2 and @xmath11 , respectively . at low enough temperatures
the first result fails and should be replaced by @xmath0 as well @xcite ; the reason is the compensation of the transverse and longitudinal contributions in the long - wavelength limit which is a consequence of the rotational symmetry of the @xmath10 exchange hamiltonian @xcite . up to
now there are no results which describe in the whole temperature region the resistivity of hfm and especially its magnetic - field dependence which is most interesting from the experimental point of view .
such expressions are obtained in sect.2 . apart from three - dimensional case studied before @xcite we consider also two - dimensional hmf keeping in mind , e.g. , layered cmr compounds like lasr@xmath9mn@xmath9o@xmath12 ( for a review see ref.@xcite ) which are almost half - metallic according to the recent band - structure calculations @xcite .
owing to peculiar band structure of hfm , an important role belongs to incoherent ( non - quasiparticle ) states which occur near the fermi level because of correlation effects @xcite . in sect.3
we treat the corresponding contributions to resistivity and discuss tunneling phenomena in hmf .
to investigate effects of interaction of current carriers with local moments we use the hamiltonian of the @xmath13 exchange model @xmath14 where @xmath15 , @xmath16 and @xmath17 are operators for conduction electrons and localized spins in the quasimomentum representation , the electron spectrum @xmath18 is referred to the fermi level @xmath19 , @xmath20 is the @xmath13 exchange parameter which will be put for simplicity @xmath21-independent , @xmath22 are the pauli matrices . in the spin - wave region
we have @xmath23 the zero - order hamiltonian includes non - interacting electrons and magnons , @xmath24 with @xmath25 being the spin splitting which is included in @xmath26 , @xmath27 are the holstein - primakoff boson operators . in the half - metallic case the spin - flip processes do not work in the second order in @xmath28 since the states with one spin projection only are present at the fermi level . at the same time , we have to consider the renormalization of the longitudinal processes in higher orders in @xmath28 ( formally , we have to include the terms up to the second order in the quasiclassical small parameter @xmath29 ) . to this end
we eliminate from the hamiltonian the terms which are linear in the magnon operators by using the canonical transformation @xcite , @xmath30 with @xmath31 then we obtain the effective hamiltonian @xmath32 here @xmath33 is the @xmath10 scattering amplitude which vanishes at @xmath34 and thereby takes properly into account the rotational symmetry of electron - magnon interaction .
more general interpolation expression for the effective amplitude which does not assume the smallness of @xmath35 or @xmath29 was obtained in ref.@xcite by a variational approach ; it does not differ qualitatively from simple expression ( [ amp ] ) . in the case of a considerably @xmath21-dependent exchange parameter , which may be relevant for real itinerant magnets including hfm , one has in ( [ hef ] ) @xmath36 the most general and rigorous method for calculating the transport relaxation time is the use of the kubo formula for the conductivity @xmath37
@xcite @xmath38 where @xmath39 @xmath40 @xmath41 is the current operator , @xmath42 is the electron velocity . representing the total hamiltonian in the form @xmath43 , the correlator in ( [ kru ] )
may be expanded in the perturbation @xmath44 @xcite . in the second order
we obtain for the electrical resistivity @xmath45[{\cal h}^{^{\prime } } , j_x]\rangle\ ] ] where @xmath46 is calculated with the hamiltonian @xmath26 .
provided that the perturbation hamiltonian has the form @xmath47 we obtain @xmath48 \label{rho}\ ] ] with @xmath49 this approach is equivalent to the solution of the boltzmann transport equation by the variational method @xcite . in the hfm situation
the band states with one spin projection only , @xmath50 are present at the fermi level @xcite . below we consider the case @xmath51 @xmath52 and omit the spin indices in the electron spectrum .
we find from eq.([rho ] ) the following expression for the transport relaxation time @xmath53 defined by @xmath54@xmath55 where @xmath56 and @xmath57 are the bose and fermi functions .
a similar expression has been derived first in ref.@xcite , but with the replacement of the effective amplitude just by @xmath28 .
after some transformations we obtain @xmath58 averaging over the angles of the vector @xmath21 leads to the result @xmath59 with @xmath60 where @xmath61 for @xmath62 and @xmath63}^2}{({\bf p - q)}^2q_0 ^ 2}\,\,\,(p , q\ll q_0).\ ] ] the wavevector @xmath64 determines the boundary of the region where @xmath65-dependence of the amplitude become important , so that @xmath66 at @xmath67
. in the case @xmath68 the simple perturbation theory fails and we have to take into account the spin splitting by careful collecting the terms of higher orders in @xmath28 . in the simple one - band model of hmf
where @xmath69 one has @xmath70
( @xmath71 is the conduction bandwidth , lattice constant is put to unity ) @xcite . generally speaking
, @xmath64 may be sufficiently small provided that the energy gap is much smaller than @xmath71 , which is the case for real hmf systems .
a `` crossover '' wavector may exist in principle even for the narrow - band case ( where , instead of spin splitting , the spin subbands have different widths ) provided that the fermi level is close to the gap edge for the spin projection @xmath72 .
the quantity @xmath64 determines a characteristic temperature and energy scale @xmath73 where @xmath74 is the spin - wave stiffness defined by @xmath75 , @xmath76 is the curie temperature .
note that in the case of an usual ferromagnetic metal the scale for existence of one - magnon processes is smaller , @xmath77 . when estimating temperature dependences of resistivity one has to bear in mind that each power of @xmath78 or @xmath79 yields @xmath80 at very low temperatures @xmath1 small quasimomenta @xmath81 yield main contribution to the integrals . averaging the quantity ( [ lam ] ) over the angle between the vectors @xmath82 and @xmath65 we derive @xmath83 with @xmath84
. then we obtain for the resistivity @xmath85 such a dependence was obtained in the narrow - band case ( double - exchange model with large @xmath35 ) , where the scale @xmath4 is absent @xcite , and by the diagram approach in the broad - band case @xcite . at the same time , at @xmath86 the function @xmath87 in eq .
( [ lam ] ) can be replaced by unity to obtain @xmath88 this result is in agreement with the old works @xcite .
now we treat the two - dimensional ( @xmath89 ) situation which may be appropriate for layered manganites like la@xmath90ca@xmath91mn@xmath9o@xmath12 @xcite . at low temperatures we obtain @xmath92 at the same time , for @xmath86
we obtain after replacing the scattering amplitude by unity a logarithmically divergent integral which should be cut at @xmath4 .
thus we get @xmath93 to calculate the magnetoresistivity we introduce the gap in the magnon spectrum , @xmath94 provided that the external magnetic field @xmath95 is large in comparison with the anisotropy gap , @xmath96 is proportional to @xmath95 . in the @xmath97 case the resistivity at @xmath98 is linear in magnetic field , @xmath99 the situation at @xmath86 is more interesting since the quantity @xmath100 contains a divergence which is cut at @xmath101 or @xmath4 .
we have at @xmath102 @xmath103^{1/2}}\ ] ] ( of course , at @xmath104 the resistivity is exponentially small ) .
a negative @xmath95-linear magnetoresistance was observed recently in cro@xmath9 @xcite .
in the @xmath89 case we obtain @xmath105 where @xmath106 this integral diverges logarithmically at @xmath107 and as @xmath108 at @xmath109 taking into account the lower limit cutoff we derive @xmath110 we see that simple replacement of the electron - magnon scattering amplitude by @xmath28 does not enable one to describe correctly magnetoresistance even at @xmath111
now we treat the impurity contributions to transport properties in the presence of potential scattering ( they were considered first in @xcite , see also @xcite ) . to second order in the impurity potential @xmath112
we derive for the electron green s function @xmath113\ ] ] where @xmath114^{-1}\ ] ] is the exact green s function for the ideal crystal . in the second order in @xmath115
the electron self - energy has the form @xmath116 with @xmath117 the fermi function . neglecting vertex corrections and averaging over impurities we obtain for the transport relaxation time @xmath118 thus the contributions under consideration are determined by the energy dependence of the density of states @xmath119 for the interacting system near the fermi level .
the most nontrivial dependence comes from the non - quasiparticle ( incoherent ) states with the spin projection @xmath120 which are present near @xmath19 ( fig.1 ) .
they originate from the imaginary part of the electron self - energy @xcite .
we obtain at @xmath121 @xmath122 the contribution ( [ inc ] ) is asymmetric and vanishes at @xmath19 ( figs.1,2 ) . near the fermi level
it is determined by the magnon density of states @xmath123 and follows a power law , @xmath124 is the maximum magnon frequency , @xmath125 is the step function , @xmath126 is referred to @xmath19 ; we have @xmath127 and @xmath128 for @xmath97 and @xmath89 cases , respectively .
the corresponding correction to resistivity reads @xmath129 the contribution of the order of @xmath130 with @xmath131 ( which is not too far from 3/2 ) has been observed recently in the temperature dependence of the resistivity for nimnsb @xcite .
the incoherent contribution to magnetoresistivity is given by @xmath132 so that we obtain a temperature - independent term in the @xmath89 case .
the non - quasiparticle states in hmf can be probed also by nuclear magnetic resonance ( nmr ) since they lead to the unusual temperature dependence for the longitudinal nuclear magnetic relaxation rate , @xmath133 instead of the @xmath134-linear korringa contribution which is absent in hmf @xcite .
another useful tool is provided by tunneling phenomena @xcite , especially by andreev reflection spectroscopy for a hmf - superconductor tunnel junction @xcite .
the most direct way is probably the measurement of a tunnel current between two pieces of hmf with the opposite magnetization directions . to this end
we consider a standard tunneling hamiltonian ( see , e.g. , ref .
@xcite ) where @xmath136 are the hamiltonians of the left ( right ) half - spaces , respectively , @xmath21 and @xmath82 are the corresponding quasimomenta , and spin projections are defined with respect to the magnetization direction of a given half - space ( spin is supposed to be conserving in the `` global '' coordinate system ) .
carrying out standard calculations of the tunneling current @xmath137 in the second order in @xmath138 one has ( cf .
@xcite ) to conclude , we have considered peculiarities of transport properties of half - metallic ferromagnets which are connected with the unusual electronic structure of these systems .
further experimental investigations would be of great importance , especially keeping in mind possible role of hmf for applications @xcite .
the research described was supported in part by grant no.00 - 15 - 96544 from the russian basic research foundation , by russian science support foundation , and by the netherlands organization for scientific research ( grant nwo 047 - 008 - 16 ) . | peculiarities of transport properties of three- and two - dimensional half - metallic ferromagnets are investigated , which are connected with the absence of spin - flip scattering processes .
the temperature and magnetic field dependences of resistivity in various regimes are calculated .
the resistivity is proportional to @xmath0 for @xmath1 and to @xmath2 for @xmath3 @xmath4 being the crossover temperature for longitudinal scattering processes .
the latter scale plays also an important role in magnetoresistance@xmath5 the contribution of non - quasiparticle ( incoherent ) states to the transport properties is discussed .
it is shown that they can dominate in the temperature dependence of the impurity - induced resistivity and in the tunnel junction conductivity . |
in supersymmetric ( susy ) hybrid inflation models , inflation is associated with the breaking of a gauge symmetry @xmath2 to the minimal supersymmetric standard model ( mssm ) .
the symmetry breaking scale @xmath3 is fixed by the amplitude of the primordial curvature perturbation , and turns out to be of order @xmath4 gev , remarkably close , if not identical , to the supersymmetric grand unification scale @xcite .
a particularly simple and compelling example of @xmath2 is provided by the standard model gauge group supplemented by a gauged @xmath5 symmetry which requires , from anomaly cancellations , the presence of three right handed neutrinos .
other , more elaborate examples of @xmath2 in which susy hybrid inflation @xcite and its extensions , namely shifted hybrid inflation @xcite and smooth hybrid inflation @xcite , have been implemented include so(10 ) @xcite and its various subgroups . in the hybrid and shifted hybrid inflation models , the scalar spectral index @xmath6 , whereas in the smooth case , @xmath7 . furthermore , @xmath8 , while the tensor to scalar ratio @xmath9 turns out to be of order @xmath10 or less @xcite .
the main focus of this paper is to estimate the reheat temperature that is required to generate sufficient lepton asymmetry following hybrid , shifted or smooth hybrid inflation . although reheating and leptogenesis in these models has previously been addressed @xcite , the soft susy breaking terms and their impact on inflation were not adequately included .
it turns out that the new terms are significant for estimating the minimum reheat temperature in these models .
we also present revised estimates for the symmetry breaking scale associated with inflation as well as the scalar spectral index .
the layout of the paper is as follows . in section [ chap2 ]
we provide a brief review of susy hybrid inflation and its extensions .
we compute the allowed range of the dimensionless coupling in the superpotential and the dependence of the spectral index on this coupling , in the presence of canonical sugra corrections and soft susy violating terms . in section [ chap3 ]
we investigate reheating and the generation of matter following inflation in these models . for hierarchical right
handed neutrinos , we obtain a lower bound on the reheat temperature from the observed baryon asymmetry .
we also briefly comment on the resolution of the mssm @xmath11 problem and its impact on the reheat temperature .
the susy hybrid inflation model @xcite is realized by the renormalizable superpotential @xcite @xmath12 where @xmath13 denote a conjugate pair of superfields transforming as nontrivial representations of some gauge group @xmath2 , @xmath14 is a gauge singlet superfield , and @xmath15 @xmath16 is a dimensionless coupling . a suitable @xmath17 r - symmetry , under which @xmath18 and @xmath14 transform the same way , ensures the uniqueness of this superpotential at the renormalizable level @xcite . in the absence of supersymmetry
breaking , the potential energy minimum corresponds to non - zero ( and equal in magnitude ) vacuum expectation values ( vevs ) @xmath19 for the scalar components in @xmath20 and @xmath21 , while the vev of @xmath14 is zero .
( we use the same notation for superfields and their scalar components . )
thus , @xmath2 is broken to some subgroup @xmath22 which , in many interesting models , coincides with the mssm gauge group . in order to realize inflation , the scalar fields
@xmath20 , @xmath21 , @xmath14 must be displayed from their present minima . for @xmath23 ,
the @xmath20 , @xmath21 vevs both vanish so that the gauge symmetry is restored , and the tree level potential energy density @xmath24 dominates the universe . with supersymmetry
thus broken , there are radiative corrections from the @xmath25 supermultiplets that provide logarithmic corrections to the potential , with additional contributions to the inflationary potential arising from @xmath26 supergravity . with a minimal kahler potential one contribution to the inflationary potential
is given by @xcite @xmath27\,,\ ] ] where @xmath28 gev is the reduced planck mass .
there are additional contributions to the potential arising from the soft susy breaking terms . in @xmath26 sugra
these include the universal scalar masses equal to @xmath29 ( @xmath30 tev ) , the gravitino mass . however , their effect on the inflationary scenario is negligible , as discussed below .
the more important soft term , ignored as far as we can tell in all earlier calculations , is @xmath31 . for convenience ,
we write this as @xmath32 , where @xmath33 .
the effective potential is approximately given by the radiative corrections @xcite plus the leading sugra correction @xmath34 and this soft term : @xmath35 + a\,m_{3/2 } \kappa m^2|s|\ , , \end{aligned}\ ] ] where @xmath36 , @xmath37 is the dimensionality of the @xmath20 , @xmath21 representations , and @xmath38 is a renormalization mass scale .
we perform our numerical calculations using this potential , taking @xmath39=1 tev , and using the well known equations in the slow roll approximation : the number of e - folds after the comoving scale @xmath40 has crossed the horizon is given by @xmath41 here @xmath42 is the normalized real scalar field , @xmath43 is the value of the field at the comoving scale @xmath40 and @xmath44 is the value of the field at the end of inflation .
the amplitude of the curvature perturbation @xmath45 is given by @xmath46 which we evaluate for the comoving wavenumber @xmath47 mpc@xmath48 .
it is instructive to discuss small and large @xmath15 limits of eq .
( [ potential ] ) . for @xmath49 , eq . ( [ potential ] ) becomes @xmath50\,,\ ] ] to a good approximation .
this potential has been analyzed in @xcite , and the presence of the sugra correction was shown to lead to a blue spectrum for @xmath51 . for @xmath52 , @xmath53 where @xmath54 is the value of the field at @xmath55 , i.e. @xmath56 .
( note that due to the flatness of the potential the last 55 or so e - folds occur with @xmath57 close to @xmath3 . ) from eqs .
( [ potential ] , [ perturb ] ) , as @xmath58 @xmath59 the denominator of eq .
( [ spert ] ) contains the radiative , sugra and the soft terms respectively . comparing them , we see that the radiative term can be ignored for @xmath60 . for a positive soft term ( @xmath61 )
, the maximum value of @xmath45 as a function of @xmath3 is found to be @xmath62 setting @xmath63 @xcite , we find a lower bound on @xmath15 ( @xmath64 ) . for larger values of @xmath15
, there are two separate solutions of @xmath3 for a given @xmath15 . for @xmath65 , there are again two solutions , but for the solution with a lower value of @xmath3 , the slope changes sign as the inflaton rolls for @xmath60 and the inflaton gets trapped in a false vacuum .
the second solution in principle allows @xmath66 , but this is not very natural since it requires a delicate cancellation between two large terms in the denominator of eq .
( [ spert ] ) .
there is also a soft mass term @xmath67 in the potential , corresponding to an additional term @xmath68 in the denominator of eq .
( [ spert ] ) . we have omitted this term , since it is insignificant for @xmath69 .
the dependence of @xmath3 on @xmath15 as well as the allowed range of @xmath15 is shown in fig .
[ fig:21 ] .
note that the soft term depends on @xmath70 , so it should be checked whether @xmath70 changes significantly during inflation .
numerically , we find that it does not , except for a range of @xmath15 around @xmath10 . for this range ,
corresponding to the grey segments in the figure , if the initial value of the @xmath14 field is greater than @xmath3 by at least a factor of two or so , the soft term and the slope become negative even if they were initially positive , before inflation can suitably end ( fig .
[ farg ] ) .
as mentioned above , @xmath71 during the last 55 or so e - folds , so strictly speaking this range is not excluded , although the required initial conditions may look contrived .
qualitatively , the duration of inflation is given by @xmath72 , where @xmath73 denotes the total number of e - folds , and the hubble parameter @xmath74 .
the rate of change in @xmath70 is @xmath75 .
therefore one expects that @xmath70 would change significantly if @xmath76 . for the range of @xmath15 where the radiative term dominates , @xmath77 @xcite .
however , for @xmath60 , the soft term dominates the slope of the potential , and eq .
( [ spert ] ) then has solutions with higher values of the slope , and the duration of inflation is shorter .
consequently , @xmath70 stays fixed also in the left segments of the curves .
the dependence of @xmath78 on @xmath15 is displayed in fig .
[ fig:22 ] .
the segment with @xmath79 for small @xmath15 corresponds to the solution with a high symmetry breaking scale .
the running of the spectral index is negligible , with @xmath80 .
the experimental data disfavor @xmath78 values in excess of unity on smaller scales ( say @xmath81 mpc@xmath48 ) , which leads us to restrict ourselves to @xmath82 for @xmath83 . may be allowed in models where dissipative effects are significant .
such effects become important for large values of @xmath15 , provided the inflaton also has strong couplings to matter fields @xcite . ]
thus , the vacuum energy density during inflation is considerably smaller than the symmetry breaking scale .
indeed , the tensor to scalar ratio @xmath84 .
the inflationary scenario based on the superpotential @xmath18 in eq .
( [ super ] ) has the characteristic feature that the end of inflation essentially coincides with the gauge symmetry breaking .
thus , modifications should be made to @xmath18 if the breaking of @xmath2 to @xmath22 leads to the appearance of topological defects such as monopoles , strings or domain walls . as shown in @xcite , one simple resolution of the topological defects problem
is achieved by supplementing @xmath18 with a non - renormalizable term : @xmath85 where @xmath86 is comparable to the susy grand unified theory ( gut ) scale @xmath87 gev and @xmath88 is an effective cutoff scale .
the dimensionless coefficient of the non - renormalizable term is absorbed in @xmath89 .
the presence of the non - renormalizable term enables an inflationary trajectory along which the gauge symmetry is broken .
thus , in this ` shifted ' hybrid inflation model the topological defects are inflated away .
the inflationary potential is similar to eq .
( [ potential ] ) @xcite : @xmath90+a\,m_{3/2}\kappa v^2|s|{}\ , .
\end{aligned}\ ] ] here @xmath91 with @xmath92 , @xmath93 , and @xmath94 is replaced by @xmath95 in the expression for @xmath96 .
the slow roll parameters ( and therefore @xmath78 , @xmath97 , and @xmath9 ) are similar to the susy hybrid inflation model ( fig .
[ fig:22 ] ) .
the vev @xmath3 at the susy minimum is given by @xcite @xmath98 and is @xmath99 gev depending on @xmath15 and @xmath89 .
the system follows the inflationary trajectory for @xmath100 @xcite , which is satisfied for @xmath69 if the effective cutoff scale @xmath101 . for lower values of @xmath89
, the inflationary trajectory is followed only for higher values of @xmath15 , and @xmath3 is lower for a given @xmath15 ( fig .
[ fig:21 ] ) .
a variation on these inflationary scenarios is obtained by imposing a @xmath102 symmetry on the superpotential , so that only even powers of the combination @xmath103 are allowed @xcite : @xmath104 where the dimensionless parameter @xmath15 is absorbed in @xmath86 .
the resulting scalar potential possesses two ( symmetric ) valleys of local minima which are suitable for inflation and along which the gut symmetry is broken . as in the case of shifted hybrid inflation , potential problems associated with topological defects are avoided . the vev @xmath3 at the susy minimum is given by @xmath105 .
for @xmath106 , the inflationary potential is @xmath107\,,\ ] ] where the last term arises from the canonical sugra correction .
the soft terms in this case do not have a significant effect on the inflationary dynamics . in the absence of the sugra correction @xmath108 @xcite .
the sugra correction raises @xmath109 to above unity for @xmath110 gev , as shown in fig .
[ fig:24 ] ( this figure is slightly different from the figure published in @xcite , due to a computational error in the latter ) .
an important constraint on susy hybrid inflation models arises from considering the reheat temperature @xmath111 after inflation , taking into account the gravitino problem which requires that @xmath112@xmath113 gev @xcite .
this constraint on @xmath0 depends on the susy breaking mechanism and the gravitino mass @xmath29 . for gravity mediated susy breaking models with unstable gravitinos of mass @xmath1141 tev , @xmath115@xmath116 gev @xcite , while @xmath117 gev for stable gravitinos @xcite . in gauge mediated models the reheat temperature is generally more severely constrained , although @xmath118@xmath119 gev is possible for @xmath120100 gev @xcite . finally , the anomaly mediated symmetry breaking ( amsb ) scenario may allow gravitino masses much heavier then a tev , thus accommodating a reheat temperature as high as @xmath113 gev @xcite .
after the end of inflation in the models discussed in section [ chap2 ] , the fields fall toward the susy vacuum and perform damped oscillations about it .
the vevs of @xmath21 , @xmath20 along their right handed neutrino components @xmath121 , @xmath122 break the gauge symmetry .
the oscillating system , which we collectively denote as @xmath123 , consists of the two complex scalar fields @xmath124 ( where @xmath125 , @xmath126 are the deviations of @xmath121 , @xmath122 from @xmath3 ) and @xmath14 , with equal mass @xmath127 .
we assume here that the inflaton @xmath123 decays predominantly into right handed neutrino superfields @xmath128 , via the superpotential coupling @xmath129 or @xmath130 , where @xmath131 are family indices ( see later for a different scenario connected to the resolution of the mssm @xmath11 problem ) .
their subsequent out of equilibrium decay to lepton and higgs superfields generates lepton asymmetry , which is then partially converted into the observed baryon asymmetry by sphaleron effects @xcite .
the right handed neutrinos , as shown below , can be heavy compared to the reheat temperature @xmath0 . without this assumption , the constraints to generate sufficient lepton asymmetry would be more stringent @xcite .
guts typically relate the dirac neutrino masses to that of the quarks or charged leptons .
it is therefore reasonable to assume the dirac masses are hierarchical .
the low - energy neutrino data indicates that the right handed neutrinos in this case will also be hierarchical in general . as discussed in ref .
@xcite , setting the dirac masses strictly equal to the up - type quark masses and fitting to the neutrino oscillation parameters generally yields strongly hierarchical right handed neutrino masses ( @xmath132 ) , with @xmath133 gev .
the lepton asymmetry in this case is too small by several orders of magnitude .
however , it is plausible that there are large radiative corrections to the first two family dirac masses , so that @xmath134 remains heavy compared to @xmath0 .
a reasonable mass pattern is therefore @xmath135 , which can result from either the dimensionless couplings @xmath136 or additional symmetries ( see e.g. @xcite ) . the dominant contribution to the lepton asymmetry is still from the decays with @xmath137 in the loop , as long as the first two family right handed neutrinos are not quasi degenerate . under these assumptions ,
the lepton asymmetry is given by @xcite @xmath138 where @xmath139 denotes the mass of the heaviest right handed neutrino the inflaton can decay into .
the decay rate @xmath140 @xcite , and the reheat temperature @xmath0 is given by @xmath141 ( we have ignored the effect of preheating in hybrid inflation @xcite , which does not seem to change the perturbative estimate for @xmath0 significantly @xcite . ) from the experimental value of the baryon to photon ratio @xmath142 @xcite , the required lepton asymmetry is found to be @xmath143 @xcite . using this value , along with eqs .
( [ nls ] , [ reheat ] ) , we can express @xmath0 in terms of the symmetry breaking scale @xmath3 and the inflaton mass @xmath127 : @xmath144 here @xmath127 is given by @xmath145 , @xmath146 and @xmath147 respectively for hybrid , shifted hybrid and smooth hybrid inflation .
the value of @xmath127 is shown in figs .
[ minf ] and [ minf2 ] .
we show the lower bound on @xmath0 calculated using this equation ( taking @xmath148 ev ) in figs .
[ ktr ] , [ sm_mtr ] .
( [ reheat ] ) also yields the result that the heaviest right handed neutrino the inflaton can decay into is about 300 ( 5 ) times heavier than @xmath0 , for hybrid inflation with @xmath149 ( @xmath150 ) . for shifted hybrid inflation ,
this ratio does not depend on @xmath15 as strongly and is @xmath151 @xcite .
this is consistent with ignoring washout effects as long as the lightest right handed neutrino mass @xmath134 is also @xmath152 .
both the gravitino constraint and the constraint @xmath153 favor smaller values of @xmath15 for hybrid inflation , with @xmath154 gev for @xmath155 .
similarly , the gravitino constraint favors @xmath15 values as small as the inflationary trajectory allows for shifted hybrid inflation , and @xmath156 gev for @xmath101 .
smooth hybrid inflation is relatively disfavored since @xmath157 and @xmath158 for @xmath159 gev ( @xmath160 gev ) . is lower and consequently lower
reheat temperatures are allowed . ]
there are ways to evade these bounds on @xmath0 .
having quasi degenerate neutrinos increases the lepton asymmetry per neutrino decay @xmath161 @xcite and thus allows lower values of @xmath0 corresponding to lighter right handed neutrinos .
provided that the neutrino mass splittings are comparable to their decay widths , @xmath161 can be as large as @xmath162 @xcite .
the lepton asymmetry in this case is of order @xmath163 where @xmath164 gev for @xmath155 , and sufficient lepton asymmetry can be generated with @xmath0 close to the electroweak scale .
for other scenarios that yield @xmath0 of order @xmath165 gev in hybrid inflation , see @xcite ( without a @xmath166 symmetry ) and @xcite .
we end this section with some remarks on the @xmath11 problem and the relationship to @xmath0 in the present context .
the mssm @xmath11 problem can naturally be resolved in susy hybrid inflation models in the presence of the term @xmath167 in the superpotential , where @xmath168 contains the two higgs doublets @xcite .
( the ` bare ' term @xmath169 is not allowed by the @xmath17 r - symmetry . ) after inflation the vev of @xmath14 generates a @xmath11 term with @xmath170 , where @xmath171 is required for the required vacuum .
the inflaton in this case predominantly decays into higgses ( and higgsinos ) with @xmath172 . as a consequence the presence of this term significantly increases the reheat temperature @xmath0 .
following @xcite , we calculate @xmath0 for the best case scenario @xmath173 .
we find a lower bound on @xmath0 of @xmath174 gev in hybrid inflation , see fig .
@xmath157 gev for shifted hybrid inflation with @xmath101 for shifted hybrid inflation .
some scalars belonging to the inflaton sector acquire negative mass@xmath175 if @xmath176 is smaller .
@xmath177 corresponds to @xmath178 . ] , and @xmath179 gev for smooth hybrid inflation .
an alternative resolution of the @xmath11 problem in these models which has no impact on @xmath0 invokes an axion symmetry @xcite .
supersymmetric hybrid inflation models , through their connection to the grand unification scale , provide a compelling framework for the understanding of the early universe .
such models can also meet the gravitino and baryogenesis constraints through non - thermal leptogenesis via inflaton decay .
sugra corrections and previously ignored soft susy violating terms in the inflationary potential lead to lower bounds on @xmath15 and therefore , assuming hierarchical right handed neutrinos , on the reheat temperature @xmath0 .
the lower bounds on @xmath0 are summarized in table [ ttable ] . vs. the allowed range of @xmath15 , for susy hybrid inflation with @xmath180 ( solid ) , with @xmath181 ( dashed ) , and for shifted hybrid inflation ( dot - dashed for @xmath101 , dotted for @xmath182 gev ) .
light grey portions of the curves are for @xmath65 , where only the segments that do not overlap with the solutions for @xmath61 are shown .
the grey segments denote the range of @xmath15 for which the change in @xmath70 is significant.,width=453 ] changes as @xmath14 rolls down , for susy hybrid inflation with @xmath181 .
left : @xmath70 exceeds @xmath183 before the field reaches the waterfall point , and the field relaxes in a false vacuum .
right : the field reaches the waterfall point without a significant change in @xmath70 . the initial value of @xmath184
, @xmath185 is taken to be zero.,width=566 ] vs. the allowed range of @xmath15 , for susy hybrid inflation with @xmath180 ( solid ) , with @xmath181 ( dashed ) , and for shifted hybrid inflation with @xmath101 ( dot - dashed ) .
the grey segments denote the range of @xmath15 for which the change in @xmath70 is significant.,width=453 ] vs. the allowed range of @xmath15 ( @xmath186 ) , for susy hybrid inflation with @xmath180 ( solid ) , with @xmath181 ( dashed ) , and for shifted hybrid inflation ( dot - dashed for @xmath101 , dotted for @xmath182 gev ) .
the grey segments denote the range of @xmath15 for which the change in @xmath70 is significant.,width=472 ] vs. the allowed range of @xmath15 ( @xmath186 ) , for susy hybrid inflation with @xmath180 ( solid ) and for shifted hybrid inflation ( dot - dashed for @xmath101 , dotted for @xmath182 gev ) .
the segments in the top left part of the figure correspond to the bounds in the presence of a @xmath187 coupling .
the grey segments denote the range of @xmath15 for which the change in @xmath70 is significant.,width=453 ] | the allowed range of parameters for supersymmetric hybrid inflation and its extensions are investigated .
the lower bound on the reheat temperature @xmath0 in these models with hierarchical right handed neutrinos is found to be @xmath1 gev .
( @xmath0 as low as 100 gev is possible for quasi degenerate right handed neutrinos . )
we also present revised estimates for the scalar spectral index and the symmetry breaking scale associated with inflation .
ba-04 - 13 + |
the theory of particle acceleration in pulsar magnetospheres has been under development for almost three decades .
although it was well known that rotating magnetic dipoles would induce electric fields in vacuum ( deutsch 1955 ) , it took several years after the discovery of pulsars to realize that the huge vacuum fields could not in practice be available for particle acceleration .
the electric field parallel to the magnetic field is at least partly screened by particles supplied from the stellar surface ( goldreich & julian 1969 ) or by electron - positron avalanches ( sturrock 1971 ) . the true accelerating voltage of a pulsar must be determined by departures from the corotation , or goldreich - julian charge density , that could completely screen the parallel electric field .
several types of models have studied pulsar acceleration due to charge deficits at different locations in the magnetosphere .
polar cap ( pc ) models consider the formation of a parallel electric field in the open field region near the magnetic poles , while outer gap models consider accceleration in the outer magnetosphere , near the null charge surface ( see mestel 1998 for the most recent and comprehensive review of pulsar electrodynamics ) .
ruderman & sutherland ( 1975 ; hereafter rs75 ) introduced a pc model invoking a vacuum gap due to the trapping of ions in the neutron star ( ns ) crust .
the calculations by jones ( 1985 , 1986 ) , and neuhauser et al .
( 1986 , 1987 ) seem to favor a low value for the work function ( at least a factor of 10 less than it was thought earlier ) in the ns surface with a strong magnetic field . the important implication of this study is that the possibility of free ejection of charges ( actually of both signs ) from the ns surface can be now , at least theoretically , justified . in this paper , we concentrate on a space - charge limited flow model ( implying very low work function in the ns surface ) based originally on the work of arons & scharlemann ( 1979 ; hereafter as79 ) , who determined the electric field produced by the small deparature from the goldreich - julian charge that grows above the surface due to the geometry of the open dipole field .
the electric field accelerating electrons in this model developed along only field lines that curved toward the rotation axis ( favorably " curved field lines ) , so that acceleration occurred over half of the pc .
the parallel field is shorted - out at a height above the surface where the @xmath0-rays from accelerated particles produce sufficient electron - positron pairs in the strong magnetic field .
the accelerating potential is thus limited by such a pair formation front ( pff ) .
these initial calculations of electron - positron pffs assumed that the primary electrons began accelerating at the ns surface and that curvature radiation ( cr ) was the only mechanism for providing pair - producing photons ( arons 1983 ; hereafter a83 ) . in recent years , it has become clear that inverse - compton scattering ( ics ) of soft thermal x - ray photons from the hot ns surface by the primary electrons is also an important mechanism above the pc . as well as being a significant energy loss ( kardashv et al .
1984 , xia et al . 1985 , daugherty & harding 1989 , sturner 1995 ; hereafter s95 ) and radiation ( sturner & dermer 1994 ) mechanism , ics can also provide photons capable of producing pairs .
pulsed x - rays have been detected from a number of pulsars which are consistent with blackbody spectra at temperatures around @xmath2 k ( ogelman 1995 ) .
zhang & qiao ( 1996 ) and zhang et al . ( 1997 ; hereafter zqlh97 ) first explored the effect of the pairs from inverse - compton photons on the acceleration in a ruderman - sutherland type model .
they found that ics photons may produce a pff sooner ( at a lower altitude ) than the cr photons would from the same accelerating electrons .
in fact in this case , the electrons will stop accelerating before they can emit significant cr .
the standard models of pc acceleration thus need substantial revision .
another effect which has never been included in pc acceleration models is the formation of a lower pff due to the positrons that are turned around and accelerated downward from the electron pff .
although the number of positrons which are accelerated downward is small compared to the number of primary electrons and even to the charge deficit near the upper pff , the multiplicity of the downward cascades is quite large ( as we will discuss in section 3.1 ) .
thus , the amount of charge produced by only a small number of downward moving positrons may be sufficient to establish a second pff . although downward going cascades have been discussed in previous papers ( see e.g. as79 ) , their effect on the acceleration of primaries has not been investigated .
daugherty & harding ( 1996 , hereafter dh96 ) qualitatively discussed the effects of pair cascades by returning positrons , their creation of pairs within the acceleration zone and the need for a self - consistent model of pc acceleration . in this paper , we present a detailed study of the acceleration of primary electrons and secondary ( downward - moving ) positrons above a pulsar pc , assuming space - charge limited flow ( free emission ) of particles from the ns surface ( see harding & muslimov 1998 , for a review ) . we include the general relativistic effect of inertial frame - dragging , which induces a much larger electric field than that expected in the flat spacetime and is not limited to favorably - curved field lines ( section [ sec : ell ] ) .
this is important because , as has been concluded earlier in papers by fawley , arons & scharlemann ( 1977 ) and a83 , the potential drops ( derived for flat space - time ) are not sufficient to account for oberved pulsar @xmath0-rays .
both electrons and positrons suffer energy loss and emit photons from cr and ics .
the treatment of ics of both upward and downward moving particles requires revisions from previous studies of only upward moving particles ( dermer 1990 ; hereafter d90 ; s95 ) , which are presented in section [ sec : ic ] .
we then compute the location of both electron and positron pffs due to one - photon pair production as a function of magnetic colatitude and height of the lower pff .
we also discuss in section [ sec : pfraction ] the fraction of positrons that are turned around at the upper pff .
when the electrons are assumed to accelerate from the ns surface , we find that ics photons produce the pffs ( section 3.1 ) , in agreement with the results of zqlh97 .
however , we also find that there is substantial difference between the scattering of upward - going electrons and downward - going positrons by the same thermal x - ray photons : the electrons scatter these photons at angles less than @xmath3 , while the positrons scatter the photons head - on .
electrons produce pairs through resonant scattering , while positrons produce pairs by scattering above the cyclotron resonance .
the photons scattered by positrons are therefore more energetic and produce pairs in a shorter distance .
these pairs may screen the accelerating field up to some distance above the surface .
we find that stable , double pffs can form only when cr photons produce them , i.e. at a height where cr losses overtake ics losses .
we compute the height of these stable pffs as a function of pulsar period and surface field strength ( section 3.1 ) .
one interesting result is that the acceleration voltage limited by cr - controlled pffs is only a function of magnetic colatitude ( i.e. geometry of the open field lines ) , ranging between @xmath4 and @xmath5 , and is insensitive to pulsar parameters such as period and surface value of the magnetic field strength and even to the height of the acceleration .
the stable location of the lower pff depends primarily on surface magnetic field , temperature and size of the hot polar cap , ranging between 0.5 and 1.0 stellar radius , but is insensitive to period
. implications of these results for high - energy pulsar emission are discussed in section [ sec : dis ] .
we consider a test particle approach to the determination of pff locations .
the test particles in this case are electrons or positrons that gain energy through electrostatic acceleration and lose energy through radiative losses .
photons are created by cr and ics and destroyed by magnetic pair creation .
we assume that the electric field is completely screened at the point where the first pair is produced .
this is a good assumption for several reasons .
first , as discussed in more detail below , the electric field arises due to a small imbalance between the actual charge density and the local , rotation - induced , goldreich - julian charge density .
it therefore does not require much additional charge to short - out this field .
second , the onset of pair cascading occurs very quickly ( dh96 ) , so that the number of pairs produced per primary particle increases rapidly over small distances .
thus , as found in a83 , the width of the pff ( the screening distance of the electric field ) is very small compared to other dimensions of the problem .
figure 1 illustrates the geometry of the calculation .
suppose that the pff results from pairs produced by @xmath0-rays of energy @xmath6 ( in units of @xmath7 ) radiated by particles of energy @xmath8 .
the distance of the pff from the starting point of the particle acceleration at @xmath9 is then : [ sc ] s_c = min[s_a(_min ) + s_p(_min ) ] where @xmath10 is the distance required to accelerate the particle until it can radiate a photon of energy @xmath11 , and @xmath12 is the pair attenuation length of the photon .
the acceleration distance , @xmath10 , is determined by first integrating the equation of motion of the particle to determine its energy as a function of its pathlength @xmath13 : [ dgamma ] cdds = emce _ - _ _ ic - _ _ cr , where @xmath14 is the electric field induced parallel to the magnetic field , @xmath15 and @xmath16 are the loss rates for ics and cr .
discussion of these processes will be given in sections [ sec : ell ] and [ sec : rloss ] .
the pair production attenuation length of photons radiated by the particle through either ics or cr is then determined .
this attenuation length @xmath17 , defined to be the path length over which the optical depth is unity , is given by [ tau ] ( ) = _ 0^s_p ( ) t_pp(_kb , ) ds = 1 , where is the pathlength differential along the photon momentum vector , @xmath18 is the attenuation coefficient for one - photon pair production and @xmath19 is the angle between and the local magnetic field direction .
computation of @xmath17 will be discussed in section [ sec : pair ] . as we will discuss in section [ sec : numres ] , when ics causes the pff@xmath20 and pff@xmath21 , @xmath22 is smaller than @xmath23 , due to different modes of ics .
the pff@xmath20 will then not coincide with the start of the electron acceleration at @xmath9 .
we use the electric field due to inertial frame dragging above the ns surface , first calculated by muslimov & tsygan ( 1990 , 1992 ; hereafter mt90 , 92 ) .
the regime under which the generation of this field occurs is actually the same as implied in the electric field computations in flat spacetime by as79 , that of space - charge limited flow .
an electric field must be present above the ns surface because as charges flow along open magnetic field lines , the corotation , or goldreich - julian charge density @xmath24 , can not be maintained .
even though @xmath25 and therefore @xmath26 at the surface , the curvature of the field lines causes the area of the open field region , through which the particles flow , to increase faster than @xmath24 , and a charge deficit grows with distance .
thus , the @xmath14 grows with height up to about one stellar radius above the surface , and then drops off .
it is important that for a nearly aligned ns in flat spacetime , the space - charge density of the outflowing particles ( electrons ) proved to be almost exactly compensated by the effective goldreich - julian charge density , thus resulting in the substantial suppression of the electric field in the region of open field lines .
general relativity causes a significant modification of the mechanism of the electric field induction , through the effect of dragging of inertial frames , a consequence of the distortion of spacetime by a rotating gravitating body .
an observer near a rotating mass experiences a force and must corotate to remain in an inertial frame , with an angular velocity that decreases with distance from the rotating mass .
the charge density above a ns surface must be computed in a local inertial frame that is rotating with respect both to the ns and to an observer at infinity .
thus the goldreich - julian charge density , which is the charge density required to make magnetospheric particles drift in corotation with the star , will differ from that in flat spacetime .
this charge difference enhances @xmath14 over what it would be in flat spacetime , by a factor of 50 - 100 for a typical 1 s pulsar . the frame - dragging contribution to @xmath14 that signifies the striking difference between the general relativistic and classical treatments of ns electrodynamics is proportional to @xmath27 ( see equation [ [ ell_1 ] ] below ) , where @xmath28 is the angle between the magnetic and spin axes of the pulsar .
[ obviously , there is also a frame - dragging contribution to @xmath14 which is proportional to @xmath29 , but it is comparable to that produced in a flat spacetime limit . ]
particle acceleration may therefore occur throughout the entire open field line region , with the relative contribution of the frame dragging component to @xmath14 being strongest for pulsars with small obliquities .
muslimov & harding ( 1997 ; hereafter mh97 ) derived expressions for the @xmath14 due to frame dragging in two limits : close enough to the ns surface such that @xmath30 , where @xmath31 is the pc half - angle , and far from the surface such that @xmath32 .
these expressions were derived from solutions to poisson s equations assuming boundary conditions @xmath26 , as well as the potential @xmath33 at the stellar surface and along the last open field line . we have adapted these solutions for use in this paper by incorporating the screening effect of an upper boundary at @xmath34 , i.e. of the pair formation front , where @xmath35 .
although this approach is not fully self - consistent , in section 2.1.1 we discuss how the screening of the electric field at the upper boundary could be included in our calculations in a self - consistent way . in this paper
we also explore the situation where the positrons flowing back to the pc surface initiate electron - positron cascades .
although this possibility has been discussed since the very first papers on pulsar electrodynamics ( see e.g. as79 ) , it has never been addressed at an appropriate quantitative level . here
we attempt to approach this and some other related problems from the quantitative point of view which , we believe , can advance our understanding of the pc electrodynamics and its relevance to the modeling of @xmath36-ray emission from pulsars .
we note that the derivation of formulae for the electric field ( potential ) we exploit in the present analysis implies that the angle @xmath31 ( magnetic colatitude of the polar field line ) is small , which is a very good approximation for the region of open magnetic field lines in most pulsars .
however , it is instructive to look at the general - relativistic expression for the goldreich - julian charge density that contains a contribution of order of @xmath37 to the main term .
this expression reads _
gj = - b_0 f ( ) \ { + 32 h ( ) } , where @xmath38 is the ns rotation frequency , @xmath39 is the surface value of the magnetic field strength at the magnetic pole , @xmath40 is the red - shift function , @xmath41 , @xmath42 is the gravitational radius of the ns , r is the stellar radius , @xmath43 is the dimensionless radial coordinate , @xmath44 , @xmath45 and @xmath46 are the moment of inertia and mass of the ns , respectively , and the functions @xmath47 and @xmath48 are defined below ( see equations [ 8 ] , [ 16 ] ) . in formula ( 4 ) the functions @xmath49 and @xmath50 should be evaluated at @xmath43 and @xmath51 . in our derivation of the general - relativistic electrodynamic equations we neglect the contribution to the electric field produced by the flaring of the magnetic field lines , which is a factor of @xmath52 ( where @xmath53 is the pulsar spin period ) smaller than that produced by the frame dragging .
this contribution results from the second - order terms like that in the square bracket of equation [ 4 ] .
[ in section 2.4 we will refer to this equation to clarify yet another important issue .
] the small - angle approximation may not be accurate enough for the millisecond pulsars , for which @xmath54 .
we also note that the effect of the electric field generation by the frame dragging does not depend on the particular configuration of the stellar magnetic field , simply because while the distribution of the real space charge in the acceleration region is mainly determined by the flaring of the magnetic field lines , the effect of frame dragging on the goldreich - julian charge density is independent of the geometry of the polar field lines .
this means that one can not mimic the frame dragging effect simply by distorting the polar magnetic flux tube .
the possibility of electron - positron cascades near the stellar surface initiated by the backflowing positrons unavoidably implies that the bottom of the polar magnetic flux tube should be treated as a highly - conducting layer of an electron - positron plasma on the top of the pc surface , rather than as a regular ns surface made of iron or hydrogen atoms in the strong magnetic field . in a more extreme case
( see e.g. wang et al .
1998 ) , which seems to be less justified at the moment , the whole ns surface may be covered by a thin electron - positron plasma layer .
note that the presence of such a layer even on the top of the pc should substantially affect the electrodynamics and therefore the acceleration of charged particles in the region of open magnetic field lines .
it will allow the free supply of charges into the acceleration region , a requirement for space - charge limited flow .
also , we shall discuss in section 3 , a self - consistent regime with double pffs that may significantly differ from the standard one implying a single pff above the stellar surface . in this section and
the next , we discuss the principal modifications one must introduce into the treatment of the acceleration region to incorporate the effects resulting from the occurrence of double pffs .
first of all , the appropriate distance variables in @xmath14 should be rescaled by an effective " ns radius @xmath55 ( cf .
fig 1 ) , which is now assumed to be the lower boundary of the acceleration , such that @xmath56 , where @xmath57 is the dimensionless altitude above the effective surface .
the expressions for the electric field are thus symmetric between the lower and upper boundaries , in the sense that they allow the proper treatment of e.g. both primary electron and secondary positron acceleration , which is an essential requirement for the theory . to clarify our analytic solutions ( see section 2.1.2 ) for the rescaled electric potential ( accelerating electric field ) produced between the lower and upper pair fronts , let us consider the expression for the electric potential in terms of the difference between the actual and goldreich - julian charge densities @xmath58 . because @xmath59 and @xmath24 are just linear combinations of @xmath29 and @xmath27 , we can write the potential @xmath60 in the simple form ( at distances from the stellar surface greater than the pc size ) as [ phi ] = s ( r ) ( 1-^2 ) , where @xmath61 ^ 2 $ ] is the cross - sectional area of the polar magnetic flux tube at the radial distance @xmath62 and @xmath63 and @xmath64 are the coefficients of @xmath27 and @xmath29 in the expression for @xmath65 ( see equations [ [ rho ] ] and [ [ rhogj ] ] )
. it also says that the local value of the electric potential is of order of the local value of the effective space charge ( equal to the difference between the real space charge and induced goldreich - julian space charge ) divided by the characteristic longitudinal length scale .
the expression ( [ phi ] ) does not incorporate the effect of the screening of the electric field at the upper boundary ( pff ) .
this effect is included in the corresponding expressions we discuss later in section [ sec : epff ] . the relation ( [ phi ] )
gives the correct solution for the electric potential and longitudinal component of the electric field for altitudes greater than the pc size and satisfying the @xmath33 and @xmath66 boundary conditions at the ( effective ) stellar surface , where @xmath67 .
this expression is very useful for illustrating the modification of the potential and longitudinal component of the electric field resulted from the redefining of the position of the lower zero - electric field boundary ( set e.g. by the lower pff produced by the backflowing positrons ) .
it explains the physical meaning of such a modification , which is simply a readjustment of the true charge outflow ( in the regime of self - limitation ) to the local value of the goldreich - julian charge density at the effective surface , and allows us to better understand the corresponding exact analytic solution .
we introduce the radial coordinate scaled by effective radius @xmath68 , @xmath69 , and transverse coordinate @xmath70 , the magnetic colatitude scaled by the half - opening angle of the polar magnetic flux tube , ( ) = _ 0 ^ 1/2 at radius @xmath71 , where _ 0 = ^1/2 is the pc half - angle at the effective surface @xmath72 , and [ f ] f ( ) = -3()^3is the correction factor for the dipole component of the magnetic flux through the magnetic hemisphere of radius @xmath62 in a schwarzchild metric ( mh97 ) .
+ note that in equation ( 5 ) s(r ) .
the expressions for @xmath73 and @xmath74 , valid for any radius , can be written as ( see e.g. mh97 ) [ rho ] = -(r ) , [ rhogj ] _ gj = -(r ) . in these expressions ( r ) ( c ) ^2
r^3 b_0 1 , where @xmath75 ( @xmath76 , as it will be shown below ) .
thus , after substituting the above expressions for @xmath73 and @xmath74 into equation ( 5 ) we get the following formulae for the rescaled electric potential @xmath77 \xi \sin \chi \cos \phi \right\ } ( 1-\xi^2),\end{aligned}\ ] ] and the accelerating component of the electric field e _ = - e _ 0 _ 0 ^ 2 _ ^2 ( 1-^2 ) , where @xmath78 , and @xmath79 .
the quantities @xmath80 and @xmath81 in all expressions derived in this section for the rescaled radial coordinate now read , .
finally , the functions @xmath82 ^{-1 } , \\ & & \delta(\eta ) = { \partial \over { \partial \eta } } \ln{[h(\eta)\theta(\eta ) ] } \end{aligned}\ ] ] are both evaluated at @xmath83 .
we now include the effect of the electric field screening at the upper pff , at @xmath84 , by obtaining the solution to poisson s equation with the upper boundary condition , @xmath85 . in the limit where @xmath86 and @xmath87 the accelerating component of the electric field reads @xmath88 \cos\chi + \right .
\nonumber \\ & & \left .
\left [ \sum_{i=1}^{\infty}\,b_i\,j_1(\tilde k_i\xi)\right ] \sin\chi\cos\phi\right\},\end{aligned}\ ] ] where , [ a_i , b_i ] a_i = 32 f_i(z,_i ) , b_i = 38_0h(1)(1)f_i(z,_i ) , [ f(z , gamma ) ] f(z , ) = -[a_1(-1)e^z + a_2(+1)e^-z + a_1(1- ) - a_2(1+)]/(a_1+a_2 ) , a_1 = ( _ c+1)e^-z_c - - 1 , a_2 = - 1 - ( _ c-1)e^z_c , and _ i , and _ i , where @xmath89 and @xmath90 are the positive roots of the bessel functions @xmath91 and @xmath92 , respectively . in expressions ( 20 ) , ( 21 ) @xmath93 or @xmath94 should be used . in the limit , where @xmath95 and @xmath96 ( cf .
appendix a , equation [ a4 ] ) , @xmath97 \cos\chi \right . \nonumber \\ & & + { 3\over 8}\,\delta(\eta ) h(\eta)\theta(\eta)\,\left [ \xi(1-\xi^2)- { \eta_c \over \eta } { { \theta(\eta_c)\delta(\eta_c)h(\eta_c ) } \over { \theta(\eta)\delta(\eta)h(\eta ) } } \right . \nonumber \\ & & \left .
\sum_{i=1}^{\infty}\,{16 j_1(\tilde k_i\xi)\over \tilde k_i^3\,j_2(\tilde k_i)}\,e^{-\tilde\gamma_i(\eta_c ) ( \eta_c-\eta)}\right ] \sin\chi\cos\phi\right\ } \end{aligned}\ ] ] where now , _
i = k_i , and _ i = k_i simple analytic expressions can be derived in several limiting cases some of which are summarized in appendix a ( for the radial coordinate scaled by the true stellar radius ) . the solutions ( [ ell_1 ] ) and ( [ ell_2 ] ) thus incorporate both the effect of the rescaled lower boundary and screening of the electric field at the upper boundary , and we will use them in our numerical calculations described later on in section ( [ sec : numres ] ) . in this section ,
we describe our treatment of the radiation processes that affect the polar cap particle acceleration .
this includes energy losses due to cr and ics , and pair production by the photons from these processes .
although the particles are radiating a full spectrum of photons , we are concerned here only with the pair producing photons . in the interest of making our numerical code as efficient as possible ( but with some loss of accuracy ) , we do not model the entire radiation spectrum of each process , but compute pair production attenuation lengths for a single representative photon energy at each step along the particle path .
the cr loss rate for particles of charge @xmath98 moving along a magnetic field with radius of curvature @xmath99 is [ gdot_cr ] -__cr = 3 e^22mc^2c_c^2 ^ 4 . the radius of curvature of the magnetic field in a schwarzschild metric is [ rhoc ] _ c ( cr)^1/2r g(x ) , and g(x ) = ( 1-x)^3/2f^5/2(x)9 - 2f(x)(4 - 3x ) in the small angle limit , where @xmath100 , @xmath101 is the gravitational radius , and the function @xmath102 has been defined in equation ( [ f ] ) .
formula ( [ rhoc ] ) gives a slightly larger radius of curvature than the flat - space formula does ( by a factor of g , about @xmath103 , for the ns parameters we assume in this paper ) .
this can be easily understood , since the strong gravity of the ns tends to increase the flaring of the polar field lines , and therefore the last open field lines ( i.e. those reaching the light cylinder at their `` turning - points '' ) should emanate from slightly smaller magnetic colatitudes . in other words , the polar field lines get slightly straighten out , and the effective pc radius slightly decreases ( by the same @xmath104 , see also gonthier & harding 1994 ) .
the spectrum of cr photons from a particle with energy @xmath36 is a power law at low frequencies , with an exponential decline at high frequencies : [ crspec ] n__cr ( ) \ { ll ^1/3 , & _
cr + , & > _ cr . where @xmath105 is the critical frequency .
since the one - photon pair attenuation coefficient ( equation [ [ eq : ppasymp ] ] ) increases sharply with energy as @xmath106 , the photons from the exponential tail of the curvature spectrum will produce the bulk of the pairs . here , and in subsequent sections , the local dipole magnetic field strength is [ b ] b = ( b_0b_cr)(rr_e)^3 . where @xmath107 g is the critical field strength .
the overlap of the curvature spectrum and the pair attenuation coefficient will fall in a narrow frequency band , which we can approximate in a steepest descents analysis as a gaussian of width , @xmath108 , with a mean energy @xmath109 .
photons will pair produce roughly when @xmath110 for @xmath111 and as soon as threshold @xmath112 , is reached , for @xmath113 ( daugherty & harding 1983 ; hereafter dh83 ) .
we therefore have lll _
p = 13.3_cr , & _
p = 0.273_p , & b < 0.1 + _ p = 4_cr/3b , & _ p = 0.87b^1/2_p , & b > 0.1 at each step along the particle acceleration path , pair attenuation lengths are computed for up to five cr photons with energies evenly distributed between @xmath114 and @xmath115 . the first finite value ( if any ) of the pair attenuation length is taken as @xmath116 for that step .
as the particles are accelerated , they may scatter soft photons from the hot ns surface . due to the strong magnetic field ,
the ics cross section is resonant at frequencies where electrons may be excited to higher landau states and is strongly suppressed ( for extraordinary mode photons ) below the fundamental . although the full qed scattering cross section for this process has been calculated ( e.g. daugherty & harding 1986 , bussard et al .
1986 ) , it is quite unwieldy for numerical calculations . however , the magnetic scattering cross section in the thomson limit ( canuto et al .
1971 ) is much simpler and has been used in place of the full qed cros section in almost all astrophysical applications .
it includes only the resonance at the cyclotron fundamental ( i.e. scattering in the ground state with no excitation ) and approaches the thomson cross section at frequencies above the cyclotron energy . in this paper , we follow the treatment of s95 , who derived simple expressions for electron energy loss rates for magnetic compton scattering in the thomson limit , expanding the treatment of d90 to include the klein - nishina suppression above the resonance .
dermer s approach was to divide the compton scattering loss rates into three components , each due to the different behavior of the cross section in the regions below the resonance ( which he called angular " scattering due to the strong angle dependence ) , in the resonance ( resonant " scattering ) and above the resonance ( non - resonant " scattering ) .
he assumed that the resonant part of the cross section was a @xmath117-function at the cyclotron energy , and that the soft photon source was a blackbody of temperature @xmath118 radiated by a hot pc of radius @xmath119 .
he derived the scattering loss rates for electrons moving away from the ns surface , so that the soft photon density decreases with height , due to a dilution factor .
we have expanded this treatment by computing the associated loss rates for particles ( in our case , positrons , although the sign of charge makes no difference for compton scattering ) moving toward the ns surface . in this case , the particles approach the soft photons
head - on " , scattering them to higher energies than the downward moving particles . the compton scattering energy loss rate for a particle of energy @xmath0 moving through a photon field , @xmath120 , that is uniform between angles @xmath121 , from equation ( 12 ) of d90 , is [ gdot ] -__ic = c_+ - _ - _ 0^[(1-_-)]^-1 d n_ph()__-^_+d(1- ) _
-1 ^ 1d_s_0^_s , maxd_s ( dd_s d_s)(_s - ) , where @xmath122 is the maximum scattered energy in the particle rest frame . here ,
primes are used to denote quantities in the particle rest frame and the subscript @xmath13 denotes scattered quantities .
thus , @xmath123 is the rest frame scattering cross section .
based on these expressions , the loss rates for the three parts of the cross section , @xmath124 , used in equation ( [ gdot ] ) can be written [ gdot_ang ] -_ang = 46.1t_6 ^ 4(1-_c)f_ang s^-1 , [ gdot_res ] -_res = 4.9 10 ^ 11 f_res s^-1 , [ gdot_kn ] -__kn = 3.7 10 ^ 11 ( t_6 ^ 2 ) _
_-^_+d[_0 + _ 0/ - ] f_kn s^-1 , where @xmath125 k is the ns surface temperature , @xmath126 is the local dipole magnetic field strength in units of @xmath127 g. equation ( [ gdot_kn ] ) assumes a @xmath117-function at energy @xmath128 , where @xmath129 , for the distribution of soft thermal photons , and [ fang ] f_ang = ^2 - 2 ^ 2 - 1 - ( _ + ^3-_-^3 ) 3(_+ -_- ) - ( _ + + _
-)2 ^ 2 + ^3 ^ 4(_+ - _ - ) , @xmath130 } + \left({2\over 3}-{1\over \eb}\right ) { 1\over \gamma } \nonumber \\ & & \left\{[w_+ \ln(1-e^{-w_+})-w_- \ln(1-e^{-w_- } ) - \theta_t[li_2(e^{-w_+ } ) - li_2(e^{-w_-})]\right\},\end{aligned}\ ] ] @xmath131 } - \nonumber \\ & & { 2(1+\erg')(\erg'^2 - 2\erg'-1)\over \erg'(1 + 2\erg')^2 } - \left[{\erg'^2 - 2\erg'-3\over \erg'^2}\right]\ln{(1 + 2\erg ' ) } - { 2\over \erg ' } .\end{aligned}\ ] ] in the above expressions , @xmath132 $ ] , @xmath133 and @xmath134 is the local cyclotron energy in units of @xmath7 . equations ( [ gdot_ang ] ) , ( [ gdot_res ] ) and ( [ gdot_kn ] ) are identical to those given by s95 , but the expressions for @xmath135 , @xmath136 and @xmath137 have been generalized to allow for scattering by downward moving particles .
specifically , the corresponding equations given by s95 have specialized to the case of upward moving particles ( @xmath138 , see equation ( [ mupm ] ) below ) .
equation ( [ fang ] ) is the same as d90 s equation ( 16 ) , which is general enough to compute the loss rate of either upward or downward moving particles .
equation ( [ fres ] ) includes , to first order , the rest - frame particle recoil @xmath139 $ ] and is a major improvement over previous expressions for resonant compton scattering that assume no recoil ( @xmath140 ) .
the first term in equation ( [ fres ] ) gives d90 s equation ( 54 ) , and the remaining terms result from recoil , where @xmath141 is the dilogarithm function .
the treatment of recoil in resonant scattering is necessary in this calculation to compute accurate energies for photons scattered by the electrons and positrons . given the generality of the above expressions , the only difference in computing scattering loss rates for upward - moving electrons and downward - moving positrons lies in the integration limits over polar angle @xmath142 in the laboratory frame . for the case of semi - isotropic blackbody radiation at height @xmath143 above a hot pc of radius @xmath144 , this translates into the following values for @xmath145 where _
c = _ c = h , ( when @xmath146 ) gives the dilution factor that is present in the soft photon density , [ n_ph ] n_ph ( ) = n_bb()(1-_c2 ) where n_bb ( ) = 8_c^3 ^ 2 [ -1 ] is a blackbody spectrum of temperature @xmath147 and @xmath148 is the electron compton wavelength . to determine the energy of the scattered photons that may produce pairs , the scattering kinematics of the electron and positron must be treated accurately . for non - resonant scattering in the thomson limit , i.e. when @xmath149 , we can assume there is no recoil in the particle rest frame , and the maximum scattered energies in the laboratory frame are ll _ s = 2 ^ 2(1-_c)_0 & , + _ s = 4 ^ 2_0 & . since the electrons and the soft photons are both moving away from the ns surface , @xmath150 depends on @xmath151 , which decreases with height .
however , the positrons scatter the soft photons head - on " with a maximum angle of @xmath152 , and @xmath150 is independent of height .
when @xmath153 , or for resonant scattering , we must include recoil in the rest frame to compute the scattered energies .
the average scattered energy in the laboratory frame is computed from the scattering rates , which for @xmath154 ( equation [ [ gdot_res ] ] ) and @xmath155 ( equation [ [ gdot_kn ] ] ) explicitly include recoil , and the loss rates , so that _ s for each component , where [ ric ] r_ic = c_+ - _ -_0^[(1-_-)]^-1 d n_ph()__-^_+d(1- ) _ -1 ^ 1d_s_0^_s , maxd_s ( dd_s d_s ) .
@xmath156 is thus the weighted average of @xmath157 , where @xmath158 for scattering by relativistic particles .
since photons scattered by the angular " part of the cross section below the resonance do not have energies above pair threshold , for our calculation we need only @xmath159 and @xmath160 . using equation ( [ ric ] ) and following dermer s ( d90 ) method of computing @xmath154
, we have r_res = 2.18 10 ^ 13 ( t_6b_12 ^ 2 ) s^-1 . for @xmath161 , the resonant scattered photon energy ,
@xmath162 is independent of incident angle , and is thus the same for upward and downward moving particles .
however , the energy of the incident photon in the particle rest frame , @xmath133 will not be the same , so the upward and downward moving particles will not necessarily both undergo resonant scattering . using equation ( [ ric ] ) and following sturner s ( s95 ) method of computing @xmath163 , we have r_kn = 3.7 10 ^ 11 ( t_6 ^2 ) _
_-^_+dr__kn s^-1 , where r__kn = 4 + 2(1+ ) ( 1 + 2)^2 + ( ^2 - 2-2 ) ^2 .
our numerical calculations show that @xmath164 approaches @xmath0 for @xmath165 , as expected .
thus , @xmath166 for @xmath167 .
the method we use to compute the electron - positron pair attenuation length of the photons , @xmath168 , has been described in detail in harding et al .
( 1997 ) . using equation ( [ tau ] ) ,
@xmath168 is computed by integrating the pair production attenuation coefficient of the photon along its path through the dipole field .
the photon is assumed to pair produce at the point where @xmath169 , and @xmath168 is then set to that path length .
the two main inputs needed are the energy of the photon and its angle to the magnetic field , @xmath19 .
the energies of the radiated photons were given in sections [ sec : cr ] and [ sec : ic ] .
for the angle of the radiated photons at the emission point , we assume that @xmath170 for compton scattered photons ( d90 ) and @xmath171 for cr photons . due to the curvature of the field lines
, @xmath19 will grow as the photon propagates , roughly as @xmath172 . to evaluate the pair production attenuation coefficient at each point along the path of the photon
, we lorentz - transform the photon energy and local magnetic field to the frame in which the photon propagates perpendicular to the local field .
this is the center - of momentum frame for the created pair , where the attenuation has its simplest form and the photon energy is @xmath173 .
the one - photon pair attenuation coefficient is considered in two regimes . for @xmath174 ,
photon pair produce far above threshold , where the asymptotic expression in the limit of large numbers of kinematically available pair landau states ( tsai & erber 1974 , dh83 ) can be used : [ eq : ppasymp ] t^pp _ , = 12 b _ , ( ) , _ , ( ) \ { lr ( 0.31 , 0.15 ) & 1 + + ( 0.72 , 0.48 ) ^-1/3 & 1 . [
eq : ppratlim ] where @xmath175 , @xmath176 is the fine - structure constant , @xmath177 is the dipole field strength at point @xmath13 along the photon path ( see equation [ [ b ] ] ) .
when @xmath178 , pair production will occur near threshold , where the above expression is not accurate .
we thus include the factor @xmath179 in @xmath28 , introduced by dh83 , as an approximation to the near - threshold attenuation coefficient . in this paper
we compute the attenuation length averaged over photon polarization .
soon after the concept of electron - positron pair production had been introduced into pulsar models ( sturrock 1971 ) , it became clear that the precipitation of positrons ( or , in some models , electrons ) onto the stellar surface may have some important effects .
for example , sturrock has noticed that ( see p. 531 ) `` if the positrons were all returned to the surface , the resulting space charge would reverse the sign of the electric field at the surface , cutting off the flux of primary electrons '' .
then he added that `` one possibility is that the configuration so adjusts itself that the primary flux is never quite cut off , so that the flow is steady , and another possibility is that the flow is oscillatory '' .
thus , he drew attention to the existence of an intrinsic feedback between the pair formation and the electric field accelerating primary particles . later on ,
several authors ( see e.g. rs75 , jones 1978 , as79 , and arons 1981 ) have discussed similar ideas focusing on the possibility of pc heating by the flow of energetic positrons ( electrons ) .
dh96 discussed the effect of downward moving positrons on the particle acceleration and the possibility of a lower pff .
the most detailed study has been performed by as79 who calculated the fraction of returning positrons , structure of the pff with positron trapping near but below the lower boundary of the pff , and briefly discussed the positron - initiated cascade very near the stellar surface .
they estimated the fraction of returning positrons as @xmath180 ( @xmath181 , where @xmath182 is the half - opening angle of the polar flux tube at the stellar surface ) of the goldreich - julian charge density . the main argument underlying
this estimate is that the contribution of the second term in the square bracket in equation ( 4 ) increases with altitude ( due to the flaring of the magnetic field lines emanating from the pc ) , and that ( see as79 , p. 867 ) `` the additional negative charge density needed to achieve @xmath66 , over and above that of the electron beam , is therefore a fraction @xmath183 of the negative charge density already present in the electron stream '' . and , `` as the pairs begin to form near but below some height ( at which the electric field shuts off ) , the residual @xmath184 of the diode causes the secondary electron - positron plasma to polarize and shield these `` external '' charges .
the formation of this polarization charge requires a dynamical response which , under most conditions , leads to the formation of the downward directed positron stream whose flux at the stellar surface is @xmath185 '' .
however , the main problem with this reasoning is that if we assume the zero - electric field boundary condition at the stellar surface , then `` the additional negative charge density '' is nothing else but the imbalance between the real and goldreich - julian charge densities that produces the electric field above the stellar surface in the first place .
this imbalance ( or charge deficit ) increases with altitude and reaches some maximum value .
thus , if the backflowing positrons shield these `` external '' charges at some height , then they will also screen the electric field all the way down to the stellar surface , and we should unavoidably come up with the situation envisaged by sturrock , where the returning positrons cut off the flux of primary electrons , thus resulting in the oscillatory regime . in this section ,
we explore the possibility of occurrence of a steady regime of positron backflow that is not capable of disrupting the current of primary electrons .
we also discuss the efficiency of pc heating produced by the precipitating relativistic positrons and high - energy quanta .
we also discuss the accounting of the feedback between the returning positron flux and electric potential distribution in the acceleration region in the self - limiting regime of particle flow .
the primary ( electron ) traversing the region of the electric potential drop emits a @xmath186ray photon that produces an electron - positron pair at the altitude @xmath187 above the stellar surface .
the electron - positron pairs keep moving forward , with some positrons of these pairs decelerating and eventually reversing the direction of their motion at some altitude @xmath188 .
it is very likely that the same effect tends to occur with the positrons from both additional photons emitted by the same electron or from a second generation of pairs .
an accurate determination of the fraction of backflowing positrons would require a self - consistent calculation of the screening of @xmath184 by the cascade pairs and the modification of the pair energy and spatial distribution by the screened @xmath184 .
short of carrying out such a detailed calculation here , we shall derive an upper limit on the fraction of returning positrons .
however , since the cascades develop very fast and the number density of particles increases in an avalanche , the altitudes at which the positrons from the higher - order cascades get decelerated should be very close to but slightly above @xmath189 .
the deceleration of positrons results in a charge separation in the initially quasi - neutral electron - positron beam , thus effectively shielding the electric field where the bulk of the pair front is produced .
we can justifiably assume that @xmath66 at @xmath190 .
the backflowing positrons slightly suppress the voltage all the way down to the bottom of the polar magnetic flux tube , so that the electric field @xmath184 ( and also potential @xmath191 ) vanishes at a height greater than the start - off height of the primary electrons .
this occurs because the flux of backflowing positrons is equivalent to a corresponding enhancement of the total electron current and therefore of the maximum number of electrons per second to be ejected into the acceleration region .
thus , in compliance with the zero - electric field boundary condition , the ejection radius should fix itself at the larger value corresponding to a higher effective `` rate of supply of goldreich - julian charge '' ( see also expressions [ 4 ] and [ [ ietae ] ] proving this statement ) .
let us introduce the goldreich - julian current ( associated with the corresponding charge density ) or the rate of ejection of charges per element of the solid angle into the region of open magnetic field lines = c | _ gj | ( r)^2 , where @xmath192 , and @xmath193 is an element of the solid angle in the pc region . in the rest of this section we present our estimate of the maximum power carried by the backflowing positrons .
using the general - relativistic expression for @xmath74 ( equation [ [ rhogj ] ] ) we can write [ ietae ] di_gj(__e ) = i_0 , where @xmath194 and @xmath195 .
the zero - electric field boundary condition at @xmath196 requires the magnitude of the primary electron current flowing from the pc surface to be equal to @xmath197 .
this expression implies ( compare with expression [ 4 ] ) that an enhanced electron current is ejected from the effective radius in the regime of self - limitation or , equivalently , allows for the additional current of positrons flowing from the upper pff downward .
it is important that in the steady - state regime the total current of actual charges ( electrons and positrons ) remains constant along the field lines , and is fixed by the goldreich - julian current at the zero - electric field boundary ( i.e. , in our case , at the effective radius @xmath68 ) .
the electric field above @xmath68 ( in the regime of self - limitation ) is produced by the imbalance between the actual current and the local value of the goldreich - julian current . at distances greater than the effective radius @xmath68 for the goldreich - julian current we can write [ ieta ] di_gj ( > _ e ) .
let us consider the situation when the fraction of positrons returning from the upper pff is not only sufficient to produce a lower pff , but is also big enough to affect the electric field accelerating the primary electrons .
for example , the backflowing positrons tend to reduce the accelerating electric field above the effective surface of radius @xmath68 , simply because the positron current is equivalent to the enhanced electron current which unavoidably requires the electrons to be ejected from the greater heights above the effective surface .
thus , the returning positrons may eventually screen the accelerating electric field and trigger an oscillating regime with alternating phases of developing and collapsing of double pair fronts .
as has been mentioned above , the oscillatory regime akin to that we discuss here was anticipated in his classical paper by sturrock ( 1971 ) , even though he did not discuss the lower pff .
the occurrence of the oscillatory regime depends on the fraction of backflowing positrons , which is determined by the particle kinematics within the upper pff and the penetration depth of the electrostatic field into the electron - positron plasma cloud . a quantitative description of this process would be possible through a detailed analysis of the electron - positron cascades and their feedback on the electrodynamics of the pair formation region .
we can now estimate the maximum positron current needed to screen the electric field accelerating the primary electrons .
for this purpose we assume that the bulk of the lower pair front sets up at the effective radius @xmath68 , and that the upper pair front establishes at the radial distance @xmath198 .
then , for the maximum positron current we can write , using equations ( [ ietae ] ) and ( [ ieta ] ) , @xmath199 = \nonumber \\ & & { { { \cal i}_0}\over { 2\pi } } \left [ \kappa \eta _ { \ast } ^3 \cos \chi + { 3\over 2 } h(\eta ) \theta ( \eta ) \xi \sin \chi \cos \phi \right],\end{aligned}\ ] ] where @xmath200 . the maximum value of the electric potential can be estimated as ( cf .
equation [ 13 ] ) _ max _ ( 1-^2 ) .
now we can derive the maximum total power put into the backflowing positrons ( cf .
mh97 , equations [ 76]-[78 ] ) : \ { l_e^+ } _ max _ _ _ max \ { di_e^+ } _ max \ { _ + } _ max i_0 ^ 2 \ { _ + } _ max l_sd , where [ lam+ ] \ { _ + } _ max = 34 _
( ^2 _ ^5 ^2 + 316 h^2 ^2 ^2 ) , and l_sd is the spin - down luminosity of a pulsar . a similar estimate for the maximum efficiency of acceleration of primary electrons yields [ lam- ] \ { _ - } _ max = 34 , which amounts to 10 @xmath201 ( for a small obliquity and @xmath202 ) . for relatively small obliquities ( or small magnetic polar angles and non - orthogonal rotator ) and typical pulsar spin periods of 0.1 - 1
s the first term in equation ( [ lam+ ] ) dominates , and we get [ lam+max ] \ { _ + } _ max ^2 _ ^6 . for a 1.4 solar mass ns and for a broad range of realistic equations of state of dense matter ( see e.g. lorenz , ravenhall , & pethick 1993 and ravenhall & pethick 1994 for the calculations of the ns moment of inertia for various equations of state ) @xmath203 . thus , for the ns of 1.4 solar mass and 8 - 10 km radius ( which is consistent with the most realistic stellar models ) we can estimate @xmath204 .
given @xmath205 ( see section [ sec : numres ] ) , equation ( [ lam+max ] ) yields [ lam+num ] \ { _ + } _ max ( 310 ^ -4 - 210 ^ -2 ) ( 0.15 ) ^2 .
this estimate combined with the recently observed x - ray luminosities ( that include both pulsed and non - pulsed components and imply isotropic emission ) of pulsars ( see becker & trmper 1997 ) may have rather interesting implications .
if the x - ray fluxes in some of these pulsars are dominated by the photons from the heated ( e.g. by the backflowing positrons ) pc , then luminosities higher than given by equation ( [ lam+num ] ) would indicate that these pulsars operate in the oscillatory regime discussed above .
according to the estimates by becker & trmper the pulsed x - ray luminosities ( for the case of isotropic emission ) , @xmath206 , for e.g. crab , vela , geminga , and psr 0656 + 14 are , respectively , @xmath207 , @xmath208 , @xmath209 , and @xmath210 ( where @xmath211 , @xmath38 and @xmath212 are the pulsar spin frequency and its time derivative , respectively ) .
although some of these pulsars have x - ray luminosities that lie above @xmath213 , the x - ray emission may not be isotropic ( see zavlin et al .
if we assume that at least for vela , geminga , and psr 0656 + 14 ( see e.g. harding & muslimov 1998 for the modeling of the soft x - ray and @xmath214ray emission for these pulsars ) the x - ray emission is beamed into a solid angle of @xmath215 1 steradian , then for these pulsars we can estimate that @xmath216 , @xmath217 , and @xmath218 , respectively .
the corresponding pc temperature for these pulsars can be estimated as t_pc ~(0.6 - 1 ) 10 ^ 6 ( 10 ^ 5 \{_x^p}_anis ) ^1/4 ( b_0 ) ^1/2 ( r ) ^3/4 ( p ) ^-3/4 k. thus , the estimated value of @xmath219 may indicate that the backflowing positrons precipitate onto the effective area smaller than that of the standard pc ( e.g. the returning positrons focus toward the magnetic axis as discussed below ) , and/or that the fraction of the returning positrons is well below the maximum possible one . then the latter would support the quasi - steady state ( mh97 ) rather than the oscillatory ( sturrock 1971 ) regime of pulsar operation .
the energetics of the pc heating is mainly determined by the energetics of the backflowing ( primary ) positrons .
their energy is eventually redistributed between the high - energy photons they emit and electron - positron pairs created by photons in the magnetic field .
let us consider the last open field line of the dipole magnetic field .
the photons emitted by backflowing positrons produce pairs on adjacent innermost field lines , which may reduce the effective area of the pc subject to the heating . here
we shall present very rough estimate of the geometrical filling factor for the pc heating qualitatively illustrating this effect .
the photons emitted by backflowing positron ( moving along the last open field line ) at height @xmath220 above the effective radius @xmath68 produces a pair at the effective radius and at the angular distance from the magnetic axis r^ = ( r_e+s_c^+)- s_c^+ , where @xmath221 is the magnetic colatitude of the point at which the photon is emitted , and @xmath222 is the angle between the tangent to the field line at the point of photon emission and the normal to the stellar surface ( here we may justifiably neglect the surface curvature ) . in a flat space limit elementary geometrical consideration yields the relationship @xmath223 , where @xmath224 is the angle between the tangent to the field line and the radius - vector of the point on the field line with the magnetic colatitude @xmath221 . for a dipole field , @xmath225 , so that @xmath226 , and we get r^ = r_e ( 1-s_c^+ ) .
since @xmath227^{1/2}$ ] , where @xmath228 is the angular size of a standard pc , we can write for the effective angular size ( at the actual stellar radius ) of the heated pc ^ = _ pc ( 1+s_c^+ ) ^1/2 ( 1-s_c^+ ) . thus , in the case where @xmath229 , and @xmath230 , we get @xmath231 the effective area of the heated pc may therefore be up to 50 @xmath201 less than the area of a standard pc .
the above estimate can be additionally justified by the fact that @xmath220 increases toward the pc rim , while @xmath68 remains practically constant .
we thus suggest that for some pulsars this effect may be worthy of discussion .
following the procedure outlined in sections [ sec : ell]-[sec : pair ] , we have made numerical calculations of the pffs and the total acceleration length , @xmath232 , from equation ( [ sc ] ) , for both upward moving electrons and downward moving positrons .
the parameters of the thermal radiation from the ns surface are somewhat uncertain .
although measured temperatures of pulsed x - rays lie in the range @xmath233 k , the distribution of the radiation is not known . for the calculations of this section
, we have assumed that the radiation is isotropically emitted from a hot pc of radius @xmath234 and that @xmath235 km .
other possibilities , and how they might affect these results will be discussed in section [ sec : dis ] . an electron is started at height @xmath9 above the ns surface with @xmath236 . for a given value of @xmath9 and thus , @xmath237 , a
first guess " value of @xmath238 , and thus also of @xmath23 , sets the initial acceleration length .
equation ( [ dgamma ] ) is integrated in discrete steps upward from the starting point , computing @xmath14 ( from either equation [ [ ell_1 ] ] or [ [ ell_2 ] ] ) , @xmath15 and @xmath16 at each step . at each step , the pair attenuation lengths , @xmath168 , of both cr and ics test photons radiated by the particle of energy @xmath239 are computed from equation ( [ tau ] ) .
the pair attenuation length , and thus the value of @xmath23 , also computed at every step , is initially infinite , because the energy of the photons is small , but decreases with distance as the energy of the radiated photons increases .
although the photon attenuation length continues to decrease , the particle acceleration length is increasing and @xmath23 has a minimum .
this minimum value of @xmath240 is adopted as the new value of the electron pff for the assumed value of @xmath9 .
the electron is accelerated again with the new value of @xmath238 , producing a new pff at the next value of @xmath240 .
the process is repeated , converging to a self - consistent value of @xmath238 . when @xmath241 , we find that the value of @xmath23 for ics photons is smaller than that for cr photons
. pffs at low altitudes are therefore ics - controlled .
figure [ fig : prim1 ] shows examples of self - consistent solutions of the ics - controlled electron pff near the ns surface ( left panel ) and the corresponding ics - controlled positron pff ( right panel ) . in the left panel of fig .
[ fig : prim2 ] , an electron starts very near the ns surface ( at height @xmath242 ) , where the density of soft thermal photons and the loss rate due to ics is high . in this case
, photons from resonant ics produce pairs well before cr photons and thus the ics photons establish the pff . in fact , the loss rate for cr is orders of magnitude smaller than that for ics at the pff . in the right panel of fig .
[ fig : prim1 ] , the loss rates and integrated energy of the positron , accelerated by the same electric field as the electron , is shown as a function of the acceleration length .
it this case , however , the larger angles between the thermal x - ray photons and the positrons allow them to accelerate through the resonant part of the ics losses at lower energies , and non - resonant ics of photons above the cyclotron energy produce the pff . since at pulsar field strengths , scattering above the cyclotron energy occurs in the kn regime , the scattered photon energies are much larger than those of photons scattered in the resonance .
therefore , the ics photons producing the positron pff have higher energies than the ics photons producing the electron pff ( see fig [ fig : epsmin ] ) and the positron pff forms in a shorter distance .
this is the major cause of the difference between the electron and positron ics - controlled pffs , which will be discussed in more detail in connection with fig .
[ fig : pffep ] . figure [ fig : prim2 ] shows an example of the self - consistent solution for a cr - controlled electron pff .
the electron begins accelerating at a significant fraction of a stellar radius above the surface , where the soft photon density has dropped , cr losses dominate and cr photons establish the pff .
because cr is much less efficient than ics , the particles must accelerate to higher energies to produce pairs and the total acceleration length @xmath232 is longer .
we first explore the solutions for the electron pff , as a function of pulsar parameters , for acceleration from the surface .
figures [ fig : pcacch ] and [ fig : pffxb ] show the height of the electron pff and the maximum acceleration voltage ( @xmath243 ) as a function of the scaled colatitude @xmath244 for cr - controlled pffs , assuming that ics is turned off " .
these results can be directly compared with fig .
5 of a83 , who computed the pff due only to curvature photons .
figure [ fig : pcacch ] shows the strong effect of obliquity @xmath28 on the shape of the electron pff and on the acceleration voltage , due primarily to the dependence of @xmath14 on @xmath28 . since the frame - dragging component of @xmath14 proportional to @xmath27 ( cf .
equation [ [ ell_1 ] ] ) is much stronger near the surface than the component proportional to @xmath29 , it dominates at small and intermediate values of @xmath28 .
the frame - dragging electric field becomes comparable to that in a flat space for orthogonal rotators ( @xmath245 ) , where our solutions match very well with those of a83 s fig .
5 . the height of the cr - controlled pff ( @xmath23 ) increases and @xmath243 decreases with increasing @xmath28 , as @xmath14 decreases .
the height of the pff increases sharply ( and in fact goes to infinity ) both at the magnetic pole ( @xmath246 ) and at the pc rim ( @xmath247 ) .
these features , referred to as slot gaps by arons , form at the pole and equator due to different effects .
the gap at the pole occurs because the radius of curvature of the magnetic field lines is infinite there , causing the pair attenuation length to go to infinity .
the electrons keep accelerating to high altitudes and thus reach high energies . at the pc rim ,
the boundary condition on the potential ( @xmath33 ) , screens @xmath14 close to the rim , preventing electron acceleration to high enough energy to produce any pairs .
thus , @xmath243 also goes to zero at the rim .
figure [ fig : pffxb ] shows the dependence of the pff height and acceleration voltage on surface magnetic field strength @xmath248 . while the height of the pff decreases with increasing @xmath248 , due both to an increase in @xmath14 and a decrease in pair attenuation length , @xmath243 is roughly constant even though @xmath248 varies over two decades .
this is because the longer total acceleration distance for lower field strengths compensates for the lower rate of acceleration .
figure [ fig : pffsx ] shows examples of ics - controlled pffs near the ns surface .
we find that both the height of the pff and the acceleration voltage is much lower than for the cr - controlled pffs .
this is due to the higher efficiency of ics in producing photons of pair - producing energy .
zqlh97 obtained a similar result for ics - controlled pffs using the electric field of rs95 .
electrons having lorentz factors of only @xmath249 , depending on @xmath244 , are capable of radiating photons of about @xmath250 to almost @xmath251 of their energy , and these photons will produce pairs .
again , the maximum lorentz factor , @xmath252 , is a very weak function of field strength .
the ics - controlled pff height , however , behaves very differently near the magnetic pole . instead of forming a slot gap , like the cr - controlled pffs
, the height of the pff actually decreases near the pole .
ics photons are radiated at much larger angles to the field and thus have shorter pair attenuation lengths than curvature photons .
the above determinations of the electron pff are not fully self - consistent in that they have neglected pairs produced by the positrons that slow down to screen @xmath14 and accelerate down toward the stellar surface .
even if the number of returning positrons is small compared to the number of primary electrons , the multiplicity of the downward cascades is high ( dh96 ) , due to the increasing field strength . from our simulations of downward going cascades using the code of dh96
, we find that the multiplicity of a positron accelerated toward the pc from a height of one stellar radius at the pc rim is @xmath253 . if the fraction of positrons that are accelerated downward from the electron pff is @xmath254 ( see equations [ [ lam- ] ] and [ [ lam+ ] ] ) , then there are @xmath255 pairs produced by downward cascades for each primary electron .
this would seem to be sufficient to screen @xmath14 and form a lower pff . to compute the location of the lower pff ,
a test positron is started at the upper ( electron ) pff with an energy of @xmath256 . in reality
, there is a spread in energies of the pairs created near the electron pff , so that the most energetic ones will decelerate but not turn - around , while the less energetic ones will decelerate before they reach the pff and accelerate downward with varying initial energies .
we have thus neglected the detailed kinematics of this turn - around process .
equation ( 2 ) is then integrated to follow the downward acceleration of the positron and its energy losses due to both ics and cr , in the same manner as for the electron . likewise ,
the attenuation lengths of the photons and the total acceleration length of the positron are computed .
although the pffs of the positrons , like the electrons , are produced by pairs from the ics process near the ns surface in the presence of thermal radiation from a hot pc , we find that the positron pff is always above the start of the electron acceleration .
that is , pair front formation by the ics process is not symmetric for upward and downward going particles .
this is because the positrons scatter the thermal photons at much larger angles than the electrons . while the ics photons producing the electron pff scatter in the cyclotron resonance to an energy of @xmath257 , photons producing the positron pff scatter above the cyclotron resonance , in the kn regime ( see fig . [
fig : prim1 ] ) , where the scattered energy is @xmath258 .
this asymmetery will occur for field strengths @xmath259 , above which positrons will form pffs through resonant ics .
as discussed further in section [ sec : analres ] , stable ics - controlled double pffs may exist for high - field pulsars because the scattered energy for resonant scattering is independent of incident angle .
figure [ fig : epsmin ] shows the difference between the electron , @xmath260 , and positron , @xmath261 , energies that produce scattered photons of energy @xmath262 and @xmath263 that form the minimum pffs , as a function of @xmath9 . since their ics is more inefficient , the electrons must accelerate to much higher energies in order to produce scattered photons that pair - produce .
the efficiency of electron scattering , @xmath264 , decreases with height , @xmath9 , above the surface , while the efficiency of positron scattering increases .
therefore , the positrons are able to radiate pair - producing photons after traveling a shorter acceleration path . since the scattered energies of the positrons , at a given height , are greater that those of the electrons , these photons have shorter pair attenuation lengths .
thus , the total positron acceleration length , @xmath265 is less than the total electron acceleration length , @xmath266 .
figure [ fig : pffep ] shows total positron and electron acceleration lengths , @xmath266 and @xmath267 as a function of the electron starting height @xmath9 , for both ics- and cr - controlled pffs . in these calculations , we have allowed the self - consistent pffs of electrons and positrons to form with only one of the radiation mechanisms operating .
when only ics produces pairs , @xmath266 is always significantly less than @xmath267 and the difference increases with height .
however , when ics is suppressed and only cr is allowed to operate , @xmath266 and @xmath267 are equal .
that is because cr for electrons and positrons is the same , the only difference being in radius of curvature between the radiation points , which is negligible since @xmath268 . while ics can not provide self - consistent solutions to the double pffs , cr can .
we suggest that stable , self - consistent double pffs can only exist when they are formed by cr . since ics will dominate near the surfaces of nss radiating thermal soft x - rays , evidence of which has been observed in many pulsars , stable pffs can only form at a height above the surface where cr becomes dominant .
we take a rough estimate for the height of cr - control of the electron and positron pffs to be where cr energy loss of the electrons dominates over ics energy loss .
if electrons of energy @xmath269 radiate more curvature photons than ics photons , then the ics pff will disappear , and the electrons will continue accelerating until a curvature pff is established . when the positrons accelerating downward from the electron curvature pff also establish curvature photon pffs , then stable acceleration can occur . in reality , the pffs will switch from ics to cr control when the number of pairs produced by ics photons becomes too small to screen @xmath14 . but to determine the number of pairs necessary for establishing a lower pff , we would need to know the number of returning positrons .
figure [ fig : gdot ] shows an example of the electron energy loss rates due to ics and cr at energy @xmath269 as a function of height of the lower pff . at low altitudes , ics losses dominate by many orders of magnitude , but decrease with @xmath9 mostly due to the decrease in the thermal photon density ( cf . equation [ [ n_ph ] ) .
the energy @xmath269 of electrons that are forming the ics pff increase with altitude due to the decreasing efficiency of ics .
their cr losses , which are a strong function of energy , therefore increase .
the height at which electron ics and cr losses are equal , which roughly set the location of stable acceleration , will depend on the pc thermal temperature and radius ( cf .
equation [ [ h_0 ] ] ) . we have computed the height of the stable acceleration zones where cr losses dominate control of pffs for various pulsar parameters .
as shown in figure [ fig : pffh0 ] , the location of the lower ( positron ) pffs @xmath9 are at higher altitudes for higher surface magnetic field strengths , because the ics losses are proportional to @xmath270 for resonant scattering .
the height of the lower pff decreases for increasing pulsar rotation period , @xmath53 , because at a given colatitude @xmath244 the radius of curvature increases , but the electron energy @xmath269 increases so that cr can dominate at a lower altitude .
figure [ fig : pffp ] shows the width of the cr - controlled acceleration zone , @xmath267 ( i.e. the distance between the lower and upper curvature pffs ) and the maximum voltage drop , @xmath243 , as a function of pulsar period and surface field strength .
the maximum particle energy is again remarkably insensitive to pulsar parameters , even though the acceleration zone width shows substantial variation .
the width @xmath267 tends to be smaller for larger fields because both the acceleration length and the pair attenuation length are shorter .
it has a large increase at long periods , where the pair attenuation lengths become large due to the increase in the pc radius of curvature .
in fact , there will be a maximum period for which electron pffs can form at a given colatitude .
figure [ fig : pffx ] shows solutions for the width and maximum lorentz factor of the self - consistent , stable acceleration zones that are controlled by cr double pffs .
the width of these zones , located nearly a stellar radius above the surface , are larger than cr - controlled zones at the surface ( e.g. fig .
[ fig : pcacch ] and [ fig : pffxb ] ) , because the magnetic field has fallen from its surface value , so that @xmath14 is lower , and the radius of curvature is larger .
however , the maximum lorentz factor of the higher altitude acceleration zones is roughly the same as that at the surface .
again , the size of the acceleration path has adjusted itself to maintain the same maximum energy , which depends only on the geometry of the magnetic field .
it is difficult to derive accurate analytic expressions for the acceleration zone parameters we have computed numerically , such as height and width of the stable accleration zone .
the form of the electric field changes as a function of acceleration length ( see appendix ) , starting as a linear function and then saturating at @xmath271 to become a constant . consequently
, there are no simple expressions for the processes that apply for altitudes from the surface to several stellar radii .
nevertheless , we can characterize the behavior of the solutions in different regions .
an estimate for @xmath9 , the height at which control of the pff formation switches from ics to cr , can be obtained by finding the height at which the loss rates for cr and ics of positrons with energy @xmath8 are equal : _ _
ic(_min ) = _ _ cr(_min ) . we can obtain an estimate of @xmath8 by solving equation ( [ sc ] ) . for @xmath272 , we can use the expression for @xmath14 given in equation ( a3 ) of the appendix , which is linear in @xmath273 and thus gives a quadratic energy increase in particle energy with @xmath13 .
the particle acceleration length is then : s_a ( ) = ( /a_)^1/2 where a _ ( 3e2mc^2 ) ( rc ) b_0 ( r_e ) ( rr_e)^2 .
note that we have rescaled equation ( a3 ) to the effective radius @xmath274 , at which the acceleration begins .
the pair attenuation length can be approximated as the mean free path s_p ( ) ( 0.2 _ c b ) valid for @xmath275 . equation ( [ sc ] ) may then be written , [ sceq ] s_c = min\ { ( /a_)^1/2 + 0.2_cb } , where _ c r_e^1/2(c)^1/2 and b ( b_0b_cr)(r_er)^-3 .
we wish to estimate @xmath8 when ics produces the pffs . for nearly all pulsar field strengths ( @xmath276 ) the photons which produce the positron pffs
have been blue - shifted above the resonance and scatter in the kn regime ( see fig .
[ fig : prim1 ] ) . in this case , we can approximate the scattered photon energy as @xmath277 . inserting this value of @xmath278 into equation ( [ sceq ] ) ,
we then obtain @xmath8 giving the minimum @xmath232 by setting the derivative of @xmath232 with respect to @xmath0 to zero : [ gmin_kn ] ^kn_min = ( 0.4_ca_^1/2b)^2/3 = 3.4 10 ^ 5b_12 ^ -1/3(r_er ) ( ) ^1/3 . where @xmath126 is the surface field strength .
thus , the particle energy required to radiate pair - producing photons increases with altitude , as was found in our numerical results ( cf .
[ fig : epsmin ] ) .
we can obtain an estimate for the width of the ics controlled acceleration zones by substituting @xmath279 from equation ( [ gmin_kn ] ) into equation ( [ sceq ] ) for the minimum value of @xmath232 , s^kn_c 5.3 10 ^ 3b_12 ^ -2/3 p^1/2r_6 ^ 1/2(r_er)^5/2()^-1/3 cm . we can estimate the height @xmath9 above the surface at which cr energy loss exceeds ics losses for the downward moving positrons of energy @xmath8 . from our numerical results
, we found that at the altitudes where cr begins to dominate the particle energy loss , the ics is in the k - n regime above the resonance .
we will therefore equate the curvature loss rate to the k - n loss rate to obtain an estimate for @xmath9 .
substituting @xmath8 from equation ( [ gmin_kn ] ) in equation([gdot_cr ] ) for the cr energy loss , we have _ _ cr(^kn_min ) = 1.6 10 ^ 4 b_12 ^ -4/3r_6 ^ -1p^-1(r_er)^3 ( ) ^4/3s^-1 which sharply increases with altitude , @xmath237 .
we approximate the k - n loss rate , modifying blumenthal & gould s ( 1970 ) formula for scattering of isotropic thermal photons of temperature @xmath280 , from a hot polar cap of radius @xmath119 , with a dilution factor that accounts for the fall - off in photon density with height @xmath281 : _ _ kn(^kn_min ) ( 4_min ktmc^2 ) r_t^2 h^2+r_t^2 7.6 10 ^ 8s^-1 t_6 ^ 2r_t^2h^2 for @xmath282 and ignoring the slowly varying @xmath283 factor . setting @xmath284 and assuming @xmath285 , [ h_0 ] ( h_0r)_kn 8.6b_12 ^ 4/15p^1/5r_6 ^ 1/5 t_6 ^ 2/5(r_tr)^2/5()^-4/15 .
although the above expression involves a number of approximations , it does seem to roughly reproduce our numerical results .
for example , taking the parameters @xmath286 , @xmath287 , @xmath288 , @xmath289 and @xmath290 gives @xmath291 , fairly close to the value shown in fig .
[ fig : pffp ] . at the altitude @xmath9 , where cr losses are dominant and
curvature photons produce the pffs , the width of the acceleration zone can be calculated in a similar manner to that of the ics controlled acceleration zones . using the cr critical frequency , @xmath292 in equation ( [ sceq ] )
, we can solve for the particle energy @xmath293 which produces the minimum acceleration zone width @xmath232 , [ gmin_cr ] ^cr_min = ( 6 c_p a_^1/2)^2/7= 4.710 ^ 7 b_12 ^ -1/7p^1/7 ( r_er ) ^4/7 r_6 ^ 1/7 ( ) ^1/7 , where c_p = ( 0.2_cb_cr ) .
then substituting the expression for @xmath293 back into equation ( [ sceq ] ) for @xmath232 , we have [ sc_cr ] s^cr_c 4.8 10 ^ 4b_12 ^ -4/7 p^4/7r_6 ^ 4/7(r_er)^16/7 ( ) ^-3/7 cm . the expression for @xmath293 above is also a very good estimate of @xmath294 , the maximum lorentz factor of the cr - controlled accleration zone , because @xmath295 for cr photons ( i.e. the pff is very close to where the first pairs are produced ) .
for very high field strengths ( @xmath296 ) both electron and positron ics operate in the resonant scattering regime .
the incident soft photons scatter in the cyclotron resonance and their scattered energies are approximately @xmath297 . in the thomson limit of resonant scattering ,
the energy of the scattered photons is independent of incident photon angle .
this means that @xmath298 , pff@xmath20 will coincide with @xmath9 , and stable ics - controlled acceleration zones are thus possible for high - field pulsars .
however , in the relatively high fields where resonant ics controls the positron pffs , relativistic effects on the resonant scattering cross section as well as photon splitting will become important ( as will be discussed in section [ sec : dis ] ) and our calculation of pffs is incomplete .
in this paper , we have investigated the effect of cascades from downward - accelerated positrons on the electrodynamics of the pc particle acceleration .
we find that when ics produces pairs in the acceleration zone , the positron cascades may screen the accelerating electric field and disrupt particle acceleration near the ns surface .
thus , if lower pffs can develop , the picture of steady particle acceleration from the pc surface must undergo major revision .
we suggest that a stable acceleration zone may exist at an altitude of about one stellar radius above the pc , in which pairs from cr limit the electrostatic acceleration of primaries by screening the electric field near the upper pff .
the calculations presented here are only a first attempt to describe the physics of what is a very complicated process .
we have made many assumptions and approximations to obtain our results .
while we believe that the gross qualitative results of our study are correct , there are a number of aspects which should be treated more accurately to achieve more solid quantitative results .
we have assumed that the screening of the @xmath14 occurs over the short distance determined by the upper boundary condition of poisson s equation , not by the dynamics of the pair screening .
the details of the pair screening also determine the fraction of pairs that return to the pc and we have briefly outlined in section ( [ sec : pfraction ] ) how such a calculation can be done within the electrodynamic framework set up in this paper .
a determination of the returning positron fraction would answer some interesting questions , such as whether pc heating is important relative to cooling in setting the pc temperature and if so , whether there is a feedback loop between the formation of the upper pff and the pc heating .
the characteristics of the downward cascades need further study and modeling .
while we have done preliminary modeling of these cascades near the pc rim , there will be significant variation in pair yields with magnetic colatitude .
near the magnetic pole , the pair multiplicity will drop , allowing the lower pff height to decrease or even disappear .
we have found that the nature of the pc acceleration depends strongly on the characteristics of the radiation from the hot pc .
thus , to solidify the quantitative aspects of our results , it is important to treat the thermal pc radiation and the ics process as accurately as possible .
for example , we have assumed in this paper , that the thermal radiation is uniformly emitted over a pc of size @xmath299 with an isotropic flux distribution .
however , studies of thermal radiation propagating through a strongly magnetized ns atmosphere ( pavlov et al . 1994 ) will not be isotropic , due to the anisotropy of the magnetized scattering cross section .
the expected radiation pattern consists of a pencil component , beamed along the magnetic field and a fan component perpendicular to the field .
such a beam pattern has been found to be consistent with observed thermal x - rays pulse fractions and pulse profiles for several pulsars ( shibanov et al .
1995 , hm98 ) .
in addition , our treatment of ics uses a combination of the ( non - relativistic ) magnetized cross section in the thomson limit for scattering near the fundamental cyclotron resonance , and the ( non - magnetic ) klein - nishina cross section to describe relativistic effects above the resonance . while this hybrid treatment is somewhat inaccurate for @xmath300 and therefore not completely satisfactory , the fully relativistic qed magnetic scattering cross section ( e.g. dh86 ) is too complicated for use in this type of calculation .
in particular , the quantization of the electron momentum perpendicular to the field limits the number of landau states contributing to the cross section for each incident photon energy . the magnetic qed cross section for scattering just above the fundamental cyclotron resonance , and therefore the loss rate , could therefore be substantially lower than the klein - nishina cross section we have used here .
unfortunately , there exists no simplified , approximate expression for the qed scattering cross section , which smoothly bridges , and allows a unified treatment of , the relativistic resonant and non - resonant regimes of ics . a physical process which we have neglected in this study , photon splitting , is not expected to be significant for the magnetic fields we have considered , but will be very important for pulsars having surface @xmath301 .
photon splitting , a third - order qed process in which one photon splits into two , operates only in very high magnetic fields and competes as an attenuation process with one - photon pair production ( harding et al .
1997 ) because it can occur below pair threshold .
the implications of photon splitting for pc pffs is profound .
pulsars having @xmath301 will produce fewer pairs , especially near the surface , so that the cascades from downward - moving positrons may not produce a lower pff .
this is about the same field strength where we found that double pffs controlled by resonant ics at the ns surface become possible .
thus , the stable , cr - controlled double pff structure we have studied in this paper , which we found to move to higher altitudes with increasing surface field strength , will eventually collapse back to the surface at very high @xmath39 .
a single pff will then form , controlled by ics , and the acceleration zone will have characteristics similar to that shown in figure [ fig : pffsx ] . at extremely high surface fields
, @xmath302 photon splitting will suppress pair creation completely at this single pff , and the pulsar may be radio quiet ( baring & harding 1998 ) .
however , the particle acceleration in these pulsars will operate very efficiently , free of any screening of @xmath14 , so they are expected to be observable at high energies .
bound - state pair production , where photons convert to positronium just below pair threshold rather than converting to free pairs ( usov & melrose 1995 ) , may also come into play at higher field strengths .
some of the main results that we have presented in this paper have some important implications for pulsar high - energy emission .
one of these results is the insensitivity of the acceleration voltage ( maximum particle energy ) @xmath243 to any pulsar parameters such as period , surface magnetic field strength , obliquity ( except for nearly orthogonal rotators ) , and even height .
this result showed up many times in the course of our calculations ( cf . figs .
[ fig : pcacch ] , [ fig : pffxb ] , [ fig : pffsx ] , [ fig : pffp ] and [ fig : pffx ] , and equation [ [ gmin_cr ] ] ) , and seems to be a robust characteristic of this type of pc acceleration model .
the maximum particle acceleration energy varies only within each pulsar , as a function of magnetic colatitude .
this energy , between @xmath303 ev and @xmath304 ev for cr - controlled acceleration zones , is about two to three times higher than that without frame - dragging ( e.g. a83 ) , and is consistent with the primary particle energy required in cr - initiated pc cascade models of @xmath0-ray pulsars ( e.g. dh96 ) .
the acceleration energy is very high near the magnetic poles , where the radius of curvature goes to infinity , allowing for the possibility of a narrowly beamed , hard component in @xmath0-ray pulses .
however , the radiation power emitted by these high - energy particles will be small , because the curvature radiation loss rate is proportional to @xmath305 .
the insensitivity of @xmath243 to pulsar parameters implies that the primary particles in all pulsars are accelerated to the same energy , and that the luminosity of the high - energy emission should depend only of the flux of primary particles .
this is consistent with trends in the observed @xmath0-ray pulsar luminosities ( thompson et al .
the @xmath0-rays will originate within a stellar radius of the upper stable pff , at height @xmath306 . from our results of figs .
[ fig : pffh0 ] and [ fig : pffp ] , and equations ( [ h_0 ] ) and ( [ sc_cr ] ) , the height of the @xmath0-ray production increases with period , roughly as @xmath307 .
the standard pc half - angle at height @xmath308 will be @xmath309 .
thus , the @xmath0-ray emission solid angle @xmath310 , which is expected to be , @xmath311 $ ] , will be very weakly dependent on period and field strength .
our conclusion that stable acceleration may occur in most pulsars at some altitude above the surface will have consequences not only for high - energy emission , but for radio emission as well .
if electron - positron pairs are necessary for coherent radio emission , then the dependence of the pff altitude on pulsar parameters should be taken into account when determining the radio pulsar death line " , the line on the period - period derivative diagram beyond which pulsars are incapable of producing pairs .
our calulcations in this paper suggest that pulsars with long periods do not produce pffs , and that there will be a death line " at periods below where there are observed radio pulsars .
but this is a long - standing problem of pc acceleration , most recently discussed by arons ( 1998 ) , and one that should be addressed in future studies .
our principal findings can be summarized as follows . 1 .
lower pffs may form by positrons returning to the ns surface from the upper pff .
pair creation by the ics process dominates near the stellar surface , but is not symmetric for upward and downward going particles , so that stable , double pff formation is very unlikely .
stable , self - consistent double pffs can only exist when they are formed by cr .
they can only form at a height above the surface where cr becomes dominant .
the maximum particle energy is insensitive to any pulsar parameters such as period , surface magnetic field strength , obliquity , even height , and is sufficient to power @xmath0-ray pulsars .
the main conclusion of this paper is that the cascades from positrons returning to the pc may have a significant effect on the primary particle acceleration in pulsars and should not be neglected .
it is possible that lower pffs do not form for all pulsars , and may not form over the entire pc .
a detailed study of the screening of the accelerating electric field by the returning positron cascades is beyond the scope of this paper , but will ultimately be necessary to understand pc acceleration .
these studies are needed to address the questions of the returning positron fraction and the multiplicity of downward cascades .
if we can show that the returning positrons do not screen @xmath14 , then the present assumption of acceleration right from the ns surface is valid .
but if these studies show that screening at a lower pff is effective , then the possibility of acceleration above the ns surface must be incorporated in pc models .
we thank the referee bing zhang for his very careful review and insightful comments .
we are also grateful to joe daugherty , for help in the downward cascade simulations , to steve sturner for discussions on inverse - compton scattering and matthew baring for comments on the manuscript .
here we present the explicit expressions for the electric field component parallel to the magnetic field for the various ranges of altitude .
these expressions have not been rescaled to allow for acceleration starting at an effective radius above the stellar surface , so that @xmath312 is the altitude above the surface .
+ @xmath313 , and @xmath314 : @xmath315 , \eqno ( a1)\ ] ] @xmath316 , and @xmath317 : @xmath318 , \eqno ( a2)\ ] ] @xmath319 , and @xmath320 : @xmath321 , \eqno ( a3)\ ] ] @xmath322 and @xmath323 : @xmath324 \cos\chi \right .
\nonumber \\ & & + { 1\over 4 } \theta_0 \left [ f(1)\over f(\eta ) \right ] ^{1/2}\ , f(\eta ) { 1 \over { \eta ^{1/2 } } } \left [ \xi(1-\xi^2 ) - { \eta_c \over \eta } { { f(\eta_c ) } \over { f(\eta ) } } \right .
\nonumber \\ & & \left .
\sum_{i=1}^{\infty}\,{16 j_1(\tilde k_i\xi)\over \tilde k_i^3\,j_2(\tilde k_i)}\ , e^{-\tilde\gamma_i(\eta_c)(\eta_c-\eta)}\right ] \sin\chi\cos\phi\right\ } , \nonumber \end{aligned}\ ] ] @xmath325 + which translates into @xmath326 , \eqno ( a5)\ ] ] when @xmath327 . + here @xmath328 , \nonumber\end{aligned}\ ] ] @xmath329 ^{1/2}.\ ] ] in the above expressions @xmath330 is the pc radius , @xmath331 is the dimensionless altitude of the upper boundary ( pff ) , and all other quantities are scaled with the true stellar radius .
note also that @xmath332 .
arons , j. 1981 , in iau symp .
95 , pulsars , ed .
w. sieber & wielebinski ( dordrecht : reidel ) , 69 arons , j. 1983 , , 266 , 215 : a83 arons , j. 1998 , in neutron stars and pulsars , eds .
n. shibazaki , n. kawai , s. shibata & t. kifune ( university academy press : tokyo ) , p. 339 .
arons , j. , & scharlemann , e. t. 1979 , , 231 , 854 : as79 baring , m. g. & harding , a. k. 1998 , in preparation .
becker , w. , & trmper , j. 1997 , , 326 , 682 bussard , r. w. , alexander , s. b. , & meszaros , p. 1986
, 34 , 440 canuto , v. , lodenquai , j. , & ruderman , m. 1971 , _ phys .
, 3 , 2303 daugherty , j. k. , & harding , a. k. 1983 , , 273 , 761 : dh83 daugherty , j. k. , & harding , a. k. 1986 , , 309 , 362 : dh86 daugherty , j. k. , & harding , a. k. 1989 , , 336 , 861 .
daugherty , j. k. , & harding , a. k. 1996 , , 458 , 278 : dh96 dermer , c. 1990 , , 360 , 197 : d90 deutsch , a. 1955 , ann .
dap . , 18 , 1 fawley , w. m. , arons , j. , & scharlemann , e. t. 1977 , , 164 , 529 goldreich , p. , & julian , w. h. , 1969 , , 157 , 869 gonthier , p. l. & harding , a. k. , 1994 , , 425 , 767 harding , a. k. , baring , m. g. & gonthier , p. g. 1997 , , 476 , 246 .
harding , a. k. , & muslimov , a. g. 1998 , in neutron stars and pulsars , eds .
n. shibazaki , n. kawai , s. shibata & t. kifune ( university academy press : tokyo ) , p. 311 .
( astro - ph/9802038 ) harding , a. k. , & muslimov , a. g. 1998 , , in press .
( astro - ph/9802044 ) jones , p. b. 1978 , , 184 , 807 jones , p. b. 1985 , phys .
lett . , 55 , 1338 jones , p. b. 1986 , , 218 , 477 kardashv , n. s. , mitrofanov , i. g. , & novikov , i. d. 1984 , sov .
astron . , 28 , 651 lorentz , c. p. , ravenhall , d. g. , & pethick , c. j. 1993 , phys . rev . lett . , 70 , 379 mestel , l. 1998 , stellar magnetism ( oxford : oxford university press ) muslimov , a. g. , & harding , a. k. 1997 , , 485 , 735 : mh97 muslimov , a. g. , & tsygan , a. i. 1990 , azh , 67 , 263 : mt90 muslimov , a. g. , & tsygan , a. i. 1992 , , 255 , 61 : mt92 neuhauser , d. , langanke , k. , & koonin , s. e. 1986 , phys .
, a33 , 2084 neuhauser , d. , koonin , s. e. , & langanke , k. 1987 , phys .
, a36 , 4163 .
pavlov , g. g. , shibanov , yu a. , ventura , j. & zavlin , v. e. 1974 , a & a , 289 , 837 .
ravenhall , d. g. , & pethick , c. j. 1994 , , 424 , 846 ruderman , m. a. , & sutherland , p. g. 1975 , , 196 , 51 : rs75 shibanov , yu . a. , zavlin , v. e. , pavlov , g. g. & ventura , j. 1995 , in 17th texas symposium on relativistic astrophysics , eds . h. bohringer , g. morfill & j. trumper , annals of ny acad .
, 759 , p. 291 .
sturner , s. j. 1995 , , 446 , 292 : s95 sturner , s. j. & dermer , c. d. 1994 , , 420 , l79 .
sturrock , p. a. 1971 , , 164 , 529 thompson , d. j. , harding , a.k . , hersen , w. & ulmer , m.p .
1997 , in proc . of the 4th compton symposium , ed .
dermer , m.s .
strickman & j.d .
kurfess ( aip 410 : new york ) , 39 .
tsai , w .- y .
& erber , t. 1974 , phys .
d , 10 , 492 .
usov , v.v . & melrose , d. b. 1995 , aust .
j. phys . , 48 , 571 .
wang , f. y .- h . , ruderman , m. , halpern , j. p. , & zhu , t. 1998 , , in press xia , x. y. et al .
1985 , a&a , 152 , 93 .
zavlin , s.e . , shibanov , yu . a. & pavlov , g.g .
1995 , astron .
lett . , 21 , 149 .
zhang , b. , & qiao , g. j. 1996 , a&a , 310 , 135 zhang , b. , qiao , g. j. , liu , w. p. , & han , j. l. 1997 , , 478 , 313 : zqlh97 | we investigate self - consistent particle acceleration near a pulsar polar cap ( pc ) by the electrostatic field due to the effect of inertial frame dragging .
test particles gain energy from the electric field parallel to the open magnetic field lines and lose energy by both curvature radiation ( cr ) and resonant and non - resonant inverse compton scattering ( ics ) with soft thermal x - rays from the neutron star ( ns ) surface .
gamma - rays radiated by electrons accelerated from the stellar surface produce pairs in the strong magnetic field , which screen the electric field beyond a pair formation front ( pff ) .
some of the created positrons can be accelerated back toward the surface and produce @xmath0-rays and pairs that create another pff above the surface .
we find that ics photons control pff formation near the surface , but due to the different angles at which the electron and positron scatter the soft photons , positron initiated cascades develop above the surface and screen the accelerating electric field .
stable acceleration from the ns surface is therefore not possible in the presence of dominant ics energy losses .
however , we find that stable acceleration zones may occur at some distance above the surface , where cr dominates the electron and positron energy losses , and there is up - down symmetry between the electron and positron pffs .
we examine the dependence of cr - controlled acceleration zone voltage , width and height above the surface on parameters of the pulsar and its soft x - ray emission . for most pulsars
, we find that acceleration will start at a height of 0.5 - 1 stellar radii above the ns surface .
= cmr5 = cmbx5 = cmr6 = cmbx6 = cmr7 = cmbx7 # 1@xmath1 # 1#2 |
in scale - invariant systems , a state at one density can be expressed through a state at another , via trivial rescaling of space .
scale invariance is always associated with an inability of the interaction potential to introduce a distinct length scale .
several examples emerged recently in the physics of quantum gases . in three dimensions ,
the @xmath2-interaction with infinite coupling strength , even when properly regularized , ensures the scale invariance of the unitary gases @xcite . in 2d , the unregularized @xmath2-potential , for any coupling constant , induces the scale invariance of two - dimensional bose @xcite and spin-@xmath3 fermi @xcite gases at the classical field level , which is however broken by quantization @xcite . finally , in 1d , we have the tonks - girardeau ( tg ) gas @xcite a one - dimensional quantum bose gas with an infinite strength @xmath2-interaction which is the subject of this article .
scale invariance enables a robust frequency gauge : when a scale - invariant gas is placed in a symmetric harmonic trap of frequency @xmath4 and a monopole oscillation is induced , the signal shows neither damping nor amplitude - dependent frequency shifts its frequency is fixed to @xmath5 , for all scale invariant systems and for all spatial dimensions @xcite . as a consequence ,
monopole excitations in scale - invariant systems are very sensitive to changes in the equation of state @xcite , whether produced by a quantum anomaly @xcite , by an influence of the confining dimension @xcite , or just by a small shift in the coupling constant away from the scale - invariant point . in this article
, we study the effect of a small deviation from the tgpoint on the frequency of the monopole excitations of a one - dimensional harmonically trapped bose gas @xcite , both for microscopically small and for microscopically large but macroscopically small excitation amplitudes .
our results also directly apply to the 3d @xmath1-wave - interacting waveguide - confined fermions @xcite , thanks to the mapping by granger and blume @xcite .
our object of study is the system of @xmath6 bosons of mass @xmath7 with contact interaction in a 1d harmonic trap of frequency @xmath4 . in the second quantized form ,
the hamiltonian reads @xmath8 & \qquad\qquad \left .
+ \frac{m\omega^2}{2 } x^2 \hat{\psi}_{b}^{\dagger}\hat{\psi}_{b}^ { } + \frac{g_{\mbox{\scriptsize 1d}}}{2 } \hat{\psi}_{b}^{\dagger } \hat{\psi}_{b}^{\dagger } \hat{\psi}_{b}^ { } \hat{\psi}_{b}^ { } \right\ } \ , .
\end{split } \label{1 m } \ ] ] here @xmath9 is the one - dimensional coupling constant , and @xmath10 is the bosonic quantum field . in this article , we will be interested in the monopole excitations in the vicinity of the tglimit , @xmath11 .
this regime is both easy and difficult to work in . on one hand ,
right at the limit , the model maps to free fermions , via the fermi - bose map by girardeau @xcite .
on the other hand , away from the limit even remaining infinitesimally close to it there exist conceptual difficulties in interpreting the resulting system as one governed by a hamiltonian for free fermions plus a small correction @xcite . nevertheless , an effective fermionic hamiltonian , @xmath12 & \left .
+ \frac{m\omega^2}{2 } x^2 \hat{\psi}_{f}^{\dagger}\hat{\psi}_{f}^ { } - \frac{2\hbar^4}{m^2 g_{\mbox{\scriptsize 1d } } } ( \partial_x\hat{\psi}_{f}^{\dagger } ) \hat{\psi}_{f}^{\dagger } \hat{\psi}_{f}^ { } ( \partial_x\hat{\psi}_{f}^ { } ) \right\ } \end{split } \label{4 m } \ , , \ ] ] @xcite can be proven @xcite to produce the correct eigenspectrum if used as the kernel of a variational energy functional .
one can further show that in this case , the first order of the perturbation theory with the quartic term in ( [ 4 m ] ) as a perturbation produces the correct @xmath13 correction to the eigenenergies . here , @xmath14 is the fermionic quantum field .
the fermionic field in eq .
( 2 ) can be expanded onto a series over the eigenstates of the harmonic trap : @xmath15^{1/2 } e^{-x^2/(2 \ell^2 ) } h_n ( x/\ell)$ ] where @xmath16 is the @xmath17-th hermite polynomial , and @xmath18 .
the operator @xmath19 is the fermionic annihilation operator that removes one particle from the @xmath20-th eigenstate .
the operators @xmath19 obey the standard fermionic commutation relations and the hamiltonian in fock space is of the form @xmath21 where @xmath22 .
the ground state of whole system is @xmath23 where @xmath24 stands for the vacuum with no particle at all .
the energy correction of ground state is analyzed in appendix a , where we will show our result recovers the formula in ref .
@xcite .
now we will come to the @xmath25 excitations .
the unperturbed manifold of energy @xmath26 is of two - fold degeneracy .
the set of unperturbed eigenstates is @xmath27 .
based on the perturbation theory , corrections to the energies are represented by the spectrum of the @xmath28 matrix of the perturbation term in the space spanned by the members of the manifold @xmath29
\omega_n & i_n^{(2b)}\\ \end{pmatrix } \end{split } \label{7tris}\ ] ] with @xmath30 , @xmath31 , and @xmath32 .
hence the transition frequencies for microscopically small amplitude read @xmath33 & \quad\pm\sqrt{\big(i_n^{(2a)}-i_n^{(2b)}\big)^2 + 4\omega_n^2}\ \bigg ] + { \cal o}(1/(g_{\mbox{\scriptsize 1d}})^2 ) \end{split } \label{8tris}\ ] ] where @xmath34 and @xmath35 takes the form of @xmath36 & \times{}_3 f_2\left[\begin{matrix}\frac{3}{2},-n ,- m\\[0.08 cm ] \frac{3}{2}-n,\frac{3}{2}-m\end{matrix};1\right ] \end{split } \label{16b}\ ] ] after a lengthy but straightforward calculation ( see the appendices ) , we obtain the analytical form of @xmath13 corrections to the relevant transition frequencies : @xmath37 @xmath38 & \left . \times { } _ 3 f_2\left[\begin{matrix}\frac{3}{2},1-n ,- n\\[0.08 cm ] \frac{7}{2}-n,\frac{1}{2}-n\end{matrix};1\right ] \frac{1}{\gamma_{0}(n ) } + { \cal o}(\frac{1}{\gamma_{0}^2(n ) } ) \right ) \hbar\omega \ , , \label{13 m } \ ] ] where @xmath39 , are the transition frequencies , @xmath40 is the ground state energy , and @xmath41 are the energies of the states that , in the strict tglimit , form a two - fold degenerate manifold , @xmath42 above the ground state .
the effective lieb - liniger parameter @xmath43 @xcite uses the tg(i.e .
@xmath11 ) density in the center of the trap , @xmath44 , instead of the true density . here , @xmath45 $ ] is the generalized hypergeometric function of order @xmath46
. the interpretation of the @xmath47 eigenstates can be inferred from the corresponding transition frequencies .
the first one ( @xmath48 ) is the second state of an infinite @xmath49-spaced `` dipole '' ladder : coherent wave packets formed out of the members of the ladder represent finite amplitude dipole excitations ; their frequency @xmath50 is equal to the frequency of the trap _ exactly _ , interactions notwithstanding @xcite .
the first state of the ladder is analyzed in appendix b. the zeroth state is the ground state .
the second eigenstate ( @xmath51 ) in the @xmath52 manifold , is the first ( ground state being the zeroth ) step in the `` monopole '' ladder , that corresponds to the breathing excitations of frequency @xmath53 . in the noninteracting case , the ladder ( exactly @xmath54-spaced )
can be obtained by a recurring application of the creation operator @xmath55 of an appropriate @xmath56 group to the ground state @xcite .
the excitation dynamics consists of a periodic scaling transformation of frequency @xmath57 .
the existence of this structure is a direct consequence of the scale invariance of the tggas and its free - fermionic counterpart in the harmonic potential .
a deviation from the tglimit ( and the corresponding fermion - fermion interactions ( [ 4 m ] ) ) _ breaks the scale invariance weakly_. the goal of this article is to assess the impact that this effect has on the excitations of a microscopic amplitude and compare it to the corresponding predictions for the microscopically large but macroscopically small excitations . as far as the microscopic amplitude excitations are concerned , our program is already fulfilled .
indeed , a linear combination of the ground state and a small admixture of the state @xmath51 _ is _ is already a small amplitude monopole excitation .
its frequency is given by the formula ( [ 13 m ] ) that _ constitutes the central result of this article_.
we verified that for two atoms ( @xmath58 ) , the formula ( [ 13 m ] ) for the frequency of the small amplitude monopole excitations coincides with the known exact results @xcite . in the three - body case ( @xmath59 )
we perform a diffusion monte carlo simulation of the imaginary time evolution and extract the @xmath60 transition frequency from the inverse laplace transform components of the imaginary - time dynamic structure factor . in fig .
[ fig.3 ] we compare our ( non - perturbative ) numerical three - body results with the perturbative prediction ( [ 13 m ] ) . in fig .
[ fig.2 ] , the dominant corrections to the monopole frequency for @xmath58 and @xmath59 , extracted from the non - perturbative data , are also compared to formula ( [ 13 m ] ) .
the frequency of the monopole excitations of a microscopically small amplitude can be bounded from above using the sum rules @xcite ( see fig . [ fig.1 ] ) .
the order @xmath13 correction to this bound can also be computed analytically , for large atom numbers @xcite : @xmath61 we conjecture that the upper bound ( [ 14 m ] ) actually equals the exact prediction ( [ 13 m ] ) in the limit of large @xmath6 . to test this conjecture
, we multiplied the @xmath62-term in the the bound ( [ 14 m ] ) by @xmath63 , with @xmath64 and @xmath65 being free parameters to be used to fit the @xmath62 term in the series ( [ 13 m ] ) .
indeed we found the values that support our conjecture , namely @xmath66 and @xmath67 .
one - dimensional bosons in a harmonic trap , as a function of the inverse of the effective lieb - liniger parameter @xmath68 ( see text ) .
solid line ( red online ) : the prediction of the formula ( [ 13 m ] ) for the first two terms of the expansion of frequency in powers of @xmath69 .
open squares ( blue online ) : the ab initio diffusion monte carlo simulation . ] , to the result predicted by the scale invariance , @xmath70 . filled squares ( red online ) : the analytic formula ( [ 13 m ] ) .
open circle ( blue online ) : the exact nonperturbative solution for @xmath58 @xcite .
open square ( blue online ) : the diffusion monte carlo simulation for @xmath59 . dotted line ( purple online ) : the @xmath71 limit of the sum - rule prediction @xcite ( also eq .
( [ 14 m ] ) ) . ] . solid line ( red online ) : the formula ( [ 13 m ] ) .
dotted line ( purple online ) : the sum - rule upper bound @xcite ( courtesy of chiara menotti and sandro stringari ) . filled squares ( green online ) : the numerically exact hydrodynamic simulation of the motion of a macroscopic motion of small amplitude . dashed line ( blue online ) : the hydrodynamic perturbation theory for the latter . for definition of @xmath68 , see text . ]
the monopole frequencies obtained above correspond to excitations of _ microscopically small amplitude _ : there the many - body energy of the excited atomic cloud is only a few one - body harmonic quanta above the ground state energy . _ a priori _ it is not obvious if the microscopic predictions will remain valid for _ microscopically large but macroscopically small _ excitations , whose _ spatial amplitude is smaller than but comparable to the size of the cloud_. to compare the two frequencies , we investigate the time dynamics using the hydrodynamic equations ( see e.g. eqs . ( 1 ) and ( 2 ) of ref .
we use the well - known thermodynamic limit for the dependence of the zero - temperature chemical potential @xmath72 on the one - dimensional particle density @xmath20 , for a uniform one - dimensional @xmath2-interacting bose gas ; this equation of state was obtained by lieb and liniger , using bethe ansatz @xcite .
we propagate the hydrodynamic equations numerically . to excite the monopole mode
, we quench the trapping frequency . fig .
[ fig.1 ] shows a good agreement with the large-@xmath6 asymptotics for the frequency of the microscopically small excitations ( [ 14 m ] ) . in order to obtain an analytic expression for the frequency shift ,
we apply the perturbation theory developed by pitaevskii and stringari in ref .
@xcite ( with more technical details worked out in ref .
@xcite ) for the purpose of computing an analytic expression for the dominant beyond - mean - field correction to the monopole frequency of a bec . here
we use the tgequation of state ( eos ) , @xmath73 as the unperturbed eos , and the first - order ( in @xmath74 ) correction to the eos , @xmath75 , as a perturbation .
the function @xmath76 is the so - called lieb - liniger parameter @xcite . while most of the outlined steps of the study in @xcite are universally applicable to any eos , the boundary conditions for the density mode functions @xmath77 at the edge of the atomic cloud @xmath78 are typically dictated by the specific physical properties of the system at hand .
( here @xmath79 is the thomas - fermi radius . ) in the tgcase , with or without further beyond - the - tgcorrections to the eos , those are given by @xmath80 indeed , following the analysis developed in ref .
@xcite , one can show ( i ) that the first two terms in ( [ bc ] ) correspond to the near - edge asymptotic of the two linearly independent solutions of the mode equation , and ( ii ) that when rewritten in lagrange form @xcite , the solutions that violate condition ( [ bc ] ) lead to the appearance of crossing particle trajectories , incompatible with hydrodynamics .
to our surprise , we found that the macroscopic perturbation theory leads to a frequency shift that is @xmath0 times greater in magnitude than its microscopic counterpart eq .
( [ 14 m ] ) ( see fig . [ fig.1 ] ) .
this is definitely an artifact of the perturbative treatment of the macroscopic theory rather than of the macroscopic theory per se .
indeed , our macroscopic nonperturbative numerical results are consistent with the microscopic theory .
we attribute the failure of the perturbation theory to the divergence of the spatial derivative of the steady - state density at the edge of the cloud : in a monopole excitation this will lead to an infinite time derivative of the density itself , possibly invalidating the perturbation theory . in the same plot , we also present the sum - rule bound @xcite . at weak fermion - fermion interactions
, it reproduces well the perturbative prediction eq .
( [ 14 m ] ) .
in this article , we obtained an analytic expression , eq .
( [ 13 m ] ) for the leading behavior of the deviation of the frequency of the microscopically small monopole excitations of a strongly - interacting one - dimensional bose gas from the value predicted by the scale invariance in the tglimit .
we further compare this prediction with ( a ) the known non - perturbative analytic expressions for two atoms @xcite and to ( b ) the diffusion monte carlo predictions for three atoms .
for large numbers of atoms , the prediction in eq .
( [ 13 m ] ) stands in excellent agreement with ( c@xmath81 the sum - rule bound ( [ 14 m ] ) @xcite .
it was not _ a priori _ obvious to us if our formula will also apply to _ microscopically large but macroscopically small _ excitations : they correspond to a large number of atoms ( still covered by the formula ( [ 13 m ] ) ) , and have a macroscopic magnitude ( that is formally beyond the scope of eq .
( [ 13 m ] ) ) .
we found that ( d ) the numerically propagated hydrodynamic equations produce the same leading - order frequency correction as the large-@xmath6 limit of eq .
( [ 13 m ] ) .
finally , we find that ( e ) the hydrodynamic perturbation theory , which was so successful in predicting the beyond - mean - field corrections to the monopole frequency in both the three - dimensional @xcite and the two - dimensional @xcite bose gases , fails to predict the analogous beyond - tgcorrection in our case : the hydrodynamic perturbative prediction turns out to be approximately @xmath0 higher than the ab initio numerical value it was designed to approximate .
we conjecture that the sharp boundary of the tgcloud , characterized by an infinite density gradient , renders the perturbation theory inapplicable .
experimentally , the monopole excitation frequency of the lieb - liniger gas has been already studied , in ref .
@xcite . in the range of parameters
our article is devoted to , the beyond - scale - invariance shifts are too small to be reliably compared with the experimental data .
however , we plan to extend our study of the frequency of microscopically large but macroscopically small monopole excitations to the whole range of the interaction strengths .
one can already observe that in the intermediate range , the experimental frequencies @xcite depart from the sum - rule upper bound @xcite .
it appears to be of interest to verify that the numerically propagated hydrodynamic equations can reproduce the experimental points .
a study of the _ finite amplitude _ beyond - tgcorrections to the monopole frequency may be of interest .
another possible direction is computing the higher orders of the perturbation theory for the frequency correction .
this step is challenging , however : the odd - wave fermion - fermion interaction potential in eq .
( [ 4 m ] ) can not be used as such , and it requires a prior regularization @xcite , similar to the fermi - huang regularization of the three - dimensional @xmath2-potential .
results of our work directly apply to another system : the spin - polarized @xmath1-wave - interacting fermions in a wave - guide @xcite .
the mapping between this system and the lieb - liniger gas of @xmath82-interacting bosons is provided by granger and blume , in the final formula of ref .
@xcite . in the case of @xmath83 atoms
, the @xmath1-wave scattering volume @xmath84 can be controlled at will , using an accessible @xmath85 , @xmath86 feshbach resonance at @xmath87 @xcite .
when atoms are confined to a one - dimensional harmonic waveguide , the position of the resonance is further shifted @xcite due to the presence of a confinement - induced resonance ( cir ) @xcite .
for example , an ensemble of @xmath88 @xmath83 atoms , transversally frozen to a harmonic waveguide of a confining frequency of @xmath89 and longitudinally trapped by a harmonic potential of frequency @xmath90 , will show a @xmath91 @xmath1-wave - interaction - induced shift of the monopole frequency , for the @xmath1-wave scattering volume of @xmath92 ( with the cir value situated at @xmath93 ) . to relate this value of the scattering volume to the detuning from the feshbach resonance , note that the above value of the scattering volume would correspond to a binding energy of the three - dimensional @xmath1-wave dimers @xcite of @xmath94 ; the later value occurs if the magnetic field is detuned by @xmath95 below the @xmath85 , @xmath86 resonance ( see a the caption to fig . 2 in @xcite for the slope of the binding energy vs. magnetic field curve ) .
in general , the slope of the dimer energy as a function of the magnetic field is measured to be @xmath96 . here
, @xmath97 is the reduced mass , and @xmath98 is plank s constant .
we acknowledge support from the institut francilien de recherche sur les atomes froids ( ifraf ) .
gea acknowledges financial support from by mec ( spain ) through the ramon y cajal fellowship program , dgi ( spain ) grant no .
fis2011 - 25275 and generalitat de catalunya grant no .
2009sgr-1003 . zdz and tb were supported by the nsf grant ( phy-0968905 ) .
mo was supported by grants from the office of naval research ( n00014 - 12 - 1 - 0400 ) and the nsf grant ( phy-1019197 ) .
laboratoire de physique des lasers is umr 7538 of cnrs and paris 13 university .
the ground state of the unperturbed system is @xmath99 where @xmath100 is the vacuum with no particles at all .
the first order perturbation theory correction to ground state energy @xmath101 is given by @xmath102 & \times\frac{\gamma\left(n-\frac{1}{2}\right)}{\gamma\left(n+1\right)}\ { } _ 3 f_2\left[\begin{matrix}\frac{3}{2},-n ,-
m\\[0.08 cm ] \frac{3}{2}-n,\frac{3}{2}-m\end{matrix};1\right ] \end{split } \label{6}\ ] ] and @xmath45 $ ] is the order @xmath46 generalized hypergeometric function .
this result , as well as its derivation , is essentially identical to the formula obtained in ref .
@xcite .
there is only one eigenstate , @xmath103 , in the first excited state manifold , and thus the correction to the energy is @xmath104 .
the formula for the transition frequency , @xmath105 , assumes a compact form , and it reads @xmath106 we interpret the state @xmath107 as the first state of an infinite @xmath49-spaced `` dipole '' ladder : coherent wave packets formed out of the members of the ladder represent finite - amplitude dipole excitations ( _ i.e. _ oscillations of the center of mass ) ; their frequency @xmath50 is equal to the frequency of the trap _ exactly _ , interactions notwithstanding @xcite .
the zeroth state of the ladder is the ground state .
@xmath109 \frac{7}{2}-n,\frac{3}{2}-n\end{matrix};1\right]-\frac{\left(n-\frac{1}{2}\right)\left(n-\frac{5}{2}\right)}{(n+1)(n-1)}{}_3 f_2\left[\begin{matrix}\frac{3}{2},1-n ,-
n-1\\[0.08 cm ] \frac{5}{2}-n,\frac{1}{2}-n\end{matrix};1\right]\bigg\}\\[0.24 cm ] & = \sqrt{\frac{72}{\pi^3}}\frac{\gamma\left(n-\frac{5}{2}\right)\left(n+\frac{1}{2}\right)}{\gamma(n)\gamma(n+2)}{}_3 f_2\left[\begin{matrix}\frac{3}{2},1-n ,- n\\[0.08 cm ] \frac{7}{2}-n,\frac{1}{2}-n\end{matrix};1\right ] \end{aligned } \label{10}\ ] ] which leads to @xmath114 & \qquad\times { } _ 3 f_2\left[\begin{matrix}\frac{3}{2},1-n ,- n\\[0.08 cm ] \frac{7}{2}-n,\frac{1}{2}-n\end{matrix};1\right ] \end{split } \label{16}\ ] ] where @xmath115 and @xmath35 have been already given before we define @xmath117 $ ] , then it can be shown @xmath118 & \qquad\quad\times e^{-x^2}h_{m+n-2k-1}(x ) \label{17 } \end{split}\ ] ] where @xmath119 , @xmath120 is the @xmath20-th eigenfunction of harmonic trap and @xmath121 is the @xmath20-th hermite polynomial .
then we have @xmath122 & = \frac{3}{\pi(n-2)!n!\sqrt{n^2 - 1}}\frac{1}{2\sqrt{2}}\sum_{k=0}^{n-2}\sum_{l=0}^{n-1}(-1)^{k+l}k!l!\begin{pmatrix}n-2\\k\end{pmatrix}\begin{pmatrix}n+1\\k\end{pmatrix}\begin{pmatrix}n-1\\l\end{pmatrix}\begin{pmatrix}n\\l\end{pmatrix}\gamma\left(2n - k - l-\frac{3}{2}\right)\\[0.2 cm ] & = -\frac{3}{8}\sqrt{\frac{2(n^2 - 1)}{\pi^3}}\frac{\gamma\left(n-\frac{5}{2}\right)\gamma\left(n+\frac{1}{2}\right)}{\gamma(n)\gamma(n+2)}{}_3 f_2\left[\begin{matrix}\frac{3}{2},1-n ,- n\\[0.08 cm ] \frac{7}{2}-n,\frac{1}{2}-n\end{matrix};1\right ] \end{aligned } \label{18}\ ] ] k. m. ohara , s. l. hemmer , m. e. gehm , s. r. granade , and j. e. thomas , science * 298 * , 2179 ( 2002 ) .
m. e. gehm , s. l. hemmer , s. r. granade , k. m. ohara , and j. e. thomas , phys .
a * 68 * , 011401 ( 2003 ) .
t. bourdel , j. cubizolles , l. khaykovich , k. m. f. magalhaes , s. j. j. m. f. kokkelmans , g. v. shlyapnikov , and c. salomon , phys .
lett . * 91 * , 020402 ( 2003 ) .
s. gupta , z. hadzibabic , m. w. zwierlein , c. a. stan , k. dieckmann , c. h. schunck , e. g. m. van kempen , b. j. verhaar , and w. ketterle , science * 300 * , 1723 ( 2003 ) . c. a. regal and d. s. jin , phys .
lett . * 90 * , 230404 ( 2003 ) c. l. hung ,
x. b. zhang , n. gemelke , and c. chin , nature * 470 * , 236 ( 2011 ) . f. chevy , v. bretin , p. rosenbusch , k. w. madison , and j. dalibard , phys .
88 * , 250402 ( 2002 ) .
e. vogt , m. feld , b. frohlich , d. pertot , m. koschorreck , and m. kohl , phys . rev
* 108 * , 070404 ( 2012 ) .
m. olshanii , h. perrin , and v. lorent , phys .
* 105 * , 095302 ( 2010 ) .
e. taylor and m. randeria , phys .
* 109 * , 135301 ( 2012 ) .
j. hofmann , phys .
108 * , 185303 ( 2012 ) .
m. girardeau , j. math . phys . * 1 * , 516 ( 1960 ) .
t. kinoshita , t. wenger , and d. s. weiss , science * 305 * , 1125 ( 2004 ) .
b. paredes , a. widera , v. murg , o. mandel , s. f@xmath125lling , i. cirac , g. v. shlyapnikov , t. w. h@xmath126nsch , and i. bloch , nature * 429 * , 277 ( 2004 ) . l. p. pitaevskii and a. rosch , phys .
a * 55 * , r853 ( 1997 ) .
t. n. d. silva and e. j. mueller , preprint at arxiv : cond - mat/0607491 ( 2006 ) . k. merloti , r. dubessy , l. longchambon , m. olshanii , and h. perrin , phys .
a * 88 * , 061603(r ) ( 2013 ) c. menotti and s. stringari , phys .
a * 66 * , 043610 ( 2002 ) .
e. haller , m. gustavsson , m. mark , j. danzl , r. hart , g. pupillo , and h .- c .
n@xmath126gerl , science * 325 * , 1224 ( 2009 ) .
k. gunter , t. stoferle , h. moritz , m. kohl , and t. esslinger , phys .
lett . * 95 * , 230401 ( 2005 ) .
b. e. granger and d. blume , phys .
* 92 * , 133202 ( 2004 ) .
it can be shown for example that if the energy shift induced by a single @xmath2-function scatterer situated in between two walls is reinterpreted in terms of the interaction present in the fermionic hamiltonian ( 2 ) , the second order of the taylor expansion in the powers of its prefactor , proportional to @xmath13 , turns out to be positive , contradicting the non - positivity of the second order perturbation theory shift ; see problem 4.1.11 in @xcite .
m. d. girardeau and m. olshanii , phys .
a * 70 * , 023608 ( 2004 ) .
m. girardeau and m. olshanii , preprint at arxiv : cond - mat/0309396 ( 2003 ) .
diptiman sen , j. phys .
a * 36 * , 7517 ( 2003 ) .
e. h. lieb and w. liniger , phys . rev . *
130 * , 1605 ( 1963 ) .
w. kohn , phys .
rev . * 123 * , 1242 ( 1961 ) .
t. busch , b. g. englert , k. rzazewski , and m. wilkens , found .
* 28 * , 549 ( 1998 ) .
g. e. astrakharchik , phys . rev .
a * 72 * , 063620 ( 2005 ) .
l. p. pitaevskii and s. stringari , phys .
* 81*,4541 ( 1998 )
. l. d. landau and e. lifshitz , _ fluid mechanics : v. 6 ( course of theoretical physics ) _ ( butterworth - heinemann , oxford , 1987 ) allan griffin , tetsuro nikuni , eugene zaremba , _ bose - condensed gases at finite temperatures _ ( cambridge university press , cambridge , 2009 ) c. a. regal , c. ticknor , j. l. bohn , and d. s. jin , phys .
* 90 * , 053201 ( 2003 ) .
j. p. gaebler , j. t. stewart , j. l. bohn , and d. s. jin , phys .
98 * , 200403 ( 2007 ) . m. olshanii , _ back - of - the - envelope quantum mechanics : with extension to many - body systems and integral pdes _ ( word scientific , singapore , 2013 ) f. n. c. paraan and v. e. korepin , phys .
a * 82 * , 065603 ( 2010 ) .
w. kohn , phys .
123 * , 1242 ( 1961 ) .
f. w. j. olver , d. w. lozier , r. f. boisvert , and c. w. clark , _ nist handbook of mathematical functions _ ( cambridge university press 2010 ) .
i. s. gradshteyn and i. m. ryzhik , _ table of integrals , series , and products _
, 7-th edition ( elsevier academic press , oxford , 2007 ) . | in this article , we consider the monopole excitations of the harmonically trapped bose gas in the vicinity of the tonks - girardeau limit .
using girardeau s fermi - bose duality and subsequently an effective fermion - fermion odd - wave interaction , we obtain the dominant correction to the scale - invariance - protected value of the excitation frequency , for microscopically small excitation amplitudes .
we produce a series of diffusion monte carlo results that confirm our analytic prediction for three particles . and
less expectedly , our result stands in excellent agreement with the result of a hydrodynamic simulation ( with the lieb - liniger equation of state as an input ) of the microscopically large but macroscopically small excitations .
we also show that the frequency we obtain coincides with the upper bound derived by menotti and stringari using sum rules .
surprisingly , however , we found that the usually successful hydrodynamic perturbation theory predicts a shift that is @xmath0 higher than its _ ab initio _ numerical counterpart .
we conjecture that the sharp boundary of the cloud in local density approximation characterized by an infinite density gradient renders the perturbation inapplicable .
all our results also directly apply to the 3d @xmath1-wave - interacting waveguide - confined confined fermions . |
to quantify the star formation activity in the universe , from low to high redshift or in individual galaxies we need an estimate as accurate as possible of the recent star formation rate in galaxies . among various potential estimators ,
the far - infrared and ultraviolet luminosities are commonly used .
both emissions are expected to come from young stars ( e.g. kennicutt 1998 and references therein ) .
the ultraviolet light is emitted by young stars and is theoretically directly connected to the recent star formation rate .
however , the use of the ultraviolet emission to trace the star formation is largely hampered by the presence of the dust which absorbs and scatters the uv light ; the far - infrared emission is not affected by this limitation .
in fact the ultraviolet and fir emissions are complementary : the ultraviolet light lost because of the dust absorption is re - emitted in fir by the dust .
this energy budget has been used to derive a measure of the dust attenuation @xcite .
this attenuation is known to be very large in some objects and even if it remains modest in average in the nearby universe @xcite it is crucial to correct the observed ultraviolet flux before any quantitative interpretation .
in this paper we consider a galaxy sample selected in the nuv for which most of the galaxies have been observed in fir . because of its selection
, this sample will be directly comparable to optically ( ultraviolet rest - frame ) selected samples at higher redshift .
we will derive the dust attenuation from the comparison between the far - infrared and the ultraviolet emissions and discuss variations of this attenuation with luminosity and fuv - nuv color of the galaxies .
for comparison , we will also build a fir selected sample of galaxies , most of these galaxies being detected in nuv .
we will perform the same analysis as for the nuv selected sample a similar pioneering work was already performed over only 35 deg@xmath7 and in a single ultraviolet wavelength @xcite .
complementary approachs may consist in exploring variations of the fir / uv ratio within single nearby galaxies ( boissier et al .
, popescu et al .
this volume )
here , we briefly discuss how we select our samples .
galaxy distances are calculated from their redshift measurements taken from the nasa / ipac extragalactic database ( ned ) , hyperleda ( http://leda.univ-lyon1.fr ) and the iras pscz @xcite and assuming @xmath8 .
for very nearby galaxies ( 4 sources ) their distances were found in the literature .
all magnitudes are expressed in the ab system . +
our nuv selected sample , down to @xmath9 mag , is built from the first observations of the galex all imaging survey ( ais ) .
see martin et al . and morissey et al .
( this volume ) for details regarding the galex instrument and mission .
we select only the frames with exposure times larger than or equal to 50 sec : the resulting field covers 615 deg@xmath7 .
we perform photometry using the ir0.2 calibration and apply a zero point offset of + 0.039 mag in fuv and -0.099 mag in nuv .
we mask bright stars and galaxies lying close to our targets . for each target , the sky background is measured by combining several individual nearby sky regions .
nuv fluxes are extracted using elliptical apertures enclosing the total fluxes .
fuv magnitudes are measured within the same aperture .
nuv and fuv fluxes are first corrected for galactic extinction using the schlegel et al .
( 1998 ) dust map and the cardelli et al .
( 1989 ) extinction curve .
a total of 95 sources brighter than @xmath10 mag are then selected .
we exclude objects whose ultraviolet flux can be contaminated by other sources than young stars i.e. early type ( 2 ellipticals ) and active galaxies ( 2 seyfert 1 , 2 seyfert 2 and 1 qso ) .
most of the remaining 88 sources are late - type spirals .
we search for a detection at 60 @xmath1 m for all the 88 sources using the iras fsc @xcite and the scan processing and integration facility ( scanpi ) .
4 galaxies were not observed by iras .
all other sources but 3 are detected at 60 @xmath1 m .
for the 3 non detections , we adopt a conservative upper limit of 0.2 jy at 60@xmath1 m . among the 81 galaxies with an iras detection , 22
are contaminated by cirrus or have close neighbours not resolved by iras .
we discard these measures from the quantitative analysis of the fir emission .
we are thus left with 59 sources with a nuv and a flux at 60 @xmath1 m and 3 nuv sources with only an upper limit at 60 @xmath1 m . + to construct the fir selected sample , we start with the iras pscz catalog @xcite which is complete down to 0.6 jy at 60 @xmath1 m ; we take advantage of the optical identification of the iras sources to search for a nuv detection .
we select 163 iras sources over 509 deg@xmath7 detected at 60 @xmath1 m with a reliability larger than 50% and not contaminated by cirrus .
144 targets also have a flux at 100 @xmath1 m and the remaining have only upper limits . among these 163 galaxies ,
97 have a morphological type in ned or leda .
we discard 4 ellipticals or lenticulars , 3 seyfert 1 and 5 seyfert 2 and 33 galaxies which have close neighbours not resolved by iras .
we are thus left with 118 galaxies .
nuv and fuv photometries are performed following the same prescriptions as for the nuv selected sample .
all but 1 galaxy are detected in nuv ( s / n ratio larger than 3 ) .
14 galaxies are not detected in fuv .
for the non detections in nuv and/or fuv upper limits are estimated corresponding to a s / n ratio of 3 .
it was shown previously @xcite that for ultraviolet measurements up to 2000 @xmath11 the dust to ultraviolet flux ratio is a robust tracer of dust attenuation in star forming galaxies regardless of the details of the extinction mechanisms ( dust / star geometry , dust properties ) .
hereafter @xmath12 refers to the total dust emission and @xmath13 and @xmath14 are defined for each galex band as @xmath15 where @xmath16 is expressed in @xmath17 . since the nuv band of galex lies at longer wavelengths than 2000 @xmath11 , we must test the reliability of the @xmath18 ratio to estimate the dust attenuation for various scenarii of star formation . at this aim ,
we use the population synthesis code pegase @xcite under different hytotheses of star formation rate ( sfr ) ( figure 1 ) .
the amount of dust emission is obtained by adding all the stellar emission absorbed by the dust .
we try different configurations for the dust attenuation : the screen - like calzetti s attenuation law @xcite , a homogeneous mixture of dust and stars in a slab geometry with a galactic or a lmc stellar extinction law @xcite or a clumpy medium @xcite .
we also include the time dependent scenario proposed by charlot & fall ( 2000 ) . within this frame ,
we develop another bi - variate model in which the attenuation of the light from stars younger than @xmath19 years follows the calzetti s law whereas older stars are supposed to be homogeneously mixed with the dust or to be distributed in a clumpy medium .
all these models give very similar results for a given star formation history with a dispersion lower than 5 @xmath20 .
in particular the exact form of the extinction curve ( with or without a 2175 @xmath11 bump ) does not affect the results .
each curve plotted in figure 1 is the average over all these scenarii for a specific star formation rate .
similar plots are found with the fuv flux instead of the nuv one .
only the exponential decrease of the sfr with @xmath21= 2 gyr leads to a much lower attenuation for a given dust to nuv flux ratio : the dust heating by old stars becomes important and only a fraction of the dust emission is related to the ultraviolet absorption .
such a steep decrease of the sfr is typical of elliptical or lenticular galaxies @xcite .
all the other scenarii of star formation lead to a relative error lower than 20% .
the calibration holds for both galex bands and the @xmath18 will be used to estimate the dust attenuation since more galaxies are detected in nuv than in fuv .
we perform polynomial fits on our models excluding the curves with @xmath222 gyr since the fir and the nuv selections focus on active star forming galaxies and we exclude ellipticals and lenticulars from the study .
@xmath23 where @xmath24 and @xmath25 where @xmath26 .
these fits are fully consistent with the mean relations between the dust attenuation and the dust to ultraviolet flux ratio for both galex bands proposed by kong et al .
( 2004 ) and charlot & brinchmann ( private communication ) for similar star formation histories . before estimating the attenuations with the above formulae
, we need to estimate the total dust emission . for
the galaxies observed at 60 and 100 @xmath1 m we calculate this total dust emission following dale et al .
3 galaxies from the nuv selected sample were not detected at 100 @xmath1 m : we use the mean value of the @xmath27 ratio found for the sample galaxies detected at both wavelengths to estimate their flux at 100 @xmath1 .
the mean bolometric corrections @xmath28 are equal to @xmath29 and @xmath30 for the nuv and fir selected sample respectively ( the fir flux is defined between 40 and 120 @xmath1 m @xcite ) .
these values are intermediate between the values found for starbursts and cirrus dominated galaxies @xcite . in the fir selected sample ,
the median values of the dust attenuation are found larger : @xmath33 mag in nuv and @xmath34 mag in fuv .
the galaxies not detected in fuv have a dust attenuation larger than @xmath35 mag in agreement with the large attenuations ( @xmath364 mag ) found in nuv .
the histograms of the attenuations are given in figure 2 .
in figure 3 , the dust attenuation in the nuv band is compared to the intrinsic luminosity of the galaxies expected to be the best measurement of the recent star formation rate : the nuv luminosity corrected for the extinction for the nuv selected sample ( left panel ) and the dust luminosity for the fir selected sample ( right panel ) . in both cases
the dust attenuation is found to increase with the luminosity .
this result is consistent with findings of an increase of the dust attenuation with the star formation rate in nearby galaxies @xcite .
vijh et al .
( 2003 ) proposed a law between the extinction and the luminosity at 1600 @xmath37 for bright lyman break galaxies at @xmath38 : the extrapolation of this law to fainter objects predicts lower attenuations than those found for our nuv selected sample .
this effect might be attributed to some evolution of the dust attenuation with redshift .
we can compare the dust attenuation found in our samples to the mean dust attenuation estimated from the fir and nuv ( or fuv ) luminosity densities ( e.g. buat et al .
1999 ) . at this aim
we calculate the ratio of the fir and nuv ( fuv ) luminosity densities and translate it to a mean dust attenuation .
the fir density is taken from saunders et al .
( 1990 ) : @xmath39 with h=0.7 .
the nuv and fuv luminosity density are derived from the new galex data ( wyder et al .
, this volume ) : @xmath40 and @xmath41 .
adopting @xmath42 we obtain @xmath43 which translates to dust attenuations of 1.1 mag and 1.6 mag in nuv and fuv respectively .
these mean values are intermediate between those found with our fir and nuv selection although closer to the dust attenuation found for the nuv selected galaxies .
therefore , as far as the dust attenuation is concerned , a nuv selection seems to be more representative of the mean galaxy properties in the local universe than a fir selection .
bell ( 2002 ) and more recently kong et al .
( 2004 ) showed that the slope of the uv continuum ( which can be traced by the fuv - nuv color ) is not a reliable tracer of the dust attenuation in galaxies which are not experiencing a strong starburst .
in figure 4 is plotted the fuv - nuv color against the dust to fuv flux ratio for the nuv selected and the fir selected galaxies .
we keep only galaxies brighter than 20 mag in nuv and fuv in order to have a photometric error in each band lower than or equal to 0.15 mag .
almost all the nuv selected galaxies and a large fraction of the fir selected ones lie below the curve followed by starburst galaxies @xcite : they exhibit a lower @xmath44 ( i.e. a lower attenuation ) than starbursts for a given fuv - nuv color , as already found for non starbursting galaxies .
kong et al .
( 2004 ) explain this behavior by a variation of the ratio of present to past average star formation rate : our data are roughly consistent with a ratio varying from 1 to 0.1 .
about 1/3 of the fir selected galaxies lie above the curve valid for starburst galaxies : their dust attenuation is larger than that expected from the fuv - nuv color .
this trend is similar to that for ultra - luminous infrared galaxies @xcite although the dust attenuations found here are not as large as for the ulirgs .
the fuv - nuv color seems therefore unable to trace the dust attenuation in these galaxies .
galex ( galaxy evolution explorer ) is a nasa small explorer , launched in april 2003 .
we gratefully acknowledge nasa s support for construction , operation , and science analysis for the galex mission , developed in cooperation with the centre national detudes spatiales of france and the korean ministry of science and technology .
bell , e. 2002 , 577 , 150 buat , v. & xu , c. 1996 , , 306 , 61 buat , v. , donas , j. , milliard , b. & xu , c. 1999 , , 352 , 371 calzetti , d. , kinney , a.l . & storchi - bergmann , t. 1994 , , 533 , 682 calzetti , d. , armus , l. , bohlin , r. c. , kinney , a. l. , koornneef , j. , & storchi - bergmann , t. 2000 , , 533 , 682 cardelli , j.,a . , clayton , g. , & mathis , j.,s .
1989 , , 345 , 245 charlot , s. , fall , m. 2000 , 539 , 718 dale , d. a. , helou , g. , contursi , a. , silbermann , n. a. , & kolhatkar , s. 2001 , , 549 , 215 fioc , m. , rocca - volmerange , b. 1997 , , 326 , 950 gavazzi , g. , bonfanti , c. , sanvito , g. , boselli , & scodeggio , m. 2002 , , 576 , 135 goldader , j . , d. , meurer , g. , heckman , t. , m. , seibert , m. , sanders , d. b. , calzetti , d. , steidel , c. , c. 2002 , , 568 , 651 gordon , k. d. , clayton , g. c. , witt , a. n. , & misselt 2000 , , 533 , 236 helou , g. , khan , i. r. , malek , l. , & boehmer , l. 1988 , , 68 , 151 hopkins , a. m. , connolly , a. , j. , haarsma , d. b. , cram , l. e. 2001 , 122 , 288 iglesias - pramo , j. , buat , v. , donas , j. , boselli , a. , milliard , b. 2004 , , 419 , 109 kennicutt , r. c. 1998 , , 36 , 189 kong , x. , charlot , s. , brinchmann , j. , & fall , s.m .
2004 , astro - ph/0312474 meurer , g. r. , heckman , t. m. , & calzetti , d. 1999 , , 521 , 64 moshir , m. et al .
1990 , iras faint source catalogue , version 2.0 rowan - robinson , m. 2003 , , 344 , 13 saunders , w. , rowan - robinson , m. , lawrence , a. et al .
1990 , , 242 , 318 saunders , w. et al .
2000 , , 317 , 55 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 steidel , c. c. , adelberger , k. l. , giavalisco , m. , dickinson , m. , & pettini , m. 1999 , , 519 , 1 vijh , u.p . ,
witt , a.n . , gordon , k.d .
2003 , apj 587 , 533 wang , b. , heckman , t. 1996 , apj 457 , 645 xu , c. , buat , v. 1995 , 293 , l65 ) against the fuv - nuv color for nuv and fir selected samples .
the solid line is the mean relation expected for starburst galaxies @xcite .
the error bars were estimated assuming 15% uncertainty for the iras fluxes and 40% uncertainty for the @xmath45 ratio , the errors on the fuv and nuv fluxes were supposed to be poissonian . ] | we compare the dust attenuation properties of two samples of galaxies purely selected in the near - ultraviolet ( nuv ) band ( 1750 - 2750 @xmath0 ) and in the far - infrared ( fir ) at 60 @xmath1 m .
these samples are built using the galex and iras sky surveys over @xmath2 600 square degrees .
the nuv selected sample contains 95 galaxies detected down to @xmath3 mag ( ab system ) .
83 galaxies in this sample are spirals or irregulars and only two of them are not detected at 60 @xmath1 m .
the fir selected sample is built from the iras pscz catalog complete down to 0.6 jy . among the 163 sources , we select 118 star forming galaxies well measured by iras ,
all but 1 are detected in nuv and 14 galaxies are not detected in the far - ultraviolet ( fuv ) band ( 1350 - 1750 @xmath4 ) .
the dust to ultraviolet ( nuv and fuv ) flux ratio is calibrated to estimate the dust attenuation at both wavelengths .
the median value of the attenuation in nuv is found to be @xmath5 mag for the nuv selected sample versus @xmath6 mag for the fir selected one . within both samples
, the dust attenuation is found to correlate with the luminosity of the galaxies .
almost all the nuv selected galaxies and 2/3 of the fir selected sample exhibit a lower dust attenuation than that expected from the tight relation found previously for starburst galaxies between the dust attenuation and the slope of the ultraviolet continuum .
the situation is inverse for one third of the fir selected galaxies : their extinction is higher than that deduced from their fuv - nuv color and the relation valid for starbursts . |
since the discovery of the first free - floating brown dwarfs a decade ago @xcite , surveys of young clusters and the field have found brown dwarfs in increasing numbers and at decreasing masses . a natural consequence of these growing samples of brown dwarfs has been the identification of binary systems with progressively lower total masses ( * ? ? ?
* ; * ? ? ?
* references therein ) .
the least massive binaries are potentially valuable for testing theories for the formation of objects at the bottom of the initial mass function @xcite . to produce meaningful results , these tests require reliable measurements of the basic properties of the components of substellar binaries , including spectral type , luminosity , age , and mass .
oph 1622@xmath02405 is one such low - mass binary that deserves careful scrutiny .
this system was discovered during a search for disk - bearing young brown dwarfs with the _
spitzer space telescope _
the @xmath12 pair was only partially resolved by _ spitzer _ , but was fully resolved in optical and near - infrared ( ir ) images and low - resolution near - ir spectroscopy presented by ( * ? ? ?
* see also @xcite ) . through this spectroscopy , she classified each component as a pre - main - sequence object and measured spectral types of m7.5 and m8 for the primary and secondary , respectively . by placing the two components on the hertzsprung - russell ( h - r ) diagram with theoretical evolutionary models , @xcite estimated masses of 0.06 and 0.05 @xmath6 and a system age of @xmath13 myr . in comparison , @xcite have reported spectral types of m9 and m9.5-l0 and masses of 0.013 and 0.007 @xmath6 for the components of oph 1622@xmath02405 , leading @xcite to characterize oph 1622@xmath02405 as the first known planetary - mass binary . @xcite
independently discovered the binarity of oph 1622@xmath02405 and estimated masses of 0.017 and 0.014 @xmath6 for its components .
we seek to better determine the physical properties of the components of oph 1622@xmath02405 , particularly their masses , through new optical and near - ir spectroscopy ( [ sec : obs ] ) . with the optical data , we measure the spectral types of oph 1622@xmath02405a and b with the optical classification scheme that has been most commonly applied to young late - type objects and show how these types compare to those of other young binaries through direct comparison of their optical spectra ( [ sec : class ] ) .
we use gravity - sensitive absorption lines in the optical and ir spectra of oph 1622@xmath02405a to constrain its age ( [ sec : grav ] ) .
we then estimate the masses of oph 1622@xmath02405a and b via theoretical evolutionary models and by considering the dynamical mass measurements of the eclipsing binary brown dwarf 2mass j05352184 - 0546085 ( hereafter 2 m 0535 - 0546 , [ sec : mass ] ) .
the optical and ir images of oph 1622@xmath02405a and b ( referred to as oph 1622@xmath02405n and s , respectively , in @xcite ) from @xcite are shown in figure [ fig : image ] .
the astrometry and photometry measured by @xcite for this pair are provided in table [ tab : data1 ] .
@xcite described the collection and reduction of the larger imaging survey from which these images are taken .
@xcite measured near - ir colors for oph 1622@xmath02405a and b as well .
their colors differed significantly from those of field m and l dwarfs , which they attributed to low surface gravity .
however , a color difference of that kind is not present in previous data for young late - m objects ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and the colors that we measure for oph 1622@xmath02405a and b are similar to those of both dwarfs and pre - main - sequence sources . on the nights of 2006 august 17 and 18 , we performed optical spectroscopy on oph 1622@xmath02405a and b , respectively , with the low dispersion survey spectrograph ( ldss-3 ) on the magellan ii telescope and a @xmath14 slit .
this configuration resulted in a spectral resolution of 5.8 at 8000 and a wavelength coverage of 5800 - 11000 . for each component of the binary
, we obtained one 20 min exposure with the slit aligned at the parallactic angle .
after bias subtraction and flat - fielding , the spectra were extracted and calibrated in wavelength with arc lamp data .
the spectra were then corrected for the sensitivity functions of the detectors , which were measured from observations of spectrophotometric standard stars . for comparison to oph 1622@xmath02405 in
[ sec : grav ] , we will make use of a spectrum of the eclipsing binary brown dwarf 2 m 0535 - 0546 @xcite that was obtained with ldss-3 on 2006 february 10 and spectra of gl 1111 ( m6.5v ) and lhs 2243 ( m8v ) that were obtained with the blue channel spectrograph at the mmt on 2004 december 11 and 12 , respectively . the ldss-3 configuration for 2 m 0535 - 0546 was the same as for oph 1622@xmath02405 .
the spectra of gl 1111 and lhs 2243 from blue channel have a resolution of 2.6 at 8000 and a wavelength coverage of 6300 - 8900 .
we obtained near - ir spectra of oph 1622@xmath02405a and b with the spectrometer spex @xcite at the nasa infrared telescope facility ( irtf ) on the nights of 2006 june 22 and 24 .
the instrument was operated in the sxd mode with a @xmath15 slit , producing a wavelength coverage of 0.8 - 2.5 and a resolving power of @xmath16 . with
the slit rotated to encompass both components of the pair , we obtained 10 2-minute exposures during sequences of dithers between two positions on the slit on each night . for comparison purposes ,
we also make use of a spectrum of the taurus member cfht 4 that was obtained on the night of 2006 january 4 with the same instrument configuration as used for oph 1622@xmath02405 , except with a @xmath17 slit ( @xmath18 ) .
these spex data were reduced with the spextool package @xcite and corrected for telluric absorption @xcite .
spectral types of late - m dwarfs and giants are defined at red optical wavelengths @xcite .
averages of optical spectra of standard dwarfs and giants agree well with data for late - m pre - main - sequence objects @xcite and are the basis of most of the published optical spectral types for low - mass members of nearby star - forming regions ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we have applied this classification scheme to our optical spectra of oph 1622@xmath02405a and b , arriving at spectral types of m7.25@xmath1 and m8.75@xmath1 , respectively .
if we use dwarfs alone to classify our spectra as done for this binary by @xcite , then we derive nearly the same spectral types , namely m7.5 and m9 .
[ fig : op1 ] and [ fig : op2 ] show comparisons of oph 1622@xmath02405a and b to these best - matching standards .
the averages of dwarfs and giants match the target spectra more closely than dwarfs alone , in agreement with previous work ( e.g. , * ? ? ?
thus , we adopt the types based on the comparisons to averages of dwarf and giant standards . by doing so
, our classifications can be reliably compared to types of most known young late - type objects , which have been measured in the same way . in figure [ fig : op1 ] , we compare the spectrum of oph 1622@xmath02405a to data for the composite of the young eclipsing binary 2 m 0535 - 0546 ( [ sec : obs ] ) and the primary in the young wide binary 2 m 1101 - 7732 @xcite .
this comparison demonstrates that oph 1622@xmath02405a is slightly later than 2 m 0535 - 0546a+b ( which we classify as m6.75 ) and has the same spectral type as 2 m 1101 - 7732a ( m7.25 , * ? ? ?
meanwhile , oph 1622@xmath02405b is slightly later than 2 m 1101 - 7732b ( m8.25 , * ? ? ?
* ) based on the comparison of their spectra in figure [ fig : op2 ] .
our optical spectral types of m7.25@xmath1 and m8.75@xmath1 for oph 1622@xmath02405a and b are significantly earlier than the optical types of m9 and m9.5-l0 from @xcite .
this is true even if we use dwarf standards as done in those studies .
as shown in figure [ fig : op1 ] , the spectrum of oph 1622@xmath02405a is poorly matched by both m9v and m9v+m9iii , as well as the m9 taurus member kpno 12 @xcite .
similarly , the spectrum of oph 1622@xmath02405b is earlier than kpno 4 ( figure [ fig : op2 ] ) , which is the prototypical pre - main - sequence representative of the m9.5 spectral class @xcite .
oph 1622@xmath02405b also differs significantly from m9.5v and the young l0 object 2mass 01415823 - 4633574 ( hereafter 2 m 0141 - 4633 , * ? ? ? * ) .
for instance , a defining characteristic of the transition from m to l types for dwarfs is the disappearance of tio absorption at 7000 - 7200 @xcite , and yet the tio in oph 1622@xmath02405b is strong , indicating that a dwarf - based spectral type of m9.5-l0 is not appropriate . our results for oph 1622@xmath02405 a and b are similar to those of @xcite for another object classified by @xcite , 2 m 1541 - 3345 , which is a disk - bearing source in the vicinity of the lupus clouds @xcite .
@xcite reported a spectral type of m8 for this object while @xcite classified it is [email protected] with spectra and methods like those used in this work .
we now compare our optical classifications of oph 1622@xmath02405a and b to previous near - ir observations .
our optical types of m7.25@xmath1 and m8.75@xmath1 for oph 1622@xmath02405a and b are consistent with the ir types of m7.5@xmath19 and m8@xmath19 from @xcite and m7@xmath19 and m8@xmath19 from @xcite , which were measured by comparison to young objects that have been optically classified with the same methods employed in this work . @xcite and @xcite also presented near - ir spectra for oph 1622@xmath02405a and b. @xcite did not compare their spectra to data for classification standards and thus did not measure spectral types .
@xcite found that oph 1622@xmath02405b exhibited similar ir spectral features as the young l0 source 2 m j0141 - 4633 , and thus classified the former as m9.5@xmath3 .
because the spectral features of the primary indicated a slightly earlier type , they classified it as m9@xmath3 .
however , given that @xcite did not present a comparison of oph 1622@xmath02405 to earlier types , it is unclear if earlier types would have provided better or worse matches to their data .
in fact , as shown in figure [ fig : op2 ] , our optical spectrum of oph 1622@xmath02405b is very different from the spectrum of 2 m j0141 - 4633 from @xcite , demonstrating that they do not have the same optical spectral types .
@xcite and @xcite estimated effective temperatures and surface gravities by comparing their data to synthetic spectra . however , because of known deficiencies in theoretical spectra of late - type objects @xcite , temperature and gravity estimates of this kind are subject to systematic errors , and thus the accuracies of these estimates are unknown . in addition , the temperature and gravity estimates from those studies can not be reliably compared to those of other young late - type objects unless the latter are derived with the same spectral features , model spectra , and fitting procedures .
@xcite and @xcite determined that the components of oph 1622@xmath02405 are pre - main - sequence objects rather than field dwarfs by performing low - resolution near - ir spectroscopy and detecting the presence of triangular @xmath20-band continua @xcite .
this spectral characteristic is present during most of the pre - main - sequence evolution of late - type objects , thus constraining the age of oph 1622@xmath02405 to be @xmath21 myr @xcite . to further constrain the age of this system , we examine additional gravity - sensitive absorption lines in our higher - resolution optical and near - ir spectra , namely k i , na i , and feh @xcite . in this analysis
, we consider only the primary because its data exhibit better signal - to - noise . for comparison to oph 1622@xmath02405a , we select representatives of three distinct luminosity classes and ages : members of the taurus star - forming region ( @xmath9 myr , * ? ?
* ; * ? ? ?
* ) , members of the upper scorpius ob association ( @xmath10 myr , * ? ? ?
* ) , and field dwarfs ( @xmath8 gyr ) .
we require that these objects have spectral types that are within 0.5 subclass of the spectral type of the primary so that our comparison is sensitive to variations in surface gravity alone . for taurus ,
we use the optical spectrum of 2mass 04484189 + 1703374 ( m7 , * ? ? ?
* ) from @xcite and our ir spectrum of cfht 4 ( m7 , * ? ? ?
field dwarfs are represented by an average of our optical spectra of gl 1111 ( m6.5v , * ? ? ?
* ) and lhs 2243 ( m8v , * ? ? ?
* ) and the ir spectrum of vb 8 ( m7 , * ? ? ?
* ) from @xcite . for upper sco
, we use the @xmath22-band spectrum of u sco ctio 128 ( m7 , * ? ? ?
* ) from @xcite . to enable a reliable comparison of absorption line strengths , spectra for a given wavelength range
have been smoothed to a common spectral resolution .
although optical spectra for upper sco are available from @xcite , we do not include them in this comparison because they have significantly lower spectra resolution that the other optical data we are examining , and we prefer to make these comparisons at the highest possible resolution . in figs .
[ fig : lines1 ] and [ fig : lines2 ] , we compare the spectra for taurus , upper sco , and field dwarfs to the spectrum of oph 1622@xmath02405a for wavelength ranges encompassing k i , na i , and feh . for all of these transitions , oph 1622@xmath02405a exhibits stronger lines than the taurus members and weaker lines than the field dwarfs , indicating that it is above the main sequence ( @xmath21 myr ) but older than taurus ( @xmath8 myr ) . for the subset of comparisons in which upper sco
is represented , the line strengths are similar between oph 1622@xmath02405a and the upper sco member .
if we degrade the optical spectrum of oph 1622@xmath02405a to the lower spectral resolution of optical data in upper sco from @xcite , we also find similar line strengths for the optical transitions of na i and k i. thus , the gravity - sensitive lines in the spectra of oph 1622@xmath02405a suggest an age similar to that of upper sco ( @xmath10 myr ) with an upper limit that is undetermined , but probably no more than a few tens of millions of years . using their near - ir spectra of oph 1622@xmath02405a and b ,
@xcite compared the equivalent widths of gravity - sensitive lines between oph 1622@xmath02405 and field dwarfs , demonstrating that the binary s components have lower gravities and hence younger ages than dwarfs , which is consistent with the results in this work , @xcite , @xcite , and @xcite .
however , their analysis did not further refine the age of the system and they did not claim to accurately measure the gravities and ages of oph 1622@xmath02405a and b , and instead assumed membership in ophiuchus . on the other hand , @xcite constrained the gravity of oph 1622@xmath02405a and b more tightly by comparing their data to kpno 4 ( @xmath9 myr ) and @xmath23 ori 51 ( @xmath10 myr ) .
this comparison indicated that oph 1622@xmath02405a and b are older than kpno 4 and similar in age to @xmath23 ori 51 , which is in agreement with the age constraints that we have derived in this section .
the analysis of gravity - sensitive lines in
[ sec : grav ] demonstrated that oph 1622@xmath02405a is older than taurus ( @xmath9 myr ) , which in turn is older than the stellar population within the ophiuchus cloud core ( @xmath11 myr , * ? ? ? * ) .
thus , oph 1622@xmath02405 is not a member of the current generation of stars forming within the ophiuchus cloud , which is consistent with the low extinction of this binary ( @xmath24 ) and its large angular distance from the cloud core ( @xmath25 ) . instead ,
oph 1622@xmath02405 probably is part of an older population of stars in the sco - cen complex , which contains several neighboring and overlapping generations of stars with ages from @xmath26 to 20 myr @xcite .
for instance , oph 1622@xmath02405 is within the area encompassed by known members of upper sco @xcite and is near a population of exposed young stars distributed across the front of the ophiuchus cloud , which is coeval with upper sco and may be an extension of it @xcite .
although we can not definitively identify the origin of oph 1622@xmath02405 nor measure its distance , membership in upper sco is likely based on its location and surface gravity diagnostics .
indeed , this evidence of membership in upper sco is the same as for previously reported late - type members @xcite .
therefore , for the purpose of estimating their luminosities , we assign to oph 1622@xmath02405a and b the distance of upper sco , which extends from 125 to 165 pc @xcite . based on the comparison of the optical spectra of oph 1622@xmath02405a and b to spectra of other young late - type objects in [ sec : class ] , we find that the extinction of each component is @xmath24 .
therefore , we adopt @xmath27 for measuring their luminosities .
the remaining details of the luminosity estimates are provided by @xcite .
our luminosity measurements for oph 1622@xmath02405a and b are listed in table [ tab : data2 ] . to examine the ages implied by their luminosities , we plot oph 1622@xmath02405a and b on the h - r diagram in figure [ fig : hr ] with the evolutionary models of @xcite and @xcite .
we have converted our optical spectral types to effective temperatures with a temperature scale that is compatible with these models for young objects @xcite .
the data and models in figure [ fig : hr ] imply ages of 10 - 30 myr for the primary and 1 - 20 myr for the secondary .
thus , these results are consistent with coevality for the two objects , which is expected for a binary system .
these ages are somewhat older than the canonical value of 5 myr that is usually quoted for upper sco , but the age of a young population is sensitive to how it is defined , the mass range of objects considered , and the choice of models .
to reliably compare the inferred ages of oph 1622@xmath02405a and b to those of upper sco members , the luminosities and temperatures of this binary should be compared directly to those of late - type members of upper sco .
we do this by including in figure [ fig : hr ] the late - type members of upper sco from @xcite .
the lower limits of the sequence in temperature and luminosity are reflections of the detection limits of the survey from @xcite .
an extension of the sequence below these limits in a manner that is parallel to the theoretical isochrones would encompass both components of oph 1622@xmath02405 .
in other words , the model ages of the more massive members of the upper sco sequence from @xcite are consistent with the model ages of oph 1622@xmath02405a and b. similarly , the binary components fall within the sequence of low - mass upper sco members in the color - magnitude diagram from @xcite .
in addition to ages , the evolutionary models in figure [ fig : hr ] also provide estimates of masses , implying values of [email protected] and @xmath28 @xmath6 for oph 1622@xmath02405a and b , respectively .
the quoted uncertainties reflect only the uncertainties in spectral types and luminosities .
additional systematic errors could be introduced by the adopted temperature scale and evolutionary models .
however , the sizes of these systematic errors are probably not large , as demonstrated by various observational tests @xcite .
the mass estimated for oph 1622@xmath02405a by @xcite is similar to our value , while her estimate for the secondary was twice our value because of the earlier spectral type that she derived . meanwhile ,
our estimates for oph 1622@xmath02405a and b are significantly higher than the masses of 0.013 and 0.007 @xmath6 from @xcite .
our estimate for the primary is also much higher than the value of 0.017 @xmath6 from @xcite , while our mass for the secondary is only slightly higher than their mass of 0.014 @xmath6 .
the validity of higher estimates is supported by the fact that oph 1622@xmath02405a is only slightly cooler than the composite of the eclipsing binary brown dwarf 2 m 0535 - 0546a+b ( figure [ fig : op1 ] ) and thus should have a comparable mass ( @xmath29 and 0.034 @xmath6 , * ? ? ?
the spectral types , temperatures , luminosities , and masses of oph 1622@xmath02405a and b produced by our analysis are compiled in table [ tab : data2 ] .
using optical spectroscopy , we have measured spectral types for the young binary oph 1622@xmath02405 that are significantly earlier than those reported by @xcite and @xcite . as a result ,
our mass estimates for these objects ( @xmath30 and 0.019 @xmath6 ) are higher than those from ( * ? ? ?
* ; * ? ? ?
* @xmath31 and 0.007 @xmath6 ) and ( * ? ? ?
* @xmath32 and 0.014 @xmath6 ) and are above the range of planetary masses ( @xmath7 @xmath33 , * ? ? ? * ) . through a direct comparison of their spectra
, we find that the primaries in oph 1622@xmath02405 and the young wide binary 2 m 1101 - 7732 have the same spectral types while oph 1622@xmath02405b is only slightly later than 2 m 1101 - 7732b , which strongly indicates that these two binaries have similar masses .
our analysis of gravity - sensitive absorption lines in the spectra of oph 1622@xmath02405a have demonstrated that this system is too old to be a member of the ophiuchus star - forming region ( @xmath11 myr ) . instead , the age constraints from those data combined with the position of oph 1622@xmath02405 on the h - r diagram are consistent with membership in upper sco ( @xmath10 myr ) . additional observations ( e.g. , proper motions , radial velocities ) are needed to better determine the origin and membership of this binary system . if the distance of upper sco is adopted for oph 1622@xmath02405 , then the separation of @xmath12 for this binary corresponds to @xmath34 au , making it the second young wide binary brown dwarf to be found .
thus , oph 1622@xmath02405 is very similar to 2 m 1101 - 7732 in both mass and separation but is somewhat older ( 5 myr versus 1 myr ) . given the advanced age of this system compared to most disk - bearing stars and brown dwarfs , the disk detected in oph 1622@xmath02405 by @xcite is a valuable laboratory for studying the evolution of brown dwarf disks .
we thank davy kirkpatrick for providing his spectrum of 2 m j0141 - 4633 and laird close for providing his results prior to publication .
k. l. was supported by grant ast-0544588 from the national science foundation .
we thank ivelive momcheva for providing the telescope time for the ldss-3 observations of oph 1622@xmath02405 .
this work is supported by nasa through the spitzer space telescope fellowship program , through a contract issued by the jet propulsion laboratory , california institute of technology under a contract with nasa .
llllllll a & 16 22 25.2 & -24 05 13.7 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + b & 16 22 25.2 & -24 05 15.6 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + | we present an analysis of the mass and age of the young low - mass binary oph 1622@xmath02405 . using resolved optical spectroscopy of the binary , we measure spectral types of m7.25@xmath1 and m8.75@xmath1 for the a and b components , respectively .
we show that our spectra are inconsistent with the spectral types of m9 and m9.5-l0 from jayawardhana & ivanov and [email protected] and m9.5@xmath3 from close and coworkers . based on our spectral types and the theoretical evolutionary models of chabrier and baraffe
, we estimate masses of @xmath4 and @xmath5 @xmath6 for oph 1622@xmath02405a and b , which are significantly higher than the values of 0.013 and 0.007 @xmath6 derived by jayawardhana & ivanov and above the range of masses observed for extrasolar planets ( @xmath7 @xmath6 ) .
planet - like mass estimates are further contradicted by our demonstration that oph 1622@xmath02405a is only slightly later ( by 0.5 subclass ) than the composite of the young eclipsing binary brown dwarf 2 m 0535 - 0546 , whose components have dynamical masses of 0.034 and 0.054 @xmath6 . to constrain the age of oph 1622@xmath02405 , we compare the strengths of gravity - sensitive absorption lines in optical and near - infrared spectra of the primary to lines in field dwarfs ( @xmath8 gyr ) and members of taurus ( @xmath9 myr ) and upper scorpius ( @xmath10 myr ) .
the line strengths for oph 1622@xmath02405a are inconsistent with membership in ophiuchus ( @xmath11 myr ) and instead indicate an age similar to that of upper sco , which is agreement with a similar analysis performed by close and coworkers .
we conclude that oph 1622@xmath02405 is part of an older population in sco - cen , perhaps upper sco itself . |
blazars constitute a class of active galactic nuclei ( agn ) that often show very strong and rapid flux variability over the electromagnetic spectrum .
they are widely held to contain a black hole ( bh ) with mass in the range @xmath6 , that launches relativistic jets emitting highly non - thermal radiation
. the jet transports energy in electromagnetic form , and bulk plus random kinetic energy of charged particles .
the source radiation may also show a contribution by the accretion disk ( including the big blue bump , bbb ) , the broad line region ( blr ) , and a dusty torus ( see urry & padovani 1995 ) ; lack or weakness of such contributions mark out the class of bl lac objects .
the primary energy of the jet may be supplied by power extracted from the central rotating bh via interaction with its accretion disk ( blandford & znajek 1977 ; blandford & payne 1982 ; see also discussions in cavaliere & delia 2002 and mckinney 2005 ) . the blazar 0716 + 714 is a distant bl lac at @xmath7 ( nilsson et al .
2008 ) ; its optical - uv continuum is so featureless ( biermann et al .
1981 ; stickel , fried & k@xmath8hr 1993 ) that a redshift estimate has been possible only resolving and using the host galaxy as a standard candle ( nilsson et al .
this bl lac is of the ibl type , displaying a first broad peak in the optical - uv bands and showing another broad peak near @xmath9 gev ; the crossover between the two components falls in x - ray range ( massaro et al .
2008 ; ferrero et al .
recently , has been detected by agile above 100 mev several times during the period september - october 2007 when the source was quite active in optical band with variations on 1-day timescale ( villata et al . 2008 ) .
two bright flares were detected .
the first one reached a flux of @xmath10photons@xmath0@xmath11s@xmath12 on 2007 september 11 with a photon index of @xmath13 and duration @xmath14day . in 2007 october , following another prominent optical activity detected by the webt consortium , agile and _ swift _ satellites pointed again at the source . on 2007
october 23 , another 1-day bright flare was observed with a flux comparable to that detected in september ( chen et al .
2008 ) . around this date
a bright one - day flare and strong day - variability were observed in the optical band ( villata et al .
2008 ) ; uv and soft x - rays also showed strong and fast variability , whereas modest or none variability appeared in the band 4 - 10 kev ( giommi et al . 2008 ) .
meanwhile , the radio flux showed a slow coherent increase that remarkably begins around the day of the first flare and culminates around the date of the second flare ( villata et al .
thereafter was detected going back to its ground state in all bands , with the photon index that softened toward @xmath15 ( chen et al .
2008 ) as previously observed by egret ( lin et al .
1995 ) . in this paper
, we present a physical model for the flaring states , and study their extreme energetics with implications on energy extraction from a rotating bh .
the simplest homogeneous synchrotron self - compton ( ssc ) model assumes the blazar emissions to be produced in a
blob of radius @xmath16 , containing relativistic electrons in a combination of tangled and uniform magnetic field .
the emitters move toward the observer with bulk lorentz factor @xmath17 ( see , e.g. , tavecchio , maraschi & ghisellini 1998 ) .
we assume the emitters to emerge from the injection / acceleration phase with a jet - frame distribution of the random energies @xmath18 in the form of a standard broken power - law @xmath19 where @xmath20 and @xmath21 are the spectral indices for @xmath22 and @xmath23 , respectively , @xmath24 is the lorentz factor at the break . these electrons emit a primary synchrotron spectrum ; a second contribution is then produced by inverse compton ( ic ) as the primary synchrotron photons scatter off the same electron population .
the spectral energy distribution ( sed ) behaves as @xmath25 , where @xmath26 is the energy of the received photons , and @xmath27 . for electrons in a magnetic field @xmath28 ,
the synchrotron sed peaks around @xmath29 where @xmath30 is planck s constant , @xmath31 is the redshift of the source , and @xmath32^{-1}$ ] is the bulk doppler factor due to the flow of emitters toward the observer at an angle @xmath33 relative to the line of sight ; the sed at the synchrotron peak is @xmath34 as to the ic component , its sed contribution peaks at @xmath35 with a peak value of @xmath36 if the scattering takes place in the thomson regime with the density of target photons scaling as @xmath37 .
the relativistic motion toward the observer amplifies the emitted power by the factor @xmath38 , and allows it to vary on a timescale @xmath39 close to or shorter than the crossing - time @xmath40 . due to the synchrotron and ic losses the electrons cool with timescale @xmath41 where @xmath42 is the thomson cross section ,
@xmath43 and @xmath44 is the energy density of radiation before scattering .
this sets a typical cooling break at @xmath45 beyond which the electrons cool rapidly . in the following ,
we take into account this constraint on the particle distributions . _
simultaneous _ multi - frequency observations can provide the five quantities @xmath46 and @xmath24 with the five equations ( 2)-(6 ) . for bl lacs with high frequency peaks ( hbls ) requiring electrons of higher energies ( @xmath47 ) the scattering approaches the klein - nishina ( kn ) regime with a blob - frame photon energy @xmath48 . in the extreme kn regime
the ic sed peaks at @xmath49 ; the dependence on @xmath28 and @xmath24 progressively weakens as the two latter parameters grow .
additional target photon can be provided by a source external to the jet ( see dermer et al .
2009 ) . in this case
, the high - energy component of the spectra is due to the electrons that compton - scatter the external photons ( ec ) ; the sed now peaks at energies @xmath50 and the corresponding sed value is @xmath51 in this ec process two new ingredients enter : @xmath52 and @xmath53 , respectively , the energy and the density at peak of the external photons as seen by the moving blob .
this has two main consequences : + * i * ) the model contains two further degrees of freedom and the parameter evaluation may be degenerate ; + * ii * ) these news quantities are related to @xmath54 and @xmath55 in the observer frame by means of the bulk lorentz factor @xmath17 in a manner that depends on the geometry of the system ( dermer & schlickeiser 2002 ) , causing an additional dependence on @xmath17 in the ec spectra .
dermer & schlickeiser ( 1993 ) discuss sed dependences on @xmath17 varying from @xmath56 to @xmath57 , for photons entering into the blob from behind or head - on , respectively
. equations .
( 2 ) - ( 8) show that , in the synchrotron - ic framework , variabilities of the first and second peaks are correlated , possibly with a lag @xmath58 . when the energy fluxes around these peaks @xmath59 are simultaneously monitored , we can compare the corresponding light curves @xmath60 in the respective band energy @xmath61 and @xmath62 , possibly with a time - lag @xmath63 .
then given two times @xmath64 and @xmath65 , the ratio @xmath66 shows a dependence on @xmath67 that is related to the emission mechanism .
here we consider some relevant cases . if the variability is mainly due to electron injection / acceleration , @xmath68 and/or @xmath24 changes ; then , @xmath69 results if ssc dominates the ic emission .
if instead ec dominates then @xmath70 applies ( compare equation ( 3 ) with equations ( 5 ) and ( 8) ) .
if variations are mainly due to changes in @xmath28 , then @xmath70 holds for ssc , and @xmath71 applies for ec . if instead @xmath17 varies , we have @xmath70 in the ssc ; in the ec framework we have different behaviors depending on the geometric of the jet relative to the external radiation : @xmath72 with @xmath73 ranging from 3/4 for seed photons entering from behind , to 3/2 for photons entering head - on the flow . .variability patterns . [ cols="^,^,^,^,^ " , ] previous radio monitoring by the very long baseline array ( vlba ) telescope of showed the presence of more super - luminal components relative to the active state of 2003 - 2004 ( rastorgueva et al .
moreover , the absence of the signatures of ic catastrophe provided a lower limit @xmath74 for the doppler factor ( ostorero et al .
2006 ; fuhrmann et al . 2008
; see also wagner et al .
1996 ) ; bach et al .
( 2005 ) also argue for high doppler factors in these fast variable components and for a viewing angle @xmath75 .
we note that modeling the sed of with a standard one - zone ssc model would fail to reproduce the simultaneous radio , optical , x - ray and data for the october flare .
in particular , the crossover in the x - ray band would not be well reproduced , and the hard x - ray flux would be overestimated by a factor @xmath763 ; furthermore it would be difficult to model the hardness of the spectrum during the september flare ( figure 1 red dashed lines ) .
a one - component model also hardly explain together the slow trends of the radio , optical and hard x - ray bands and the faster variability observed in the optical , soft x - ray and bands ( see villata et al .
2008 , giommi et al .
2008 , chen et al .
hence , we adopt a two - component model : the first produces the slowly variable radio and hard x - ray bands , whereas the second is responsible for the faster variability in optical - uv , soft x- , and bands . despite these problems , the possibility to fully constrain by observations the parameters
stimulate us to also show for comparison a one - component model .
the sed of all these models are shown in figure 1 and the parameters are listed in table 2 ; a viewing angle * @xmath77 * is adopted according to bach et al .
under the assumption of isotropic emission , the observed power radiated from a source with luminosity distance @xmath78 is @xmath79 .
the jet transports a total power @xmath80 contributed by intrinsic radiated power , kinetic energy flow of the electrons and of the cold protons ( with one proton per emitting electron ) , and poynting flux , respectively : @xmath81 see also celotti & ghisellini ( 2008 ) .
the latter authors show that in bl lacs @xmath82 tends to match the sum of the other contributions .
in fact , for the two flares of at redshift @xmath83 we find @xmath84 , and from our two - component ssc model ( with the parameters listed in table 2 ) we obtain a total jet power @xmath85 with @xmath86 . under this condition the total jet power is minimized and the details of cooling does not affect materially the global energetics , being the radiated luminosity @xmath82 mainly contributed by peaks emission . moreover , the uncertainty in @xmath87 is mainly due to the observed flux error . for the one - component model , we obtain @xmath88and @xmath89 holds . in this case , the total jet power is not dominated by the radiated one , but the parameters now are well constrained by the observation and the uncertainty is still due to the flux error . here ,
we compare the jet powers provided by our models with the bz mechanism ; this set a limit on the power extractable from a rotating bh @xmath90 under conservative values of @xmath28 , as discussed in cavaliere & delia ( 2002 , and references therein ; see also our discussion ) .
estimating the redshift and the bh mass of is not trivial because of the lack of emission lines .
recently , however , nilsson et al . ( 2008 ) pinpointed the host galaxy of this source and reported a determination of its redshift at @xmath91 .
then , for this host , a bh of mass @xmath92 should accord with the fundamental plane of bl lacs ( falomo et al .
2003 ) .
on the other hand , some authors using micro - variability of the optical flux estimate the mass of the central bh for by @xmath93}.\ ] ] considering @xmath94s as the shortest variability time found by sasada et al .
( 2008 ) one obtains @xmath95 .
gupta , srivastava & wiita ( 2009 ) obtain the more stringent constraint @xmath96m@xmath5 for the case of a kerr bh .
however , these authors discuss that the micro - variability in may be due to a small region in the jet or due to internal disk modes , so that the bh mass would be higher .
we note that a bh of mass @xmath97m@xmath5 in would imply in equation ( 13 ) a power limit @xmath98 inconsistent with the power @xmath82 emitted during the flares .
we modeled the flares of 2007 september 11 and october 23 with a two - component ssc model , similarly to tavecchio @xmath99 ghisellini ( 2009 ) . despite the larger number of model parameters
, we believe that our two - component modeling reproduces the complex variability and the hard spectra of better than a one - component model .
our figure 2 indicates a quadratic dependence between the synchrotron and ic fluxes .
this concurs with the lack of emission lines and bbb to rule out models involving external sources of seed photons that produce a linear dependence , and to strongly support electrons radiating in the thomson regime within the ssc framework as the most likely radiation process .
we refer to our figures 2 and 1 in chen et al .
( 2008 ) to show that the 2007 september data lie around the maximum ; in time , the trajectory starts from the lower @xmath100 value , attains a maximum and then falls down . in 2007 october
, the trajectory describes the decline of a flare starting from the higher @xmath100 value .
rises and falls both occur on 1-day timescale , and the trajectories are quadratic .
this suggests the electron acceleration and cooling to occur on similar timescales .
the crossover between the synchrotron and ic branches of the sed is located in x - rays ( see also foschini et al . 2006 ; ferrero et al .
2006 ) , and is contributed by both the slowly variable component ( i ) and the faster component ( ii ) . as to the faster components ( ii )
adopted for the two flares , they are marked by high electron energies @xmath101 with a sharp low energy cutoff @xmath102 ( see also tsang & kirk 2007 ) .
moreover , high bulk lorentz factor @xmath103 ( that is @xmath104 ) is used in accord with wagner et al .
( 1996 ) . in table 2
, it is shown that our choice of parameters implies @xmath105 for the _ fast _ components ii : those quench very rapidly causing strong variability in the optical - uv , soft x- and bands .
little or no variability results in radio and hard - x ray bands produced by the rising part of the _ slow _ components i. this behavior is in agreement with the complex multi - band variability reported by chen et al .
( 2008 ) , giommi et al .
( 2008 ) , and villata et al .
( 2008 ) . considering the redshift @xmath106 of
, we find that its intrinsic radiative luminosity is of order @xmath107 . on adding the other jet components in equations ( 10 ) - ( 12 )
, the total power becomes @xmath108 for the two - component model , and @xmath109 for the one - component model : the source exceeds the bz limit @xmath110 ( see figure 3 ) .
this obtains for a maximally rotating bh of @xmath111(spun by past accretion ) via interaction with a disk magnetic field @xmath112 sustained by gas or radiation pressure , in the absence of ongoing accretion ( see discussion by cavaliere & delia 2002 ) .
the _ simultaneous _ nature of our multifrequency data during the 2007 flares of is a crucial ingredient of our result .
other bl lacs with weak or negligible accretion disk or blr contributions have been reported to attain ( model dependent ) high total luminosities , e.g. , 2032 + 107 with an inferred @xmath113 ( see celotti & ghisellini 2008 ) .
corbel & reyes ( 2008 , atel 1744 ) report the high peak flux recently attained by 0235 + 164 at @xmath114 for which , however , ec contributions can not be ruled out .
however , to our knowledge , only the simultaneous data obtained by our group for provide a model independent evaluation of the total jet power from a bl lac being close or just above the bz limit .
we also show in figure 3 the intrinsic radiated power for other bl lacs with intense and tev emissions as for the 1997 flare of bl lacertae with peak flux around @xmath115 but with lower redshift @xmath116 ( bloom et al .
1997 ; see also ravasio et al .
2002 ) and intermittent evidence of lines and thermal emissions .
we also report w comae with redshift @xmath117 but flux at levels of @xmath118 ( bttcher , mukherjee & reimer 2002 ) .
the increasing trend of @xmath119 with @xmath31 is likely to arise from the malmquist bias , while the decrease at low @xmath31 may result from sampling a limited cosmological volume .
nevertheless , we stress that up to now no bl lac source has sharply exceeded the bz limiting power .
it will be worthwhile to keep under close watch the most distant bl lacs ( despite the obvious monitoring difficulties ) to catch powerful emissions .
if violations will be found , these may be possibly discussed in terms of the blandford - payne mechanism ( blandford & payne 1982 ) that , however , requires ongoing accretion not supported in the case of .
alternatively , higher powers may be attained with higher magnetic fields up to @xmath120 related to the plunging orbits ( see meier 1999 ) , which however imply short source lifetimes in the absence of accretion . in conclusion
we find the total power transported in the jet of 0716 + 714 to be @xmath121erg@xmath0s@xmath12 , and so to approach the limit of the bz mechanism for a bh up to @xmath122m@xmath5 with _ conservative _ values of @xmath28 .
such high powers in constitute an unescapable consequence of two observed facts : * the flux attains high levels in the sed during the two flares , with an emitted power comparable with the optical one as shown by simultaneous observations ( see figure 1 ) ; * the lack of external sources of seed photons related to emission lines or bbb concurs with the observed quadratic dependence @xmath69 to rule out ec contributions to the high energy hump ( see figure 2 ) and considerable ongoing accretion . | bl lac objects of the intermediate subclass ( ibls ) are known to emit a substantial fraction of their power in the energy range 0.110 gev . detecting emission from such sources
provides therefore a direct probe of the emission mechanisms and of the underlying powerhouse .
the agile satellite detected the remarkable ibl s5@xmath0(@xmath1 ) during a high state in the period from 2007 september - october , marked by two very intense flares reaching peak fluxes of @xmath2 above 100 mev , with simultaneous optical and x - ray observations .
we present here a theoretical model for the two major flares and discuss the overall energetics of the source .
we conclude that 0716 + 714 is among the brightest bl lac s ever detected at @xmath3-ray energies .
because of its high power and lack of signs for ongoing accretion or surrounding gas , the source is an ideal candidate to test the maximal power extractable from a rotating supermassive black hole via the _ pure _ blandford - znajek ( bz ) mechanism .
we find that during the 2007 flares approached or just exceeded the upper limit set by bz for a black hole of mass @xmath4m@xmath5 . |
the quantitative understanding of structure , function , dynamics and transport of biomolecules is a fundamental theme in contemporary life sciences .
geometric analysis and associated biophysical modeling have been the main workhorse in revealing the structure - function relationship of biomolecules and contribute enormously to the present understanding of biomolecular systems .
however , biology encompasses over more than twenty orders of magnitude in time scales from electron transfer and ionization on the scale of femtoseconds to organism life spanning over tens of years , and over fifteen orders of magnitude in spatial scales from electrons and nuclei to organisms . the intriguing complexity and extraordinarily large number of degrees of freedom of biological systems
give rise to formidable challenges to their quantitative description and theoretical prediction .
most biological processes , such as signal transduction , gene regulation , dna specification , transcription and post transcriptional modification , are essentially intractable for atomistic geometric analysis and biophysical simulations , let alone _ ab - initio _ quantum mechanical descriptions .
therefore , the complexity of biology and the need for its understanding offer an extraordinary opportunity for innovative theories , methodologies , algorithms and tools .
the study of subcellular structures , organelles and large multiprotein complexes has become one of the major trends in structural biology .
currently , one of the most powerful tools for the aforementioned systems is cryo - electron microscopy ( cryo - em ) , although other techniques , such as macromolecular x - ray crystallography , nuclear magnetic resonance ( nmr ) , electron paramagnetic resonance ( epr ) , multiangle light scattering , confocal laser - scanning microscopy , small angle scattering , ultra fast laser spectroscopy , etc .
, are useful for structure determination in general @xcite . in cryo - em experiments ,
samples are bombarded by electron beams at cryogenic temperatures to improve the signal to noise ratio ( snr ) .
the working principle is based on the projection ( thin film ) specimen scans collected from many different directions around one or two axes , and the radon transform for the creation of three - dimensional ( 3d ) images .
one of major advantages of cryo - em is that it allows the imaging of specimens in their native environment .
another major advantage is its capability of providing 3d mapping of entire cellular proteomes together with their detailed interactions at nanometer or subnanometer resolution @xcite .
the resolution of cryo - em maps has been improved dramatically in the past two decades , thanks to the technical advances in experimental hardware , noise reduction and image segmentation techniques . by further taking the advantage of symmetric averaging ,
many cryo - em based virus structures have already achieved a resolution that can be interpreted in terms of atomic models .
there have been a variety of elegant methods @xcite and software packages in cryo - em structural determination @xcite .
most biological specimens are extremely radiation sensitive and can only sustain a limited electron dose of illumination . as a result ,
cryo - em images are inevitably of low snr and limited resolution @xcite .
in fact , the snrs of cryo - tomograms for subcellular structures , organelles and large multi - protein complexes are typically in the neighborhood of 0.01 @xcite .
to make the situation worse , the image contrast , which depends on the difference between electron scattering cross sections of cellular components , is also very low in most biological systems .
consequently , cryo - em maps often do not contain adequate information to offer unambiguous atomic - scale structural reconstruction of biological specimens .
additional information obtained from other techniques , such as x - ray crystallography , nmr and computer simulation , is indispensable to achieve subnanometer resolutions .
however , for cryo - em data that do not have much additional information obtained from other techniques , the determination of what proteins are involved can be a challenge , not to mention subnanometer structural resolution . to improve the snr and image contrast of cryo - em data , a wide variety of denoising algorithms has been employed @xcite .
standard techniques , such as bilateral filter @xcite and iterative median filtering @xcite have been utilized for noise reduction .
additionally , wavelets and related techniques have also been developed for cryo - em noise removing @xcite . moreover ,
anisotropic diffusion @xcite or beltrami flow @xcite approach has been proposed for cryo - em signal recovering .
however , cryo - em data denoising is far from adequate and remains a challenge due to the extremely low snrs and other technical complications @xcite .
for example , one of difficulties is how to distinguish signal from noise in cryo - em data . as a result
, one does not know when to stop or how to apply a threshold in an iterative noise removing process .
there is a pressing need for innovative mathematical approaches to further tackle this problem . recently
, persistent homology has been advocated as a new approach for dealing with big data sets @xcite . in general ,
persistent homology characterizes the geometric features with persistent topological invariants by defining a scale parameter relevant to topological events .
the essential difference between the persistent homology and traditional topological approaches is that traditional topological approaches describe the topology of a given object in truly metric free or coordinate free representations , while persistent homology analyzes the persistence of the topological features of a given object via a filtration process , which creates a family of similar copies of the object at different spatial resolutions .
technically , a series of nested simplicial complexes is constructed from a filtration process , which captures topological structures continuously over a range of spatial scales .
the involved topological features are measured by their persistent intervals .
persistent homology is able to embed geometric information to topological invariants so that birth " and death " of isolated components , circles , rings , loops , pockets , voids or cavities at all geometric scales can be monitored by topological measurements .
the basic concept of persistent homology was introduced by frosini and landi @xcite and by robins @xcite in 1999 independently .
edelsbrunner et al .
@xcite introduced the first efficient computational algorithm , and zomorodian and carlsson @xcite generalized the concept .
a variety of elegant computational algorithms has been proposed to track topological variations during the filtration process @xcite .
often , the persistent diagram can be visualized through barcodes @xcite , in which various horizontal line segments or bars are the homology generators lasted over filtration scales .
it has been applied to a variety of domains , including image analysis @xcite , image retrieval @xcite , chaotic dynamics verification @xcite , sensor network @xcite , complex network @xcite , data analysis @xcite , computer vision @xcite , shape recognition @xcite and computational biology @xcite . the concept of persistent homology has also been used for noise reduction .
it is generally believed that short lifetime events ( or bars ) are of less importance and thus regarded as `` noise '' while long lifetime ones are considered as `` topological signals '' @xcite , although this idea was challenged in a recent work @xcite . in topological data analysis ,
pre - processing algorithms are needed to efficiently remove these noise .
depending on the scale of a feature , a simple approach is to pick up a portion of landmark points as a representative of topological data @xcite .
the points can be chosen randomly , spatially evenly , or from extreme values .
more generally , certain functions can be defined as a guidance for node selection to attenuate the noise effect , which is known as thresholding .
clustering algorithms with special kernel functions can also be employed to recover topological signal @xcite .
all of these methods can be viewed as a process of data sampling without losing the significant topological features .
they rely heavily on the previous knowledge of the geometric or statistic information .
in contrast , topological simplification @xcite , which is to remove the simplices and/or the topological attributes that do not survive certain threshold , focuses directly on the persistence of topological invariant .
in contrast , gaussian noise is known to generate a band of bars distributes over a wide range in the barcode representation @xcite .
thank to the pairing algorithm , persistence of a homology group is measured through an interval represented by a simplex pair .
if the associated topological invariant is regarded less important , simplices related to this simplex pair are reordered .
this approach , combined with morse theory , proves to be a useful tool for denoising @xcite , as it can alters the data locally in the region defined as noise .
additionally , statistical analysis has been carried out to provide confidence sets for persistence diagram .
however , persistent homology has not been utilized for cryo - em data noise reduction , to our knowledge .
a large amount of experimental data for macroproteins and protein - protein complexes has been obtained from cryo - em . to analyze these structural data ,
it is a routine procedure to fit them with the available high - resolution crystal structures of their component proteins .
this approach has been shown to be efficient for analyzing many structures and has been integrated into many useful software packages such as chimera @xcite . however , this docking process is limited by data quality .
for some low resolution data , which usually also suffer from low snrs , there is enormous ambiguity in structure fitting or optimization , i.e. , a mathematically ill - posed inverse problem .
sometimes , high correlation coefficients can be attained simultaneously in many alternative structures , while none of them proves to be biologically meaningful .
basically , the fitting or optimization emphasizes more on capturing `` bulk '' regions , which is reasonable as greater similarities in density distributions imply higher possibility .
however , little attention is paid to certain small `` linkage '' regions , which play important roles in biological system especially in macroproteins and protein - protein complexes .
different linkage parts generate different connectivity , and thus directly influence biomolecular flexibility , rigidity , and even its functions . since persistent homology is equally sensitive to both bulk regions and small linkage regions , it is able to make a critical judgment on the model selection in structure determination , however , nothing has been reported on persistent homology based solution to ill - posed inverse problems , to our knowledge .
although persistent homology has been applied to a variety of fields , the successful use of persistent homology is mostly limited to characterization , identification and analysis ( cia ) .
indeed , persistent homology has seldom employed for quantitative prediction .
recently , we have introduced molecular topological fingerprints ( mtfs ) based on persistent homology analysis of molecular topological invariants of biomolecules @xcite .
we have utilized mtfs to reveal the topology - function relationship of macromolecules .
it was found that protein flexibility and folding stability are strongly correlated to protein topological connectivity , characterized by the persistence of topological invariants ( i.e. , accumulated bar lengths ) @xcite .
most recently , we have employed persistent homology to analyze the structure and function of nano material , such as nanotubes and fullerenes .
the mtfs are utilized to quantitatively predict total curvature energies of fullerene isomers @xcite .
the overall objective of this work is to explore the utility of persistent homology for cryo - em analysis .
first , we propose a topology based algorithm for cryo - em noise reduction and clean - up .
we study the topological fingerprint or topological signature of noise and its relation to the topological fingerprint of cryo - em structures .
we note that the histograms of topological invariants of the gaussian random noise have gaussian distributions in the filtration parameter space .
contrary to the common belief that short barcode bars correspond to noise , it is found that there is an inverse relation between the snr and the band widths of topological invariants , i.e. , the lower snr , the larger barcode band width is .
therefore , at a low snr , noise can produce long persisting topological invariants or bars in the barcode presentation .
moreover , for cryo - em data of low snrs , intrinsic topological features of the biomolecular structure are hidden in the persistent barcodes of noise and indistinguishable from noise contributions . to recover the topological features of biomolecular structures ,
geometric flow equations are employed in the present work .
it is interesting to note that topological features of biomolecular structures persist , while the topological fingerprint of noise moves to the right during the geometric flow iterations . as such
, `` signal '' and noise separate from each other during the geometric flow based denoising process and make it possible to prescribe a precise noise threshold for the noise removal after certain iterations .
we demonstrate the efficiency of our persistent homology controlled noise removal algorithm for both synthetic data and cryo - em density maps .
additionally , we introduce persistent homology as a new strategy for resolving the ill - posed inverse problem in cryo - em structure determination .
although the structure determination of microtubule data emd 1129 is used as an example , similar problems are widespread in other intermediate resolution and low resolution cryo - em data . as emd 1129
is contaminated by noise , a preprocess of denoising is carried out by using our persistent homology controlled geometric flow algorithm .
a helix backbone is obtained for the microtubule intermediate structure .
based on the assumption that the voxels with high electron density values are the centers of tubulin proteins , we construct three different microtubule models , namely a monomer model , a two - monomer model , and a dimer model .
we have found that all three models give rise to essentially the same high correlation coefficients , i.e. , 0.9601 , 0.9607 and 0.9604 , with the cryo - em data .
this ambiguity in structure fitting is very common with intermediate and low resolution data .
fortunately , after our topology based noise removal , the topology fingerprint of microtubule data is very unique , which is true for all cryo - em data or data generated by using other molecular imaging modalities .
it is interesting to note that although three models offer the same correlation coefficients with the cryo - em data , their topological fingerprints are dramatically different .
it is found that the topological fingerprint of the microtubule intermediate structure ( emd 1129 ) can be captured only when two conditions are simultaneously satisfied : first , there must exist two different types of monomers , and additionally , two type of monomers from dimers .
therefore , based on topological fingerprint analysis , we can determine that only the third model is a correct model for microtubule data emd 1129 .
the rest of this paper is organized as follows .
the essential methods and algorithms for geometric and topological modelings of biomolecular data are presented in section [ methods ] .
approaches for geometric modeling , which are necessary for topological analysis , are briefly discussed .
methods for persistent homology analysis are described in detail .
we illustrate the use of topological methods with both synthetic volumetric data and cryo - em density maps .
their persistence of topological invariants is represented by barcodes .
the geometric interpretation of the topological features is given .
section [ sec : ph_noise ] is devoted to the persistent homology based noise removal .
the characterization of gaussian noise is carried out over a variety of snrs to understand noise topological signature .
based on this understanding , we design a persistent homology monitored and controlled algorithm for noise removal , which is implemented via the geometric flow .
persistent homology guided denoising is applied to the analysis of a supramolecular filamentous complex . in section [ sec : ph_microtubule ] , we demonstrate topology - aided structure determination of microtubule cryo - em data .
several aspects are considered including helix backbone evaluation , coarse - grained modeling and topology - aided structure design and evaluation .
we show that topology is able to resolve ill - posed inverse problem .
this paper ends with a conclusion .
persistent homology has been utilized to analyze biomolecular data , which are collected by different experimental means , such as macromolecular x - ray crystallography , nmr , epr , etc . due to their different origins
, these data may be available in different formats , which requires appropriate topological tools for their analysis .
additionally , their quality , i.e. , resolution and snr varies for case to case , and thus , a preprocessing may be required . moreover , although biomolecular structures are not a prerequisite for persistent homology analysis , the understanding of biomolecular structure , function and dynamics is crucial for the interpretation of topological results . as a consequence , appropriate geometric modeling @xcite is carried out in a close association with topological analysis .
furthermore , information from geometric and topological modelings is in turn , very valuable for data preprocessing and denoising .
finally , topological information is shown to be crucial for geometric modeling , structural determination and ill - posed inverse problems . [ cols="^ " , ] in the @xmath0 panel of fig .
[ fig:1129_barcodes_polish ] , bars can be roughly groups into three parts from the top to the bottom , i.e. , an irregular hair - like " part on the top , a narrow regular body " part in the middle and a large regular base " part in the bottom .
topologically , these parts represent different components in the microtubule intermediate structure .
the irregular hair - like " part corresponds to the partial monomer structures located on the top and the bottom boundaries of the structure . as can be seen in fig .
[ fig : microtubule_fitting ] , each monomer has lost part of the structure at the boundary regions . the regular body " and base " parts
are basically related to two types of monomers in the middle region where the structure is free of boundary effect . from the barcodes
, it can be seen that
body " part has a later birth " time and earlier death " time compared with the base " barcode part .
this is due to the reason that this type of monomers has relative lower electron density .
as the filtration is defined to go from highest electron density values to lowest ones , their corresponding barcodes appear later .
their earlier death time , however , is due to the reason that they form dimers with the other type of monomers represented by the base " barcode part .
it can be derived from these nonuniform behavior that monomers are not equally distributed along the helix backbone structure .
instead , two adjacent different types of monomers form a dimer first and then all these dimers simultaneously connect with each other as the filtration goes on . moreover , from the analysis in the previous section , it is obvious to see that the body " and base " parts are topological representations of type ii " monomer and type i " monomer , respectively . for the @xmath1 panel of fig .
[ fig:1129_barcodes_polish ] , there also exists a consistent pattern when the denoising process passing a certain stage . two distinctive types of barcodes can be identified in the fingerprint , i.e. , a shorter band of barcodes on the top and a longer band of bars on the bottom .
topologically , these @xmath1 bars correspond to the rings formed between two adjacent helix circles of monomers or dimers . during the filtration
, dimers are formed between type i " and type ii " monomers and soon after that , all dimers connect with each other and form the helix backbone . as the filtration goes on , type i " monomers from the upper helix circle first connect with type ii " monomers at the lower circle .
geometrically , this means six monomers , three (
i - ii - i " ) from the upper layer and three ( ii - i - ii " ) from the lower layer , form a circle .
as the filtration goes further , this circle evolves into two circles when two middle monomers on two layers also connect .
however , these two circles do not disappear simultaneously . instead , one persists longer than the other .
this entire process generates the unique topological fingerprint in @xmath1 barcodes .
the topological fingerprint we extracted from the denoising process can be used to guide the construction and evaluation of our microtubule models . to this end
, we analyze the topological features of three theoretical models .
our persistent homology results for three models are demonstrated in figs .
[ fig:1129_fit_barcodes ] * a * , * b * and * c * , respectively . it can be seen that all the three models are able to capture the irregular hair " region in their @xmath0 barcode chart . from the topological point of view , the first model is the poorest one .
it fails to capture the regular fingerprint patterns in both @xmath0 and @xmath1 panels of the original cryo - em structure in fig .
[ fig:1129_barcodes_polish ] * d*. with two different weight functions to represent two types of monomers , the second model delivers a relatively better topological result .
it is able to preserve part of the difference between type i " and type ii " barcodes in the @xmath0 panel . in @xmath1 panel ,
some nonuniform barcodes emerges .
the persistent homology results are further improved in the third model when the intra - dimer and inter - dimer interactions are considered . in our third model ,
fingerprint patterns of the cryo - em structure in both @xmath0 and @xmath1 panels of fig .
[ fig:1129_barcodes_polish ] * d * are essentially recovered by those of fig .
[ fig:1129_fit_barcodes ] * c*. even though their scales are different , their shapes are strikingly similar .
the essential topological features that are associated with major topological transitions of the original cryo - em structure are illustrated in figs .
[ fig:1129_homology ] * a@xmath2 * , * a@xmath3 * , * a@xmath4 * and * a@xmath5*. as shown in figs .
[ fig:1129_homology ] * b@xmath2 * , * b@xmath3 * , * b@xmath4 * and * b@xmath5 * , these features have been well - preserved during the denoising process . our best predicted model is depicted in figs .
[ fig:1129_homology ] * c@xmath2 * , * c@xmath3 * , * c@xmath4 * and * c@xmath5*. in these figure labels , subscripts @xmath6 and @xmath7 denote four topological transition stages in the filtration process , namely hetero - dimmer formation , large circles formation , evolution of one large circle into two circles , and finally death of one of two circles . by the comparison of denoising results ( figs .
* b@xmath2 , b@xmath3 , b@xmath4 * and * a@xmath5 * ) with original structures ( figs . *
a@xmath2 , a@xmath3 , a@xmath4 * and * a@xmath5 * ) , it is seen that , in the noise reduction process , although some local geometric and topological details are removed , fundamental topological characteristics are well preserved . as illustrated in fig .
[ fig:1129_barcodes_polish ] , using the persistent homology description , these fundamental topological characteristics are well - preserved in topological persistence patterns , which are further identified as fingerprints of the microtubule intermediate structure .
we believe that topological fingerprints are crucial to the characterization , identification , and analysis of the biological structure . as demonstrated in figs .
[ fig:1129_homology ] * c@xmath2 * , * c@xmath3 * , * c@xmath4 * and * c@xmath5 * , once our model successfully reproduces the topological fingerprints , the simulated structure is able to capture the essential topological characteristics of the original one .
moreover , through the analysis in section [ sec : mode_evaluation ] , it can be seen that to reproduce the topological fingerprint of emd 1129 , two conditions are essential .
the first is the creation of two types of monomers .
the second is the differentiation of intra - dimers and inter - dimers .
biologically , these requirements means : 1 ) there are two types of monomers , i.e. , @xmath8-tubulin monomers and @xmath9-tubulin monomers ; and 2 ) intra - dimers and inter - dimers should behave differently from hetero - dimers . it also should be noticed that a higher correlation coefficient may not guarantee the success of the model , especially when the original data is of low resolution and low snr . as can be seen in section [ sec : three_models ] , our three theoretical models have very similar fitting coefficients .
the second model even has a slightly higher cross - correlation coefficients .
however , only the third model is able to reproduce the essential topological features of the original cryo - em data .
this happens as topological invariants , i.e. , connected components , circles , loops , holes or void , tend to be very sensitive to tiny " linkage parts , which are almost negligible in the density fitting process , compared to the major body part .
we believe these linage parts play important roles in biological system especially in macroproteins and protein - protein complexes .
different linkage parts generate different connectivity , thus can directly influence biomolecular flexibility , rigidity , and even its functions . by associates topological features with geometric measurements , our persistent homology analysis is able to distinguish these connectivity parts
. therefore , persistent homology is able to play a unique role in protein design , model evaluation and structure determination .
cryo - electron microscopy ( cryo - em ) is a major workhorse for the investigation of subcellular structures , organelles and large multiprotein complexes . however , cryo - em techniques and algorithms are far from mature , due to limited sample quality and or stability , low signal to noise ( snr ) , low resolution , and the high complexity of the underlying molecular structures .
persistent homology is a new branch of topology that is known for its potential in the characterization , identification and analysis ( cia ) of big data . in this work ,
persistent homology is , for the first time , employed for the cryo - em data cia .
methods and algorithms for the geometric and topological modeling are presented . here , geometric modeling ,
such the generation of density maps for proteins or other molecules , is employed to create known data sets for validating topological modeling algorithms .
we demonstrate that cryo - em density maps and fullerene density data can be effectively analyzed by using persistent homology . since topology is very sensitive to noise , the understanding of the topological signature of noise is a must in cryo - em cia .
we first investigate the topological fingerprint of gaussian noise .
we reveal that for the gaussian noise , its topological invariants , i.e. , @xmath0 , @xmath1 and @xmath10 numbers , all exhibit the gaussian distribution in the filtration space , i.e. , the space of volumetric density isovalues . at a low snr , signal and noise
are inseparable in the filtration space .
however , after denoising with the geometric flow method , there is clear separation between signal and noise for various topological invariants . as such , a simple threshold can be prescribed to effectively remove noise . for the case of low snr ,
the understanding of noise characteristic in the filtration space enable us to use persistent homology as an efficient means to monitor and control the noise removal process .
this new strategy for noise reduction is called topological denoising .
persistent homology has been applied to the theoretical modeling of a microtubule structure ( emd 1129 ) .
the backbone of the microtubule has a helix structure .
based on the helix structure , we propose three theoretical models .
the first model assumes that protein monomers form the helix structure .
the second model adopts two types of protein monomers evenly distributed along helix chain .
the last model utilizes a series of protein dimers along the helix chain .
these models are fitted with experimental data by the least square optimization method .
it is found that all the three models give rise to similar high correlation coefficients with the experimental data , which indicates that the structural optimization is ill - posed . however , the topological fingerprints of three models are dramatically different . in the denoising process ,
the cryo - em data of the microtubule structure demonstrate a consistent pattern which can be recognized as the intrinsic topological fingerprint of the microtubule structure .
by careful examination of the fingerprint , we reveal two essential topological characteristics which discriminate the protein dimers from the monomers . as such ,
we conclude that only the third model , i.e. , the protein dimer model , is able to capture the intrinsic topological characteristics of the cryo - em structure and must be the best model for the experimental data .
it is believed that the present work offers a novel topology based strategy for resolving ill - posed inverse problems .
this work was supported in part by nsf grants dms-1160352 and iis-1302285 , nih grant r01gm-090208 and msu center for mathematical molecular biosciences initiative .
the authors acknowledge the mathematical biosciences institute for hosting valuable workshops .
shawn q. zhenga , bettina keszthelyi , eric branlunda , john m. lyleb , michael b. braunfelda , john w. sedatb , and david a. agarda . : an integrated software suite for real - time electron microscopic tomographic data collection , alignment , and reconstruction . , 157:138147 , 2007 .
t. hrabe , y. chen , s. pfeffer , l.k .
cuellar , a.v .
mangold , and f. forster . : a python - based toolbox for localization of macromolecules in cryo - electron tomograms and subtomogram analysis . , 178:178188 , 2012 .
r. s. pantelic , c. y. rothnagel , r.and huang , d. muller , d. woolford , m. j. landsberg , a. mcdowall , b. pailthorpe , p. r. young , j. banks , b. hankamer , and g. ericksson .
the discriminative bilateral filter : an enhanced denoising filter for electron microscopy data .
, 155:395c408 , 2006 .
p. van der heide , x. p. xu , b. j. marsh , d. hanein , and n. volkmann .
efficient automatic noise reduction of electron tomographic reconstructions based on iterative median filtering .
, 158:196c204 , 2007 .
v. de silva and g. carlsson .
topological estimation using witness complexes . in _ proceedings of the first eurographics conference on point - based graphics _ , pages 157166 .
eurographics association , 2004 .
r. j. adler , o. bobrowski , m. s. borman , e. subag , and s. weinberger .
persistent homology for random fields and complexes . in _
borrowing strength : theory powering applications a festschrift for lawrence d. brown _ , volume 6 , pages 124143 .
institute of mathematical statistics , 2010 .
m. lysaker , a. lundervold , and x. c. tai .
noise removal using fourth - order partial differential equation with application to medical magnetic resonance images in space and time .
, 12(12):15791590 , 2003 .
q. qiao , c. h. yang , c. zheng , l. fontn , l. david , x. yu , c. bracken , m. rosen , a. melnick , e. h. egelman , and h. wu .
structural architecture of the carma1/bcl10/malt1 signalosome : nucleation - induced filamentous assembly .
, 51(6):766779 , 2013 . | in this work , we introduce persistent homology for the analysis of cryo - electron microscopy ( cryo - em ) density maps .
we identify the topological fingerprint or topological signature of noise , which is widespread in cryo - em data . for low signal to noise ratio ( snr ) volumetric data , intrinsic topological features of biomolecular structures
are indistinguishable from noise .
to remove noise , we employ geometric flows which are found to preserve the intrinsic topological fingerprints of cryo - em structures and diminish the topological signature of noise .
in particular , persistent homology enables us to visualize the gradual separation of the topological fingerprints of cryo - em structures from those of noise during the denoising process , which gives rise to a practical procedure for prescribing a noise threshold to extract cryo - em structure information from noise contaminated data after certain iterations of the geometric flow equation . to further demonstrate the utility of persistent homology for cryo - em data analysis
, we consider a microtubule intermediate structure ( emd-1129 ) .
three helix models , an alpha - tubulin monomer model , an alpha- and beta - tubulin model , and an alpha- and beta - tubulin dimer model , are constructed to fit the cryo - em data .
the least square fitting leads to similarly high correlation coefficients , which indicates that structure determination via optimization is an ill - posed inverse problem .
however , these models have dramatically different topological fingerprints .
especially , linkages or connectivities that discriminate one model from another one , play little role in the traditional density fitting or optimization , but are very sensitive and crucial to topological fingerprints . the intrinsic topological features of the microtubule data are identified after topological denoising . by a comparison of the topological fingerprints of the original data and
those of three models , we found that the third model is topologically favored .
the present work offers persistent homology based new strategies for topological denoising and for resolving ill - posed inverse problems .
key words : cryo - em , topological signature , geometric flow , topological denoising , topology - aided structure determination . |
let @xmath4 be an affine kac - moody lie algebra with index set @xmath5 , and @xmath6 the corresponding quantum affine algebra without the degree operator . in @xcite
, it was conjectured that a certain subfamily of finite - dimensional @xmath6-modules known as kirillov - reshetikhin ( kr for short ) modules @xmath7 has crystal bases @xmath8 called kr crystals . here
the index @xmath9 corresponds to a node of @xmath10 where @xmath11 is the node specified in @xcite , and @xmath12 is a positive integer .
this conjecture has been confirmed in many cases , in particular when @xmath4 is nonexceptional @xcite .
let @xmath13 be a tensor product of kr crystals . in @xcite
, the authors defined the one - dimensional sum @xmath14,\ ] ] where @xmath15 is a dominant integral weight of the simple lie subalgebra @xmath16 whose dynkin nodes are @xmath17 , @xmath18 is a certain @xmath19-function on @xmath20 called the energy function , and @xmath21 are kashiwara operators .
then they conjectured that @xmath22 has an explicit expression @xmath23 called the fermionic form , where @xmath24 is the sequence of elements of @xmath25 corresponding to @xmath20 .
this conjecture is called the @xmath1 conjecture , which was confirmed in many instances @xcite , but a proof in full generality has not been available except for the type @xmath26 in @xcite . it should be mentioned that this conjecture was recently settled for nonexceptional @xmath4 under the large rank hypothesis by combining the results in @xcite , @xcite and @xcite , and for general @xmath4 if @xmath27 @xcite .
recently , an identity was proved in @xcite which connects one - dimensional sums of nonexceptional type with demazure operators .
let us recall the result briefly . for simplicity , we assume @xmath4 is of type @xmath2 or @xmath3 here .
denote by @xmath28 the weight lattice of @xmath4 and by @xmath29 for @xmath30 the demazure operator on the group ring @xmath31 $ ] defined by @xmath32 where the simple reflection @xmath33 acts on @xmath31 $ ] by @xmath34 .
let @xmath35 be the weyl group of @xmath4 , @xmath36 the extended affine weyl group , and @xmath37 the subgroup of the group of dynkin automorphisms such that @xmath38 ( for the precise definition , see section [ nontwisted affine lie algebra ] ) .
then for each @xmath39 and @xmath40 with a reduced expression @xmath41 , @xmath42 is defined by @xmath43 ( this definition does not depend on the choice of the expression ) .
let now @xmath44 be a tensor product of kr crystals such that @xmath45 , @xmath46 the set of dominant integral weights of @xmath47 , and @xmath48 the irreducible @xmath47-module with highest weight @xmath49
. then the following identity was proved in @xcite , where @xmath50 is some constant and we set @xmath51 : @xmath52 here @xmath53 is the null root , @xmath54 is the fundamental weight of @xmath4 associated with the node @xmath11 , @xmath55 are the fundamental weights of @xmath47 , @xmath56 is the longest element of the weyl group of @xmath47 , and @xmath57 are the translations .
the goal of this article is to show the identity ( [ eq : intro ] ) with the one - dimensional sums replaced by the corresponding fermionic forms .
namely , we will prove the following result as a corollary of theorem [ thm : intro ] stated below ( corollary [ cor : m - side ] ) : [ cor : m - side_in_intro ] assume that @xmath4 is of nontwisted type and @xmath47 is of @xmath0 type .
let @xmath58 be a sequence of elements of @xmath59 such that @xmath45 .
then we have @xmath60 with some constant @xmath50 , where we set @xmath51 . then as an immediate consequence of ( [ eq : intro ] ) and this corollary , the @xmath1 conjecture for @xmath2 and @xmath3 is settled ( theorem [ thm : x = m_conj ] ) .
our strategy of the proof of corollary [ cor : m - side_in_intro ] is to show the isomorphism between two modules whose characters are equal to the left hand side and the right hand side respectively .
let us describe what these modules are .
the module corresponding to the left hand side is the fusion product of kirillov - reshetikhin ( kr ) modules for the current algebra @xmath61 $ ] .
kr modules for @xmath61 $ ] , which we denote by @xmath62 ( @xmath63 , @xmath64 ) in this article , were defined in @xcite in terms of generators and relations .
they can also be obtained from @xmath7 ( kr modules for @xmath6 ) by specialization and restriction , which is why they are so named .
the fusion product is a refinement of the usual tensor product defined in @xcite , which constructs a cyclic graded ( @xmath61$])-module from some such modules .
it was proved by di francesco and kedem in @xcite that the fusion product of kr modules has the required ( graded ) character .
namely , they proved the following character formula for the fusion product of kr modules : for a sequence @xmath65 of elements of @xmath59 , we have @xmath66 where @xmath67 denotes the fusion product . to present the module corresponding to the right hand side , we recall the definition of joseph functors introduced by joseph @xcite .
let @xmath68 be a cartan subalgebra of @xmath4 , @xmath69 a borel subalgebra of @xmath4 containing @xmath68 , and @xmath70 for @xmath30 the parabolic subalgebra associated with the set @xmath71 containing @xmath69 .
we denote by @xmath72 ( resp .
@xmath73 ) the category of finite - dimensional @xmath68-semisimple left @xmath69-modules ( resp .
@xmath70-modules ) .
then the functor @xmath74 is defined as the left adjoint functor of the restriction functor @xmath75 . for @xmath40 with a reduced expression @xmath76 , the joseph functor @xmath77 is defined by @xmath78 which does not depend on the choice of the expression .
we also define @xmath79 for @xmath40 and @xmath39 by @xmath80 , where @xmath81 denotes the pull - back functor associated with the lie algebra automorphism @xmath82 of @xmath69 .
then the @xmath69-module corresponding to the right hand side of corollary [ cor : m - side_in_intro ] is constructed using these functors as follows : @xmath83 from the results of @xcite and @xcite , we can easily see that @xmath84 holds for @xmath85 if a @xmath69-module @xmath86 has a demazure flag ( see definition [ def : demazure_flag ] ) .
it is checked from this fact that the above module in fact has the required character .
now , what we have to prove is the following isomorphism , which is our main theorem in this article ( theorem [ thm : main_theorem ] ) : [ thm : intro ] assume that @xmath4 is of nontwisted type and @xmath47 is of type @xmath0 . let @xmath87 be a sequence of kr modules such that @xmath45 .
then there exists an isomorphism of @xmath69-modules @xmath88 with some constant @xmath50 , where the left hand side is naturally regarded as a @xmath69-module .
as explained above , corollary [ cor : m - side_in_intro ] follows as a direct consequence of this theorem .
it should be remarked that this theorem for @xmath89 is already proved by feigin and loktev in @xcite .
the plan of this article is as follows . in section [ section : demazure ] , we review the results on demazure modules and joseph functors . in section [ nontwisted affine lie algebra ] , we prepare some notation and elementary lemmas concerning a nontwisted affine lie algebra . in section [ section :
kr ] , we recall the definition of kr modules for a current algebra . for the later convenience ,
we define them as @xmath90-modules in this article , where @xmath90 denotes the parabolic subalgebra associated with @xmath17 . in section [ fusion product ] , we recall the definition of fusion products
. then we state our main theorem in section [ section : main ] , and explain in section [ x = m_conjecture ] how the @xmath1 conjecture for type @xmath2 and @xmath3 is deduced from this theorem .
the last two sections are devoted to prove the main theorem . in section [ section : b - fusion ] , we introduce the @xmath69-fusion product , which is some modified version of the fusion product constructing a @xmath69-module from some @xmath69-modules . then in section [ section : proof ] , we give the proof of the main theorem using @xmath69-fusion products .
+ * acknowledgments : * the author would like to express his gratitude to r. kodera , s. naito and y. saito for a lot of helpful discussions and comments .
let @xmath4 be a complex symmetrizable kac - moody lie algebra with index set @xmath5 and chevalley generators @xmath91 , @xmath92 its borel subalgebra , and @xmath93 its cartan subalgebra .
let @xmath94 @xmath95 be the simple roots , @xmath96 @xmath95 the simple coroots , and @xmath35 the weyl group of @xmath4 with simple reflections @xmath33 ( @xmath30 ) . denote by @xmath97 the weight lattice and the set of dominant integral weights respectively .
for a @xmath68-module @xmath86 , we denote by @xmath98 for @xmath99 the weight space @xmath100 we call a vector @xmath101 of @xmath86 a _ weight vector _ if @xmath102 for some @xmath99 .
denote by @xmath103 the irreducible highest weight @xmath4-module with highest weight @xmath104 . for @xmath105 ,
let @xmath106 be a nonzero vector of the one - dimensional weight space @xmath107 . for @xmath104 and @xmath40 ,
the @xmath69-submodule @xmath108 is called the _ demazure module _ associated with @xmath109 and @xmath110 . for a subset @xmath111
, we denote by @xmath112 the parabolic subalgebra associated with @xmath113 , which is the lie subalgebra of @xmath4 generated by @xmath69 and @xmath114 .
note that @xmath115 .
if @xmath116 , we denote @xmath112 by @xmath70 . in this article , we denote by @xmath117 the category of finite - dimensional @xmath68-semisimple left @xmath112-modules . if @xmath118 , then the demazure module @xmath119 is preserved by the action of @xmath120 .
hence @xmath119 belongs to @xmath117 if @xmath121 holds for all @xmath122 .
[ def : demazure_flag ] let @xmath123 .
a filtration @xmath124 is called a _ demazure flag _ of @xmath86 if every successive quotient @xmath125 is isomorphic to some demazure module . for every @xmath30
, it is known that the restriction functor @xmath75 has the left adjoint functor @xmath74 @xcite .
we often regard @xmath126 as a functor from @xmath72 to @xmath72 in the obvious way .
let @xmath40 , and @xmath41 be its arbitrary reduced expression .
then the functor @xmath127 does not depend on the choice of the expression .
the functor @xmath128 is called the _
joseph functor _ associated with @xmath40
. note that for every @xmath30 and @xmath123 , the canonical @xmath69-module homomorphism @xmath129 is defined as the image of the identity under the bijection @xmath130 .
the following lemma was proved in ( * ? ? ?
* subsection 2.7 and lemma 2.8 ( iv ) ) for finite - dimensional @xmath4 , and the proof goes without any change for general @xmath4 : [ lem : elementary_for_demazure ] let @xmath30 .
+ ( i ) for every @xmath131 , the canonical @xmath69-module homomorphism @xmath129 is an isomorphism . in particular , we have @xmath132 for every @xmath133 . + ( ii ) assume that @xmath134 is an exact sequence of objects of @xmath72 and @xmath135 is isomorphic to some @xmath69-submodule of a finite - dimensional @xmath70-module .
then the sequence @xmath136 is exact . for @xmath99 ,
we denote by @xmath137 the one - dimensional @xmath69-module spanned by a weight vector with weight @xmath109 on which @xmath138 ( @xmath30 ) acts trivially .
note that we have @xmath139 for @xmath104 .
now we recall the following theorem : [ thm : construction ] for every @xmath104 and @xmath40 , we have @xmath140 as @xmath69-modules .
[ cor : demazure_flag ] assume that @xmath123 has a demazure flag .
then @xmath141 has a demazure flag for every @xmath40 .
it is enough to show the assertion for @xmath142 .
let @xmath143 be a demazure flag of @xmath86 .
we show the assertion by the induction on @xmath144 . for a demazure module @xmath119
, we have from lemma [ lem : elementary_for_demazure ] ( i ) and theorem [ thm : construction ] that @xmath145 where @xmath12 denotes the length function .
hence the assertion for @xmath146 follows .
assume @xmath147 and write @xmath148 .
recall that @xmath149 is defined as a @xmath69-submodule of @xmath150 .
hence this module is also a @xmath69-submodule of the @xmath70-module @xmath151 , which is finite - dimensional since @xmath150 is integrable .
hence the sequence @xmath152 is exact by lemma [ lem : elementary_for_demazure ] ( ii ) , and then the assertion follows from the induction hypothesis .
let @xmath31 $ ] denote the group ring of @xmath28 with basis @xmath153 ( @xmath154 ) , and define for @xmath30 a linear operator @xmath29 on @xmath31 $ ] by @xmath32 where @xmath33 acts on @xmath31 $ ] by @xmath34 .
we call @xmath29 the _ demazure operator _ associated with @xmath155 .
the following lemma is proved by elementary calculations such as @xmath156 .
[ lem : invariance ] assume that @xmath157 $ ] is @xmath33-invariant
. then we have @xmath158.\ ] ] in particular , we have @xmath159 . for a finite - dimensional semisimple @xmath68-module @xmath86 such that @xmath160 , we denote the character of @xmath86 by @xmath161.\ ] ] for every reduced expression @xmath41 of @xmath40 , the operator @xmath162 on @xmath31 $ ] is independent of the choice of the expression ( * ? ? ?
* corollary 8.2.10 ) , and it is known that the character of a demazure module is expressed as follows : [ thm : character_formula ] for every demazure module @xmath119 , we have @xmath163 [ cor : flag_character ] assume that @xmath123 has a demazure flag .
then for every @xmath40 , we have @xmath164 by corollary [ cor : demazure_flag ] , it is enough to show the assertion for @xmath142 . let @xmath143 be a demazure flag of @xmath86 .
we show the assertion by the induction on @xmath144 .
assume @xmath146 , and write @xmath165 .
if @xmath166 , we have @xmath167 by theorem [ thm : construction ] and theorem [ thm : character_formula ] .
if @xmath168 , we have @xmath169 by ( [ eq : demazure ] ) , and @xmath170 by ( [ eq : demazure2 ] ) and lemma [ lem : invariance ] .
hence the assertion for @xmath146 follows .
assume @xmath147 and write @xmath171 .
then as proved in the proof of corollary [ cor : demazure_flag ] , the sequence @xmath152 is exact .
hence the assertion follows from the induction hypothesis .
the following theorem is obtained by taking the classical limit of ( * ? ? ?
* theorem 5.22 ) : [ thm : joseph ] assume that @xmath4 is symmetric ( i.e. the cartan matrix of @xmath4 is symmetric ) .
then for every @xmath172 and @xmath40 , the @xmath69-module @xmath173 has a demazure flag . in ( * ? ? ?
* theorem 5.22 ) , a given kac - moody lie algebra is assumed to be simply laced .
this assumption , however , is used only in ( * ? ? ?
* lemma 3.14 ) to apply a positivity result of lusztig , and we can check that the proof of this positivity result in @xcite goes without any change for every symmetric kac - moody lie algebra .
hence ( * ? ? ?
* theorem 5.22 ) holds for symmetric kac - moody lie algebras , and so does the above theorem . from the theorem
, we have the following corollary since the functor @xmath174 is exact : [ cor : joseph ] assume that @xmath4 is symmetric .
if @xmath123 has a demazure flag , then @xmath175 for @xmath104 has a demazure flag .
from this section , we assume that @xmath4 is a nontwisted affine lie algebra with @xmath176 unless stated otherwise . let @xmath177 be the dynkin diagram of @xmath4 , @xmath178 the cartan matrix of @xmath4 , @xmath179 the root system of @xmath4 , and @xmath180 the set of positive roots corresponding to @xmath69 . in this article
, we use the kac s labeling of nodes of @xmath177 in ( * ? ? ?
* section 4.8 ) .
let @xmath181 ( resp .
@xmath182 ) be the unique sequence of relatively prime positive integers satisfying @xmath183 let @xmath184 be the degree operator , which is any element satisfying @xmath185 for @xmath30 , @xmath186 the canonical central element , and @xmath187 the null root . for each @xmath30 ,
let @xmath188 be the fundamental weight , which satisfies @xmath189 note that we have @xmath190 let @xmath191 be the @xmath35-invariant symmetric bilinear form on @xmath192 defined by @xmath193 we denote by @xmath194 ( resp .
@xmath195 ) the lie subalgebra of @xmath4 generated by @xmath196 ( resp .
@xmath197 ) , and by @xmath198 for @xmath199 the root space of @xmath4 .
let @xmath200 , and @xmath16 be the simple lie subalgebra generated by @xmath201 with cartan subalgebra @xmath202 and weyl group @xmath203 .
let @xmath204 be the root system of @xmath47 , @xmath205 the set of positive roots , @xmath206 the weight lattice of @xmath47 , @xmath207 the set of dominant integral weights , @xmath208 the root lattice of @xmath47 , and @xmath209 .
denote by @xmath210 and @xmath211 ( @xmath212 ) the fundamental weights and the fundamental coweights of @xmath47 respectively .
for the notational convenience , we set @xmath213 and @xmath214 . we often regard @xmath215 as a subspace of @xmath192 by setting @xmath216 , @xmath217 .
then we have @xmath218 .
we denote by @xmath219 the highest root of @xmath47 , and by @xmath56 the longest element of @xmath220 .
let @xmath221 ( resp .
@xmath222 ) be the lie subalgebra of @xmath47 generated by @xmath223 ( resp .
@xmath224 ) . set @xmath225 and @xmath226 for @xmath212 , and for each @xmath227 fix nonzero vectors
@xmath228 , @xmath229 so that @xmath230 = \ga^{\vee } , \ \ \ [ \ga^{\vee } , e_{\pm \ga } ] = \pm 2 e_{\pm\ga}.\ ] ] recall that there exists a unique lie algebra isomorphism @xmath231 \oplus \c k \oplus
\c d\ ] ] satisfying @xmath232 the lie algebra structure of @xmath233 \oplus \c k \oplus \c d$ ] is defined by @xmath234 \\ & = [ x , y ] \otimes t^{m+n } + n b_1 y \otimes t^{n } - m b_2 x \otimes t^{m } + m\gd_{m , -n } ( x , y)k.\end{aligned}\ ] ] in the sequel , we always identify these two lie algebras via the above isomorphism .
it should be noted that we have @xmath235 \oplus \c k \oplus \c d.\ ] ] set @xmath236 $ ] , @xmath237 note that we have @xmath238 and @xmath239 .
let @xmath240 denote the canonical projection , and set @xmath241 .
since @xmath35 fixes @xmath53 , @xmath35 acts on @xmath242 and @xmath243 . for @xmath244 , we denote by @xmath245 the subset @xmath246 of @xmath243 . as ( * ? ? ?
* ( 6.5.2 ) ) , we define for @xmath247 an endomorphism @xmath248 of @xmath192 by @xmath249 the map @xmath250 defines an injective group homomorphism from @xmath251 to the group of linear automorphisms of @xmath192 orthogonal with respect to @xmath191 .
let @xmath252 for @xmath212 , and define the sublattices @xmath86 and @xmath253 of @xmath251 by @xmath254 let @xmath255 and @xmath256 be the subgroups of @xmath257 defined by @xmath258 it is known that @xmath259 ( * ? ? ?
* proposition 6.5 ) .
define the subgroup @xmath260 of @xmath257 by @xmath261 which is called the _ extended affine weyl group_. the action of @xmath36 preserves @xmath262 , and @xmath263 and @xmath264 satisfy @xmath265 by @xmath266 we denote the group of automorphisms of the dynkin diagram @xmath177 , that is , the group of permutations @xmath267 of @xmath5 satisfying @xmath268 for all @xmath269 .
let @xmath270 , @xmath271 be the fundamental chamber , and @xmath272 the subgroup consisting of elements preserving @xmath50 .
then we have @xmath273 since @xmath39 preserves the set of simple roots , @xmath267 induces a permutation of @xmath5 ( also denoted by @xmath267 ) by @xmath274 , which belongs to @xmath275 since @xmath191 is @xmath267-invariant . by abuse of notation , we denote by @xmath37 both the subgroups of @xmath36 and @xmath275 .
[ lem : transform ] let @xmath267 be an arbitrary element of @xmath37 , and @xmath276 the unique element of @xmath220 such that @xmath277
. then we have @xmath278 as @xmath279 mod @xmath280 for every @xmath281 and @xmath282 , we have @xmath283 mod @xmath280 .
this forces @xmath284 for @xmath212 since @xmath267 preserves @xmath285 and @xmath286 .
now the assertion follows since @xmath287 we define an action of @xmath37 on @xmath4 by letting @xmath39 act as a lie algebra automorphism defined by @xmath288 this action obviously preserves @xmath69 and @xmath289 . for a module @xmath86 , we denote by @xmath290 the pull - back of @xmath86 with respect to @xmath267 . the image of @xmath291 under the canonical linear isomorphism @xmath292 is denoted by @xmath293 .
note that this isomorphism maps the weight space @xmath294 onto @xmath295 .
we prepare some notation here .
let @xmath85 be an arbitrary element , and take unique elements @xmath296 and @xmath297 so that @xmath298 . then define a functor @xmath299 by @xmath300 and a linear operator @xmath301 \to \z[p]$ ] by @xmath302 where @xmath267 acts on @xmath31 $ ] by @xmath303 .
we set @xmath304 then we have the following lemma : [ lem : affine_version ] ( i ) let @xmath39 , @xmath305 and @xmath40 .
then we have @xmath306 ( ii ) assume @xmath123 has a demazure flag .
then for every @xmath85 , @xmath141 has a demazure flag and we have @xmath307 since @xmath308 is an integrable highest weight @xmath4-module with highest weight @xmath309 , we have @xmath310 .
moreover , we see that @xmath311 is the @xmath69-submodule of @xmath312 generated by the weight space with weight @xmath313 .
hence the assertion ( i ) follows .
then , since the functor @xmath314 is exact , the first assertion of ( ii ) follows from ( i ) and corollary [ cor : demazure_flag ] .
since we have @xmath315 , the second one follows from corollary [ cor : flag_character ] .
let @xmath316 $ ] denote the group algebra of @xmath243 with basis @xmath317 ( @xmath318 ) , and by @xmath319 we also denote the projection from @xmath31 $ ] to @xmath316 $ ] defined by @xmath320 . for each @xmath40 , a linear operator @xmath321 on @xmath316 $ ]
is defined by @xmath322 [ lem : final_lem ] let @xmath323 $ ] , and assume @xmath324 is @xmath220-invariant .
then we have @xmath325 \ \text{and } \ w
\in w.\ ] ] since @xmath323 $ ] , we have @xmath326 , where @xmath327 denotes the reflection associated with @xmath328 .
hence @xmath324 is @xmath33-invariant for all @xmath30 , and then the assertion is easily proved from lemma [ lem : invariance ] .
we need the following elementary lemmas later : [ lem : trivial_lem ] let @xmath123 . for every @xmath85 and @xmath329
, we have @xmath330 it is enough to show the assertions for @xmath331 and @xmath332 .
the first case is obvious , and the second one follows since we have @xmath333 for every @xmath334 .
[ lem : trivial_lem2 ] let @xmath335 be an object of @xmath72 which is generated by a weight vector @xmath336 , @xmath337 an object of @xmath72 , and @xmath338 a homomorphism of @xmath289-modules from @xmath335 to @xmath337 which maps @xmath336 to a weight vector @xmath339 .
( i ) for some @xmath329 , @xmath338 induces a homomorphism of @xmath69-modules from @xmath340 to @xmath337 .
+ ( ii ) assume further that @xmath337 extends to an object of @xmath73 for some @xmath30 .
then there exists a homomorphism @xmath341 of @xmath342-modules from @xmath343 to @xmath337 satisfying @xmath344 , where @xmath345 is the canonical homomorphism of @xmath69-modules .
let @xmath346 be the respective weights of @xmath347 , and @xmath348 .
it is easily seen that the induced map @xmath349 preserves the weights , and hence it is a homomorphism of @xmath69-modules .
the assertion ( i ) is proved . then under the assumption of ( ii )
, there exists a homomorphism @xmath350 of @xmath70-modules such that @xmath351 since @xmath126 is left adjoint to the restriction functor @xmath75 . since @xmath352 holds by lemma [ lem : trivial_lem ] , required @xmath341 is obtained by restricting @xmath353 to @xmath342 .
following @xcite , we define the following @xmath90-modules : [ def : kr_module ] for @xmath63 and @xmath354 , let @xmath62 be the @xmath90-module generated by a nonzero vector @xmath355 with relations @xmath356 ) v_{r,\ell } & = 0 , \ \ \ ( \fh_0 \otimes t\c[t])v_{r,\ell } = 0 , \ \ \
hv_{r,\ell } = \langle \ell \varpi_r , h \rangle v_{r,\ell } \ \ \text{for } \ h \in \fh , \\ f_r^{\ell+1}v_{r,\ell } & = ( f_r\otimes t)v_{r,\ell } = 0 \ \ \
\text{and } \ \ \ f_{i}v_{r,\ell } = 0 \ \ \text{for } \ i \in i_0 \setminus \{r\}. \end{aligned}\ ] ] we call @xmath62 the _ kirillov - reshetikhin module _ ( kr module for short ) for @xmath90 associated with @xmath9 and @xmath12 .
\(i ) kirillov - reshetikhin modules were originally defined in @xcite as ( @xmath357$])-modules . since we would like to consider @xmath62 as a @xmath69-module later , we adopt the above definition in this article .
it is obvious that the restriction of @xmath62 to @xmath357 $ ] coincides with the original one .
+ ( ii ) as the name indicates , @xmath62 has strong connections with the kirillov - reshetikhin module @xmath7 for the quantum affine algebra @xmath6 @xcite .
we return to this topic in section [ x = m_conjecture ] .
it is easily seen that the @xmath47-submodule @xmath358 is isomorphic to the irreducible module with highest weight @xmath359 . for each @xmath360 ,
take and fix a nonzero vector @xmath361 of this @xmath47-submodule whose weight is @xmath362 , and set @xmath363 .
[ lem : kr ] ( i ) let @xmath263 .
for @xmath364 and @xmath365 , we have @xmath366 ( ii ) as a @xmath69-module , @xmath62 is generated by @xmath367 .
the assertion ( i ) for @xmath368 follows by definition . since @xmath62 is finite - dimensional ( * ? ? ?
* theorem 1.2.2 ) , its weight set and dimensions of weight spaces are @xmath220-invariant . from this , ( i ) for general @xmath110 is proved . then ( i ) for @xmath369 implies ( ii ) . the following theorem , which easily follows from ( * ? ? ?
* theorem 4 ) , connects kr modules with demazure modules .
[ thm : kr - demazure ] let @xmath9 be an arbitrary element of @xmath17 and @xmath12 a positive integer satisfying @xmath370 .
then there exists an isomorphism @xmath371 of @xmath372-modules which maps @xmath373 to @xmath374 .
if @xmath47 is of @xmath0 type , then @xmath375 holds for all @xmath63 .
hence the above isomorphism follows for all kr modules in this case .
the fusion product was defined in @xcite as a construction of a graded cyclic ( @xmath61$])-module . here we slightly reformulate it as a construction of a @xmath90-module .
note that @xmath376 has a natural grading defined by @xmath377 = kx \},\ ] ] from which we define a natural filtration on @xmath376 by @xmath378 let @xmath86 be a cyclic @xmath372-module with a generator @xmath101 , and denote by @xmath379 for @xmath380 the subspace @xmath381 of @xmath86 .
( note that @xmath382 . )
then the associated @xmath90-module @xmath383 is defined by @xmath384 where @xmath385 acts on @xmath386 as multiplication by @xmath144 .
now we recall the definition of fusion products .
let @xmath387 be a sequence of objects of @xmath388 such that each @xmath389 is generated by a weight vector @xmath390 , and @xmath391 pairwise distinct complex numbers . for each @xmath392
, define @xmath372-module @xmath393 by the pullback @xmath394 , where @xmath395 is an automorphism of @xmath372 defined by @xmath396 for @xmath397 , we denote by @xmath398 its image under the canonical map @xmath399 . as shown in (
* proposition 1.4 ) , @xmath400 is a cyclic @xmath372-module generated by @xmath401 , and we define a @xmath90-module @xmath402 by @xmath403 when the parameters or generators are clear from the context , we write simply as @xmath404 , @xmath405 or @xmath406 . the @xmath90-module @xmath407 is called the _ fusion product_. [ lem : elementary ] ( i ) as @xmath408-modules , @xmath409 ( ii ) let @xmath410 be weight vectors of @xmath411 respectively , and assume that @xmath412 then we have @xmath413 ( iii ) for any @xmath414 , we have @xmath415 ( iv ) @xmath416 ( the fusion product of a single module @xmath417 ) is isomorphic to @xmath417 as a @xmath372-module . the assertions ( i ) and ( ii ) easily follow from the definition . as the automorphism @xmath418 preserves the subspace @xmath419 , we have @xmath420 since the left hand side is equal to the filtration of @xmath421 , ( iii ) is proved . when @xmath422 , ( iv ) obviously follows
. then the assertion follows in general by ( iii ) .
now we state the main theorem of this article .
this is a generalization of ( * ? ? ?
* theorem 2.5 ) , in which the case @xmath423 is proved .
[ thm : main_theorem ] assume that @xmath47 is of @xmath0 type .
let @xmath87 be a sequence of kr modules such that @xmath45 , and set @xmath424 for @xmath392 .
then for arbitrary pairwise distinct complex numbers @xmath425 , there exists an isomorphism of @xmath69-modules @xmath426 with some constant @xmath50 .
we postpone the proof of this theorem to the latter part of this article .
we see from corollary [ cor : joseph ] and lemma [ lem : affine_version ] ( ii ) that the right hand side of ( [ thm : main_theorem ] ) has a demazure flag .
hence we can prove inductively using lemma [ lem : affine_version ] ( ii ) that the following equation holds : [ cor : main_cor ] under the notation and the assumptions of theorem [ thm : main_theorem ] , we have @xmath427 the right hand side of ( [ eq : main_theorem ] ) also appeared in @xcite . in the article , it was proved that this module , which was called a _ generalized demazure module _ ,
is isomorphic to the space of global sections of a certain line bundle on a _ bott - samelson variety_.
in this section , we give an important application of theorem [ thm : main_theorem ] , the proof of the @xmath1 conjecture for type @xmath2 and @xmath3 . here
, we assume that @xmath4 is a general ( possibly twisted ) affine lie algebra . for
a sequence @xmath428 of elements of @xmath59 and @xmath49 , denote by @xmath429 $ ] the associated _ fermionic form _
( see @xcite for definition , in which the fermionic form is denoted by @xmath430 with @xmath431 ) .
the most important result concerning fermionic forms in this article is the following theorem proved by di francesco and kedem using the result of @xcite : [ thm : dfk ] assume that @xmath4 is of nontwisted type . for a sequence @xmath58 of elements of @xmath59 and pairwise distinct complex numbers
@xmath432 , we have @xmath433 where we set @xmath51 and denote by @xmath48 the irreducible @xmath47-module with highest weight @xmath15 . in @xcite ,
the above theorem was proved under the assumption that the dimension of each @xmath434 is equal to that of the corresponding kr - module @xmath435 for the quantum affine algebra @xmath6 .
we can see from the pentagon of identities in ( * ? ? ?
* subsection 1.2 ) that this assumption in fact holds in general . from corollary [ cor : main_cor ] and theorem [ thm : dfk ]
, we have the following corollary : [ cor : m - side ] assume that @xmath4 is of nontwisted type and @xmath47 is of @xmath0 type .
let @xmath58 be a sequence of elements of @xmath59 such that @xmath45 .
then we have @xmath60 with some constant @xmath50 , where we set @xmath51 .
next , we recall the definition of one - dimensional sums
. denote by @xmath8 the _ kirillov - reshetikhin crystal _
( kr crystal for short ) associated with @xmath436 and @xmath437 . for nonexceptional @xmath4 , it is known that @xmath8 is perfect if and only if @xmath438 @xcite .
let @xmath13 be a tensor product of kr crystals , and @xmath439 the _ energy function _
defined on @xmath20 . for the definitions of these objects ,
see @xcite .
the _ one - dimensional sum _ @xmath440 $ ] for @xmath49 is defined by @xmath441 where @xmath21 are kashiwara operators . in (
* corollary 7.3 ) , the following proposition was proved : [ prop : x - side ] assume that @xmath4 is of nonexceptional type .
let @xmath442 be a tensor product of perfect kr crystals such that @xmath45 .
then we have @xmath443 with some constant @xmath444 , where we set @xmath51 .
now we show the following theorem from the above results , which is the @xmath1 conjecture presented in @xcite .
this result for @xmath3 is new .
this has already been proved for @xmath2 in @xcite , but our approach is quite different from theirs : [ thm : x = m_conj ] assume that @xmath4 is of nontwisted , nonexceptional type and @xmath47 is of @xmath0 type ( i.e. @xmath445 or @xmath3 ) .
let @xmath428 be a sequence of elements of @xmath59 , and @xmath44 .
then for every @xmath49 , we have @xmath446 where @xmath447 denotes the unique element of @xmath20 whose weight is @xmath448 .
recall that energy functions and one - dimensional sums are invariant under reordering of the given sequence by ( * ? ? ?
* proposition 2.15 ) , and so are fermionic forms by definition .
hence we may assume @xmath45 .
then as @xmath449 holds for all @xmath436 , we have from corollary [ cor : m - side ] and proposition [ prop : x - side ] that @xmath450 with some constant @xmath451 , which implies @xmath452 for every @xmath15 since the characters of irreducible @xmath47-modules are linearly independent .
it remains to show @xmath453 .
let @xmath454 .
since @xmath455 holds by definition , the coefficient of @xmath317 in the left hand side of ( [ eq : identity ] ) is equal to @xmath456 . on the other hand
, we easily see from theorem [ thm : dfk ] that the coefficient of @xmath317 in the right hand side is @xmath457 .
hence we have @xmath458 , which implies @xmath453 .
the theorem is proved .
we devote the rest of this article to prove theorem [ thm : main_theorem ] . in this section
, we introduce a construction of a @xmath69-module , which we call the @xmath69-fusion product , defined by modifying the definition of the fusion product in section [ fusion product ] .
this construction is essentially used in the proof of the theorem .
here we only assume that @xmath4 is a nontwisted affine lie algebra ( that is , @xmath47 is allowed to be of type @xmath459 ) since the definition of @xmath69-fusion products makes sense in this setting .
let @xmath460 be a sequence of objects of @xmath388 such that each @xmath389 is generated ( as a @xmath90-module ) by a weight vector @xmath390 , and @xmath461 an object of @xmath72 which is generated ( as a @xmath69-module ) by a weight vector @xmath462 . for pairwise distinct _
nonzero _ complex numbers @xmath463 , define a @xmath372-module @xmath464 as in section [ fusion product ] .
[ lem : cyclicity ] @xmath465 is generated by the vector @xmath466 as a @xmath289-module .
the proof is similar to that of ( * ? ? ?
* proposition 1.4 ) .
here we give it for completeness .
let @xmath467 and @xmath468 for @xmath469 be arbitrary vectors , and set @xmath470 ( recall that @xmath471 denotes the image of @xmath472 under the canonical map ) .
first we show that for every @xmath473 , @xmath474 and @xmath475 , there exists a vector @xmath476 \in \fb'$ ] such that @xmath477\wti{w}= z \otimes w_1'\otimes \cdots \otimes \big((x\otimes t^q ) w_j\big ) ' \otimes \cdots \otimes w_p'.\ ] ] since @xmath461 and @xmath389 are finite - dimensional and @xmath68-semisimple , there exists a sufficiently large positive integer @xmath478 such that @xmath479 $ ] acts trivially on them .
since @xmath476 = 0 $ ] satisfies the above equation for @xmath480 , we may assume @xmath481 .
note that we have for @xmath482 that @xmath483 consider the matrix of the coefficients : @xmath484 since the determinant of this matrix is equal to @xmath485 we see that this matrix is invertible from the assumption that @xmath432 are nonzero and pairwise distinct . hence @xmath476 $ ] can be obtained as a linear combination of @xmath486 ( @xmath487 ) , and our assertion is proved . from this , we can also prove that for every @xmath488 , there exists a vector @xmath489 satisfying @xmath490 now the lemma obviously follows .
define the subspace @xmath491 for @xmath365 similarly as @xmath419 . by considering the filtration @xmath492 of @xmath465
, we define @xmath493_{\fb}$ ] as the associated @xmath69-module @xmath494_{\fb } = \mathrm{gr}_{\wti{f}_{u \otimes v_1'\otimes \cdots \otimes v_p'}}(n\otimes m_{c_1}^1 \otimes \cdots \otimes m_{c_p}^p),\ ] ] which we call the _ @xmath69-fusion product_. we sometimes omit the parameters or the generators when they are clear from the context .
it should be noted that in the definition of the @xmath69-fusion product , only the leftmost module is allowed to be a @xmath69-module , and the others are assumed to be @xmath90-modules . by definition
, it is easily seen for every @xmath495 that @xmath496_{\fb } \cong \left[(\c_{\ell \gl_0 } \otimes n ) * m^1 * \cdots * m^p\right]_{\fb}.\ ] ] in some special cases , an original fusion product is connected to a certain @xmath69-fusion product by the following lemma : [ lem : isomorphim_of_fusion ] assume that each generator @xmath390 of @xmath389 is annihilated by @xmath497 .
then we have the following isomorphisms of @xmath69-modules : + ( i ) @xmath498_{\fb } , \ \text{and}\ ] ] ( ii ) @xmath499_\fb,\ ] ] where @xmath500 denotes the trivial module . from the assumption ,
@xmath417 is generated by @xmath336 as a @xmath69-module .
hence the right hand side of the isomorphism ( i ) makes sense .
the isomorphisms easily follow from the definition since we have @xmath501 by the assumption and the poincar - birkhoff - witt theorem
. the isomorphism ( ii ) of the above lemma does not hold in general .
for example , let @xmath417 be a finite - dimensional irreducible @xmath47-module with a highest weight vector @xmath336 , which is considered as a @xmath90-module via the evaluation map @xmath502 . then by lemma
[ lem : elementary ] ( iv ) , the fusion product @xmath416 is isomorphic to @xmath417 .
however , we easily see that the degree @xmath11 space of @xmath503_{\fb}$ ] is one - dimensional , and hence they are not isomorphic unless @xmath417 is trivial . on the other hand , @xmath504_{\fb}$ ]
is isomorphic to @xmath417 if a lowest weight vector of @xmath417 is chosen as a generator . as seen from this example , the @xmath69-fusion product is sensitive to the choice of generators .
[ lem : important ] let @xmath30 .
the @xmath69-module @xmath505_{\fb}$ ] extends to a @xmath70-module if @xmath461 extends to a @xmath70-module and either of the following conditions is satisfied : + ( i ) @xmath212 and all @xmath390 and @xmath462 are annihilated by @xmath120 , or + ( ii ) @xmath506 , @xmath507 acts trivially on @xmath508 , and each @xmath390 ( resp .
@xmath462 ) is annihilated by @xmath509 $ ] ( resp .
@xmath510 ) . the case ( i ) is easily proved since @xmath465 is a @xmath342-module and we have @xmath511 for each @xmath144 .
let us prove the case ( ii ) .
since @xmath507 acts trivially , a @xmath512-module structure is defined on each @xmath513 by letting @xmath514 act by @xmath515 note that the above sum is finite since @xmath516 .
hence @xmath465 extends to a @xmath517-module .
moreover the equality ( [ eq : coincidence ] ) also holds in this case from the assumption .
hence the assertion also follows in this case .
take arbitrary vectors @xmath468 ( @xmath469 ) and @xmath467 .
let @xmath144 be the unique integer such that @xmath518 and denote by @xmath519 the vector of @xmath505_{\fb}$ ] which is the image of @xmath520 under the projection @xmath521 note that @xmath522 is a generator of @xmath505_{\fb}$ ] .
the following lemma , which obviously follows by definition , is important for the later arguments .
[ lem : annihilating ] let @xmath523
. then @xmath524 annihilates @xmath525 if and only if there exists some @xmath526 satisfying @xmath527
now , we begin the proof of theorem [ thm : main_theorem ] .
assume that @xmath4 is a nontwisted affine lie algebra and @xmath47 is of type @xmath0 . for a given sequence of kr modules @xmath528
, we set @xmath529 and @xmath530 for @xmath474 for short , and write @xmath531 for @xmath263 ( defined in section [ section : kr ] ) .
we shall show the theorem by the induction on @xmath532 .
the assertion of the theorem for @xmath533 follows from lemma [ lem : elementary ] ( iv ) , theorem [ thm : construction ] , theorem [ thm : kr - demazure ] , and lemma [ lem : trivial_lem2 ] ( i ) .
assume @xmath534 .
by lemma [ lem : elementary ] ( iii ) , we may ( and do ) assume @xmath535 , which implies @xmath536 are nonzero .
first we show the following lemma : since @xmath539 holds , we have from lemma [ lem : elementary ] ( ii ) that @xmath540 whose right hand side is isomorphic to @xmath541_{\fb}$ ] by lemma [ lem : isomorphim_of_fusion ] ( i ) .
hence we have using ( [ eq : level ] ) and theorem [ thm : kr - demazure ] that @xmath542_{\fb}\\ & \ \cong\big [ v_{t_{w_0(\varpi_{r_p})}}(\ell_p\gl_0)_{u_{t_{w_0(\varpi_{r_p})}(\ell_p\gl_0 ) } } * m^{p-1}_{v_{p-1}^{w_0 } } * \cdots * m^1_{v_1^{w_0}}\big]_{\fb}. \end{aligned}\ ] ] the isomorphism ( [ eq : isom1 ] ) is proved .
let us prove ( [ eq : isom2 ] ) .
by the induction hypothesis , there exists an isomorphism @xmath543 whose right hand side is isomorphic to @xmath544_\fb \end{aligned}\ ] ] by lemma [ lem : elementary ] ( ii ) and lemma [ lem : isomorphim_of_fusion ] ( ii ) .
hence we have using ( [ eq : level ] ) that @xmath545_{\fb}\big ) \\ & \cong \cd_{t_{w_0(\varpi_{r_p})}}\left[\c_{\ell_p\gl_0 } * m^{p-1}_{v_{p-1}^{w_0 } } * \cdots * m^1_{v_1^{w_0}}\right]_{\fb}. \end{aligned}\ ] ] the isomorphism ( [ eq : isom2 ] ) is proved . by lemmas [ lem : translation ] and [ lem : trivial_lem2 ] ( i ) , in order to prove the theorem it suffices to show the following isomorphism of @xmath289-modules : @xmath546_{\fb } \nonumber \\ & \ \ \ \ \
\cong \cd_{t_{w_0(\varpi_{r_p})}}\big [ \c_{\ell_p\gl_0 } * m^{p-1}_{v_{p-1}^{w_0 } } * \cdots * m^1_{v_1^{w_0}}\big]_{\fb}.\end{aligned}\ ] ] let @xmath40 and @xmath39 be unique elements satisfying @xmath547 , and @xmath548 a reduced expression .
for @xmath549 , let @xmath550 , and @xmath551 be the unique element of @xmath220 satisfying @xmath552 ( note that @xmath553 ) .
we also write @xmath554 for @xmath555 .
then since @xmath556 holds , we see that the isomorphism ( [ eq : translated_isom ] ) is deduced by the induction on @xmath557 from the following two propositions , and hence theorem [ thm : main_theorem ] is established : [ prop2 ] for each @xmath559 , we have @xmath560_{\fb}\\ & \cong \cd_{i_q}\big [ v_{w^{q-1}}(\ell_p\gl_0)_{u_{w^{q-1}(\ell_p\gl_0 ) } } * m_{v_{p-1}^{\ol{w}^{q-1}w_0}}^{p-1 }
* \cdots * m_{v_1^{\ol{w}^{q-1}w_0}}^1 \big]_{\fb } \end{aligned}\ ] ] as @xmath289-modules .
[ lem : kac ] let @xmath86 be a finite - dimensional @xmath372-module .
then for every @xmath263 , there exists a linear automorphism @xmath561 of @xmath86 satisfying @xmath562 where @xmath563 are some nonzero complex numbers which do not depend on @xmath86 .
[ lem:1 - 0 ] the annihilating ideal of @xmath565 in @xmath566 is generated by @xmath567 , \ \ \
\fh_0 \otimes t\c[t],\ \ \
h- \langle w_0(\ell\varpi_r ) , h \rangle \ \ ( h \in \fh ' ) , \\ & e_{\bar{r}}^{\ell+1},\ \ \
e_{\bar{r}}\otimes t , \ \
\ \text{and } \ \ e_i \ \ ( i \in i_0\setminus\{\bar{r}\ } ) , \end{aligned}\ ] ] where @xmath568 is the node of @xmath17 such that @xmath569 . [ lem:1 - 1 ] for @xmath414 ,
the annihilating ideal of @xmath570 in @xmath571 is generated by @xmath572,\ \ \
\fh_0 \otimes ( t - c)\c[t ] , \ \ \
h- \langle w_0(\ell\varpi_r ) , h \rangle \ \ ( h \in \fh'),\\ & e_{\bar{r}}^{\ell+1 } , \
e_{\bar{r } } \otimes ( t - c ) , \ \
\ \text{and}\ \ e_i \ \
( i \in i_0 \setminus \{\bar{r}\ } ) . \end{aligned}\ ] ] let @xmath5 be the subspace of @xmath571 spanned by the above vectors . from lemma [ lem:1 - 0 ] , we see that the annihilating ideal of @xmath573 in @xmath566 is equal to @xmath574 .
we have to prove that @xmath575 since @xmath576 holds by the poincar - birkhoff - witt theorem , it suffices to show that @xmath577 then since we have @xmath578 and @xmath579 is generated by @xmath224 , it is enough to prove that @xmath580 holds for @xmath212 , which is proved by elementary calculations .
[ lem : annihilators ] there exists a nonzero complex number @xmath581 satisfying the following statement : for every kr module @xmath62 and @xmath414 , there exists a homomorphism @xmath582 of @xmath289-modules which maps @xmath573 to @xmath583 .
since @xmath584 is generated by @xmath573 as a @xmath289-module , it suffices to show for suitable @xmath585 that if @xmath586 annihilates @xmath587 , then @xmath588 annihilates @xmath589 . since the @xmath590-weights of @xmath591 and @xmath592 coincide
, we may assume @xmath593 .
let @xmath594 , and @xmath595 be an algebra homomorphism defined by @xmath596 using lemma [ lem:1 - 1 ] , we easily check that @xmath597 also annihilates @xmath591 .
let @xmath598 be the linear automorphism of @xmath584 given in lemma [ lem : kac ] .
then since @xmath599 holds by lemma [ lem : kr ] , @xmath600 annihilates @xmath601 , which is equivalent to that @xmath602 annihilates @xmath603 ( @xmath418 is defined in section [ fusion product ] ) .
it is easy to see from lemma [ lem : transform ] that there exists some @xmath604 for each @xmath30 such that @xmath605 define a linear automorphism @xmath606 on @xmath62 by @xmath607 where @xmath608 are the fundamental coweights of @xmath4 .
@xmath609 annihilates @xmath610 since it is a weight vector .
set @xmath611 .
it is easily checked that @xmath612 which implies @xmath613 hence we see that @xmath614 annihilates @xmath610 , which is equivalent to that @xmath588 annihilates @xmath589 .
the assertion is proved .
now , we give the proof of proposition [ prop1 ] : + _ proof of proposition [ prop1 ] .
_ note that the right hand side of theorem [ thm : main_theorem ] does not depend on the parameters @xmath463 .
hence from the induction hypothesis on @xmath532 and the proof of ( [ eq : isom2 ] ) , we see that the @xmath69-module @xmath615_{\fb}$ ] also does not depend on the parameters , and in particular we have @xmath616_{\fb } \cong \big[\c_{\ell_p\gl_0 } * m_{b^{-1}c_{p-1},v_{p-1}^{w_0}}^{p-1 } * \cdots * m_{b^{-1}c_1,v_1^{w_0}}^1\big]_{\fb},\ ] ] where @xmath581 is the complex number given in lemma [ lem : annihilators ] .
hence the proposition is equivalent to the following isomorphism of @xmath289-modules : @xmath558_{\fb } \\ &
\cong ( \tau^{-1})^*\left[\c_{\ell_p\gl_0 } * m_{b^{-1}c_{p-1},v_{p-1}^{w_0}}^{p-1 } * \cdots * m_{b^{-1}c_1,v_1^{w_0}}^1 \right]_{\fb}. \end{aligned}\ ] ] let us prove this .
let @xmath617 and @xmath618 be nonzero vectors . since dimensions of two modules are equal , it suffices to show there exists a surjective homomorphism of @xmath289-modules from the right hand side to the left hand side mapping @xmath619 to @xmath620 , which is equivalent to show that if @xmath586 annihilates @xmath621 , then @xmath588 annihilates @xmath620 .
we may assume @xmath622 for some @xmath623 and @xmath624 , where we set @xmath625 = \langle \gg , h \rangle z \ \text{for } \ h \in \fh_0 \right\}.\ ] ] by lemma [ lem : annihilating ] , there exists @xmath626 such that @xmath627 and we may assume @xmath628 .
we see from lemma [ lem : annihilators ] that there exists a homomorphism of @xmath289-modules @xmath629 which maps @xmath630 to @xmath631 . from this and ( [ eq :
trivial ] ) , we have @xmath632 moreover , we have @xmath633 and @xmath634 from lemma [ lem : transform ] , where we set @xmath635 .
hence by lemma [ lem : annihilating ] , @xmath588 annihilates @xmath636 . the assertion is proved .
+ it remains to prove proposition [ prop2 ] : + _ proof of proposition [ prop2 ] .
_ we abbreviate @xmath637 , @xmath638 and @xmath639 .
fix @xmath640 , and assume when @xmath641 that the assertion of the proposition holds for @xmath642 .
then we see that @xmath643_{\fb}$ ] has a demazure flag from the induction hypothesis on @xmath532 , corollary [ cor : joseph ] , lemma [ lem : affine_version ] ( ii ) , and proposition [ prop1 ] .
recall that , by the definition of demazure modules , there exists a canonical embedding @xmath644 .
let @xmath645 denote the image of @xmath646 under this embedding .
it should be noted that @xmath647_{\fb}$ ] is not a generator .
+ + _ claim _ 1 .
there exists a homomorphism of @xmath289-modules @xmath648_{\fb } \to \big [ ( v_q)_{u^q } * m_{v_{p-1}^{q}}^{p-1 } * \cdots * m_{v_1^q}^1\big]_{\fb}\ ] ] which maps @xmath649 to @xmath650 .
+ it suffices to show that , if @xmath586 annihilates @xmath649 , then @xmath524 also annihilates @xmath651 .
we may assume that @xmath652 for some @xmath624 .
then there exists some @xmath626 satisfying @xmath653 by lemma [ lem : annihilating ] .
obviously , @xmath654 also holds .
let @xmath461 be the unique integer such that @xmath655 and @xmath656 and @xmath657 be vectors such that @xmath658 then we have from ( [ eq : x - y ] ) that @xmath659 since @xmath660 @xmath661 holds by lemma [ lem : annihilating ] . on the other hand , we have by definition that @xmath662 hence @xmath663 holds , and claim 1 is proved . + set @xmath664 . by lemma [ lem : important ] , the @xmath69-module @xmath665_{\fb}$ ] extends to a @xmath70-module . then by lemma
[ lem : trivial_lem2 ] ( ii ) , there exists a homomorphism of @xmath342-modules @xmath666_{\fb } \to \big [ ( v_q)_{u^q } * m_{v_{p-1}^{q}}^{p-1 } * \cdots * m_{v_1^q}^1\big]_{\fb}\ ] ] which makes the following diagram commutative : @xmath667_{\fb } \ar[r]^(.52){\phi } \ar[d ] & \left [ ( v_q)_{u^q } * \cdots * m_{v_1^q}^1\right]_{\fb } \\ \cd_i\left [ ( v_{q-1})_{u^{q-1 } } * \cdots * m_{v_1^{q-1}}^1\right]_{\fb } \ar[ru]^(.50){\wti{\phi } } % & } , \ ] ] where the vertical map is the canonical one .
+ + _ claim _ 2 .
the homomorphism @xmath341 is surjective .
+ it suffices to show the image of @xmath341 contains the generator @xmath668 , whose @xmath590-weight @xmath109 is equal to @xmath669 note that the image of @xmath341 contains @xmath670 , whose @xmath590-weight @xmath15 is @xmath671 as the image is a @xmath342-module , its weight set contains @xmath672 . since @xmath673_{\fb}$ ] is isomorphic to @xmath674 as a @xmath408-module , it is easily checked that the weight space with weight @xmath109 is one - dimensional .
hence the image contains the generator , and claim 2 is proved .
+ now , the following claim completes the proof of the proposition : + + _ claim _ 3 .
the dimensions of the both sides of @xmath341 are equal . + the dimension of the right hand side is equal to @xmath675 let us calculate the dimension of the left hand side . as stated at the beginning of this proof , @xmath676_{\fb}$ ] has a demazure flag .
hence by corollary [ cor : flag_character ] , the character of the left hand side of @xmath341 is equal to @xmath677_{\fb}.\ ] ] for a @xmath590-semisimple module @xmath86 whose @xmath590-weight set is contained in @xmath678 , denote by @xmath679 $ ] the @xmath590-character of @xmath86 . since each
@xmath389 is a finite - dimensional ( @xmath680)-module on which @xmath507 acts trivially , @xmath681 belongs to @xmath682 $ ] and is @xmath220-invariant .
hence we have from lemma [ lem : final_lem ] that @xmath683_{\fb } & = \ol{d}_i \ , \bch \big [ v_{q-1 } * m^{p-1 } * \cdots * m^1\big]_{\fb } \\
= \ol{d}_i \ , \bigg(\bch v_{q-1 } \times \prod_{1\le j \le p-1 } \bch m^j\bigg ) & = \prod_{1\le j \le p-1 } \bch{m^j } \times \ol{d}_i \,\bch v_{q-1 } \\ & = \prod_{1 \le j \le p-1 } \bch m^j \times \cl \circ d_i \ , \ch v_{q-1}. \end{aligned}\ ] ] since @xmath684 by theorem [ thm : character_formula ] , we see that the dimension of the left hand side is equal to ( [ eq : dimension ] ) .
hence claim 3 is proved , and the proof of the proposition is complete .
+ as stated above , theorem [ thm : main_theorem ] is now established from propositions [ prop1 ] and [ prop2 ] . in theorem [ thm : main_theorem ] , it is assumed that @xmath47 is of @xmath0 type .
the author , however , expects the theorem to be true for general types if all given kr modules satisfy the assumption of theorem [ thm : kr - demazure ] .
in fact , all the proof of the theorem can also be applied in this case , except for the joseph s theorem ( theorem [ thm : joseph ] ) which is needed in the final step of the proof . hence to prove the theorem for non - simply laced type by our approach , it is needed to prove the joseph s theorem for this type .
since proposition [ prop : x - side ] has already been proved for nonexceptional type , this would also imply the @xmath1 conjecture for perfect kr crystals of type @xmath685 and @xmath686 .
b. feigin and s. loktev . on generalized kostka polynomials and the quantum verlinde rule . in _
differential topology , infinite - dimensional lie algebras , and applications _ ,
volume 194 of _ amer .
2 _ , pages 6179 .
soc . , providence , ri , 1999 .
g. fourier , m. okado , and a. schilling .
perfectness of kirillov - reshetikhin crystals for nonexceptional types . in _ quantum affine algebras ,
extended affine lie algebras , and their applications _ , volume 506 of _ contemp .
_ , pages 127143 .
soc . , providence , ri , 2010 .
g. hatayama , a. kuniba , m. okado , t. takagi , and z. tsuboi .
paths , crystals and fermionic formulae . in _ mathphys odyssey , 2001 _ , volume 23 of _ prog .
_ , pages 205272 .
birkhuser boston , boston , ma , 2002 .
g. hatayama , a. kuniba , m. okado , t. takagi , and y. yamada .
remarks on fermionic formula . in _
recent developments in quantum affine algebras and related topics ( raleigh , nc , 1998 ) _ , volume 248 of _ contemp .
_ , pages 243291 .
soc . , providence , ri , 1999 . | in this article , we show in the @xmath0 case that the fusion product of kirillov - reshetikhin modules for a current algebra , whose character is expressed in terms of fermionic forms , can be constructed from one - dimensional modules by using joseph functors . as a consequence ,
we obtain some identity between fermionic forms and demazure operators .
since the same identity is also known to hold for one - dimensional sums of nonexceptional type , we can show from these results the @xmath1 conjecture for type @xmath2 and @xmath3 . |
the recent results from the relativistic heavy ion collider ( rhic ) @xcite indicate the formation of a thermalized medium endowed with large collective flow and very low viscosity @xcite .
these findings suggest that quark gluon plasma ( qgp ) is a strongly interacting system for temperatures close to its transition temperature ( @xmath2 ) .
apart from the experimental indications , the most convincing evidence in favour of the existence of a strongly interacting qgp comes form the lattice qcd simulations @xcite .
these non - perturbative studies show that the thermodynamic quantities , like pressure and energy density , deviate form there respective ideal gas ( of free quarks and gluons ) values by about @xmath3 even at temperature @xmath4 . on the other hand ,
other lattice studies indicate the smallness of the viscous forces in qgp @xcite . all these results point to the fact that close to @xmath2 nature of qgp is far from a gas of free quarks and gluons . in order to uncover the nature of qgp in the vicinity of @xmath2 and also to understand the underlying physics of these lattice results many different suggestions have been made over the last decade .
descriptions in terms of various quasi - particles @xcite , resummed perturbation theories @xcite , effective models @xcite are few among many such attempts . apart from all these
, the newly proposed model of shuryak and zahed @xcite has generated considerable amount of interest in the recent years .
motivated by the lattice results for the existence of charmonium in qgp @xcite , this model proposed a strongly interacting chromodynamic system of quasi - particles ( with large thermal masses ) of quarks , anti - quarks and gluons along with their numerous bound states .
as different conserved charges , , baryon number ( @xmath5 ) , electric charge ( @xmath6 ) , third component of isospin ( @xmath7 ) , are carried by different flavours ( @xmath8 ) of quarks , in the conventional quasi - particle models , conserved charges come in strict proportion to number of @xmath9 , @xmath10 , @xmath11 quarks .
thus conserved charges are strongly correlated with the flavours and the flavours have no correlations among themselves . on the other hand , in the model of @xcite , presence of bound states
demand correlations among different flavours .
hence correlations between conserved charges and flavours depend on the mass - spectrum of the bound states and the strong correlations among them are lost .
based on the above arguments , in @xcite , it has been suggested that the quantity c_bs = -3 , can be used to probe the degrees of freedom of qgp . here
@xmath12 is the net baryon number and @xmath13 , @xmath14 , @xmath15 are the numbers of net ( quarks minus anti - quarks ) up - quarks , down - quarks and strange - quarks respectively .
the notation @xmath16 denotes average taken over a suitable ensemble .
it has been argued in @xcite that for qgp where quarks are the effective degrees of freedom , , where correlations among @xmath13 , @xmath14 and @xmath15 are absent , @xmath17 will have a value of @xmath18 for all temperature @xmath19 . on the other hand , for the model of @xcite @xmath20 at @xmath21 , while for a gas of hadron resonances @xmath22 .
thus the knowledge of @xmath17 helps to identify the degrees of freedom in qgp . by extending the idea of @xcite ,
recently in @xcite , many ratios like c_(kl)/l= , [ eq.ratio ] have been calculated using lattice qcd simulations with two flavours of dynamical light quarks and three flavours ( two light and one heavy ) of valance quarks . here
@xmath23 and @xmath24 denote the susceptibilities corresponding to conserved charge @xmath25 and correlation among conserved charges @xmath26 and @xmath25 respectively .
the physical meaning of the ratios like @xmath27 can be interpreted as follows create an excitation with quantum number @xmath25 and then observe the value of a different quantum number @xmath26 associated with this excitation .
thus these ratios identify the quantum numbers corresponding to different excitations and hence provide information about the degrees of freedom .
the calculations of @xcite found no evidence for the existence of bound states @xcite even at temperatures very close to @xmath2 .
these finding are consistent with the results of @xcite , where the hypothesis of @xcite has been tested by investigating the ratios of higher order baryon number susceptibilities obtained from lattice simulations . as these lattice studies @xcite involved simulations with dynamical quarks
, they were done using small lattices having temporal lattice size @xmath28 . by comparing with the results from quenched simulations
it has been shown @xcite that @xmath27 do not depend on @xmath29 for temperature @xmath30 .
it is clearly important to verify whether the same conclusion holds even close to @xmath2 .
furthermore , it is known that in the case of quenched qcd with standard staggered quarks the diagonal quark number susceptibilities ( qns ) have strong dependence on the lattice spacing even for the free theory @xcite . on the other hand ,
the off - diagonal qns are identically zero for an ideal gas and acquires non - zero value only in the presence of interactions .
so the lattice spacing dependence of the off - diagonal qns is likely to be more complicated , as opposed to that for the diagonal qns where these corrections are dominated by the lattice artifacts of the naive staggered action .
thus if these two qns become comparable the ratios mentioned in eq .
( [ eq.ratio ] ) can have non - trivial dependence on the lattice spacing @xmath31 and hence the continuum limit of these ratios can be different from that obtained using small lattices . since the perturbative expressions for diagonal and off - diagonal qns ( for vanishingly small quark mass and chemical potential ) are respectively @xcite 1 + o(g^2 ) - g^6 g^-1 , [ eq.pert ] it is reasonable to expect that the off - diagonal qns may not be negligible at the vicinity of @xmath2 where the coupling @xmath32 is large . as the contributions of the bound states in the qns become more and more important as one approaches @xmath2 @xcite , on the lattice it is necessary to investigate the continuum limit of the these ratios of in order to verify the existence of bound states in a strongly coupled qgp
at present a continuum extrapolation of this kind can only be performed using quenched approximation due to the limitations of present day computational resources . a quenched result for these ratios will also provide an idea about the dependence of these ratios on the sea quark mass .
the aim of this work is to carefully investigate the continuum limit of the ratios of the kind @xmath27 for temperatures @xmath0 using quenched lattice qcd simulations .
the plan of this paper is as follows in section [ sc.results ] we will give the details of our simulations and present our results . in the section [ sc.discussion ] we will summarise and discuss our results .
the partition function of qcd for @xmath33 flavours , each with chemical potential @xmath34 and mass @xmath35 , at temperature @xmath36 has the form ( t,\{_f},\{m_f } ) = e^-s_g(u ) _ f m_f ( t,_f , m_f ) , where @xmath37 is the gauge part of the action and @xmath38 is the dirac operator .
we have used standard wilson action for @xmath37 and staggered fermions to define @xmath38 .
the temperature @xmath36 and the spatial volume @xmath39 are expressed in terms of lattice spacing @xmath31 by the relations @xmath40 and @xmath41 , @xmath42 and @xmath43 being the number of lattice sites in the spatial and the euclidean time directions respectively .
the flavour diagonal and the flavour off - diagonal quark number susceptibilities ( qns ) are given by _ ff & = & ( ) = ( ) , + _ ff & = & ( ) = ( ) ( m_f^-1m_f)*tr*(m_f^-1m_f ) , respectively . here
the single and double primes denote first and second derivatives with respect to the corresponding @xmath34 and the angular bracket denote averages over the gauge configurations . in this paper
we report results of these susceptibilities on lattices with @xmath44 , for the temperatures @xmath45 , chemical potential @xmath46 and using quenched approximations .
the details of our scale setting procedure are given in @xcite .
we have generated quenched gauge configurations by using the cabbibo - marinari pseudo - heatbath algorithm with kennedy - pendleton updating of three @xmath47 subgroups on each sweep . since for @xmath48
qns are almost independent of the bare valance quark mass ( @xmath49 ) @xcite , we have used @xmath50 for the light @xmath9 and @xmath10-flavours .
motivated by the fact that for the full theory @xmath51 we have used @xmath52 for the heavier @xmath11-flavour .
the fermion matrix inversions were done by using conjugate gradient method with the stopping criterion @xmath53 , @xmath54 being the residual after the @xmath55-th step and @xmath56 @xcite .
the traces have been estimated by the stochastic estimator @xmath57 , where @xmath58 is a complex vector whose components have been drawn independently from a gaussian ensemble with unit variance .
the square of a trace has been calculated by dividing @xmath59 vectors into @xmath25 non - overlapping sets and then using the relation @xmath60 .
we have observed that as one approaches @xmath2 from above these products , and hence @xmath61 , become more and more noisy for larger volumes and smaller quark masses .
so in order to reduce the errors on @xmath61 number of vectors @xmath59 have been increased ( for the larger lattices and the smaller quark masses ) with decreasing temperature .
details of all our simulations are provided in table [ tb.simulation ] .
.the couplings ( @xmath62 ) , lattice sizes ( @xmath63 ) , number of independent gauge configurations ( @xmath64 ) and number of vectors ( @xmath59 ) that have been used for our simulations are given for each temperature .
the gauge configurations were separated by @xmath65 sweeps . [ cols="^,^,^,^,^,^ " , ] in the following sections we present our results .
the notations we use are same as in @xcite .
since we use equal masses for the two light @xmath9 and @xmath10 flavours , the flavour diagonal susceptibilities in this context are @xmath66 and the flavour off - diagonal susceptibilities are @xmath67 . for the heavy flavour @xmath11 the flavour diagonal susceptibility is denoted as @xmath68 and the flavour off - diagonal susceptibilities are @xmath69 .
expressions for all the susceptibilities used here have been derived in the appendix of @xcite .
( @xmath70 ) dependence of @xmath71 ( squares ) and @xmath72 ( circles ) for @xmath73 ( top panel ) and for @xmath74 ( bottom panel ) .
the continuum extrapolations ( linear fits in @xmath75 ) are shown by the lines.,title="fig : " ] ( @xmath70 ) dependence of @xmath71 ( squares ) and @xmath72 ( circles ) for @xmath73 ( top panel ) and for @xmath74 ( bottom panel ) .
the continuum extrapolations ( linear fits in @xmath75 ) are shown by the lines.,title="fig : " ] in order to understand the cut - off dependence of @xmath27 let us start by examining the same for the diagonal and off - diagonal qns .
we have found that the for all the temperatures the diagonal qns ( @xmath76 and @xmath77 ) depend linearly on @xmath78 , , the finite lattice spacing corrections to the diagonal qns have the form @xmath79 . as an illustration of this
we have shown our data for @xmath73 and @xmath74 in fig .
[ fig.u - s ] .
similar variations were found for the other temperatures also .
we have made continuum extrapolations of the diagonal qns by making linear fits in @xmath75 .
our continuum extrapolated results match , within errors , with the available data of @xcite at @xmath80 and @xmath81 . in fig .
[ fig.ud-us ] we present some of our typical results for the off - diagonal qns .
note that here the scales are @xmath82 magnified as compared to fig .
[ fig.u - s ] .
the sign of our off - diagonal qns is consistent with the perturbative predictions of @xcite , as well as with the lattice results of @xcite .
the order of magnitude of our off - diagonal qns matches with the results of @xcite which uses the same unimproved staggered fermion action as in the present case . as can be seen from fig .
[ fig.ud-us ] , within our errors , we have not found any perceptible dependence @xmath61 on the lattice spacing @xmath31 .
hence to good approximation @xmath83 . also for the other temperatures , which are not shown in fig . [ fig.ud-us ] , similar variations were found .
results of our continuum extrapolations of the diagonal and off - diagonal qns are listed in table [ tb.cont-qns ] .
@xmath84 & & & & + & @xmath31&@xmath85&@xmath86&@xmath31&@xmath85&@xmath86 & @xmath87&@xmath86&@xmath87&@xmath86 + 1.1&0.79(1)&11.3(5)&0.3&0.33(1)&5.4(1)&0.1&-4(4)&0.5&-6(4)&0.1 + [ 2pt ] 1.25&0.84(1)&15(1)&0.5&0.45(1)&10(1)&0.8&-0.2(1.0)&0.1&-0.7(1.0)&0.6 + [ 2pt ] 1.5&0.83(1)&17.3(3)&0.5&0.55(1)&12.5(5)&0.9&-7(5)&0.1&2(2)&0.6 + [ 2pt ] 2.0&0.86(2)&19.7(2)&0.7&0.70(2)&17(2)&0.8&2(3)&0.5&-0.1(1.0)&0.8 + dependence of the off - diagonal qns @xmath88 at @xmath80 ( top panel ) and @xmath89 at @xmath73 ( bottom panel ) have been shown.,title="fig : " ] dependence of the off - diagonal qns @xmath88 at @xmath80 ( top panel ) and @xmath89 at @xmath73 ( bottom panel ) have been shown.,title="fig : " ] ( squares ) and @xmath90 ( circles ) have been shown . ] for the sake of completeness we also present our continuum extrapolated results for the two very important quantities , the baryon number susceptibility ( @xmath91 ) and the electric charge susceptibility ( @xmath92 ) .
these quantities are related to the event - by - event fluctuations of baryon number and electric charge @xcite which have already been measured at rhic @xcite .
the definitions that we use for @xmath91 and @xmath92 are @xcite _ b = ( 2_u + _ s + 2_ud + 4_us ) , _ q = ( 5_u + _ s - 4_ud - 2_us ) . in fig .
[ fig.b - q ] we show the continuum results for @xmath93 and @xmath90 .
continuum extrapolations have been performed by making linear fits in @xmath94 .
continuum limit of these quantities were also obtained in @xcite for @xmath95 , though using different definitions for these quantities .
nevertheless , given the compatibility of our diagonal qns with that of @xcite and the smallness of the off - diagonal qns for @xmath95 our continuum results for @xmath91 and @xmath92 are compatible with that of ref .
@xcite , for any chosen definitions for these quantities . ) with changing lattice spacings has been shown for @xmath73 .
the lines indicate the @xmath96 error band of a constant fit to this data . in the bottom panel
we show our continuum results for @xmath97 ( see text for details).,title="fig : " ] ) with changing lattice spacings has been shown for @xmath73 .
the lines indicate the @xmath96 error band of a constant fit to this data . in the bottom panel
we show our continuum results for @xmath97 ( see text for details).,title="fig : " ] wroblewski parameter ( @xmath97 ) @xcite is a quantity of extreme interest due to its relation to the enhancement of strangeness production in qgp @xcite .
the rate of production of quark pairs in a equilibrated plasma is related to the imaginary part of the complex qns by fluctuation - dissipation theorem .
if one assumes that the plasma is in chemical ( and thermal ) equilibrium and the typical energy scales for the production of @xmath9 , @xmath10 and @xmath11 quarks are well separated from the inverse of the characteristic time scale of the qcd plasma , then using kramers - kroing relation one can relate @xmath97 to the ratio of qns @xcite _ s = = .
( in the above equation @xmath98 should be interpreted as quark number density and not as quark anti - quark condensates . )
we have found that @xmath97 , which is a ratio of two diagonal qns , remains constant ( within @xmath99 ) with varying lattice spacings for all temperatures in @xmath100 .
we have illustrated this in the top panel of fig .
[ fig.lam_s ] by plotting @xmath97 with @xmath75 for the temperature @xmath73 .
these results are somewhat surprising since the order @xmath101 corrections are not negligible for the individual diagonal qns . but for the ratio of the diagonal qns for two different bare valance quark masses these order @xmath101 corrections happen to be negligible and thus seems to be quark mass independent .
this indicates that the finite lattice spacing corrections to the diagonal qns is constrained to have the form @xmath102 $ ] , as opposed to the more general form @xmath79 .
our continuum results for the wroblewski parameter have been shown in the bottom panel of fig .
[ fig.lam_s ] . in view of the constancy of @xmath97
we have made the continuum extrapolations by making a constant fit to @xmath94 .
our continuum limit for @xmath97 are consistent with the previously reported @xcite continuum values for @xmath95 .
our continuum results for @xmath97 are very close to the results of @xcite for the whole temperature range of @xmath0 .
closeness of our quenched results with the results from the dynamical simulations of @xcite suggest that the wroblewski parameter has practically no dependence on the mass of the sea quarks .
these observations along with the fact that @xmath97 has very mild dependence on the valance quark mass @xcite shows that the present day lattice qcd results for the wroblewski parameter are very reliable .
the robustness of the wroblewski parameter is very encouraging specially since in the vicinity of @xmath2 the lattice results for this quantity almost coincides with the value ( @xmath103 ) extracted by fitting the experimental data of rhic with a hadron gas fireball model @xcite .
dependence of @xmath104 at @xmath80 .
the bottom panel shows the same for @xmath105 at @xmath73 , title="fig : " ] dependence of @xmath104 at @xmath80 .
the bottom panel shows the same for @xmath105 at @xmath73 , title="fig : " ] after examining the ratio of the diagonal qns let us focus our attention on the ratios of off - diagonal to diagonal qns .
given our results for the diagonal and off - diagonal qns it is clear that these will have the form @xmath106[1-b(t)a^2]$ ] .
since @xmath107 is positive , , @xmath108 decreases with decreasing lattice spacing , this ratio is expected to decrease ( as @xmath61 is negative ) and move away from zero . however , due smallness of these ratios itself , within our numerical accuracies , we have been unable to identify any such effect .
this has been exemplified in fig .
[ fig.ud_u-us_s ] where @xmath104 at @xmath80 ( top panel ) and @xmath105 at @xmath73 ( bottom panel ) have been shown .
+ following the main theme of this paper we now present the lattice spacing dependence of ratios the like @xmath27 .
two such ratios that can directly probe the degrees of freedom in a qgp are @xcite c_bs & & -3c_(bs)/s = -3 = = 1 + , + c_qs & & 3c_(qs)/s = 3 = = 1 - . [ eq.cxs ]
these quantities probe the linkages of the strangeness carrying excitations to baryon number ( @xmath17 ) and electric charge ( @xmath109 ) and hence give an idea about the average baryon number and the average electric charge of all the excitations carrying the @xmath11 flavours .
these ratios are normalized such that for a pure quark gas , , where unit strangeness is carried by excitations having @xmath110 and @xmath111 , @xmath112 .
a value of @xmath17 and @xmath109 significantly different from @xmath18 will indicate that the qgp phase may contain some other degrees of freedom apart form the quasi - quarks .
similar ratios can also be formed for the light quark sector @xcite , , for the @xmath9 flavour the ratios c_bu & & 3 c_(bu)/u = 3 = = 1 + + , + c_qu & & 3 c_(qu)/u = = = 2 - - [ eq.cxu ] quantifies the average baryon number ( @xmath113 ) and and the average electric charge ( @xmath114 ) of all the excitations carrying @xmath9 quarks . for a medium of pure quarks , , where the @xmath9 flavours are carried by excitations with baryon number @xmath115 and electric charge @xmath116 , @xmath117 and @xmath118 .
similar ratios can also be formed for the @xmath10 quarks @xcite .
as can be seen seen from eqs .
( [ eq.cxs ] , [ eq.cxu ] ) the lattice spacing dependence of @xmath17 are governed by the cut - off dependence of the ratios @xmath119 . since we have already emphasised that , within our numerical accuracies , the ratios @xmath119 are almost independent of lattice spacings
it is expected that the same will also happen for the ratios @xmath27 . in accordance to this expectation
we have found that for temperatures @xmath120 these ratios are independent of lattice spacings within @xmath121 errors , see fig .
[ fig.cxy ] .
note that these ratios are not only independent of the lattice spacings but also acquire values which are very close to their respective ideal gas limits .
( top panel ) and @xmath122 ( bottom panel ) .
the lines indicate the ideal gas values for these quantities .
see text for details.,title="fig : " ] ( top panel ) and @xmath122 ( bottom panel ) .
the lines indicate the ideal gas values for these quantities .
see text for details.,title="fig : " ] in fig .
[ fig.cxy-cont ] we present our continuum results for @xmath122 ( bottom panel ) and @xmath123 ( top panel ) , where @xmath124 . since these ratios remain almost constant with changing @xmath75 ( see fig .
[ fig.cxy ] ) we have made continuum extrapolations by making constant fits of our data to @xmath75 . for the whole temperature range of interests ( @xmath125 )
these ratios have values which are compatible with that for a gas of pure quarks .
this is exactly what has been found in @xcite using partially quenched simulations with smaller lattices . for the @xmath10 quarks
also we have found similar results .
in this paper we have made a careful investigation of the continuum limit of different ratios of off - diagonal to diagonal susceptibilities in quenched qcd using lattices with large temporal extents ( @xmath126 ) , for a very interesting range of temperature ( @xmath0 ) and for vanishing chemical potential .
we have found that for this whole range of temperature the lattice results for the ratios like @xmath17 , @xmath109 are robust , , they are almost independent ( within @xmath99 ) of the lattice spacing .
we have also arrived at the same conclusion for the wroblewski parameter which is of interest to the experiments in rhic and large hadron collider ( lhc ) . at this point , it is good to have some idea about how unquenching may change our results
. it has been found @xcite that in the temperature range @xmath127 there is only @xmath128 change in the qns in going from quenched to @xmath129 dynamical qcd . on the other hand ,
since the order of the phase transition depends strongly on the number of dynamical flavours the change in qns is likely to be much larger in the vicinity of the transition temperature for the quenched theory which has a first order phase transition . though this may be true for the individual qns
, their ratios may have very mild dependence on the sea quarks content of the theory .
given the good compatibility of our results of @xmath17 , @xmath109 with the results of @xcite it is clear that indeed these ratios have very mild dependence on the sea quark content of the theory .
it is also known that @xcite for bare valance quark mass of @xmath48 the dependence of the qns on the valance quarks mass is very small .
hence our results show that the ratios like @xmath27 are robust not only in the sense that they do not depend on the lattice spacings but also they have very mild dependence on the quark masses .
all the results presented in this paper are for spatial lattice sizes @xmath131 , , for aspect ratios @xmath132 . in view of
the fact that quenched qcd has a first order phase transition it is important to have some idea about the volume dependence of our results , specially in the vicinity of the transition temperature @xmath2 . to check this dependence
we have performed simulations using lattices having aspect ratios @xmath133 , for our smallest temporal lattice @xmath130 and at temperature @xmath73 . in these simulations
we have not found any significant volume dependence of any quantity which have been presented in this paper . as an illustration , in fig.[fig.vol - dep ] , we have shown the dependence of @xmath105 on the aspect ratio , for @xmath130 at @xmath73 .
the volume dependence is expected to be even smaller as one goes further away from first order phase transition point .
also the agreement of our results with that of @xcite , where an aspect ratio of @xmath134 have been used , shows that the these ratios have almost no volume dependence for @xmath135 .
while the closeness of @xmath123 and @xmath122 ( @xmath124 ) to their respective ideal gas values do support the notion of quasi - particle like excitations in qgp , a significant deviation of these ratios from their ideal gas values neither rule out the quasi - particle picture nor confirms the existence of the bound states proposed in @xcite .
large contributions from the chemical potential dependence of the quasi - particle masses may lead to significant deviation of these ratios , especially in the vicinity of @xmath2 .
it has already been pointed out @xcite that , near @xmath2 , the chemical potential dependence of the quasi - particle masses becomes crucial for the baryonic susceptibilities .
nevertheless , it may be interesting to compare our results with the predictions of the bound state model of @xcite . based on the model of @xcite ( and assuming that the mass formulae given in @xcite hold right down to @xmath2 ) the predicted values of @xmath17 are approximately @xmath136 at @xmath80 @xcite , @xmath137 at @xmath74 and almost zero at @xmath73 @xcite . clearly , as can be seen form fig .
[ fig.cxy-cont ] ( bottom panel ) , these values are very much different from our continuum results . however , it has been argued in @xcite that apart from all the bound states mentioned in @xcite , baryon like bound states may also exist in qgp .
these baryons make large contributions to the baryonic susceptibilities , especially close to @xmath2 @xcite . taking account the contributions from the strange baryons may increase the value of @xmath17 . in @xcite
it has also been argued that for two light flavours if one considers the contributions of the baryons only then close to @xmath2 the ratio of 2-nd order isospin susceptibility ( @xmath138 ) to the 2-nd order baryonic susceptibility ( @xmath139 ) is @xmath140 .
clearly this is inconsistent with our results since a value of @xmath141 gives a positive @xmath104 ( @xmath142 ) .
whereas , the lattice results for @xmath104 are negative and much smaller in magnitudes .
this suggest that the contribution of the mesons ( also possibly of the quarks , diquarks and @xmath143-states ) are definitely important in the isospin susceptibility @xmath144 .
if one takes into account of the contributions of the mesons ( pions and rhos ) and assumes that the boltzmann weight of the mesons are equal to that of the baryon one gets a lower bound for @xmath145 , namely @xmath146 @xcite .
but this lower bound gives @xmath147 and hence very far from our results .
moreover , very recently it has been argued @xcite that one can carefully tune the densities of the baryon and meson like bound states in the model of refs .
@xcite to reproduce the lattice results for off - diagonal qns .
but even those carefully tuned values fail to reproduce @xcite the lattice results for higher order susceptibilities . in view of all these , the lattice results of @xcite favours a quasi - particle like picture of qgp , as opposed to the bound state model of @xcite .
the results of this paper show that these lattice results are really robust in the sense that they have very mild dependence on the lattice spacing and sea quark content of the theory .
the author is grateful to rajiv gavai for his constant encouragement , many illuminating discussion and careful reading of the manuscript .
the author would also like to thank sourendu gupta for many useful comments and discussion .
part of this work was done during a visit to ect * , trento . the financial support from the doctoral training programme of ect * , trento is gratefully acknowledged .
99 k. adcox [ phenix collaboration ] , , a 757 ( 2005 ) 184 ; + i. arsene [ brahms collaboration ] , , a 757 ( 2005 ) 1 ; + b. b. back [ phobos collaboration ] , , a 757 ( 2005 ) 28 ; + j. adams [ star collaboration ] , , a 757 ( 2005 ) 102 .
d. teaney , , c 68 ( 2003 ) 034913 .
r. v. gavai , s. gupta and s. mukherjee , hep - lat/0506015 , pos lat2005 173 ( 2005 ) ; g. boyd , , 75 ( 1995 ) 4169 , , b 469 ( 1996 ) 419 ; + f. karsch , e. laermann and a. peikert , , b 478 ( 2000 ) 447 .
a. nakamura and s. sakai , , 94 ( 2005 ) 072305 .
p. levai and u. heinz , , c 57 ( 1998 ) 1879 ; + a. peshier , a 702 ( 2002 ) 128 .
m. bluhm , b. kampfer and g. soff , , b 620 ( 2005 ) 131 .
p. arnold and c. zhai , , d 50 ( 1994 ) 7603 , _ ibid .
_ , d 51 ( 1995 ) 1906 ; + e. braaten and a. nieto , , d 53 ( 1996 ) 3421 ; +
blaizot , e. iancu and a. rebhan , , d 63 ( 2001 ) 065003 ; + k. kajantie , , 86 ( 2001 ) 10 , , 0304 ( 2003 ) 036 ; + j. o. andersen , , d 66 ( 2002 ) 085016 .
a. dumitru , j. lenaghan and r. d. pisarski , , d 71 ( 2005 ) 074004 ; + c. ratti , m. a. thaler and w. weise , , d 73 ( 2006 ) 014019 .
e. shuryak and i. zahed , , d 70 ( 2004 ) 054507 , , c 70 ( 2004 ) 021901 .
m. asakawa and t. hatsuda , , 92 ( 2004 ) 012001 ; + s. datta , f. karsch , p. petreczky and i. wetzorke , , d 69 ( 2004 ) 094507 .
v. koch , a. majumder and j. randrup , , 95 ( 2005 ) 182301 .
r. v. gavai and s. gupta , , d 73 ( 2006 ) 014004 .
s. ejiri , f. karsch and k. redlich , , b 633 ( 2006 ) 275 .
r. v. gavai and s. gupta , , d 67 ( 2003 ) 034501 .
r. v. gavai , _ nucl .
_ , 119 ( 2003 ) 529 . j. p. blaizot , e. iancu and a. rebhan , , b 523 ( 2001 ) 143
. j. liao and e. shuryak , , d 73 ( 2006 ) 014509 .
r. v. gavai and s. gupta , , d 71 ( 2005 ) 114014 .
s. ejiri , , d 71 ( 2005 ) 054508 .
m. asakawa , u. heinz and b. muller , , 85 ( 2000 ) 2072 ; + s. jeon and v. koch , , 85 ( 2000 ) 2076 . k. adcox [ phenix collaboration ] , , 89 ( 2002 ) 082301 ; + j. adams [ star collaboration ] , , c 68 ( 2003 ) 044905 .
a. wroblewski , _ acta .
_ , b 16 ( 1985 ) 379 . j. rafelski and b. muller , , 48 ( 1982 ) 1066 , _ erratum - ibid _ 56 ( 1986 ) 2334 .
r. v. gavai and s. gupta , , d 65 ( 2002 ) 094515 .
s. gupta and r. ray , , d 70 ( 2004 ) 114015 .
j. cleymans , _ j. phys .
_ , g 28 ( 2002 ) 1575 .
r. v. gavai , s. gupta and p. majumdar , , d 65 ( 2002 ) 054506 .
we are grateful to abhijit majumder for supplying us with these values .
we thank jinfeng liao for pointing out this lower bound and for other useful correspondences .
a. majumder and b. muller , nucl - th/0605079 | using quenched lattice qcd simulations we investigate the continuum limit of baryon - strangeness correlation and other related conserved charge - flavour correlations for temperatures @xmath0 . by working with lattices having large temporal extents ( @xmath1 )
we find that these quantities are almost independent of the lattice spacing , , robust .
we also find that these quantities have very mild dependence on the sea quark mass and acquire values which are very close to their respective ideal gas limits .
our results also confirm robustness of the wroblewski parameter . |
the kondo model , initially introduced to describe with great success the behavior of diluted magnetic impurities in metals @xcite , has been an exciting playground for condensed matter physicists in the last two decades . indeed , the extreme simplicity of this many - body hamiltonian was key to some unexpected progresses in the field , such as the discovery of the effective theory describing the mott transition @xcite , and as a prototypal description of non fermi liquid behavior in multi - channel @xcite and pseudogap extensions @xcite .
localized quantum impurities also provide the building entity to describe heavy fermion materials @xcite , and this is still a strong motivation for studying kondo models on the lattice , most interestingly in relation to the violation of fermi liquid behavior associated to the breakdown of kondo screening close to a quantum critical point @xcite , as discovered in a wide class of @xmath5-electron metals such as cecu@xmath6au@xmath7 @xcite .
interest in the kondo model was also revived by the possibility of building and controlling precisely quantum impurity states in semiconducting devices , leading to the realization of the kondo effect in quantum dots @xcite .
the possibility of observing non fermi liquid behavior in a multi - channel setup was recently advocated @xcite , and provides further incentive for a complete understanding of quantum impurity problems .
quite recently , peculiar attention was given to the kondo problem by coleman and pepin @xcite , in the study of an underscreened model which showed relatively strong deviations from fermi liquid behavior in the specific heat ( we point out however that related effects can be traced back to the physics of the ferromagnetic kondo model @xcite , see also @xcite ) .
this surprising result was interpreted as due to anomalous scattering of conduction electrons onto the remaining unscreened spin degrees of freedom , giving rise to a `` singular '' fermi liquid fixed point , which differs from the intermediate fixed point @xcite associated to usual non fermi liquid by the fact that the effective ferromagnetic kondo coupling renormalizes to zero at low energy @xcite .
this insightful work was however limited to a range of frequencies higher than the zeeman energy , and concentrated mainly on thermodynamic quantities . in order to study in more detail the dynamics due to these anomalous excitations in the underscreened kondo model
, we would like to find a simple approach that would be able to grasp the full crossover from the local moment regime at high temperature down to the underscreened state in which nearby conduction electrons are tightly bound to the impurity . in the realistic case of su(2 )
spins , we note that the deviation from exact screening can be in principle tuned @xcite by changing the size @xmath8 of the impurity spin and/or the number of screening channels @xmath9 , obtaining a transition from underscreening at @xmath10 to overscreening at @xmath11 ( with perfect screening at @xmath12 ) . a simple route to capture this physics consists in establishing a large @xmath1 limit of the problem , generalizing the quantum spin to the su(@xmath1 ) group .
a crucial step lies however in the choice of the su(@xmath1 ) representation in which the spin is considered .
it is known that fermionic representations at large @xmath1 only allow perfect screening when @xmath13 @xcite and overscreening otherwise @xcite .
interestingly , bosonic representations of su(@xmath1 ) preserve the distinction between `` small '' and `` large '' spin as found by parcollet and georges @xcite , and allow to study the underscreened situation
. however , this case presents some pathologies : the t - matrix scales as @xmath14 , and the singular behavior as found in @xcite seems to be absent of the solution @xcite .
the previous discussion illustrates the need for an alternative large @xmath1 limit to describe the underscreened kondo effect , but also gives momentum to the idea we will pursue in the following .
indeed , we can understand that overscreening and underscreening are somewhat `` dual '' in the sense that , while the former situation is reached in the presence of many screening channels , the latter case should be obtained by considering many `` spin channels '' .
a simple way to formalize this is to strongly enlarge the symmetry group of the impurity spin , thus forcing underscreening by the bath of conduction electrons .
in fact , if one considers a generalized kondo model involving a _ single bath _ of su(@xmath1 ) electrons interacting with a localized su(@xmath0 ) spin , where @xmath0 is larger than @xmath1 , one expects underscreening , _ independently _ of the representation chosen for the impurity . because the @xmath9-channel large @xmath1 limit which describes overscreening is obtained for @xmath15 @xcite , we can guess that a reasonable large @xmath1 limit of underscreening can be found when @xmath16 ( @xmath17 being the number of `` spin channels '' ) .
we will show that the saddle - point equations derived in this way at @xmath3 present a `` dual '' structure to the non crossing approximation associated to the overscreened case @xcite , and reveal non fermi liquid behavior with a continuously varying exponent parametrized by the ratio @xmath18 .
the lattice version of this generalized underscreened kondo model is also straightforwardly solvable , due to the absence of generated rkky interactions .
because singular scattering of delocalized electrons on the degenerate manifold of unscreened spin occurs , coherence can not establish over the lattice , in contrast to the exactly screened kondo lattice @xcite .
this proves the existence of a generic insulating - like state with a pseudogap density of states , parametrized by the same anomalous exponent unveiled in the single impurity underscreened kondo model .
we conclude the paper by future prospects regarding the theory of underscreened kondo models , and by the possible observation of their physical consequences in strongly correlated materials , pointing out some similarities with the cerium based compound cecuas@xmath4 , on which transport measurements were recently reported .
the remainder of the paper is organized as follows : in section [ sec : derivation ] we demonstrate how the new large @xmath1 limit of the underscreened kondo impurity model can be performed , followed by a physical discussion of the results in section [ sec : study ] .
the kondo lattice extension is examined in section [ sec : lattice ] .
our aim in this section is to introduce a kondo model in which the underscreened aspect is built in from the beginning , while a simple large @xmath1 limit of the problem can be found by proper rescaling of the parameters .
the basic idea , which was presented in the introduction , is to consider a single - channel kondo hamiltonian involving conduction electrons with @xmath1 spin flavors and interacting with a su(@xmath0 ) spin @xmath19 localized at the origin : @xmath20 where @xmath21 , @xmath22 .
the matrix of coupling constants @xmath23 will be specified in the following .
we choose a completely antisymmetric representation of the spin using abrikosov fermions : @xmath24 where the second equation is the necessary constraint to enforce the spin size @xmath25 ( this scaling , where @xmath26 is of order 1 , is necessary to get a large @xmath1 limit , see below ) . as emphasized previously
, @xmath27 leads to underscreening , and a likely scaling to obtain a solvable limit is @xmath28 . let us therefore set @xmath29 , with @xmath30 , expressing fermions in a double index notation : @xmath31 where @xmath32 represents the doublet of indices @xmath33 , with @xmath34 .
neglecting potential scattering terms , which only contribute to next order in @xmath14 , the hamiltonian now reads : @xmath35 this is now completely general , and we would like to chose the simplest coupling between the itinerant su(@xmath1 ) fermions and the localized su(@xmath36 ) spin . in the spirit of having @xmath37 `` spin channels '' , we set : @xmath38 and obtain the hamiltonian @xmath39 associated with the _ single _ constraint eq .
( [ new_constraint ] ) .
this will be our starting point for the large @xmath1 solution .
one might argue that coupling a band of fermions to a spin with a different symmetry group is not a very transparent concept on a physical point of view .
indeed , one can try to view alternatively the hamiltonian ( [ hfinal ] ) as the standard su(@xmath1 ) kondo model with a localized su(@xmath1 ) spin taken in a _
representation @xcite of size @xmath40 : @xmath41 usual kondo coupling to the su(@xmath1 ) electronic bath leads indeed to the hamiltonian ( [ hfinal ] )
. we will however not pursue this route in the present work , as this type of representation asks to consider a _ set _ of @xmath42 constraints : @xmath43 which makes the large @xmath1 limit impracticable .
we will come back to the differences between the large group technique ( developed in this work ) and the rectangular representation approach in section [ sec : comparison ] , while discussing the physical results . in order to solve the model ( [ hfinal]-[new_constraint ] ) ,
we first write the imaginary time action of the problem with inverse temperature @xmath44 : @xmath45 \\ & + & { \int_0^\beta \!\!\!\ ! { { { \mathrm{d}}}}\tau}\ ; \bigg [ \frac{j}{n } \sum_{kk'{\sigma}{\sigma}'\alpha } { c^\dagger_{k\sigma}}c^{{\phantom{\dagger}}}_{k'\sigma ' } \ ; f^\dagger_{\alpha{\sigma } ' } { f^{{\phantom{\dagger}}}_{\alpha\sigma}}- q_0 p \lambda \bigg ] \end{aligned}\ ] ] introducing a lagrange multiplier @xmath46 to enforce the constraint ( [ new_constraint ] ) . the precise form of the action ( [ action ] ) is hinting that a large @xmath1 solution is possible , which we perform now .
the following step is to decouple the interaction with a bosonic field @xmath47 : @xmath48 \\ \nonumber & + & { \int_0^\beta \!\!\!\ ! { { { \mathrm{d}}}}\tau}\bigg [ \sum_\alpha \frac{{b^\dagger_{\alpha}}{b^{{\phantom{\dagger}}}_{\alpha}}}{j } - q_0 p \lambda + \sum_{k \alpha { \sigma } } { c^\dagger_{k\sigma}}{f^{{\phantom{\dagger}}}_{\alpha\sigma}}\frac{{b^{{\phantom{\dagger}}}_{\alpha}}}{\sqrt{n } } + h.c .
\bigg ] \end{aligned}\ ] ] and integrate out the fermions @xmath49 : @xmath50 \\ \nonumber & + & { \int_0^\beta \!\!\!\ ! { { { \mathrm{d}}}}\tau}{\int_0^\beta \!\!\!\ ! { { { \mathrm{d}}}}\tau } ' \frac{1}{n } g_{f0}(\tau-\tau ' ) \sum_{kk ' \alpha { \sigma } } \big({c^\dagger_{k\sigma}}{b^{{\phantom{\dagger}}}_{\alpha}}\big)_{\tau } \big({b^\dagger_{\alpha}}c^{{\phantom{\dagger}}}_{k'\sigma } \big)_{\tau'}\end{aligned}\ ] ] where @xmath51 , with @xmath52 .
we finally introduce the electron sitting at the impurity center , @xmath53 , so that : @xmath54 \\ \nonumber & - & { \int_0^\beta \!\!\!\ ! { { { \mathrm{d}}}}\tau}\!\ ! { \int_0^\beta \!\!\!\ ! { { { \mathrm{d}}}}\tau } ' g_{c0}^{-1}(\tau\!-\!\tau ' ) \sum_{\sigma}{c^\dagger_{\sigma}}(\tau ) { c^{{\phantom{\dagger}}}_{\sigma}}(\tau ' ) \\ \nonumber & + & { \int_0^\beta \!\!\!\ ! { { { \mathrm{d}}}}\tau}{\int_0^\beta \!\!\!\ ! { { { \mathrm{d}}}}\tau } ' \frac{1}{n } g_{f0}(\tau-\tau ' ) \sum_{\alpha { \sigma } } \big({c^\dagger_{\sigma}}{b^{{\phantom{\dagger}}}_{\alpha}}\big)_{\tau } \big({b^\dagger_{\alpha}}c^{{\phantom{\dagger}}}_{\sigma } \big)_{\tau'}\end{aligned}\ ] ] where @xmath55 . because @xmath56 scales as @xmath1 , the existence of a saddle - point is manifest in the previous expression .
following @xcite , we find the integral equations : @xmath57 where @xmath58 is a bosonic matsubara frequency .
let us first comment on the formal analogy that the system of equations ( [ gc]-[sb ] ) shares with the usual non crossing approximation ( nca ) @xcite .
we notice indeed that the roles of the local bath electron @xmath59 and of the abrikosov fermion @xmath49 are simply exchanged with respect to the usual nca structure , which , we recall , is associated to overscreening @xcite .
in particular , the self - energy now involves the propagator @xmath51 instead of @xmath55 in the nca , and this difference will be shown below to radically affect the physics , with the occurence of underscreening instead of overscreening .
this interesting `` duality '' is one of the main results of the present work .
it is also very appealing to remark that the computation of physical observables related to the itinerant band does not involve directly pseudo - fermion degrees of freedom , contrarily to the multichannel case .
indeed , the present theory is formulated directly in terms of the bath propagator , to which the t - matrix is simply related by the relation : @xmath60 if we sum the previous expression over all momenta , we find : @xmath61 where ( [ gc ] ) was used to obtain the second expression .
interestingly , we note that the t - matrix ( [ t_mat_kondo ] ) is of order @xmath62 in the present scheme , whereas it happens to scale as @xmath14 in the multi - channel large @xmath1 limit @xcite , a certain drawback for the theory of overscreening .
however , by the `` duality '' argument , we expect to find that the spinon propagator @xmath63 only shows @xmath14 corrections to the free impurity limit ( @xmath64 ) , which can indeed be easily checked by a direct computation : @xmath65 this means that , although the t - matrix is conveniently captured by the present scheme , computation of e.g. the spin susceptibility involves necessarily @xmath14 corrections .
this remark allows however to solve explicitely the constraint equation ( [ lambda ] ) at the leading order : @xmath66
for simplicity , we start with the assumption of particle - hole symmetry , @xmath67 ( the free bath @xmath68 will always be assumed symmetric in the following ) .
this implies that the constraint ( [ lambda ] ) is fulfilled provided @xmath69 , and we have : @xmath70 i.e. long range correlations in the self - energies ( [ sc]-[sb ] ) . by analogy with the low - frequency analysis
usually performed in studying the nca equations @xcite , we assume power law behavior of the self - energies and plug into the saddle - point equations
. self - consistency can be achieved at zero temperature ( see appendix [ app1 ] ) and we obtain for real frequency quantities : @xmath71 where @xmath72 and @xmath73 are undetermined constants .
this analytical result is well borne out by the numerical solution of the saddle point equations as shown by figure [ fig : power ] ( in the following computations , a semi - circular density of states with half - width @xmath74 was chosen to model the bath of conduction electrons ) .
we would like to contrast this peculiar low energy behavior with the one observed in a standard fermi liquid ( as in the fully screened kondo model ) .
the interpretation of these results can be made clearer by considering the t - matrix of a related anderson model : @xmath75 where the local fermion @xmath76 has an hybridization @xmath77 to the @xmath78-electrons and is subject to a local coulomb repulsion term @xmath79 ( and possibly to a strong hund s rule if one wants to stabilize a @xmath2 spin in order to make the connection to the underscreened case ) , giving rise to the term @xmath80 in the previous equation . comparing ( [ t_mat_kondo ] ) and ( [ t_mat_anderson ] )
, we see that the inverse self - energy of the bath electrons is simply related to the impurity self - energy by @xmath81/v^2 $ ] . in a fermi liquid , we know @xcite that the impurity self - energy obeys at low frequency : @xmath82 where @xmath83 is the quasiparticle residue and @xmath84 the inelastic scattering rate .
the previous general identification between @xmath85 and @xmath86 provides the self - energy of the local electron in the bath for a fermi liquid : @xmath87 the first term corresponding to regular elastic scattering from the impurity . the fact that this self - energy diverges as a power law at low frequency in the underscreened case ( [ sigma_c ] ) signals anomalous scattering on the remaining unscreened degrees of freedom , which ultimately violates the fermi liquid description of the problem .
this result can be more clearly witnessed from the impurity self - energy itself @xmath88 showing a power law scattering rate for the localized state which even dominates the linear part found in the self - energy of a fermi liquid , equation ( [ eq : sfl ] ) . indeed , we deduce from this behavior a frequency dependent quasiparticle residue @xmath89^{-1 } \propto |{\omega}|^{1-\alpha}$ ] which _ vanishes _ at low energy , demonstrating a clear breakdown of fermi liquid theory . a final way to put into evidence the singular energy dependence of this underscreened impurity is to compute the t - matrix .
we find indeed that the kondo resonance , whose development is illustrated on figure [ fig : spectral ] , displays a cusp at low frequency : @xmath90 using in equation ( [ t_mat_kondo ] ) the low frequency behavior of the free green s function , @xmath91 .
this expression shows that , although the unitary limit is recovered in the t - matrix at @xmath92 , fermi liquid behavior is nevertheless violated by the non - analytic expansion a low frequency .
here we discuss briefly the case @xmath93 , and investigate whether the previous results remain valid when particle - hole symmetry is broken .
the propagator appearing in the bubble giving the self - energies , equations ( [ sc]-[sb ] ) , is given by : @xmath94 and decays exponentially .
we argue that the long range correlations which are crucial to maintain the non - trivial power laws are still present , because @xmath95 saturates at low temperature , from ( [ lambda_dominant ] ) .
this would imply that the non fermi liquid state survives the introduction of the asymmetry parameter @xmath96 , as can also be verified from the numerics .
however , the structure of the saddle point is such that the constraint ( [ new_constraint ] ) scales as @xmath42 instead of @xmath1 , and hints that it might be necessary to account for @xmath14 corrections to the result ( [ lambda_dominant ] ) .
we do not find that this possibility really modifies the previous result , @xmath97 , although we think this question deserves further clarification .
the present approach seems well suited to describe the underscreened situation of quantum impurity models .
we can however point out two kinds of limitations , which are in fact also inherent to the non crossing approximation of multichannel models as well @xcite .
first , although the t - matrix can now be naturally extracted at the saddle - point level , some physical quantities such as the spin susceptibility appear only at next to leading order in 1/@xmath1 .
second , motivated by some puzzles raised in heavy fermion compounds , it remains of great interest to develop faithfull large @xmath1 expansion methods that can work in the exactly screened case too ( an interesting step in this direction was however taken in @xcite ) .
it was in fact already acknowledged @xcite that a simple generalization of the kondo model is quite suitable for overcoming both of these difficulties : it consists in taking the localized spin in a rectangular representation of su(@xmath1 ) , as indeed done in the present work ( if one were able to take good care of the necessary constraints , see discussion in section [ sec : alt ] ) , while assuming also a large number ( of order @xmath1 ) of screening channels ( in this paper we only considered the single channel case ) .
this extension of the formalism , also known as `` matrix kondo model '' , leads however to technical difficulties due to proliferation of feynman graphs at large @xmath1 .
it is actually possible to resum this complicated diagrammatics using a soft constraint approach inspired by a recent study @xcite of the interplay of kondo effect and coulomb blockade ( viewed now as a purely mathematical trick ) .
such a theory would be a bridge between the usual large @xmath1 approach of overscreening and the present work , both appearing as simple limitating cases of it , thus finding an unity behind our `` duality '' picture .
such a theory will be considered in a forthcoming publication @xcite as we have demonstrated from the analytical and numerical solution of the new large - n equations , the underscreened kondo impurity model displays violations of low energy fermi liquid behavior , although those are less severe as the one observed in the overscreened case ( where the deviation from the unitary limit associated to finite lifetime of the electrons at zero temperature even dominates the anomalous frequency dependence of the scattering rate ) .
the underscreened impurity is nevertheless associated to non fermi liquid behavior : the local quasiparticle weight vanishes at low energy and the system presents additional degeneracy due to the non fully screened spin . it is interesting to compare this result to previous exact knowledge from the bethe ansatz solution of underscreened kondo models .
although the general features we have sketched so far are indeed obtained from the bethe ansatz analysis , one crucial difference concerns the precise nature of the singular frequency dependence of the physical quantities : while we observe universal non - trivial exponents in the large n solution , the @xmath2 su(2 ) kondo model rather displays logarithmic behavior @xcite ( although power laws are actually found in the anisotropic case @xcite ) .
we think this difference is a minor point , that we would naively like to attribute to special features associated to the large n generalization of the model .
this point is actually more subtle . indeed , the hamiltonian ( [ hfinal ] ) we have solved can be argued ( see section [ sec : alt ] ) to be equivalent to a su(@xmath1 ) kondo model with spin in a rectangular representation , which is actually diagonalized by bethe ansatz @xcite , and found to display also logarithmic singularities ! as stressed previously , we have however enforced a different type of constraint in the present approach , and we would ultimately pin point this difference to be the origin of the discrepancy in the low energy behavior as compared to the exact solution .
this discussion would tend to illustrate in any case the fact that the physics can be sensitive in how the constraint is implemented , but we will reinforce our judgment that the present method captures the essential features of underscreening .
as we will now demonstrate , the main interest of our approach is that it can apply to problems where no exact solution is available , in particular in the many impurities extension of the underscreened kondo model .
we introduce here the lattice extension of the previous single impurity model , which consists of a dense network of impurities carrying a su(@xmath0 ) spin on which a band of su(@xmath1 ) itinerant electrons scatter : @xmath98 here @xmath99 labels sites , with @xmath100 , @xmath101 is the chemical potential in the @xmath102-band and other conventions are similar to previously .
this model is not very realistic in many respects , and to date lacks a physical realization in condensed matter systems ( see however the ending discussion in section [ discussion ] ) . on a purely theoretical point of view
, we can raise however two interesting questions .
first , what is the behavior of the electronic band below the kondo temperature associated to underscreening ? for this to make sense , we need to argue that the magnetic processes involving the remanent spin degrees can be ignored , for example due to frustration ( maybe by analogy to the compound liv@xmath4o@xmath103 ) .
second , when the partially - screened local moments start to order , and quench their macroscopic entropy , how does the system behaves ? in the following large @xmath1 analysis of this problem , we will see that intersite magnetic correlations are absent , and that the first of these questions can be answered .
the large @xmath1 limit is derived following the same steps performed in section [ sec : derivation ] , introducing local bosons @xmath104 to decouple the kondo term on each site , and integrating the abrikosov fermions : @xmath105 we can simply read off this expression the final saddle point equations , which are completely identical to ( [ gb]-[sb ] ) , except that the local bath propagator now reads : @xmath106 the new system of integral equations ( [ gb]-[sb]),([gc_latt ] ) is remarkably simple in its structure , a fact due to the absence of intersite correlations at this level of approximation .
indeed , as the action ( [ action_lattice ] ) shows , no magnetic rkky interaction is generated at the saddle - point level , which is expected since the additional quantum number @xmath107 carried by the localized spins can not be transported from site to site by the itinerant fermions .
we point out however that this is not a general feature of underscreened models , and one can easily check that the rkky interaction is indeed mediated by the electronic bath in the finite @xmath8 su(2 ) case .
besides , expression ( [ gc_latt ] ) is reminiscent of a self - consistent t - matrix approximation and signals that electrons in the bath are still anomalously scattered by the localized spins , acting independently of each other . indeed ,
if we assume that @xmath108 is divergent as in the single impurity case , equation ( [ sigma_c ] ) , we see that this singular self - energy dominates the @xmath109-summation in ( [ gc_latt ] ) at low energy , so that : @xmath110 with the _ same _ exponent ( [ alpha ] ) as found previously .
this means that in the underscreened kondo lattice , there is no important distinction between the cases of dense and diluted impurities ( on the point of view of the bath electrons that feel the kondo interaction ) , and that the itinerant electron density of states always shows a pseudogap at low energy .
this is quite different from the situation of exactly screened models , where a hard hybridization gap would open at half - filling . in the perfectly screened case , coherence
can always be re - established upon doping , despite the fact that each individual impurity scatters strongly the electrons .
the result ( [ g_c_lattice ] ) would however let us think that electronic degrees of freedom always remain confined in the underscreened kondo lattice , as we now check on the numerical solution of the saddle - point equations
. we will be again interested in the t - matrix , which is related to the ( translation invariant ) @xmath102-electron green s function @xmath111 by : @xmath112 this relation can be best understood from an equivalent periodic anderson model , where @xmath113 is proportional to the momentum- and frequency - dependent green s function of the localized electrons . from the effective action ( [ action_lattice ] )
we have : @xmath114 so that the t - matrix can be expressed as : @xmath115 again , we identify @xmath116 as the impurity self - energy @xmath80 ( up to a factor @xmath117 ) .
the local t - matrix , @xmath118 is easily calculated from ( [ t_lattice ] ) after the numerical solution of the saddle - point equations , and is shown in figure [ fig : lattice ] ( here also a semi - circular density of states for the @xmath102-electron was taken ) . from the numerical solution , and in agreement with the previous analytical analysis , we see that an _ hybridization pseudogap _ opens in the spectrum , irrespective of the filling of the @xmath102-band .
this prevents coherence to be reached over the lattice at zero temperature ( note that figure [ fig : lattice ] is performed at finite temperature , so that the pseudogap in the density of states is filled by thermal excitations ) .
therefore , the strongly underscreened kondo lattice is strictly speaking an insulator , with a resistivity likely showing a power law increase as temperature is lowered , instead of the activated or metallic behavior expected respectively in usual kondo insulators or in heavy fermion metals . in order to shed some light onto this remarkable result , we put it into the perspective of previous theoretical work and experimental measurements on a newly discovered @xmath5-electron material .
the main physical result of the paper is that the underscreened kondo lattice is an insulator with a pseudogap density of states above the ordering temperature of the partially screened moments , for all filling of the electronic band .
theoretically , we can ask whether this behavior is peculiar to the large @xmath1 solution performed here or can be expected to be generic . to answer this question ,
it is interesting to notice two facts . first , previous work by k. le hur in a one dimensional model of underscreened kondo ladder @xcite demonstrated infact similar insulating behavior ( this quite peculiar model was chosen in order to prevent the haldane spin gap associated to the 1d @xmath2 chain , which would pollute the interesting physics ) .
second , our large @xmath1 solution displays actually some high dimensional features @xcite , such as a local self - energy , as seen in equation ( [ local_self ] ) .
because this insulating behavior is found in two such extreme cases , we could expect for a continuity of this non - metallic behavior of underscreened lattice in intermediate dimension also .
we now turn to the nature of this insulating state .
as discussed in the single impurity case , section [ sec : comparison ] , the large @xmath1 solution predicts that the dynamics is governed by universal power law behavior rather than the logarithmic singularities found in the @xmath2 su(2 ) case and extensions , although the physics of these anomalies can be expected to be similar .
this discussion extends naturally to the underscreened lattice , and deserves clearly further theoretical studies .
we think that a next step could be taken by solving the @xmath2 kondo lattice model in a combination of dmft @xcite and nrg techniques , and would help both to confirm the generic insulating behavior found here , and investigate in more detail the physical properties that are associated to it .
we finally discuss the possible experimental realization of underscreened kondo physics . concerning the single impurity case , the stability of @xmath2 for quantum dots in even valleys of coulomb blockade was discussed both theoretically and experimentally ( for a review ,
see @xcite ) , although no emphasis was put onto the singular behavior that we have discussed previously .
another way to observe kondo underscreening would be to study large @xmath8 atoms deposited on a metallic surface and scanned by an stm tip .
finding an @xmath5-electron material related to the underscreened kondo lattice seems to be a more daunting task .
indeed , not only the intersite magnetic correlations would need to be small in such a system , but a mechanism to form large spin ( or spin - orbital degeneracy ) of the @xmath5-level should be invoked . typically , crystal fields leave unfortunately an effective kramers doublet below temperature of several dozens of kelvins . however , we think that the present work is interesting since it highlights the peculiar physical properties of underscreened lattices , and may be useful in strengthening an eventual experimental observation .
in fact , very recent transport measurements of a new cerium based compound , cecuas@xmath4 , revealed weakly insulating behavior of the resistivity , with a power law divergence at low temperature @xcite . while we are careful in connecting hastily this observation to our theoretical findinds , we hope that these remarks would stimulate further studies of underscreening effects in kondo systems .
the underscreened kondo model was investigated in this paper by means of a specially developed large @xmath1 technique .
although the strong coupling fixed point in which itinerant electrons are tightly bound to the uncompensated spin is stable in the renormalization group sense , we have shown that non fermi liquid behavior appears in the form of anomalous power laws in the physical observables .
the universal exponent was computed analytically and checked over the numerical solution of integral equations , which show an interesting duality with respect to the previous theory of overscreening in multichannel models .
the extension to a finite dimensional lattice of underscreened magnetic moments was also considered , and a pseudogap ( weakly insulating ) regime was discovered , with similar power laws governing the dynamics of the electrons .
we have also tried to point out several directions for future research , both concerning the theoretical aspects of underscreened models and their experimental realization , for which cecuas@xmath4 might be a candidate .
i wish to thank matthias vojta for some useful suggestions at the early stages of this work , and for comments on the manuscript .
i am also grateful to e. v. sampathkumaran for communicating his experimental results before publication @xcite
. financial support by the center for functional nanostrutures is also acknowledged . during the completion of the present paper ,
a similar large @xmath1 limit of the underscreened kondo model was proposed by p. coleman and i. paul @xcite , with results complementary to ours in the single impurity case . a very detailed investigation of the singularities occuring in the @xmath2 su(2 ) kondo model
was also carried out simultanenously by p. mehta et al .
@xcite using a combination of bethe ansatz and numerical renormalization group techniques .
we present here the derivation of the non - trivial exponent governing the low frequency behavior of the physical quantities , following @xcite .
let us assume the ansatz : @xmath119 at zero temperature . using the spectral decomposition : @xmath120 and the expression for the self - energies ( [ sc]-[sb ] ) with @xmath121
, we find : @xmath122 going back to frequency , we have simply : @xmath123 to determine the real part in the previous self - energies , we use an analyticity argument , which gives for complex frequency @xmath124 : @xmath125 } |z|^{-\alpha_b } \\
\sigma_b(z ) & = & - \frac{a_c}{2 } \frac{e^{i\alpha_c\pi/2}}{\sin[\alpha_c\pi/2 ] } |z|^{-\alpha_c}\end{aligned}\ ] ] and similarly for @xmath126 and @xmath127 .
finally , from dyson s equation ( [ gc])-([gb ] ) , we have @xmath128 and @xmath129 , providing relations between amplitudes @xmath130 , @xmath131 and exponents @xmath132 , @xmath133 . after some manipulations
, we find : @xmath134 | we study a generalization of the kondo model in which the impurity spin is represented by abrikosov fermions in a rotation group su(@xmath0 ) _ larger _ than the su(@xmath1 ) group associated to the spin of the conduction electrons , thereby forcing the single electronic bath to underscreen the localized moment .
we demonstrate how to formulate a controlled large @xmath1 limit preserving the property of underscreening , and which can be seen as a `` dual '' theory of the multichannel large @xmath1 equations usually associated to overscreening . due to the anomalous scattering on the uncompensated degrees of freedom ,
the fermi liquid description of the electronic fluid is invalidated , with the logarithmic singularities known to occur in the @xmath2 su(2 ) kondo impurity model being replaced by continuous power laws at @xmath3 .
the present technique can be extended to tackle the related underscreened kondo _
lattice _
model in the large @xmath1 limit .
we discover the occurence of an _ insulating pseudogap regime _ in place of the expected renormalized metallic phase of the fully screened case , preventing the establishement of coherence over the lattice . this work and
the recent observation of a similar weakly insulating behavior on transport in cecuas@xmath4 should give momentum for further studies of underscreened impurity models on the lattice . |
for @xmath51 we approximate a complicated pattern in @xmath121-integration defined by conditions @xmath122 by a simpler splitting into @xmath55 and @xmath54 .
this is justified as the above - mentioned conditions are not strict , but rather crossovers . here
we give a detailed calculation for @xmath54 [ the opposite case has been considered in the main text , eq .
( [ bigk ] ) ] .
the expression to be evaluated reads @xmath123 we intruduce polar angles @xmath66 for @xmath124 and @xmath125 for @xmath121 as well as @xmath126 . then the integral takes the form : @xmath127 here @xmath128 comes from expanding the numerator in eq .
( [ initialintegral ] ) into harmonics in @xmath66 and reads : @xmath129+{}\\ \left[-\frac{1}{2}(a_x^2-a_y^2)(x^2 - 1)+(a_x^2-a_y^2)x(x-1)\cos^2{\phi}\right]\cos{2\alpha}+{}\\ \left[\frac{(a_x - a_y)^2(1+x)^2}{8}-\frac{(a_x - a_y)^2(1 + 2x)x}{2}\cos^2{\phi}+(a_x - a_y)^2x^2\cos^4{\phi}\right]\cos{4\alpha}.\end{gathered}\ ] ] note that only the first term in the first brackets survives in the isotropic case .
now , using @xmath130 where @xmath131 is the hypergeometric function , we obtain @xmath132 and @xmath133 this leads to @xmath134+\cos{4\alpha}\frac{(a_x - a_y)^2(1+x)^2}{8}\left[\frac{(x-1)^2}{2}f_2-x^2f_1+\frac{(1+x)^2}{2}f_0\right]\biggr\}\end{gathered}\ ] ] expanding the expressions in brackets for @xmath135 yields @xmath136 which means that the corresponding parts of the integrand decrease at least as @xmath137 for large @xmath114 , so the integrals converge well and can be extended to infinity .
the rest of the integrand decreases as @xmath138 , so the convergence for @xmath64 is quite poor .
we add and subtract @xmath139 to @xmath140 in the first term and a unity to @xmath141 in the second term , extend the @xmath114-integral to infinity for well - converging parts , and keep the upper limit for the integral of @xmath142 .
evaluating the integrals with hypergeometric functions numerically and eventually adding contribution from eq .
( [ bigk ] ) , we obtain eq .
( [ smallk ] ) .
note that the subleading , @xmath58-independent term stemming from the integration over @xmath54 has the same dependence on anisotropic material constants as the contribution from @xmath55 [ eq .
( [ bigk ] ) ] .
here we give details of the evaluation of the integral appearing in eq .
( [ sigma_2 ] ) for @xmath143 .
first we take the @xmath66-integral , in which we keep only the leading term in @xmath144 .
the denominator in the integrand has a sharp minimum at @xmath145 for small @xmath84 , so we can replace @xmath146 by @xmath147 and @xmath148 by @xmath149 in the numerator .
then the @xmath66-integral is evaluated straightforwardly for both @xmath68 and @xmath51 using the identity in eq .
( [ cosint ] ) , and neglecting 1 compared to @xmath150 .
this yields the @xmath151 behavior seen in eq .
( [ eqsigma ] ) .
the @xmath58-integrals for both regions are now trivial .
to obtain @xmath91 , we now have to collect all contributions and substitute @xmath85 ( the latter gives an extra @xmath152-dependence ) .
the results for @xmath51 are here @xmath155 is some numerical constant of the order of unity that comes from the logarithms in eq .
( [ bigq ] ) .
if we use that expression , then @xmath156 . making any corrections here
is meaningless , as the formal derivation of eq .
( [ bigq ] ) implied @xmath157 and here we have to deal with @xmath158 , so just ignoring the logarithm seems to be reasonable .
furthermore , @xmath159 is about 5 times smaller than @xmath160 as can be seen from the values of the material constants given in the main text .
so , luckily , calculating any corrections to @xmath155 is irrelevant for our work and in our calculations we replaced it with 2 . | flexural thermal fluctuations in crystalline membranes affect the band structure of the carriers , which leads to an exponential density - of - states ( dos ) tail beyond the unperturbed band edge .
we present a theoretical description of this tail for a particular case of holes in single - layer black phosphorus , a material which exhibits an extremely anisotropic quasi - one - dimensional dispersion ( @xmath0 ) and , as a result , an enhanced van hove singularity at the valence band top .
the material parameters are determined by _ ab initio _ calculations and then are used for quantitative estimation of the effect of two - phonon ( flexural ) processes have on the charge carrier dos .
it is shown that unlike the isotropic case , the physics is determined by the phonons with wavevectors of the order of @xmath1 , where @xmath1 determines the crossover between harmonic and anharmonic behavior of the flexural phonons .
the spectral density of the holes in single - layer black phosphorus at finite temperatures is calculated .
few - layer black phosphorus attracts a lot of attention now as a prospective two - dimensional ( 2d ) semiconducting material with a tunable energy gap @xcite .
being , as other 2d materials , an atomic - thick crystalline membrane , it is affected by thermal flexural fluctuations which results in anomalous structural and mechanical properties at finite temperatures @xcite .
it was suggested @xcite that these thermal fluctuations can essentially influence on the states of charge carriers in 2d semiconductors and even lead to an autolocalization , that is , formation of a self - trapped state ( flexuron ) . that general consideration assumed an isotropic electronic spectrum whereas black phosphorus is highly anisotropic , with dramatically different effective masses , @xmath0 . especially , in the hole - doped case the situation is quite peculiar , with a formation of van hove singularity near the top of the valence band which is essentially enhanced by mass anisotropy @xcite . one can expect that the holes in single - layer black phosphorus ( also called phosphorene ) are quasi - one - dimensional quasiparticles and thus are strong candidates for a pronounced effect of localization due to interaction with bosonic modes . in general , deformation of the lattice in 2d materials results in distortion of the carrier dispersion .
for different types of deformations this effect can be accounted for by different phonon modes interacting with fermionic charge carriers .
the interaction parameters then can be determined either from experiment or from different theoretical approaches . in 2d materials like graphene or black phosphorus
it is a well - justified approach to separate in - plane from out - of - plane , or flexural phonon modes for not too small wavevectors @xcite .
on the other hand , for smaller wavevectors these modes begin to interact which results in changing of the functional form of the correlation function dependence on the wavevector @xcite .
this will be discussed in more detail below , in particular for the anisotropic case . in what follows
we calculate the spectral function and the density of hole states ( dos ) in black phosphorus renormalized due to the interaction with flexural phonons . to this end
, we use the diagrammatic approach and calculate the hole self - energy using the ward identity as a low momentum transfer approximation for the vertex .
this seems to be the minimal model that allows one to get an exponential dos tail and is thus a more suitable approach than e.g. the self - consistent born approximation .
this approximation was earlier suggested in a context of magnetic semiconductors near the critical point @xcite .
its relation to the path - integral approach used in ref . is discussed in ref . .
the fermionic hamiltonian is given by @xmath2 here @xmath3 , @xmath4 are the creation and annihilation operators of the fermions with momentum @xmath5 ( at point @xmath6 ) , @xmath7 is the dispersion of the fermions and @xmath8 is the effective potential acting on them due to lattice distortion . for flexural deformations in systems with symmetry plane ( or gliding plane as in case of black phosphorus ) , the effective potential is quadratic in gradients of the lattice atoms shift @xmath9 in the direction , perpendicular to the sample plane @xmath10 @xcite : @xmath11 the mixed @xmath12 term is suppressed for the orthorhombic symmetry and we disregard it in our consideration .
@xmath13 is the fourier transform of the flexural lattice displacement field .
correlators of the @xmath14 field are found from the elasticity theory equations for membranes . for isotropic membranes
the well - known result is @xcite @xmath15 here @xmath16 denotes gibbs averaging , @xmath17 is the temperature , @xmath18 is the material constant , determining the flexural rigidity of the membrane , @xmath19 is the critical exponent ( @xmath20 in the self - consistent screening approximation @xcite and @xmath21 in monte carlo simulations for graphene @xcite ) and @xmath1 is the characteristic wavevector , such that for @xmath22 the interplay between flexural and in - plane deformations becomes crucial and softens the flexural modes .
neglecting this effect would lead to unphysical low wavevector divergence in @xmath23 . in the isotropic case @xmath1
is given by @xmath24 where @xmath25 is the young modulus @xcite .
before we proceed , note that averaging of @xmath26 yields a constant that just shifts the whole fermionic band and has no non - trivial effects . to account for those we have to calculate the second cumulant of the bosonic field
@xmath27 .
then , neglecting higher cumulants , we can write for the fermionic self - energy @xmath28 where @xmath29 is the fermion s green s function , @xmath30 is the three - leg vertex and @xmath31 is the energy of the fermion ( cf
. refs . ) .
the approximation @xmath32 corresponds to the self - consistent born approximation which is supposed to be enough for positive energies ( that is , within the band ) . to take into account also the fluctuation dos tail
, vertex corrections should be taken into account . at this stage
we argue that for energies far enough from the band edge only small momentum transfers play a role @xcite , so that we neglect momentum dependence of the vertex and of the self - energy and use the ward identity for the vertex : @xmath33 ending up with the following closed differential equation for the self - energy : @xmath34 now we go on to estimate @xmath35 for the anisotropic case , which is a question of its own significance .
for @xmath36 it has been done recently in ref .
, so here we do the calculations for small wavevectors .
a consistent theory of anisotropic membranes is to our knowledge not yet fully developed , so in this paper we make plausible assumptions for the low-@xmath37 behavior of the cumulant @xmath38 providing their consistence with eq .
( [ cumulant_isotropic ] ) . following the reasoning of ref .
, we model the anisotropy of the bending rigidity by writing for large @xmath37 : @xmath39 as a next step , we rescale the @xmath40-components of the wavevector introducing @xmath41 , and write @xmath42 with @xmath43 , @xmath44 and @xmath45 is the young moduli in @xmath40 directions . finally ,
introducing @xmath46 and shifting @xmath47 , we obtain for @xmath35 : @xmath48 ^ 2\left < h_{\mathbf{k}+\frac{\mathbf{q}}{2}}h_{-\mathbf{k}-\frac{\mathbf{q}}{2}}\right > \left
< h_{\frac{\mathbf{q}}{2}-\mathbf{k}}h_{\mathbf{k}-\frac{\mathbf{q}}{2}}\right>. \quad\end{aligned}\ ] ] to get this expression , we take @xmath26 given by eq .
( [ potential ] ) and calculate the fourier transform of @xmath49 over @xmath50 , which is done using the wick s theorem to decouple the average of the product of four @xmath13 s .
the latter corresponds to the self - consistent screening approximation @xcite .
finally , the rescaling of the components of all wavevectors is made as described above .
as the conditions @xmath51 , @xmath52 are rather crossovers than strict boundaries , we can within the same accuracy split the @xmath53-integral into two parts ( 1 ) @xmath54 and ( 2 ) @xmath55 . for the first one we take the anharmonic expression for the @xmath56 correlators , and for the second part we use the harmonic one .
moreover , we limit ourselves to calculating only the leading terms in @xmath57 , so for @xmath55 we just neglect the @xmath58-dependence to obtain the contribution to @xmath59 : @xmath60 the integration over @xmath54 is more involved , as we have to keep the @xmath61-dependence of the integrand , which leads to anisotropic contributions . here , apart from the leading term @xmath62 , we also keep the subleading term @xmath63 , as for @xmath64 they are numerically of the same order of magnitude for not extremely small @xmath58 . evaluating the integral @xcite and adding it up with the contribution given by eq .
( [ bigk ] ) , we arrive at @xmath65 here the numerical factors in the parentheses are results of numerical integration , for which the specific value for @xmath64 was used ; @xmath66 is the polar angle of the @xmath61 vector ( not the @xmath67 vector ! ) . the case @xmath68 has been considered in ref . .
the result is : @xmath69 where @xmath70 is some number determined by details of @xmath71 behavior for @xmath72 . after obtaining the above expressions for @xmath59 , our immediate goal is to evaluate the integral in eq .
( [ sigma_equation ] ) . as it has been mentioned above
, the dispersion of the holes in black phosphorus is quasi - one - dimensional .
it has been evaluated using _ ab initio _ numerical methods in our previous work @xcite and can be approximated in atomic units by @xmath73 in these units , the averaged @xmath74 for room temperature .
note that if the coefficient at @xmath75 were exactly zero we would have an enhanced van hove singularity in the bare density of states @xmath76 at the band edge , instead of a step in generic 2d case . as it will be shown below , the integral in eq .
( [ sigma_equation ] ) is determined by @xmath72 .
for these values the first term in the dispersion dominates , so we can take @xmath77 we perform the same variable change @xmath78 and introduce @xmath79 to rewrite our equation as @xmath80 consider first @xmath51 .
as @xmath81 and @xmath82 with @xmath83 , we easily confirm , that for small @xmath84 the @xmath58-integral is determined by the upper limit @xmath85 .
analogously , for @xmath68 , the @xmath58-integral is determined by the lower limit , as @xmath86 .
note that for conventional 2d - dispersion , this reasoning does not work and the integral acquires its value on @xmath87 for @xmath88 .
estimation of the integral for @xmath89 allows us to recast eq .
( [ sigma_2 ] ) as @xmath90 where @xmath91 is a dimensionless constant , depending on material constants . using the numerical values from the first - principles calculations @xcite ( @xmath921.31 ev / @xmath93 , @xmath945.68 ev / @xmath93 , @xmath951.29
ev , @xmath965.62 ev , @xmath972.11 ev , @xmath980.89 ev , @xmath99 , @xmath100 being the free electron mass ) it can be estimated that @xmath101 @xcite .
introducing @xmath102 ^ 2,\ ] ] we arrive at the ricatti equation for the function @xmath103 @xmath104 at this point we note that the inequality @xmath105 holds reasonably well , what justifies the above calculations .
the last equation can be turned into a linear one by the substitution @xmath106 .
the corresponding equation reads @xmath107 which is the airy equation .
thus the general solution for @xmath103 is given by @xmath108 here ai and bi are the airy functions and @xmath109 is an arbitrary complex constant , which can be determined by assuring we get correct density of states for @xmath110 , or in other words @xmath111 @xcite . using the asymptotic behavior of the airy functions
it can be checked that for @xmath112 , @xmath113 for large positive @xmath114 which yields the correct behavior for @xmath84 . and
400 k. for reference , the dos corresponding to bare dispersion without the inclusion of the flexuron tail is shown .
the van hove singularity in the latter manifests the aforementioned quasi - one - dimensionality of the holes in black phosphorus .
it is to a large extent smeared by the flexural modes . ] and @xmath115 axis for @xmath116 k. note the different scale for the two plots . ]
now it is straightforward to calculate the dos : @xmath117 with @xmath118 taken from the _ ab initio _
@xmath119 calculations @xcite . the result for different temperature
is shown in fig.[dos_plot ] .
the spectral function @xmath120 for room temperature is shown in fig .
[ spectr ] .
according to the results presented on these figures the effect of flexural phonons on the density of states near the top of the valence band ( the bottom of the band in hole representation ) is quite noticeable and can be probed by conventional spectroscopic tools like angle - resolved photoemission or scanning probe microscopy and other tunneling methods .
quasi - one - dimensional character of the charge carrier dispersion makes hole - doped black phosphorus the best candidate to study these effects .
such experiments would be of a great interest , as an electronic probe of complicated physics of fluctuating anisotropic membranes .
the research has received funding from the european union horizon 2020 programme under grant no .
696656 graphene core1 , and from the stichting voor fundamenteel onderzoek der materie ( fom ) , which is financially supported by the nederlandse organisatie voor wetenschappelijk onderzoek ( nwo ) . |
recently there has been a significant attention to the chaotic behaviour of the solutions to partial differential equations . for example , chaos was found in a perturbed sine - gordon equation @xcite , nonlinear schrdinger equation @xcite , and later for korteweg - de vries - burgers equation @xcite and its generalization on high - order nonlinearities and the case of kadomtsev - petviashvili equation @xcite .
it was shown how chaos can appear in such systems which were completely integrable without perturbation via the appearance of subharmonics and homoclinic tangles .
for the kdv equation incorporating dissipation and instability @xcite it was shown that for the strongly dissipative case the overall evolution of solutions is chaotic with irregular soliton interactions . in our previous paper @xcite
we ve considered the influence of time - periodic hamiltonian external perturbations on a kdv system and showed , how the stochastic layer can appear on a phase plane near the unperturbed separatrix .
the width of this layer was calculated with the help of chirikov criterion for the overlap of resonances . in this work
we continue this study on an example of one - harmonic perturbation .
the fact that this perturbation is taken explicitly provides the possibility to obtain an exact expression for the melnikov function in order to determine the transition to a chaotic behaviour . to study the system in a near - separatrix region , we transform the governing ode into a map for which the width of stochastic layer can be easily obtained . finally we plot poincar sections of chaotic behaviour and calculate the corresponding lyapunov exponents .
we start with the perturbed korteweg - de vries equation taken in the form : @xmath0 here @xmath1 is assumed to be periodic on its last argument and will be taken simply as @xmath2 . transforming coordinates in a moving frame @xmath3 and considering steady waves @xmath4 , one obtains ( the primes are omitted ) : @xmath5 where @xmath6 ( supercritical case ) or @xmath7 ( subcritical case ) , @xmath8 , @xmath9 is the integration constant which can be chosen equal to zero from the condition that the solution is bounded at infinity . without loss of generality
we consider here only the supercritical case .
therefore we have : @xmath10 or with @xmath11 @xmath12 the outline of the rest of the paper is as follows . in the next section the unperturbed case of the korteweg - de vries equation
is considered and the explicit expressions for the solutions are obtained . later in section 3
conditions of chaos are derived on the basis of studying subharmonic and homoclinic melnikov functions .
then , in section 4 , we transform our ode ( [ 1.4 ] ) into a mapping , for which phase plots are presented together with the calculation of the width of the stochastic layer .
section 5 contains poincar plots numerically found in the vicinity of a saddle along with the lyapunov exponents which prove to be positive thereby supporting our conclusion about chaotic behaviour in the system .
finally , section 6 contains a summary and conclusions .
for the unperturbed system @xmath13 we simply have @xmath14 this is an equation of motion for a nonlinear oscillator in a potential field @xmath15 with the hamiltonian @xmath16 if we now look for the solutions of ( [ 2.1 ] ) in the form of elliptic jacobi functions of modulus @xmath17 : @xmath18 where @xmath19 , @xmath20 and @xmath21 are depending on @xmath17 constants , then the following expressions can be easily obtained : @xmath22 here @xmath23 is the period of oscillations and @xmath24 is the complete elliptic integral of the first kind . in the limit @xmath25 ( [ 2.5 ] ) yields the separatrix @xmath26 phase plot of the system ( [ 2.1 ] ) as well as the potential energy ( [ 2.2 ] ) is presented in figure 1 .
the source of chaotic motion in our system is the homoclinic orbit on a phase plane . in the unperturbed case
this orbit is formed by coinciding stable and unstable manifolds of the saddle .
when there is a perturbation , the homoclinic orbit can be broken to yield the transverse intersection of stable and unstable manifolds , which gives rise to chaotic behaviour near the separatrix .
as far as one of the precursors of chaos is the appearance of subharmonics , we ll start with the subharmonic melnikov function defined for the periodic orbits as @xmath27 where @xmath28 is the unperturbed vector field corresponding to the solution @xmath29 with the period @xmath30 for @xmath31 , @xmath32 relatively prime integers , @xmath33 is the small vector of perturbation , and the wedge product is defined by @xmath34 . in considering our problem
we substitute the term @xmath35 for @xmath36 , and the term of perturbation @xmath37 for @xmath33 into the subharmonic melnikov function , therefore obtaining : @xmath38 calculations with the help of fourier representation of elliptic functions @xcite finally give : @xmath39 where @xmath40},\mbox { } k'(m)=k(1-m).\ ] ] as it is clearly seen from ( [ 3.3 ] ) @xmath41 has simple zeros , what proves the presence of subharmonic bifurcations @xcite . to move further the homoclinic melnikov function @xmath42 can be introduced as a limit of @xmath41 when @xmath43 , @xmath44 , and @xmath45 : @xmath46dx}=\frac{12\pi\omega^{2}f_{0}}{\sinh{\pi\omega\sqrt{\frac{\beta}{v}}}}\sin{\omega\theta}.\ ] ] surely this function also has simple zeros on @xmath47 and therefore proves the existence of transverse homoclinic orbits .
the presence of these orbits means , via smale - birkhoff theorem , the appearance of a smale horseshoe in the vicinity of an unperturbed saddle and thereby homoclinic chaos in this region of a phase space @xcite .
we note here that unlike the situation with dissipation ( like in @xcite or @xcite ) , when melnikov theory resulted in some relation between the perturbation amplitude , frequency and the dissipation coefficient , in our case there are no such restrictions .
it should be also noted that although the presence of a stochastic layer near the separatrix can be stated for rather generic hamiltonian perturbation making system near - integrable @xcite , the transverse intersection of stable and unstable manifolds is still a question which should solved by melnikov theory in every particular case .
to study the chaotic motion near the separatrix we shall use the whisker mapping ( separatrix mapping ) @xcite . to derive it we note first that the change in the unperturbed energy is given by : @xmath48 where @xmath49 is the momentum .
since the perturbation is a periodic function of time , we may introduce the phase of perturbation : @xmath50 which is an equivalent of time . in order to obtain the desired separatrix mapping one
has to consider the discretized time scale @xmath51 , and the mapping will involve @xmath52 and @xmath53 as variables . the change of phase after one period of oscillations is equal to @xmath54 , while energy change @xmath55 .
now we can estimate the change in the unperturbed energy @xmath56 per period of motion in proximity separatrix at time @xmath51 : @xmath57dx}\approx\beta f_{0}\int_{-\infty}^{\infty}{z_{0}(x)\cos\left[\omega(x+x_{n})\right]dx}.\ ] ] here we have approximated @xmath56 by evaluating the integral ( [ 4.2 ] ) on the unperturbed separatrix in accordance with the standard procedure @xcite . as it is clearly seen from the comparison of ( [ 3.5 ] ) and ( [ 4.2 ] ) , @xmath58 .
expansions of @xmath59 and @xmath60 in the vicinity of @xmath61 give finally the following correlation between them : @xmath62 $ ] . using this approximate period we can write the separatrix mapping in the following form : @xmath63
if to measure the stretching of a small phase interval with the parameter @xmath64 @xcite , one can detect the border of chaotic motion as an appearance of a local instability in phase : @xmath65 .
for the mapping ( [ 4.4 ] ) this condition gives the following approximation for the width of the stochastic layer @xmath66 : @xmath67 typical form of this dependence is presented in the fig .
2 . in fig . 3 ,
4 we have plotted the phase plane of the mapping ( @xmath68 , @xmath69 ) for different values of the perturbation amplitude @xmath70 .
3(right ) shows how the stochastic layer arises around the unperturbed separatrix .
then the width of this layer increases together with @xmath70 , leading finally to a chaotic sea , as demonstrated in fig .
4(right ) .
in this section we present the poincar sections of ( [ 1.4 ] ) in the vicinity of the saddle @xmath71 , which show homoclinic crossings in this region .
the methods used were runge - kutta - fehlberg and adams - bashford - moulton ( `` predictor - corrector '' ) method .
local chaotic behaviour is detected by the numerical calculation of the dominant lyapunov exponent .
it was computed using the method due to @xcite .
the system of equations , which defines the value of this exponent , was integrated with the runge - kutta fourth - order algorithm , while the equation ( [ 1.4 ] ) for a fiducial trajectory was run with fourth - order symplectic scheme @xcite . in figures 5 - 8
the poincar sections are presented for the system ( [ 1.4 ] ) with the forcing amplitude @xmath72 and different values of the frequency : @xmath73 , @xmath74 , @xmath75 and @xmath76 .
the value of the largest lyapunov exponent in these cases is @xmath77 , @xmath78 , @xmath79 and @xmath80 respectively , and so we can state that we really have chaotic motion in a close vicinity of a separatrix , as it was predicted
. numerical calculation of the lyapunov exponent in the region of regular dynamics with initial conditions at @xmath81 yields @xmath82 , and therefore our results can be taken as significant .
in this work we have considered the stationary wave reduction of the korteweg - de vries equation under small hamiltonian perturbations . on the basis of melnikov theory
it was shown how chaos can occur in this situation via the appearance of subharmonics and a further homoclinic tangle .
theoretical predictions about chaotic behaviour are supported numerically by plotting poincar sections and calculation of the corresponding lyapunov exponents . in a near - separatrix region
the governing ode was transformed into a mapping what allowed one to find the width of the stochastic layer .
our future work will focus on the development of kinetic description of chaotic behaviour for the obtained mapping @xcite , and with the derivation of new methods for the symplectic integration of the kdv equation under hamiltonian perturbations following the framework of @xcite , where the same was done for sin - gordon system .
author would like to thank prof .
m. bestehorn for stimulating discussions , kai neuffer for help with numerics , and dr .
s. reich for the information about symplectic integration methods .
zheng , w.j .
yeh , o.g .
symko : _ phys .
lett . _ a * 140 * , 225 - 228 ( 1989 ) h.t . moon : _ phys .
lett . _ * 64 * , 412 - 414 ( 1990 ) r. grimshaw , x. tian : _ proc .
london _ a * 445 , * 1 - 21 ( 1994 ) q. cao , k. djideli , w.g .
price , e.h.twizell : _ physica _ d * 125 * , 201 - 221 ( 1999 ) t. kawahara , s. toh : _ on some properties of solutions to a nonlinear evolution equation inlcuding long - wavelength instability _ , in _ nonlinear wave motion _ * 43 * , ed .
a. jeffrey , longman , new york , 1989 k.b .
blyuss : _ rep .
phys . _ * 46 * , 47 - 54 ( 2000 ) m. abramowitz , i. a. stegun : _ handbook of mathematical functions _ , dover , new york , 1972 j. guckenheimer , p. holmes : _ nonlinear oscillations , dynamical systems , and bifurcations of vector fields _ , 3rd printing , springer , new york , 1990 g.m .
zaslavsky , r.z .
sagdeev , d.a .
usikov , a.a .
chernikov : _ weak chaos and quasi - regular patterns _ , cup , cambridge , 1991 m. latka , b.j . west : _ phys . rev .
_ e * 52 * , 3252 - 3255 ( 1995 ) s. habib , r.d .
ryne : _ phys .
lett . _ * 74 * , 70 - 73 ( 1995 ) e. forest , r.d .
ruth , _ physica _
d * 43 * , 105 - 117 ( 1990 ) x. lu , r. schmid : _ mathematics and computers in simulation _ * 50 * , 255 - 263 ( 1999 ) | stationary wave solutions of the perturbed korteweg - de vries equation are considered in the presence of external hamiltonian perturbations .
conditions of their chaotic behaviour are studied with the help of melnikov theory . for the homoclinic chaos poincar sections
are constructed to demonstrate the complicated behaviour , and lyapunov exponents are also numerically calculated .
+ keywords : korteweg - de vries equation , hamiltonian perturbations , melnikov theory , poincar sections .
* chaotic behaviour of the solutions to a perturbed korteweg - de vries equation * + + k.b .
blyuss + _ department of mathematics and statistics + university of surrey , gu2 5xh , guildford , uk _ + e - mail : [email protected] |
this paper is mainly concerned with the asymptotic limit of the following linear kinetic transport equation as @xmath0 , @xmath1 here , @xmath2 denotes the density of particles at time @xmath3 in the phase space @xmath4 , and @xmath5 is the macroscopic density , @xmath6 the velocity distribution @xmath7 is given . for the sake of clarity
, we focus here on the case of the gaussian distribution with variance @xmath8 , @xmath9 our methodology can be applied to a wide range of distributions , as discussed at the end of this introduction . the time renormalizing factor @xmath10 in front of the bgk velocity operator in is chosen so as to obtain a nontrivial limit in the asymptotic regime of large deviations , _
i.e. _ in order to capture the vanishing exponential tails of the density @xmath11 .
we emphasize that the classical diffusive limit would involve an additional factor @xmath8 in front of the time derivative @xmath12 .
[ [ scaling . ] ] scaling .
+ + + + + + + + starting from the dimensionalized kinetic transport equation @xmath13 where @xmath14 is the rate of reorientation , and @xmath15 is the variance of the velocity distribution , we can always reduce to @xmath16 after the change of variables @xmath17 , @xmath18 , and @xmath19 .
then , we are interested in the asymptotics @xmath20 with the appropriate scaling .
it appears that the correct scaling in order to capture the small exponential tails of the distribution is such that @xmath21 and @xmath22 .
accordingly , we set @xmath23 for some ( small ) parameter @xmath24 that drives the large scale asymptotics when it vanishes .
we recover in the new variables @xmath25 .
note that the appropriate rescaling is specific to the choice of a gaussian velocity distribution .
[ [ diffusive - limit . ] ] diffusive limit .
+ + + + + + + + + + + + + + + + to draw an instructive parallel , let us consider the diffusive limit of , namely the heat equation ( see @xcite and the references therein ) @xmath26 when the effective diffusion coefficient @xmath27 is small ( say @xmath28 ) or , equivalently , when investigating the large scale asymptotics @xmath29 for small @xmath8 , one deals with the same equation with vanishing viscosity , @xmath30 it is well - known that , under appropriate conditions , @xmath31 converges uniformly locally towards a viscosity solution of the following hamilton - jacobi equation @xmath32 our main purpose here is to obtain a similar result for the kinetic transport equation .
[ [ large - deviations . ] ] large deviations .
+ + + + + + + + + + + + + + + + + our work can be viewed as a preliminary contribution to the theory of large deviations for simple velocity jump processes .
we follow the lines of evans and ishii @xcite ( see also @xcite ) , in which pde techniques were successfully applied to reformulate the ideas of fleming @xcite in the context of viscosity solutions .
fleming s work was devoted to applying the logarithmic transformation , and ideas from stochastic control , to the freidlin - wentzell theory for stochastic differential equations with a small noise .
[ [ logarithmic - transformation . ] ] logarithmic transformation . + + + + + + + + + + + + + + + + + + + + + + + + + + + as in the diffusive case , we perform the following hopf - cole transformation , @xmath33 the function @xmath34 satisfies @xmath35 [ [ limit - system . ] ] limit system .
+ + + + + + + + + + + + + our first result can be stated informally as follows : under suitable conditions ( see theorem [ hjlimit ] below ) , @xmath34 converges locally uniformly towards a viscosity solution of the following non - local hamilton - jacobi equation , @xmath36 where we have used the following notation , @xmath37 to avoid possible boundary layers at @xmath38 as @xmath0 , we assume that the initial condition for @xmath11 is of the form @xmath39 .
this set @xmath40 as the initial condition for and for the limit problem .
it is worth making some comments concerning the structure of the system .
first of all , it is not a standard hamilton - jacobi equation as , and the one obtained in the case of bounded velocities @xcite .
moreover , we notice that the first equation of does not contain enough information due to the apparition of @xmath41 for which extra dynamics are required .
although it seems somehow sparse , the two additional ( in)equations @xmath42 are sufficient to determine a unique solution of the cauchy problem , as stated in the comparison principle below ( theorem [ theo : comp ] ) . to the best of our knowledge ,
system is of a new kind .
we refer to it as a hamilton - jacobi problem by analogy with which was obtained via a similar procedure .
moreover , in the case of a compactly supported velocity distribution @xmath43 , the same procedure leads to a standard hamilton - jacobi problem @xcite . we should mention that non - local hamilton - jacobi equations of a very different type have been studied in the context of dislocations by g. barles , p. cardaliaguet
, o. ley , r. monneau , and a. monteillet in a series of papers , see _
e.g. _ @xcite and the references therein .
[ [ viscosity - solutions . ] ] viscosity solutions .
+ + + + + + + + + + + + + + + + + + + + equation should be read as a coupled system of hamilton - jacobi equations on @xmath44 and @xmath45 .
accordingly , we define viscosity solutions of using a couple of test functions @xcite . [ def : subsol ]
let @xmath46 .
a upper semi - continuous function @xmath47 is a * viscosity sub - solution * of on @xmath48 if and only if : + _ ( i ) _ @xmath49 . + _
( ii ) _ it satisfies the constraint @xmath50 + _ ( iii ) _ for all pair of test functions @xmath51 , if @xmath52 is such that both @xmath53 and @xmath54 have a local maximum at @xmath55 , then @xmath56 [ def : supersol ] let @xmath46 .
a lower semi - continuous function @xmath57 is a * viscosity super - solution * of on @xmath48 if and only if : + _
( i ) _ @xmath58 . + _
( ii ) _ for all @xmath59 , @xmath60 is a global minimum of @xmath61 .
moreover , @xmath60 is locally uniformly isolated : for any compact set @xmath62 , there exists @xmath63 such that @xmath64 for all @xmath65 . + _ ( iii ) _ for all pair of test functions @xmath51 , if @xmath52 is such that both @xmath66 and @xmath67 have a local minimum at @xmath68 , then @xmath69 let mention that the mimimality ( _ resp . _
maximality ) in the definition of the super- ( _ resp .
_ sub- ) solution arises with respect to variables @xmath70 only .
this is consistent with the fact that there is no derivative in the velocity variable in .
[ def : sol ] let @xmath46 .
a function @xmath44 is a * viscosity solution * of on @xmath48 if its upper ( _ resp . _
lower ) semi - continuous envelope is a sub- ( _ resp . _
super- ) solution in the sense of definitions [ def : subsol ] and [ def : supersol ] above .
[ [ convergence - and - uniqueness - for - the - limit - system . ] ] convergence and uniqueness for the limit system .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the following theorem states a comparison principle for viscosity ( sub / super-)solutions of the system .
this establishes uniqueness of viscosity solutions as a corollary .
the proof is contained in section [ sec : comp ] .
[ theo : comp ] let @xmath47 ( resp .
@xmath57 ) be a viscosity sub - solution ( resp .
super - solution ) of on @xmath71 .
assume that @xmath47 and @xmath57 are such that @xmath72 then @xmath73 on @xmath48 .
this result is extended for sub- and super- solutions with quadratic spatial growth at infinity , in section [ sec:621 ] .
this growth condition is compatible with the fundamental solution of the limit system , as we shall see below . in section [ sec : conv ]
, we prove the convergence of the sequence @xmath74 as @xmath0 , towards the unique viscosity solution of .
[ hjlimit ] assume that the initial condition @xmath75 satisfies the following properties : @xmath76 } \quad & b_0 = { u}_0 - \frac{\vert v\vert^2}{2 } \in w^{1,\infty}({\mathbb{r}}^{2n})\,,\\ \label{eq : initial condition 2 } \textbf{[b ] } \quad & - \det\left ( \text{hess}_{v}({u}_0(x , v ) ) \right ) \neq 0 , \qquad d_{v}^{(3 ) } u_0 \in l_{\text{loc}}^{\infty}.\end{aligned}\ ] ] let @xmath77 be the solution of , with the initial data @xmath78
. then , @xmath77 converges locally uniformly towards @xmath44 , which is the unique viscosity solution of , as @xmath0 .
in particular , for all @xmath79 , @xmath60 is an isolated global minimum of @xmath80 . however , we were not able to extend this convergence result to unbounded initial data with respect to space variable .
we leave this issue for future work .
[ [ heuristics . ] ] heuristics .
+ + + + + + + + + + + it is useful to give some heuristics , in order to understand how the supplementary condition @xmath81 appears in the limit @xmath0 , with equality if @xmath82 is reduced to the singleton @xmath83 .
first , notice that the constraint @xmath84 immediately follows from if the left - hand - side is bounded with respect to @xmath8 .
as a consequence , the minimum of @xmath44 with respect to velocity is necessarily attained at @xmath85 .
then , integrating against the probability measure @xmath86 we obtain the following continuity equation , @xmath87 the probability measure @xmath88 is expected to concentrate on the minimum points of @xmath44 as @xmath0 .
let assume that we do have in some sense , @xmath89 where the weights satisfy @xmath90 .
we notice that the constraint at each @xmath91 is clearly unsaturated , in the sense that @xmath92 .
there , we expect to see the last contribution of vanish .
this would lead to @xmath93 for each such @xmath94 . plugging this into , and using ,
we obtain successively , @xmath95 as we have formally @xmath96 by the chain rule , we expect eventually that @xmath97 and even @xmath98 if @xmath99 , that is , somehow @xmath100 . obviously , all this reasoning is formal , but we shall make it rigorous in section [ sec : conv ] . [ [ comparison - with - the - case - of - bounded - velocities . ] ] comparison with the case of bounded velocities .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we extend previous results obtained in @xcite to the case of unbounded velocities .
the main difference lies of course in the velocity rescaling which is not admissible in the case of bounded velocities .
this has important consequences .
firstly , in the case of bounded velocities , the asymptotic limit @xmath101 does not depend on @xmath102 , due to some averaging process arising in the velocity variable .
secondly , the limit equation is a standard hamilton - jacobi equation , associated with an effective hamiltonian .
in fact , the hamilton - jacobi equation obtained in the limit reads as @xmath103 under appropriate integrability conditions , where the limit is taken as @xmath104 from above ( in order to preserve the positivity of the denominator ) .
caillerie has extended this result to the general case , without such integrability condition @xcite . ] .
the case of unbounded velocities is by far more subtle , since there is no averaging with respect to velocity .
high velocities play a prominent role in the dispersion process .
the case of bounded velocities is analog to large deviations estimates for slow - fast systems as in @xcite , and references therein . in our case ,
the role of the fast variable is played by velocity , whereas the space variable is the slow one . in this work ,
we follow the hamiltonian viewpoint , focusing on the value function @xmath44 , solution of the hamilton - jacobi equation .
there is a dual viewpoint , focusing on the trajectories of the underlying piecewise deterministic markov process ( pdmp ) .
we refer to @xcite for further reading .
[ fig : min intro ] [ [ the - fundamental - solution . ] ] the fundamental solution .
+ + + + + + + + + + + + + + + + + + + + + + + + + in sections [ sec : fund ] and [ sec : explicit ] , we compute explicitly the fundamental solution of the non - local hamilton - jacobi equation _ in space dimension one _ , for @xmath105 .
we follow a time discrete iteration scheme based on the duhamel formulation of . since the problem is not translation invariant with respect to velocity , it is necessary to compute the solution for all initial data of the form @xmath106 where @xmath107 we obtain the following expression for the fundamental solution , @xmath108 the minimum value with respect to velocity is given by @xmath109 the function @xmath110 is plotted in figure [ fig : min intro ] for successive values of time .
a striking feature is that the solution does not converge to zero as @xmath111 , as compared to the fundamental solution of the hamilton - jacobi equation associated with the heat equation .
[ [ comparison - with - the - heat - equation . ] ] comparison with the heat equation .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + at first glance , the asymptotics of the kinetic equation , is linked with the asymptotics of the heat equation with vanishing viscosity , as the latter is the limit of the former under appropriate diffusive rescaling , yet different from : @xmath112 . however , as shown in in figure [ fig : min intro ] , the two behaviours are radically different
. we may informally present our results as follows : the scaling @xmath113 leads to the heat equation with vanishing viscosity @xmath24 after a simple chapman - enskog expansion in order to preserve the diffusive scaling @xmath114 .
the fundamental solution of the associated hamilton - jacobi equation , associated with the initial data @xmath115 , is @xmath116 . in particular
, it converges to zero in long time , uniformly on compact intervals . on the other hand ,
we have computed the fundamental solution of the limit system , obtained after the more appropriate rescaling @xmath117 . for the sake of comparison ,
it is better to describe @xmath118 ( which corresponds to the macroscopic density @xmath119 via the hopf - cole transform ) .
surprisingly enough , it does not converge to zero , but to the function @xmath120 .
we interpret this as follows : at a larger scale than the standard hyperbolic rescaling @xmath113 ( both space and velocity are larger in ) , we get non trivial asymptotics , but the density @xmath11 remains uniformly exponentially small far from the origin , of the order @xmath121 . as compared to the heat equation , lacks scaling invariance ( with respect to velocity ) .
this is emphasized by the fact that the fundamental solution of does not have a single - line expression , see below .
furthermore , picking a high velocity from a gaussian distribution is a rare event , that can be completely reset at the next velocity jump .
this leads to the predominance of low velocities that slows down the dispersion .
this has to be compared with the fokker - planck equation having the same stationary velocity distribution , @xmath122 for which we expect that the large deviations potential in the scaling @xmath123 behaves like a solution of . technically speaking
, we expect that the limit function is independent of the velocity variable due to the additional drift that compactifies the velocity space .
this is under investigation by the first author .
[ [ accelerated - fronts - in - reaction - transport - equations . ] ] accelerated fronts in reaction - transport equations .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + as an application of this work , we investigate quantitatively front propagation in reaction - transport equations in section [ sec : acc ] .
we focus on with an additional monostable reaction term : @xmath124 this models a population of particles that change velocity at rate one , pick up a random new velocity following a gaussian distribution , and divide at rate @xmath125 .
moreover , new particles pick up their initial velocity from the same gaussian distribution .
this type of model has been studied in @xcite in the case of bounded velocities , and in @xcite in the case of possible unbounded velocities . in the case of bounded velocities , there exist traveling waves with constant speed @xcite . moreover ,
any solution to the cauchy problem with sufficiently decaying initial data spreads with the minimal speed .
the classical fisher - kpp equation is obtained in the diffusive limit , provided that the rate of division is assumed to be of order @xmath126 , @xmath127 it is well known that front propagation occurs at constant speed , here @xmath128 @xcite , for any suitably decaying initial data ( otherwise , see @xcite ) . beyond this classical problem
, there has been some recent interest for describing accelerating fronts in reaction - diffusion equations .
garnier has investigated integro - differential equations , where the spreading operator is given by the convolution with a fat - tailed kernel @xcite , @xmath129 here , fat - tailed means that the kernel @xmath130 decays slower than exponentially .
there , the level lines of the solution spread super linearly , depending on the decay of the convolution kernel @xmath130 .
cabr and roquejoffre have studied the fisher - kpp equation where the diffusion operator is replaced with a fractional diffusion operator @xcite , @xmath131 for some exponent @xmath132 .
they describe quantitatively the acceleration of the front , which occurs at exponential rate , namely @xmath133 in a weak sense .
this seminal work was continued by coulon and roquejoffre in @xcite .
more recently , mirrahimi and mlard have described the limit of the reaction - diffusion with a fractional diffusion operator in dimension @xmath134 , after suitable rescaling and the hopf - cole transform @xcite . as opposed to our results
, the limit function has the following simple expression , with separation of variables : @xmath135 .
recently , spreading in the so - called cane toads equation has been studied intensively .
@xmath136 when the variable @xmath137 is unbounded , accelerated propagation has been proved independently by berestycki , mouhot and raoul @xcite , and by the first author , henderson and ryzhik @xcite .
there is a formal analogy between , and our problem , or rather .
indeed , acceleration also happens due to the influence of the microscopic variable @xmath137 , which plays a similar role as the velocity variable in this paper .
this is another example of a nonlinear acceleration phenomena appearing in a structured model .
both equations and describe spatial jumps of particles , with a focus on the distribution of long - range jumps .
equation describes a velocity - jump process , with a focus on the distribution of high velocities , that correspond somehow to long - range spatial jumps .
it was established in @xcite that solutions to behave in the long - time asymptotics as accelerating fronts due to the ( rare ) occurence of high velocities that send particles far from the bulk .
furthermore , the location of the front is of the order of @xmath138 , in accordance with the scaling limit .
however , the exact location of the front was not determined accurately , but it was estimated as @xmath139 in a weak sense ( see ( * ? ? ?
* theorem 1.11 ) for details ) . from the knowledge of the quantitative scaling limit procedure in the kinetic dispersion operator ( sections 2 to 5 ) , we aim to apply the same procedure as in @xcite , respectively for the classical fisher - kpp equation , and for the fractional fisher - kpp equation .
however , we face several issues , two of them are still open . starting from , after appropriate rescaling , we prove that @xmath140 converges towards a non - local hamilton - jacobi equation , very similar to , complemented with the additional constraint , @xmath141 so far , the limit is rigorously proven for bounded data @xmath142 only
. however , this rules out compactly supported initial data @xmath143 . on the other hand , in order to estimate quantitatively the location of the front , as the boundary of the support of the limit function @xmath144 , some condition on the extremal trajectories is needed : the so - called freidlin condition @xcite .
we propose an extension of this condition in our context , taking into account the non local feature of the constraint .
we verify that this condition is verified , but only on a subset of @xmath145 .
this rules out an immediate application of our comparison principle .
we conjecture that the edge of the front is located around @xmath146 naturally , this lies between the two bounds obtained in @xcite . finally , we claim that our work can be extended in a straightforward way to include the velocity distribution @xmath147 for @xmath148 ( this last assumption appearing for convexity reasons ) .
the rate of expansion should be given in the limit @xmath0 by the following expression , @xmath149
in this section , we prove the comparison principle stated in theorem [ theo : comp ] . we perform a classical doubling of variables
however , much attention has to be paid to the velocity variable .
this is the main concern of this proof . in particular ,
the velocity variable is not doubled , which is consistent with the fact that there is no gradient with respect to velocity in the limit system .
however , due to the unboundedness of the velocity space , it is mandatory to confine the velocities in a way to get absolute extrema .
let us define , for @xmath150 , the minimum values with respect to the velocity variable : @xmath151 let @xmath152 , @xmath153 .
let @xmath154 to be suitably chosen below . since the limit system requires two test functions , we shall define @xmath155 notice that @xmath156 penalizes high velocities .
we denote @xmath157 . the assumption will give a confinement on the minima of @xmath158 as follows .
[ lem : minconf ] let @xmath159 . for all @xmath160 , for all @xmath150 , we have @xmath161 .
fix @xmath150 .
suppose that there exists @xmath162 such that @xmath163 .
then at @xmath70 , @xmath164 and this can not hold .
therefore we have @xmath161 .
let @xmath165 .
we can now introduce the following maximum values , @xmath166 these two quantities are clearly finite from .
[ lem : oomega ] we have @xmath167 .
let @xmath168 be a maximum point of @xmath169 .
let @xmath170 .
the following sequence of inequalities hold true , @xmath171 we now prove that for suitably chosen parameters @xmath172 , the supremum @xmath173 of @xmath156 is attained at @xmath38 . for that
, we distinguish between two cases : @xmath174 , and @xmath175 . * # case 1 : @xmath176 . * we denote by @xmath177 a maximum point of @xmath156 , such that @xmath178 .
this point exists thanks to the penalization in the three variables and the fact that @xmath156 is upper semi - continuous .
we emphasize that this point depends on the parameters @xmath179 and @xmath180 , but for legibility we omit this dependency .
we aim to prove that @xmath181 .
we argue by contradiction , and assume that @xmath182 . we know
show that @xmath94 necessarily satisfies _ a priori _ a bound independent of @xmath180 .
[ lem : velconf ] we have @xmath183 the evaluation @xmath184 gives @xmath185 we define some auxiliary function with twice the number of variables , except for the velocity , as follows @xmath186 let @xmath187 which realizes the maximum of @xmath188 .
[ lem : limit t0 ] the following limit holds true , @xmath189 the evaluation @xmath190 yields @xmath191 we deduce the following estimates , @xmath192 therefore , the sequence @xmath187 converges as @xmath0 , up to a subsequence .
a closer look shows that the limit is necessarily @xmath193 .
we denote by @xmath194 an accumulation point .
we have @xmath195 and @xmath196 . on the other hand , passing to the limit @xmath0 in the inequality @xmath197
, we deduce @xmath198 therefore , we have necessarily @xmath199 , and @xmath200 .
[ lem : constraint ] there exists @xmath201 such that for all @xmath202 , @xmath203 the evaluation @xmath197 gives the following piece of information , @xmath204 we rewrite as @xmath205 since @xmath47 is a subsolution and @xmath206 is verified for @xmath8 small enough , we get that @xmath207 as @xmath0 , by upper semi - continuity , the last contribution in the r.h.s .
of is such that @xmath208 therefore , there exists @xmath201 such that for all @xmath202 , @xmath209 since we have assumed in this case that @xmath210 .
we now use the test function @xmath211 associated to the supersolution @xmath57 at the point @xmath212 .
notice that the condition @xmath213 is verified for @xmath8 small enough . by using the definition [ def : supersol ] of a super - solution , this yields @xmath214 on the other hand , using the test function @xmath215 associated to the subsolution @xmath47 at the point @xmath216
, we obtain @xmath217 by using the definition [ def : subsol ] of a sub - solution . by substracting to , we obtain @xmath218 letting @xmath0 , this gives @xmath219 .
recall that from the estimates in the proof of lemma [ lem : limit t0 ] , we have that @xmath220 holds true for some constant @xmath221 , and for @xmath24 small enough .
recall also lemma [ lem : velconf ] .
we obtain @xmath222 where @xmath223 depends only on @xmath224 .
hence , the choice @xmath225 establishes a contradiction .
we conclude that @xmath181 in this case .
* # case 2 : @xmath226 .
* we denote by @xmath227 a maximum point of @xmath169 .
we aim to prove that @xmath228 .
we argue again by contradiction and we suppose that @xmath229 . the evaluation @xmath230 yields immediately the following bound : @xmath231 * # # case 2.1 : we first consider the case where @xmath232 . * we introduce the following auxiliary function @xmath233 let @xmath234 which realizes the maximum value of @xmath235 .
we can prove as in lemma [ lem : limit t0 ] , that the following limit holds true , @xmath236 we use the test function @xmath237 associated to the supersolution @xmath238 at the point @xmath239 .
notice that the condition @xmath240 is verified for @xmath8 small enough . in order to apply the second criterion in ,
it is required that the set @xmath241 is reduced to @xmath83 , provided @xmath8 is sufficiently small .
[ lem:26 ] there exists @xmath242 such that the set @xmath241 is reduced to @xmath83 , provided that @xmath243 .
we argue by contradiction .
assume that there exists a sequence @xmath244 such that @xmath245 contains some nonzero @xmath246 .
it is required in the definition [ def : supersol ] of a viscosity super - solution that @xmath60 is a locally uniformly isolated minimum over @xmath247 .
hence , there exists @xmath125 such that @xmath248 for all @xmath249 . up to extraction of a subsequence
, we can assume that @xmath250 converges to some nonzero @xmath251 .
the following inequalities are satisfied , @xmath252 there is some subtlety here , because the lower semi continuity of @xmath238 would only imply @xmath253 in full generality .
however , the following argument establishes that such a @xmath254 is in fact a true limit . from the maximality of @xmath255
, we have @xmath256 now using the upper semi - continuity of @xmath257 and the lower semi - continuity of @xmath238 , we deduce that @xmath258 as a consequence of the previous inequalities , all inequalities are equalities .
thus , the following limit is well - defined , @xmath259 we conclude from that @xmath260 this yields the existence of @xmath261 such that @xmath262 , which is a contradiction . the second criterion in writes as follows , @xmath263 on the other hand , using the test function @xmath264 associated to the subsolution @xmath257 at the point @xmath265
, we obtain @xmath266 by substracting to , we obtain @xmath267 passing to the limit @xmath0 , we get a contradiction . *
# # case 2.2 : there exists some nonzero @xmath268 .
* the following lemma describes the equality cases in lemma [ lem : oomega ] .
[ lem : oomega eq ] assume @xmath269 , and let @xmath270 . then , 1 .
@xmath271 realizes the supremum of @xmath156 , 2 .
@xmath94 is a minimum velocity also for @xmath272 .
we repeat the proof of lemma [ lem : oomega ] with @xmath273 . by examining the case of equality
, we realize that @xmath274 similarly as in case @xmath275 , we define some auxiliary function with twice the number of variables , except for the velocity , as follows @xmath276 let @xmath187 which realizes the maximum value of @xmath188 . we can prove as in lemma [ lem : limit t0 ] , that the following limit holds true , @xmath277 one key observation is that the following strict inequality holds true @xmath278 by the very definition of @xmath279 . as @xmath280 as @xmath0
, we expect that this inequality is also strict for @xmath8 small enough .
however , this is not compatible with the _ a priori _ lower semi - continuity of @xmath57 as it may exhibit some negative jump when passing to the limit @xmath281 .
the following lemma resolves this difficulty ( see also the proof of lemma [ lem:26 ] ) .
[ lem : constraint2 ] there exists @xmath201 such that for all @xmath202 , @xmath203 from the maximality of @xmath282 , we have @xmath283 using the upper semi - continuity of @xmath47 and the lower semi - continuity of @xmath57 , we deduce that @xmath284 as a consequence of the previous inequalities , all inequalities are equalities .
thus , the following limit is well - defined , @xmath285 now define @xmath286 and take @xmath8 sufficiently small such that both inequalities hold true , @xmath287 the former is a consequence of , while the latter is deduced from lower semi - continuity of @xmath238 .
then , one has @xmath288 and the lemma follows . therefore , we can use the test function @xmath289 associated to the supersolution @xmath57 at the point @xmath212 .
notice that the condition @xmath213 is verified for @xmath8 small enough .
this yields @xmath290 on the other hand , using the test function @xmath291 associated to the subsolution @xmath47 at the point @xmath216 , we obtain @xmath292 by substracting to , we obtain @xmath293 passing to the limit @xmath0 , we get @xmath294 similarly as in case 1 , the choice @xmath295 yields a contradiction .
we conclude that @xmath181 in this case also .
we are in position to conclude . by using that @xmath296
, we deduce that @xmath297 , that is @xmath298 we obtain the comparison by taking limits @xmath299 , and @xmath300 .
in this section , we shall prove that @xmath77 converges locally uniformly towards the unique viscosity solution of the system .
we first establish lipschitz bounds that allow to extract a locally uniformly converging subsequence .
then , we prove that this limit is a solution of . * ( uniform estimates).*[estimates ] let @xmath301 be a solution of equation and define @xmath302 as @xmath303
. if @xmath304 satisfies condition , then @xmath302 is uniformly locally lipschitz .
precisely the following _ a priori _ bounds hold for all @xmath305 : @xmath306 the function @xmath302 satisfies @xmath307 let us notice that we obtain a unique solution @xmath308 , and thus a unique @xmath77 , from a fixed point method on the duhamel formulation of : @xmath309 [ [ proof - of - the - bound - i . ] ] proof of the bound @xmath310 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we define , for @xmath311 to be chosen later : @xmath312 for any @xmath313 , @xmath314 attains a maximum at point @xmath315 .
suppose that @xmath316 .
then , we have @xmath317 as a consequence , we have at the maximum point @xmath315 : @xmath318 moreover , the maximal character of @xmath315 also implies @xmath319 it yields that @xmath320 and @xmath321 . introducing these last inequalities in yields @xmath322 and thus
@xmath323 one can choose @xmath223 such that , for sufficiently small @xmath180 , this last inequality is impossible since the r.h.s is @xmath324 when @xmath325 : @xmath326 as a consequence @xmath327 , and we have , @xmath328 passing to the limit @xmath325 , we obtain @xmath329 .
one can carry out the same argument with @xmath330 to get the lower bound .
[ [ proof - of - the - bound - ii . ] ] proof of the bound @xmath331 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + to find lipschitz bounds , we use the same ideas on the difference @xmath332 .
the equation for @xmath333 reads as follows , @xmath334 using the same argument as above with the correction function @xmath335 , we end up with @xmath336 using the @xmath337 bound on @xmath302 , we find @xmath338 taking again @xmath223 suitably , we get that the previous inequality is impossible for small @xmath180 . as a consequence , @xmath339 \times { \mathbb{r}}^{2n } , \qquad b_h^{{\varepsilon}}(t , x , v ) \leq \sup_{(x , v ) \in
{ \mathbb{r}}^{2n}}\left ( b_0(x+h , v ) - b_0(x , v ) \right)\ ] ] the same argument will apply to @xmath340 .
indeed , one can write an equation on @xmath340 on the following form , @xmath341 such the maximum principle works since the r.h.s has the right sign when @xmath340 attains a maximum . finally , @xmath339 \times { \mathbb{r}}^{2n } , \qquad \vert b_h^{{\varepsilon}}(t , x , v ) \vert \leq \sup_{(x , v ) \in
{ \mathbb{r}}^{2n}}\left \vert b_0(x+h , v ) - b_0(x , v ) \right \vert \leq \left \vert \nabla_x b_0 \right\vert_{\infty } \vert h \vert.\ ] ] from which the estimate follows .
[ [ proof - of - the - bound - iii . ] ] proof of the bound @xmath342 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath343 and @xmath344 . to obtain the local bound on the time derivative on @xmath345 ,
let us differentiate with respect to time , @xmath346 we shall multiply by @xmath347 and split the r.h.s into two parts : @xmath348 and estimate both parts separately .
first , re - using the equation on @xmath302 and the estimate on the space derivative , we get @xmath349 e^{\frac{2\vert b^{\varepsilon}\vert_\infty}{{\varepsilon } } } dv ' : = c({\varepsilon}),\end{gathered}\ ] ] where @xmath350 is uniformly bounded ( and vanishes when @xmath0 ) since the ball @xmath351 contains the centered ball of radius @xmath352 .
assume now that @xmath353 has a positive local maximum at @xmath354 . then at this point
@xmath355 we thus deduce that @xmath356 , and thus @xmath339 \times { \mathbb{r}}^n \times { \mathcal{b}_r } , \qquad \vert \partial_t b^{\varepsilon}\vert \leq c({\varepsilon})t + \sup_{(x , v ) \in { \mathbb{r}}^n \times { \mathcal{b}_r } } \vert \partial_t b^{\varepsilon}\vert(0,x , v ) \,.\ ] ] we now recall to compute @xmath357 . indeed , we have @xmath358 the laplace method gives that @xmath359 is uniformly bounded in @xmath8 , locally in @xmath360 as soon as any minimum point in velocity @xmath102 of @xmath75 is non degenerate ( in the sense that @xmath361 ) and that the third order derivatives in velocities are locally uniformly bounded , which is exactly hypothesis @xmath362 $ ] .
as a consequence , we may bound @xmath363 by a uniform constant in @xmath8 , locally uniformly in @xmath364 .
[ [ proof - of - the - bound - iv . ] ] proof of the bound @xmath365 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + to obtain regularity in the velocity variable , we differentiate with respect to @xmath102 , @xmath366 where @xmath367 .
multiplying by @xmath368 , we obtain @xmath369 from which we deduce @xmath370 thanks to proposition [ estimates ] , the sequence @xmath77 converges locally uniformly towards @xmath371 .
we first start with a lemma that shows how the supplementary equation giving the evolution of the minimum in velocity in the system appears in the limit @xmath0 .
[ lem : convmin ] let @xmath372 .
one has , locally uniformly in @xmath373 , @xmath374 we shall find the limit of @xmath375 thanks to the local uniform convergence of @xmath77 , we know that @xmath376 converges locally uniformly to @xmath377 .
we shall prove that the remaining part of the r.h.s converges to @xmath378 .
for this purpose , we will prove the following estimate : @xmath379 for some @xmath380 suitably chosen .
we start with the l.h.s .. for @xmath381 ( possibly depending on @xmath8 at the end ) , we write @xmath382 where @xmath383 .
thanks to the lipschitz bound on the derivatives in @xmath102 , @xmath384 has non empty interior .
indeed , for all @xmath385 , for all @xmath386 , one has @xmath387 next , as @xmath386 , one has @xmath388 , and thus @xmath389 . hence , @xmath390 , we deduce @xmath391 take @xmath392 such that @xmath393 .
the previous inequality gives that @xmath394 coming back to , we get @xmath395 which gives the first bound that we want to prove .
let us come to the second part , namely the r.h.s . of . for all @xmath396
, one has @xmath397 we write @xmath398 for legibility .
we deduce @xmath399 as a consequence , @xmath400 ( where @xmath401 denotes the volume of the unit ball in @xmath402 ) . gathering the inequalities
, we find @xmath403 which gives the conclusion
. we shall start by the following lemma : [ lem : convmeas ] for @xmath404 , @xmath405 , consider the probability measure on @xmath402 : @xmath406 assume that @xmath407 .
then for any @xmath381 , one has @xmath408
let us rewrite @xmath409 one has : @xmath410 where @xmath411 comes from .
we estimate again @xmath412 the estimation of the remaining part will use the condition on the minimum points . since @xmath413 only achieves its minimum at @xmath60 , since @xmath34 converges locally uniformly towards @xmath413 and the minimas of @xmath34
are uniformly localized , we have necessarily some @xmath414 such that @xmath415 thus , @xmath416 gathering everything , the result follows . to prove the theorem [ hjlimit ]
, we need to go through the two following steps .
* # step 1 : viscosity sub - solution . *
let @xmath417 .
one wants to prove that @xmath371 is a viscosity sub - solution of on @xmath418 , in the sense of definition [ def : subsol ] .
let @xmath419 , such that @xmath420 and @xmath421 have a local maximum in the @xmath70 variables at the point @xmath422 , with @xmath423 .
we shall prove that the following conditions are fulfiled at @xmath52 : a. [ ( i ) ] @xmath424 , b. [ ( ii ) ] @xmath425 , c. [ ( iii ) ] @xmath426 .
inequality comes from the fact that one can pass to the limit in the viscosity sense in the following inequality @xmath427 to prove the constraint , assume by contradiction that @xmath428 thanks to the local uniform convergence of @xmath77 , for @xmath8 sufficiently small , @xmath429 since @xmath430 is locally bounded uniformly in @xmath8 , there exists a uniform interval @xmath431 centered at a minimal point such that @xmath432 forgetting the dependence in @xmath70 in order to shorten the notations , we deduce that at the point @xmath52 , @xmath433 which contradicts the estimates of proposition [ estimates ] since @xmath434 is bounded with respect to @xmath8 .
it remains to prove inequality .
this is the main originality in this procedure .
let @xmath313 be any positive number .
let us fix @xmath435 .
for all @xmath385 , let us integrate over @xmath436 to find : @xmath437 one can write , using @xmath438 : @xmath439 as a consequence , introducing the probability measure @xmath440 , we obtain : @xmath441 we conclude by investigating the following integral @xmath442 we can now pass to the limit in the viscosity sense with lemma [ lem : convmin ] , to conclude that there exists @xmath221 such that for all @xmath313 , @xmath443 we now recall the crucial fact that for any @xmath444 , any sub - solution achieves its minimum in velocity at @xmath60 , since @xmath445 . as a consequence , for any @xmath313 , for any @xmath444 , @xmath446 finally , we can pass to the limit @xmath325 , and get @xmath447 in the viscosity sense .
* # step 2 : viscosity super - solution .
* we now need to prove that @xmath371 is a super solution of in the viscosity sense @xmath448 to prove the first line of , let us take @xmath449 such that @xmath450 .
thus , thanks to the locally uniform convergence and the confinement of the minima of @xmath34 , for @xmath8 sufficiently small , one has @xmath451 .
thus , recalling @xmath452 , we have @xmath453 next , assume that @xmath454 , that is , @xmath378 is the only minimum of @xmath455 . for all @xmath385 , let us integrate over @xmath402 to find : @xmath456 for any fixed @xmath313 , we now split the last r.h.s into two parts : @xmath457 we estimate the first part exactly as in the sub - solution case : @xmath458 to estimate the second part , we use lemma [ lem : convmeas ] : @xmath459 as a consequence , passing to the limit in @xmath8 and then in @xmath180 , and recalling again lemma [ lem : convmin ] , we get @xmath460 in the viscosity sense .
it remains to check that @xmath85 is an isolated minimum for @xmath3 .
this is the purpose of the final lemma , [ lem : isolated ] for all @xmath3 and @xmath461 , @xmath85 is an isolated minimum point of the function @xmath80 .
we introduce the minimal value function @xmath462 .
we shall prove that for all @xmath70 such that @xmath3 , there exists a neighbourhood @xmath463 of @xmath378 , such that @xmath464 this would immediately imply that @xmath378 is an isolated minimum point .
passing to the limit @xmath0 in the duhamel formula , @xmath465 we deduce that the functions @xmath44 , and @xmath466 satisfy the following identity , @xmath467 the claim is equivalent to proving the two following estimates : @xmath468)\quad m(t - s , x - sv ) + s + \dfrac{\vert v \vert^2}2\geq m(t , x ) + \dfrac{|v|^2}2 & ( \star\star ) \end{cases}\ ] ] the proof of @xmath469 goes as follows .
first recall that @xmath466 is time decreasing .
thus , it is sufficient to establish that @xmath470 for all @xmath471 and small @xmath102 .
this holds for @xmath85 . on the other hand , the uniform spatial lipschitz estimate in proposition [ estimates](ii )
guarantees that the minimal value @xmath466 is lipschitz continuous with respect to @xmath472 , with the same bound , _
i.e. _ @xmath473 .
we deduce that @xmath474 as a conclusion , @xmath469 holds true for @xmath475 .
the proof of @xmath476 is similar .
firstly , we have @xmath477 the latter is nonnegative , provided that @xmath478 the latter condition suitably defines the neighbourhood @xmath479 .
our aim in this section is to compute some particular solutions of the cauchy problem . for this purpose
we first present a recursive numerical scheme which is based on the representation of the solution as an iterated semi - group .
we propose a time - discretization of this formulation which enables to compute the solution .
we justify a posteriori that the solution that we exhibit via the discretization procedure is the solution of the cauchy problem .
let us focus for a little while on the spatially homogeneous problem : @xmath480 its solution is given by @xmath481 where @xmath482 is the fundamental solution of the system , starting from the measure @xmath483 . since in this case @xmath484 , one can compute this fundamental solution straightforwardly : @xmath485 after performing the hopf - cole transform @xmath486 in the scaling @xmath487 , we obtain with the fundamental solution in the hj framework , @xmath488 where @xmath489 if @xmath490 , @xmath378 if @xmath491 .
we thus have the following representation for the associated cauchy problem : @xmath492 we now come back to equation .
the duhamel formula for a positive @xmath24 writes @xmath493 we deduce that the functions @xmath494 , @xmath495 satisfy formally in the limit @xmath0 , @xmath496 the minimal value function @xmath497 satisfies a closed equation : @xmath498 the time discretization of this implicit semi - group formulation is not trivial .
we propose to use an explicit method for the last recursive contribution in .
we approximate the function @xmath497 by @xmath499 at time @xmath500 , where @xmath501 is a given time step .
our recursive formula for @xmath502 writes @xmath503 we consider the initial data @xmath505 .
this corresponds in fact to the initial condition @xmath506 which has been instantaneously projected in order to satisfy the constraint @xmath507 , see the spatially homogeneous case of section [ sec : homog ] for a discussion .
[ lem : minvalue ] the minimum value satisfies @xmath508 the formula is clearly true for @xmath509 . to proceed with the recursion at range @xmath510
, it is sufficient to establish the following identity : @xmath511 by the induction hypothesis , we have @xmath512 the key convexity inequality is simply the following , @xmath513 the second inequality is an equality if and only if @xmath514 and @xmath515 , or @xmath516 and @xmath85 .
we deduce that @xmath517 we deduce from the more explicit formula : @xmath518 with the convention that @xmath519 if @xmath516 .
therefore , the solution is an envelope of parabolas . for the sake of convenience , we prove directly that is the solution of the recursive identity .
it is sufficient to compute the minimum value for @xmath520 such that @xmath521 , @xmath522 with equality if and only if @xmath514 or @xmath516 .
let @xmath524 .
we consider the initial data @xmath525 .
this corresponds in fact to the initial condition @xmath526 which has been instantaneously projected in order to satisfy the constraint @xmath507 , as in the previous case .
let us compute first the primary evolution of the initial data , which is essential for the comprehension of the dynamics .
it amounts to compute the following minimum value : @xmath527 the first contribution is new in comparison with the case @xmath515 .
this corresponds to the deposition of mass at point @xmath528 , with an initial level of @xmath529 .
it is expected that this mass will create secondary parabolas , as the ones which are already located around @xmath530 .
therefore , the solution should consist of the envelope of parabolas centered at @xmath530 , and secondary parabolas centered at @xmath528 , having some delay with respect to the central parabolas . we infer from this reasoning the following lemma . let us introduce the set of indices @xmath531 defined by @xmath532 [ lem : expmin ] the solution of the recursive identity starting from the initial condition @xmath525 is given by the formula @xmath533 with the convention that @xmath534 if @xmath516 .
the formula coincides with the initial condition for @xmath509 , namely @xmath535 .
we inject into the recursive formula .
the first contribution gives : @xmath536 this corresponds to the two extreme cases in the formula , namely @xmath537 and @xmath538 . on the other hand ,
the second contribution when is plugged into is the following , @xmath539 with equality if and only if @xmath514 or @xmath516 .
it remains to verify that the rule of induction on the sets @xmath540 is compatible with this recursion .
we have already seen that the two extreme points @xmath541 and @xmath542 are taken into account . for @xmath516 in the preceding equality
, we obtain the set @xmath543 .
then , for all @xmath544 , we obtain @xmath545 from the choice @xmath514 , where @xmath546 , _ i.e. _ @xmath547 .
this gives the additional set of indices @xmath545 such that @xmath548 .
this completes the set @xmath549 .
we infer from the preceding recursive numerical scheme that the fundamental solution should write in the continuous setting as follows , @xmath550 for the sake of clarity , we notify the dependency of @xmath497 with respect to the parameter @xmath551 ( initial concentration in the velocity space ) .
it is the envelope of a two - parameters family of parabolas . to conclude
, we plug the expression into the formula for the full fundamental solution @xmath552 , @xmath553 we deduce from the knowledge of the fundamental solution , and the comparison principle theorem [ theo : comp ] a useful representation formula .
let @xmath554 be a bounded continuous function .
then , the viscosity solution of is given by the following formula : @xmath555 interestingly enough , by exchanging the minima , the minimum value has a similar representation : @xmath556 notice that is not written as an inf - convolution with respect to velocity .
this reflects the lack of invariance of the problem with respect to this variable .
in this section , we now stick to dimension 1 in space and velocity .
this is to make explicit computations without having a too big computational cost , which appears to increase when the dimension is greater .
moreover , this is reasonable for our final purpose which is the study of propagation fronts .
we can derive a more explicit formulation of the minimization problem .
this is done after a discussion of several possible cases . in particular , if @xmath515 , it is better to choose @xmath558 , and we have @xmath559 we now assume that @xmath560 .
we introduce some notations for the sake of convenience , @xmath561 ^ 2 , \ ;
0\leq s_1 + s_2 \leq t , \qquad j(s_1,s_2 ) = \dfrac{\vert x - s_1 w \vert^2}{2 s_2 ^ 2 } + s_1 + s_2 \ , .\ ] ] a. assume that the minimum is attained at an interior point @xmath562 , that is such that @xmath563 then , computing the gradient yields the first order condition and the possible points of minimum : @xmath564 therefore , under the conditions to be an interior point , which read equivalently , @xmath565 the first candidate for the minimum value is @xmath566 b. assume that the minimum is attained on the edge @xmath567 , @xmath568 $ ] .
recall that we have by convention @xmath569 therefore , under the condition @xmath570 , including the equality @xmath571 , the second candidate for the minimum value is @xmath572 we notice immediately that this candidate is always better than .
c. assume that the minimum is attained on the edge @xmath573 , @xmath574 $ ] .
we have from @xmath575 } j(0,s_2 ) = \min_{s_2\in [ 0,t ] } \left ( \dfrac{\vert x \vert^2}{2 s_2 ^ 2 } + s_2 \right ) = \begin{cases } \dfrac32 |x|^{2/3}\ , , & \text { if } |x|\leq t^{3/2}\smallskip\\ \dfrac{\vert x \vert^2}{2 t^2 } + t\ , , & \text { if } |x|\geq t^{3/2 } \end{cases}\label{eq : candidate 3}\ , , \ ] ] and the optimal @xmath576 is given by @xmath577 .
d. finally , assume that the minimum is attained on the edge @xmath578 , with @xmath5790,t[^2 $ ] .
then minimizing @xmath580 reduces to minimizing on @xmath5810,t[$ ] the following function @xmath582 if @xmath583
, then @xmath584 is constant and its value is always worse than .
we can now assume that @xmath585 since we are done if not .
we have @xmath586 the minimizing @xmath576 is @xmath587 , if and only if @xmath5880,t[$ ] , that is @xmath5890,t[$ ] .
the value of the minimum is @xmath590 in this case .
else , the minimum can not be achieved on the interior of the edge . in any case
, we notice that this choice is always worse than under the same restriction @xmath5890,t[$ ] .
therefore , only three possible candidates remain after this discussion , @xmath591 we plot in figure [ fig : fund ] the final result that we obtain after minimizing between these three candidates .
[ fig : fund ] starting from an initial data of the form @xmath526 , a particle arriving at @xmath364 at time @xmath593 follows the following trajectory in the phase plane @xmath594 in order to compute the fundamental kernel @xmath552 given by , we first fully determine : @xmath595 on the set @xmath596 ^ 3 \ ; : \ ; 0\leq s_1 + s_2 + s_3 \leq t \right\rbrace.$ ] [ prop : traj ] we have @xmath597 define @xmath598 .
the lagrangian trajectories are given in the phase space by : * if @xmath599 , then depending on the position on the @xmath70-plane ( see figure [ fig : fund ] ) , we have either 1 .
@xmath600 2 . @xmath601 * if @xmath602 , then depending on the position on the @xmath70-plane ( see figure [ fig : fund ] ) , we have either 1 .
@xmath603 2 .
@xmath604 in particular , it shows that to reach its final point , it is better for a particule to redistribute its velocity only once .
if @xmath102 and @xmath551 have the same sign , positive say , then @xmath605 we should now discuss the values of the three times @xmath606 that minimize the travelling energy .
we can make a bunch of reductions before computing the minimum .
we may assume that @xmath607 and @xmath608 , as well as @xmath609 . indeed , in the three opposite cases , the work is already done through the minimization of the functional @xmath130 above .
the values of the minimum are then respectively given by @xmath610 , @xmath557 and @xmath611 and the travel times are obtained through the minimization of the functional @xmath130 above .
a. assume that the minimum is attained at an interior point @xmath612 : @xmath613 then , computing the gradient yields the first order condition , @xmath614 + since we work with @xmath609 , there is no solution to the first order condition , so the minimizer can not be an interior point .
b. assume that the minimum is attained on the edge @xmath615 , @xmath616 $ ] .
we have by convention @xmath617 + the latter quantity is finite for values of @xmath618 ^ 2 $ ] and @xmath616 $ ] such that @xmath619 can be achieved .
the minimum of the quantity @xmath620 on this set is necessarily attained on the boundary of the triangle .
thus , * the minimizer @xmath621 is available only if @xmath622 $ ] , and then a candidate is @xmath623 .
* the minimizer @xmath624 is available only if @xmath625 $ ] , and then a candidate is @xmath626 .
* the minimizer @xmath627 is available only if @xmath628 $ ] , and then a candidate is @xmath593 .
+ we check that the latest possibility is always worse than the two others . as a consequence
, one has to take the minimum between the two remaining of these possibilities , * the minimizer @xmath621 is available only if @xmath622 $ ] , and then a candidate is @xmath623 , * the minimizer @xmath624 is available only if @xmath625 $ ] , and then a candidate is @xmath626 , + and there is no possible minimum else .
we see that the value of the minimum in these zones is nevertheless given by @xmath629 .
c. assume that the minimum is attained on the edge @xmath630 , @xmath631 $ ] .
we have @xmath632 and the optimal times are given through @xmath497 .
d. assume that the minimum is attained on the edge @xmath633 , @xmath634 $ ] .
we have @xmath635 and the optimal times are given through @xmath497 .
e. finally , assume that the minimum is attained in the interior of the edge @xmath636 .
we have to minimize for @xmath6370,t[^2 , \qquad 0 < s_1 + s_3 < t,\ ] ] the quantity @xmath638 the first order condition reads @xmath6390,t[^2 , \qquad 0 < s_1 + s_3
< t , \qquad x - s_1 w - s_3 v = 0 , \\ \qquad \longleftrightarrow \qquad \exists u
\in \ ; ] 0,t [ , \qquad \exists s_1 \in \ ; ] 0,u [ , x = u \left ( \frac{s_1}{u } w + \left ( 1 - \frac{s_1}{u } \right ) v \right ) , \\
\longleftrightarrow \qquad \exists
u \in \ ; ] 0,t [ , \qquad x \in u \cdot [ v , w].\end{gathered}\ ] ] then the corresponding value of the minimum is @xmath593 .
we observe that this possibility is always worse than @xmath610 or @xmath557 .
discussing all the different cases , we see that @xmath640 and that the optimal travel times and trajectories are given by the statement of proposition [ prop : traj ] . with the knowledge of the trajectories in hand , we can now simplify the expression of the fundamental kernel @xmath641 .
[ prop : kernel ] for all @xmath642 , one has @xmath643 except in the very special case when @xmath644 , for which @xmath645 we point out the discontinuity in the fundamental kernel due to the initial condition when @xmath560 . from proposition [ prop : traj ]
, recall that we have @xmath597 thus , @xmath646 we can now distinguish various cases , depending on the value of @xmath647 and @xmath648 . 1 . first , if @xmath649 , we have directly @xmath650 .
2 . now assuming that @xmath651
, one has @xmath652 1 . if @xmath653 , then @xmath654 2 .
now , if @xmath655 , then @xmath656 finally , @xmath657 unless @xmath658 for which @xmath659 .
[ fig : min ]
we consider a kinetic reaction - transport problem .
we focus on equation after rescaling @xcite : @xmath660 in this section , we extend the comparison principle ( theorem [ theo : comp ] ) , and the convergence of @xmath668 to a viscosity solution of the corresponding non - local hamilton - jacobi equation ( theorem [ hjlimit ] ) for bounded initial data .
we further extend theorem [ theo : comp ] for solutions having locally uniform quadratic growth both in space and velocity ( section [ sec:621 ] ) .
this class of solutions appears naturally in our problem ( section [ sec:622 ] ) .
however , we were not able to prove the convergence theorem [ hjlimit ] for unbounded initial data .
in particular , we were not able to adapt the proof of uniformly isolated minima ( lemma [ lem : isolated ] ) due to the lack of uniform lipschitz regularity in space .
then , we extend the representation formula to the case @xmath125 , but ignoring the quadratic nonlinearity in .
this nonlinear term yields a constraint in the limit problem .
usually , this additionnal constraint is handled by means of the so - called freidlin condition which asserts that the solution to the constrained problem can be obtained by truncating the representation formula of the unconstrained problem so that it fulfils the constraint @xmath669 .
the latter condition is an immediate consequence of .
we now define the non - local hamilton - jacobi system obtained in the limit @xmath0 .
a function @xmath44 is a solution of the limit system if , and only if it is both a sub and a super - solution as described in the following definitions .
1 . @xmath671 2 .
if @xmath672 , + _ ( i ) _ it satisfies the constraint @xmath673 + _ ( ii ) _ for all pair of test functions @xmath51 , if @xmath52 is such that both @xmath53 and @xmath54 have a local maximum at @xmath55 , then @xmath674 3 .
if @xmath675 then , for all @xmath676 , one has @xmath677 . 1 .
@xmath58 2 . for all @xmath676 ,
one has @xmath678 .
3 . for all @xmath59 , @xmath60 is a global minimum of @xmath61 .
moreover , @xmath60 is locally uniformly isolated : for any compact set @xmath62 , there exists @xmath63 such that @xmath64 for all @xmath65 .
4 . for all pair of test functions
@xmath51 , if @xmath52 is such that both @xmath66 and @xmath67 have a local minimum at @xmath679 , then @xmath680 we consider the linearized version of , ignoring the quadratic nonlinearity , @xmath681 we point out that and are equivalent after the following change of variables : let @xmath682 defined by @xmath683 , then @xmath682 solves . as a consequence , the corresponding hamilton - jacobi system for @xmath684
is written @xmath685 we define accordingly the fundamental solution @xmath686 by performing the same change of variables in .
we refer to it as the _ unconstrained problem _ , because the definition of viscosity sub - solution ( resp .
super - solution ) coincide with definition [ def : nlsub ] ( resp . definition [ def : nlsup ] ) except that the constraint @xmath687 is ignored , meaning that item 2 is ignored in definition [ def : nlsub ] ( resp . in definition [ def : nlsup ] ) .
although we were not able to prove that this truncation procedure is valid , this would have the following consequences on the spatial spreading rate of accelerating fronts for .
in fact , we can compute the exact location of the level sets associated with the minimal value of the fundamental solution @xmath686 . although we were not able to prove this conjecture , we discuss in the rest of this section several intermediate results supporting it .
first , in order to prove that the truncation procedure might yield a representation of the solution of equation ( [ eq : nlsys ] ) , one needs to check that this solution is unique .
we thus first show a comparison principle for bounded initial data in subsection [ subsec : bdd ] .
we then extend in subsection [ subsec : unbdd ] this comparison principle to unbounded initial data with an appropriate quadratic growth rate , this rate being satisfied by the fundamental solution after positive time .
we then discuss in subsection [ subsec : freidlin ] the validity of the truncation by checking a freidlin - type condition for equation ( [ eq : nlsys ] ) , although we are not able to prove that this condition yields the conjecture .
[ prop : compr ] let @xmath47 and @xmath57 be respectively a viscosity sub - solution and a viscosity super - solution of equation ( [ eq : nlsys ] ) in the sense of definitions [ def : nlsub ] and [ def : nlsup ] .
assume that @xmath47 and @xmath57 satisfy
. then @xmath73 .
we follow the same lines as for the proof of theorem [ theo : comp ] . for the sake of clarity
, we do not reproduce all the details , but we focus on the arguments specific to the constrained problem .
this is the discussion on the minimum @xmath696 .
* # case 1 : @xmath176 . *
if @xmath699 , then argue as in section [ sec : comp ] . on the other hand , if @xmath700 , then by definition we have @xmath701 .
but we also have @xmath702 .
therefore , we have a contradiction at @xmath703 .
[ thm : cv sec 6 ] let @xmath77 be the solution of , with the initial data @xmath709 .
we assume that the initial data satisfies conditions and .
then , as @xmath0 , @xmath77 converges towards @xmath44 , the unique viscosity solution of the limit system , characterized by definitions [ def : nlsub ] and [ def : nlsup ] .
we now show how to pass to the limit in the presence of the nonlinearity .
we again follow the same lines as for the linear case , and do not reproduce the technicalities to focus only on the main features of the argument .
we start with reproducing the lipschitz estimates .
then , we derive the viscosity procedure . the function @xmath302 satisfies : @xmath710 we shall rewrite this equation on the form : @xmath711 the fact that @xmath712 combined with a maximum principle argument for @xmath302 gives the uniform bound . for the space derivative ,
we write : @xmath713 e^{\frac{u^{\varepsilon}(v ) - u^{{\varepsilon}}(v')}{{\varepsilon } } } dv ' \\ - \frac{r}{{\varepsilon } } \int_{{\mathbb{r}}^n } \left [ \nabla_x u^{\varepsilon}(v)- \nabla_x u^{\varepsilon}(v')\right ] \left(m_{\varepsilon}(v ) e^{\frac{u^{\varepsilon}(v)}{{\varepsilon } } } - 1 \right ) e^{\frac{-u^{{\varepsilon}}(v')}{2}}dv ' - \frac{r}{{\varepsilon } } \left ( \int_{{\mathbb{r}}^n } e^{-\frac{u'^{\varepsilon}}{{\varepsilon } } } dv ' \right ) \nabla_x u^{\varepsilon},\end{gathered}\ ] ] and the result follows from a maximum principle argument . to bound the time derivative , we write the same type of equation as for
the space derivative and redo the argument used for the linear case .
the first inequality comes from the fact that @xmath716 and then the same argument as for the linear case works .
let us set @xmath717 , we get @xmath718 since the r.h.s is now mass conservative , the same procedure as for the linear case applied to @xmath719 gives @xmath720 and thus the result follows .
since @xmath725 is bounded uniformly , the same proof as in the linear case gives @xmath726 the maximum principle on @xmath11 yields that in the limit , one has necessarily @xmath727 .
we deduce @xmath728 take first a point @xmath70 such that @xmath729 .
the previous inequality gives @xmath730 .
take finally a point @xmath70 such that @xmath731 . then in this case @xmath732 goes to zero .
thus , passing to the limit in the viscosity sense in @xmath733 gives the first inequality of the limit system .
the proof in the linear case can be reproduced to give @xmath734 here , we explain how to extend the comparison principle section [ sec : comp ] beyond hypothesis , _ i.e. @xmath735_. specifically , we consider sub- and super- solutions having locally uniform quadratic growth at infinity both in space and velocity .
this is motivated by the fact that the fundamental solution that was derived in section [ sec : explicit ] has precisely this large scale behaviour after positive time ( see also figure [ fig : fund ] ) .
the method can not handle asymptotic behaviour bigger than quadratic .
[ theo : comp2 ] let @xmath47 ( resp .
@xmath57 ) be a viscosity sub - solution ( resp .
super - solution ) of on @xmath71 .
assume that @xmath47 and @xmath57 are such that there exist constants @xmath736 such that for all @xmath364 , @xmath737 then @xmath73 on @xmath48 .
we shall not reproduce all the details of proof of the comparison principle , but only indicate the major changes that have to be made in comparison with section [ sec : comp ] .
the major discrepancy concerns the localization of minima with respect to the velocity variable .
indeed , they are no more confined uniformly with respect to @xmath472 as in section [ sec : comp ] , see .
we have to adapt the penalization terms in @xmath169 and @xmath156 accordingly . a side effect
is that a short time condition @xmath739 is required .
however , this does not affect the conclusion since we can iterate the comparison principle on time intervals of length @xmath740 .
let us define @xmath741 let @xmath739 .
let @xmath152 , @xmath153 .
let @xmath154 to be suitably chosen below .
we define @xmath742 for some suitable constants @xmath743 to be chosen below .
the quadratic penalty term in has been turned into a quartic one in order to ensure the existence of a minimum with respect to space variable .
the exponential prefactor is inspired from @xcite .
also , the penalization with respect to velocity has been extended in order to take into account the nonuniform velocity confinement .
we define @xmath697 as in . to establish that @xmath744
, it is enough to check that the penalty term @xmath745 vanishes at the minima of @xmath57 with respect to velocity .
this is indeed a consequence of and the short time condition .
the following lemma replaces the confinement lemma [ lem : minconf ] .
we now present the test function step @xmath761 associated to the supersolution @xmath57 at the point @xmath212 .
notice that the condition @xmath213 is verified for @xmath8 small enough .
this yields @xmath214 we strongly emphasize a crucial cancellation .
namely , the free transport operator cancels the velocity penalization @xmath762 .
this is the main motivation for considering such an expression . on the other hand ,
using the test function @xmath763 associated to the subsolution @xmath47 at the point @xmath216 , we obtain @xmath764 by substracting to , we obtain @xmath765 we deduce from lemma [ lem : velconf a ] and @xmath766 that @xmath767 for @xmath8 small enough . now observe that @xmath768 for some explicit constant @xmath769 , provided that @xmath770 . by taking the limit @xmath0
, we deduce that @xmath771 the choice @xmath772 yields a contradiction . in this case ,
the spatial derivative of the transport equation is not used , however the time derivative of the test function being different , one has to perform slight changes .
the new function @xmath235 is @xmath773 the second criterion in writes as follows , @xmath775 on the other hand , using the test function @xmath776 associated to the subsolution @xmath257 at the point @xmath265 , we obtain @xmath777 by substracting to , we obtain @xmath778 passing to the limit @xmath0 , we get a contradiction .
similarly as in case @xmath275 , we define the following auxiliary function , @xmath779 let @xmath187 which realizes the maximum value of @xmath188 .
we can prove as in lemma [ lem : limit t0 a ] , that the following limit holds true , @xmath277 therefore , we can use the test function @xmath780 associated to the supersolution @xmath57 at the point @xmath212 .
notice that the condition @xmath213 is verified for @xmath8 small enough .
this yields @xmath781 on the other hand , using the test function @xmath782 associated to the subsolution @xmath47 at the point @xmath216 , we obtain @xmath783 we conclude as in case 1 , except that the confinement estimate on @xmath94 is given by rather than lemma [ lem : velconf a ] .
the proof of the comparison principle can be ended exactly as in section [ sec : comp ] .
the purpose of this section is to provide crude _ a priori _ bounds for both the fundamental solution in the case @xmath784 , and the solution of the cauchy problem for compactly supported initial data , in the case @xmath785 .
we show that in both cases @xmath786 is naturally bounded above by @xmath787 .
thus , it would be natural to extend the uniqueness result ( theorem [ theo : comp ] ) to such unbounded solutions , in order to prove that the solution of the cauchy problem coincides with the solution built from the representation formula . in this direction , we establish such a uniqueness result for solutions satisfying a uniform quadratic bound involving @xmath788 on @xmath789\times { \mathbb{r}}\times { \mathbb{r}}$ ] .
we deduce immediately from the expression that the fundamental solution @xmath552 satisfies @xmath790 for this , simply choose @xmath791 in the minimization problem .
note that a similar bound can be derived for @xmath686 when @xmath125 . for the sake of simplicity ,
we restrict to initial conditions compactly supported in @xmath472 , with a gaussian velocity distribution , @xmath792.\left [ { \varepsilon}^{1/2 } m_{\varepsilon}(v)\right]\ , , \ ] ] for some is some open interval @xmath793 .
first , we claim that we can restrict to the case without the nonlinear term ( @xmath797 ) , since the contribution @xmath798 is nonnegative . from the duhamel formula
, we deduce that @xmath11 is bounded below by the damped free transport equation , @xmath799 hence , we have @xmath800 up to elementary spatial rescaling , we assume that @xmath793 contains the interval @xmath801 . plugging this bound into the duhamel formula with @xmath797 [ eq : duhamel formula ] , we deduce @xmath802 this estimate from below , combined with yields the following estimate on @xmath34 , @xmath796 we now continue our discussion about the limit system .
this limit system appears to be an obstacle problem , subject to the constraint @xmath669 .
it is analogous to the following constrained hamilton - jacobi equation , @xmath803 which can be derived from the fisher - kpp equation using the same methodology as described above @xcite ( see also @xcite for earlier works using a different framework based on large deviations ) .
the so - called freidlin condition allows to solve this obstacle problem by the following truncation procedure : first , solve the unconstrained problem @xmath806 then , truncate the solution @xmath807 by taking into account the constraint @xmath808 : @xmath809 the solution @xmath807 can be written using the hopf - lax representation formula , @xmath810 ds + \tilde u_0(x(0 ) ) \ ; \big\vert \ ; x(t ) = x \right\rbrace\\ & = \inf_{x \in
\mathsf{h}_{loc}^1((0,t),{\mathbb{r}}^n ) } \left\lbrace \int_0^t \left [ \frac{\vert \dot x(s ) \vert^2}{2 } - r \right ] ds \ ; \big\vert \ ; x(0 ) \in \overline{g_0}\;,\ ; x(t ) = x \right\rbrace.\end{aligned}\ ] ] the freidlin condition states that , starting from a final point @xmath70 such that @xmath811 , then tracing backward a minimizing trajectory @xmath812 , it verifies @xmath813 for all @xmath814 .
this implies the following modification of the hopf - lax formula , @xmath815 ds \ ; \big\vert \ ; x(0 ) \in \overline{g_0}\;,\ ; x(t ) = x \ , , \ , ( \forall s)\ ; \tilde u(s , x(s ) ) > 0 \right\rbrace,\ ] ] alternatively speaking , trajectories ending in the unconstrained area ( _ i.e. _ @xmath816 ) have gone through the unconstrained area only .
the freidlin condition holds true for the unconstrained problem . in section [ sec : explicit ] , we derived explicit formulas for extremal trajectories .
let define the following set , @xmath822 an important observation is that extremal trajectories ending in @xmath823 remain in @xmath823 in backward time .
let @xmath824 .
clearly , @xmath825 is an optimal choice in .
we deduce from proposition [ prop : traj ] that an extremal trajectory joining @xmath826 and @xmath364 is necessarily given by one of the two following choices , the choice between the first case ( referred to as trajectory 1 ) , and the second case ( referred to as trajectory 2 ) is set up according to proposition [ prop : traj ] , see also figure [ fig : fund opt ] for a sketchy diagram .
we propose the following extension of the freidlin condition : let @xmath830 be an extremal trajectory such that the ending point @xmath831 belongs to @xmath823 . assume that @xmath25 is such that @xmath832 , then for all @xmath833 , @xmath834 we refer to it as a non - local freidlin condition , as the sign condition is required for all @xmath551 along the projected backward trajectory @xmath835 .
we prove below that this condition holds true .
let @xmath836 be such an extremal trajectory belonging to @xmath823 .
we denote by @xmath826 its starting point .
the condition @xmath832 is rewritten using @xmath837 and the requirement is : @xmath838 we can conclude , provided that the function @xmath839 has the following monotonicity for @xmath840 : increasing , then decreasing .
indeed , it is nonnegative at @xmath841 and at @xmath842
. therefore , it is positive for @xmath840 .
there is some subtlety here , because , the trajectory is associated with some parameter @xmath844 , which is generally different from @xmath551 .
it means that the typical trajectories shown in figure [ fig : fund opt ] are not necessarily in phase with background areas delimited by the plain red curves .
there is quite a number of cases to discuss .
however , we can overcome the complexity of the picture , by remarking that the trajectory always finish with a constant part , @xmath845 . on this part , the function @xmath846 is clearly decreasing .
to conclude , it remains to prove that the function @xmath847 is concave on the linear part of the trajectory .
let it be @xmath848 , for some positive speed @xmath849 .
on each of the zones @xmath850 , @xmath851 is concave : it is linear on @xmath852 and @xmath223 , and it is equal to @xmath853 on @xmath224 . in addition , there are two possible junctions : the one from @xmath852 to @xmath224 , and the one from @xmath223 to @xmath224 ( see figure [ fig : junction ] ) .
* junction from @xmath852 to @xmath224 ( figure [ fig : junction]a ) : * let @xmath854 be the time of transition from @xmath852 to @xmath224 .
it satisfies @xmath855 .
the slope at @xmath856 is equal to 1 .
the slope at @xmath857 is @xmath858 there is no slope discontinuity at this junction .
* junction from @xmath223 to @xmath224 ( figure [ fig : junction]b ) : * let @xmath854 be the time of transition from @xmath223 to @xmath224 .
it satisfies @xmath859 .
the slope at @xmath856 is @xmath860 the slope at @xmath857 is @xmath861 the slope is necessarily decreasing at this junction . we conclude that @xmath851 is globally concave on the linear part of the trajectory , wherever it goes through .
then , it is decreasing on the constant part of the trajectory .
consequently , @xmath851 has the required monotonicity .
g. barles , p. cardaliaguet , o. ley , and a. monteillet .
_ existence of weak solutions for general nonlocal hamilton - jacobi equations _
, nonlinear analysis : theory , methods & applications , 71(78):2801 - 2810 " , 2009 e. bouin , v. calvez , n. meunier , s. mirrahimi , b. perthame , g. raoul , r. voituriez , _ invasion fronts with variable motility : phenotype selection , spatial sorting and wave acceleration _ , c. r. math . acad .
paris , * 350 * ( 2012 ) , 761766 .
e. bouin , v. calvez , and g. nadin .
_ front propagation in a kinetic reaction - transport equation : travelling waves and accelerating fronts_. archive for rational mechanics and analysis , ( 2015 ) , to appear .
l. c. evans , h. hishii , _ a pde approach to some asymptotic problems concerning random differential equations with small noise intensities _ , annales de linstitut henri poincar ( c ) analyse non linaire ( * 2 * ) , ( 1985 ) .
kolmogorov , i.g .
petrovsky , n.s .
piskunov , _ etude de lquation de la diffusion avec croissance de la quantit de matire et son application un problme biologique _ , moskow univ .
math . bull .
* 1 * ( 1937 ) , 125 .
s. mlard and s. mirrahimi , _ singular limits for reaction - diffusion equations with fractional laplacian and local or non - local nonlinearity _ , communications in partial differential equations , vol .
40.5 , ( 2015 ) , pp .
957 - 993 .
p. e. souganidis .
front propagation : theory and applications . in _ viscosity solutions and applications ( montecatini terme , 1995 ) _ , volume 1660 of _ lecture notes in math .
_ , pages 186242 .
springer , berlin , 1997 . | we establish a large deviation theory for a velocity jump process , where new random velocities are picked at a constant rate from a gaussian distribution . the kolmogorov forward equation associated with this process
is a linear kinetic transport equation in which the bgk operator accounts for the changes in velocity .
we analyse its asymptotic limit after a suitable rescaling compatible with the wkb expansion .
this yields a new type of hamilton jacobi equation which is non local with respect to velocity variable .
we introduce a dedicated notion of viscosity solution for the limit problem , and we prove well - posedness in the viscosity sense .
the fundamental solution is explicitly computed , yielding quantitative estimates for the large deviations of the underlying velocity - jump process _
la freidlin - wentzell_. as an application of this theory , we conjecture exact rates of acceleration in some nonlinear kinetic reaction - transport equations . * key - words : * large deviations , piecewise deterministic markov processes , hamilton - jacobi equations , viscosity solutions , scaling limits , front acceleration . + * ams class .
no : * |
beta - minus decay and proton emission take a nucleus in almost opposite directions on a nuclear chart , so @xmath0-delayed proton emission ( where beta decay feeds excited states that subsequently emit a proton ) is forbidden in all but a few nuclei where it is heavily suppressed as the available energy is @xcite @xmath4 where @xmath5 is the neutron separation energy of the nucleus .
we describe here an experiment to detect this decay mode from the one - neutron halo nucleus @xmath1be that is believed to be the most favourable case @xcite due to the single - particle behaviour of halo nuclei @xcite that may favour this decay mode and due also to the relatively long halflife that is caused by the normal beta - decay of @xmath1be being hindered since a level inversion gives it a @xmath6 ground state rather than a @xmath7 .
beta - delayed particle emission is in general a prominent decay mode for nuclei close to the dripline , see @xcite for recent reviews .
the energetically open channels for @xmath1be are @xmath8 , @xmath9 t , @xmath9p and @xmath9n with corresponding @xmath10-values of @xcite @xmath11 kev , @xmath12 kev , @xmath13 kev and @xmath14 kev .
the low decay energy implies that the branching ratio for beta - delayed proton emission is low , typical estimates are slightly above @xmath15 @xcite . to detect the process experimentally ,
it is therefore essential to keep contaminants at a very low level .
the @xmath9p decay mode may be expected preferentially in one - neutron halo nuclei , partly due to the requirement of low neutron separation energy , partly due to the more pronounced single - particle behaviour of halo nuclei .
two - neutron halo nuclei are in a similar way candidates for beta - delayed deuteron emission , which has so far been observed only in the nuclei @xmath16he and @xmath1li @xcite .
for @xmath1li the decay has a branching ratio of order @xmath17 , the low value again caused by a small energy window , whereas cancellation effects reduces the branching ratio for @xmath16he down to the @xmath18 level
. it may be more useful to consider the standard measure for the strength of a decay , the reduced matrix element squared @xmath3 , that is found from the relation @xcite @xmath19 where @xmath20 is the beta - decay phase space , @xmath21 s and @xmath22 . converting the observed spectra for beta - delayed deuteron emission from the two - neutron halo nuclei
@xmath16he @xcite and @xmath1li @xcite gives total @xmath3 values within the observed energy range of about 0.0016 and 0.75 .
( note , however , that the @xmath16he decay to the @xmath16li ground state has been described as an effective di - neutron to deuteron decay , it is a super - allowed transition with a @xmath3 of 4.7 .
this may be a reflection of a general trend for super - allowed decays to occur in very neutron - rich nuclei @xcite . ) for comparison , the sum of @xmath3 for all currently known transitions in the @xmath1be decay is 0.27 .
the radioactive @xmath1be nuclei were produced at the isolde facility at cern .
searching for protons with a kinetic energy of a few hundred kev with relative intensity @xmath15 is challenging in a radioactive beam environment , so we instead detect the decay product , @xmath23be with a halflife of @xmath24 y , that exists only in minute quantities on earth . to reach the needed sensitivity we must employ state - of - the - art ams .
it is also crucial to limit the amount of contaminants in the samples so sample collection took place at isolde s high - resolution mass separator .
the resolution from the magnetic separation stage is supplemented by the electrostatic beam transport at isolde similar to , but at lower resolution than , the separation stages in ams facilities .
a first attempt was made in 2001 and the results were published recently @xcite .
the signal was not sufficiently strong to be clearly separated from background and gave a @xmath9p branching ratio of @xmath25 , significantly above the published theoretical expectations . due to improvements both in production of @xmath1be and ams detection of @xmath23be ,
the current collection was performed in december 2012 and resulted in three samples .
the @xmath1be activity was produced by bombarding a uc target with 1.4 gev protons .
the products were ionized in a laser ion source , which provided element selectivity , mass separated in the isolde high - resolution separator , and guided through several collimators to the collection point where they were implanted at 60 kev in a small copper plate ( 15@xmath2620@xmath262 mm ) .
a high - purity coaxial ge - detector placed 40 cm downstream behind a lead shielding monitored the collection rate .
the ge - detector was energy and efficiency calibrated with standard sources of @xmath27co , @xmath28eu and @xmath29th .
the main lines in the @xmath30 spectrum recorded during @xmath1be collection are the 2124 kev line from the decay of @xmath1be and the 511 kev line from positron annihilation . the overall efficiency at 2124 kev is found to be @xmath31 . a second line from the decay at 2895 kev was also used to check the overall amount of @xmath1be
the two determinations gave about the same precision , the one from the 2124 kev line being dominated by systematic uncertainties in the efficiency and the one from the 2895 kev being dominated by statistical uncertainties , and were internally consistent leading to a final value for the amount of collected @xmath1be in the main sample ( s1 ) of @xmath32 .
this includes a correction for dead time of 2.8% , determined from the ratio of accepted to total number of triggers . as cross - checks
two other samples were collected : sample s2 at the mass position of @xmath1li ( 0.02 mass units heavier than @xmath1be ) where an upper limit of @xmath33 could be determined for the number of atoms collected ( corresponding to a @xmath1li yield below 625/s which is reasonable ) and , for one second only , sample s3 at the @xmath23be mass position where an estimate of the current of 3.5 pa ( uncertain by a factor two ) converts into @xmath34 atoms . according to srim calculations @xcite
about 6% of all be ions implanted in cu at 60 kev energy will backscatter out of the sample .
most of the backscattered ions are expected to remain close to the sample so @xmath30-rays from their decays will be seen as well , although the decay products are not retained in the sample .
this gives a correction which we estimate to be @xmath35% .
the @xmath23be accelerator mass spectrometry ( ams ) measurements were performed at the vienna environmental research accelerator ( vera ) at the university of vienna .
vera is a dedicated ams facility based on a nec 3 mv pelletron tandem accelerator
. a new scheme for @xmath23be using a passive foil absorber in front of a gas ionization chamber detector was employed . in this way the detection efficiency for @xmath23be atoms is increased significantly . according to trim simulations @xcite the maximum implantation depth of @xmath1be in our copper plate catcher was below 1 @xmath36 m . to reduce the amount of material to be dissolved only the surface layer of each irradiated copper plate was leached in nitric acid .
a second leaching was performed to verify the blank level of the irradiated copper plate .
the second leaching of sample s3 did not produce enough beo for a measurement . for samples
s1 and s2 the values of the second leachings were consistent with a blank sample .
this shows that the material was sitting in the surface , as expected for an implanted sample , and not due to a bulk contamination .
an amount of 359 @xmath36 g ( uncertainty 3% ) @xmath37be carrier was added to the solution to reach a @xmath23be/@xmath37be isotopic ratio in the range of 10@xmath3810@xmath39 . in the next step
the solution was treated with ammonium hydroxide to precipitate the beryllium as beryllium hydroxide ( be(oh)@xmath40 ) .
the dissolved copper remains in the solution in this step .
the beryllium hydroxide was dried out by heating in an oven at 900@xmath41c for at least 8 hours forming beryllium oxide ( beo ) .
the beo powder was mixed 1:1 with high purity copper powder and pressed into sample holders and mounted together with standard and blank material in a mc - snics type cesium sputter ion source .
blank is the pure phenakite material directly pressed into a sample holder . a separate sample , s - blank , went through the chemistry preparation to check for the amount of @xmath23be introduced during the chemical sample preparation .
beo@xmath42 was extracted from the ion source and stripped in the terminal of the tandem accelerator to be@xmath43 , resulting in a total ion energy of 2.4 mev .
after further mass separation by a sector magnetic analyzer and an electrostatic analyzer the remaining particles are sent to a gas ionization chamber detector with a two - split anode for particle identification .
a silicon nitride foil stack as a passive absorber was installed in front of the detector .
this foil stack prevents the isobaric background @xmath23b from entering into the detector : the energy loss of boron in the foil stack is slightly larger compared to beryllium . by selecting the right foil thickness and carefully tuning the particle energy the boron ions are stopped in the foil stack whereas the beryllium ions can enter the detector .
.results of the ams measurement .
s1 to s3 denote the irradiated samples .
1st and 2nd correspond to the first or second leaching .
blank and s - blank are control samples without activity . [ cols="<,^,^",options="header " , ] the final results are given in table [ tab : ams ] .
the amount of atoms in sample s3 agrees with the estimation from the implantation current .
the number for sample s2 is consistent with the lack of observed @xmath30-rays from the decay of @xmath1li .
the number for sample s1 , the @xmath1be sample , is @xmath44 .
be position .
the beta activity measured is shown versus the mass with positions indicated for @xmath1be and the possible contaminant @xmath1li .
the horizontal line marks the detection limit of 0.2/s.,width=264 ] contaminations in our sample might arise due to tails of the neighbouring activities @xmath23be or @xmath1li , whose decay also produces @xmath23be .
both possibilities are ruled out by the low recorded number of atoms for the @xmath1li sample ( s2 ) .
the isolde mass separator profile was found by measuring the beta activity as the mass settings were changed around the nominal @xmath1be mass , see figure [ fig : scan ] .
the release function of this specific target and ion source combination was measured first , which allows to combine measurements with different collection times relative to proton impact on target . in this way
the sensitivity was increased and the activity could be followed down to the @xmath45 level that occured at a mass difference of 0.05 mass units .
the only remaining way for @xmath23be to appear on the @xmath1be position is as the molecule @xmath23be@xmath46h , but this molecule is unlikely to be formed in the target and to survive through the laser ion source since its ionization energy of 8.22 ev @xcite is much higher than its dissociation energy of 3.26 ev .
nevertheless , we have re - checked the data from an earlier experiment on @xmath47be @xcite and were able to put limits on the amount of @xmath1be@xmath46h ( from the @xmath8 branch ) that would correspond in our current case to a @xmath23be@xmath46h intensity less than @xmath48 of @xmath1be .
our conditions should be better , partly due to higher laser ionization power , partly due to the beam passing through a gas - filled rfq cooler , both effects that would enhance molecular break - up .
we therefore conclude that we have observed the @xmath1be(@xmath9p ) decay via detection of the final nucleus @xmath23be .
the observed intensity converts to a branching ratio of @xmath49 .
the experimentally found branching ratio is surprisingly large , but consistent with the outcome of the first experiment .
if the strength in @xmath1be(@xmath9p ) was as broadly distributed as in @xmath1li(@xmath9d ) we would expect the @xmath3 within the q - window to be less than 0.1 , which would not be sufficient to explain the decay rate .
we therefore turn to a simple model for the decay along the lines of the direct decay calculations in @xcite , details of the calculations are reported elsewhere @xcite .
the basic assumption is that the beta decay proceeds as an essentially detached decay of the halo neutron into a proton .
the initial and final state wavefunctions are calculated as single - particle states in square - well or woods - saxon potentials with the final state spectrum discretized by imposing a large confining radius at 1000 fm .
the overlap of the wavefunctions gives the beta strength @xmath3 and the decay rate is found from equation ( [ eq : ft ] ) .
the final total branching ratio for beta - delayed proton emission depends strongly on the strength of the potential between the final state proton and @xmath23be .
for most potential strengths the branching ratio will indeed be a few times @xmath15 , as in other calculations , but in a limited range the beta strength will be concentrated within the q - window .
effectively , in this range the proton formed in the decay interacts strongly with the remaining @xmath23be and forms a resonance - like structure ; as a consequence it emerges with a quite well defined energy .
the branching ratios obtained for this set of parameters are shown in figure [ fig : th_br ] as a function of the energy of the resonance .
be is shown as a function of the centre - of - mass energy of the resonance .
the stars mark results of calculations with square well potentials , the filled circles are results from woods - saxon potentials .
the curves arise from integrating the r - matrix expression in equation ( [ eq : rmatrix ] ) for different widths for other decay channels .
the horizontal band indicates the experimentally found branching ratio . ]
the simple model neglects isospin .
the lowest @xmath50 states are situated slightly more than 1 mev above the @xmath51 window .
they are members of isospin multiplets that include the @xmath1be ground state and first excited state neutron halos .
the data indicate @xcite that the intermediate states ( @xmath52 of 1/2 ) in these multiplets have good total isospin rather than a composition with just one proton ( or neutron ) plus core .
we therefore expect that realistic final state wave functions in our case , with @xmath53 , also should have good isospin .
standard isospin coupling then predicts that the state should be proton plus @xmath23be with weight 2/3 and neutron plus @xmath23b(t=1 ) with weight 1/3 .
our calculated decay probabilities must therefore be corrected by a factor 2/3 .
a further reduction factor about 0.7 is due to the initial @xmath1be wavefunction containing several configurations @xcite .
the overall scaling factor on the theory , included in figure [ fig : th_br ] , is therefore about 0.5 .
could this be an established resonance in @xmath1b ?
the known states @xcite in this region mainly couple to the @xmath54-particle channel ( with partial widths around 100 kev ) and only one , a state at @xmath55 kev , may have spin - parity that allows emission of an s - wave proton the others will have angular momentum barriers that will suppress proton emission . decays through levels that have other sizeable decay channels ( @xmath54 emission or , for very narrow levels , @xmath30 emission ; in principle triton emission could also occur ) would therefore only contribute to the proton channel with probability @xmath56 .
since @xmath57 for the @xmath6 , @xmath50 state at 12.55 mev ( that apart from isospin should be similar in structure to our state ) is about 10 ev @xcite , and even a small admixture into our state of other @xmath6 levels is likely to give a @xmath58 at least of the same magnitude , we shall assume the width for other decay channels @xmath59 to be larger than 0.01 kev . to take these effects into account calculations were also made within the r - matrix approach @xcite ,
but in a simplified version where e.g. other decay channels are approximated as having a constant width @xmath58 over the energy window , see @xcite for details .
converting the decay rate into a differential branching ratio gives the following expression : @xmath60 where @xmath61 , @xmath62 , @xmath63 is the standard ( energy - dependent ) penetrability factor and @xmath64 the maximal reduced width .
integration over the q - window gives the total branching ratios shown in figure [ fig : th_br ] as a function of resonance position @xmath65 for different values of @xmath66 .
the branching ratios agree well with the ones from the simple model .
it is clear that all known levels are too wide to fit and that a @xmath66 above 0.01 kev gives a lower limit on the @xmath3 of about 0.3 with an upper limit given by the theoretical maximum of 3 .
we have observed beta - delayed proton emission for the first time in a neutron - rich nucleus .
the unexpectedly high decay rate can only be understood within current theory if the decay proceeds through a new single - particle resonance in @xmath1b that is strongly fed in beta - decay .
the @xmath3-value could be as high as that of a free neutron decay .
a natural interpretation would be peripheral beta decay of the halo neutron in @xmath1be into a single - proton state .
this appears to be a simpler process than the @xmath9d decays of the two - neutron halo nuclei @xmath16he and @xmath1li .
although the halo structure must be important for the @xmath9p decay mode , the large value of @xmath3 may be related to large values found in other ( non - halo ) near - dripline nuclei @xcite and point to a more widespread change of beta - decay patterns at least in light nuclei in line with some predictions @xcite .
we thank the isolde group for the successful operation of the hrs separator at very high resolution and aksel jensen for discussions on the theoretical interpretation . we acknowledge support from the european union seventh framework through ensar ( contract no .
262010 ) , from austrian science fund ( fwf ) p22164-n20 , from spanish mineco through projects fpa2010 - 17142 and fpa2012 - 32443 , and cpan consolider csd-2007 - 00042 .
88 b. jonson and k. riisager , nucl . phys .
a * 693 * ( 2001 ) 77 .
d. baye and e.m .
tursonov , phys . lett .
b * 696 * ( 2011 ) 464 .
borge et al . , j. phys .
g * 40 * ( 2013 ) 035109 .
jensen , k. riisager , d.v .
fedorov and e. garrido , rev .
* 76 * ( 2004 ) 215 .
i. tanihata , h. savajols and r. kanungo , prog . part .
nucl . phys . * 68 * ( 2013 ) 215 .
k. riisager , physica scripta t**152 * * ( 2013 ) 014001 .
m. pftzner , m. karny , l.v .
grigorenko and k. riisager , rev .
* 84 * ( 2012 ) 567 .
b. blank b and m.j.g .
borge , prog .
* 60 * ( 2008 ) 403 .
m. wang et al . , chinese phys .
c * 36 * ( 2012 ) 1603 .
t. nilsson , g. nyman and k. riisager , hyperfine int .
* 129 * ( 2000 ) 67 .
d. anthony et al . , phys .
c * 65 * ( 2002 ) 034310 .
r. raabe et al .
* 101 * ( 2008 ) 212501 .
borge et al . , z. phys .
a * 340 * ( 1991 ) 255 .
ziegler , particle interactions with matter .
( jan . 20 , 2014 ) s. bubin and l. adamowicz , j.chem.phys . * 126 * ( 2007 ) 214305 .
diget et al . , nucl .
phys . a * 760 * ( 2005 ) 3 .
zhukov , b.v .
danilin , l.v .
grigorenko and j.s .
vaagen , phys .
c * 52 * ( 1995 ) 2461 .
k. riisager , submitted to nucl . phys .
a ( 2014 ) , arxiv:1312.0479 b. jonson and k. riisager , phil . trans .
lond . a * 356 * ( 1998 ) 2063 .
schmitt et al . , phys .
* 108 * ( 2012 ) 192701 .
kelley et al . , nucl .
phys . a * 880 * ( 2012 ) 88 .
f.c . barker and e.k .
warburton , nucl .
phys . a * 487 * ( 1988 ) 269 .
h. sagawa , i. hamamoto and m. ishihara , phys .
b * 303 * ( 1993 ) 215 . | we have observed @xmath0-delayed proton emission from the neutron - rich nucleus @xmath1be by analysing a sample collected at the isolde facility at cern with accelerator mass spectrometry ( ams ) . with a branching ratio of @xmath2
the strength of this decay mode , as measured by the @xmath3-value , is unexpectedly high .
the result is discussed within a simple single - particle model and could be interpreted as a quasi - free decay of the @xmath1be halo neutron into a single - proton state .
beta decay , halo nucleus , @xmath1be |
the purpose of the paper is two - fold : to introduce a multivariable creative telescoping method , and to apply it in a problem of quantum topology : namely the computation of the non - commutative @xmath0-polynomial of twist knots .
our multivariable creative telescoping method allows us to compute linear recursions for sums of the form @xmath1 given a recursion relation for @xmath2 a the hypergeometric kernel @xmath3 .
general theory implies the existence of a recursion relation for @xmath2 . however , in practice the computation is not manageable for twist knots , and there is no guarantee that the recursion relation will be of minimal order .
our method does not guarantee a minimal order recursion relation either , however ( unlike the known methods ) it is manageable and produces a minimal order recursion relation the non - commutative @xmath0-polynomial for twist knots of @xmath10 twists .
the non - commutative @xmath0-polynomial encodes the unique monic , linear , minimal order @xmath6-difference equation satisfied by the sequence of colored jones polynomials of the knot .
our results give a new proof of the aj - conjecture for those knots . in this section
we recall the relevant laurent polynomial invariants of knots , such as the jones polynomial and its colored cousins . in 1985
v. jones introduced the famous _ jones polynomial _ of a knot @xmath11 in 3-space , @xcite .
the jones polynomial ( an element of @xmath12 $ ] ) is a powerful knot invariant which amongth other things detects cheirality , and it can be extended to a sequence @xmath13 of laurent polynomials by taking parallels of a knot @xmath11 .
technically , @xmath14 $ ] is the quantum group invariant of the 0-framed knot using the @xmath15-dimensional representation of @xmath16 , and normalized by @xmath17 . for a detailed definition ,
see @xcite and also @xcite . with this normalization
, we have that @xmath18 , and @xmath19 is the jones polynomial of @xmath11 . for a given knot @xmath11 , the sequence of laurent polynomials @xmath13 is not random .
to be precise , @xmath13 is @xmath6-_holonomic _ i.e. , it satisfies a linear @xmath6-difference equation ( which of course , depends on the knot ) with coefficients in @xmath20 .
this fact , proven in @xcite , is an easy consequence of two facts : * @xmath21 is a finite multisum of a proper @xmath6-hypergeometric term , as follows from the state - sum definition of the colored jones function ; see @xcite .
* ultisums of proper @xmath6-hypergeometric terms are @xmath6-holonomic , as follows from the wz theory of wilf - zeilberger ; see @xcite .
a @xmath6-holonomic sequence is annihilated by a unique monic homogeneous linear @xmath6-difference equation of smallest degree , and the corresponding monic polynomial in two @xmath6-commuting variables @xmath22 and @xmath23 is an invariant ( the so - called _ characteristic polynomial _ of the @xmath6-holonomic sequence .
we define the _ non - commutative @xmath0-polynomial _
@xmath24 of a knot @xmath11 to be the characteristic polynomial of @xmath13 . in @xcite
, it was conjectured by the first author ( the so - called aj conjecture ) that the specialization @xmath25 of @xmath24 should agree with the _ @xmath0-polynomial _ of a knot .
the latter is an important invariant that parametrizes the @xmath26 character variety of the knot complement , as viewed from the boundary torus . for a detailed definition of the @xmath0-polynomial ,
its properties and its applications to the geometry and topology of the knot complement , see @xcite .
thus , @xmath24 can be thought of as a deformation ( or quantization ) of the character variety .
the jones polynomial of a knot is easily computable via skein theory with knots with , say , @xmath8 crossings ; see for example @xcite . on the other hand ,
the @xmath0-polynomial of a knot is much harder to compute , and at present it is unknown for some knots with @xmath9 crossings .
there are two general methods to compute the @xmath0-polynomial : an exact ( primarily elimination , and puiseux expansions ) developed by boyd @xcite and a numerical one developed by culler @xcite .
the non - commutative @xmath0-polynomial and its possible relation with the @xmath0-polynomial of a knot is an important ingredient to the hyperbolic volume conjecture and its generalization . for theoretical as well as experimental reasons it would be good to have explicit formulas for the non - commutative @xmath0-polynomial .
so far , an explicit formula has been given for torus knots in @xcite ( using properties of the kauffman bracket skein module of the solid torus ) , as well as for the simplest hyperbolic @xmath27 knot in @xcite ( using an explicit single - sum formula for the colored jones function ) .
the wz algorithm has been implemented ( see @xcite ) and together with explicit state - sum formulas for the colored jones function of an arbitrary planar projection given in @xcite , in principle one can obtain a linear @xmath6-difference equation for the colored jones function of an arbitrary knot .
there are two problems with this approach : * the number of summation variables in the multisum formulas is generally two less than the number of crossings , and the @xmath6-multisum algorithms appear to be slow for the current machines .
* there is no guarantee that the various @xmath6-multisum algorithms will give a minimal order linear @xmath6-difference equation .
in fact , in many cases ( where symmetry is involved ) , it has been observed that they fail to give the minimal order @xmath6-difference equation .
see @xcite for well - known examples of this failure . with respect to the first problem , we were unable to use the sofware of @xcite to compute a @xmath6-difference equation for our double sums .
one might wonder whether problem ( b ) really occurs for the state - sums that originate in knot theory . as expected
, this problem _ does _ occur .
the knots @xmath28 and @xmath29 have double - sum formulas for their colored jones function .
an application of the @xmath6-multisum package of @xcite was done by takata in @xcite who found out an explicit inhomogeneous @xmath6-difference equation of degree @xmath30 and @xmath30 respectively . on the other hand , as we shall see , there exist inhomogeneous @xmath6-difference equations of degree @xmath31 and @xmath32 respectively . in a different direction ,
le used geometric methods of the kauffman bracket skein module and was able to prove the aj conjecture for most 2-bridge knots , as well as give a linear algebra algorithm that in principle computes the non - commutative @xmath0-polynomial ; see ( * ? ? ?
* thm.1 ) and ( * ? ? ?
* sec.5.6.3 ) .
the algorithm was implemented in maple by the second author , but proved to be too slow to run for the @xmath28 and @xmath29 knots .
the main goal of the paper is to give an explicit formula for the non - commutative @xmath0-polynomial of twist knots with @xmath33 twists , where @xmath34 .
let us recall the _ twist knots _
@xmath35 for integer @xmath33 , shown in figure [ twist ] .
the planar projection of @xmath35 has @xmath36 crossings , @xmath37 of which come from @xmath33 full twists , and @xmath38 come from the negative _
clasp_. @xmath39 for small @xmath33 , these knots may be identified with ones from rolfsen s table ( see @xcite and @xcite ) as follows : @xmath40 @xmath41 let @xmath22 and @xmath23 denote the operators that act on a sequence @xmath42 of laurent polynomials @xmath43 $ ] by : @xmath44 note that @xmath45 .
let @xmath46 denote the _ inhomogeneous _ non - commutative @xmath0-polynomial of @xmath35 .
that is , @xmath47 is monic and minimal degree ( with respect to @xmath22 ) that satisfies the equation @xmath48 where @xmath49 . to convert the inhomogeneous equation above to a homogeneous one ,
see section [ sec.reviewct ] .
\(a ) for @xmath50 we have : @xmath51 ( b ) for @xmath52 we have : @xmath53 the formulas quickly become too lengthy to type . for more information ,
see appendix [ sec.p-33 ] for @xmath54 as well as the data file @xcite for @xmath34 .
theorem [ thm.0 ] gives a new proof of the aj - conjecture for twist knots with @xmath55 twists .
in section [ sec.strategy ] we outline the main strategy .
the idea is to use the recursion relation of the cyclotomic function @xmath56 of the twist knot @xmath35 ( from @xcite ) as well as a single - sum relation between @xmath57 and @xmath58 , together with some new ideas of creative telescoping and some guessing . in section [ sec.reviewct ]
we review the method of creative telescoping and in section [ sec.hybrid ] we present a multi - certificate version that takes into account the product of a hypergeometric summand with a @xmath6-holonomic one .
we conclude with three appendices : in appendix [ sec.gf ] we present an alternative method that uses generating functions ( that was kindly communicated to us by zeilberger ) . in appendix [ sec.p-33 ]
we give the non - commutative polynomial of twist knots @xmath35 for @xmath59 , and in appendix [ sec.acomm ] we give the @xmath0-polynomial of the same knots .
the authors wish to thank d. zeilberger for encouragement and enlightening conversations .
in this section we will describe our strategy to obtain a formula for @xmath60 .
* we consider the cyclotomic function @xmath61 $ ] introduced by habiro in @xcite , who used the notation @xmath62 . * the relation between the cyclotomic and the colored jones functions is given by : @xmath63 where the _ cyclotomic kernel _
@xmath3 is a proper @xmath6-hypergeometric term given for @xmath64 by : @xmath65 where @xmath66 and the _ quantum factorial _ is defined by : @xmath67 ( x ; q)_n = \
{ + [ cols= " < , < " , ] p. paule and a. riese , _ a mathematica @xmath6-analogue of zeilberger s algorithm based on an algebraically motivated approach to @xmath6-hypergeometric telescoping _ , in special functions , @xmath6-series and related topics , fields inst .
* 14 * ( 1997 ) 179210 . | the purpose of the paper is two - fold : to introduce a multivariable creative telescoping method , and to apply it in a problem of quantum topology : namely the computation of the non - commutative @xmath0-polynomial of twist knots .
our multivariable creative telescoping method allows us to compute linear recursions for sums of the form @xmath1 given a recursion relation for @xmath2 a the hypergeometric kernel @xmath3 . as an application of our method
, we explicitly compute the non - commutative @xmath0-polynomial for twist knots with @xmath4 and @xmath5 crossings .
the non - commutative @xmath0-polynomial of a knot encodes the monic , linear , minimal order @xmath6-difference equation satisfied by the sequence of colored jones polynomials of the knot .
its specialization to @xmath7 is conjectured to be the better - known @xmath0-polynomial of a knot , which encodes important information about the geometry and topology of the knot complement . unlike the case of the jones polynomial , which is easily computable for knots with @xmath8 crossings , the @xmath0-polynomial is harder to compute and already unknown for some knots with @xmath9 crossings . |
in the paper we investigate a stochastic version of a linear volterra equation of the general form @xmath0 where @xmath1 , @xmath2 , @xmath3 , @xmath4 is a linear operator and @xmath5 some mapping . to fix our attention
we shall consider the equation ( [ eq1 ] ) in a separable hilbert space @xmath6 with a scalar product @xmath7 , a norm @xmath8 and a complete orthonormal system @xmath9 .
the equation ( [ eq1 ] ) creates a big class of equations and generalizes heat and wave equations and even linear navier - stokes equation .
we refer to the excellent monograph @xcite for a rich survey .
that kind of volterra equation has been studied by many authors in connection with problems arising in mathematical physics , particularly in viscoelasticity , heat conduction in materials with memory , energy balance and termoviscoelasticity . in order to take into account random fluctuations
, we have to consider the equation ( [ eq1 ] ) with random external force .
there are our first considerations concerning numerical treatment of stochastic volterra equations , so we will be grateful for readers remarks and advices .
next , we plan to study the probabilistic features of family of trajectories , take into account different noises and develope numerical schemes for cases where the analytic form of reselvent is not known .
assume that @xmath10 is a probability space with a complete right - continuous filtration and @xmath11 , is a cylindrical wiener process with values in the space @xmath6 .
let us omit , for convenience , the space variable @xmath12 in the equation ( [ eq1 ] ) and introduce the process @xmath11 , instead of function @xmath5 .
hence , we arrive at the following stochastic volterra equation @xmath13 in this part of the paper we recall some results concerning solutions to ( [ eq2 ] ) .
we restrict our considerations to paper containing so - called resolvent approach to volterra equation . the notion _ resolvent _ or _ fundamental solution _ for volterra equation ( [ eq1 ] ) probably comes from friedman and shinbrot @xcite who studied deterministic volterra integral equations in banach space . for recent survey
we refer again to @xcite . in the sequel
we shall assume that the equation ( [ eq1 ] ) is _ well - posed _ , that is , that ( [ eq1 ] ) admits resolvent @xmath14 . as in deterministic case ,
the _ mild solution _ to the stochastic volterra equation ( [ eq2 ] ) is of the form @xmath15 where @xmath14 , is the resolvent family for the equation ( [ eq1 ] ) determined by the operator @xmath4 and the function @xmath16 . in order to study solution ( [ eq3 ] ) it is enough to consider the stochastic convolution @xmath17 where the stochastic integral is defined according to particular case under consideration .
stochastic volterra equations with resolvent approach have been treated by several authors , see e.g. @xcite,@xcite,@xcite,@xcite,@xcite and recently @xcite and @xcite . in the first three papers stochastic volterra equations are studied in connection with viscoelasticity and heat conduction in materials with memory .
the paper @xcite is particularly significant because the authors were the first who have extended the well - known semigroup approach , applied to stochastic differential equations , to the equation ( [ eq2 ] ) .
the resolvent approach is a natural way of extension the semigroup approach which is well - known from the theory of evolution equations .
that approach enables to follow some results and schemes obtained for semigroups .
unfortunately , some results are not valid in our case because resolvent family , @xmath14 , does not satisfy semigroup property .
clment and daprato studied stochastic volterra equation ( [ eq2 ] ) where @xmath4 was self - adjoint , negative operator in the space @xmath6 , such that @xmath18 they considered stochastic volterra equation ( [ eq2 ] ) driven by the noise term @xmath19 of the form @xmath20 where @xmath21 was a sequence of real - valued , independent wiener processes .
they assumed that the kernel function @xmath16 is completely positive .
the consequence of completely positiveness of the function @xmath16 is that the solution @xmath22 , to the following equation @xmath23 is nonnegative and nonincreasing for any @xmath24 .
in fact , @xmath25 $ ] . in @xcite regularity of stochastic convolution ( [ eq4 ] )
is studied and hlderianity of the corresponding trajectories is proved .
[ hyp1 ] + * a is a self - adjoint negative operator and @xmath26 . *
@xmath16 is completely positive . *
@xmath27 .
[ hyp2 ] there exists @xmath28 and @xmath29 such that , + for all @xmath30 we have @xmath31 @xmath32 ^ 2 d\sigma \leq c_\theta\mu^{\theta } |t-\tau|\theta\ ] ] and @xmath33 [ hyp3 ] there exists @xmath34 such that @xmath35 where @xmath36 is a bounded open subset of @xmath37 .
clment and daprato proved the following results . [ 3.1 ] ( @xcite , theorem 2.2 ) assume that hypothesis [ hyp1 ] holds .
then for any @xmath38 the series @xmath39 is convergent in @xmath40 to a gaussian random variable @xmath41 with mean 0 and covariance operator @xmath42 determined by @xmath43 [ 3.2 ] ( @xcite , proposition 3.3 ) under hypotheses [ hyp1 ] and [ hyp2 ] , for every positive number @xmath44 , the trajectories of @xmath45 are almost surely @xmath46-hlder continuous . [
3.3 ] ( @xcite , theorem 4.1 ) under hypotheses [ hyp1 ] , [ hyp2 ] and [ hyp3 ] , the trajectories of @xmath45 are almost surely @xmath46-hlder continuous in @xmath47 for any @xmath48 . in the paper @xcite , white noise perturbation of an integro - differential equation arising in the study of evolution of material with memory is studied . in the next paper @xcite
, the authors considered evolutionary integral equations as appearing in the theory of linear parabolic viscoelasticity forced by white noise .
as earlier , they studied the stochastic convolution that provides regular solutions .
additionally , under suitable assumptions the authors proved that the samples are hlder - continuous . in the remaining part of the paper @xcite ,
the results obtained of that paper were put in a wider perspective by consideration of equations with fractional derivatives . in the paper @xcite , the authors first proved that @xmath41 is a gaussian random variable for any @xmath38 .
next , the transition function @xmath49 , associated with @xmath50 was considered .
when @xmath50 is the resolvent operator of a stochastic volterra equation , the convolution @xmath51 , is not a markov process .
this fact has the consequence that @xmath52 , is not a semigroup and then it is not possible to associate to @xmath53 a kolmogorov equation .
however , the authors characterized those transition functions such that @xmath54 was differentiable for any uniformly continuous and bounded function @xmath55 .
there are some other regularity results concerning stochastic convolution ( [ eq4 ] ) . in @xcite
was studied the case when the equation ( [ eq2 ] ) was driven by a correlated , spatially homogeneous wiener process @xmath19 with values in the space of real , tempered distributions @xmath56 .
let @xmath57 be the covariance of @xmath58 and the associated spectral measure be @xmath59 .
we considered existence of the solutions to ( [ eq2 ] ) in @xmath60 and derived conditions under which the solutions to ( [ eq2 ] ) were function - valued and continuous .
in that case , the initial value @xmath61 , @xmath16 is a locally integrable function and @xmath4 is an operator given in the fourier transform form @xmath62 we introduce the following hypothesis .
[ h ] 1 . for any @xmath63 ,
the equation ( [ eq6 ] ) has exactly one solution @xmath64 locally integrable and measurable with respect to both variables @xmath63 and @xmath38 .
moreover , for any @xmath65 , @xmath66 } \ ; \sup_{\gamma\geq 0 } |s(t,\gamma)|<+\infty .$ ] for some special cases the function @xmath67 may be found explicitly .
for instance @xmath68 , & ~t\geq 0,\hspace{1ex } \gamma\geq 0 .
\end{array}\ ] ] in that case , the resolvent family @xmath14 , determined by the operator @xmath4 and the function @xmath16 is given by the formula ( [ eq7 ] ) and has the form @xmath69 where @xmath70 the following results for stochastic convolution are consequences of properties of stochastic integral . [ 3.4 ] ( @xcite , theorem 2 ) let @xmath19 be a spatially homogeneous wiener process and @xmath71 , the resolvent for the equation ( [ eq2 ] ) . if hypothesis [ h ] holds then the stochastic equation @xmath72 is a well - defined @xmath56-valued process . for each @xmath38
the random variable @xmath73 is generalized , stationary random field on @xmath37 with the spectral measure @xmath74 : @xmath75 \mu(d\lambda)\;.\ ] ] [ 3.5 ] ( @xcite , theorem 3 ) assume that hypothesis [ h ] holds .
then the process @xmath73 is function - valued for all @xmath38 if and only if @xmath76 if for some @xmath77 and all @xmath38 , @xmath78 then , for each @xmath38 , @xmath73 is a sample continuous random field . as we have already written , the volterra equation ( [ eq1 ] ) creates a big class of equations . in particular cases ,
when the operator @xmath4 and the function @xmath16 in the stochastic volterra equation ( [ eq2 ] ) are fixed , there is possible to obtain some additional regularity results .
this is obvious because in particular cases we may use some extra features of solutions to ( [ eq2 ] ) .
for instance , apparently well - known is integrodifferential equation which interpolates heat and wave equations , that is the volterra equation ( [ eq1 ] ) , where @xmath79 , the laplace operator and @xmath80 , where @xmath57 is the gamma function .
recently , deterministic version was studied in detail by fujita @xcite and , independently , by schneider and wyss @xcite and stochastic version of that integrodifferential equation was treated in @xcite and @xcite . in this paper
we shall demonstrate numerical results obtained for that equation in the deterministic version and the stochastic one , as well . in the theory of stochastic volterra equations we are interested not only in the existance , uniqueness and regularity of solutions but in some asymptotics , too .
the paper @xcite is concerned with a limit measure of stochastic volterra equation driven by very general noise in a form of a spatially homogeneous wiener process , with values in the space of tempered distributions @xmath56 .
that paper provides necessary and sufficient conditions for the existence of the limit measure and additionally , it gives a form of any limit measure .
let us summarize the results cited above .
the resolvent operators @xmath14 , corresponding to the volterra equation ( [ eq1 ] ) do not form any semigroup . therefore it is not possible to obtain such strong results as in the case of evolution equations with semigroup generators .
in the case of volterra equations one can not use _ the fractional method of infinite dimensional stochastic calculus_. that method , used for demonstration of continuity with respect to @xmath81 for convolutions with semigroups , enables to obtain only some estimates for the convolutions ( [ eq4 ] ) .
moreover , it is possible only in some special cases , see e.g. @xcite .
it is clear that the volterra equation ( [ eq1 ] ) , in particular its stochastic version ( [ eq2 ] ) , is difficult to study .
it results from the fact that equation ( [ eq1 ] ) contains a wide class of equations .
an essential role is played by the kernel function which , in general , is assumed to be a locally integrable function . as the function @xmath16 and operator @xmath4
determine the resolvent @xmath14 , the type of the function @xmath16 is very important . that is reason why it is so difficult to obtain in a general case the continuity of the convolution ( [ eq4 ] ) and some other theoretical results .
for more general convolution , significant in many applications , like @xmath82 , where @xmath83 is an appropriate process , it becomes even far more difficult .
therefore , in many cases a numerical support of theoretical ( analytical ) considerations is demanded .
we need computations for obtaining estimates in regularity results , choosing some paramenters , choosing function @xmath16 with required properties and for visualization of solutions obtained .
numerical analysis is particularly important when we are not able to obtain analytical results .
additionally , numerical schemes are especially useful for studying the asymptotics .
in this section we construct a scheme for numerical solution of the volterra equation ( [ eq2 ] ) without random part , that is , for @xmath84 we shall consider the case when @xmath4 is the laplace operator . denoting by @xmath85 we can write ( [ eq1a ] ) in the standard form @xmath86 in galerkin method one introduces the complete set of orthonormal functions @xmath87 on the interval @xmath88 $ ] , that is fulfilling conditions @xmath89 where @xmath90 is the scalar product .
the set @xmath91 spans a hilbert space .
the approximate solution is then postulated in the form of an expansion of the unknown true solution in the subspace @xmath92 determined by @xmath93 first basis functions @xmath94 inserting ( [ e3 ] ) into ( [ e2 ] ) we obtain @xmath95 where the function @xmath96 represents the approximation error . from ( [ e5 ] ) we have @xmath97 from ( [ e6 ] ) and ( [ e3 ] ) it can be written as @xmath98 the coefficient functions @xmath99 are determined by the requirement that the error function @xmath96 has to be orthogonal to the subspace @xmath92 @xmath100 then for @xmath101 the following equations hold @xmath102 \phi_j(\tau)d\tau \\ & - & \int_0^t \left [ \int_0^\tau k(x,\tau , s ) \sum_{k=1}^n
c_k(x)\,\phi_k(s)ds \right ] \phi_j(\tau)d\tau\;. \nonumber\end{aligned}\ ] ] the first integral on the r.h.s , due to ( [ e3 ] ) , is very simple @xmath103\phi_j(\tau)d\tau = \sum_{k=1}^n c_k(x ) \int_0^t \phi_k(\tau)\phi_j(\tau)d\tau = c_j(x ) \;.\ ] ] the second one we calculate in our particular case , @xmath104 as follows @xmath105 \phi_j(\tau)d\tau \hspace{25ex } \\
\hspace{18ex } = \sum_{k=1}^n \frac{d^2 c_k(x)}{dx^2 } \int_0^t \left [ \int_0^\tau a(\tau - s)\phi_k(s)ds \right ] \phi_j(\tau)d\tau \nonumber \;.\end{aligned}\ ] ] denoting by @xmath106 d\tau \end{aligned}\ ] ] we arrive at the set of coupled differential equations for the functions @xmath99 @xmath107 this set can be solved numerically ( aproximately ) by discretization , on a grid @xmath108 . applying the difference form for the second derivative @xmath109 , \quad\quad h= x_{i}-x_{i-1}\ ] ] one obtains from ( [ e12 ] ) the following set of linear equations @xmath110 \;,\end{aligned}\ ] ] where @xmath111 .
those equations can be written in the matrix form @xmath112 here , @xmath113-dimensional vectors @xmath114 i @xmath115 have the following structures @xmath116 where by @xmath117 , @xmath118 we denoted the consecutive @xmath119-dimensional blocks of vectors @xmath114 and @xmath115 , respectively .
then we can write the matrix @xmath120 in the block form @xmath121 & \ldots & \left[a_{1n}\right ] \\
\vdots & \cdots & \vdots \\
\left[a_{n1}\right ] & \ldots & \left[a_{nn}\right]\end{array}\right ) \ ; , \end{aligned}\ ] ] where every block is a tridiagonal matrix .
the diagonal blocks have the following structure @xmath122 = \left ( \!\begin{array}{ccccccc } 1\!+\!\frac{2}{h^2}a_{ii } & -\frac{1}{h^2}a_{ii } & 0 & 0 & 0 & \ldots & 0 \\ -\frac{1}{h^2}a_{ii } & 1\!+\!\frac{2}{h^2}a_{ii } & -\frac{1}{h^2}a_{ii } & 0 & 0 & \ldots & 0 \\ 0 & -\frac{1}{h^2}a_{ii } & 1\!+\!\frac{2}{h^2}a_{ii } & -\frac{1}{h^2}a_{ii } & 0 & \ldots & 0 \\
\vdots & & & & & & \vdots \\ 0 & 0 & 0 & \ldots & -\frac{1}{h^2}a_{ii } & 1\!+\!\frac{2}{h^2}a_{ii } & -\frac{1}{h^2}a_{ii } \\ 0 & 0 & 0 & 0 & \ldots & -\frac{1}{h^2}a_{ii } & 1\!+\!\frac{2}{h^2}a_{ii } \end{array}\!\right),\end{aligned}\ ] ] and nondiagonal ones @xmath123 = \left ( \begin{array}{ccccccc } \frac{2}{h^2}a_{ij } & -\frac{1}{h^2}a_{ij } & 0 & 0 & 0 & \ldots & 0 \\ -\frac{1}{h^2}a_{ij } & \frac{2}{h^2}a_{ij } & -\frac{1}{h^2}a_{ij } & 0 & 0 & \ldots & 0 \\ 0 & -\frac{1}{h^2}a_{ij } & \frac{2}{h^2}a_{ij } & -\frac{1}{h^2}a_{ij } & 0 & \ldots & 0 \\ \vdots & & & & & & \vdots \\ 0 & 0 & 0 & \ldots & -\frac{1}{h^2}a_{ij } & \frac{2}{h^2}a_{ij } & -\frac{1}{h^2}a_{ij } \\ 0 & 0 & 0 & 0 & \ldots & -\frac{1}{h^2}a_{ij } & \frac{2}{h^2}a_{ij } \end{array}\right).\end{aligned}\ ] ] the set of linear equations ( [ e14 ] ) can be solved by standard methods , for instance the lu decomposition @xcite .
the essential part of the mild solution ( [ eq3 ] ) is the stochastic convolution ( [ eq4 ] ) .
we present here a particular case when the resolvent @xmath50 is known in analytical form .
let us focus the attention on stochastic volterra equation ( [ eq2 ] ) with the function @xmath16 in the form @xmath124 .
this is the integrodifferential equation @xcite,@xcite,@xcite . for three particular cases , @xmath125 the analytical form of the resolvent @xmath126
is known : @xmath127 where the last form comes from the convolution property .
we shall illustrate the applicability of our numerical algorithms with two cases of the above @xmath16 functions , the case with @xmath128 and with @xmath129 .
then the function @xmath130 in ( [ eq19 ] ) takes the following form @xmath131 we assume the process in the form @xmath132 .
then , the algorithm for an approximate construction of the stochastic integral ( [ eq4 ] ) can be built in the following way .
let us introduce a time grid @xmath133 on @xmath134 $ ] , i.e. @xmath135 and next a finite sequence of independent random variables @xmath136 , with standard normal distribution .
the approximation for the convolution ( [ eq4 ] ) can be written in the form @xmath137 \\ & = & \tau^{\half } \sum_{i=0}^{i-1 } \zeta_i \int_{-\infty}^{\infty } \phi_\alpha ( t - s_i , x - y)\,w_2(y ) dy \nonumber\end{aligned}\ ] ] for further specification we choose @xmath138 ( the constant @xmath139 represents a strength of stochastic forces ) . with this assumption , after performing the integral ( [ eq21 ] ) for particular @xmath130 one obtains @xmath140 and these explicite forms were inserted into the numerical code .
the sequence of independent random variables @xmath136 , with standard normal distribution was generated using subroutines gasdev and ran1 from @xcite .
we illustrate the efficiency of the numerical approach on two examples of the function @xmath16 , mentioned earlier . as the initial value of the @xmath142 we chose the gaussian @xmath143 .
the grid in @xmath12 variable contained @xmath144 intervals with @xmath145 , covering the interval @xmath146 $ ] .
the dimension of the approximation subspace in the galerkin method was chosen as @xmath147 .
the resulting dimension of the matrix @xmath148 was then 1208@xmath1491208 and calculations were performed up to @xmath150 . in fig .
[ err ] we show the errors of the numerical solutions to the deterministic equation ( [ eq1a ] ) obtained in cases @xmath128 ( top ) , and @xmath129 ( bottom ) . in both cases the errors are relatively small . the top part of the fig .
[ a1ds ] displays the solution of the deterministic volterra equation [ eq1 ] for the case @xmath128 as function of time , @xmath151 $ ] .
the bottom part of the fig . [ a1ds ] shows the example of a single stochastic trajectory ( i.e. the sum of deterministic solution and stochastic integral ) for the same case .
clment and g. daprato , some results on stochastic convolutions arising in volterra equations perturbed by noise , _ rend . math .
lincei _ , s. 9 , * 7 * ( 1996 ) , 147 - 153 . ph .
clment and g. daprato , white noise perturbation of the heat equation in materials with memory , _ dynamic systems and applications _ * 6 * ( 1997 ) , 441 - 460 . ph . clment and g. daprato ,
stochastic convolutions with kernels arising in some volterra equations , in : volterra equations and applications ( arlington , tx , 1996 ) 55 - 65 , _ stability control theory methods appl .
_ * 10 * , gordon and breach , amsterdam , 2000 .
clment , g. daprato and j. prss , white noise perturbation of the equations of linear parabolic viscoelasticity , _ rendiconti trieste _ , 1997 .
a. friedman and m. shinbrot , volterra integral equations in banach spaces , _ trans .
soc . _ * 126 * , ( 1967 ) , 131179 .
y. fujita , integrodifferential equations which interpolates the heat equation and the wave equation , _ osaka j. math . _
* 27 * ( 1990 ) , 309321 .
y. fujita , a probabilistic approach to volterra equations in banach spaces , _ diff .
eqs . _ * 5 * ( 1992 ) , 769 - 776 .
a. karczewska , on the limit measure to stochastic volterra equations , _ j. int
. eqs .
appl . _ * 15 * ( 2003 ) , 59 - 77 .
a. karczewska , function valued stochastic convolutions arising in integro - differential equations , submitted .
a. karczewska , the fractional calculus used to linear stochastic volterra equations , submitted .
a. karczewska and j. zabczyk , regularity of solutions to stochastic volterra equations , rend .
s. 9 , * 11 * no.3 ( 2001 ) 141154 .
w.h . press , s.a .
teukolsky , w.t .
vetterling and b.p .
flannery , _ numerical recipes in fortran _ sec.ed .
, cambridge university press , new york , 1992 j. prss , j. , _ evolutionary integral equations and applications _ ,
birkhuser , basel , 1993 .
schneider and w. wyss , fractional diffusion and wave equations , _ j. math .
phys . _ * 30 * ( 1989 ) , 134 - 144 . | the aim of the paper is to demonstrate the use of the galerkin method for some kind of volterra equations , determininistic and stochastic as well .
the paper consists of two parts : the theoretical and numerical one . in the first part we recall some apparently well - known results concerning the volterra equations under consideration . in the second one
we describe a numerical algorithm used and next present some examples of numerical solutions in order to illustrate the pertinent features of the technique used in the paper .
stochastic and deterministic volterra equations , galerkin method . 60h20 , 65c30 , 65r20 , 60h05 , 45d05 |
the study of charm and beauty production is a central topic of research at hera . in @xmath0 collisions , charm and beauty quarks
are predominantly produced by the photon gluon fusion process , which makes the production rates directly sensitive to the gluon density in the proton .
the mass of these heavy quarks provides a hard scale so that calculations in perturbative qcd are expected to be reliable .
however , this seemingly simple picture of photon gluon fusion with a hard scale given by the quark mass is considerably complicated by two issues : on one hand by the presence of additional hard scales such as the quark s transverse momentum , which means that heavy flavour production at hera is inherently a multi - scale problem ; on the other hand by the hadronic structure of the photon that leads to additional contributions such as flavour excitation , which pose additional challenges to theoretical calculations . while it is possible to take the stance that the theoretical issues are sufficiently well under control that one can infer from the measured charm and beauty production rates the gluon density of the proton , a complementary approach is to use the gluon density as determined from other processes , i.e. fits to inclusive structure function data , in particular from hera , and investigate the extent to which the formation of heavy quarks is theoretically understood in a rgime where multiple scales and qcd dynamics play an important rle . in the following , new measurements of charm and beauty production in photoproduction at hera will be presented .
photoproduction is characterized by a near zero virtuality and is driven by the detector acceptance for scattered electrons . ]
@xmath2 of the exchanged photon , which essentially behaves like a real photon .
to correct experimental data for detector effects , monte carlo generators based on leading order ( @xmath3 ) matrix elements and matched parton showers are used .
the pythia program @xcite is based on collinear factorization and dglap evolution of parton densities .
in contrast , cascade @xcite is based on the @xmath4 factorization ansatz , which uses a @xmath4 unintegrated gluon density that is evolved according to the ccfm evolution . for calculations at next to leading order
, the fmnr program @xcite , based on the calculation by nason , dawson and ellis @xcite , is available . in this program , which does not include parton showers , the fragmentation to hadrons is usually modelled using an independent fragmentation approach ; a new development is the fmnr @xmath5 pythia program @xcite , which augments fmnr with full lund string fragmentation .
the h1 collaboration has recently released a new measurement @xcite of charm photoproduction based on an integrated luminosity of @xmath6 of data taken at hera-2 .
charm events were tagged by the presence of a @xmath1 meson decaying in the golden decay channel @xmath7 , @xmath8 ( and charge conjugates ) .
the data were taken using a new trigger @xcite that allows a full mass reconstruction of the @xmath1 candidates at the third trigger level within @xmath9 .
the data covers a kinematic range defined by photon virtualities of @xmath10 and photon - proton centre of mass energies of @xmath11 .
the @xmath1 meson is measured for transverse momenta @xmath12 and pseudorapidities @xmath13 .
differential distributions in @xmath14 , @xmath15 and @xmath16 as well as double differential cross sections in @xmath14 and @xmath15 have been measured and compared to various qcd models .
in particular , two models at nlo qcd are available , namely fmnr and a recent calculation in the gm - vfns @xcite .
both models show similar behavour ( fig .
[ fig:1 ] ) : the @xmath14 spectrum is well described , with a slight deficiency at high @xmath17 in the case of the gm - vfns model , with sizeable uncertainties from scale variations , in particular at low @xmath17 .
when the @xmath15 spectrum is considered , the theoretical uncertainties are several times larger than the experimental ones ; nevertheless , both models show a tendency to fall below the data at large values of pseudorapidity .
a similar behavour is found in the deep - inelastic regime @xcite .
the measurement of beauty photoproduction at hera is difficult due to the fact that beauty production corresponds only to about @xmath18 of the total photoproduction cross section .
two basic techniques have been used to enrich beauty production events : the measurement of track impact parameters or secondary vertices , using a silicon vertex detector , which enrich beauty production due to the relatively large lifetime of beauty hadrons , and measurements based on leptons ( muons and/or electrons ) , which are based on the large semileptonic branching fraction of beauty hadrons . the h1 and zeus collaborations have recently released results from several analyses based on lepton tags .
two analyses , one from h1 @xcite and one from zeus @xcite , are based on samples with at least one muon and two jets in the final state ; in both cases , jet transverse momenta above @xmath19 are required .
the resulting samples are enriched to about @xmath20 of beauty events by the requirement of two high-@xmath17 jets in conjunction with a highly energetic muon .
the actual fraction of beauty events is then determined from fits to the distribution of the impact parameter @xmath21 of the muon track and the relative transverse momentum @xmath22 of the muon with respect to the nearest jet axis . in these measurements ,
the @xmath22 distribution allows the separation of beauty induced events from those of light ( uds ) and charm quarks with good precision ; the impact parameter distribution on the other hand allows in addition the separation of light and charm quark induced events and thus serves not only as a cross check of the @xmath22 based measurement , but also helps to reduce the systematic uncertainty which would otherwise be incurred by the uncertainty of the charm quark contribution to the @xmath22 spectrum . both measurements cover a range of @xmath23 in inelasticity , momentum transfer @xmath24 , muon transverse momentum @xmath25 and jet pseudorapidity @xmath26 ; however , the zeus data cover a range @xmath27 of muon pseudorapidities , while the h1 analysis is restricted to @xmath28 .
the zeus analysis is based on @xmath29 of data taken in 2004 and 2005 , while the h1 result uses @xmath30 of data from the years 2006 and 2007 .
both analyses show good agreement with perturbative qcd calculations performed with the fmnr program , after corrections for hadronisation effects have been applied to the fmnr predictions , which are available only at parton level .
the excess of data over nlo predictions at low values of jet and muon transverse momenta that was reported in an earlier h1 analysis of hera-1 data @xcite and was in disagreement with zeus data @xcite , is not confirmed by the new h1 analysis .
zeus has also published a new measurement based on events with semileptonic decays to electrons @xcite . in this analysis ,
events with two jets of @xmath31 and at least one electron are selected from @xmath32 of data taken at hera-1 .
the beauty contribution was determined from a fit to a likelihood variable based on quantities that are sensitive to the quality of electron identification , such as the specific ionisation @xmath33 , as well as quantities that allow the separation between electrons from light quark or charm induced processes from bottom production .
these quantities are again @xmath34 , and @xmath35 , the azimuthal angle difference between the electron flight direction and the direction of the total missing transverse momentum . again
, reasonable agreement with the nlo qcd calculation fmnr is observed , for the total cross section as well as differential distributions .
r0.5 an upcoming publication by zeus @xcite is based on @xmath36 of data taken at hera-1 and utilizes events with two muons in the final state .
contrary to other work discussed here , no jet requirement is imposed in the selection , which is possible due to the high beauty fraction of a dilepton sample , and the analysis is performed down to muon transverse momenta of @xmath37 , or even @xmath38 for high - quality muon candidates .
the analysis is thus sensitive down to to the threshold of beauty production , where the transverse momentum of the beauty quarks is close to zero .
therefore a measurement of the total beauty production cross section is possible with relatively small extrapolation errors .
the total cross section for the process @xmath39 at @xmath40 has been determined to be @xmath41 , to be compared to the nlo qcd prediction of @xmath42 . within the large uncertainties ,
in particular of the nlo calculation , the nlo prediction is consistent with the data .
[ fig:2 ] summarizes all currently available beauty photoproduction measurements at hera , in comparison to the qcd prediction at nlo .
it can be seen that nlo qcd describes the data well in shape and normalization , albeit with rather large uncertainties , which are dominated by the choice of scale .
note that for this plot the central value of the qcd prediction is derived for a choice of @xmath43 for the renormalization and factorization scales @xcite , while the individual measurements sometimes use twice this value , i.e. @xmath44 for the central value .
the study of the photoproduction of charm and beauty quarks at hera remains a challenging testing ground for perturbative qcd . with the increased statistical precision provided by the hera-2 data set , and improved detector understanding that results in more precise measurements from hera-1 , measurements of charm and beauty production are now available with experimental uncertainties that are significantly smaller than the theoretical uncertainties of qcd calculations , even at next to leading order .
using gluon densities based on fits to inclusive structure function measurements , these calculations in general describe the data reasonably well in shape and normalization , which is a great success of perturbative qcd .
however , differential distributions such as the @xmath45 distribution of @xmath1 mesons indicate deficiencies in the calculations which may need further work to resolve .
t. sjstrand _ et al .
_ , comput .
commun .
* 135 * ( 2001 ) 238 [ hep - ph/0010017 ] .
h. jung and g. p. salam , eur .
j. c * 19 * ( 2001 ) 351 [ hep - ph/0012143 ] ; + h. jung , comput .
commun .
* 143 * ( 2002 ) 100 [ hep - ph/0109102 ] .
m. l. mangano , p. nason and g. ridolfi , nucl .
b * 373 * ( 1992 ) 295 - 345 ; + s. frixione , m. l. mangano , p. nason and g. ridolfi , nucl .
b * 412 * ( 1994 ) 225 - 259 [ arxiv : hep - ph/9306337 ] ; + s. frixione , p. nason and g. ridolfi , nucl .
b * 454 * ( 1995 ) 3 [ hep - ph/9506226 ] .
p. nason , s. dawson and r. k. ellis , nucl .
b * 303 * ( 1988 ) 607 ; + _ idem , _ nucl .
b * 327 * ( 1989 ) 49 [ erratum - ibid .
b * 335 * ( 1990 ) 260 ] ; + r. k. ellis and p. nason , nucl .
b * 312 * ( 1989 ) 551 .
a. geiser and a. e. nuncio quiroz , arxiv:0707.1632 [ hep - ph ] .
h1 collaboration , _ measurement of @xmath1 meson production in photoproduction , _
h1prelim-08 - 073 , http://www-h1.desy.de/h1/www/publications/htmlsplit/h1prelim-08-073.long.html ; + k. urban , phd thesis univ .
heidelberg , in preparation .
a. schning [ h1 collaboration ] , nucl .
instrum .
meth . a * 518 * ( 2004 ) 542 .
g. kramer and h. spiesberger , eur .
j. c * 38 * ( 2004 ) 309 [ arxiv : hep - ph/0311062 ] ; + b. a. kniehl _ et al .
_ , phys .
d * 71 * ( 2005 ) 014018 [ arxiv : hep - ph/0410289 ] ; + h. spiesberger , private communication .
h1 collaboration , _ a measurement of beauty photoproduction through decays to muons and jets at hera - ii , _
h1prelim-08 - 071 , http://www-h1.desy.de/h1/www/publications/htmlsplit/h1prelim-08-071.long.html ; + m. krmer , phd thesis univ .
hamburg , in preparation .
a. aktas _ et al .
_ [ h1 collaboration ] , eur .
j. c * 41 * ( 2005 ) 453 [ arxiv : hep - ex/0502010 ] .
s. chekanov _ et al . _
[ zeus collaboration ] , phys .
d * 70 * ( 2004 ) 012008 [ erratum - ibid .
d * 74 * ( 2006 ) 059906 ] [ arxiv : hep - ex/0312057 ] .
s. chekanov _ et al . _
[ zeus collaboration ] , arxiv:0805.4390 [ hep - ex ] .
s. chekanov _ et al . _
[ zeus collaboration ] , _ measurement of beauty production using dimuon events with zeus at hera , _
desy-08 - 128 . | after the completion of data taking at hera-2 , a large data set is at hand to study the photoproduction of charm and beauty quarks in @xmath0 collisions .
new measurements of charm production based on @xmath1 meson tagging and beauty production based on muon and electron reconstruction test perturbative qcd calculations with improved accuracy . in general , qcd calculations at nlo describe the data well .
the scale uncertainties are , however , large and dominate over the experimental uncertainties , calling for more precise calculations . |
with _ chandra _ just launched , and xmm and astro - e due to follow suit soon , it is the right time to review what we have learned on seyfert 1 galaxies from previous generation satellites like asca and bepposax , and point out the still open questions , hopefully to be answered or at least addressed by the new generation missions . in particular , i will concentrate here on bepposax results , not only because i worked myself on them but also because it is only now that a general picture from these observations is emerging . before discussing the bepposax observations , however , let me at least mention briefly the recent results from asca and xte which i consider most significant .
the most exciting result in recent years on variability in seyfert galaxies is probably the discovery of periodicity in the flux of the compton
thin seyfert 2 galaxy iras 18325 - 5925 ( iwasawa et al . 1998 ) .
the discovery has been obtained with an imaging satellite , asca ; when the field is observed with a better spatial resolution instrument ( rosat ) , no evidence for a confusing source is found .
the chance of the periodicity being due to a cataclysmic variable ( like in the infamous case of ngc 6814 , madejski et al .
1993 ) or some other confusing source is therefore very low . as periodicity has been observed so far only in one ( or two , see below ) sources , it is possible that it is a transient event , e.g. a flaring region orbiting around the black hole with a lifetime longer than usual ( or being so strong to outshine all other flares simultaneously present ) .
iras 18325 - 5925 has many characteristics typical of the subclass of narrow line seyfert 1 galaxies ( see boller , 1999 , for a review ) : large amplitude variation , steep spectrum , possibly ionized disc .
one can then argue that periodicities are better to be searched for in nls1s rather than in ` classical ' seyferts .
however , a recent claim of a 33 h periodicity in mcg6 - 30 - 15 by lee et al .
( 2000 ) could indicate that periodicity may be a not so rare phenomenon , after all , and that it has been generally missed so far due to lack of long enough observations .
thanks mostly to rxte , long monitoring campaigns in x
rays simultaneously with optical and uv have been performed on a few selected objects , with quite surprising and important results . in ngc 7469 ( nandra et al .
1998 ) , ngc 5548 ( chiang et al .
1999 ) and ngc 3516 ( maoz et al .
2000 ) , optical / uv and x ray emissions seem to be uncorrelated or , if any correlation exists , is the optical / uv that leads x rays , opposite to what expected in the reprocessing scenario .
this scenario ( collin souffrin 1991 ) , in which optical and uv emissions arises as reprocessing in the accretion disc of the hard x
rays , was suggested to explain the too tight correlation ( and lack of delays ) between optical and uv fluxes ( courvoisier & clavel 1991 ) in some seyfert 1s ( ngc 5548 being one of the best examples ) .
so , the situation is rather confused and puzzling , as in the same source both the viscous dissipation and the reprocessing scenario appear to be unviable . regarding even longer term variability ,
the most spectacular result recently has been the switching off of the swan s song " galaxy , ngc 4051 ( guainazzi et al .
when observed by bepposax in may 1998 , only the compton reflection component and the iron line were visible , while the primary radiation had disappeared . when put in the context ( uttley et al .
1999 ) , this observation demonstrated that the observed reflected emission occurred in matter distant at least a few light
months from the nucleus .
so far , this is the most convincing evidence for the presence of large amount of cold circumnuclear matter in seyfert 1 galaxies .
warm absorbers are detected in at least half of seyfert 1s ( reynolds 1997 , george et al .
1998 ) , indicating that the covering factor of the ionized material is fairly large .
the reason why warm absorbers are not observed in all seyferts is not clear
. it may be due either to a spread in column density and/or ionization structure , or to a covering factor of the ionized matter less than unity . in the first case , an improved sensitivity , as for instance that provided by xmm , is necessary to search for small edges due either to small columns of oxygen or to less abundant elements ( e.g. neon ) . in the second case ,
the ionized material may be observed via emission lines ( e.g. netzer 1993 ) , and high resolution detectors ( gratings , _ chandra _ and xmm , or calorimeters , astro - e ) are needed .
the most interesting recent discovery possibly related to warm absorbers is the 1 kev absorption feature observed in some nls1s ( vaughan et al .
, 1999a ; leighly 1999 and references therein ) . when interpreted as a blend of several resonant lines , mainly from iron l ( nicastro , fiore & matt 1999 ) , this implies that the ionization structure of narrow and broad lines seyferts is rather different , not surprisingly given the different ionizing continuum , especially in soft x rays .
other possible explanations include blueshifted absorption edges ( leighly et al .
1997 ) and oxygen smeared edge from ionized discs ( fabian , private communication ; see also vaughan et al .
xmm should easily distinguish between competing models .
the iron line in seyfert 1 galaxies is usually broad ( nandra et al . 1997 ) , likely originating in a relativistic disc ( e.g. fabian et al .
in the two best studied sources , mcg6 - 30 - 15 ( tanaka et al .
1995 ; guainazzi et al .
1999 ) and ngc 3516 ( nandra et al .
1999 ) , the observation of a double horned and asymmetric line profile , expected in presence of fast orbital motion and strong gravity effects ( e.g. fabian et al . 1989
; matt et al . 1992 ) , makes this interpretation quite strong . in principle , from the line profile and short term variability in response to variations of the continuum ( i.e. the so called reverberation mapping ) , it is possible to determine the spin of the black hole and the geometry of the accretion disc and the illuminating source ( e.g. martocchia , karas & matt 1999 and references therein ) .
large sensitivity , coupled with good energy resolution , it is required for these goals .
it is possible that even the next missions will not sufficient for this purpose , and we have probably to wait for constellation x ( e.g. young & reynolds 2000 ) and xeus .
it is worth noting , however , that already with asca it has been possible to start studying variations of the iron line in mcg6 - 30 - 15 ( iwasawa et al . 1996 ; 1999 ) .
even if the results are not yet conclusive , these works demonstrated the potentiality of the method .
while most of the iron lines observed in seyfert 1s can be explained in terms of relativistic discs , it is not clear whether emission from more distant matter ( let us call it the torus " ) is also present .
evidence for this component is scanty , and the only unambiguous cases so far are those of ngc 4151 ( e.g. piro et al .
1999 ) and of ngc 4051 discussed in sec.2.1 . in the latter case , the discovery was due more to a stroke of luck rather than to systematic searches , and the probability to find other sources switched off seems not very high .
more promising , in view of missions like xmm and expecially astro e with large sensitivity coupled with good energy resolution , is to search for narrow iron lines ( as the torus should be at a distance such that relativistic effects are negligible ) . while there is wide consensus on the fact that the power law index ( assuming that a power law is actually a good enough description of the primary continuum : see petrucci et al . , 1999 )
is typically around 1.82 ( nandra & pounds 1994 ) and that the compton reflection is a common ingredient , there are still several open questions : is there a correlation between the photon index and the amount of compton reflection , as suggested by zdziarski et al ( 1999 ) ? how common is the soft excess in broad lines seyfert 1s ( there is actually no doubt that it is common in narrow lines seyfert 1s ) ?
what is the typical ( if any ) cut off energy in the hard x ray spectra ? to address these questions , let me discuss the bepposax observations of a dozen of moderately bright seyfert 1s .
in table 1 , the observed sources and the main spectral parameters are listed .
when the sources have been observed more than once , the results refer to the combined spectrum if there is not spectral variability , or otherwise to the observation when the source was brightest . for all sources ,
the spectral model includes : a power law with exponential cut off ; a gaussian line ( or a relativistic line , when it provides a significantly better fit ) ; the compton reflection continuum . a warm absorber and/or a soft excess
have been added if required by the data .
the spectral parameters reported in the table refer to the best fit model , whatever its complexity .
r & ew ( ev ) & e@xmath0 ( kev ) & & & & & & & & & & ngc 5548@xmath1 & no & 1.63@xmath2 & 0.44@xmath3 & 120@xmath440 & 160@xmath5 & & & & & & & & & & ngc 3783@xmath6 & no & 1.71@xmath7 & 0.37@xmath8 & 190@xmath465 & 160@xmath9 & & & & & & & & & & ngc 7469@xmath6 & no ? & 2.04@xmath10 & 1.1@xmath11 & 180@xmath490 & @xmath12 230 & & & & & & & & & & ngc 4151@xmath6 & ? & 1.35@xmath13 & 0.45@xmath13 & 240@xmath470 & 70@xmath420 & & & & & & & & & & fairall 9 & ? & 2.05@xmath14 & 1.2@xmath15 & 220@xmath16 & @xmath12 220 & & & & & & & & & & ngc 5506 & ? & 2.03@xmath17 & 1.5@xmath18 & 175@xmath445 & @xmath12 380 & & & & & & & & & & ngc 4593@xmath19 & no & 1.87@xmath10 & 1.1@xmath20 & 190@xmath21 & @xmath12 150 & & & & & & & & & & ic 4329a@xmath22 & no & 1.86@xmath23 & 0.55@xmath24 & 110@xmath25 & 270@xmath26 & & & & & & & & & & mrk 509@xmath27 & yes & 1.58@xmath28 & 0.6@xmath29 & 110@xmath30 & 70@xmath31 & & & & & & & & & & mcg-8 - 11 - 11@xmath27 & no & 1.84@xmath10 & 1.0@xmath32 + & 130@xmath33 & 170@xmath34 & & & & & & & & & & mcg-6 - 30 - 15@xmath35 & no & 2.06@xmath23 & 1.2@xmath36 & 200@xmath37 & 160@xmath30 & & & & & & & & & & mrk 841 & yes ? & 2.16@xmath38 & 3.9@xmath39 & 290@xmath40 & @xmath12 150 & & & & & & & & & & ngc 3516@xmath41 & no & 2.04@xmath23 & 1.4@xmath20 & 100@xmath42 & @xmath12 350 & & & & & @xmath6 piro et al .
1999 ; derosa et al . 1999 ; @xmath43 nicastro et al . 2000 ; @xmath19 guainazzi et al . 1999a ; @xmath22 perola et al . 1999 ; @xmath27 perola et al . 2000 ; @xmath35 guainazzi et al .
1999b ; @xmath44 costantini et al .
2000 . in fig .
1 the distributions of @xmath45 , @xmath46 ( the solid angle , in units of 2@xmath47 , subtended by the reflecting matter , assumed to be observed face on ) and of the equivalent width of the iron line , are given .
the results are in agreement with those found by nandra & pounds ( 1994 ) and nandra et al .
( 1997 ) and based on _ ginga _ and asca observations , respectively . before the launch of bepposax , a high energy cut off was unambiguously detected only in one source , ngc 4151 ( jourdain et al .
, 1992 ) . from the analysis of the spectrum obtained summing several sources observed with osse , gondek et al
. suggested an average cut off energy of a few hundreds kev . from table 1 , it can be seen that a high energy cut off
is positively detected in about half of the sample .
the detected values range from 70 kev ( ngc 4151 ) to 270 kev ( ic4329a ) .
when only a lower limit can be obtained , the values are generally consistent with the measured ones .
the plot of the @xmath48-folding energies vs. the power law index is shown in fig . 2 .
while there is no obvious correlation between the two parameters , there is a tendency for flat sources to avoid high values of the cut off energy . whether this reflects something intrinsic to the spectra , or simply the fact that it is easier to determine high energy cut offs in flat sources ,
can not be said with the present data . only two out of 13 sources ( namely mkn 509 and mkn 841 ) in the bepposax sample discussed here present evidence for a soft excess . even if for three sources ( ngc 5506 and ngc 4151 , which have a large intrinsic absorption , and fairall 9 , for which the lecs instrument was switched off , for technical problems , for almost all the observation ) we can not say anything , the soft excess appears to be the exception rather than the rule . for mkn 509
, the inclusion of a soft excess ( in the form of a steeper power law dominating below about 1 kev ) improves significantly the goodness of the fit and provides a more astrophysically sound global solution ( perola et al .
2000 ) . for mkn 841
, a warm absorber fits the spectrum almost as well as a black body soft excess , but from asca it seems that the warm absorber is not very relevant in this object ( see e.g. reynolds 1997 ) . in both sources ,
the shape and luminosity of the soft component remains rather undetermined , and what can be said is that a steepening of the spectrum occurs below 1 kev or so . whether there is a curvature in the primary x
ray emission , or there are two different components , it is impossible to establish . in fig .
3 the iron line ew is plotted against @xmath46 . in the reprocessing scenario ,
the two quantities are expected to be strongly correlated ( george & fabian 1991 ; matt , perola & piro 1991 ) .
the error bars , however , are too large to really check this hypothesis . the relatively small probability ( 3.4% ) that they are not correlated
is mainly due one point , i.e. mkn 841 ( but see next paragraph ) .
more interesting and instructive is the comparison between the average values of the two quantities ( fig . 1 ) , as it can be used to estimate the iron abundance , which turns out to be consistent ( say , within a factor of 2 or so ) with the cosmic value . in fig .
4 @xmath46 is shown against the photon index . a correlation between the two quantities
is evident , the probability that they are not correlated being only 1.3% if mkn 841 is excluded , and less than 0.1% when it is included .
however , a word of caution is necessary here .
the two parameters are strongly correlated in the fit procedure ( see also the discussion in nandra et al . 2000 ) . if , for some reason , the photon index is miscalculated , and e.g. a value steeper than real is found , a larger @xmath46 is immediately obtained .
let me use the case of mkn 841 as an example .
if the soft excess is modeled with a black body , the results for @xmath45 , @xmath46 and the iron ew listed in table 1 and presented in the figures are obtained . an even steeper spectrum ( @xmath45=2.27 ) and ,
correspondingly , larger @xmath46 ( 4.6 ) is obtained with the warm absorber model .
conversely , if the soft excess is modeled by a power law , the photon index is significantly flatter ( ( @xmath45=1.75 ) and @xmath46 reduces to 1.5 ( but the line ew remains fairly large , e.e .
350 ev ) .
i believe that the issue of the reality of the @xmath45@xmath46 correlation is still open , and can be further addressed only by much more sensitive missions like xmm ( but the energy band may be not broad enough ) or constellation x .
the main results from the bepposax observations of a sample of seyfert 1 galaxies can be summarized as follows : * in most sources there is no evidence for a soft excess . in only two sources , mkn 509 and mkn 841 , a steepening of the spectrum below about 1 kev seems required by the data .
whether it is due to a curvature of the primary emission , or to an altogether distinct component , is impossible to say . * a high energy cut off with @xmath48folding energy typically in the range @xmath49100200 kev is observed , when the quality of the data is good enough . * the power law index , @xmath45 , and the amount of reflection component
are clearly correlated , confirming the findings of zdziarski et al .
however , it is possible that such a correlation is not real , but partly or entirely due to the fact that the two parameters are strongly correlated in the fit procedure .
let us suppose that some values of @xmath45 are miscalculated , either because there are other components not taken into account , or because a power law is a too simple model ( e.g. petrucci et al . 1999 ) . in this case ,
a correlation between @xmath45 and @xmath46 would be immediately found , as the quality of present ( and of course past ) data is not sufficient to measure the amount of compton reflection independently of the shape of the primary continuum .
the bepposax results have been presented on behalf of large teams led by g.c .
perola and l. piro .
i thank e. costantini , a. de rosa , f. nicastro , l. piro and especially g.c .
perola for many useful discussions , and for kindly providing the results of their analysis before publication .
i acknowledge financial support from the italian space agency , and from murst ( grant cofin980232 ) . | recent results on the x ray spectrum and variability of seyfert 1 galaxies are reviewed .
new spectral results from bepposax observations are also presented and briefly discussed . |
this manuscript has been authored by ut - battelle , llc under contract no . de - ac05 - 00or22725 with the u.s .
department of energy .
the united states government retains and the publisher , by accepting the article for publication , acknowledges that the united states government retains a non - exclusive , paidup , irrevocable , world - wide license to publish or reproduce the published form of this manuscript , or allow others to do so , for united states government purposes .
the department of energy will provide public access to these results of federally sponsored research in accordance with the doe public access plan(http://energy.gov / downloads / doepublicaccess - plan ) .
the quantum spin liquid ( qsl ) holds particular fascination as a state of matter that exhibits strong quantum entanglement yet is devoid of long - range order @xcite .
these exotic states can possess topologically protected fractionalized excitations , with possible implications for quantum information science @xcite .
a prototypcal example is the kitaev model on a honeycomb lattice@xcite , which can be solved exactly and has a qsl ground state .
an effective hamiltonian with kitaev terms consisting of bond - directional ising couplings can arise in spin - orbit assisted mott insulators with @xmath8 moments in an edge - sharing octahedral environment @xcite . a strong push for the experimental realization of quasi-2d honeycomb lattices showing kitaev physics initially focused on iridate materials with chemical formula a@xmath9iro@xmath1 @xcite , and more recently there has been much interest in @xmath0-rucl@xmath1 .
each of these compounds orders magnetically at low temperatures in a zig - zag or incommensurate phase @xcite , and the effective low energy hamiltonian is believed to be described by a generalized heisenberg - kitaev-@xmath10 model @xcite . despite the appearance of long - range order , broad scattering continua observed via inelastic neutron or raman scattering in @xmath0-rucl@xmath1 @xcite and the iridates @xcite
match the predicted signatures of itinerant majorana fermions in pure kitaev calculations @xcite , suggesting that kitaev interactions play an important role in these materials and that they are proximate to the qsl state .
in this letter we report the evolution of the magnetic ground state in @xmath0-rucl@xmath1 with magnetic ru@xmath4 substituted by nonmagnetic ir@xmath4 , and determine a complete phase diagram as a function of temperature and dilution . the motivation to understand the response of qsl - adjacent systems to perturbations is two - fold :
first , to account for the role of unavoidable defects in real materials and secondly , to explore avenues towards suppression of long - range order and the realization of the pure qsl or other exotic states .
numerous theoretical studies predict the emergence of novel superconductivity with hole doping in the strong kitaev limit @xcite while bond disorder @xcite , dislocations @xcite , magnetic impurities @xcite , and spin vacancies @xcite have also received theoretical attention .
experimentally , substitution of both magnetic and non - magnetic cations for ir rapidly led to spin glass freezing in na@xmath9iro@xmath1 and li@xmath9iro@xmath1 @xcite .
isoelectronic substitution within the solid solution ( na , li)@xmath9iro@xmath1 decreased the magnetic ordering temperature , although phase separation has hampered efforts to completely suppress long - range order @xcite . with no intervening alkaline earth cation between layers ,
the chemically simpler binary compound @xmath0-rucl@xmath1 provides an excellent framework to explore the effects of various perturbations on the relevant kitaev physics .
low - spin 5d@xmath11 ir@xmath12 represents a non - magnetic impurity in the @xmath8 ru@xmath4 magnetic sublattice while the identical ionic radii ( 0.68 ) of ru@xmath4 and ir@xmath4 preserve a regular mcl@xmath13 environment .
the weakly van der waals bonded honeycomb layers in @xmath0-rucl@xmath1 are very susceptible to stacking faults , and these are known to affect the magnetic properties .
two zig - zag magnetic phases , distinguished by out - of - plane periodicity ( abc vs abab ) and t@xmath14 ( 7 k vs 14 k ) can be observed depending on growth conditions and sample history @xcite .
single crystals with pristine monoclinic stacking exhibit a single 7 k transition , while the appearance of a second transition at 14 k is associated with the presence of stacking faults @xcite . in ru@xmath2ir@xmath3cl@xmath1
we observe a rapid suppression of the 7 k ( abc - type ) zig - zag antiferromagnetic transition of the clean @xmath0-rucl@xmath1 system with increasing @xmath6 , demonstrating that site dilution represents a viable approach to destabilizing long - range magnetic order .
ru@xmath2ir@xmath3cl@xmath1 also exhibits a secondary magnetic phase of abab type that is less sensitive to ir content and completely suppressed only when the 2d percolation threshold is crossed .
ir@xmath3cl@xmath1 ( @xmath6 values indicated in legend ) single crystals with a magnetic field of @xmath15 1 t applied in the ab plane .
inset : magnification of low - temperature range .
red arrow and black / grey arrows mark the evolution of @xmath16 and @xmath17 , respectively . ( b )
field - dependent magnetization at 2 k. inset : curie - weiss intercepts @xmath18 for various compositions with a field of 2 t. ( c ) heat capacity curves of low - doped crystals .
solid lines are an estimate of the lattice contribution . dashed lines are a guide to the eye . ] single crystals of ru@xmath2ir@xmath3cl@xmath1 were prepared using vapor transport techniques from commercial powders of @xmath0-rucl@xmath1 and @xmath0-ircl@xmath1 .
polycrystalline ru@xmath2ir@xmath3cl@xmath1 was also synthesized from direct reaction of the starting materials .
the composition of each crystal was checked by energy dispersive x - ray spectroscopy , and @xmath6 hereafter refers to the observed value averaged over a number of sites .
single crystal x - ray diffraction experiments confirm a monoclinic c2/@xmath19 space group in small ru@xmath2ir@xmath3cl@xmath1 crystals consistent with the end members , @xmath0-rucl@xmath1 and @xmath0-ircl@xmath1 @xcite , with ir substituting for ru on the @xmath20 site . in line with previous studies
@xcite we adopt trigonal notation in the following discussion for simplicity ; this affords a more natural representation of the honeycomb planes and , in fact , properly describes the room temperature structure of many larger ( @xmath21 1 mm diameter ) @xmath0-rucl@xmath1 crystals @xcite .
single crystal neutron diffraction was carried out at the hb-1a triple axis spectrometer ( [ h 0 l ] scattering plane ) and hb-3a four - circle diffractometer at the high flux isotope reactor at ornl with incident neutron wavelengths of 2.36 and 1.546 , respectively .
muon spin rotation experiments on powder samples were performed at the triumf m20 beamline in the lampf @xmath22sr spectrometer .
figure 1(a ) shows the in - plane magnetic susceptibility of a series of ru@xmath2ir@xmath3cl@xmath1 single crystals .
in addition to the @xmath23 k zig - zag transition of the parent compound , a second feature at @xmath24 k for @xmath25 ( black arrow , fig.1a inset ) indicates that doped crystals obtained from the current growth process are not completely free of stacking faults .
the sharp cusp in the susceptibility at @xmath16 is rapidly shifted toward lower temperatures at non - zero @xmath6 ( red arrow ) denoting a global suppression of the ordering temperatures with dilution . as doping increases
, the minority phase contribution ( @xmath17 ) becomes more visible and is the only transition evident above 2 k by @xmath5 .
further ir substitution leads to a rounding of the cusp at @xmath17 ( grey arrow ) , which decreases continuously with @xmath6 and disappears at the percolation limit of the honeycomb lattice , @xmath26 @xcite .
strong irreversibility between zero - field - cooled and field - cooled curves was not observed at any level of site dilution , in notable contrast to the spin - glass - like characteristics reported in the honeycomb iridates with non - magnetic substitution @xcite .
the low - temperature susceptibility does not decrease monotonically with @xmath6 , but shows an increase for @xmath27 .
the origin of this behavior is unexplained , however it is interesting to note that vacancies in a site diluted kitaev model have been shown to lead to an increase in the local susceptibility @xcite .
large effective paramagnetic moments of the @xmath28 ru ions in @xmath0-rucl@xmath1 have been previously noted and are attributed to a mixed spin - orbital character @xcite . for ru@xmath2ir@xmath3cl@xmath1
the effective paramagnetic moment per formula unit determined from curie - weiss fittings of the high - temperature data decreased from 2.4 @xmath29 at @xmath30 to 1.9 @xmath29 at @xmath31 , consistent with s = 0 ir ions within systematic uncertainties .
normalized to ru content , the moment per ru ion in ru@xmath2ir@xmath3cl@xmath1 retains large values in the range of 2.3 - 2.7 @xmath29/ru@xmath4 ; a significant trend with @xmath6 can not be established from the current experiment , however some small systematic dependence also can not be ruled out .
curie - weiss intercepts @xmath18 ( inset , fig .
2b ) are positive in all samples and follow an overall decreasing trend with doping , although the values appear to level off at @xmath32k for @xmath33 after initially dropping more rapidly .
in the parent compound the field - dependent magnetization at 2 k shows an upward curvature approaching a crossover out of the zig - zag phase near 8 t @xcite . for @xmath34 0.09 , the field - dependent magnetization curves in fig .
2(b ) develop opposite concavity from the clean limit .
interestingly , this coincides with the region of increase in the low temperature susceptibility mentioned above .
the turnover in the magnetization curves at large @xmath6 is not captured by a simple brillouin function for free or nearly free spins . through characteristic magnetic reflections * * q**@xmath35(1/2 0 1 ) and * * q**@xmath36(1/2 0 3/2 ) for a 50 mg single crystal with @xmath37 .
20 k ( @xmath38 ) data are subtracted as a background .
( b ) temperature scans of the scattering intensity at * * q**@xmath39 and * * q**@xmath9 .
wavevector scans along [ 1/2 0 l ] at ( c ) 4 k and ( d ) 0.3 k in a 20 mg single crystal with @xmath40 .
50 k data are subtracted as a background . ]
anomalies in the specific heat at @xmath16and @xmath17 mirror the downward trend with doping observed in susceptibility measurements ( fig .
away from the transition points all curves can be described by @xmath41 , an approximation of the 2d debye law characteristic of other van der waals bonded materials @xcite .
the 2d character of the materials is thus not strongly affected by ir substitution .
weak low temperature features with a magnetic origin ( indicated by magnetic field dependence ) are observed at larger @xmath6 of 0.2 and 0.3 , however , the absence of a @xmath42 - like anomaly is not consistent with a long - range ordering transition above 2 k ( see supplementary information [ si ] ) .
sr spectra for ru@xmath2ir@xmath3cl@xmath1 ( @xmath43 ) powders .
solid lines are a fit to a two - component relaxation of the form @xmath44cos@xmath45 ( @xmath46 ) and @xmath44 ( @xmath47 ) .
( b ) longitudinal field curves for @xmath48 at 1.8 k show a decoupling of the fast component in lf @xmath49 g with persisting exponential relaxation due to spin fluctuations . ] the magnetic ground state of lightly - doped crystals with @xmath50 was characterized in neutron diffraction experiments .
as in the undoped system with stacking faults @xcite we observe two magnetic propagation vectors * * q**@xmath39 = ( 1/2 0 1 ) and * * q**@xmath51 , corresponding to zig - zag order with abc and abab magnetic layer stacking , respectively , in ir - substituted single crystals measured at hb-1a ( fig .
2 ) . for @xmath37
the relative peak intensities @xmath52(**q**@xmath39)/@xmath52(**q**@xmath9 ) @xmath53 at 1.5 k show that the majority of the sample adopts the abc - type structure . the onset of scattering intensity at * * q**@xmath39 is spread over several kelvin compared to the parent compound , which could indicate distribution in @xmath16 related to compositional variation [ si ] .
the center of the transition at @xmath54 5.5 k is consistent with the location of features in @xmath55 and @xmath56 for similarly doped samples .
an extensive survey of reciprocal space using the four - circle diffractometer did not reveal additional peaks associated with alternative magnetically ordered states in the h - k-@xmath10 phase diagram .
an ordered moment of 0.32(2 ) @xmath29 in the modulated abc zig - zag phase was determined after collecting a large set of nuclear reflections to yield the scale factor for normalization of the magnetic peak intensity , approximately two - thirds of the value in pure @xmath0-rucl@xmath1 @xcite .
an increase in ir - content from @xmath37 ( fig .
2a - b ) to @xmath40 ( fig .
2c - d ) strongly affected @xmath16 as well as the relative amount of the two magnetic phases .
scans along ( 1/2 0 l ) show a single peak at l = 3/2 at 4 k ( fig . 2c ) .
a second peak at l = 1 was resolved only in lower temperature scans at 1.6 k and 0.3 k. the intensity of the peak at * * q**@xmath39 dropped precipitously with respect to * * q**@xmath9 , yielding @xmath52(**q**@xmath39)/@xmath52(**q**@xmath9 ) @xmath57 at the experimental base temperature of 0.3 k ( fig .
no additional peaks were observed at wavevectors characteristic of other candidate magnetic structures that could be accessed in the ( h 0 l ) scattering plane .
it is thus unlikely that a distinct magnetically ordered phase arises as the abc - type zig - zag order is suppressed in the dilute system .
the temperature dependence of the scattering at * * q**@xmath9 yielded @xmath58 10.2(2 ) k and 10.0(9 ) k for @xmath37 and 0.06 , respectively
. the relatively slower decline of @xmath17 with ir content compared to @xmath16 is consistent with the bulk measurements in fig . 1 .
@xmath22sr measurements were performed on ru@xmath2ir@xmath3cl@xmath1 powders to characterize the weakly magnetic state of the dilute system .
polycrystalline rucl@xmath1 undergoes a single broad transition at 14 k with abab magnetic structure @xcite ; ir - substituted powders also exhibit a single transition of the abab type that disappears near the percolation limit [ si ] .
the relaxation of the muon asymmetry @xmath59 reflects the amplitudes and fluctuations of local magnetic fields @xcite .
@xmath22sr spectra of ru@xmath2ir@xmath3cl@xmath1 with @xmath46 show only nuclear contributions to the relaxation at 20 k. a fast - relaxing component emerges below @xmath60 and well - defined oscillations signifying long - range magnetic order develop by 1.8 k ( fig .
two distinct frequencies @xmath61 mhz and @xmath62 mhz are observed , in agreement with a recent report in @xmath0-rucl@xmath1 powders that found two inequivalent muon sites in the pseudo - bilayer zig - zag phase @xcite .
in contrast , at @xmath48 no coherent muon precession frequencies that would indicate long - range order are resolved at any temperature down to 1.8 k. the 1.8 k spectrum is comprised instead of exponential terms @xmath44 , consistent with the fast - fluctuation limit of distributed internal magnetic fields @xcite .
longitudinal field experiments , in which spin depolarization due to static internal fields is suppressed , ( fig .
3b ) were used to distinguish between static and dynamic processes in the @xmath48 sample .
a static origin can be assigned to the early rapid depolarization in the zero - field spectrum ( @xmath63 9 mhz ) , which is decoupled in a longitudinal field lf @xmath49 g ( fig .
however , the persistence of the slow exponential decay term @xmath64 0.1 mhz in the lf spectra is a clear demonstration of dynamic magnetism in the fast - fluctuation limit in the remaining volume of the sample .
it is unknown whether or not these two components are also present in the single crystals .
transition temperatures @xmath16 ( red symbols ) and @xmath17 ( blue symbols ) determined from magnetic susceptibility , heat capacity , and neutron diffraction mark the boundaries of the abc - stacked ( afm1 ) and abab - stacked ( afm2 ) zig - zag phases in regions of pristine monoclinic structure and a high degree of stacking faults , respectively .
linear fitting to @xmath65 ( dashed lines ) gives a critical dilution level for @xmath66 of @xmath67 .
the cusp at @xmath17 is driven below the experimental base temperature of 2 k by @xmath68 ; at larger @xmath6 @xmath17 is estimated by extrapolation to @xmath69 ( grey symbols ) .
@xmath22sr measurements indicate a dynamic ground state in polycrystalline samples at @xmath70 with rapidly fluctuating internal fields . ]
the phase diagram of ru@xmath2ir@xmath3cl@xmath1 determined from the current study is summarized in figure 4 .
the neel temperatures of both abc and abab stacked phases are universally decreased by incorporation of nonmagnetic ir@xmath4 .
linear fits to the normalized critical temperatures give initial suppression rates @xcite @xmath71/dx = 8.1(7)$ ] for @xmath16 and 4.4(3 ) for @xmath17 .
the more sensitive response to magnetic site dilution in the abc phase is likely related to a higher degree of frustration in magnetic interactions for 3-layer magnetic stacking than for 2-layer stacking .
this disparity is evident in the different @xmath60 values for the two phases in the parent compound .
extrapolating linearly from small @xmath6 , the suppression of the zig - zag phase with abc periodicity to @xmath72 occurs for @xmath67 , with @xmath73 for abab - type . for the latter magnetic correlations
fall off more slowly at large @xmath6 and persist until the percolation threshold of the 2d honeycomb lattice is crossed .
a polycrystalline sample at this dilution level ( @xmath47 ) exhibited fast magnetic fluctuations in @xmath22sr measurements .
future careful experiments at lower temperatures near the critical dilution will show if the extinction of the long - range order reveals any critical behavior .
this phase diagram illustrates one method for suppressing magnetic order in a system exhibiting kitaev physics , and shows a possible avenue for approaching the qsl phase .
suggestive indications including the unusual behavior of the susceptibility with doping and the persistent dynamics at the percolation threshold merit further experimental and theoretical attention .
in particular , detailed low temperature spectroscopic investigations are called for to determine whether or not qsl physics is present .
* acknowledgments * p.l.k and d.m . were supported by the gordon and betty moore foundations epiqs initiative through grant gbmf4416 .
. acknowledges support from the u.s .
department of energy , office of science , office of basic energy sciences , materials sciences and engineering division .
the work at ornl hfir was sponsored by the scientific user facilities division , office of basic energy sciences , u.s .
department of energy .
we thank iain mckenzie for local support at triumf , bryan chakoumakos and feng ye for assistance with single - crystal xrd , and brian sales , michael mcguire , and andrew may for helpful discussions during the course of this research .
71ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @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 * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( )
single crystals of ru@xmath2ir@xmath3cl@xmath1 were grown by a transport method at high temperatures ( 900 - 1100@xmath74c ) from commercial powders of rucl@xmath1 ( furuya metals ) and ircl@xmath1 ( american elements ) in the presence of nacl .
single crystal diffraction ( rigaku , mo k@xmath0 radiation ) was performed for the compositions @xmath46 and @xmath75 on small crystals with @xmath76 m in - plane dimension .
the structure was refined using shelx to the monoclinic c2@xmath77 m space group , consistent with similarly - sized rucl@xmath1 single crystals .
individual planes of the larger ( @xmath54 mm ) sized plate - like crystals used in squid and heat capacity measurements were checked by x - ray diffraction on a panalytical xpert pro powder diffractometer.indexed in the trigonal p3@xmath3912 setting , the out - of - plane lattice parameter of these crystals as determined from the ( 0 0 l ) peaks decreases linearly with increasing ir substitution ( fig .
ru@xmath2ir@xmath3cl@xmath1 was prepared by direct reaction of the commercial starting materials in an evacuated quartz ampoule at 800@xmath74c .
xrd patterns ( fig .
s1b ) show strong ( 0 0 l ) texturing in a fraction of the powder samples , and warren lineshapes are characteristic of the quasi-2d nature of the system . minority oxide and binary metal phases ( @xmath78 wt .
% ) identified in the powder patterns likely indicate a partial decomposition of the starting materials on heating .
the c - axis lattice parameters determined from xrd are in good agreement with those of the single crystal samples while the in - plane lattice parameter grows systematically with incorporation of ir ( fig .
s1a ) . ir@xmath3cl@xmath1 single crystals ( filled symbols ) and powders ( half - filled symbols ) in the trigonal setting .
lines are a guide to the eye .
( b ) offset x - ray diffraction patterns of ru@xmath2ir@xmath3cl@xmath1 powders .
average ir content and distribution is determined by eds measurements in single crystals ; ir doping in the powdered materials is taken as the nominal value . ]
the elemental compositions of the ru@xmath2ir@xmath3cl@xmath1 single in this study were checked by energy dispersive x - ray spectroscopy ( eds ) .
variation of ru@xmath77ir ratio over @xmath54 mm length scales was noted in samples large enough for bulk measurements ; the distribution of @xmath6 was characterized in each sample by measuring a large number of sites spanning the crystal with a spot size of @xmath79 m .
representative histograms for twelve crystals are shown in fig .
standard deviation from the mean value of @xmath6 ranged from 0.3 mol % ir to 2 mol% ir , but remained less than 1% in nearly all of @xmath2120 samples tested .
ir@xmath3cl@xmath1 as determined by eds measurements .
solid lines represent a gaussian fit to the distribution in each sample . ]
magnetic transition temperatures were estimated either from the cusp ( @xmath80 ) in the temperature - dependent susceptibility curves or by linear extrapolation of data above and below the turnover in @xmath55 where @xmath17 appears as a shoulder no well - defined cusp exists ( e.g. @xmath17 in 2-phase samples or large @xmath6 ) , as in fig .
s3(a - b ) .
critical temperatures determined in this way are in good agreement with heat capacity and neutron diffraction measured in select samples ( fig .
s3c - f ) .
@xmath81 to estimate @xmath17 below the experimental base temperature of 2 k. ( c ) offset heat capacity curves for various ir substitution levels .
arrows indicate @xmath16 and @xmath17 ( d ) intensity vs. temperature for the ( 1/2 0 1 ) magnetic bragg peak showing @xmath16 in @xmath82 and @xmath83 crystal .
the doped crystal shows a smearing of the transition compared to the parent compound and intensity is not well described a power law .
dashed line is a guide to the eye .
solid line is a fit to @xmath84 for @xmath82 .
( e),(f ) intensity vs. temperature for the ( 1/2 0 3/2 ) magnetic bragg peak indicating @xmath17 for ( e ) @xmath83 and ( f ) @xmath85 single crystals .
solid lines are a power law fit . ]
to check for the presence of spin - glass - like freezing and/or thermal hysteresis at low temperatures , susceptibility was measured in a magnetic field of 1 t under zero - field - cooled ( zfc ) , field - cooled cooling ( fcc ) and field - cooled warming ( fcw ) protocols for several compositions spanning the range investigated .
no splitting was observed between the zfc / fc curves in fig .
s4(a ) or in fcc / fcw curves ( not shown ) at this field that would indicate frozen moments or a strongly first - order phase transition , respectively .
ir@xmath3cl@xmath1 single crystals collected under zero - field - cooled warming ( zfc ) and field - cooled cooling ( fc ) protocols with a magnetic field of 1 t applied in plane .
( b ) susceptibility curves , normalized to @xmath55 ( 20 k ) for three ir - doped crystals from the same batch indicating a variation of the abc and abab magnetic phase fraction between samples , as characterized by the relative prominence of @xmath16 and @xmath17 . ]
the presence of two magnetic phases in rucl@xmath1 is closely tied to structural details ( see discussion in [ 1 ] ) and thus their volume fraction can be expected to vary sensitively with stacking faults , strain , etc . in general we observe both magnetic transitions in ru@xmath2ir@xmath3cl@xmath1 crystals , however the relative prominence of @xmath16 and @xmath17 transitions in heat capacity and susceptibility ( fig .
s3 ) indicate that the volume fractions of these phases vary from one sample to the next .
s4(b ) shows susceptibility curves for three as - grown doped crystals from the same batch , qualitatively showing ( i ) dominant abc phase , ( ii ) comparable contributions , and ( iii ) dominant abab phase .
that similarly doped samples ( i ) and ( iii ) exhibit nearly single - phase abc and abab characteristics , respectively , indicates that ir - substitution does not necessarily lead to the abab - type phase ( at least for low substitution levels @xmath86 ) .
furthermore the proximity of @xmath16 and @xmath17 in the single and multiphase ( ii ) samples in fig .
s4(b ) suggests that ir is not segregated in regions with abab phase .
the multiphase tendency in the doped crystals is thus most likely a result of the modified growth processes necessary to incorporate ir in transported crystals .
the anomalies at @xmath16 and @xmath17 in heat capacity curves were not affected by a magnetic field applied perpendicular to the @xmath88 plane ( the only configuration available in our experimental setup ) for small @xmath6 ( fig .
s5a - c ) , in agreement with a previous study on the parent compound , in which magnetic field dependence of the anomalies was only observed for @xmath89 [ 2 ] . at large substitution levels
@xmath90 and 0.3 , where susceptibility measurements indicate the magnetic transitions are suppressed below 2 k , an upturn is visible in @xmath91 vs. @xmath7 curves approaching the base temperature of 2 k ( fig .
s6d - e ) , and @xmath56 departs from a nearly @xmath92 behavior below @xmath54 5 k ( fig .
s5f ) . magnetic field partially suppresses this upturn ( insets , fig .
s5 ) , indicating that it has a magnetic origin .
the feature may be related to short - range magnetic correlations , or represent the start of a very broad transition centered below 2 k. the emergence of field dependence with @xmath93 points to a possible change in the magnetic anisotropies of the parent compound at large @xmath6 , however a more detailed study with variation of the applied field direction in the crystals is required to explore this point . vs. @xmath7 for in zero field and with an 8 t field applied perpendicular to the honeycomb plane in ru@xmath2ir@xmath3cl@xmath1 crystals with ( a ) @xmath94 , ( b ) @xmath95 , ( c ) @xmath96 , ( d ) @xmath90 , and ( e ) @xmath47 .
insets are magnifications of the low - temperature region .
( f ) @xmath56 vs. @xmath7 for larger doping @xmath90 and @xmath47 showing a departure from @xmath41 behavior below @xmath54 6 k where curves exhibit field dependence . ]
field - dependent magnetization curves for ru@xmath2ir@xmath3cl@xmath1 single crystals at 2 k are shown in fig .
s6 over several doping ranges .
magnetization initially decreases with x up until @xmath54 10% substitution and retains an upward curvature ( fig .
s6a ) . in the range @xmath97 magnetization increases and a switch in the concavity of @xmath98 occurs , with an approximately linear field dependence in the @xmath99 sample ( fig .
s6b ) . at higher dopings ,
@xmath98 is concave down and magnetization decreases continuously ( fig .
we note that in the parent compound rucl@xmath1 , field - dependent magnetization curves in the ordered state switch from concave up to concave down around @xmath100 t , shown in the inset of fig .
the ir - induced curvature change in the substituted samples could thus be explained by a downward shift in the critical field of the parent compound , with @xmath101 for @xmath67 , or alternatively by additional contributions to the magnetization from vacancy - induced moments in the kitaev framework .
ir@xmath3cl@xmath1 crystals with ( a ) @xmath102 , ( b ) @xmath103 , and ( c ) @xmath104 .
data were collected on two squid instruments with maximum magnetic fields of 5 t and 7 t. inset : high - field dc magnetization curve for @xmath82 measured on a 14 t ppms with an acms insert . ]
relaxation of the muon asymmetry due to nuclear moments in a polycrystalline sample is given by the well- known kubo - toyabe function , @xmath105\ ] ] where @xmath106 is the width of the local distribution of static fields . at 20 k in the ru@xmath2ir@xmath3cl@xmath1 powders ,
weak gaussian relaxation can be described by eqn .
1 , with @xmath107 and 0.095(9 ) @xmath22s@xmath108 for@xmath109 and @xmath47 , respectively . in the sample with @xmath46 ,
a fast - relaxing component emerges below @xmath60 ( see fig .
s8a ) at 11 k and well - defined oscillations develop by 1.8 k. the asymmetry is generally modeled as @xmath110\ ] ] where @xmath111 represent the fast - relaxing / oscillating components and @xmath112 is a slowly relaxing exponential tail .
the fitted parameters are given in table s1 .
at 1.8k , the total oscillating volume fraction @xmath113 is close to the expected value of 2/3 for a fully ordered long - range magnetic system . in the @xmath47 sample , two - component gaussian plus fast exponential spectra
are observed at 11 k and 5 k , modeled as @xmath114\ ] ] and a two - component slow exponential plus fast exponential spectrum at is observed at base temperature given by @xmath115.\ ] ] the fitted parameters are given in table s2 .
ir@xmath3cl@xmath1 powders with @xmath116 t. zero - field @xmath22sr spectra at various temperatures in ru@xmath2ir@xmath3cl@xmath1 powders with ( b ) @xmath46 and ( c ) @xmath48 .
solid lines are fits to relaxation models described in the text . ]
[ b ] .fitted parameters of the @xmath22sr spectra of polycrystalline ru@xmath2ir@xmath3cl@xmath1 with @xmath46 , see eqn . ( 2 ) . [ cols="^,^,^,^,^,^,^,^,^ " , ]
[ 1 ] h.b . cao et al . , phys . rev .
b 93 , 134423 ( 2016 ) . | quantum fluctuations in a 2d honeycomb kitaev model completely suppress long - range order to produce a quantum spin liquid ( qsl ) ground state . while @xmath0-rucl@xmath1 is positioned on the verge of the kitaev qsl phase , fragile long - range magnetic order arises from non - kitaev terms in the hamiltonian . in a study of ru@xmath2ir@xmath3cl@xmath1 single crystals
, we find that spin vacancies in the form of nonmagnetic ir@xmath4 successfully destabilize magnetic order in @xmath0-rucl@xmath1 .
neutron diffraction and bulk characterization demonstrate a suppression of the primary zig - zag magnetic phase transition in ru@xmath2ir@xmath3cl@xmath1 single crystals to below 2 k by @xmath5 , potentially paving the way towards the sought - after pure kitaev limit .
stacking disordered regions exhibit less sensitivity to dilution , and sustain weak zig - zag correlations to larger @xmath6 , near the percolation threshold for the honeycomb lattice . the phase diagram as a function of @xmath6 and @xmath7
is determined . |
the theory of lyapunov exponents of hard - ball systems has a long history .
it started with the pioneering work of krylov @xcite , was rigorously developed by sinai @xcite and collaborators , and completed ( to some extent ) by van beijeren , dorfman and co - workers @xcite .
the analytical calculation of , e.g. , the largest lyapunov exponent of a dilute rigid - sphere gas , is based on the fact that the dynamics consists of free rectilinear motions interrupted by instantaneous elastic collisions @xcite ; the expressions so - obtained agree quantitatively with the numerical experiments @xcite .
the case of a dilute gas with finite - range interactions can be handled in close analogy with the rigid - sphere problem : though the collisions are not trivial any more , the dynamics is still ruled by occasional pairwise encounters @xcite .
however , when one considers long - range interactions ( or short - range interactions and high densities ) , the theoretical approach must be substantially modified .
in the general case we must deal with the full system of coupled differential equations that govern the evolution of multidimensional tangent vectors @xmath2 .
consider for instance a gas of @xmath3 particles in three dimensions described by the hamiltonian @xmath4 where @xmath5 and @xmath6 , are conjugate position - momentum coordinates . assuming @xmath7 , tangent vectors evolve according to @xmath8 ( dot meaning time derivative ) , where @xmath9 is the hessian matrix of the potential @xmath10 , namely @xmath11 the hessian depends explicitly on time because it is calculated along a reference trajectory @xmath12 .
once initial conditions @xmath13 and @xmath14 have been specified , one can find @xmath2 from eq .
( [ tangent ] ) .
then the lyapunov exponent @xmath0 is obtained by calculating the limit @xcite @xmath15 assuming ergodicity on the energy - shell , @xmath0 becomes independent of initial conditions @xmath16 , which can then be chosen randomly according to the microcanonical distribution .
there will also be no dependence on initial tangent vectors , because if @xmath14 is also chosen randomly , it will have a non - zero component along the most expanding direction .
it is this average over @xmath16 and @xmath14 that permits to treat equations ( [ tangent ] ) formally as a system of stochastic differential equations @xcite .
moreover , if the dynamics can be thought of as free motion plus weak interactions , then perturbative techniques , like the cumulant expansion @xcite , can be invoked .
so , the theory attempts to calculate the average @xmath17 however , in practice , it is much simpler to develop an estimate for the generalized lyapunov exponent @xcite @xmath18 this is essentially the approach followed by barnett et al @xcite , pettini et al @xcite , and the present authors @xcite . in situations of weak intermittency both exponents are expected to be close . if one wishes to use a theoretically calculated @xmath1 as an approximation to @xmath0 , then a numerical check must be done first to verify that both exponents coincide . the cumulant expansion to be discussed below offers an analytical expression for @xmath0 via the replica trick .
note , however , that the difficulties involved in such a calculation are much greater than those we shall face when dealing with @xmath1 .
though there are some differences among the formulations of the three just - mentioned groups , it may be said that the main theoretical conclusion extracted from that body of work is : if one combines the cumulant expansion with some kind of isotropy approximation ( which may be fully justified in some cases ) , the original problem of @xmath19 differential equations can be reduced to a system of only two equations for a representative " single degree of freedom : @xmath20 in this kind of mean - field approximation , the curvature " @xmath21 is a scalar stochastic process , whose cumulants can be related to the ( operator ) cumulants of the hessian @xmath22 ( see , e.g. , @xcite ) .
the comparison of theoretical results obtained with the cumulant approach versus numerical simulations has met mixed success .
the agreement is very good for a many - particle system with bounded weak interactions @xcite and for the fermi - pasta - ulam system @xcite .
however , the results for the 1d - xy model @xcite , for a dense one - component plasma @xcite , and for a dilute lennard - jones gas @xcite are not so satisfactory .
the purpose of this paper is to investigate the limits of validity of the cumulant approach for the lyapunov exponent of a many particle system .
we choose as a starting point the simplified mean - field setting ( [ tangent2 ] ) and consider two possibilities for @xmath21 .
it has been argued @xcite that , for typical chaotic many - body systems , @xmath21 should be close to gaussian white noise ; this is the first case we shall consider . in the white - noise case
the second - order expansion for @xmath1 is exact , thus this case is ideally suited for analyzing the difficulties that appear in the _ numerical _ calculation of @xmath1 .
next , we keep the gaussian and markov properties but allow for finite correlation times , leading to the ornstein - uhlenbeck process ; in this case we calculate the fourth cumulant contribution to @xmath1 .
though it will not be considered here , we also mention the interesting situation of @xmath21 being a poisson process , which appears to be the appropriate choice for modeling the tangent - vector dynamics in a dilute gas with short - range interactions .
formally , eq . ( [ tangent2 ] ) describes a harmonic oscillator with a random frequency @xmath23 such that @xmath24 ( kubo oscillator ) .
it is worth generalizing this model a bit to account for the possibility of damping , i.e. , we shall consider an oscillator described by the dynamical equation @xmath25 setting @xmath26 , @xmath27 , @xmath28 , recovers ( [ tangent2 ] ) . some analytical results for the lyapunov exponent of the kubo oscillator ( [ kuboosc ] ) can be found in the literature ( see , e.g. , @xcite ) .
here we shall restrict ourselves to the analytical calculation of the generalized exponent @xmath1 .
for this purpose we must consider the dynamics of second moments : @xmath29 let us think that , in principle , both parameters @xmath30 and @xmath31 are stationary stochastic processes . if fluctuations are small enough ( in a sense that will be discussed later ) , one can obtain the average of the second - moment vector using the first terms of the cumulant expansion , which works as follows @xcite .
first we split the stochastic matrix as an average plus fluctuations : @xmath32 for long times one has : @xmath33 where @xmath34 is the @xmath35 matrix given by the operator cumulant expansion @xcite @xmath36 dots stand for third and higher cumulants ( some explicit expressions can be found in @xcite ) .
the exponent @xmath1 is related to the eigenvalue of @xmath34 that has the largest real part : @xmath37 with @xmath38 the eigenvalues of @xmath39 .
when the entries of the fluctuation matrix @xmath40 are gaussian white noise ( and only in this case @xcite ) the cumulant expansion stops at the second order , i.e. , eq .
( [ cum ] ) is exact ( without the ellipsis ) .
this is the case we consider now .
let us first study the situation where the damping @xmath30 is a constant and @xmath41 where @xmath42 is a zero - mean gaussian white noise .
its correlation function reads @xmath43 with these definitions one has @xmath44 after substitution into eq .
( [ cum ] ) we readily obtain @xmath45 the generalized exponent @xmath1 can now be calculated from eq .
( [ lambdastar ] ) . a closed expression for the standard lyapunov exponent can be found in the literature @xcite . as an example , fig .
[ fig : rf0 ] displays analytical results for both exponents .
we also show the outcomes of numerical simulations . given that the theoretical expressions are exact , fig . [
fig : rf0 ] constitutes a test for our numerical calculations .
numerical details , including a discussion about the difficulties found in the calculation of @xmath1 , will be presented in sec .
[ sec : diff ] .
harmonic oscillator with random frequency . shown
are the lyapunov exponents @xmath0 and @xmath1 as a function of the noise strength @xmath46 .
solid lines correspond to the theoretical expressions given by inserting ( [ k_rf ] ) into ( [ lambdastar ] ) , and in ref .
@xcite for @xmath1 and @xmath0 , respectively .
symbols are the results of numerical simulations ( averaged over @xmath47 trajectories ) .
we chose @xmath26 and @xmath48 .
, scaledwidth=45.0% ] now we consider an harmonic oscillator with constant frequency but in an environment with fluctuating damping coefficient @xmath49 where @xmath42 is also in this case a zero - mean gaussian white noise .
the corresponding stochastic differential equation ( [ kuboosc ] ) will be taken in stratonovich sense .
therefore , the matrix @xmath50 in eq .
( [ b ] ) can be decomposed as @xmath51 hence , substitution into ( [ cum ] ) yields @xmath52 from the eigenvalues of @xmath39 we obtain @xmath1 following eq .
( [ lambdastar ] ) . a theoretical expression for @xmath0 can be found in ref .
[ fig : rd0 ] exhibits numerical and analytical results for both exponents , as a function of the noise intensity . harmonic oscillator with random damping .
shown are the lyapunov exponents @xmath0 and @xmath1 as a function of the noise strength @xmath46 .
solid lines correspond to the theoretical expressions given by ( [ lambdastar],[k_rd ] ) and ref .
@xcite , for @xmath1 and @xmath0 , respectively .
symbols are the results of numerical simulations ( averaged over @xmath47 trajectories ) .
we chose @xmath53 and @xmath54 . , scaledwidth=45.0% ] note that in the cases considered above @xmath0 and @xmath1 do not coincide .
we have checked that the difference between them ( which is a quantifier of the degree of intermittency of the dynamics ) may be controlled by suitable choice of the parameters of the oscillator .
we preferred to consider intermittent cases , because it is in these regimes that the numerical difficulties arise , as we discuss in the following section .
we remark that there are situations of interest where both exponents practically coincide , e.g. , for a dilute lennard - jones gas @xcite . in such cases
a theory capable of estimating @xmath1 will also produce a good estimate for the standard lyapunov exponent @xmath0 .
numerical simulations were performed by means of the euler algorithm with time step @xmath55 . for each trajectory
we computed the norm @xmath56 as a function of time @xmath57 .
the lyapunov exponent is approximated by the average over initial conditions of the finite - time exponents : @xmath58 where @xmath57 is large enough to guarantee the convergence of the average to the desired precision . in order to obtain the asymptotic value of the generalized exponent ( [ liapstarav ] ) , in principle
, one must calculate the squared - norm averaged over several realizations at a given large time .
however , we must keep in mind that such an average is dominated by the extreme positive values of the local exponent @xmath59 .
hence , direct averaging over @xmath60 may yield spurious results whenever the variance of @xmath59 fails to vanish with time fast enough . to avoid this problem , instead of the straightforward averaging , we preferred to estimate the local generalized exponent from the cumulants of the distribution of @xmath59 : @xmath61 where @xmath62 are the @xmath63th - order cumulants of the distribution of local exponents @xmath59 .
[ fig : diff ] illustrates , for the white - noise random frequency oscillator , the @xmath64 as a function of time ( first - order truncation of ( [ local ] ) ) , as well as the expansion ( [ local ] ) truncated at the second and third orders .
for comparison , also plotted is the crude estimate ( [ liapstarav ] ) . clearly , the expansion ( [ local ] ) has to be considered in order to properly estimate @xmath1 . in figs .
[ fig : rf0 ] and [ fig : rd0 ] , @xmath1 was numerically computed from the third - order truncation , because the next ( noisier ) terms do not contribute significantly .
numerical difficulties in the calculation of @xmath1 .
we plot the finite - time exponent @xmath65 as a function of time ( red line ) for the random frequency oscillator with @xmath26 and @xmath66 .
averages were computed over @xmath67 realizations . also plotted
are the corrections arising from the second and third cumulants of the distribution of finite - time lyapunov exponents eq .
( [ local ] ) ( blue ) . for comparison
we also show the straightforward average @xmath68 ( light gray ) .
dashed lines correspond to the theoretical asymptotic values .
, scaledwidth=45.0% ]
for white noise fluctuations , either in the frequency or in the damping , we have verified in sec . 3 ( see figs . [
fig : rf0 ] and [ fig : rd0 ] ) , that our theory for @xmath1 is in agreement with numerical results , provided the later are obtained by means of the procedure described in the preceding section . the analysis in sec .
3 also allows to quantify the discrepancy between @xmath0 and @xmath1 , which typically increases with increasing amplitude of the fluctuations .
now we shall analyze the effect of introducing noise correlations .
we consider again the case of a random frequency , as in eq .
( [ kappa ] ) , but now the noise is a zero - mean ornstein - ulhenbeck process , i.e. , with correlation function @xmath69 for simplicity we set @xmath26 and @xmath70 . by inserting eq .
( [ b_rf ] ) into eq .
( [ cum ] ) , the second - cumulant matrix @xmath71 becomes @xmath72 notice that in the limit @xmath73 the white - noise case is recovered . in the presence of correlations the second - order truncation of the cumulant expansion ( [ cum ] ) is not exact . in order to improve the theory one must calculate higher cumulants .
for the present case the third cumulant is null .
explicit expressions for the fourth cumulant were given by fox @xcite and by breuer et al @xcite .
so , the fourth cumulant can be calculated without great effort ( with the aid of algebraic manipulation programs ) .
the fourth order approximation to @xmath39 reads @xmath74 the comparison between the theoretical results for @xmath1 ( with the second ( blue ) and fourth ( dark blue ) order corrections ) and numerical outcomes is shown in fig .
[ fig : rfou ] . notice that in numerical estimates the fourth - order correction is very small in comparison with the third - order one , suggesting that the cumulant expansion is rapidly converging .
harmonic oscillator with correlated random frequency .
we show @xmath0 and @xmath1 as a function of the noise amplitude @xmath46 .
solid lines correspond to the theoretical results obtained from ( [ lambdastar],[k_2 ] ) ( blue ) and ( [ lambdastar],[k_4 ] ) ( dark blue ) for @xmath1 and to the approximate expression ( using the decoupling ansatz ) following ref .
@xcite for @xmath0 .
symbols are the results of numerical simulations ( averaged over @xmath67 trajectories ) , corresponding to the second ( circles ) , third ( squares ) and fourth ( triangles ) order corrections of ( [ local ] ) .
parameters are @xmath26 , @xmath70 and @xmath75 . , scaledwidth=45.0% ]
the perturbation parameter controlling the convergence of the cumulant expansion is the so - called kubo number @xmath76 .
general considerations led van kampen @xcite to conclude that the kubo number is the product of the amplitude of the fluctuations and the correlation time , that is @xmath77 . however , in the present case it is clear that such a combination is not adimensional .
the correct kubo number is instead @xmath78 this can be checked explicitly from the second and fourth cumulants above .
consider , for instance , the element @xmath79 , which dominates the lyapunov exponent for small correlation times : @xmath80 in the white - noise limit , i.e. , @xmath81 with @xmath46 fixed , the kubo number tends to zero as it should be .
we have taken the first step towards the application of the cumulant expansion to calculate the largest lyapunov exponent of a dilute gas .
the case of white - noise fluctuations ( either in the frequency or in the damping ) was considered first .
this study was very useful to understand the difficulties behind the numerical calculation of the generalized exponent @xmath1 .
it was verified that @xmath1 can be obtained with a satisfactory precision by using the cumulant expansion for the distribution of the finite - time lyapunov exponent .
we also analyzed briefly the case of correlated noise .
for the ornstein - uhlenbeck noise we were able to obtain the fourth cumulant contribution to the analytical @xmath1 , which showed an improvement with respect to the second order truncation , when compared with numerical outcomes . moreover , we showed that the correct perturbative parameter for the present problem is the product @xmath82 , and not @xmath77 , as a literal reading of van kampen s discussion @xcite would suggest .
it is expected that the present results will be helpful for the correct application of the cumulant approach in higher dimensionality systems , as well as for the numerical checking of its validity .
we acknowledge brazilian agencies faperj and cnpq for partial financial support . + | we consider a one - dimensional harmonic oscillator with a random frequency , focusing on both the standard and the generalized lyapunov exponents , @xmath0 and @xmath1 respectively . we discuss the numerical difficulties that arise in the numerical calculation of @xmath1 in the case of strong intermittency .
when the frequency corresponds to a ornstein - uhlenbeck process , we compute analytically @xmath1 by using a cumulant expansion including up to the fourth order
. connections with the problem of finding an analytical estimate for the largest lyapunov exponent of a many - body system with smooth interactions are discussed . |
the problem of multiple hypothesis testing in the setting of quantum channels can be formulated as follows .
assume that @xmath0 is an unknown quantum channel , but some a priori information is available , in the sense that @xmath0 is one of given channels @xmath1 , with probabilities @xmath2 .
the task is to find a procedure that determines the true channel , with the greatest possible probability of success . for quantum states ,
this problem was formulated by helstrom @xcite and since then has been the subject of active research , see e.g. @xcite for an overview and further references . here , an ensemble @xmath3 is given , where @xmath4 are quantum states with prior probabilities @xmath2 , with a similar interpretation as above .
a testing procedure , or a measurement , for this problem is described by a positive operator valued measure ( povm ) @xmath5 , defined as a collection of positive operators @xmath6 summing up to the identity operator @xmath7 . the value @xmath8 is interpreted as the probability that the procedure chooses @xmath9 while the true state is @xmath10 .
the task is to maximize the average success probability @xmath11 over all povms . in the case
@xmath12 , it is well known that the optimal povm is projection valued , given by the projections onto the positive and negative parts of the operator @xmath13 , @xcite . for @xmath14
, there is no explicit expression for the optimal povm in general , but it is known that a povm @xmath5 is optimal if and only if it satisfies @xmath15 this condition was obtained in @xcite using the methods of semidefinite programming . in the case of quantum channels ,
a most general measurement scheme is described by a triple @xmath16 , where @xmath17 is an ancilla , @xmath18 a ( pure ) state on @xmath19 and @xmath20 is a povm on @xmath21 . for @xmath22 , the value @xmath23 is interpreted as the probability that @xmath24 is chosen when the true channel is @xmath25 .
the average success probability is then @xmath26 the task is to maximize this value over all triples @xmath16 .
it was observed @xcite that using entangled input states may give greater success probability and it was shown in @xcite that every entangled state is useful for some channel discrimination problem . however , there are situations when e.g. the maximally entangled input state does not give an optimal success probability .
it is therefore important to find out whether an optimal scheme with a given input state exists . in the broader context of generalized decision problems , conditions for existence of an optimal scheme with an input state having maximal schmidt rank
were found in @xcite , a related problem was studied in @xcite . in the present paper , we show that these conditions can be obtained using the methods of semidefinite programming .
such methods were already applied before in the context of discrimination of quantum channels , see @xcite . compared to these works , we are more concerned with the choice of an optimal input state .
it is an easy observation that if a given scheme @xmath27 is optimal , then @xmath5 must be an optimal measurement for the ensemble @xmath28 .
we show that , at least in the case that the input state is assumed to have maximal schmidt rank , the optimality condition for a channel measurement can be divided into the condition for this ensemble and an additional condition that ensures optimality of the input state .
if the schmidt rank of the input state is not maximal we obtain a comparably weaker result , but show an example where the use of such an input state is required . as an important special case
, we get a necessary and sufficient condition for existence of an optimal scheme with a maximally entangled input state .
if this condition is not fulfilled , we give an upper bound on the optimal success probability . for discrimination of two channels , we use the known form of an optimal povm for two states to obtain a relatively simple condition in terms of choi matrices of the involved channels , which we call the condition .
we also derive an upper bound on the diamond norm , which is tighter than the previously known bound given e.g. in @xcite , see remark [ rem : bounds ] below .
the results are applied to discrimination of covariant channels , qubit channels , unitary channels and simple projective measurements .
the paper is organized as follows : in the next section we rewrite the problem as a problem of sdp from which we obtain necessary and sufficient conditions for optimal solution and derive an upper bound on the optimal success probability . in section [ sec : two ] we investigate the condition and the related bounds .
it the last two sections , we study special cases of channels and present some examples demonstrating the results .
let @xmath29 be a finite dimensional hilbert space .
we denote by @xmath30 the set of positive operators and by @xmath31 the set of states , that is , positive operators of unit trace . a completely positive trace preserving map @xmath32
is called a channel , we will denote the set of all channels by @xmath33 .
any linear map @xmath32 is represented by its choi matrix @xmath34 , defined in @xcite as @xmath35 here @xmath36 is a fixed orthonormal basis of @xmath29 . note that @xmath37 if and only if @xmath38 is positive and @xmath39 .
an alternative description of a channel measurement is given in terms of process povms @xcite ( or testers @xcite , see also @xcite ) .
a process povm is a collection @xmath40 of positive operators in @xmath41 with @xmath42 for some state @xmath43 . for any triple @xmath16
, there is a process povm @xmath44 such that for all @xmath45 and @xmath46 , @xmath47 conversely , for any process povm @xmath44 , one can find some ancilla @xmath17 , a pure state @xmath18 and a povm @xmath5 such that ( [ eq : ppovm ] ) holds @xcite . to see this ,
let @xmath48 , for @xmath49 and observe that by schmidt decomposition , we have @xmath50 for some unitary operator @xmath51 and @xmath52 for every @xmath53 . denoting @xmath54 we get @xmath55 since the channel in acts only on the first part of the system we get @xmath56 and holds with @xmath57 .
conversely , let @xmath58 for @xmath43 .
let @xmath59 be defined on the support of @xmath60 and 0 elsewhere , then @xmath61 is a povm on @xmath62 where now @xmath63 and holds as before . using the description by process povms , we will show that the maximization of the success probability can be written as a problem of semidefinite programming : @xmath64 here @xmath65 , @xmath66 is the canonical basis of @xmath67 and @xmath68 is any basis of the ( real ) linear subspace @xmath69 to see this , note that according to , equation can be rewritten as : @xmath70 put @xmath71 .
then : @xmath72 and the problem of maximizing @xmath73 can be understood as the problem of maximizing @xmath74 .
we have @xmath75 and from @xmath76 it follows @xmath77 .
note also that since @xmath78 is block - diagonal , we may extend the maximization over all positive elements @xmath79 with @xmath80 , @xmath43 ( and not only over block - diagonal ones ) . to rewrite this to the more usable form stated above
, we need to note that @xmath81 with @xmath82 if and only if @xmath83 for all @xmath84 and @xmath85 . to prove this statement ,
let us first assume that @xmath81 , then for any @xmath86 , @xmath87 and @xmath88 .
conversely , assume that @xmath83 for all @xmath89 and @xmath85 .
consider @xmath90 as a hilbert space with hilbert - schmidt inner product , then there is an orthonormal basis @xmath91 in @xmath90 , where each @xmath92 is a self - adjoint operator such that @xmath93 . with respect to this basis ,
each @xmath84 can be expressed as : @xmath94 with some @xmath95 . from the condition
@xmath96 we obtain @xmath97 . expressing @xmath98 and using the condition
@xmath83 we get @xmath99 since there is no restriction on @xmath100 for @xmath101 , we must have @xmath102 for all @xmath101 , and hence @xmath103 . to conclude the proof , from the condition @xmath85 we get @xmath104 .
moreover , it is worth realizing that from the condition @xmath77 we get @xmath105 , hence @xmath106 .
the following result is obtained using standard methods of semidefinite programming ( see e.g. @xcite ) .
the expression for maximal success probability was obtained also in @xcite , in a more general setting .
[ thm : opt_cond ] let @xmath107 be a process povm . then @xmath107 is optimal if and only if there is some @xmath108 and some @xmath109 , such that for all @xmath53 , @xmath110 and @xmath111 moreover , in this case , the maximal success probability is @xmath112 as first , we will formulate the dual problem .
let @xmath113 , @xmath114 be some basis of @xmath115 and let @xmath116 , then dual problem is : @xmath117 let @xmath118 be dual feasible , then since @xmath119 and @xmath120 , we must have @xmath121 .
if we denote @xmath122 , then @xmath123 and from @xmath124 we obtain @xmath125 .
hence there is some channel @xmath126 , such that @xmath127 . from the condition
@xmath128 we obtain @xmath129 for all @xmath53 .
from here we see , that the dual problem may be formulated as : @xmath130 now let @xmath131 , then @xmath132 is a primal feasible plan .
moreover @xmath132 belongs to the interior of the cone of positive operators , therefore by slater s condition we obtain that the duality gap is zero , in other words @xmath133 or @xmath134 , where by @xmath135 we denote the primal optimal plan and by @xmath136 we denote the dual optimal plan . since @xmath135 is feasible , we have @xmath137 and we get : @xmath138 as @xmath139 and @xmath140 , the sum may be zero if and only if all summands are zero .
moreover , trace of the product of two positive matrices is zero if and only if their product is zero . to see this
let @xmath141 and @xmath142 .
we have @xmath143 and hence @xmath144 and @xmath145 . by the above argumentation
, we get from @xmath146 on the other hand , the condition must hold for any dual feasible plan , but if for some primal and dual feasible plans the condition holds , then the duality gap for these plans is zero and they are optimal .
this concludes the proof . using this result
, we can characterize optimality of measurement schemes with input states of maximal schmidt rank .
[ coro : optimcond ] let @xmath147 be a pure state such that @xmath148 is invertible .
then a measurement scheme @xmath149 is optimal if and only if a. @xmath150 majorizes @xmath151 for all @xmath53 [ item : optimcondmajor ] b. @xmath152 .
[ item : optimcondptr ] let @xmath48 and let @xmath153 be the operator such that @xmath154 , so that the process povm corresponding to @xmath149 is given by @xmath155 , see . note that by our assumptions , @xmath156 is invertible , @xmath157 and @xmath158 .
assume that @xmath107 is optimal , then by theorem [ thm : opt_cond ] , there must be some @xmath159 and @xmath109 such that @xmath160 and @xmath161 summing up over @xmath53 , we obtain @xmath162 multiplying the above equality by @xmath163 from the left and by @xmath164 from the right , we get using the above expression for @xmath165 , @xmath166 the two conditions follow easily from this equality . assume conversely that the conditions , are satisfied .
put @xmath167 , then ( i ) and ( ii ) imply that @xmath168 and @xmath169 .
it follows that there is some positive number @xmath170 and @xmath109 such that @xmath171 .
moreover , implies that @xmath160 for all @xmath53 and @xmath172 it follows that @xmath173 and this implies the optimality condition of theorem [ thm : opt_cond ] , exactly as in its proof .
note that is the optimality condition for a povm in discrimination of the ensemble @xmath174 , where @xmath175 .
in other words , if @xmath176 is an optimal povm for this ensemble and @xmath177 the majorization @xmath178 is satisfied .
it follows that the existence of an optimal scheme with the given input state is equivalent to the condition ( ii ) .
clearly , in this case , @xmath179 is the optimal scheme and the optimal success probability is @xmath180 .
next , we show that the conditions of corollary [ coro : optimcond ] are necessary for a general pure input state .
[ coro : notfullrankstate ] let @xmath147 be a pure state such that @xmath148 .
then a measurement scheme @xmath149 is optimal only if the the conditions , from the previous corollary hold . we will show that the measurement scheme is optimal for some problem with reduced input space .
let us denote by @xmath181 the support of @xmath182 .
since @xmath18 is pure , it must be of the form @xmath183 for some schmidt decomposition of @xmath184 .
from here we see that @xmath185 , where @xmath186 is a subspace isomorphic to @xmath181 .
let @xmath187 be the restriction of @xmath24 to @xmath188 and let @xmath189 be the projection onto @xmath190 , then it is clear that @xmath191 , moreover , @xmath192 defines an optimal measurement scheme for the reduced channels , with full schmidt rank input state .
the rest follows from the previous corollary . in general
, the opposite implication does not hold .
that is because if we limit the problem to some subspace @xmath193 of the original hilbert space @xmath29 , then in general we do nt have a guarantee that the optimal input state will be supported on a subspace of the form @xmath194 , or in other words we would have to maximize the average success probability over all choices of the subspace @xmath193 .
we demonstrate this by the following simple example .
let @xmath195 , where @xmath196 and let @xmath197 , where @xmath198 is the optimal povm for discrimination of the ensemble @xmath199 .
by we have @xmath200 and @xmath201 .
it is easy to see that both conditions and are satisfied , but as argued in @xcite , there are cases when entangled input states give strictly larger probability of success than any separable state , so that a scheme of the form @xmath149 can not be optimal .
it seems that optimality of input states strongly depends on the structure of the channels . in some cases it is even necessary to use an input state with lower schmidt rank , because using maximal schmidt rank input state would `` waste '' some normalization of the input state on parts of the channels where it is unnecessary , as will be demonstrated in example [ ex : spm ] .
it is an open question whether some stronger conditions for general input states can be obtained .
see also @xcite for a discussion of a similar problem in the case of qubit pauli channels .
we will next present an upper bound for @xmath202 in the case that condition is violated .
we assume that the input state @xmath18 is maximally entangled , but a similar bound can be obtained for any input state having a maximal schmidt rank .
[ thm : errorestimate ] let @xmath5 be an optimal povm for discrimination of the ensemble @xmath203 and let @xmath204 , @xmath205 .
let @xmath206 denote the operator norm .
then the optimal success probability @xmath202 satisfies @xmath207 note that @xmath208 is the largest success probability that can be obtained by the maximally entangled input state , this implies the first inequality .
further , note that we have @xmath209 by optimality of the povm @xmath5 .
if now @xmath121 and @xmath37 are such that @xmath210 , then @xmath211,@xmath0 correspond to a dual feasible plan , hence @xmath212 by . to obtain the tightest upper bound in this way , we put @xmath213 by the choi isomorphism , there is some completely positive map @xmath214 , such that @xmath215 .
as it was shown in @xcite ( see corollary 2 and section 3.1 ) , @xmath216 , where the diamond norm is defined as @xmath217 moreover , since @xmath218 is completely positive , this norm simplifies to @xmath219\psi = \lvert\ptr_\ka z \rvert.\end{aligned}\ ] ] in general , the bound that we obtain in this way does not have to be meaningful , that is , it may happen that @xmath220 .
but , as will be demonstrated by the examples in the last section , there are cases when the bound is meaningful , or even tight .
[ rem : epsilon ] note that if @xmath221 , then @xmath222 and the value of @xmath223 indicates how much the condition ( ii ) is violated .
it is easy to see that @xmath224 , this shows that if @xmath225 is small , the maximally entangled state is close to optimal .
let @xmath12 and @xmath226 . the following notation will be used throughout .
let @xmath227 , then we put @xmath228 and @xmath229 let @xmath230 be the maximally entangled state and consider any two - outcome povm on @xmath231 , given by @xmath232 for some operator @xmath233 on @xmath231 .
the average success probability for the triple @xmath149 as defined by equation is : @xmath234 the optimal povm is obtained if @xmath5 is the projection onto the support of the positive part of @xmath235 . in this case , @xmath236 and @xmath237 [ coro : maxent ] an optimal measurement scheme @xmath149 with a pure maximally entangled input state @xmath18 exists if and only if the choi operators satisfy @xmath238 by the remarks below corollary [ coro : optimcond ] , such a scheme exists if and only if @xmath239 , equivalently , @xmath240 .
since we always have @xmath241 and @xmath242 the condition can be rewritten as stated .
the following corollary describes the upper bound of the optimal probability .
[ coro : bounds ] we have the following bounds @xmath243 if the condition is satisfied , the inequalities become equalities .
we only have to note that if the condition is satisfied , then @xmath244 .
[ rem : bounds ] it is well known that @xmath245 is related to the diamond norm as @xmath246 . to our knowledge , the only known bounds on the diamond norm in terms of the choi matrices are the following @xmath247 ( see e.g. ( * ? ? ?
* lemma 6 ) ) which is quite coarse . as in remark
[ rem : epsilon ] , we obtain from corollary [ coro : bounds ] the following new upper bound : @xmath248 where @xmath249 .
this shows that if is nearly satisfied , the above bounds are quite precise . to show that the upper bound given by is better than the bound @xmath250 we will show that in general @xmath251 we have @xmath252 since @xmath253 is a state
this implies the above inequality .
we also see that this inequality is strict unless @xmath254 is of rank 1 .
we apply the results of the previous section to the problem of discrimination of covariant channels , unitary channels , qubit channels and measurements . in the case of covariant channels and unital qubit channels ,
similar results were obtained in @xcite for more general decision problems on families of quantum channels .
let @xmath255 denote the unitary group of @xmath29 .
for @xmath256 , let @xmath257 let @xmath258 be a group and let @xmath259 and @xmath260 be unitary representations .
assume that @xmath261 and @xmath262 are covariant channels , that is , @xmath263 irreducibility of @xmath264 plays a strong role , as we will see . in this case , the only non - zero projection that commutes with all @xmath265 is @xmath7 , see e.g. @xcite .
let @xmath266 be channels satisfying .
assume that the representation @xmath264 is irreducible .
then the condition is satisfied for any @xmath227 .
let @xmath267 denote the transpose of @xmath51 with respect to the fixed basis @xmath36 .
let @xmath227 and let @xmath268 be as in .
we will prove the proposition by showing that @xmath269 is invariant under @xmath270 and by the discussion above this implies that @xmath271 .
for every @xmath272 we have @xmath273 in case the representation @xmath51 is reducible , let us sketch an upper bound of the optimal probability . by the previous proof , we have @xmath274 , hence @xmath275 where @xmath276 and @xmath277 are projections onto the subspaces of the irreducible representation , orthogonal sum of which is @xmath51 .
let @xmath278 then @xmath279 and we have @xmath280 let @xmath281 and let us denote @xmath282 . let @xmath283 be the werner - holevo channel , where @xmath284 denotes the transpose map with respect to the canonical basis @xmath285 . then @xmath286 is a unitary channel , given by the unitary @xmath51 such that @xmath287 it can be easily checked that @xmath288 and @xmath289 . if @xmath290 is a linear map such that there is some @xmath291 , satisfying @xmath292 then @xmath293 where @xmath294 .
moreover , for a self - adjoint @xmath295 , @xmath296 if and only if @xmath297 .
let @xmath261 and @xmath262 be two qubit channels and let @xmath268 for @xmath227 as before . by the previous remarks ,
the condition is equivalent to @xmath298 .
we are now going to investigate this equality .
note that @xmath299 satisfies ( [ eq : tracepres ] ) with @xmath300 .
since @xmath286 is a unitary channel , we have @xmath301 using further properties of @xmath286 and @xmath299 , we get @xmath302^t.\end{aligned}\ ] ] the last equality follows from the fact that @xmath303 , where @xmath304 denotes transpose with respect to the product basis @xmath305 , and that @xmath306 for any @xmath307 .
thus we have proved : for a pair of qubit channels , the condition holds if and only if @xmath308 in particular , this is true if @xmath309 .
if both channels are unital , this holds for any @xmath211 , hence maximally entangled input state is optimal , as it was already observed in @xcite and in @xcite in the case of qubit pauli channels . if @xmath310 , then the condition is satisfied if @xmath311 , even if the channels are not unital .
let @xmath312 and let @xmath313 , @xmath314 be the corresponding unitary channels . as it was proved in @xcite ,
it is not necessary to use entangled inputs for optimal discrimination of two unitary channels .
nevertheless , it is an interesting question whether a maximally entangled state is also optimal , this will be addressed in this paragraph . let @xmath315 .
since any input state @xmath18 may be replaced by @xmath316 , it is clear that discrimination of @xmath317 and @xmath318 is equivalent to discrimination of @xmath319 and the identity channel , and that a maximally entangled input state is optimal for one problem if and only if it is optimal for the other .
we may therefore assume that @xmath320 and @xmath321 .
since the unitaries are given only up to a phase , we may also assume that @xmath322 .
put @xmath323 , @xmath324 are the choi matrices of the unitary channel @xmath319 and identity . by the results of the appendix
it is clear that @xmath325 if and only if @xmath326 , where @xmath327 and @xmath328 since the transpose is a linear map and @xmath329 , we see that is equivalent to @xmath330 if @xmath331 this condition is obviously satisfied .
if @xmath332 it is equivalent to @xmath333 the unitary @xmath334 has a spectral decomposition @xmath335 where @xmath336 $ ] . from the condition
we see that @xmath337 must be constant with respect to @xmath338 , or in other words there must exist @xmath339 $ ] and numbers @xmath340 such that @xmath341 for every @xmath338 . by the assumption
@xmath322 we must have @xmath342 this implies that either @xmath343 and @xmath344 , or @xmath334 has exactly two eigenvalues , each of the same multiplicity .
the fact that in our calculation the eigenvalues are complex conjugate of each other is simply caused by the choice @xmath322 and does not have to be generally required .
we have proved the following : [ prop : unit - chan ] let @xmath313 , @xmath314 be unitary channels .
put @xmath345 and let @xmath227 . then holds if and only if either @xmath346 or @xmath334 has at most two different eigenvalues , each of the same multiplicity .
note that if @xmath347 is odd , mei holds iff @xmath348 , in which case the two channels are perfectly distinguishable .
a special case of a channel is a measurement , which is given by a povm @xmath349 .
one shot discrimination of quantum measurements was investigated in @xcite , where it was proved that entangled input states are necessary in some cases .
the corresponding channel @xmath350 is defined as @xmath351 and the choi matrix has the form @xmath352 .
let @xmath353 , @xmath354 for two povm s @xmath5 , @xmath355 with @xmath356 outcomes . in this case , the condition mei has the form @xmath357 we will further investigate simple projective measurements .
let @xmath358 and @xmath359 be two orthonormal bases in @xmath29 an let @xmath360 and @xmath361 .
we will also assume that @xmath362 .
assume that @xmath363 for some @xmath338 .
then the condition mei is satisfied if and only if @xmath364 .
in this case , the condition is @xmath365 since @xmath366 for all @xmath367 , this can be true
only if @xmath368 , that is , @xmath364 . from now on we will always assume that @xmath369 , equivalently , @xmath370 , for all @xmath53 .
then @xmath371 , where @xmath372 and @xmath373 is the projection onto @xmath374 , so that the condition mei becomes @xmath375 note that if @xmath376 , then @xmath377 , so that the condition trivially holds . in this case ,
@xmath378 and @xmath379 are unital qubit channels , so that this follows also by previous results .
put @xmath380 then @xmath381 .
the condition ( [ eq : simplemea ] ) is equivalent to @xmath382 for all @xmath338 , or @xmath383 for all @xmath384 .
note that the diagonal matrix @xmath385 , where @xmath386 , is invertible .
indeed , @xmath387 for some @xmath338 implies that @xmath388 so that @xmath389 for all @xmath367 .
but then @xmath390 , which is a contradiction .
hence @xmath391 for all @xmath338 and @xmath385 is positive definite .
let us begin from .
let us denote @xmath392 .
clearly @xmath334 is a unitary matrix .
it is straightforward to see , that @xmath393 so the condition becomes @xmath394 which can be written as a matrix equation of the form @xmath395 at this point we are ready to settle the case @xmath396 .
let @xmath397 , then the condition holds if and only if the matrix @xmath334 , defined as above , has of one of the following forms @xmath398 for @xmath399 . in other words ,
one povm is a cyclic permutation of the other .
consider the equation and let @xmath400 , then we obtain @xmath401 for @xmath402 . by putting @xmath403
, we obtain @xmath404 .
it follows that some off - diagonal elements of the matrix @xmath334 must be zero . on the other hand , for @xmath405
the equation becomes @xmath406 which shows that some off - diagonal elements must be non - zero .
putting , say , @xmath407 , the above equalities imply that also @xmath408 and all other off - diagonal elements are nonzero . from the condition @xmath409
, we obtain that @xmath410 and @xmath411 , this implies that @xmath334 is of the form @xmath412 and the basis @xmath413 and @xmath414 are just cyclic permutations of each other , modulo phase change .
similarly , assuming that @xmath415 , we obtain that @xmath334 is of the form @xmath416 , which is just the other possible cyclic permutation of the basis , modulo complex phase . for the converse , it is easy to check that both @xmath412 and @xmath416 satisfy .
see also the remark following the proof of proposition [ prop : const ] .
the basis vectors @xmath358 and @xmath417 are given only up to a phase , note that also the condition is invariant under phase changes .
we may therefore assume that @xmath418 for all @xmath53 .
then the matrix @xmath419 has only real non - negative entries , hence it is self - adjoint . from @xmath420 we can express @xmath421 and hence @xmath422 . since both @xmath78 and @xmath419 are self - adjoint matrices , their functions are well defined by the spectral theorem . multiplying out the right hand side of the equation and using the the obtained expression for @xmath419
we obtain : @xmath423 clearly the operators @xmath424 and @xmath425 are self - adjoint .
let us denote @xmath426 the real linear space of self - adjoint operators on @xmath427 , endowed with a hilbert - schmidt scalar product , which we denote as @xmath428 .
let us choose an orthonormal basis on @xmath429 composed of the operator @xmath430 and of operators @xmath113 , such that @xmath431 .
from equation we can see that : @xmath432 where @xmath433 is some real linear combination of @xmath113 , hence @xmath434 and @xmath435 .
it is easy to compute @xmath436 as follows : @xmath437 now it is easy to see , that : @xmath438 because from equation we see that @xmath439 as @xmath440 .
we have proved the following : let @xmath441 and @xmath442 for some orthonormal bases @xmath413 and @xmath414 . then condition is equivalent to the equations , .
the obtained equations look rather complicated , but they yield some results in specific cases .
[ prop : const ] with the above notations , assume that @xmath443 for all @xmath53
. then 1 .
if @xmath444 and @xmath347 is odd , then can not be satisfied .
if @xmath444 and @xmath347 is even , then is satisfied if and only if @xmath445 , where @xmath258 is a symmetric unitary matrix with zero diagonal .
3 . if @xmath446 , then always holds .
let s assume that @xmath447 for all @xmath53 .
immediately we see , that @xmath448 and @xmath449 . as a direct consequence of this we obtain @xmath450 from the equation . using this to our advantage in the equation
we obtain that in this case , is equivalent to @xmath451 as in the proof of proposition [ prop : unit - chan ] , we see that @xmath334 has exactly two eigenvalues with the same multiplicity , this also implies ( i ) . if @xmath347 is even , it is clear that these eigenvalues must be equal to @xmath452 .
let @xmath189 be the eigenprojection corresponding to @xmath453 , then @xmath454 , with @xmath455 .
conversely , it is easy to see that if @xmath334 is of this form , then holds , this finishes the proof of ( ii )
. now assume that @xmath456 , which means that @xmath457 for all @xmath53 .
again we immediately see that conditions and are satisfied , with @xmath450 .
this implies ( iii ) .
note that we can conclude from the last statement that any two bases , such that the unitary matrix mapping one basis to the other is hollow satisfy the condition .
especially , this happens if the basis @xmath458 is a permutation of @xmath459 , leaving no element fixed .
a particular case of the situation described in proposition [ prop : const ] is when the bases are mutually unbiased , that is when @xmath460 for all @xmath461 .
then @xmath462 is a hadamard matrix , see e.g. @xcite for more information on complex hadamard matrices .
as we have seen , such bases can satisfy the condition only in even - dimensional hilbert spaces .
below , we provide a further result for @xmath463 .
recall that two hadamard matrices are equivalent if one can be turned into the other by multiplication by diagonal unitaries and permutations . up to equivalence , any hadamard matrix can be turned into a dephased form , with all elements in the first row and column equal to unity . in this way ,
any four - dimensional hadamard matrix is equivalent to a member of a one - parameter family , containing a unique matrix @xmath464 with real entries . since
the vectors @xmath413 and @xmath465 are given only up to a phase , the next result shows that this is the only case when is satisfied .
let @xmath466 , and let the bases @xmath467 and @xmath414 be mutually unbiased .
then the condition is satisfied if and only if the corresponding hadamard matrix satisfies @xmath468 , where @xmath469 , @xmath470 are diagonal unitaries and @xmath471 it is clear that @xmath472 . by proposition [ prop : const ] , we see that we need to search for a @xmath473 hollow symmetric unitary matrix @xmath258 such that each off - diagonal element has modulus @xmath474 . a general form for such a matrix can be found by a straightforward computation .
we find that @xmath475 for @xmath476 , so that @xmath477 , where @xmath478 and @xmath479 .
conversely , suppose that @xmath468 for diagonal unitaries @xmath469 and @xmath470 .
since we also assume that all diagonal elements of @xmath334 are equal to 1 , we obtain that we must have @xmath480 and @xmath481 .
it is now easy to check that @xmath482 has the form , with @xmath483 , @xmath484 and @xmath485 .
in this section , we present examples based on the results of the previous sections . to underline how maximally entangled input states may be used for discrimination of two qubit channels , we will present an example of discrimination of identity channel @xmath486 and amplitude damping channel @xmath487 .
the amplitude damping channels are not unital and hence is not necessarily satisfied , and we will see that the maximally entangled state is indeed not optimal . the identity channel may be replaced by a unitary channel with some changes to the following calculations .
we will set @xmath488 . , optimal success probability @xmath202 and the upper bound @xmath489 given by thm .
[ thm : errorestimate ] on the parameter @xmath490 in discrimination of the amplitude damping channel @xmath487 and the identity channel ( example [ ex : qubit ] ) . ]
[ ex : qubit ] let @xmath29 denote complex hilbert space with @xmath376 and @xmath285 will denote some orthonormal basis of @xmath29 .
the amplitude damping channel is represented by kraus operators @xmath491 , @xmath492 , defined as @xmath493 where @xmath494 $ ] is a parameter .
note that for @xmath495 the amplitude damping channel becomes the identity .
we will proceed as follows : we will find the choi matrices corresponding to @xmath486 and @xmath487 to obtain @xmath208 and the upper bound on the optimal success probability .
then we will find @xmath202 to compare it with the upper bound .
let @xmath496 and @xmath497 .
we will be interested in the matrix @xmath498 which is of the form @xmath499 since @xmath500 we can already find @xmath208 as a function of @xmath490 .
it is easy to see that @xmath501 and @xmath502 are eigenvalues of @xmath503 .
finding the other two eigenvalues is easy since the problem reduces to finding eigenvalues of a @xmath504 matrix , they are @xmath505 and @xmath506 .
let us denote @xmath507 then we have @xmath508 .
moreover after a tedious calculation it can be seen that @xmath509 and we have @xmath510 . since @xmath511 and @xmath512 we have @xmath513 to verify the upper bound will find @xmath202 , for which we only need to find the value of @xmath514 , which we have done by numeric methods . according to @xcite the problem of finding a diamond norm can be formulated as following sdp problem : @xmath515 where @xmath516 , @xmath517 .
we used matlab and the package cvx for solving the convex program , @xcite .
we have computed the numerical values for @xmath518 values of the parameter @xmath490 , homogeneously distributed on the interval @xmath519 $ ] . between these points
straights lines were drawn , hence the figure looks like a continuous line .
the resulting expressions of @xmath208 , @xmath489 and numerical data of @xmath202 as functions of @xmath490 are plotted in the fig .
[ fig : prob - theta ] . even though it shows that maximally entangled input state is not optimal , notice that the upper bound is close to @xmath245 for small values of the parameter @xmath490 . , optimal success probability @xmath202 and the upper bound @xmath489 given by thm .
[ thm : errorestimate ] on the parameter @xmath218 in discrimination of the unitary channel and the identity ( example [ ex : unitary ] ) . ]
[ ex : unitary ] let @xmath520 and @xmath397 . without loss of generality
, we may always assume that we discriminate the identity channel against a unitary @xmath319 .
as our first example , we consider the unitary matrix @xmath521 at this point , it is easy to compute @xmath208 and @xmath489 , as we can express them as functions of @xmath412 , which is only dependent on @xmath218 , so the bounds are functions of @xmath218 .
moreover it is possible to compute @xmath202 by numerical methods described in example [ ex : qubit ] .
the expressions for @xmath208 and @xmath489 are long and messy , they are plotted in the fig .
[ fig : prob - xi ] as well as the computed values of @xmath202 . from the figure it is once again clear that on one hand as @xmath208 rises the the upper bound becomes meaningless , but on the other hand if @xmath208 is small then the upper bound and @xmath202 is small as well .
another case we consider is matrix of the form @xmath522 again , it is straightforward to obtain @xmath208 and @xmath489 and to compute @xmath202 numerically , the obtained functions are plotted in fig . [
fig : prob - xi-2 ] .
again it can be nicely seen that as @xmath208 rises towards @xmath523 the upper bound becomes meaningless . , optimal success probability @xmath202 and the upper bound @xmath489 given by thm .
[ thm : errorestimate ] on the parameter @xmath218 in discrimination of the unitary channel and the identity ( example [ ex : unitary ] ) . ]
let us demonstrate our results once more , this time for simple projective measurements .
[ ex : spm ] let @xmath310 , @xmath524 , @xmath525 , @xmath526 , @xmath527 for @xmath528 .
according to our previous results the maximally entangled input state is not optimal for discriminating these two measurements .
as first we need to find @xmath529 .
we get @xmath530 .
the operator @xmath531 is diagonal in the orthonormal basis @xmath532 , @xmath533 , where @xmath534 the following calculation is straightforward .
we obtain @xmath535 now it is easy to see that @xmath536 i.e. the bound is meaningful and the same for all @xmath347 , even though @xmath208 tends to @xmath537 from above in the formal limit @xmath538 . to underline the correctness of the upper bound we will find the optimal state for discrimination of the channels
let us denote @xmath539 , @xmath540 the channel corresponding to simple projective measurement corresponding to the set of projectors @xmath541 , @xmath542 respectively and denote @xmath543 .
notice that the channels @xmath539 and @xmath540 can be separated as follows @xmath544 where @xmath18 is a state on @xmath29 and @xmath545 , @xmath546 and @xmath547 .
notice that @xmath548 where @xmath549 .
this leads to the following @xmath550 in other words , it is sufficient to consider only input states of the form @xmath551 , where @xmath182 and @xmath552 are states , such that @xmath553 , @xmath554 and @xmath555 .
since @xmath556 where @xmath5 , @xmath557 is the povm we use for discrimination of the channels , it is obvious that for @xmath202 to be maximal we also have to set @xmath558 and according to corollary [ coro : notfullrankstate ] the problem reduces to discriminating the channels @xmath559 and @xmath560 .
both @xmath559 and @xmath560 are unital qubit channels and by our previous results the optimal input state @xmath561 is @xmath562 in the product basis given by @xmath563 , @xmath564 , we have @xmath565 this matrix has two eigenvalues @xmath566 and @xmath567 each with multiplicity @xmath568 .
we get @xmath569 and @xmath570 which is exactly the same as our upper bound .
we presented necessary and sufficient conditions for optimality of a process povm in channel discrimination , especially for a process povm corresponding to a measurement scheme with full schmidt rank input state .
in particular , a necessary and sufficient condition for existence of an optimal measurement scheme with a given full schmidt rank input state were found . in the case of maximally
entangled input states , an upper bound of the optimal success probability was given if the optimality condition is not satisfied .
for discrimination of two channels , we obtained a simple condition in terms of the choi matrices of the channels and a new upper bound on the diamond norm .
the results were applied to discrimination of four types of channels . for covariant channels , known results for the irreducible case were extended to some reducible cases and an upper bound on the optimal success probability was found . for qubit channels , the obtained condition generalizes previously known results to some pairs of non - unital channels . we proved that for discrimination of unitary channels , maximally entangled input states are optimal only in some very special cases . to our best knowledge , the results obtained for unitary channels and simple projective measurements are new .
an interesting open question is whether it is possible to obtain a similar condition for input states with lower schmidt rank . as it was shown , there are cases when such input states are optimal and there may even be no optimal full rank input states .
it is not only of question what the schmidt rank of the optimal input state may be but also how to select the subspace of the input hilbert space that will form the support of the partial trace of the input state .
another possible future directions of research is to investigate optimal discrimination of more complex quantum processes .
we thank the anonymous referee for the valuable feedback and comments that made this paper more readable and better organized and also for pointing out an easier proof of prop . [
prop : unit - chan ] .
this research was supported by grant vega 2/0069/16 .
s. m. barnett and s. croke , quantum state discrimination , adv .
photonics , vol . 1 , no . 2 , pp .
238 - 278 ( 2009 ) a. o. barut , r. rczka , _ theory of group represetations and applications _
, warszava : pwn - polish scientific publishers , 1980 a. barvinok , _ a course in convexity _ , ams , 2002 f.g.s.l .
brandao , m. piani , p. horodecki , nature communications 6 , art .
7908 ( 2015 ) g. m. dariano , p.lo presti , m. g. a. paris , phys .
87 , 270404 ( 2001 ) g. m. dariano , p.lo presti , m. g. a. paris , j. opt .
b 4 , 273 ( 2002 ) g. chiribella , optimal networks for quantum metrology : semidefinite programs and product rules , new j. phys . , vol .
14 , pp . 1 - 19 , ( 2012 ) g. chiribella , g. m. dariano , and p. perinotti , phys .
lett . 101 , 180501 ( 2008 ) m.d .
choi , completely positive maps on complex matrices , lin .
285 - 290 ( 1975 ) cvx research , inc .
cvx : matlab software for disciplined convex programming , version 2.1 .
http://cvxr.com/cvx , june 2015 .
m. grant , s. boyd , graph implementations for nonsmooth convex programs , recent advances in learning and control ( a tribute to m. vidyasagar ) , v. blondel , s. boyd , and h. kimura , editors , pp .
95 - 110 , lecture notes in control and information sciences , springer , 2008 .
gus gutoski , j. math .
3 , ( 2012 ) gus gutoski and john watrous , toward a general theory of quantum games . in _ proceedings stoc 2007 _ , pp .
565574 c. w. helstrom , _ quantum detection and estimation theory _ , academic press , 1976 a. s. holevo , journal of multivariate analysis 3 , pp .
337394 ( 1973 ) a. jenov , j math .
55 , 022201 ( 2014 ) a. kitaev , quantum computations : algorithms and error correction .
russian mathematical surveys , 52 , pp .
11911249 ( 1997 ) k. matsumoto , arxiv:1209.2392 m. piani and j. watrous , phys .
lett . , 102 ( 25 ) : 250501 , ( 2009 ) m. f. sacchi , optimal discrimination of quantum operations , phys . rev . a 71 , 062340 ( 2005 ) ; m. f. sacchi , entanglement can enhance the distinguishability of entanglement - breaking channels , phys .
rev . a - at .
1 , pp . 12 ( 2005 ) m. f. sacchi , minimum error discrimination of pauli channels , j. opt . b - quantum semiclassical opt .
s333s336 ( 2005 ) m. sedlk , m. ziman , phys .
a 90 , 052312 ( 2014 ) w. tadej and k. zyczkowski , open systems and information dynamics , vol .
133 - 177 ( 2006 ) j. watrous , chicago journal of theoretical computer science , vol .
2013 , pp . 1 - 19
( 2013 ) h. yuen , r. kennedy and m. lax , ieee transactions on information theory , it-21 ( 1975 ) m. ziman , phys .
a 77 , 062112 ( 2008 )
let @xmath571 be unit vectors , @xmath572 , @xmath573 .
let @xmath227 and @xmath574 then @xmath575 where @xmath576 are the eigenvalues and @xmath577 the corresponding eigenvectors of @xmath578 . moreover , we have @xmath579 and since we do not worry about a phase
, we may suppose that @xmath580 . from @xmath581
, we obtain @xmath582 where @xmath583 .
it follows that @xmath584 , so that @xmath585 where @xmath586 .
we obtain @xmath587 | we find optimality conditions for testers in discrimination of quantum channels .
these conditions are obtained using semidefinite programming and are similar to optimality conditions for discrimination of quantum states .
we get a simple condition for existence of an optimal tester with any given input state with maximal schmidt rank , in particular with a maximally entangled input state . in case when maximally entangled state is not optimal an upper bound on the optimal success probability is obtained .
the results for discrimination of two channels are applied to covariant channels , qubit channels , unitary channels and simple projective measurements . |
Subsets and Splits